Pediatric Orthopedic Deformities: Basic science, Diagnosis, and Treatment

Preface

Pediatric Orthopedic Deformities: Basic Science, Diagnosis, and Treatment provides a detailed understanding of major

ments used, and results of treatment, stressing relationships with the underlying pathobiology at each step of the way. Chapter 3 describes developmental dysplasia of the hip; Chapter 4, Legg-Calve-Perthes disease; Chapter 5, coxa vara in developmental and acquired abnormalities of the femur within which slipped capital femoral epiphysis, proximal femoral focal deficiency, infantile coxa vara, coxa vara with congenital short femur, and coxa vara with the skeletal dysplasias are discussed; and Chapter 6 reviews epiphyseal disorders of the knee encompassing the distal femur, proximal tibia, and proximal fibula. Care is taken throughout the book to review in detail the studies done so as to place the descriptions on as rigid a scientific basis as possible. Part III discusses entities crucial to an understanding of the more complicated growth-related deformities in pediatric orthopedics. These disorders produce deformities by affecting primarily the epiphyses and metaphyses. In Chapter 7, epiphyseal growth plate fracture-separations are reviewed, because this subset of childhood fractures has the potential to contribute in the most major way to limb deformity. Understanding of these fracture-separations as it has evolved over several decades is reviewed. The underlying cell biology and histopathology of growth plate injuries are reviewed such that the various classifications and the presence or absence of negative growth sequelae can be understood. The various pathoanatomic classifications are reviewed in detail, as is the pathophysiologic approach, which is dependent to a great extent on understanding the histopathology and assessing it by newer imaging modalities. Treatments specific for each of the fracture patterns at each epiphysis are reviewed. The importance of understanding the structure and blood supply of epiphysis and adjacent metaphyis, the specific fracture patterns that occur, the biologic rationale for the treatments now used and being developed, and the newer investigative technologies involving CT scanning and MR imaging are stressed. In Chapter 8 the entire entity of lower extremity length discrepancy is described in detail, including the natural history of the specific disorders that cause the discrepancies, the negative sequelae of length discrepancies, methods to project the eventual discrepancies at skeletal maturity, a review of the developmental patterns likely to occur in each of the

pediatric orthopedic deformities showing how normal developmental bone biology and abnormal pathobiology relate to the occurrence of these deformities, their diagnosis, and their treatments. An understanding of normal developmental bone biology as outlined by several research disciplines is vital to an understanding of abnormal bone development. Many orthopedic deformities of childhood worsen with growth, while others have the potential for correction either spontaneously or with appropriate therapeutic interventions over the years remaining until skeletal maturity. Management of pediatric orthopedic deformities encompasses the need to understand biologic and mechanical contributions to skeletal development. Part I of this book provides basic information on the developing skeleton and current imaging methods used to assess it. Chapter 1 describes developmental bone biology as outF,ned by several investigational disciplines. These include histology at the light and electron microscopic levels; molecular biology outlining the wide array of genetic and molecu!ar controls for skeletal tissue differentiation, growth, and synthesis of structural molecules; mechanical-biophysical effects on the developing skeleton; and basic radiologic parameters of growth such as appearance of secondary ossifi,:ation centers and times of physeal fusion. Chapter 2 has been written by Dr. Diego Jaramillo, a col":eague of several years, who outlines the rationale for diagnosis of normal and abnormal skeletal development by many imaging techniques including plain radiographs, ultrasonoglaphy, bone scanning, computerized tomography, and magnetic resonance imaging. Part II discusses disorders of the developing hip and knee. Each chapter begins with a clear description of the terminology used for the disorder discussed and then provides an outline of the basic biology relevant to the entity. For each entity, detailed review of the clinical findings, diagnostic techniques, associated surgical and nonsurgical treatments, and eventual results are provided. The literature for each entity has been presented both critically and in detail in relation to a wide spectrum of findings, including pathoanatomy, clinical presentation, diagnostic techniques, various treat-

xvii

xviii

Preface

various disorders, and extensive overviews into shortening, lengthening, and transphyseal bone bridge resection treatments. In Chapter 9 the skeletal dysplasias are discussed, including clinical and radiographic descriptions of the entities, the molecular abnormalities recently identified, the specific orthopedic deformities, and the specific orthopedic treatments. The chapter is designed to address and correct the fragmentation of presentation of the skeletal dysplasias in the literature, in which the disorder tends to be discussed only from the viewpoint of each disciplinemradiology, molecular biology, or orthopedics. Chapter 10 reviews entities known to be associated with pediatric orthopedic deformities but also discussed in general in isolated fashion. The chapter presents these entities from

the point of view of the basic underlying biology, the pathoanatomy of deformation, and the principles and results of medical and orthopedic management. The entities described focus on the epiphyseal and metaphyseal effects of involvement with rickets of the various types, juvenile rheumatoid arthritis, benign and malignant neoplasms of the epiphysealmetaphyseal regions, pyogenic and tuberculous infections that affect primarily the epiphyseal-metaphyseal regions, and major hematologic disorders including the hemophilias. The book has two main premises: (1) current orthopedic treatments of growth deformities of the developing skeleton are most effective if based on understanding and relating to the pathobiology and (2) future treatments can be best developed if based on the underlying primary and secondary pathobiology.

Frederic Shapiro

Acknowledgments

and revising the manuscript, tables, and references; Emily Flynn Mclntosh of the Children's Hospital for artwork; James Koepfler of the Children's Hospital for medical photography; George Malatantis of the Children's Hospital for histology preparations and printing; and Taft Paschall, Joanna Dinsmore, Judy Meyer, and the entire Academic Press team for production of the book.

The author gratefully acknowledges the extensive efforts of several individuals who helped bring this work to publication: Dr. Diego Jaramillo of the Massachusetts General Hospital, Boston, for contributing Chapter 2 and MR imaging studies throughout; Dr. Frank Rand of the New England Baptist Hospital, Boston, for collaborative work in Chapters 7 and 8; Mary Doherty and Joanne Hutchinson for typing

Frederic Shapiro

xix

CHAPTER

1

Developmental Bone Biology I. II. III. IV.

Terminology Early Scientific Understandings of Bone Growth

Embryology of the Limbs Bone Development at the Light Microscopic Level Following Delineation of the Cell Theory and Advances in Microscopy and Histochemistry V. More Detailed Histologic Studies of Bone Formation VI. Fate of the Hypertrophic Chondrocyte as Interpreted from Light Microscopic Studies VII. Structural Development of the Epiphyseal Regions Including the Joints, the Metaphyses, and the Diaphyses

I. T E R M I N O L O G Y A. Overview During the embryonic period, limb buds filled with undifferentiated mesenchymal cells form from the lateral side walls. The earliest sites of skeletal formation are characterized by condensation or close packing of mesenchymal cells followed by early cartilage differentiation. Each of the long bones is preformed as a cartilage model. Bone tissue is then deposited beginning in the middle of the model on the calcified cartilaginous core using the endochondral ossification mechanism and directly by the surrounding periosteum using the intramembranous ossification mechanism. Bone deposition progresses toward each end with intramembranous bone formation at the periphery slightly in advance spatially of internally or centrally positioned endochondral bone formation. The proportional involvement of each of the three regions of a developing bone--diaphysis, metaphysis, and epiphysis--is established by the early fetal stage and remains more or less unchanged until skeletal maturity. The central part of the bone is the diaphysis or shaft; the furthermost bone extension of the diaphysis is the metaphysis; and the developing cartilaginous end of each bone is the epiphysis. Normal bone development occurs in conjunction with the proliferation and differentiation of cells, the synthesis and interaction of specific molecules, and the generation of intrinsic and extrinsic biophysical forces. Primary genetic blueprints and secondary epigenetic and inductive phenomena lead to the patterns characteristic of each bone through-

VIII. IX. X. XI.

Axes along Which Bones Are Patterned Gene and Molecular Controls of Limb Development Chemistry of the Extracellular Matrix Mineralization

XH. Epiphyseai Growth XIII. Responses of Developing Bones and Epiphyses to Mechanical Stresses XIV. Radiographic Characteristics in Development of Major Long Bone Epiphyses XV. Why Epiphyses Form and the Evolution of Epiphyses

out the skeleton. The epiphyses are responsible for long bone longitudinal growth, transverse growth at the ends of the bone, and the shape of the articular surfaces. In this book we will outline established, newly evolving, and theoretical information on normal bone development and abnormal development as it relates to the many skeletal growth disorders of childhood with their particular concentration at the epiphyseal regions.

B. Theories of Embryogenesis--Preformationism and Epigenesis Prior to the nineteenth century the most widely accepted theory of embryogenesis was preformationism, the doctrine that the entire adult individual was present in miniature in the egg or sperm and development simply involved an increase in size. With more detailed observation however, the concept of development by epigenesis became accepted more widely. Epigenesis refers to the sequential development of morphological complexity in the embryo by the gradual and progressive differentiation of homogenous material; each stage in development is considered to be dependent on and directed by the stage immediately preceding it. The works of Caspar Friedreich Wolff (373), Pander (303), and Karl Ernst von Baer (13) were particularly noteworthy in establishing the scientific validity of development by epigenesis. Wolff, in 1759, reported his observations on hen's egg development, which showed blood vessels appearing where none existed previously and intestine forming from a flat plate. He concluded that epigenesis was real:

CHAPTER 1 ~

Developmental Bone Bioloyy

"each part is first of all an effect of the preceding parts, and itself becomes the cause of the following part." The structural basis of embryogenesis was revealed more clearly by Pander, who defined the three germ layers in 1817, and then by von Baer (13), who is widely considered the founder of embryology. Von Baer performed elaborate descriptions of chick embryo development and its similarity to the development of several other vertebrate types and reported them in 1828 and 1837 (13). He recognized the significance of the three germ layers in development in all vertebrates and the truly epigenetic mechanism of development from the general to the specific. He recognized that the general path of differentiation proceeds in three sequential stages: the primary formation of the germ layer, histological differentiation of cell and tissue types within the germ layers, and morphological differentiation to early organ formation. The mode of development is from simple to complex and from undifferentiated cell masses to new organs9 The law of biological development is progressive differentiation of homogeneous, coarsely structured, and general to heterogeneous, finely structured, and specific. It is now recognized that much of development occurs in a self-assembly or "automatic" mode based on chemical and biophysical phenomena. Development is epigenetic in which only the early prepatterns are rigidly determined, after which each subsequent step is a combination of gene synthesis and automatic self-assembly based on the physical pressure of certain molecules9 Genes give approximate direction only. The detailed structure of multicellular organisms occurs on the basis of many intermediate levels of interaction each with its own immediate properties such that "the edifice is virtually entirely epigenetic."

C. Epiphysis The term epiphysis refers to the entire developing end of the bone (320, 321). This region initially is formed completely in cartilage and subsequently subdivides during development into three histologically distinct regions: (1) the cartilage immediately adjacent to the joint, referred to as articular cartilage; (2) the cartilage adjacent to the metaphysis, referred to variously as the growth plate, the epiphyseal growth plate, or the physis (it is the functionally and cytologically specialized region where the bulk of longitudinal growth occurs, and it encompasses the area from the reserve zone of cells to the end of the hypertrophic cell layer); and (3) the cartilage between the articular cartilage and the growth plate cartilage, referred to as the epiphyseal cartilage. Eventually this is transformed entirely into bone and marrow following the appearance and enlargement of what is variously referred to as the secondary ossification center, the bony nucleus, the ossific nucleus, or the bony epiphysis (Fig. 1). The epiphysis is sometimes referred to as the chondroepiphysis, but use of this term should be restricted to the time prior to formation of the secondary ossification center.

9

--..

9 .

r .

-',-

,,'.

9

"

p

9

.

E GP

M

FIGURE 1 The histologic structure of the epiphysis of the proximal tibia is illustrated. The entire developing end of the bone from the articular cartilage surface to the last cells of the hypertrophic zone of the growth plate is the epiphysis (E). This encompasses three regions that are initially cartilage: (1) the articular cartilage (AC), (2) the growth plate (GP), also referred to as the epiphysealgrowth plate or the physis, and (3) the epiphyseal cartilage (EC), which refers to the cartilage mass between the articular cartilage and the growthplate cartilage. It is within the epiphysealcartilagethat the secondaryossificationcenter (SOC), also referred to as the bony nucleus, the ossific nucleus, or the bony epiphysis, forms and expands. [Reprinted from Shapiro (1987), New Engl. J. Med. 317: 1702-1710, with permission. Copyright 9 1987 MassachusettsMedical Society.All rights reserved.]

D. Endochondral Ossification The cartilage models of a developing bone and ultimately the epiphyses form bone by a mechanism referred to as endochondral bone formation. The major characteristics of this mechanism of bone formation involve growth of the cartilage by interstitial expansion involving chondrocyte proliferation, matrix formation, and chondrocyte hypertrophy. At a certain stage of development, the cartilage matrix adjacent to the hypertrophic cells mineralizes, and there is vascular invasion of the lacunae in which the hypertrophic cells reside. This vascular invasion is accompanied by mesenchymal cells that shortly differentiate to osteoblasts and synthesize a bone matrix on the calcified cartilage cores. The calcified cartilage thus is serving as a scaffold on which bone is deposited initially at the center of the developing cartilage model of the bone and eventually at the lower regions of the physis merging into the metaphysis and also within the epiphyseal cartilage where the bone formed is referred to as the secondary ossification center. The epiphyseal cartilage immediately surrounding the secondary ossification center and undergoing chondrocytic hypertrophy can be considered as the physis of the secondary ossification center.

E. Intramembranous Ossification Long bone formation is also characterized by a mechanism referred to as intramembranous ossification, which occurs from the surrounding periosteum. In this mechanism, bone tissue is formed directly from mesenchymal cells without the mediation of a cartilage scaffold phase. The initial site of

SECTION II ~ Early Scientific Understandings o f Bone G r o w t h

F I G U R E 2 Photomicrograph of the developing end of a rabbit metatarsal is shown. The secondary ossification center is shown centrally at the top. EC, epiphyseal cartilage; P, physeal cartilage. The two closed white arrows mark the perichondrial ossification groove of Ranvier. The two open white arrows demonstrate the cortex, which is formed by the intramembranous ossification mechanism. The metaphyseal bone just below the physis is formed by the endochondral ossification mechanism as is the bone of the secondary ossification center. [Reprinted from Shapiro et al. (1977). J. Bone Joint Surg. 59A: 703-723, with permission.]

bone formation, referred to as the primary center of ossification, is at the periosteum surrounding the center of the diaphysis. The periosteum is composed of a specific structure. It has two layers: an outer fibrous layer and an inner osteogenic or cambial layer. The inner cambial layer also displays an organized cellular differentiation pattern, although it is not as structurally specific as the physis. The outermost part of the inner layer is composed of undifferentiated mesenchymal cells; these then begin to secrete and surround themselves with an osteoid matrix as they differentiate from preosteoblasts to osteoblasts. Further toward the cortex, osteoblasts line the surface of the bone to synthesize osteoid preferentially on the bone surface. When the osteoid matrix surrounds a cell completely and then becomes mineralized, that cell is referred to as an osteocyte.

F. Perichondrial Ossification Groove of Ranvier In all long and most flat bones both mechanisms of bone formation, endochondral and intramembranous, are present. They relate intimately and specifically to one another at the periphery of the growth plate in a region referred to as the perichondrial ossification groove of Ranvier (Fig. 2). The tissues comprising the intramembranous ossification mechanism circumferentially ensheathe and support the physeal cartilage at this area. A circumferential groove indenting the cartilage is present, whose deepest part is opposite the epiphyseal cartilage-physeal cartilage junction. It contains three tissue components: (1) an outer fibrous layer that is

5

continuous with the outer fibrous layer of the periosteum and inserts beyond the physeal region into the epiphyseal cartilage; (2) a zone of densely packed cells that is a continuation of the inner cambial layer of the periosteum and is present into the depths of the groove as far as the resting zone of physeal cartilage (this collection of dense cells synthesizes osteoid and intramembranous bone directly); and (3) a collection of loosely packed cells between the outermost reaches of the zone of dense cells and the fibrous tissue layer that adds chondrocytes to the periphery of the epiphysis just beyond the physis itself. The intramembranous bone synthesized by the tissues within the groove region is sometimes referred to as the bony bark of Lacroix. Frequently it is discontinuous with the cortical bone of the diaphysis and metaphysis in those areas where the metaphyseal cutback zone is extensive. The perichondrial ossification groove of Ranvier and its fibrous chondroprogenitor and osteoprogenitor cells are an integral part of the epiphyseal region (320).

II. E A R L Y S C I E N T I F I C U N D E R S T A N D I N G S

OF BONE GROWTH Clarification of the mechanisms of bone growth, representing early examples of the application of scientific method to biological phenomena, began to show definitive advances early in the eighteenth century (24, 74, 81, 88-92, 140, 169, 185,242) (Table I).

A. Hales and Belchier In 1727, Stephen Hales showed that the bones grew in length by the addition of new tissue at their ends rather than interstitially (140). He measured the leg bones of a young chicken 2 months after drilling two holes in the shaft to act as markers. The holes were no further apart although the bone itself had lengthened considerably. Bone growth studies were enhanced by John Belchier (24), a London surgeon who observed that the bones of growing pigs and fowl that had been fed on madder were colored red. The vital staining dye of the madder plant was subsequently recognized as alizarin (12). The alizarin red stain is still used today in whole mount embryo studies. Staining with methylene blue outlines cartilage structures whereas alizarin red stains the bone (322). Subsequent studies by Duhamel, Hunter, and Flourens further demonstrated that bones grow in length by continuous increments at the ends.

B. Nesbitt The formation of bone by the two now familiar endochondral and intramembranous mechanisms was noted early. Robert Nesbitt (1736) is recognized as being the first to describe the two methods of ossification in human fetal bone, one occurring directly in a membrane and the other occur-

CHAPTER 1 ~ Developmental Bone Bioloyy TABLE I .

1727 1731-1736

1736 1739-1743

1740 1815

1841-1847

1850 1852-1853

1858 1860

1864

.

Early O b s e r v a t i o n s on Bone G r o w t h .

.

.

Bones grow in length by the addition of new tissue at their ends (Hales). There are two mechanisms of bone formation, one occurring directly in a membrane (intramembranous ossification) and one via preexisting cartilage (endochondral ossification). Vascularization immediately precedes bone formation in cartilage (Nesbitt). Growing bone is stained red by madder (alizarin) in diet (Belchier). A long bone grows in length from its ends and in thickness by formation of new bone on its outer surface. The osteogenic function of the periosteum thickens the bone on its outer surface, based on madder feeding studies (Duhamel). Increased diameter of marrow cavity with growth is due to absorption of preexisting bone internally. Bone development includes both bone deposition and bone absorption (Hunter). Interplay of cartilage and bone formation in long bone development. Cartilage vascular canals in epiphyses. Gross and crudely magnified evidence of physeal cartilage. Vascularization of inner periosteum precedes earliest site of bone formation. Cartilage model determines the shape of the future bone and establishes ossification within it. Mechanical pressure variously modified is the principal agent in effecting progressive changes of structure in growing bone (Howship). Bone is formed in the periosteum, grows in thickness by superimposing new layers externally, grows in length by adding new layers at the growth cartilage at each end, and has the marrow cavity formed by resorption of the inner bone layers. There is formationresorption-reformation of the bone with growth accompanied by constant change in bone substance (Flourens). Resorption of bone as part of developmental sequence is mediated by osteoclasts (Kolliker). Tissue at epiphyseal-diaphyseal (metaphyseal) junction is characterized by differing layers as determined by careful gross and early lower power microscopic examination (Broca; Tomes and de Morgan). Detailed microscopic-histologic structural recognition of physeal layers and endochondral sequence (MUller). Bone description as a tissue (bone cells plus calcified matrix) and as an organ (encompassing bone tissue, marrow, periosteum, articular and epiphyseal cartilage, vessels, and nerves). Detailed cellular description of physeal sequence including hypertrophic chondrocyte fate (Virchow). Bone forming cells first referred to as osteoblasts (Gegenbaur).

fjring via preexisting cartilage (81,242). In his book Human Osteogeny, published in 1736, and based on lectures he gave in 1731 he indicated that he would "show the ancient and common notion of all bones being originally cartilaginous to be a vulgar error." He went on to indicate that there were two methods, or species in his terminology, of ossification. "The bony particles in the fetuses begin to be deposited or to shoot either between membranes or within cartilages." Nesbitt clearly noted that "the periosteum is a delicate fine and strong membrane which is spread on and covers not only all the bones in general but is also continued over the cartilages that have any connection with them; where from its situation it acquires the name of perichondrium." By formation of bone and membrane, therefore, he clearly was referring to periosteal new bone formation. He also recognized both the inner and outer layers, or strata as he referred to them, of the periosteum. Although bone formation from preexisting cartilage models had previously been appreciated, Nesbitt also showed that "some bones begin and continue to increase until they arrive at maturity without the least appearance of cartilage in or around them." The first

species of ossification, therefore, was intramembranous bone. "The texture of that species of ossification which is produced between membranes by a careful and proper examination may be seen to be of small particles so conjoined together as to form fine bony threads or fibres which are disposed differently according to the particular formation of each bone and its several parts. This is most visible in thin and broad bones, especially in some of those which form the cranium." He noted that "you may observe the bony particles to be gradually multiplied and so conjoined in contact as to produce the appearance of small fine bony threads or fibres which then appear a little like radii shooting from a centre." With time there was an increase in the number of bony fibers that became "pressed so close together to form a single lamina or plate of bone." Membrane bone was that type formed in the cranium and it was also seen in cylindrical bones. He defined well the periosteal intramembranous bone sequence by noting that "their ossifications begin while the circumference of the part is not larger than a small pin in the form of a broad flat ring which surrounds the internal periosteum and is surrounded by the external. As these tings

SECTION II 9 Early Scientific Understandings o f Bone G r o w t h

increase in breadth their fibres shoot toward both extremities of the part, not always in straight lines, but according to the particular figure the bone is designed by nature to be." He was aware that "the other species of ossification which first appears within a cartilage begins late." Bone formation occurred in close association with blood vessels. He described formation of the secondary ossification centers and noted that vascularization immediately preceded bone formation in the epiphyseal cartilage. "The first small corpuscles of bone which become visible are always in that part of the cartilage which has the greatest quantity of red fluid appearing in it and they are not always placed close together but often at small distances from each other." In a large series of accompanying diagrams Nesbitt indicated that "there are often 3 or more very considerable vessels going to and penetrating the ossifications" and commented that "near the ossification you will rarely miss feeling by the point of a knife bony particles." In a series of drawings of the formation of the secondary ossification center of the distal femur, he notes in one early section that vessels only are seen in the cartilage, and in others as ossification increases centrally various additional vessels appear. Thus, even before the development of the cell theory there were clear descriptions involving the transformation of cartilage to bone and the close relationship of vascularization to bone formation.

C. Duhamel Henri-Louis Duhamel of France demonstrated in papers published from 1739 to 1743 that madder colored only those parts of the skeleton that were being formed at the time of its administration (88-92). When madder feeding was suspended for several weeks before sacrifice, the bone at the extreme ends of the shafts of long bones was uncolored as was that of the most peripheral cortical bone surrounding the midshaft region. By varying the time of madder feeding, he inferred that a long bone grew in length from its ends and in thickness by the progressive development of new bone on its outer surface. The discovery of the osteogenic function of periosteum is credited to Duhamel on the basis of his interpretation of madder feeding and subsequent patterns of bone staining. Duhamel also found that bone grew in length at its extremities by drilling two holes at a measured distance in the shaft of a growing bone, filling them with metal plugs, and finding no difference in their position with time. He was unable to explain the increasing width of the marrow cavity with growth.

D. Hunter The biological cause of the increased diameter of the marrow cavity with growth was first understood in detail by John Hunter, who from 1740 onward recognized both the growth of bone in length at its ends and the deposition of new bone by the periosteum on the outer surface of the shaft, as well

7

as the absorption of preexisting bone that must occur within the marrow cavity and on the external surface of the expanded metaphyses (169). He referred to this phenomenon as "modeling absorption" and clearly expressed the dynamic component of bone formation. Hunter used madder staining and diaphyseal hole drilling approaches to assess subsequent growth in pigs and fowls. He also discussed the modeling of the head and neck of the femur. His work was the first to clearly recognize that absorption of bone was as essential to overall bone growth as deposition of bone.

E. Howship Howship demonstrated the interplay of cartilage formation and bone formation in human and animal embryos on the basis of studies with a solar compound microscope (165). His text and illustrations presented in 1815 defined the embryonic 8-week human hand, showing the primary ossification "rings of bone" of the metacarpals and phalanges. He identified cartilage canals in the epiphyseal ends of the bones and was able to oudine the laminar structure of cortical and trabecular bone tissue. In a remarkable tissue section of the distal femur of a newborn child, the epiphyseal-metaphyseal junction was magnified. The edge of the newly formed bone, by which he refers to the metaphyseal region, "exhibited an appearance of small short pointed villi shooting forward from the surface of the bone into the substance of the cartilage." He noted that "all sections exhibited an apparent alteration in the texture of the cartilage upon the surface connected with the bone. In many instances the cartilage seemed to be more opaque here than elsewhere, this slight opacity forming a line equal to 120th of an inch in breadth." The latter refers to the physeal cartilage, which indeed can be distinguished from the adjacent epiphyseal cartilage at low powers of magnification. Howship further studied the cartilage-bone junction in a distal femur from a 3-week-old child, preparing the tissue for examination by maceration, cleaning, and a form of decalcification. He examined the bone from the diaphyseal region toward the growth region. "It was observed that in proceeding from the middle of the cylindrical bones, where the medullary spaces are larger and the cancellated structure stronger towards the more recendy formed extremities of the bone, the ossific masses become more numerous, of a lighter substance, and a thinner texture; the same gradation being continued up to the margin of the newly ossified surface, where the structure is most curiously wrought, and so exquisitely fine as scarcely to admit a description." His description clearly relates to the changes at the lower end of the physis and the farthest reaches of the adjacent metaphysis. "It was ascertained that the first and earliest state in which the particles of ossific matter become apparent, after they have formed a mass by their cohesion, may be considered as an assemblage of the finest and thinnest fibers, molded into the form of short tubes, arranged nearly parallel to each other, and opening externally upon the surface connected

CHAPTER 1 9 Developmental Bone Biology with the cartilage. These tubes appeared to correspond in number to the villi noticed in the last examination." Similar studies were also performed in animals. Howship clearly defined cartilaginous canals within the cartilage tissue at the ends of the bone. He noted that the principle of bone formation involving the appearance of the cartilage and ossifying bones "was in every respect precisely similar" in many species and that "the same purpose of ossification is accomplished by one and the same means." Howship concluded that the first rudiments of ossification in the long bones were associated with vascularization and occurred "upon the internal surface of the periosteum, which produced a portion of a hollow cylinder; this form of bone having been found antecedent to the evolution of any cartilaginous structure." Howship mentioned the importance of circulation both for cellular growth and for providing the means of calcification. He noted the value of the cartilaginous mode of bone formation, indicating that "at a certain stage of the process the mode of operating is changed in order that it may proceed more expeditiously. A cartilage is formed, which, by the nature of its organization, and by admitting of a specific provision of cavities and canals lined with vascular membranes, which secrete an abundant store of gelatinous matter, is adapted to this particular purpose; while at the same time it serves to determine the future figure of the extremity of the bone by establishing and conducting the ossification within its own substance." He indicated that "from the period when the ossification proceeds in the mode above described by the medium of cartilage the process is continued in the same uniform manner until it has completed the growth of the bone. The growth of the epiphyses at the ends of the bone are also effected by the same means." He also noted the simultaneous formation of bone peripherally in cylindrical bones being "deposited primarily in the form of fine thin tubular plates: a mode of deposition of all others the most favorable for their being subsequently remodeled and for facilitating all the subsequent changes of structure they are destined to undergo." He commented on the mechanical aspects of bone development, noting that "the principal agent in extending the cylinder and in effecting the subsequent progressive changes of structure which in a growing bone are continually taking place appears to be simply the mechanical pressure exerted by the fluid secretions within the medullary cavities of bone, this power operating successively in different directions according to the particular determination given by the circulation." "The particular simplicity observable in the mode of production of the bones of the skull affords a strong argument in favor of the opinion that pressure variously modified constitutes one of the most efficient instruments in the hand of nature."

F. Flourens Flourens also emphasized that longitudinal growth of the long bone took place at the ends (Fig. 3A). He outlined six major principles of bone growth based on extensive refer-

ence to the work of Duhamel as well as his own experiments, which repeated and augmented Duhamel's work (106-110). His six principles of bone formation follow: (1) bone is formed in the periosteum; (2) bone grows in thickness by superimposing new layers externally; (3) bone grows in length by adding layers at each end; (4) the medullary cavity grows by resorption of the inner layers of bone; (5) the ends of the bone are first formed, then resorbed, and then reformed as the bone grows; and (6) bone development is accompanied by a constant change in the bony substance with the gaining of new molecules and loss of older molecules. Oilier also demonstrated the principle of long bone longitudinal growth from either end (Fig. 3B) (258).

III. EMBRYOLOGY OF THE LIMBS A. Timing and Staging of Human Limb Development The limb buds form as outpouchings of the embryo lateral plate mesoderm mass and are composed initially of undifferentiated mesenchymal cells uniformly packed throughout the entire extent of each bud and continuous with the undifferentiated mesenchyme of those regions that will become the shoulder and pelvis. There is a craniocaudal differential time gradient in development, with the upper limb buds appearing in the lower cervical region in the human on day 24 and the lower limb buds in the lower lumbar region on day 28. By 33 days the hand plate is seen, and by the end of week 6 all upper and lower limb segments can be seen. Digital rays appear in the upper limb during week 6 and in the lower limb during week 7. By the end of week 8 components of each of the upper and lower limb bones are formed in cartilage. The embryonic period comprises the first 8 postovulatory weeks, with limb morphogenesis in the human occurring between weeks 4 and 8. At 8 weeks ossification of the humeral diaphysis begins, a time at which embryonic development is arbitrarily considered to be over and fetal development, which involves the growth of fully established models, begins (102, 210, 261-264). Embryonic staging terminology follows. The human embryonic phase is divided into 23 stages using the Carnegie system. This system, adopted in the early 1970s, is a refinement of what were previously referred to as Streeter's horizons. Much human embryologic and fetal study is categorized on the basis of the crown-rump length expressed in millimeters. The Carnegie staging system incorporates 23 stages and relates them to crown-rump length and age in postovulatory days. It is printed in Table IIA, listing some general correlations in the human embryonic periods along with developmental features particularly related to limb development. A more detailed outline of human limb development is listed in Table liB. O'Rahilly (264) makes several points in terms of previous descriptions referable to embryonic staging in the human. These include the following: (1) The term horizon used by

SECTION III ~ Embryology o f the Limbs

9

FIGURE 3 Earlyexperimental illustrations showing that long bones grow in length from their ends. (A) This reproduction of an illustration from Flourens (109) depicts studies showing long bone growth followinginsertion of two metal pins in the diaphysealregion and one in each epiphysis of the rabbit tibia. The distance between diaphyseal pins two and three always remains the same, whereasthe distance between epiphyseal-diaphysealpins one and two (proximal) and three and four (distal) progressivelyincreases with time. The upper and lower sets initially are equidistant at slightly over 6 mm in this magnification, but in Figure 5 the distance is 20 mm above and 16 mm below, indicating not only that growth in length has occurred but also that proximal growth activity is slightly greater than distal. These studies, along with those done earlier by Hales, Duhamel, and Hunter, led to the realization that bone grows in length by the addition of tissue at either end rather than interstitially. As the bone grows in length, the distance between the two diaphyseal pins remains the same, but the distance between the diaphyseal pins and those in the epiphyses continually increases. (B) An illustration from Oilier (258) demonstratesthe same principle: the two diaphyseal pins remain the same distance apart even though extremegrowth in length of the rabbit tibia has occurred.

Streeter is no longer used and has been replaced by the term stage. (2) Roman numerals from the old Streeter classification have been replaced by Arabic numerals to denote what are now referred to as the Carnegie stages. (3) The most useful single measurement of an embryo or a fetus is the crown-rump (C-R) length, which is expressed in millimeters. (4) The c r o w n - r u m p lengths used by embryologists agree closely with those determined ultrasonically. (5) The length of an embryo is not a stage and when used in a descriptive mode should simply be reported as, for example, 15 mm. (6) Stages within the embryonic period are expressed as postovulatory weeks or days. (7) Ages previously described by Streeter are incorrect for the human. Finally, (8) the 23 stages of the Carnegie system refer to the embryonic period only, that is, the first 8 postovulatory weeks; no widely accepted staging system has been devised for the fetal period.

B. Outline of Embryonic Development of Long Bones Mesenchymal condensation has outlined the scapula and humerus of the upper limb and the pelvis and femur of the

lower limb by the end of week 5. By early in week 6 the developing models of the more distal limb bones are seen and chondrification has begun in humerus, ulna, and radius. By late in week 6 carpal and metacarpal chondrification has begun; by the middle of week 6 the femur, tibia, and fibula chondrify, with tarsals and metatarsals following by late week 6. By the end of week 7 all upper extremity bones are chondrifying as are all bones of the lower extremity except the distal phalanges, which do so in week 8. The appearance of the diaphyseal primary ossification centers also follows a regular sequence: clavicle, early week 7, followed by humerus, radius, and ulna; femur and tibia, week 8; scapula and ilium, week 9; ischium, week 15; calcaneus, week 16; and pubis, week 20. The developing model of each long bone is preformed in cartilage (102, 120, 131). The undifferentiated mesenchymal cells, which have undergone condensation and started to outline specific bone shapes, then differentiate, surround themselves with a cartilaginous matrix, and take on the conformation of round chondrocytes. Use of histochemical stains such as Safranin O-fast green shows the pinkish development of the matrix, indicating the presence

CHAPTER 1 9 Developmental Bone Biology

10

TABLE IIA Correlation of Timing Systems Used for Human Embryos (Weeks I-8)" Week

5

Day

Length (mm)

Carnegie stage

1 1.5-3 4 5-6 7-12

0.1-0.15 0.1-0.2 0.1-0.2 0.1-0.2 0.1-0.2

1 2 3 4 5

13 16 18

0.2 0.4 1-1.5

6 7 8

20

1.5-2.5

9

22

2-3.5

10

24

2.5-4.5

11

26

3-5

12

28

4-6

13

32

5-7

14

33

7-9

15

37

8-11

16

41

11-14

17

44

13-17

18

47

16-18

19

50 52

18-22 22-24

20 21

54

23-28

22

56

27-31

23

Features Fertilization First cleavage divisions (2-16 cells) Blastocyst free in uterus Blastocyst hatches, begins implanting Blastocyst fully implanted Primary stem villi appear; primitive streak develops Notochordal process forms; gastrulation commences Neural plate and neural folds appear; primitive pit forms; vasculature begins to develop in embryonic disk Caudal eminence and first somites form; neuromeres appear in presumptive brain vesicle; primitive heart tube forming Neural folds begin to fuse; cranial end of embryo undergoes rapid flexion; myocardium forms and heart begins to pump Primordial germ cells begin to migrate from wall of yolk sac; cranial neuropore closes; optic sulci form Upper limb buds appear; caudal neuropore closes; urorectal septum begins to form; pharyngeal arches 3 and 4 form Dorsal and ventral columns begin to differentiate in mantle layer of spinal cord and brain stem; lower limb buds appear; septum primum and muscular ventricular septum begin to form in heart Spinal nerves begin to sprout; semilunar valves begin to form in heart; metanephros begins to develop; lens pit invaginates into optic cup; cerebral hemispheres become visible Hand plate develops; arterioventricular valves and definitive pericardial cavity begin to form; lens vesicle forms and invagination of nasal pit creates medial and lateral nasal processes Foot plate forms on lower limb bud; major calyces of kidney begin to form and kidneys begin to ascend; genital ridges appear Finger rays are distinct; bronchopulmonary segment primordia appear; septum intermedium of heart is complete; cerebellum begins to form Skeletal ossification begins; elbows and toe rays appear; intermaxillary process and eyelids form on face Trunk elongates and straightens; pericardioperitoneal canals close; septum primum fuses with septum intermedium in heart; minor calyces of kidneys are forming Upper limbs bend at elbows Hands and feet approach each other at the midline Eyelids and auricles are more developed Definitive superior vena cava and major branches of the aortic arch established; gut tube lumen almost completely recanalized

aDerived from Larsen (210) and O'Rahilly (264).

of glycosaminoglycans (300). When the cartilage model of each of the long bones has been formed, the region in which the joint eventually will be present is still filled with cells and is referred to as the interzone area. Early shaping of the epiphyseal ends of the bone occurs prior to necrosis and

resorption of cells in the interzone area. When the latter occur, the joint cavity is formed and the complete model of the developing bone and joint has been formed. The cartilage model of the developing bone then increases in size by both interstitial and appositional growth of the chondrocytes. At

SECTION III 9 Embryology o f the Limbs

TABLE liB

11

Stages a t Which D e v e l o p m e n t a i F e a t u r e s Appear and Events O c c u r in Human Limbs a Feature

Stage for upper limb

Stage for lower limb

Limb bud Length: width = 1.1 Apical ectodermal ridge Hand plate-foot plate Mesenchymal skeleton Mesenchymal scapula-hip Mesenchymal humerus, radius, ulna-femur, tibia, fibula Chondrifying humerus-femur Chondrifying radius-tibia Chondrifying ulna-fibula Finger rays-toe rays Chondrifying metacarpus-metatarsus Chondrifying carpus (except pisiform) tarsus Chondrifying scapula-hip Chondrifying proximal phalanges Homogeneous shoulder and elbow-hip and knee Homogeneous wrist-ankle Three-layered elbow-knee Chondrifying middle phalanges Chondrifying distal phalanges Three-layered wrist-ankle Ossifying humerus and radius-femur and tibia Ossifying ulna-fibula Cavitation in shoulder and elbow-hip and knee Cavitation in wrist-ankle

12 14 14-17 15 15 16 16 16-17 17 17-18 17-18 17-18 18-19 18 18-19 19 ? ? 19-20 20-21 21 21-23 22-23 23 23

13 15 15-18 16 16 15-18 17 17-18 17-18 17-18 18 18-19 18-19 19 19-21 19 21 21 21 21-23 23 22-23 23 23 ?

aDerived from R. O'Rahilly and E. Gardner (263).

a certain stage of development, the primary center of ossification forms. There is some confusion in the literature as to the exact meaning of this term. It generally refers to the circumferential mid-diaphyseal rim of periosteum, which synthesizes the initial bone of the cortex using the intramembranous mechanism without the mediation of a cartilage phase. It can also refer to mid-diaphyseal endochondral ossification within the cartilage model. The initial site of endochondral bone formation is mid-diaphyseal although it appears to occur slightly after the initial periosteal cortical bone formation. Contemporaneous with the periosteal new bone formation is hypertrophy of cells in the mid-diaphyseal region of the cartilage model, calcification of the cartilage matrix, and vascular invasion of the hypertrophic cell lacunae accompanied by undifferentiated preosteoblast cells that then synthesize bone on the scaffold of the calcified cartilage matrix. The vascular invasion occurs in the areas of hypertrophic cells and serves to remove these and replace them

with marrow cells and newly synthesized bone. The zone of hypertrophic cells or ossification front is then extended toward either end of the long bone. The central replacement of hypertrophic chondrocytes with deposition of bone on the calcified cartilage cores encompasses what is referred as the endochondral mechanism. In the periosteal region intramembranous bone formation extends the periosteal new bone sleeve. The periosteal development always is spatially somewhat more advanced toward either end of the bone than the central endochondral development. As this developmental sequence works its way toward either end of the bone, the cartilage forms itself into a specifically structured region referred to as the epiphyseal growth plate. This is characterized by specific conformations of the chondrocytes and serves as a functionally specialized region responsible for longitudinal growth. This pattern of long bone development encompassing both endochondral and intramembranous mechanisms is illustrated in Fig. 4A-4E.

12

CHAPTER I 9

Developmental Bone Biology

F I G U R E 4 The five diagrams (A-E) illustrate the cell and tissue changes in long bone formation. [Derived from references 83, 188, 215, 234, 266, 326.] (A) The cartilage model of the developing bone is shown at left. A and B represent cross-sectional cuts of the developing bone with the tissue pattern illustrated below. In the diagram at far left, the tissue representations are the same because the

SECTION IV 9 Bone Development at the Light Microscopic Level

IV. B O N E D E V E L O P M E N T AT T H E L I G H T MICROSCOPIC LEVEL FOLLOWING DELINEATION OF THE CELL THEORY AND A D V A N C E S IN M I C R O S C O P Y AND H I S T O C H E M I S T R Y Although early studies clearly established that growth in long bones occurs by means of the cartilages at either end, progress was slower in elucidating the structure of the growth cartilage itself. With the development and widespread acceptance of the cell theory dating from 1838, combined with advances in microscopy techniques, the structure and function of the growth apparatus became clearer. Kolliker (1850) illustrated the sequential findings in growth of a long bone from embryonic to mature phases showing the upward and outward deposition of tissue and the need for both cell deposition and cell resorption during the process (193, 194). He initially identified and described the functions of the osteoclast as the cell responsible for tissue resorption. Gross examination of a neatly cut longitudinal section of a developing bone showed tissue and vascular differences between the epiphyseal cartilage area and the metaphyseal-diaphyseal regions that had been appreciated long before the era of microscopy. Broca defined these regions structurally on the basis of both gross inspection and the early use of histologic sections,

13

which enabled him to study the phenomenon down to the cellular level (37, 38). He was able to observe five layers at the physeal and periphyseal regions. The first was the "couches cartilaginous;" the second was a bluish region of the epiphyseal cartilage referred to as the "couches chondroid," which corresponded at the histological level to the columnar cell region or the "cartilage series." The third, toward the diaphysis, was called the "couches chondrospongioid" because it had gross appearing characteristics both of the chondroid layer above and the spongioid below. The fourth with a yellowish tinge was referred to as the "couches spongioid," which in histologic terminology represented the hypertrophic cell zone including the area of calcified cartilage, and the fifth was referred to as the "tissue spongieux," which represented the metaphyseal bone and was red in appearance. Tomes and deMorgan also described and illustrated the growth plate sequence (348). The earliest detailed descriptions of epiphyseal cartilage development and the growth plate mechanism that reached a coherent understanding at the light microscopic level were provided by Heinrich Mueller in 1858 (237). His illustrations of the growth plate apparatus depicted the cell structure of the epiphyseal growth plate and the epiphyseal-metaphyseal junction in great detail and in a way fully consistent with observations made today (Fig. 5). He clearly illustrated the palisading or proliferating cell zone of the growth plate and

F I G U R E 4 (continued) entire model of the bone is still in a cartilage phase. In the central illustrations, chondrocyte hypertrophy and matrix calcification are shown in the central or diaphyseal region. Chondrocyte hypertrophy starts as a central nidus both in the middle part of the shaft when considered in a longitudinal orientation and in the central part deep within the cartilage when considered in the transverse orientation. This eventually will pass from one edge of the shaft to the other as is shown in the transverse section A below. At far right, the intramembranous bone formation from the periosteum has begun surrounding the endochondral bone formation centrally. The two mechanisms of bone formation are illustrated most clearly in the transverse cut section A as seen at bottom. The initial formation of endochondral bone within and intramembranous periosteal bone at the periphery occurs at approximately the same time. This initial area of bone formation is referred to as the primary center of ossification. The periosteal site of bone formation appears to precede the endochondral by a very short time interval. In a strictly technical sense, that region that formed first would be the primary center of ossification. This would have biological significance, but in a practical sense both mechanisms can be referred to as the primary center. An accurate observation, however, and one made by most observers, is that the intramembranous or periosteal new bone formation is present spatially somewhat in advance of that occurring within the endochondral sequence centrally, as bone formation passes from the primary center of ossification toward either end of the bone. This is a relationship that is maintained even in relation to formation of the physis and the perichondrial ossification groove of Ranvier. (B) Events surrounding the primary ossification center formation are detailed. Once the cartilage within the midpart of the shaft has hypertrophied and its matrix has calcified, vascular invasion occurs from the periphery along with undifferentiated mesenchymal cells, which shortly lay down bone on the calcified cartilage cores. Endochondral bone formation is underway. Intramembranous bone formation at the periphery occurs directly without the mediation of a cartilage phase. Note also that the intramembranous bone is spatially in advance of the more central endochondral bone. There are no vessels within the developing diaphyseal cartilage prior to its hypertrophy and matrix calcification. It is only when hypertrophy has occurred that vascular invasion from the adjacent perichondrium-periosteum occurs. Thus, no cartilage canals are present in the diaphysis and metaphysis analogous to those seen later in the epiphysis. (C) Bone formation advances toward either end of the developing bone with a characteristic physeal orientation of the cartilage and endochondral sequence occurring. The physeal orientation is represented here by the slanted lines. The solid region below represents the calcified cartilage of the lower part of the hypertrophic zone and the persisting cartilage cores in the metaphysis. Note also the advance of the intramembranous bone formation and its slightly greater peripheral extent than that of the endochondral bone within. (D) Formation of the secondary ossification center above is shown. The endochondral sequence has moved the physes relatively closer to either end of the bone. The secondary ossification center has formed above, whereas at the lower end it has not begun to form. This is a characteristic feature of long bones in which one center, which can be at either end, forms before the other. Vessels passing from the periphery are present within the epiphyseal cartilage in cartilage canals for many weeks and sometimes for many months before formation of the secondary ossification center. This feature is not characteristic of the endochondral mechanism at the primary center of ossification. (E) The secondary ossification center is now seen at the lower end of the developing bone where the physis (oblique lines) is still open. At the other end, the physis has been closed completely and resorbed and there is continuity between metaphyseal and epiphyseal bone. The growth on the undersurface of the articular cartilage has also terminated and the innermost zone of articular cartilage is now calcified.

14

CHAPTER

1

~

Developmental Bone Biology

FIGURE 5 This histologic illustration of the growth plate from Mueller's 1858 article (237) shows his accurate depiction of the specific structures.

its transformation over a few cells to the hypertrophic zone. The calcified cartilage matrix of the hypertrophic cell region was identified, as was the invasion from below by vessels and bone forming cells, the deposition of new bone on the cartilage cores, the presence of osteocytes adjacent to the new region of bone formation, and the diaphyseal (metaphyseal) marrow. Mueller provided a series of illustrations of transverse cuts through the growth plate region of a human embryo 3 months of age. The first cut passed through the upper layer of small cartilage cells of the physis; the second through the region bordering the calcified cartilage where the cartilage cell cavities were somewhat larger, the third through calcified cartilage with large chondrocyte cavities (the hypertrophic zone); and the fourth through a zone characterized by central cores of cartilage trabeculae, surrounding

bone, and cellular marrow. Mueller recognized the structural changes at both the cell and matrix levels in the physeal regions. In his illustration of the outer reaches of the diaphysis, he showed both vessels and red cells. He concluded from his structural observations that elements of the marrow could be regarded as derivatives of young generations of cartilage cells that could transform themselves into bone cells. Mueller was described by Retterer (292) as showing "no doubt on this point: the elements or cells of the bone marrow and even the bone cells (osteoblasts) are derived from cartilage cells." He felt that the cell division was so rapid that it was impossible to observe. On the other hand many cartilage cells died by being included in the calcified matrix. His work stood as the standard for explaining and illustrating growth plate morphology over several decades in the latter part of the nineteenth century. Kolliker reproduced the drawings of Mueller accepting the transformation of cartilage cells into bone cells. Much later, in 1889, he utilized the same drawings but admitted the decline or decay of the cartilage cells and the budding of the perichondrial tissues, providing a source of new cells (194). u (1860) made the important distinction between bone as a tissue, which he defined as the bone corpuscles or cells plus the calcified intercellular substance, and bone as an organ, which encompassed not only the osseous tissue but also the medullary marrow tissue, the periosteum, articular cartilage, and all vessels and nerves (356). It would also encompass the epiphyseal cartilage in a growing child. Long bones grew in length from cartilage and in thickness from periosteum. He described the endochondral sequence in excellent detail. In reference to the cartilage cells he observed that "the greater the number of cells which undergo this change, the larger the cartilage will become and the height to which any one of us attains essentially depends upon the extent to which growth occurs in the individual groups of cartilage cells." Virchow felt that the enlarged cartilage cells (of the hypertrophic region) "may be converted by a direct transformation into marrow-cells and continue as such; or they may first be converted into osseous and then into medullary tissue; or lastly, they may first be converted into marrow and then into bone." He was aware of the calcification of cartilage in the endochondral mechanism, indicating that "what first takes place in the course of these processes is not the production of real osseous tissue, but only the deposition of calcareous salts . . . . There first of all takes place in the immediate vicinity of the border of the bone a calcification of the cartilage which gradually a d v a n c e s . . , so that every individual cartilage cell is surrounded by a ring of calcareous substance. This is not yet bone, it is nothing more than calcified cartilage . . . . " Bone tissue can arise out of marrow cells or directly from cartilage cells. Virchow indicated that "it is no doubt true that in the case of the normal growth in length of bone, most of the bonecorpuscles do not directly proceed from cartilage-corpuscles, but are immediately derived from marrow-cells.., but it is

SECTION IV 9 Bone Development at the Light Microscopic Level just as true that cartilage-cells can also be transformed straightway into bone-corpuscles." He felt that the isolated transformation of single cartilage cells into bone corpuscles was an accurate observation and also of great importance to the cell theory in general. This direct transformation was not associated, therefore, with death of the cartilage cell and its subsequent replacement. At that time, a cartilage corpuscle was considered to be composed of the cartilage cell and a surrounding membrane that was an integral part of it, whereas a bone corpuscle referred only to a bone cell with the lacunar or canalicular wall representing the endpoint of the adjacent matrix. He pointed out that the direct conversion of cartilage into osteoid tissue was clearly evident at points of transition from cartilage to bone where the boundaries of the different forms of tissue are "completely obliterated and all sorts of transitions between round (cartilaginous) and jagged (osteoid) cells are seen." Gegenbaur (1864) was a morphologist whose studies primarily involved comparative anatomy, but they did involve embryologic and histologic analyses. In studies on the development of primary and secondary bone he indicated that the tissue eventually formed was the same even though the site of bone formation and the tissue replaced differed. In his early works he defined the cell most closely involved in bone formation as the osteoblast (123) Oilier (1867) wrote a classic treatise on bone development and regeneration (258). This extensive experimental work defined the role of the periosteum in bone formation. He also demonstrated that bone irritation increased the rate of growth at the epiphyseal lines and that damage to the epiphyseal lines inhibited growth. He was the first to study the variable amounts of bone growth at the proximal and distal epiphyses, which we will summarize later. He also repeated the experiments of Duhamel and Flourens, showing that growth in length took place only at the epiphyseal lines (256-258). A brief review of the major landmarks in bone growth research is presented in Table I. Growth plate structure and development were further described beginning in the late nineteenth century with particularly excellent presentations by Waldeyer (362), Schafer (309), and Bidder (29). It was some time after the work Nesbitt published in 1736 before the two modes of bone formation, intramembranous and endochondral, were more widely appreciated. Kolliker wrote on the morphological significance of membrane and cartilage bone. Russell (303) noted that Reichert (1849) and Mueller (1858) (237) had pointed out that there was essentially no difference in the eventual histologic structure of bone whether it had been formed initially in cartilage or in membrane, a view still held today. Retterer (1900) described the development of bone by the endochondral mechanism in a detailed article entitled "The Development of the Transitional Cartilage" (292). His work on the histogenesis of the cartilage referred in detail to works of investigators throughout the previous century. The

15

various developmental phases from the embryonic limb bud onward were defined. Growth of the cartilage model was accomplished by both interstitial and appositional mechanisms in which the perichondrium contributed to the growth in width of the cartilage. He provided detailed descriptions and illustrations of the cells and matrix of the hyaline cartilage models long before bony ossification began. The matrix, at that time referred to as the "fundamental substance," was clearly recognized to have many components, among them connective tissue fibrils called collagens, which were very fine in nature and immersed in a matrix with acidlike chondroitin, also referred to as chondromucoid. The nature of the cells and matrix was assessed on the basis of light microscopic histologic preparations. Virchow is credited with being the first to describe the cartilage cell in hyaline cartilage, which was formed of a central nucleus, a surrounding cell body, and a capsule. Retterer concluded that the cartilage hyaline or fetal cell was composed of a large nucleus and a cell body with many organelles leading to a perinuclear granular chromophilic appearance, whereas the cell periphery had reticular fibrils. The capsule was felt to be a continuation of the peripheral protoplasm. The question of transport mechanisms for nutrition within cartilage was raised. It was evident that a diffusion process was available with evidence from many experiments showing that it occurred by diffusion from the synovial cavity (as we accept today), whereas others implied that it passed from cell to cell along cell processes. Retterer assessed the epiphyseal growth plate and the transformation of cartilage to bone with emphasis on the epiphyseal region and the diaphysis (by which we would currently understand both the metaphyseal and diaphyseal regions). A major concern at that time was what became of the cartilage cells at the lower part of the growth plate, with the question being raised as to whether they atrophied and died or whether they survived and were transformed to other cellular elements. This question has not been completely answered almost 100 years later, with more recent work on the hypertrophic cell pointing to its continuing functional role in some instances. Retterer outlined the cartilage transformation during the early stages of endochondral ossification. He demonstrated the cartilage model of the developing bone, the flattening of chondrocytes in the transverse plane at the early regions of physeal differentiation, and the swollen central cartilage cells that were much larger than other cells. He gave this region the name of hypertrophic cartilage ("cartilage hypertrophic"). It was in this zone that the calcium salts were deposited in the matrix, leading to the term calcified cartilage. A four-zone region of cartilage cell transformation was recognized. In the first zone the flattened cells of the cartilage were referred to as "cartilage series." The characteristic of cartilage cells in this region was their flattening in the transverse plane where they were 15-20 Ixm wide and only an average of 9 Ixm thick. The flattened cells were disposed in

16

CHAPTER 1 ~

Developmental Bone Biology

groups separated by a matrix whose long axis was parallel to the long axis of the developing bone. The second zone was the hypertrophic cell zone with calcified cartilage matrix. The cartilage cells enlarge in this region, being larger than in the adjacent zones and either round or polyhedral in shape. The question was raised as to whether the modifications of the hypertrophic cells represented a progressive or regressive phenomenon or, in other words, whether the cells of this zone were preparing to ultimately evolve or perish. Some of the cells histologically appeared to be undergoing atrophy, and most observers indeed were under the impression that the cells of the hypertrophic zone were undergoing a certain death. Retterer felt, however, that if one prepared the cartilage for histologic examination in a different way and did serial sections in paraffin, one was able to obtain preparations in which the hypertrophic zone showed intact cells rather than empty cell spaces. He felt he could demonstrate intact cells with cytoplasmic elements in place, revealing results "entirely different from those announced classically." The third zone, the "cartilage hyperplasie," included the medullary space created by the entrance of blood vessels and red blood cells through the transverse septae into the lower hypertrophic cell lacunae. In the fourth zone, there were trabeculae of bone representing the upper portion of the diaphyseal "spongieuse" ossification region. Retterer thus established a four-zone region of transformation: zone 1, "cartilage serie;" zone 2, the hypertrophic calcified region; zone 3, the zone of the initial medullary space (zone hyperplasie); and zone 4, the zone of ossification. The terminology current at that time was such that the physis was generally described as being diaphyseal due to its position at the extreme extent of the diaphysis and the term metaphysis was not used. Retterer was also able to distinguish clear zonal differentiations within the physis in particular with transverse sections slightly oblique to the horizontal axis, which showed adjacent regions shading into one another. He was able to show multiple and somewhat smaller cell components invading the hypertrophic cell lacunae. Some of the cells were red blood cells associated with the vascular invasion. The hyperplastic zone was characterized by multiple cell types, which merited the name of a zone of metamorphosis, though he wisely indicated that it would be too confusing to introduce yet another term to this complex region. Some cartilage cells passed through the hypertrophic zone without degenerating and then transformed in the upper regions of the medullary zone. Some of these he felt became multinucleated. Calcium was deposited in the matrix trabeculae of the hypertrophic zone, and no cell division had been noted in the hypertrophic zone. The hypertrophic zone did not consist only of an increase in the volume of preexisting cells, but rather the cells were also involved in a formative process that created both nucleus and cytoplasm different from those of the mother cell. He thus refers to an issue that is not fully understood, even today. He goes on to describe the invasion

of the lower parts of the hypertrophic zone by vascular tissue, the development of capillaries carrying blood cells, and the resorption of transverse cartilage trabeculae. He concluded that it was the metamorphosis of the cartilage cell that represented the initial phenomenon of endochondral ossification and indicated that "the transformation of the cartilage cell led to the development of a reticular and vascular tissue analogous to that of the perichondrium; these 2 tissues are capable subsequently of elaborating bone." Retterer also described the development of bone within the epiphysis, where the cartilage was transformed centrally to begin formation of the secondary ossification center at a much later stage. He observed that "the evolution of cartilage tissue of the epiphysis is in every point analogous to that which we have studied in the diaphysis in the developing skeleton of young embryos." Vascular canals were present within the epiphyseal cartilage, but he felt that the cartilage canals formed directly within the cartilage model itself rather than being derived from the perichondrial vessels.

V. M O R E D E T A I L E D H I S T O L O G I C

STUDIES OF BONE FORMATION A. Histogenesis of Bone The histogenesis of bone was studied in detail by Stump (1925), including both the endochondral and intramembranous mechanisms (337). Certain differences from current terminology must be recognized in reading earlier work on developmental bone biology; even as late as 1925, mention is rarely made of the metaphysis and indeed the growth plate is referred to by Stump as the "diaphyseal plate." His descriptions of the developmental sequences raise several points still pertinent to our understanding. The work included the concepts of bone as a tissue growing by apposition and cartilage growing by interstitial proliferation and of the perichondrium evolving into the periosteum "concurrently with changes in the sub-adjacent mesenchyme." Stump indicated in his discussion of the cartilage growth plate that the measure of its growth depended not only on the rate of division in forming the chondrocyte groups but also on "the velocity of enlargement of the cells collectively." The enlargement of cells is primarily in reference to the hypertrophic zone. Persistence of the cartilage cell after its hypertrophy was sought but not found, and he concluded that the hypertrophic chondrocytes "persistently showed the appearance of old mature cells with all the signs of structural degeneration." Calcification of the longitudinal cartilage trabeculae of the hypertropic zone was defined as was its scaffold function. The hypertrophic cell region was invaded by vessels and undifferentiated mesenchymal cells, soon to become osteoblasts. This invasion occurred into the lower regions of the hypertrophic cell mass, with evidence that the transverse, thin, noncalcified septae were readily resorbed

SECTION V 9 More Detailed Histologic Studies o f Bone Formation

but that osteoclasts or chondroclasts were not involved in this particular phase of development. Endochondral bone was transitory in nature, "serving to increase the stability of a bone at the site of its growth. It undergoes absorption, extending the area of the medullary cavity toward the diaphyseal [which we now describe as epiphyseal] cartilage." In the epiphysis itself, cartilage and bone deposition was more a feature of growth than absorption, although replacement of all minute trabeculae to meet new tension and pressure stresses was a continuous process.

B. Chondrocyte Shape and Orientation in Epiphyseal and Physeal Cartilage; Mineralization and Vascularization in the Endochondral Sequence (Dodds) Dodds was one of the earliest to focus on specific cellular orientations within the epiphyseal growth plates (82). He was particularly interested in longitudinal row formation of chondrocytes and their gradual enlargement toward the metaphyseal region. He described the early epiphyseal cartilage structure (as distinct from the physeal structure) at either end of a long bone as the endochondral sequence worked its way from the diaphysis toward its ultimate epiphyseal position. He summarized as follows. In the cartilaginous ends [the epiphyseal regions] of young bones, before the appearance of the epiphyseal centers, the condition of the cartilage is that of a primitive type of hyaline cartilage, such as would be found elsewhere in the body, with no relation to the process of ossification. The cells are small and all very much of the same size. They have a rounded form. They are undergoing mitotic divisions which take place in all planes. Following the divisions, the pairs of daughter cells slowly migrate apart and soon come to occupy separate lacunae. For this reason, the cells of this region occur singly, except for those, here and there, which, on account of recent division, occur in pairs. The growth of the cartilage in this region is not rapid, nor is it conspicuously greater in any one direction. It is simply sufficient in amount and of a suitable nature to keep the cartilaginous ends of the bones of proportionate size and of proper shape. This is the primitive type of cartilage-cell arrangement from which the (physeal) rows are derived. It might also be remarked, in passing, that in this region the cartilage also receives increments from the mesenchyme which underlies the perichondrium, as well as from the interstitial growth just described. Dodds describes the development of the epiphyseal growth plate cartilage region, the cytologically and functionally specialized area of the epiphysis. He defines differentiating features from the primitive (epiphyseal)cartilage described in the preceding paragraph: (1) a definite orientation of the mitotic figures; (2) the two cells resulting from each division remain close together; (3) instead of retaining the primitive rounded form, the two daughter cells become greatly flattened, assuming a discoid form; and (4) all of the pairs of flattened cells remain oriented in the same way with

" .....

..~S.;,.:

........

17

~ . . .

A 9 ..,

.........

. ..........

-

.

o

.

~

......................

9

B ..................

. .....

(~'""

. ...............

...~=,...

.....""~i'...

I FIGURE 6 Dodd's illustration of the genesis of row formation in the proliferating part of the cartilage growth plate from newborn dog metatarsal (82). At left (1), the mode of cell division is shown beginning with a single cell (A) at top left and progressing to a flattened chondrocyte at lower right (I). At right (2), the beginning of a row of cells by division is shown. Part (2): A, telophase; B, two daughter cells just after division; and C, the two daughter cells after flattening and final reorientation. [Reproduced from Dodds, G. S. (1930). Anat. Rec. 46:385-399, copyright 9 1930, WileyLiss, Inc., a subsidiary of John Wiley & Sons, Inc.]

their widest diameters perpendicular to the long axis of the bone. These pairs of flattened cells are the beginning of rows of chondrocytes. This area of the physis is referred to by most as the proliferating or columnar cell zone. The rows are built by the division of flat cells, which following division retain a relation "like coins in piles" (Fig. 6). The number of cells in a row is constant for most epiphyses, but different regions have different characteristic numbers. These cell divisions increase the length of the rows of cells and thus are an important, but not the only, factor in the elongation of the bone. The increase in physeal width comes by multiplication of the number of rows caused by more numerous row mother cells produced by the epiphyseal cartilage. Some would refer to this process as occurring in the germinal or resting cell zone of the physis. Cartilage matrix forms in a thin septum between the two daughter cells, with transverse partitions being much thinner than the longitudinal septae that intervene between adjacent rows. There is progressive hypertrophy of the cartilage cells. When the full number of cells in a row has been produced, the enlargement of the cells begins by generally affecting the cells in sequence, beginning at the diaphyseal ends of the rows. The growth in thickness by the physis along its longitudinal axis is greater than that in width, a disparity that tends to turn the transverse flattened cells into more discoid cells. During growth of the cells, they remain in rows such that "elongation of the cartilage which was begun by the division of the fiat cells in the rows, is continued with increased speed by the rapid enlargement of these same cells." Cells divide in the transverse plane and then rearrange in the longitudinal plane. The synthetic patterns of the chondrocytes during rapid growth of the cells in rows are recognized because there must be a consequent increase in the total length of the rows and also a corresponding increase in the length of the longitudinal trabeculae

18

CHAPTER 1 9 Developmental Bone Bioloyy

between them. These happenings represent the actual elongation of the growing bone. Dodds was one of the earliest to note that "when calcification of the cartilage matrix takes place, it affects only the longitudinal trabeculae, the transverse walls between the cells being unaffected by it. These thin, transverse, uncalcifled walls are seemingly easily destroyed by the invading marrow as are also the degenerate cartilage cells within the lacunae so that the long parallel cylindrical holes are easily produced as lining on which the first bone matrix is deposited." Formation of the primary center of ossification in each bone occurs by ossification in the central cartilage model following cell hypertrophy and vascular invasion. Dodds implies that central bone formation by the endochondral mechanism occurs "at about the time of the beginning of sub-perichondral ossification," by which he means the first evidence of deposition of periosteal intramembranous bone. Physeal structure is further summarized by Dodds in another article dealing primarily with cartilage removal by osteoclasts at the lower zone of the physis (83). Mineralization of the cartilage matrix occurs along the longitudinal septae with essentially no mineralization of the narrow or transverse septae. Invasion of the hypertrophic chondrocyte lacunae is characterized by both undifferentiated mesenchymal cells and loops of capillaries. When the lacunae have been opened by a defect in the transverse septum, the marrow cells and capillaries advance into it beside the cartilage cell and immediately adjacent to the mineralized longitudinal cartilage core. "The primitive connective tissue cells of the marrow are seen to gradually develop into osteoblasts which begin at once to deposit bone upon the persisting, longitudinal, calcified cartilage walls surrounding the cylindrical holes." The osteoclasts, which also form by a fusion of marrow cells, are not found in the advancing capillary front but are close behind it and are always applied to the persisting remains of the calcified longitudinal walls to resorb them. Osteoclasts can resorb both the bone spicules on the cartilage core or just the cartilage alone, at which time they can also be referred to as chondroclasts. The osteoclast is present only on calcified material, be it bone or calcified cartilage. A brief definition of the cells present in bone is provided.

C. Cellular Components 1. UNDIFFERENTIATED MESENCHYMAL CELLS The developing tissues of the skeleton form from undifferentiated mesenchymal cells. These are present initially in the limb buds prior to any histologic developmental differentiation. They are simply uniform appearing round to oval cells with a nucleus and cytoplasm not yet surrounded by any specific matrix. They have the potential for differentiation with the appropriate stimulus along various tissue producing lines, including becoming chondrocytes to produce cartilage, osteoblasts to produce bone, fibroblasts to produce

fibrous tissue, adipocytes to produce fat, or myoblasts to produce muscle. 2. OSTEOBLASTS Osteoblasts are active bone forming cells. They are characterized by an abundant cytoplasm filled with rough endoplasmic reticulum at the ultrastructural level. Some refer to the cell intermediate between the undifferentiated mesenchyreal cell and the osteoblast as a preosteoblast. The cells are responsible for synthesizing large amounts of collagen primarily of the type I variety, which accumulate to form the matrix of bone. We refer to two types of osteoblasts based on their topography. The mesenchymal osteoblast is surrounded completely by randomly oriented collagen fibrils and is thus responsible for the synthesis of woven bone. Surface osteoblasts line up along the surface of preexisting bone tissue and synthesize collagen fibrils along the preexisting surface in a parallel array. The surface osteoblast is thus involved in the direct synthesis of lamellar bone. As bone synthesis proceeds, the osteoblast becomes completely surrounded by matrix referred to as osteoid, and when that matrix becomes mineralized the encased cell is referred to as an osteocyte. Gene and molecular controls of osteoblast differentiation and the structural molecules synthesized by osteoblasts are reviewed in Sections IX and X. However, an osteoblast-specific transcription factor has been identified (86, 87, 195). The first osteoblast-specific transcription factor is Cbfal, one of three vertebrate homologues of the Drosophila runt and lozenge proteins. Cbfal appears to have features specific for early differentiation along the osteoblast line. Cbfal expression is initiated in the mesenchymal condensations of the developing skeleton, is strictly restricted to cells of the osteoblast lineage, and is regulated by BMP7 and vitamin D3. 3. OSTEOCYTES Osteocytes are mature bone cells. They reside in spaces referred to as lacunae, and their cell processes are connected to one another and are present in canals referred to as canaliculi. Each osteoblast and osteocyte has numerous cell processes passing from it that serve to relate to cell processes from adjacent osteoblasts and osteocytes. These cell processes link up via the gap junction mechanism. 4. CHONDROBLASTS AND CHONDROCYTES Chondroblasts and chondrocytes are cells responsible for the synthesis and maintenance of cartilage tissue. The cells surround themselves with a matrix composed primarily of type II collagen, although there are also considerable amounts of types IX, X, and XI collagen and a large array of proteoglycans. Cartilage has a high proportion of water, which composes approximately 80% by volume of its tissue mass. It is difficult to make a histologic differentiation between chondroblasts and chondrocytes because both cell types are surrounded by cartilage tissue, which unlike bone does not,

SECTION VI ~ Fate of the Hypertrophic Chondrocyte except in rare incidences, mineralize. The chondroblast or chondrocyte has an oval shape with the ultrastructure of the cytoplasmic wall having the appearance of mild scalloping. The collagen fibrils in cartilage are randomly arrayed and tend to be much thinner than those in bone, averaging 10-20 nm in diameter. 5. OSTEOCLASTS The cell type responsible for resorption of bone and cartilage tissue is the osteoclast. This is a large, multinucleated cell formed by the fusion of circulating monocytes. Thus, it is not part of the mesenchymal series but has its origin from the hematopoietic system. The same cell type can resorb cartilage or bone. If the cell is relating to cartilage exclusively then the term chondroclast can be used, although the tendency is to simply use the term osteoclast in relation to the resorption of either bone or cartilage. The osteoclast attaches itself to the underlying tissue by a characterized structural mechanism evident only by ultrastructural assessment. The cell surface has a circular region free of organelles, which is referred to as the clear zone. This serves to attach the cell in a donutlike fashion to the underlying bone and cartilage. The cell surface within the rim of the clear zone is then thrown into innumerable folds or outpouchings, forming what is referred to as the ruffled border. This mechanism serves to allow for increased secretion of lytic enzymes by the cell with the extreme increase in extent of the cell border being caused by the ruffling phenomenon. The circumferential clear zone seals the environment and allows the lytic enzymes to be present in high concentration. It is these enzymes that are responsible for resorption of the underlying mineral and then the matrix of cartilage and bone. Several factors affect osteoclast formation at differing stages of their development, including colony stimulating factor- 1 (CSF- 1 or M-CSF), interleukins-l, -6, and -11, transforming growth factor-[3, tumor necrosis factors (TNF-et, TNF-[3), vitamin D3, calcitonin, and parathyroid hormone (196). The two major molecules exclusively essential for osteoclast function, however, are macrophage colony stimulating factor (M-CSF) and the receptor for activation of nuclear factor K-B (RANK) ligand (RANKL), also known as osteoprotegerin ligand (OPGL) (196, 340). The latter is a tumor necrosis factor family molecule identified as an osteoclast differentiation factor. 6. GAP JUNCTIONS LINKING BONE CELLS Gap junctions serve as areas for direct cell-cell communication either by electrical coupling or as points of passage for small low-molecular-weight messenger molecules (325). They are intercellular channels formed by different membrane spanning proteins called connexins. Gap junctions are observed linking the cell bodies of surface osteoblasts, linking osteoblast processes passing through newly synthesized osteoid, linking osteocyte processes of adjacent cells, and on the surface of osteoblasts or of osteocytes within their lacu-

19

nae. Gap junctions are arranged in five basic shapes as defined by thin section transmission electron micrographs. These appear as linear, stacked linear, curvilinear, oval, and annular.

VI. FATE OF THE HYPERTROPHIC CHONDROCYTE AS INTERPRETED FROM LIGHT MICROSCOPIC STUDIES The fate of the hypertrophic chondrocyte is an important matter and many views have been presented, most of which continue to be debated. An important point should be made concerning the technical preparation of histologic sections of bone and cartilage for light microscopic examination. The techniques in use in the late 1800s and early 1900s in many instances allowed for the preparation of sections with better preservation of cell detail than the paraffin-embedded, hematoxylin- and eosin-stained sections that characterized much bone research of this past century. The drawings of cell and tissue appearance and improving photomicroscopy showed a degree of structural preservation supportive of some of the interpretations made. Trueta has made this point as well (349). Retterer provided a detailed assessment of previous descriptions of cartilage transformation into marrow bone in which he examined in particular the fate of the hypertrophic chondrocytes. He summed up the possibilities by which cartilage cells of the growth plate could subsequently be found in the medullary tissues of what we now refer to as the metaphysis. Many investigators of that era were comfortable describing the transformation of at least some of the hypertrophic cartilage cells into bone cells, whereas some felt that two distinct pathways were occurring: one involving cell death or one involving cell transformation in which the cartilage cell survived and underwent dedifferentiation and reemergence as a bone forming cell. Due to the current high level of interest in the fate of the hypertrophic chondrocyte, a review of previous morphologic findings and interpretations is of interest.

A. Chondrocyte Survival, Dedifferentiation, and Reemergence to a Bone Forming Cell Line Baur denied that there was any direct transformation or metaplasia of cartilage cells to bone, feeling that there was survival and proliferation of the freed cartilage cells, many of which subsequently formed cells that constituted the embryonic marrow (20). Ranvier also believed that there was no direct transformation of cartilage to bone, but rather that cartilage cells, in particular young cells, divided extensively, dissolved their capsules, and led to the formation of embryonic marrow by what appears to have been a dedifferentiation (285). Both observers felt that the medullary cells did derive directly from cells of cartilage after their surrounding capsules were dissolved. Ossification was not direct in the

20

CHAPTER I ~ Developmental Bone Biolo~ty

sense that cartilage cells did not lead directly to the formation of bone cells but underwent initial modification, or dedifferentiation using today's terminology, during which they lost their cartilage faculty. They then became embryonic marrow cells that eventually were able to form bone. Embryonic cells that had not yet taken a determined form could be seen in the development of bone from a cartilage model. The cells were thought to originate in the marrow of the bone under the influence of vessel presence. These vessels led the cartilage capsules to dissolve and the cartilage cells to be freed and then to proliferate, following which they took on an embryonic character referred to as the medullary vascularized tissue. Only later could these marrow cells align themselves along the walls or structures, at which time they were referred to as osteoblasts. The general feeling of many observers, therefore, was that cartilage cells did not transform themselves directly into bone cells, but rather that the bone forming cells developed from descendants of primordial cartilage cells. Mueller also stressed that bone cells were not derived directly from the cartilage cell capsules, but rather from their young progenitors.

B. Direct Transformation of Cartilage Cells to Bone Cells Virchow and Lieberkuhn (215, 292) stressed that cartilage transformed itself directly into bone. Bone cells were interpreted as cartilage cells that survived and bone substance was interpreted as modified cartilage matrix but not a new tissue. Schoney also thought that cells dedifferentiated in the cartilage-to-bone formation phase, although he was clear enough to note that, at the lower part of the growth plate, he had never seen a cartilage cell divide nor the protoplasm of a cartilage cell transform itself into an element of bony marrow (314). Some histologists such as Czermak (72) indicated that the connective tissues of bone and cartilage were similar and that they differed only slightly based on the nature of the surrounding matrices. Metaplastic changes therefore consisted only of a transformation of one matrix to another with one of the subtle changes involving the ability of cells to secrete calcium salts and transform themselves into bony tissue. Leser (212) and Retzius, Brachet, and Retterer (292) each described situations where growth plate chondrocytes survived and were transformed into either marrow medullary cells or osteoblasts.

C. Death of Cartilage Cells An additional view, commonly held through most of the twentieth century, was expressed by Loven and others as early as 1863. It favored the destruction of hypertrophic chondrocytes and the development of bone without any participation of previous cartilage cells (217, 292). He was impressed by the vesicular and swollen nature of the hyper-

trophic cartilage cells that occupy the calcified zone, leading to the interpretation that the cells were degenerating and that calcification led to destruction of the cartilage cells, which served to prepare a space where the bone tissue could invade and develop. He indicated that "it was the blood vessels coming from the perichondrium which contributed to the resorption of the cartilage tissue." He denied that there was any participation of the cartilage cells in the development of the embryonic marrow. Stieda also published similar observations, namely, that the marrow was a tissue whose origin was unique and not derivative from cartilage dedifferentiation (332). Stieda was never able to see a direct passage of the cell line from cartilage to medullary elements, and he felt that the medullary elements originated from the periosteal buds. It was the osteoblastic tissue of the embryonic state that invaded from the undersurface of the periosteum that formed the first bony deposits. Uranossow did not find continuity of cartilage cell division with that of the bone marrow and concluded that the medullary elements did not have their origin in cartilage (292, 354). The cell masses that gave birth to endochondral bone originally derived from the cambial layer of the periosteum. He concluded that cartilage had only a passive role in the development of bone. Levschin denied any relationship between cartilage cells and the marrow elements of long bones (214, 292). Retzius described the changes in cell shape and size as one proceeded toward the hypertrophic zone but felt that the cells were degenerating toward the ossification front (292). Strelzoff indicated that, for most bones, the cartilage was destroyed and bone itself occurred on the basis of osteoblasts derived from the periosteum (335). In some regions, however, in particular in facial bones, ossification occurred wherein cartilage cells were transformed directly to bone cells. An increasingly large number of observers thus defined the hypertrophic cartilage cells to be shriveled and withered, whereas the osteoblasts brought in by the capillary vessels appeared healthy. Others adhered to the view that cartilage cells continued to be derived from the physeal cells even though some were smaller at the lowest part of the hypertrophic zone. Seemingly all were in agreement that, below the palisading layer of the cartilage, there were no longer any mitoses. Brachet indicated that whereas in the hypertrophic zone the cartilage cells seem to be degenerating, it was as though in the region of the resorptive zone they again took on a more embryonic form and subsequently became free in the marrow (31). The question was how the cartilage cells that became hypertrophied and modified reached their ultimate goal. He noted that they essentially disappeared once the capsule had been opened and that osteoblasts filled the space along with vessels and chondroclasts. He indicated that he had never seen transitional forms between the modified chondrocytes and osteoblasts and at least stated in clearcut fashion that "it is impossible for me to determine exactly the ultimate fate of these elements." Retterer felt that most

SECTION VII 9 Structural Development of the Epiphyseal Regions

views inclined to support the classic theory, which attributed an extracartilaginous origin primarily from the blood vessels for the new osteoblasts.

D. Variable Responses Not surprisingly, opinions persisted that there were variable patterns of occurrence in relation to the final state of the hypertrophic chondrocytes. Tschistowitsch (353) indicated that some of the hypertrophic cells degenerated completely and that their space was invaded by vessels from the diaphyseal region, which brought in cells of separate origin, whereas other hypertrophic cells that appeared to be degenerating did not continue to final degeneration but persisted as pale transparent cells, some of which appeared to regenerate and persist in medullary tissue. He was not able to follow the final evolution of the cells within the medullary tissue. It is fascinating to note that today, almost 100 years later, many of the same questions about the ultimate fate of the hypertrophic cell remain. With the onset of the identification of type X collagen deposition within the hypertrophic zone matrix, and with definition by the ultrastructure of persisting organelles in the hypertrophic cells, the question has again been raised as to whether the swelling of the cells does or does not represent terminal degeneration. This phase is best referred to today as apoptosis, by which is meant genetically determined or programmed cell death. Whereas there appears to be little doubt that some and indeed most of the hypertrophic cells either die or are destroyed by vascular invasion from below, the question remains as to whether some of the cells survive and go on to contribute both bone and perhaps even marrow elements to the metaphyseal region. In works by Gerstenfeld and Shapiro (126) and by Roach et al. (297), the opinions of current cell biologists as to possible lines of cell fate in this region were reviewed. (See also Section VIIC for a discussion of the role of apoptosis and possible differentiation paths of hypertrophic chondrocytes.)

VII. STRUCTURAL DEVELOPMENT OF THE EPIPHYSEAL REGIONS INCLUDING THE JOINTS, THE METAPHYSES, AND THE DIAPHYSES Series of histologic sections and diagrams outlining the cell and tissue development in long bone formation have characterized descriptions of bone formation since the late nineteenth century. These include the works of Mathias Duval, Schafer (309), Prenant, Maillard, and Bouin (278), Renaut (289), Testut (341), Payton (271,272), Maximow and Bloom (227), Nonidez and Windle (246), Dubreuil and Baudrimont (85), Ham and Cormack (142), and Krstic (199), as well as many others. Figure 7A outlines an early yet accurate pres-

21

entation of the dynamic aspects of the developmental sequence derived from M. Duval, including the resorptive as well as synthetic phenomena in establishing bone shape and internal structure. The specific regional shaping mechanisms in long bone growth are outlined in Figs. 7B and 7C and Table IliA. The structural features of long bone development are illustrated by light microscopic photomicrographs in Figs. 8A-8J (limb bud and early endochondral and intramembranous bone formation), Figs. 9A-9L (epiphyseal formation), and Figs. 10A-10E (physeal and metaphyseal structure), and by electron micrographs in Figs. l l A - 1 1 I (physeal structure).

A. Epiphyses Once the primary ossification centers have formed and early endochondral and intramembranous bone formation has proceeded toward either end of the bone, it is at this stage that the entire cartilaginous region at the developing end of each long bone is referred to as the epiphysis. That part of the epiphysis adjacent to the joint is referred to as the articular cartilage. This merges in indistinguishable fashion histologically with the underlying epiphyseal cartilage and only reaches its definitive structure at skeletal maturity when the lowest zone of the articular cartilage calcifies, persisting in the adult as the zone of calcified cartilage. The epiphyseal growth plate is the cytologically and functionally specialized region of the epiphysis responsible for most of the longitudinal growth of a long bone. Descriptions of its cell and matrix composition vary slightly between differing authors due to the mixture of terms, some of which are purely descriptive and others imply function. Table IIIB outlines terminology for the layers of the physis that seems most consistent with current biological understanding and also summarizes common terminology used by others. The cell layer at the outermost extent of the physis is composed of resting or germinal cells, followed by the proliferating or columnar cell layer in which the chondrocytes line up in a longitudinal array, followed by the occurrence of the hypertrophic chondrocytes, and below that the peripheral margins of the metaphysis. The epiphyseal growth plate, which is also referred to as the physis or simply the growth plate, maintains its height during growth as there is a symmetrical occurrence of cell proliferation at the upper margin allowing for longitudinal growth, followed by resorption in the hypertrophic chondrocyte region allowing for the formation of metaphyseal bone and marrow.

B. Secondary Ossification Center Formation Cartilage canals that contain blood vessels are present within the epiphyseal cartilage from the early stages of epiphyseal development. At specific times, the chondrocytes within the central region of the epiphyseal cartilage undergo hypertrophy

22

CHAPTER 1 ~

Developmental Bone Biolo~ty

F I G U R E 7 Patterns and mechanisms underlying long bone growth are outlined. (A) These detailed representations illustrate nine stages of growth and development. Number 1 shows the cartilage model of the developing bone. Characteristic changes can then be seen progressing toward number 9. Illustrations from number 4 on show periosteal bone indicated by the vertical lines and endochondral bone by the horizontal lines. The dense black regions of the cortex in 7-9 represent compact cortical lamellar bone. In illustration number 6 the proximal secondary ossification center is shown, whereas in number 7 both are present. The dotted curved lines in illustrations 7-9 represent original cortical areas that have subsequently been resorbed to allow for marrow cavity formation. [Derived from Mathias Duval, modified from Prenant, Maillard and Bouin (278)]. (B) Regional shaping mechanisms in long bone growth. (C) Contributions of synthesis and resorption to long bone growth.

SECTION VII ~ Structural Development of the Epiphyseal Regions TABLE IlIA (i) Articular-epiphyseal cartilage complex (ii) Growth plate (iii) Diaphyseal cortex

(iv) Perichondrial groove of Ranvier

23

Interplay b e t w e e n Synthesis and Resorption in Long Bone Development Synthesis-expansion by interstitial cartilage growth; resorption of bone and cartilage in secondary ossification center (to shape epiphyseal regions and articular cartilage surfaces) Synthesis-expansion germinal to hypertrophic zones; resorption at lower hypertrophic zone-upper metaphysis (to maintain same growth plate thickness throughout most of the growth period) New bone formation by apposition on outer side by periosteum; resorption on inner side by osteoclasts (to maintain same relative extent of cortex throughout growth while allowing for progressive expansion of diameter of marrow cavity) Widening of physeal cartilage and farthest extent of periosteal new bone formation; resorption at lowermost groove region to remove boney ring (allows for metaphyseal funnelization)

TABLE IIIB

Structural Terms for the Epiphyseai G r o w t h Plate or Physis

Descriptive terms

1. 2. 3. 4.

Functional terms

1. Germinal zone 2. Columnar cell zone (a) Upper proliferating zone (b) Lower maturation zone 3. Hypertrophic cell zone (a) Upper part (4/5), nonmineralized matrix

Resting cell zone Columnar cell zone Hypertrophic cell zone Metaphyseal bone

(b) Lower part (1/5), mineralized matrix 4. Metaphysis outer reaches

] Zone of provisional calcification J

Ham and Cormack (142)

Zone of resting cartilage, zone of young proliferating cartilage, zone of maturing cartilage, zone of calcifying cartilage, developing trabeculae of metaphysis

Brighton (36)

Reserve zone, proliferative zone, zone of maturation, zone of degeneration hypertrophic zone, zone of provisional calcification

in association with mineralization of the cartilage matrix by a process felt to be entirely analogous to that occurring at the physeal-metaphyseal junction (182). This is followed shortly by vascular invasion of the hypertrophic region from the adjacent cartilage canals, lysis of the hypertrophic chondrocytes, and synthesis of new bone on the calcified cores of cartilage by osteoblasts, which accompany the vessels. The initial positioning of the hypertrophic cells is uniform, forming a 360 ~ continuous arc of cells. As development proceeds, this circumferential arc of chondrocytes assumes a hemispheric shape with the hypertrophic chondrocytes forming in relation to the more peripheral articular surfaces, whereas the cartilage closest to the epiphyseal growth plate stops growing and the cells assume the appearance of resting chondrocytes. The hypertrophic chondrocytes and the proliferating cells adjacent to them form what we refer to as the physis of the secondary ossification center. As growth continues the size of the cartilaginous epiphysis increases, but there is a relative increase in the amount of secondary ossi-

fication center bone formation at the expense of the epiphyseal cartilage. For each epiphysis a point is reached in which the only cartilage persisting during the remainder of growth is the physis, the articular cartilage, and its under surface, which is referred to by some as the miniplate. By this stage, the secondary ossification center has almost completely replaced the epiphyseal cartilage. We prefer the term physis of the secondary ossification center to miniplate; it is more awkward but can be used to describe this specific region throughout the developmental cycle. The cartilage region closest to the main physis no longer undergoes proliferation, and the secondary ossification center shows evidence of growth only adjacent to the side walls of the epiphysis and to the undersurface of the articular cartilage. This is also reflected by changes in the structure of the secondary ossification center. At those regions immediately adjacent to the hypertrophic cell areas there is continuing development of the endochondral sequence, with new bone synthesized on cartilage cores. At that part of the

24

CHAPTER

1 ~

Developmental Bone Biology

SECTION VII ~ Structural Development of the Epiphyseal Regions secondary center adjacent to the physes where cell proliferation is no longer occurring, bone is present almost exclusively with minimal to no traces of persisting cartilage cores. In association with this change in orientation of the physes of the secondary ossification center and its hypertrophic cell region there is also a change in the marrow pattern. That marrow immediately adjacent to the hypertrophic zone always remains hematopoietic, whereas the marrow in the regions closer to the epiphyseal bone plate demonstrates the area of earliest fatty change (94). Toward the end of active epiphyseal growth there is a slowing down of the function of the physes of the secondary ossification center. Bone formation begins to predominate over the central cartilage cores because bone synthesis from the marrow cells continues. At skeletal maturation, the lowest level of the articular cartilage becomes calcified and this layer persists through adult life. A subchondral bone plate is formed. Virtually all the marrow is now fatty. The epiphyseal growth plate itself stops its proliferative function, is gradually resorbed from the metaphyseal side, and eventually completely disappears allowing for continuity of the epiphyseal and metaphyseal bone marrow and circulation. The peripheral perichondrial ossification groove is an integral part of the epiphysis. The staging classification, including all structural aspects of epiphyseal development as outlined by Shapiro and Rivas (324), is listed in Table IV. A more detailed report will be published shortly (Rivas and Shapiro, Bone Joint Surgery [Am], in press).

C. Physis--Structure and Relation to Function The cells and matrices of the epiphyseal growth plate have been well-defined at the light microscopic and ultrastructural levels (Figs. 10-12). The resting or germinal cell layer chondrocytes are round to oval in shape and possess a central nucleus and a cytoplasm that is filled with abundant amounts of rough endoplasmic reticulum. The endoplasmic reticulum

25

is mildly to moderately dilated, a finding considered to be consistent with the synthesis of protein, which in this instance is primarily type II collagen. There is also a welldeveloped Golgi apparatus that is responsible for processing the proteoglycans. The chondrocytes of the columnar or proliferating zone are markedly flattened and align themselves into columns. There is a marked tendency for these chondrocytes to be wedge-shaped with the bases of the wedges alternating to allow the linear conformation to persist. The cells are also quite active in terms of mitoses in the proliferating upper part of the zone and in relation to protein synthesis, showing abundant amounts of dilated rough endoplasmic reticulum in the lower or maturation part of this zone. In the hypertrophic zone a characteristic cell appearance is seen with the cell volume larger than that in the above lying regions. The nucleus persists but there are many clear spaces within the cytoplasm. There is an apparent marked diminution of rough endoplasmic reticulum, and by the time one examines cells at the lower margins of the hypertrophic zone only scattered cisternae of endoplasmic reticulum persist. In the upper parts of the hypertrophic zone the cartilage matrix is not calcified. Calcification occurs in the matrix of the lower margins of the hypertrophic zone almost exclusively involving the longitudinal septae along the long axis of the bone, whereas the transverse septae are either not mineralized or only slightly mineralized. The vascular invasion from below progresses two or three cells deep into the lower margin of the hypertrophic zone, and in association with this vascular invasion of the hypertrophic cell regions there is deposition of bone by osteoblasts on the persisting calcified cartilage matrix. The sequence of events is illustrated in the adjacent figures and figure legends. Each level of the physis gradually merges with the level below; there are no sharp transitions. In addition, transitions of cell size, shape, and function occur sequentially even within each subregion. The terminology used by Buckwalter et al. (43) to describe the physis is the reserve (germinal) zone, upper and

FIGURE 8 The histologic features of the developmental sequence of the limb. These photomicrographs are from the developing limb of the rabbit, although the histologic appearance in the human is the same. (A) The developing limb bud is shown at right along with the spinal cord at upper left. At this stage the limb bud is composedof undifferentiated,closely packed mesenchymalcells with no histologic evidenceof any structure. The densely packed cells of the apical ectodermalridge are shown at far right (arrow). (B) A higher power view of the tip of the limb bud showing the mesenchymalcells (M) and the apical ectodermalridge (arrow). (C) A major segment of the developing lower limb is seen at a slightly later stage. The developingfemur (F) and tibia (T) are shown, as is the quadriceps (Q) muscle mass. The central region of the femur shows the early changes of chondrocyte hypertrophy. These bones initially were formed completely in cartilage. (D) A slightly higher power view of a developing long bone, a metatarsal,shows the cartilagemodel and central regions of cell hypertrophy. (E) A higher power view of the femur shown in part (C) highlights the hypertrophic region. There is also development of the intramembranousbone mechanism at the primary center of ossification (arrow). The popliteal artery is shown at lower left. Quadriceps muscle fibers are shown above (Q). (F) The central hypertrophic cells of the femur are shown. The outer fibrous (F) and inner osteogenic (O) layers of the developing periosteum are seen. (G) The interzone region of the developingjoint is seen at right. The cartilage models of the femur (F) and tibia (T) are shown. Cells completely fill the region between these two bone models, indicating the relativelylate time of formation of the joint. (H) A higher power view shows the dense cell accumulationwithin the interzone region. (I) The femoral cartilage is shown at left and the tibial at right. The beginning outline of density of the articular cartilage surface of the tibia is seen (arrow). (J) At a slightly later stage of developmentthe formation of the intramembranousbone sequence overlying the central endochondral sequence is seen. Endochondral bone formation is seen below and intramembranousbone above. Hypertrophic cells of the central endochondral sequence are seen to persist (white arrow). The outer fibrous layer and the inner osteogenic layer of the periosteumare marked. Surface osteoblasts overlyingnewly synthesizedcortical bone are seen (curved dark arrow).

F I G U R E 9 Some characteristic developmental features of epiphyseal formation are shown. Patterns in humans are similar to those seen in pig and rabbit. (A) The cartilaginous epiphysis at the developing end of a rabbit phalangeal bone is shown. At this stage, there is no formation of the secondary ossification center. The articular cartilage (black arrow), epiphyseal cartilage (EC), and physeal cartilage (P) are shown, as is the perichondrial groove of Ranvier (white arrow). (B) Cartilage canals containing vessels have long been recognized as an integral part of epiphyseal development. They are responsible initially for nutrition of the cartilage and only later, often several months later, are they involved in the formation of the secondary ossification center. They are shown here within the cartilage at either

F I G U R E 9 (continued) end of the developing end of human bone from an illustration by Koelliker (193) over 150 years ago. (C) The cartilage canals (Ci-Civ) from a developing proximal femoral capital epiphysis in the newborn pig are shown. [Parts Ci, Civ reprinted from Jaramillo, D., et al. (1996). Am. J. Roentgenol. 166: 879-887, with permission from the American Journal of Roentgenology.] They are never present in the articular cartilage but often traverse the physis in the fetal and early postnatal time periods. Vessels are present within a loosely packed connective tissue matrix. In parts (Cv) and (Cvi), Watermann depicts the vessel contents of a larger (Cv) and a smaller (Cvi) canal. The striated structures are arterioles, the dark solid structures are venules, and the smaller structures are capillaries. [Reprinted from Watermann, R. (1966). Zeit. f. Orthop. 101: 247-257, Georg Thieme Verlag, with permission.] (D)Central cartilage hypertrophy is seen in early formation of the secondary ossification center from a rabbit proximal humerus. Adjacent cartilage canals are shown. Articular cartilage is at top, whereas physeal (arrow) and metaphyseal tissues are at the bottom of the photograph. (E, F) Cartilage matrix mineralization followed by vascular invasion is seen in higher power photomicrographs from the same histologic section as part (D). The vascular invasion has extended from the cartilage canals. The matrix mineral is not seen because the specimens have been decalcified for better sectioning. (G) Endochondral bone formation is now seen centrally in the proximal humeral epiphysis with new bone synthesized on cartilage cores and early evidence of marrow cavitation. The hypertrophic cells of the secondary ossification center are oriented in a 360 ~ arc. (H) Early developing secondary ossification center showing a vessel at lower fight passing into the center of the region. (I) The secondary ossification center has increased in size in relation to the epiphyseal cartilage. At this stage the orientation of the hypertrophic chondrocytes of the physis of the secondary ossification center, miniplate in some terminologies,

F I G U R E 9 (continued) has changed from the 360 ~ orientation shown in part (G) to an approximately 180 ~ orientation here. The two arrows illustrate where the cell hypertrophy of this region ends, with the area between the arrows no longer showing such changes. (J) At a slightly later stage of development, the secondary ossification center moves progressively toward the articular cartilage surface. There is considerable cartilage, however, between the surface of the articular cartilage and the hypertrophic zone of the physis of the secondary ossification center. Well-formed bone with osteocytes is present now within the secondary ossification center (arrow) with only small persisting central cores of cartilage seen. The marrow (M) is hematopoietic. (K) Physeal closure shows the disappearance of proliferating and hypertrophic zones and beginning transphyseal vessel communication (arrows). A continuous bone plate of mature lamellar bone is seen at the top, indicating complete replacement of the epiphyseal cartilage. (L) Epiphyseal development classification of Shapiro and Rivas (324). [From Rivas and Shapiro, J. Bone Joint Surgery (Am), in press, with permission.]

SECTION VII ~ Structural Development of the Epiphyseal Regions

lower proliferating zones, and upper and lower hypertrophic zones. Their histomorphometric studies of proximal tibial physes in mice of varying ages show that, in transverse sections, cell profiles do not change within the same growth plate zones, but in longitudinal sections the cell profiles and profile orientations differ significantly among zones. Cell profiles in the upper and lower proliferative zones are eccentric and highly oriented, but they become more rounded and the degree of cell orientation decreases between the proliferative and hypertrophic zones. The degree of cell profile orientation decreases extensively in the upper and lower proliferative zones, decreases less in the reserve and upper hypertrophic zones, and remains unchanged in the lower hypertrophic zones at varying ages. Changes in cell profiles and in the degree of proliferative zone cell profile orientation correlate with the rate of longitudinal bone growth. Chondrocytes prior to the appearance of a growth plate have rounded profiles with no apparent orientation. Questions remain as to what mechanisms give the cells of the proliferative zone their eccentric shape and high degree of orientation. A correlation is noted between the degree of chondrocyte flattening and the rate of proliferation, with flatter cells proliferating more rapidly and less flattened cells decreasing their rate of proliferation. Others have suggested that the cell is flattened secondarily by the accumulation of matrix. Hypertrophic zone cells have a less eccentric shape and rarely if ever divide, although they do enlarge. Their height in particular is increased in relation to their width, which appears to be associated with the decreasing rate of DNA synthesis and proliferation. Cell swelling in the lower proliferative and upper hypertrophic zones tends to change the eccentrically shaped highly oriented proliferative cell into the rounded and randomly oriented hypertrophic cell. Their height is increased relative to their width, and it is felt that the growth plate matrix helps restrain transverse expansion. The flattest most highly oriented cells are those with the highest rate of proliferation, whereas development of a more spherical shape and loss of orientation seem to be associated with a decreasing rate of proliferation. A second assessment by Buckwalter et al. studied chondrocyte hypertrophy (44). It is evident that, in the hypertrophic zone, the chondrocytes enlarge and assume a more spherical shape. Morphometric analyses of electron micrographs show that, between the upper proliferative zone and the lower hypertrophic zone, the cells increase their mean volume by more than 500%. As the cells enlarge, the intercellular matrices change. The territorial matrix volume increases but the interterritorial matrix volume decreases. Between the upper proliferative zone and the lower hypertrophic zone, the absolute volume per cell of endoplasmic reticulum, Golgi membranes, and mitochondria increases by 126%, whereas the volume of cytoplasm and nucleoplasm increases by 779%, apparently by the accumulation of water. Chondrocyte enlargement thus involves some increase in organelle synthesis, but the primary mechanism of cell

29

enlargement is cytoplasmic and nuclear swelling. The chondrocytes enlarge primarily by accumulating water. The endoplasmic reticulum (ER) and mitochondria dilate, ribosomes dissociate from the endoplasmic reticulum, and distended fragments of ER or Golgi appear in the cytoplasm. This work, along with biochemical studies that identified type X collagen synthesis localized to the hypertrophic zone (147, 307, 313), again raised the question as to whether all hypertrophic cells were degenerating or whether some, and perhaps a significant number, survived and functioned deep into the physeal region and perhaps beyond. The work reviewed previously in Sections III and IV becomes relevant again. Throughout most of the first half of the twentieth century the hypertrophic chondrocyte was considered to be a degenerating cell at the terminal end of the endochondral sequence. It was recognized that its swelling helped provide longitudinal growth, but that its degeneration coincided with vascular invasion of the hypertrophic cell lacunae, following which bone was synthesized on the cartilage cores to form the metaphyseal trabeculae. Holtrop studied the hypertrophic chondrocyte at the ultrastructural level and concluded that "the results showed ultrastructural preservation of the cells that strongly suggest cell activity from the beginning of enlargement of the cell up to the stage where the lacunae breaks open" (158, 160). She felt that careful morphologic study did not confirm the long accepted view of degeneration or cell death but was, rather, consistent with cell activity. Not only did the hypertrophic chondrocytes appear to play a role in the calcification of the matrix, which was being suggested by others, but she felt that they had the potential to become osteoblasts and osteocytes. Once again, careful microscopy was raising the question not only of the function of the hypertrophic cell but also of its persistence, much as did the work of Retterer and the many authors he referenced in the late nineteenth century. Holtrop also reported on transplantation experiments of 4-week-old mice rib growth plates into leg muscles (159). These experiments were designed to assess the potency of the physeal tissue itself in relation to its presumed passive role at the terminal end of the endochondral sequence where it was resorbed. In the discussion she repeated her previous contention that "hypertrophic cartilage cells are able to transform into osteoblasts and osteocytes." She suggested that the development of endochondral ossification in cartilage transplants in the hypertrophic zone implied a more active role for the hypertrophic cells. A second study demonstrated that the function of the physeal transplant was independent of the host because younger transplants elongated more than older transplants using the same technique. The ultrastructure of the growth plate was described in detail in two subsequent articles (161, 162). Hunziker and Schenk further clarified the functions of the specific cell regions of the growth plate (172, 173, 311). The growth plate is characterized structurally by the resting or reserve zone, proliferation and hypertrophy of chondrocytes,

30

CHAPTER 1 ~

Developmental Bone Biology

F I G U R E 10 Physeal and physeal-metaphyseal junction tissues are detailed in this figure. All specimens except part (E) are from the rabbit. (A) The resting germinal cell layer of the physis is shown. This is not a particularly well-defined layer either by histologic or

SECTION VII ~ S t r u c t u r a l D e v e l o p m e n t o f t h e Epiphyseal R e g i o n s

31

calcification, a n d v a s c u l a r invasion. C e l l s w i t h i n the resting

b o r d e r b e t w e e n p r o l i f e r a t i n g a n d h y p e r t r o p h i c zones. Dis-

z o n e i m m e d i a t e l y a d j a c e n t to the e p i p h y s e a l b o n e o c c u r sing l y or in g r o u p s o f two, are d i s t r i b u t e d r a n d o m l y w i t h i n the matrix, a n d l a c k a c o l u m n a r a r r a n g e m e n t . K e m b e r has s h o w n these cells to h a v e s t e m cell f u n c t i o n a n d to u n d e r g o

t i n c t i o n is m a d e b e t w e e n the u p p e r h y p e r t r o p h i c zone, w h e r e cell e n l a r g e m e n t b e g i n s , a n d the l o w e r h y p e r t r o p h i c zone, w h e r e m o s t c h o n d r o c y t e s h a v e attained final size. T h e hypert r o p h i c c h o n d r o c y t e s are no l o n g e r c o n s i d e r e d to be d e g e n -

d i v i s i o n o n l y rarely. C e l l s a d j a c e n t to the tip o f the c o l u m n s

erate cells (69). O n the basis o f m o r e e x a c t i n g p r e p a r a t i o n

b e l o w are c o n s i d e r e d as stem cells, w h i c h f e e d d a u g h t e r cells

techniques, they m a i n t a i n a r o u g h e n d o p l a s m i c r e t i c u l u m and

into the p r o l i f e r a t i n g pool. T h e p r o l i f e r a t i n g z o n e itself has

are felt to f u n c t i o n e v e n to the l o w e s t m a r g i n s o f the zone.

a c l e a r - c u t c o l u m n a r a r r a n g e m e n t . P a r a l l e l to the l o n g axis

M i n e r a l i z a t i o n o f the m a t r i x in the l o w e r part o f the h y p e r -

o f the b o n e c h o n d r o c y t e s d i v i d e in a t r a n s v e r s e d i r e c t i o n

t r o p h i c z o n e is r e s t r i c t e d to the l o n g i t u d i n a l septae, w h i c h is

w i t h d a u g h t e r cells initially situated side b y side. T h e y then

s o m e t i m e s r e f e r r e d to as the " i n t e r t e r r i t o r i a l m a t r i x . " T h e

s w i t c h into a c o l u m n a r t r a n s v e r s e r e l a t i o n s h i p at w h i c h time

territorial m a t r i x i n v o l v e s c o l l a g e n fibrils that run trans-

m a t r i x s y n t h e s i s occurs. C h o n d r o c y t e e n l a r g e m e n t m a r k s the

v e r s e l y a r o u n d e a c h c h o n d r o c y t e . T h e r e g i o n b e y o n d the

F I G U R E 10 (continued) by autoradiographic criteria because it is present in the area at the tip of the arrows between the epiphyseal cartilage and the proliferating cell layers of the physis. It is considered to become or serve as the germinal layer for the well-structured and clearly evident proliferating cell layer of the physis below. (B) Parts (Bi) and (Bii) are presented. The proliferating (P) or columnar cell layer of the physis is shown above; these then merge into the hypertrophic cell layer (H) as the cells increase in size. Cartilage matrix is labeled (M). The metaphysis is shown below. It is recognized increasingly that these various layers merge into one another with changes in cell activity and cell size occurring gradually in a progressive fashion. The proliferating cell zone is recognized in its upper part to undergo proliferation characterized by the high uptake of tritiated thymidine, indicating DNA turnover. In its lower part, it is sometimes referred to as a zone of maturation in the sense that there is little cell turnover, but the synthesis of proteoglycans, collagen, and noncollagenous proteins occurs actively. The hypertrophic zone region also can be subdivided into an upper zone, which is characterized by a progressive increase in cell size that contributes extensively to longitudinal growth of the bone, and a lower hypertrophic zone in which the matrix mineralizes in preparation for eventual vascular invasion of the hypertrophic cell lacunae. This lower layer has been referred to by some as part of the zone of provisional calcification. Metaphyseal bone is formed in association with the advancing front of vascularization. (C) Morphological characteristics at the growth plate cartilage-metaphyseal bone interface are shown in parts (Ci-Cv). Diagrammatic illustration and surrounding high-power photomicrographs are shown. A diagrammatic outline of the major morphological and cellular events at the hypertrophic chondrocyte-metaphyseal interface is presented. RBC, red blood cell; OB, osteoblast; OC, osteocyte; OCL, osteoclast; V, vessel. The major morphological landmarks are depicted in the adjacent micrographs. Area I (upper fight): Photomicrograph illustrates a portion of an un-decalcified growth plate-metaphyseal junction from a 2-week-old rabbit metatarsal. The hypertrophic cells are seen above. Note the dark staining mineral in the longitudinal cartilage septae with very little or no mineral in the transverse septae. Vascular invasion of the hypertrophic cell lacunae has occurred (solid arrow) on one side of a mineralized cartilage trabeculum, whereas on the opposite side there are newly differentiating mesenchymal cells designed to become osteoblasts and synthesize osteoid on the calcified cartilage core. The dots represent tritiated proline in the autoradiograph, with sacrifice performed 20 min after injection. Area II (lower fight): High-power photomicrograph shows the lower regions of the hypertrophic zone at top and the vascular and mesenchymal cell invasion from below. This histologic section and those depicting areas III-V are all demineralized preparations. The hypertrophic cells from the lowest layers of the hypertrophic zone (white arrow) are seen. Red blood cells from the metaphysis are seen adjacent to the last persisting hypertrophic cell lacunae. A multinucleated osteoclast (curved arrow) is seen, as are undifferentiated mesenchymal cells that will shortly begin to synthesize an osteoid matrix on the persisting cores of calcified matrix (open arrow). (Proximal tibial growth plate-metaphyseal junction in 1-month-old rabbit; plastic embedded JB4 section stained with 1% toluidine blue.) Area III (upper left): Photomicrograph of metaphyseal tissue immediately adjacent to the hypertrophic zone of the growth plate from a 1-month-old rabbit proximal tibial metaphysis. The persisting cartilage cores (C) are black. Newly synthesized bone (B) can be seen adjacent to them. Osteoblasts (curved arrow) line the surface of the newly synthesized bone. Osteocytes are within. We refer to these as mixed trabeculae encompassing both bone and cartilage tissue. A vessel (V) is also seen. A multinucleated osteoclast is seen centrally where it is resorbing both cartilage and bone (black arrow). Area IV (lower left): Photomicrograph of tissue deeper within the metaphyseal region. There are areas of persisting cartilage (black) but much more newly synthesized light staining bone. Surface osteoblasts are seen (black arrows), as are osteocytes and osteoclasts (open arrow). Area V (lower middle): A high-power view of metaphyseal bone and cartilage trabeculae is seen. Osteoclasts (arrows) can be seen resorbing both bone and cartilage. (D) Mineralized sections of a rabbit metatarsal epiphysis and adjacent metaphysis are shown in parts (Di-Div) (see next page). Part (Di) shows mineralized longitudinal cartilage septae passing up into the lower reaches of the hypertrophic zone. At far fight the mineralized cortex is also seen. A higher power view of a part of the physeal-metaphyseal junction is shown in (Bii). It is the longitudinal septae that remain mineralized, with the transverse septae for the most part free of mineral. In parts (Diii) and (Div) a tritiated proline autoradiograph of the mineralized physeal-metaphyseal junction is shown. Centrally in (Diii) one can note the red blood cells passing into the hypertrophic cell lacunae following disruption of the transverse and oblique septae. Part (Div) adjacent to the region of part (Diii) shows, at the left side of the central cartilage septum, the proline-labeled osteoblasts, which are lying on the calcified cartilage core and shortly will begin synthesis of osteoid. (E) Photomicrograph from the metaphyseal region of a developing calf long bone. A central cartilage core is seen surrounded by newly formed bone. Young osteocytes are present within the bone matrix. Osteoblasts line the surface of the bone. Immediately adjacent to the osteoblasts one can see a multinucleated osteoclast resorbing bone and ultimately cartilage. The marrow contains two cell lines: one serves as a hematopoietic precursor, including monocytes that fuse to form osteoclasts, and the other cells, called stromal precursors, can differentiate into preosteoblasts and osteoblasts. [Parts Bi, Bii, C,

32

CHAPTER

1 ~

Developmental Bone Biology

F I G U R E 10 (continued) and Diii from Gerstenfeld and Shapiro, J. Cell. Biochem. 62:1-9, copyright 1996. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.]

SECTION Vll ~ Structural Development o f the Epiphyseal Regions

F I G U R E 11 Electron micrographs of the physeal region of the developing mouse are shown. (A) Electron micrograph from the proliferating zone region of newborn mouse physis. The flattened cells are seen above. The rough endoplasmic reticulum is highly active and dilated, indicating synthesis of protein. The Golgi apparatus in each cell can also be seen. (B) Higher power photomicrographs from the proliferating zone are seen. Note the extensive endoplasmic reticulum and Golgi apparatuses. Mitosis is being completed below. Note two nuclei. (C) Two flattened chondrocytes of the proliferating zone are seen. Note the prominent organelles for synthesis. (D) Chondrocytes from the upper hypertrophic zone continue to show well-developed active organelles. (E) Cells farther down the zone of hypertrophy show areas of markedly distended rough endoplasmic reticulum, flattened ER, and increasing areas devoid of organellar definition. (F) Hypertrophic chondrocytes lower in the zone show a distended nucleus at left, a large focus of RER, and areas devoid of specific organellar structure. (G) Only small local collections of RER remain in the lower hypertrophic zone. (I-I) Electron micrograph

33

34

CHAPTER 1 ~ Developmental Bone Biology

F I G U R E 11 (continued) from the hypertrophic region shows mineralization primarily of the longitudinal septae, with little to no mineralization of the immediately adjacent transverse septae. The white arrows represent the long axis of the bone. The cell lacunae show remnants of hypertrophic cell organelles. The space is considered to be filled with water, which is responsible for the hypertrophic cell size increase and its contribution to longitudinal bone growth. (I) Mineralization of the longitudinal septum is shown, as is the distinct absence of mineral from the transverse septum. The hypertrophic chondrocyte in the lowermost region of the physis shows a remnant of the nucleus (N), a dilated fragment of the rough endoplasmic reticulum (R), and flattened segments of rough endoplasmic reticulum (arrows).

territorial matrix thus is in the longitudinal septae. The bulk of the transverse septae remain unmineralized. The fact that the growth plate maintains its same height during most of its growth phase indicates excellent synchronization between vascular invasion and resorption from below and addition by cell proliferation and matrix production in the upper part of the cartilage. Approximately one-third of the proliferating zone is occupied by cells and two-thirds by matrix, a ratio that is reversed in the hypertrophic zone by the tremendous enlargement of chondrocytes. The marked size change during hypertrophy is an almost 10-fold increase in volume. The orientation and shape of the cell also change

from the proliferative zone in which the cell is ellipsoid and transversely oriented to the hypertrophic zone in which it is cylindrical with its long axis oriented in the longitudinal direction. Cell hypertrophy primarily causes a stretching of the cell columns in their axial direction, which thus contributes substantially to longitudinal growth. The hypertrophic cells expand by the act of transport of water and electrolytes through their cell walls. The cells continue to produce matrix. The rough endoplasmic reticulum at first glance appears scanty and dispersed, but considering the 10-fold increase in cellular volume, detailed studies actually show the organelles to be increased in extent. Longitudinal growth rate thus

SECTION VII ~ Structural Development of the Epiphyseal Regions

35

TABLE IV The General Plan of Epiphyseai and Long Bone Development a Stage 1 Stage 2 Stage 3 Stage 3a Stage 3b Stage 4 Stage 4a Stage 5 Stage 5a Stage 6 Stage 6a Stage 7 Stage 8 Stage 9 Stage 10 Stage 11 Stage 12 Stage 13 Stage 13a Stage 14 Stage Stage Stage Stage Stage

15 15a 15b 16 16a

Limb bud formation; uniform mesenchymal cell distribution; apical ectodermal ridge Mesenchymal condensation Cartilage differentiation Interzone formation Chondrocyte hypertrophy in middle part of long bone cartilage model Epiphyseal shaping Primary center of ossification Resorption of joint interzone; smooth articular cartilage surface Vascular invasion of hypertrophic chondrocyte area, mid part cartilage model; endochondral bone Physeal differentiation and peripheral groove tissue formation Farthest relative extent of epiphyseal-physeal position Epiphyseal cartilage vascularization; cartilage canals Central chondrocyte hypertrophy to form spherical mass; development of growth plate completely surrounding secondary ossification center Vascular invasion of developing secondary ossification center hypertrophic chondrocytes adjacent to mineralized cartilage matrix Bone formation and marrow cavitation in secondary ossification center; hematopoeitic marrow Increase in size of secondary ossification center; decrease in epiphyseal cartilage Central chondrocyte hypertrophy-secondary ossification center growth plate change to hemispherical orientation Fat in marrow; hematopoietic marrow adjacent to secondary ossification center growth plate Epiphyseal bone plate formation Fullest relative extent of secondary ossification center involvement in epiphyseal cartilage articular cartilage miniplate formation Thinning of the physis Articular cartilage miniplate growth cessation Subchondral bone plate formation Resorption of physis linking epiphyseal and metaphyseal circulations Calcification of lowest zone of articular cartilage and tidemark formation; all marrow fatty

aFrom Shapiro and Rivas (324) and Rivas and Shapiro, J. Bone Joint Surgery (Am), in press, with permission.

is a function of modulation in cell turnover, fluctuations in cell hypertrophy, and changes in matrix production. Hunziker and Schenk sought to determine which of the many variables contributed most to longitudinal growth. They reviewed the fact that changes in cell proliferation rate, height, volume, and matrix production had each been generally implicated in the process. On the basis of detailed histologic and histomorphometric studies of the rat physis, they concluded that "growth acceleration is achieved almost exclusively by cell shape modeling, namely increase in final cell height and a decrease in lateral diameter." They felt that the cell proliferation rate in the longitudinal direction and net matrix production per cell remained unchanged, such that the cartilage matrix itself appeared to play a subordinate role in regulating longitudinal bone growth rate. Growth thus can

be regulated most acutely by factors acting on cell shape modeling. Hunziker and Schenk noted the frequent observation that the cytoskeleton of the chondrocyte was relatively minimal compared to other cells, which again indicated a diminished role for this organelle in controlling cell shape and size. Their morphologic studies of chondrocytes supported a high degree of coordination between matrix remodeling and chondrocyte shape change. Changes in the shape of the hypertrophic cell can occur much more rapidly than those that require changes in the cell turnover rate. Thus, during both acceleration and deceleration of linear growth, changes in hypertrophic cell activities appear to play an important regulatory role rather than the previously assumed fact that linear growth was modulated principally by changes in chondrocyte proliferation activity. In a careful quantitative

36

CHAPTER

1 9

Developmental Bone Biology

analysis of physeal cell features, they concluded that a proliferating chondrocyte needed approximately 54 hr to duplicate its own volume, whereas during hypertrophy a corresponding volume increase would be achieved in a period as short as 5 hr. Hypertrophy thus appeared to be a more proficient mechanism for bringing about columnar linear growth than did cell proliferation. They also concluded that it was a hypertrophic cell change in shape and volume itself that modulated growth because they could demonstrate that by the end of its life cycle the hypertrophic chondrocyte produced neither an increase nor a decrease in its associated matrix volume. Hunziker and Schenk felt that the matrix had its primary function as a space filling role between cells to compensate for the changes in height, diameter, and volume and thus helped to maintain columnar tissue organization during linear growth. The matrix was also crucial for maintenance of the biomechanical properties of growth plate cartilage (172-174). The duration of the hypertrophic phase at approximately 48 hr also remained remarkably constant irrespective of animal age or growth rate. Hunziker et al. also pointed out that individual physeal chondrocytes actually remained in a fixed location throughout their life during which their function and shape changed (172). The two most prominent stages were those of cellular proliferation and hypertrophy. By the late hypertrophic stage, 4-fold and 10-fold increases in the mean cellular height and volume, respectively, and a 3-fold increase in the mean volume of the matrix synthesized per cell had been achieved. The continuing high metabolic activity of hypertrophic cells was also demonstrated on the basis of a 2- to 5-fold increase in the mean cellular surface area of the rough endoplasmic reticulum, the Golgi membranes, and the mean cellular mitochondrial volume. They also concluded that the hypertrophic chondrocytes had increased their volume by active transportation of fluid (water and electrolytes) across the plasma membrane into the cell itself. The work of Wilsman, Farnum, and colleagues has shown the importance of structural changes in relation to growth plate cartilage function (19, 33, 100). Their work also supports the concept that hypertrophic chondrocytes are fully viable cells that play a major role in endochondral ossification. It has not been determined whether all hypertrophic chondrocytes die or whether some may survive and modulate to bone forming cells in the metaphyseal regions. It is now almost universally accepted that the large majority of hypertrophic cells are viable and functioning in terms of continuing matrix synthesis, with some molecules such as type X collagen synthesized primarily in the hypertrophic cells and involved directly or indirectly in the mineralization process. Terminal hypertrophic chondrocytes can be found in three morphologically distinct forms: (1) fully hydrated cells with direct circumferential attachments of the plasma membrane to the pericellular matrix identical to that of the more proximal hypertrophic chondrocytes; (2) fully hydrated cells with asymmetric attachment to the last transverse septum; and

(3) a condensed cell with the same asymmetrical attachment to the last transverse septum. They hypothesize that "the initial attachment of the chondrocyte plasma membrane to the last transverse septum represents the first of a series of rapid terminal morphologic changes that represent chondrocytic death." Wilsman et al. feel that it is only the terminal chondrocyte that undergoes changes in which there is withdrawal of the plasma membrane from its attachments to the pericellular matrix, cellular condensation, and cellular vacuolization. They indicate that "cellular disintegration is complete before the opening of the chondrocytic lacunae." Hunziker has also suggested that only the terminal chondrocyte in each cell column dies and that it does so in fashion consistent with programmed cell death. There has been a constant, if small, group of individuals studying growth plate function who have hypothesized "that terminal hypertrophic chondrocytes modulate in phenotype and change not only their chemical phenotypical expression but also their ultimate biological role." The terminal chondrocytes are viable and metabolically active. This group does not go so far, however, as to indicate that they then switch to a bone phenotype within the metaphyseal area. Breur et al. demonstrated that chondrocyte enlargement began immediately following cell division in the proliferative zone and that it consisted of two morphologically distinct phases (33). The transition point between the first and second phases of chondrocytic enlargement corresponds with the junction between proliferative and maturation zones. The controlling features in longitudinal growth are. the rate of new cell production in the proliferative zone and the role of chondrocyte hypertrophy. Their study suggested that "in growth plates, chondrocytic enlargement plays a major role in the determination of longitudinal bone growth." Chondrocyte enlargement is characterized by an increase in cell volume and a modulation of cell shape. In the proliferative zone, the cells are flattened ellipsoids, whereas in the hypertrophic zone they are more rounded and spheroid in a longitudinal plane. The increase in cell volume is mainly the result of cellular swelling by absorption of water, with only some increase in cell organelles. Chondrocyte proliferation is felt to take place only in the proximal ends of the proliferative zone, and the function of chondrocytes in the proliferative zone distal to the actively proliferating cells involves actual enlargement and hypertrophy. Each region thus merges with the next rather than being strictly separate as is sometimes indicated. In a study of rat growth plate structure, Breur and colleagues demonstrated that (1) cell volume increase started immediately following cell division in the proximal portion of the proliferative zone, (2) the process of chondrocytic enlargement as it relates to longitudinal bone growth consists of two morphologically distinguishable phases, (3) the rate of cell volume increase and the rate of cell shape modulation are significantly higher during the second than during the first phase, and (4) cell volume increase during the first phase results mainly in an increase in vertical chondrocyte diame-

SECTION VII ~ Structural Development of the Epiphyseal Regions ter, whereas cell volume increase during the second phase results in a large increase in the vertical chondrocyte diameter and a smaller but significant increase in horizontal chondrocytic diameters. The cell volume increase starts immediately following cell division in the proximal portion of the proliferative zone, and chondrocytic enlargement consists of a phase of slow and then rapid cellular enlargement. The complex functions of the hypertrophic chondrocyte during endochondral bone development have been stressed by Gerstenfeld and Shapiro in a work that combines consideration of molecular data with the histologic characteristics of the physeal-metaphyseal junction (126). Endochondral bone formation is one of the most extensively examined developmental sequences within vertebrates. The major cellular events of this process include the recruitment and induction of both osseous and vascular tissues. Presumptive osteoblasts are recruited and line the trabeculae of mineralized cartilage to synthesize osteoid. The term mixed trabeculum refers to the presence of mineralized cartilage cores surrounded by newly synthesized bone. Vascular elements invade and line the empty lacunae of the lowermost hypertrophic chondrocytes, which have undergone cell death. Pertinent questions relative to the osseous and vascular induction within these zones include the nature of the functional coupling among cartilage, bone, and vascular tissues. Signals elaborated by the endochondral cells are targeted to the subsequent development of the chondrocytic components of the growth plate (autocrine regulation) and toward the osteogenic and vascular elements (paracine regulators). Extracellular matrix components of the mineralized growth plate may elaborate signals or be a permissive substrate for osseous and vascular induction. The surrounding and resorption of the mixed trabeculae by osteoclastic cells is the terminal event in the remodeling process of the endochondral tissue. Signals elaborated from the mineralized cartilage-bone trabeculae, as well as the cells lining this trabeculae, are involved in this recruitment process. This process involves the coordinated temporal-spatial differentiation of three separate tissues (cartilage, bone, and the vasculature) into a variety of complex structures. The differentiation of chondrocytes during this process is characterized by a progressive morphological change associated with the eventual hypertrophy of these cells. These cellular morphological changes are coordinated with proliferation, a columnar orientation of the cells, and the expression of unique phenotypic properties, including type X collagen, high levels of bone, liver, and kidney alkaline phosphatase, and mineralization of the cartilage matrix. Several studies indicate that hypertrophic chondrocytes also express osteocalcin, osteopontin, and bone sialoprotein, three proteins that were widely believed to be restricted in their expression to osteoblasts. Other studies suggest that the hypertrophic chondrocytes are regulated by the calcitropic hormones, morphogenic steroids, and local tissue factors. The considerations are based on the regulation by 1,25-(OH)zD3 and the retinoids of the cartilage-specific

37

genes as well as osteopontin and osteocalcin expression in hypertrophic chondrocytes. Studies further suggest that specific transcription factors mediate exogenous regulatory signals in a coordinated manner with the development of bone. Whereas it has been demonstrated to the satisfaction of some that the majority of hypertrophic chondrocytes undergo apoptosis during terminal stages of the developmental sequence, their response to specific exogenous regulatory signals and their expression of bone-specific proteins give rise to questions about whether all growth chondrocytes have the same developmental fates and identical functions. Aizawa et al. demonstrated apoptosis in hypertrophic growth plate chondrocytes in the rabbit using specific antibody staining techniques for TUNEL [terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate (dUTP)-biotin nick end labeling] (2). Gibson reviewed the active role of chondrocyte apoptosis in endochondral ossification (128). The physiological form of cell death appears not to be just a passive removal of unwanted cells but rather an event that plays an active role in initiating processes that follow cell death. The possible pivotal role of chondrocyte apoptosis is discussed for each calcification of the growth plate longitudinal septae, cartilage resorption, growth factor release, maintenance of growth plate height, and activation of vascular invasion and bone formation at the lower part of the endochondral sequence. Specific questions arise as to whether there are similar mechanisms or regulation for commonly expressed genes found in both cartilage and bone or whether these genes have unique regulatory mechanisms in these different tissues. These findings suggest that hypertrophic chondrocytes are functionally coupled during endochondral bone formation to the recruitment of osteoblasts, vascular cells, and osteoclasts. Over 100 years ago some observers, using light microscopy, continued to maintain that at least some hypertrophic chondrocytes survived and differentiated to bone forming cells in the metaphysis (see Section IV). The ability of hypertrophic chondrocytes to synthesize specific matrix molecules now is well-accepted, as is the fact that many or most of the hypertrophic chondrocytes subsequently undergo apoptosis. Surprisingly, perhaps the question of whether some hypertrophic chondrocytes do differentiate to an osteogenic line is still supported by some, now based on molecular studies. Roach et al. in a study in vitro of embryonic chick bones, noted asymmetric cell division in hypertrophic chondrocytes with diverging fates of two daughter cells, one undergoing death by apoptosis and the other surviving, dividing, and generating osteogenic cells. Their work also referred to differing views on the matter published over the previous three decades. The crucial event in differentiation to a bone line was the asymmetric division noted. This division resulted in one viable and one apoptotic cell, following which the viable cell reentered the cell cycle and gradually differentiated to the bone forming line.

38

C H A P T E R 1 ~ Developmental Bone Biology

F I G U R E 12 The cell and matrix characteristics of the perichondrial ossification groove of Ranvier are illustrated. (A) A light micrograph from the developing groove of the distal femur of the rabbit is shown. The physeal cartilage is shown at left. Three cell populations within the groove are the region of densely packed cells at the depth of the groove, the region of less densely packed cells just above it, and the outer layer of fibrous cells. The region of densely packed cells is the terminal extension of the inner osteogenic layer of the periosteum. It secretes an osteoid matrix, which ensheathes the physis and shortly begins to synthesize the intramembranous bone referred to as the bony ring or bony bark of the ossification groove. The fibrous layer is the continuation of the outer fibrous layer of the periosteum. [Reprinted from Shapiro et al. (1977). J. Bone Joint Surgery 59A:703-723, with permission.] (B) The most terminal extension of the region of densely packed cells within the depth of the groove is shown. Note the curvilinear orientation of the cells. There is virtually no matrix seen between them. These cells do not add cartilage to the transverse diameter of the epiphyseal or physeal cartilage, but rather serve as bone forming cells. (C) The fact that the cells densely packed have been differentiated along a bone

SECTION VII 9 Structural Development of the Epiphyseai Regions

D. Perichondrial Ossification Groove of Ranvier The endochondral and intramembranous bone formation systems always merge at the periphery of a developing bone in a specific series of tissue conformations referred to as the perichondrial groove of Ranvier (204, 205, 208, 285, 320, 321, 331). As we noted previously, the developmental sequence of the periosteal sleeve is always spatially and temporally slightly in advance of that of the contained endochondral sequence. This relationship persists even when the epiphyseal growth plate has established its definitive relative position at either end of the bone. The epiphyseal growth plate thus is surrounded by periosteal tissues, which serve as a structurally supportive mechanism. Three cell populations have been defined in the depths of the groove (320). The groove refers to the circumferential depression in the periphery of the epiphyseal growth plate, which has its deepest extent opposite the resting or germinal cell layer and the adjacent epiphyseal cartilage. The outer fibrous layer of the periosteum continues surrounding the epiphyseal growth plate and inserting above the deepest part of the groove of Ranvier into the cartilage of the epiphysis. The inner osteogenic cell layer of the periosteum is also continuous, serving to surround the epiphyseal growth plate and commencing as an accumulation of densely packed cells in the depth of the groove adjacent to the germinal and proliferating cell layers of the physis. At their deepest part, these cells are undifferentiated mesenchymal cells that quickly differentiate to preosteoblasts and synthesize a characteristic osteoid matrix, which shortly is mineralized to form actual intramembranous bone. These spicules of intramembranous bone surround the epiphyseal growth plate well into the level of the proliferating zone. Although they are not sufficiently wide to be seen radiographically, they are clearly present at a histological level. The bone ring is sometimes referred to as the perichondrial bony bark or bony ring of Lacroix. The third region of cells specific to the groove of Ranvier is undifferentiated cells that are more loosely packed and that lie between the outer fibrous layer and the region of densely packed cells

39

that are osteoblast precursors. These cells appear to add to the periphery of the epiphyseal cartilage and therefore serve as chondrocyte progenitors. The cells are considered to be responsible for the increase in width of the epiphysis. Once deposited in the epiphyseal cartilage with growth, they are encompassed within the epiphyseal growth plate and then pass through the developmental stages from germinal to proliferating to hypertrophic cells. An integral part of development of the periphyseal region is the cut-back zone, which is responsible for the shaping of the metaphysis, a process sometimes referred to as funnelization. As this region is narrower than that immediately above it, it is evident that osteoclastic resorption must have occurred to remove the excess tissue. Histologic slides reveal abundant amounts of these osteoclasts, which are responsible for the funnelization and shaping of the developing bone in the metaphyseal regions. In the large majority of instances, the bone bark is resorbed at the metaphyseal level along with adjacent metaphyseal endochondral bone, serving to leave the circumferential bone ring as a structure anatomically separate from the cortex of the metaphyseal and diaphyseal regions. The outer fibrous layer of the periosteum, however, always remains intact. Burkus and Ogden described the perichondrial ossification groove region of the human distal femur from the fetal time period to skeletal maturation at 16 years (48). The cell regional structure including the bony ring was most evident in the young fetus. They felt that appositional chondrocyte growth was most evident during the first 5 months of gestation, after which its activity diminished. The structural characteristics of the perichondrial groove region are illustrated in Figs. 12A-12H, Figs. 13A-13C, and Figs. 14A, 14B.

E. Periosteum and Its Relationship to the Epiphyses, Metaphyses, and Diaphyses The periosteum is widely recognized to play a major role in cortical bone formation by the intramembranous mechanism.

F I G U R E 12 (continued) forming line is shown by this alkaline phosphatase histochemical stain. The adjacent physis is seen at right. The region of densely packed cells is well-outlined by the dark staining alkaline phosphatase at left. (D) Tritiated thymidine autoradiography outlines the high cell turnover activity in the region of densely packed cells. Uptake initially is concentrated in the upper end and also along the outer margins, which corresponds to the presence of the rapidly dividing nondifferentiated cells. The physeal cartilage is at left. (E) Another tritiated thymidine autoradiograph shows extensive uptake in the region of densely packed cells. One also notes uptake in the proliferating layer of the physis, indicating interstitial growth there, uptake in the germinal zone of the physis that would serve to add more columns at the periphery and allow for interstitial widening of the physis, and also a cell right at the point of juncture of the epiphyseal cartilage and the region of less densely packed cells, indicating that this region contributes to transverse growth of the epiphyseal cartilage and ultimately of the physis by adding chondrocytes at the periphery. (F) Passage of the outer fibrous layer into the epiphyseal cartilage (arrows) beyond the physeal cartilage is shown. This is where the periosteum is firmly attached to the developing bone. The region of less densely packed cells (i, curved arrow) is seen between the densely packed cells below and the outer fibrous layer above. (G) The groove components closer toward the metaphysis show increasing osteoid synthesis by osteoblasts (arrow). These cells synthesize the bony ring. (H) The bony ring of the groove (arrows) is shown adjacent to the metaphysis. It is an example of direct intramembranous bone synthesis. The outer fibrous layer of the periosteum is shown at left. At the metaphyseal region, there is discontinuity between the bony ring of the ossification groove and the cortex of the metaphyseal-diaphyseal area due to resorption by multinucleated osteoclasts. [Reprinted from Shapiro et al. (1977). J. Bone Joint Surgery 59A:703-723, with permission.]

CHAPTER 1 9 Developmental Bone Biology

40

c

A

io f

"oE

\ 9. . . . . . .

2o

/

.

...... z :. :. .2. . .... .. .. .. .

k a

.!i

I

.o ~,, M tli _~', i~l 1~!1 lit~i1 / / ~ o ~

L ~ ..........~'/"

v~ ~,,~ 0h I 24

I 16

I 24

I 48

I 72

~!] ~ ~t~50~" r3~ .~ " - - ~ 0~"

HOURS AFTER INJECTION

Bii

Bi

CELL SPECIALIZATION I. FIBROUS CONNECTIVETISSUE

~

0 (~ ~:9

RANVIER'S GROOVE

GROWTH PLATE

CELL ACCUMULATION

- encoche d'ossification

~

Ilk

I~~

.

i/!, , ~ ~

~

t

GROWTHJ_.__.~ DIRE

OFF

i I ~0

oo,_ 2 T "R A INGLEA R E A

- commencement of perichondrium

5. PERIOSTEAL PROGENITORREGIOI~ - densecel, occumulotion - + olkoline phosphotose

- high :SH-thymidineuptoke

OSTEOID PERICHONDRIALBONY RING OF OSSIFICATION GROOVE (LACROIX) or

NAPKIN RING OF RANVIER - ecorce

F I G U R E 13

~

BONE METAPHYSIS

OSTEOCLAST RESORPTION

(A) Tritiated thymidine labeling of progenitor cells in the region of densely packed cells is illustrated. The study assessed the number of labeled cells and their presence from 2 to 72 hr after injection. Four comparably sized regions of the densely packed cells were measured with area 1 being at the uppermost part and area 4 along the metaphyseal region. At 2 - 4 hr, most of the cells are concentrated in area 1, and they progressively appear in area 4 with time as a reflection of their presence within the groove, which itself is growing upward and outward. [Reprinted from Shapiro et al. (1977). J. Bone Joint Surg. 59A:703-723, with permission.] (B) The contributions of the various cell populations to growth at the peripheral regions of the physis are shown in parts (Bi) and (Bii). [Part Bi reprinted from "Skeletal Growth and Development," (J.A. Buckwalter et al., eds.), Ch. 28, Fig. 2B, copyright 1998 by the American Academy of Orthopedic Surgeons, with permission.] (C) Illustration showing the relative changes of position of the perichondrial ossification groove and in particular the bony ring bark with growth. The bony ring tends to be positioned in an outward and upward direction with growth. New bone is added at the top, and bone is resorbed from the bottom at the metaphyseal cut-back zone. The physis would be to the left.

SECTION VII 9 Structural Development of the Epiphyseal Regions It is composed of an outer fibrous layer and an inner osteogenic or cambial layer. Less widely appreciated, however, is the fact that the periosteum also has a major support role in relation to stability at the physeal-metaphyseal junction as well as a role in applying appropriate tensile forces to the physis during the growing years. The outer fibrous layer of the periosteum passes beyond the physis and attaches into the epiphyseal cartilage. It serves as a continuous layer from the epiphyseal cartilage of the proximal end of a long bone to the epiphyseal cartilage of the distal end of that bone. The periosteum is quite loosely attached to the underlying cortical bone in the developing child. The reasons for this relate to the differing growth rates within the bone itself and within the periosteum. The bone grows by apposition of tissue at either end, but the periosteum has been shown to grow uniformly throughout its length by interstitial cell mechanisms (364). In addition, one end of any long bone grows more rapidly than the opposite end. The loose attachment of the periosteum to the underlying bone enables the differential growth mechanisms to occur simultaneously without difficulty. The interstitial growth of the periosteum also serves to maintain the relationship of the muscle attachments to the periosteum, an occurrence that would be much more difficult if the periosteum itself grew only at its proximal and distal ends. The outer fibrous layer of the periosteum is continuous from epiphyseal cartilage to epiphyseal cartilage, whereas the inner osteogenic layer often is discontinuous at the region of the metaphyseal cut-back zone particularly where this zone is quite angled. The periosteum is firmly adherent to the growing bone at either epiphyseal end. Lacroix indicates that the only area between these regions in which periosteal elongation and bone elongation are the same is at the so-called "null point" of periosteal growth, which is farthest away from the most active growth plate and nearest to the least active growth plate (204). This would occur, for example, in the tibia at about 35% of the tibial length above the growth plate because only 35% of tibial growth occurs at the distal end of the bone. The extrinsic support that the periosteum provides for the growth plate at the periphery of the groove of Ranvier region is considerable. John Poland, in his classic treatise on epiphyseal growth plate fractures, reports an experiment by John Wilson in the 1820s in which weights were applied to anatomic specimens of human distal childhood femurs (273). When the circumferential periosteal tissues were removed from the growth plate region, the amount of weight required to dislodge the epiphysis from the metaphysis was only one-fifth as great as when the tissues were intact. Considerable structural support is provided by the periosteal and perichondrial tissues. Amamilo and associates also showed that a consistently higher force was needed in rats to produce epiphyseal displacement with the periosteum intact (5). Alexander (4) documented the occurrence of distal radial epiphyseal fractures at times of most rapid growth and im-

41

plicated changing mechanical features of the open physes at different ages. The muscles and tendons are attached directly to periosteum in the growing child rather than to the underlying cortical bone. There is a distinct change in adults, however, in whom the periosteum is much thinner, is firmly adherent to the underlying cortex, and demonstrates muscle and tendon fibrils that pass through it to gain direct attachment to the underlying cortex by Sharpey's fibers. It has been postulated that there is a strong fibroelastic periosteal sleeve effect on the physis that not only applies a certain degree of tension across it but may serve as a check to unconstrained longitudinal growth. It has long been recognized and continues to be shown that circumferential division of the periosteal sleeve, especially if it is performed close to the metaphyseal-epiphyseal regions, will allow for increased longitudinal growth of those bones (61, 80, 151). What is unclear, however, is whether the increased growth is due to the diminution of mechanical constraint during the time that the periosteal sleeve is discontinuous or due to an increase in vascularity in the peri-epiphyseal region that occurs consequent to injury and during the repair phase. The absence of overgrowth when longitudinal cuts were made in the periosteum though supports the mechanical effects (80). A medial hemicircumferential division of the proximal tibial periosteum leads to medial overgrowth and valgus deformation (64, 164). When periosteal removal was done circumferentially in 4-mm-wide strips, in the mid-diaphyseal region of 4-week-old rats, overgrowth was seen but it was minimal: only 1.5% greater than the opposite side (119). Haasbeek et al. have shown that, when periosteum is thickened adjacent to a physis, it serves as a tether to cause angular deformity (134). They demonstrated the phenomenon in two clinical cases and experimentally. In summary, the periosteum is shown to affect growth of the physes mechanically because it ensheathes the physes and inserts beyond them into the epiphyseal cartilage. When periosteal tension is reduced the longitudinal bone growth is increased, and when the tension is increased growth slows slightly.

F. Cortical (Diaphyseal) Bone Formation-Woven Bone and Lamellar Bone The histologic appearance underlying cortical (diaphyseal) bone formation is well-understood. In fact, the book Osteologia Nova or Some New Observations of the Bones by Clopton Havers published in London in 1691, in which he described longitudinal pores passing from one end of the bone to the other, is widely considered to represent the initial scientific description of bone (145). These longitudinal passages have come to be referred to as Haversian canals, even though Havers did not describe blood vessels within them. He also described transverse canals and clearly noted the internal structure of bone to be lamellar.

42

CHAPTER

1 ~

Developmental Bone Biology

SECTION VII ~ Structural Development of the Epiphyseal Regions Beginning with the primary center of ossification, and its increased development during the fetal time, woven bone is synthesized initially by the inner cambial layer of the periosteum. Very shortly, once an appropriate amount of scaffolding has been synthesized, new bone formation is lamellar in nature. The lamellar bone is deposited on the woven bone cores. With increasing development the lamellae become more compacted, and by the late fetal period Haversian systems have clearly formed. This development continues throughout the postnatal period to skeletal maturation. With the increasing length and diameter of the long bone, there is synthesis particularly on the outer layer of the cortex and resorption internally as the marrow cavity forms and widens. The term woven bone refers to the randomly oriented positioning of the collagen fibrils in the newly synthesized matrix. Any time there is rapid synthesis of new bone in a spatial area where bone did not previously exist, the osteoblasts secrete the collagenous fibrils in all directions and themselves become enmeshed among the fibrillar mass. The newly formed collagen prior to mineralization is referred to as osteoid. Histologic characteristics of woven bone involve relatively large and numerous cells in relation to the amount of matrix present. We refer to the osteoblasts of woven bone as mesenchymal osteoblasts because they are newly formed along the undifferentiated cell line to preosteoblasts and then to osteoblasts. The term lamellar bone refers to bone tissue in which the collagen fibrils are well-oriented in a parallel array. The sheets of fibrils form lamellae, which also relates to the term lamellar bone. Adjacent lamellae are at fight angles to one another although both are in a parallel array. This is sometimes referred to as an orthogonal arrangement. Lamellar bone is synthesized on preexisting cores of tissue, which in cortical bone formation is primarily the initially synthesized woven bone. A structural characteristic of lamellar bone formation is that the osteoblasts responsible for it are present on the surface and they basically secrete the fibrils not only in a parallel array but also in a directional sense only along the adjacent surface of the underlying bone. This is in counter distinction to the mesenchymal osteoblasts, which form woven bone in a 360 ~ spatial pattern around the cell. Lamellar bone tissue is synthesized in intimate relationship to the associated vessels. In a general sense, the vessels of bone run in parallel position along the longitudinal axis of

43

the developing cortex. The vessels are associated with mesenchymal cells of the osteoblast line, the cells referred to in lamellar bone formation as surface osteoblasts. Bone tends to form circumferentially around the individual vessels, forming circumferential lamellae and synthesizing tissue internally until only the vessel and a small number of adjacent cells are present. This leads to what is referred to as compaction of bone, which is also referred to as bone of increased density. The term Haversian system refers to the central vessel surrounded by the lamellae of bone that derives its nutrition from that vessel. This is sometimes referred to as the functional unit of bone and is also referred to as an osteon. The osteons and the central Haversian canal vessels tend to run along the long axis of the bone, although few are strictly parallel to it. Vessels that connect adjacent Haversian systems in a plane transverse to the long axis of the bone are present in Volkmann's canals. They typically are illustrated as being along the transverse axis of the cortex and at fight angles to adjacent Haversian systems, although in fact they tend to be somewhat oblique in orientation. The osteocytes are present in lacunae, and the osteocyte cell processes that link adjacent osteocytes are present in canaliculi. The lacunae and canaliculi in woven bone have no specific orientation consistent with the relationship of the various mesenchymal osteoblasts and young osteocytes. In lamellar bone the orientation of the cells and cell processes becomes more structured. The lacunae are oval structures flattened along the long axis of the lamella, and the canaliculi also tend to pass either parallel to the lamellae or at fight angles to them. Two additional terms referring to the structure of bone can best be introduced here. Bone is referred to as being either dense-compact or cancellous. Cancellous bone refers to spatial regions of the bone in which there is relatively more space filled with cells and vessels than there is mineralized bone tissue. Fine cancellous bone refers to this situation at a light microscopic level of resolution, and it is commonly seen in the developing cortices of the fetus. Gross examination does show what would appear to be uniformly structured cortex, but microscopic examination shows woven and early lamellar bone to be associated with relatively large spaces between the bone tissue filled with cells and vessels, leading to the cancellous description. Coarse cancellous bone refers to the appearance upon gross examination

FIGURE 14 Photomicrographsillustrate the cell and matrix contributions to the physis and the surrounding ossification groove. (A) A portion of the physis, including the perichondrial ossification groove from a rabbit metatarsal, is shown. Note the hypertrophic cell invasion by the red blood cells below and immediatelyto the fight of the bony ring of the ossificationgrooveextendingwell beyond the level of endochondral bone formation. (B) A series of transversecuts is shown, indicating the changingcell and matrixrelationships: (Bi) transverse cut through the epiphyseal cartilage; (Bii) transverse cut through the proliferating and hypertrophic cell region of the physis; (Biii) section through the hypertrophic zone; (Biv) section through the upper reaches of the metaphysis; (Bv) section through the metaphysis at the cut-back zone. Multinucleatedresorptiveosteoclastsare seen. The outer fibrous layer of the grooveremainsintact, but the inner osteogeniclayer and the bone bark are incomplete. (Bvi) Section through the lower regions of the metaphysis shows the outer fibrous layer intact but the bone bark completelyresorbed. Multinucleatedresorptive osteoclasts are prominent. The cortex will be reestablishedat the metaphyseal-diaphysealjunction as intramembranousbone from the periosteumis synthesized. [Parts A, Bi, Bii, Biv, Bv, Bvi reprinted from Shapiro et al. (1977). J. Bone Joint Surgery 59A:703-723, with permission.]

44

CHAPTER

1 *

Developmental Bone Biology

in which there is relatively more space than bone tissue. The term is applied to the noncortical metaphyseal regions in which the trabeculae of bone are surrounded by relatively large spaces filled with cells and vessels.

G. Development of J o i n t s - - G e n e r a l Description Development occurs by a series of cellular changes that are regional in scope. A series of undifferentiated cells forms initially, which then undergoes tissue differentiation and pattern formation from the very general to the specific. Work on joint development initially was described by von Baer (1837) (13), who observed that each element of the appendicular skeleton originally was laid down as a separate cartilage and that it was the unchondrified tissue between these more differentiated elements that eventually formed the joints. Appendicular development moves in a wavelike fashion from the proximal part of the extremity toward the distal part, and development in the upper limb precedes that in the lower limb by several hours to a few days. The various bones are formed by initial centers of chondrification at the central region of the developing bone, followed by progressive chondrification to each end in a well-patterned fashion. The joints form relatively late in development after the patterns of the two adjacent bones, including their epiphyseal regions and articular cartilages, have been established. Development of the joints in the human extends from the first appearance of the joint rudiments at 11 mm C - R length to the appearance of the cavities at 30-34 mm (120, 210, 263). The skeletal development of the shoulder and pelvic girdles and larger bones continues as centers of chondrification in the interior of the blastema, with one for each element at 11 mm. Muscle condensation slightly precedes chondrification. As chondrification expands toward the presumptive ends of each bone, the unchondrified mesenchymal tissues remaining between the cartilages gradually become thinned to form a series of cellular collections referred to as the interzones. It is these regions that form the first morphological indication of the joints. The cavities of the larger joints appear at or soon after the onset of periosteal ossification in the long bones. Joint development thus can be followed through clear stages of (i) homogeneous interzone; (ii) three-layered ihterzone, (iii) the stage of early liquefaction, and (iv) the stage of full separation (Fig. 16). The three-layered structure was described by Bernays (26), Schulin (315), and Kazzander (184) but was best illustrated by Hesser (152). Retterer (290, 291,293) described the histologic aspects of joint formation in detail. Haines believes that the intermediate layer of the interzone is always liquefied soon after the synovial mesenchyme differentiates, so that separation of two adjacent cartilage surfaces is complete by the 34-mm stage at the shoulder and radial-humeral joints, by 44 mm at the knee and probably at the hip, and by 45 mm at the metacarpophalangeal joints and ankle joints, whereas the wrist and carpal joints separate only after 50 mm (138).

It is only later that the joint cavities spread peripherally to develop the complex recesses they subsequently show. Hesser has indicated that the articular surfaces are shaped at a stage when their interzones are still homogeneous, such that little or no movement of adjacent developing cartilage surfaces could occur. The synovial cavities initially are formed partly in the tissues of the interzone and partly in the synovial mesenchyme, so that the articular surfaces are composed centrally of chondrogeneous layers of the interzones and peripherally of tissues of similar constitution that originally are part of the intracapsular perichondrium. Near the margins of the articular surfaces there is a zone of transition between those articular surfaces and the perichondrium. With full chondrification of the articular surfaces, the remains of the liquefied tissues of the interzone or synovial mesenchyme come to form a thin layer overlying the cartilage containing flattened cells, some of them necrotic. Eventually these cells disappear. It is felt that the capsules and the region adjacent to the joints are new formations. Joint development is outlined by photomicrographs of histologic sections in Figs. 15A-15D.

1. DETAILED DESCRIPTIONOF JOINT DEVELOPMENT Joints are first defined in the interzone regions at 11 mm and the synovial cavities of the larger joints appear at 34 mm. The joints first appear as interzones, which are formed from the remains of the skeletal blastema between the cartilages. The interzones form the more central parts of the articular cartilages and synovial cavities. Each interzone passes from a homogeneous stage to a three-layered stage with chondrogeneous layers, an intermediate loose layer, and the stage where the interzone layer breaks down and the chondrogeneous layers become fully chondrified. The fibrous capsule near the joints cuts the mesenchyme into two regions, one of which forms the synovial mesenchyme and the other the perichondrium, which can be intracapsular. The intracapsular perichondrium is partly transformed into the more peripheral parts of the articular cartilage, whereas the remainder persists throughout life. The synovial mesenchyme forms the more central part of the synovial cavities, synovial and subsynovial tissues, and all intracapsular structures including ligaments, tendons, and fibrous cartilages. The interzones of the larger joints appear at 11-12 mm, the fibrous capsule at 16 mm, the interzones become three-layered at 21-26 mm, liquefaction of the synovial mesenchyme begins at 30-34 mm, and the separation of the articular cartilages is complete by 40 mm. Eventually, when the synovial cavities are formed, the remnants of the liquefying tissue are totally destroyed. By 12 mm, the cartilages at the shoulder and elbow have adopted their characteristic shapes and the interzones are relatively thick but quite distinct, with each formed by a mass of undifferentiated blastemal tissue between the cartilage. At the knee the interzone at the first stage is indistinct because chondrification of the distal femur and proximal tibia have not progressed as far, relatively speaking, as those

SECTION VII ~ S t r u c t u r a l D e v e l o p m e n t o f t h e Epiphyseal R e g i o n s

F I G U R E 15 A series of photomicrographs shows the stages of joint development. (A) In stage 1 there is a homogeneous interzone with undifferentiated cells filling the space between the developing cartilage model of two adjacent long bones. (B) A three-layered interzone is formed with relatively more dense accumulations of tissue outlining the articular regions (arrows) of the adjacent epiphyses and slightly less dense but still homogeneous cell accumulations filling what will eventually become the joint space. (C) The stage of early liquefaction is shown here in the developing elbow joint. Although there are still cells in the interzone region, their density is diminished and areas of a cellularity are seen. (D) The fourth stage of joint development represents the stage of full separation with a clear synovial cavity seen.

45

46

CHAPTER 1 ~ Developmental Bone Biology

of the upper extremity at the elbow. At the hip the pelvis is still represented by dense blastemal tissue continuous without form with the upper end of the femur with the interzone not yet in evidence. By 13 mm at the shoulder, hip, and knee, each interzone is now formed by a distinct plate of condensed tissue whose cells pass without interruption into the cartilages on either side. Peripherally the interzone is continuous with the perichondrium of the skeletal elements. At the hip the interzones, perichondrial tissues, and condensed tissues are in unbroken continuity. There is virtually no differentiation within the interzone area during the time in which considerable differentiation of periarticular tissues is occurring. The interzone shows little sign of the shape of any future articular surfaces. At the hip and knee in particular the interzones are fiat with the eventual shape not yet distinguishable. By 14 mm the ends of the cartilages have assumed their characteristic shapes. By 16 mm the interzones of the larger joints are still homogeneous and continuous with the perichondrium surrounding the cartilages. Differentiation of fibrous capsule and ligaments is beginning to be seen, and the interzone region is continuous at its periphery with the intracapsular perichondrium, which in turn is continuous with the extracapsular perichondrium. The early synovium forms between the capsule and extracapsular perichondrium in its own less condensed tissue. This eventually will differentiate into the synovial and subsynovial tissues and such intercapsular structures as are formed. The early derivatives of the blastema, including the cartilages, perichondrial tissues, and interzones, are free of blood vessels at this stage. Vessels ramify over the outer surfaces of the perichondrial tissues. Vessels do not enter the perichondrium until this layer alters in structure in preparation for the formation of periosteal bone. Some vessels, however, appear in the fibrous capsules of the joints and are seen lying in the synovial mesenchyme. By 19 mm the pattern of the interzone, synovial mesenchyme, capsule, and perichondrium is more distinct. The interzones remain fully homogeneous, dense structures composed of tightly packed cells at 18-20 mm, with the cells continuing to undergo frequent mitotic division. As development proceeds, a characteristic three-layered interzone is formed particularly in the intracapsular region. At 21 turn at the elbow both the humeral-radial and humeralulnar joints are clearly three-layered, with two dense cartilage layers that are destined to form the articular surfaces separated by an intermediate loose cell layer. At the periphery of the presumptive joint, the intermediate layer of the interzone is continuous with the synovial mesenchyme but the interzone tissue is always avascular. Embryologists refer to the cartilage layer of the developing ends of the bones as true perichondria, which are continuous with the extracapsular perichondrium. There are clear differences, however, from the perichondrial tissues, which can be removed from underlying surfaces and are well-vascularized from those within the joint itself, which never have a blood supply.

In the shoulder joint at 21 mm there is early loosening of the middle layer of the interzone, with the capsule and the adjacent tendons differentiated from the dense mesenchyme. At the knee dense interzonal tissue still intervenes between each femoral condyle and the articular surface of the tibia. By 23 mm the interzone of the shoulder has attained a typical three-layered structure and a fibrocartilaginous labrum is seen. At 24 mm at the hip and knee the interzones are still dense and homogeneous. At 26 mm the interzones of the knee are in the early three-layered stage. In the knee the interchondral interval is filled with an abundant vascular synovial mesenchyme, from which the cruciate ligaments appear as well as the menisci. At 29 mm, just before the joint cavities appear, the distinction of the layers of the interzone becomes sharper. In the intermediate layer the cells become flattened and lie with their surfaces parallel to the surfaces of the interzone. The matrix of this tissue becomes clearer in preparation for resorption. At the shoulder, Which is advanced in development from other joints, the synovial mesenchyme on either side of the joint is breaking down and small cavities are being formed. By 29 mm all of the larger joints have reached the stage in which the interzone is threelayered. The times at which this appearance is attained vary: shoulder, 23 mm; elbow, 21 mm; wrist and most of the carpal joints, 25 mm; all of the interzones of the fingers, 27 mm; hip, 30 mm; knee, 26 mm; ankle, 27 mm; smaller intertarsal joints, 32 mm; toes, 32 mm. By 30 mm the middle layer of the interzone and the inner portion of the synovial mesenchyme are softening and begin to break down to form the first joint cavities. Between the cartilage surfaces there is a loose liquefying tissue. At 34 rnrn the cavity of the humeroradial joint is well-formed but still contains a few cells in its interior. At the hip the cavity is spreading around the head of the femur and the ligamentum teres lies in this synovial mesenchyme accompanied by conspicuous blood vessels, which will later supply the cartilage canals of part of the head of the femur. At the knee the menisci are sharply differentiated and joint cavities have developed between the anterior parts of the menisci and the femoral condyles.

H. Epiphyseal Blood Supply 1. GENERAL DESCRIPTIONmDUAL PHYSEAL BLOOD SUPPLY FROM EPIPHYSEAL AND METAPHYSEAL VESSELS The blood supply of developing and mature bones has been well-described and is illustrated in Figures 16Ai,ii (39, 73, 255, 349-351). The growth plate itself is avascular but has a dual blood supply, receiving its nutrition from two separate sources: the epiphyseal vessels that supply the germinal, proliferating (columnar), and upper hypertrophic cell layers by diffusion and the metaphyseal vessels that supply the zone of calcification beginning in the lower hypertrophic

SECTION VII ~ Structural Development of the Epiphyseal Regions cell layers by passing only two to three cells deep into the hypertrophic cell lacunae. The epiphyseal vessels thus are responsible for permitting longitudinal growth to occur, whereas the metaphyseal vessels, accompanied by the osteoprogenitor cells, are responsible for laying down bone on the calcified cartilage matrix cores. The specific roles of the epiphyseal vessels and the metaphyseal vessels have been studied for some time in the works of Haas (1917) (132, 133), Trueta and Amato (1960) (351), and Brashear (1963) (32), the latter being most helpful. Each of these works has documented the fact that the proliferation of epiphyseal cartilage cells to enhance longitudinal growth depends on the epiphyseal circulation (Fig. 16B) and that damage to this blood supply causes not only necrosis of the secondary ossification center but marked changes in the epiphyseal plate and longitudinal growth. The diffusion of nutrients also encompasses the entire thickness of the plate, including nutrition into the hypertrophic zone. The extremely rich blood supply on the metaphyseal side of the physis plays essentially no role in the growth process, but has as its main role the calcification of the matrix in the lowermost part of the hypertrophic zone, the invasion of the hypertrophic cell lacunae, and the transport of associated osteoprogenitor cells that synthesize bone on the calcified cartilage cores. The experimental evidence points to the fact that blood carried by the metaphyseal side vessels is of no nutritional importance to the hypertrophic cells. Indeed, when the metaphyseal circulation is markedly ablated, the hypertrophic zone not only persists but increases in size due to the continuation of diffusion from the epiphyseal vessels above. The dual vascular supply to the physeal regions has been well-demonstrated by Trueta and Morgan using vascular perfusion studies (349). They demonstrated the epiphyseal vessels passing through the bone plate of the well-developed secondary ossification center and then ramifying over the surface of the germinal zone of cartilage of the physis without passing into the physeal cartilage itself. They noted that the epiphyseal arteries crossed through canals in the bone plate and expanded into terminal spurs before turning back as large veins, although not always through the same canal. Each terminal expansion covered the space corresponding to from 4 to 10 physeal cell columns, although the vessels themselves did not penetrate into the physeal cartilage. Transverse histologic sections immediately under the bone plate and thus on the surface of the germinal zone of physeal cartilage showed a very rich vascularity, with the whole of the vascular expansions forming a ceiling over the physeal cartilage. The second set of vessels for the dual blood supply represented those from the metaphyseal side, with approximately four-fifths of the vessels reaching the growth plate from the metaphyseal side consisting of the last ramifications of the nutrient artery distributed over the most central parts of the growth plate, with the outer fringe of the lower plate supplied from the system of large perforating metaphy-

47

seal arteries from the adjacent periosteum. These two systems merged, however, such that there was no detectable distinction between the end vessels of the two sources on the metaphyseal side. Both repeatedly divided into ever finer arterioles. 2. EXTRINSIC BLOOD SUPPLY

There are two basic extrinsic patterns of blood supply to the epiphyses (73). In the large majority of epiphyses, vessels pass through the perichondrium of the side walls into the epiphyseal cartilage throughout the circumference of the epiphysis from the level of the articular cartilage to that of the physeal cartilage (Fig. 16B). In epiphyses that are covered almost entirely by articular cartilage, namely, those of the proximal femur and the proximal radius, the blood supply is more tenuous and vessels enter the epiphyseal cartilage in a small region between the articular cartilage and the growth plate cartilage. In the former instance the physeal regions are extracapsular in position and in the latter they are intracapsular (Fig. 16C). The vessels entering the hypertrophic zone are derived from and are the termination of both the nutrient and metaphyseal arteries. Arsenault and Hunter also studied the microvascular organization of the epiphyseal-metaphysealjunction in rats by light microscopy, serial section reconstructions, and scanning electron microscopy (10, 170). The metaphyseal and nutrient arteries undergo extensive arborization and anastomosis near their terminal ends. The microvascular system that projects into the hypertrophic cell region consists of saccular and bulbous terminal arteriole extensions. It is not formed of capillary loops returning arteriole blood to formed venules; venous return occurs below the hypertrophic zone of the epiphysis within anastomosing blood vessels, which drain into large blood sinuses. 3. INTRINSIC BLOOD SUPPLY OF EPIPHYSEAL CARTILAGE VIA CARTILAGE CANALS

Each of the major epiphyseal cartilages in the human is vascularized beginning from the third to the seventh fetal month, which represents a long period of time in each epiphysis prior to formation of the secondary ossification center (39, 48). The primary function of the canals is nutrition, and only later and secondarily do they participate in forming the secondary ossification centers. The vessels within the epiphyseal cartilage are present in cartilage canals in which a central artery or arteriole, capillaries, and a plexus of veins are enclosed in a connective tissue matrix (18, 29, 39, 49, 62, 63, 67, 68, 75, 76, 118, 135, 144, 146, 154, 168, 175, 178, 202, 213, 219, 292, 323, 328, 333, 363, 365-367, 370, 371) (Figs. 9B and 9C). In small canals there is only a capillary plexus. The vessels within the canals do not, therefore, directly contact the cartilage matrix nor do the arteries and veins run separately from each other. The canals and their contents appear to derive from and be continuous with the

Ai

F I G U R E 16 The blood supply of a developing bone is illustrated. (Ai) The epiphyseal vessels (E) are responsible for supplying the epiphyseal cartilage, the secondary ossification center, and also the growth plate, but only by diffusion from above. The lower part of the growth plate is supplied by the metaphyseal vessels coming in from the periphery and by the terminal ramifications of the nutrient (N) artery. The periosteal (P) vessels supply the outer one-third of cortical bone, whereas the nutrient artery supplies the inner two-thirds.

SECTION VII ~ Structural Development of the Epiphyseal Regions

perichondrium, although Delgado-Baeza et al. (76) commented that vessels, perivascular cells, or perichondrium was not necessary for canal morphogenesis and that on the basis of structural cell differences the canals were not a continuation of the perichondrium. Two mechanisms for formation of cartilage canals have been proposed, with support still expressed for both. One mode of origin of the cartilage canals is referred to as the theory of inclusion, whereby the vessels are considered to be present within the cartilage in a passive sense: as the cartilage matrix grows outward by perichondrial cells becoming chondrocytes it incorporates the perichondrial vessels, which continue to function. A more widely supported origin for the canals involves an active invasive or chondrolysis component to the cartilage canals, called the theory of invasion, as they position themselves into relatively large areas in need of nutrition not satisfied by cartilage canals assessed by tritiated thymidine autoradiography. It was concluded that active vessel invasion and proliferation occurred (221). Tritiated thymidine labeling studies to assess cell proliferation activity were done by using 2 Ixc/g body weight intraperitoneal injections into newborn and 3-, 4-, and 7-day-old New Zealand white rabbits that were killed 1 hr after the injection. Proximal humeral, distal femoral, and third metatarsal epiphyses were assessed by histology and serial section autoradiography. Cartilage canals were seen in each epiphysis. Transphyseal vessels were seen in each epiphysis continuous from the epiphysis to the metaphysis or present within the physis traversing the proliferating and hypertrophic cell zones (Fig. 16D). Histologic sections showed vessels from the perichondrium continuous from those of the epiphyseal cartilage canals at proximal humeral, distal femoral, and metatarsal epiphyses. Serial sections showed vascular buds and connective tissue cells lying in indentations at the periphery of and present within the epiphyseal cartilage (Fig. 16E). Autoradiographic studies showed extensive labeling of vessel wall cells and

49

surrounding connective tissue cells of the cartilage canals (1) within the epiphyseal cartilage, (2) traversing the physis, and (3) within the epiphyseal cartilage but continuous with the perichondrial vessels (Fig. 16F). The labeling was always far more extensive than in the surrounding chondrocytes and was always present throughout the entire extent of the canals. The cell labeling activity strongly supports an active dynamic phenomenon underlying the vascularization of epiphyseal and physeal cartilage. Kugler et al. studied cartilage canals in the rat as they related to formation of the secondary ossification center, observing that cartilage canals advanced within the cartilage matrix by chondroblastic absorption (202). One autoradiograph showed thymidine labeling of cartilage canal connective tissue and blood vessel cells. Although some cells similar to chondroclasts can be seen within cartilage canals, our observations indicate that the large majority of canals are not associated with such cells. Fibroblast-like cells and uninuclear macrophages appear to be responsible for the chondrolysis (62, 63). Cole and Wezeman (67) noted chondroclasts in mouse epiphyseal cartilage canals but felt they were related to calcified matrix resorption specifically rather than to the invasive stage of canal positioning. Chondrolytic aspects of cartilage cell positioning do not appear to be mediated exclusively or even predominantly by chondroclasts. The primary function of the canals is to provide nutrition to the epiphyseal cartilage. Only weeks to months later do they provide vascularization for endochondral bone formation of the secondary ossification centers. Many have noted cartilage canal presence from the third fetal month. According to Haines and others there are no anastomoses between separate canals within the cartilage, and the vessels function as end vessels (135). Haines defines the patterns of cartilage canals as (1) simple branched or unbranched canals, which project into the cartilage and end blindly; (2) double or multiple rooted canals, which take their origin by two or

F I G U R E 16 (continued) There is, however, a rich anastomosis within the cortex. The tissues of the groove are supplied by the perichondrial groove artery (GA). (Aii) A cross-sectional illustration of blood supply to cortex and marrow from the mid-diaphyseal region. [Parts Ai, Aii reprinted from "Frazer's Anatomy," A. S. Breathnach, p. 10, 1965, by permission of the publisher Churchill Livingstone.] (B) A light power photomicrograph shows a perichondrial vessel (upper left) passing into the epiphyseal cartilage above the physis and well before the formation of the secondary ossification center. (C) In those epiphyses that are fully intracapsular (A, proximal femur), the epiphyseal region is covered almost completely by articular cartilage and the articular cartilage and the physeal cartilage are quite close (arrow). The vessels must enter in a restricted circumferential region through a narrow area between the articular and physeal cartilages. This pattern is seen at the proximal femoral and proximal radial epiphyses. The vessels are vulnerable to any shift of the epiphysis by acute trauma. This type of blood supply is referred to as type A by Dale and Harris (73). The more common type of blood supply to the epiphyseal regions is shown at right, referred to as type B by Dale and Harris. The vessels enter circumferentially through the side walls of the epiphysis in a broad region between the articular cartilage (arrow) and the physeal cartilage (arrow). These epiphyses are extracapsular in position. [Parts Ci, Cii reprinted from Shapiro, F. (1982). Orthopedics 5:720-736, with permission.] (D) A transphyseal vessel is seen from the proximal humeral epiphysis of rabbit. (E) Perichondrial vessels at the periphery of epiphyseal cartilage mass are illustrated from a 1-week-old distal femur (El) and newborn metatarsal (Eli) (rabbit). The curved arrow in (El) indicates the indentation at cartilage periphery. The perichondrium (P) and fibrous tissue are at upper left. The vessels in the indented areas generally are associated with a cellular connective tissue front (Eii). (F) Tritiated thymidine autoradiography demonstrates the high degree of cell proliferation in the vessels and connective tissue cells of the epiphyseal cartilage canals continuous with the perichondrial vessels in the distal femur at 1 week. This cartilage canal was continuous with a perichondrial vessel. [Parts D - F from Shapiro, F. (1998). Anat. Rec. 252:140-148, copyright 9 1998. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.]

50

C H A P T E R 1 9 Developmental Bone Biology

more separate roots from the perichondrium, but join within the cartilage and then act as simple canals; (3) tunnel canals, which traverse peripheral tonguelike projections of cartilage from edge to edge; (4) dividing and reuniting canals, which end as simple canals; and (5) transphyseal communicating canals, which originate from the perichondrium and end in the bone marrow of the shaft. After the ossification center has formed there are then (6) nutrient canals, which supply the bony center from the perichondrium, and (7) centrifugal canals, which arise from the bony center but behave as simple canals. Hurrell has noted that the vascular pattern of the cartilage canals is always different even in corresponding epiphyses in different fetuses (175). Cartilage canals are present in the human and in vertebrates such as pig, lamb, calf, and rabbit. They are not present in the mouse (111) and rat (23) in which vascular invasion of the epiphysis from the perichondrium occurs only at the time of secondary ossification center formation. 4. TRANSPHYSEAL COMMUNICATING CARTILAGE CANALS In the human in the fetal period epiphyseal cartilage canals occasionally occur, which originate from the perichondrium, pass through the physeal cartilage, and communicate with the metaphyseal marrow. The large majority of these are obliterated either before birth or soon after a secondary center is ossified. Such communicating canals are more prominent in other species such as the lamb in which several can often be seen in one section across the width of a physis. Histologically they are noted to contain red blood cells, indicating participation in blood flow, but the extent of their participation is not quantified. Many observers have noted transphyseal vessel communication in the human, including Hurrell (175), Hintsche (154), Haines (135), Bidder (29), Bardeen (15), Brookes (39), Trueta and Morgan (349), Watermann (365-367), Levene (213), Wang (363), Chappard et al. (62-63), Gray and Gardner (131), and Gardner and Gray (120). Brookes (39) studied human fetal lower extremity bone by barium sulfate microvascular injection techniques. Vascular cartilage canals were detected in epiphyses in rudimentary form as early as 16 cm-20 weeks, but they were well-developed in both ends of femur and tibia by 22 cm C - R length. Transphyseal connections were seen from 22 cm C - R length onward and were more numerous in older fetuses. They were seen at both upper and lower femoral ends. Transphyseal vessels are also present in other species. Levene (213) showed transphyseal vessels in lamb, cat (partial), rabbit, goat, and human, Parsons (268) found them in the young deer, and Hayashi (146) reported them in the fetal rabbit proximal femur. Lutfi (219) and Hunt et al. (168) showed vessels in the proliferating and hypertrophic layers of the chicken. Histologically they are noted to contain red blood cells, indicating participation in blood flow, but the extent of their participation is not quantified. We described transphyseal vessels in the New Zealand white rabbit in new-

bom, and 3-, 4-, and 7-day-old proximal humeral, distal femoral, and metatarsal bones (293) (Fig. 16D). Brookes (39) has shown that the lower end of the femur is vascularized by four groups of canals entering at the intercondylar notch, the suprapatellar surface, and the collateral aspects of the condyle. Multiple sections allow for demonstration of a vascular cartilage canal from the epiphyseal cartilage passing through the physeal cartilage into the metaphyseal area. The upper end of the tibia was canalized by vessels penetrating the intercondylar eminence and the whole circumference of the superior tibial perichondrial cartilage region. The vascular pattem showed a convergence toward the center of the epiphysis. The distal tibial epiphysis also showed arteries entering from all aspects circumferentially, with some entering the medial surface of the malleolus as a separate group. At the proximal femur, the vessels enter the femoral neck and head region circumferentially early on at 22 cm C - R length. Later two major groups are distinguished, a superior and an inferior group. Hurrell (175) felt, unlike others, that the transphyseal communication was passive, with epiphyseal vessels persisting with growth and being encompassed by the physeal cartilage and then metaphyseal marrow. The connection thus was felt to be "accidental and functionless." He particularly stressed that no vascular canal grew actively from the diaphyseal-metaphyseal shaft into the epiphysis. He noted vessels communicating at birth in upper humerus and upper tibia. Haines noted perforating transphyseal canals in several mammals at certain stages of their development. He also felt that the development of such canals was passive. Obliteration of the canals was a constant phenomenon at a slightly later stage of growth. Haines noted them to be "very conspicuous" in the early stages of ossification. The articular cartilage was always avascular. Smaller mammals such as the mouse and rat have no cartilage canals and develop their secondary centers by direct communication from perichondrial vessels at the time of bone formation. 5. CARTILAGE CANALS AND SECONDARY OSSIFICATION CENTER FORMATION Hurrell (175) and Retterer (292) have provided detailed reviews of the many descriptions of cartilage canals. The cartilage canals tend to become obliterated with time but many persist to supply the secondary ossification centers. Haines (135) believed from his serial reconstructions that the ossification center always appears centrally in an avascular zone between sets of cartilage canals. The canals are present in relatively large epiphyses and serve as sources of nutrition where diffusion from the periphery alone would be ineffective. Once an area that is not served by cartilage canals reaches a certain stage, the chondrocytes hypertrophy and degenerate, the cartilage matrix calcifies, and invasion by vessels from the cartilage canals follows bringing in osteoprogenitor cells to form the seconday center of ossification.

SECTION VII ~ Structural Development of the Epiphyseal Regions Bone is then synthesized on calcified cartilage cores. According to Haines: There can be no possibility of the osteoblasts of the epiphysis being derived from periosteal buds growing up from the bone marrow of the shaft, for at the time of ossification all communications between the shaft and the epiphysis have disappeared. It is highly likely that they are derived from the connective tissue of the cartilage canals . . . . The connective tissues of the cartilage canals and the perichondrium are exactly similar and are continuous with one another. The cartilage canals (1) determine by their distribution the position of a clear space or avascular lamina, which is where the ossification center forms, and (2) they give origin to the osteoblasts and marrow of the center. Hintsche (154), using serial section reconstructions, noted a uniform cartilage canal vascularity in epiphyseal cartilage and no relation of vessel pattern to site of secondary center formation or of center formation to avascular regions. Wilsman and van Sickle (371) noted a direct relation of cartilage canal glomerular ends to foci of calcified cartilage, such that the endochondral sequence of the secondary center was the same as that at the undersurface of the physis. They also found cartilage canals to be very evenly spaced.

I. Development of the Articular Cartilage 1. STAGES IN ARTICULAR CARTILAGE DEVELOPMENT The articular cartilage initially is being shaped during the times of epiphyseal and interzone formation. Once resorption of the interzone region has occurred, however, the smooth, free surface of the articular cartilage is evident. Four stages in the development of the articular cartilage can be defined (Fig. 17A). In the second stage the articular cartilage merges imperceptibly at a histologic level of resolution with the underlying epiphyseal cartilage. Other than the fact that the superficial one to two cell layers of the articular cartilage tend to be somewhat flattened parallel to the long axis of the surface, the underlying articular chondrocytes are dispersed randomly and demonstrate no different pattern or orientation from the chondrocytes of the epiphysis. Mankin has demonstrated differing cell proliferation zones, however, using tritiated thymidine autoradiography (223, 224). Thymidine localization in the cartilaginous epiphysis occurs in two distinct layers, indicative of regions of high cell proliferation. One zone was subadjacent to the articular surface so that it contributed to its growth and the other zone was just peripheral to the secondary ossification center, contributing to the epiphyseal expansion and laying the groundwork for enlargement of the ossific nucleus. The third stage of development occurs once the secondary ossification center has been formed and has reached its greatest relative extent referable to replacing the epiphyseal cartilage. At this time the undersurface of the articular cartilage merges with the physis of the secondary ossification center, which is referred to by

51

some as the miniplate. The physis of the secondary ossification center or the miniplate represents the mechanism by which the epiphyseal region of the bone continues to grow on the one hand and to transform to secondary ossification center bone on the other. Strictly speaking, the miniplate is part of the epiphyseal cartilage rather than being an integral part of the articular cartilage, although again there is no line of demarcation at a structural level between the lowermost portions of the articular cartilage and the outermost portions of the epiphyseal cartilage. The final stage in the development of the articular cartilage occurs at skeletal maturation at the same time that the physis closes. Both the physeal and miniplate sequences lose their proliferative capacity. The physeal cartilage is resorbed, allowing for bony continuity between the marrow of the secondary ossification center and that of the metaphysis. At this time, the innermost layer of the articular cartilage calcifies, forming the calcified layer of the articular cartilage, which persists throughout life. The articular cartilage has now reached its final development and is structurally composed of four layers, which cover the subchondral bone plate.

2. LAYERS WITHINARTICULAR CARTILAGE The outermost layer of the articular cartilage is the tangential zone in which two to three cell thicknesses of chondrocytes are flattened with their long axes parallel to the surface of the cartilage. This region is characterized by a relatively high concentration of type I collagen such that it will invariably stain green with Safranin O-fast green due to the relatively diminished amount of proteoglycan. The zone beneath this is called the second or transitional zone referring to the orientation of the collagen fibers as they pass from the superficial tangential zone where they are virtually parallel to the surface to the third or radial zone where they run perpendicular to the surface of the articular cartilage. Between the radial zone and the fourth or calcified zone is a thin, dense staining line referred to as the tidemark (288). The chondrocytes remaining within the calcified layer are viable based on autoradiographic studies, although their activity level is extremely low and perhaps nonexistent. Below the calcified zone of articular cartilage lies the subchondral bone plate. This is composed of lamellar bone and is much more dense than the marrow trabeculae of the mature epiphysis. Polarizing microscopy studies have helped to define the general orientation of the collagen fibrils of articular cartilage (25, 330). Benninghoff (1925) described their orientation as a series of arcades with the fibers of the radial zone perpendicular to the surface, anchored at the base, oblique in the transitional zone, and often parallel to the surface in the tangential zone (25). Subsequent high-power studies at the ultrastructural level failed to reveal this degree of collagen fibrillar orientation, but more recent studies have confirmed their general accuracy. The fibrils, however, do not form continuous sheaths such as can be seen in the outer fibrous layer of the periosteum. As shown at the ultrastructural level,

52

CHAPTER

1 9

Developmental Bone Biology

F I G U R E 17 Articular cartilage development in the rabbit along with the tissue of the epiphysis immediately adjacent to it is shown in photomicrographs at progressively older time periods. (A) At upper left (Ai) the interzone region (I) of the embryonic rabbit knee is shown with femur above and tibia below. The presumptive joint remains filled with cells. The area destined to become articular cartilage of the proximal tibia shows cells packed slightly more densely (arrow). At middle (Aii), the articular cartilage merges imperceptibly with the epiphyseal cartilage. The secondary ossification center at this stage is just beginning to form with hypertrophic cells (H) seen. In panel (Aiii), the secondary ossification center has formed but is still relatively far from the articular cartilage. The articular and epiphyseal cartilages still merge. At lower left (Aiv), most of the epiphyseal cartilage has been converted to secondary ossification center bone. Some still remains. The undersurface of the articular cartilage appears to be continuous over a short distance with the endochondral sequence at the periphery of the enlarged secondary ossification center. The latter region can be referred to as the miniplate at this stage. In panel (Av), the articular cartilage at skeletal maturity has its lowermost zone calcified, separating it from the subchondral bone (B) of the mature epiphyseal region. (The photomicrographs Ai-Av are not magnified to scale.) (B) Tritiated proline autoradiograph shows the nutritive pathway of immature articular cartilage. The proline was dropped onto the surface of the rabbit distal femur cartilage at open operation. Uptake (black dots) 2 hr postinjection is concentrated over the articular chondrocytes, but much label is seen in matrix. Note that osteoblasts of underlying bone have also taken up the label and synthesized collagen, with dense label indicative of new osteoid on subchondral bone surfaces (arrows). The proline is incorporated into newly synthesized collagen. At skeletal maturity the label does not pass the tidemark, indicating no nutrition of subchondral bone from synovial fluid after skeletal maturity.

SECTION VIII ~ Axes along Which Bones Are Patterned the fibrils are short and discontinuous but their cumulative orientation is consistent with the arcade pattern. 3. NUTRITION OF ARTICULAR CARTILAGE The articular cartilage is always avascular throughout the entire period of its development and in adult life. In the human and in the other relatively large species in which cartilage canals form, they are never present within the articular cartilage regions. They are not seen in random or serial histologic sections or by angiographic studies. Articular cartilage thus receives its nutrition by diffusion from the synovial fluid at all ages. In the time period prior to skeletal maturity, the cartilage of the lowermost part of the articular cartilage and the outermost part of the epiphyseal cartilage forms a miniplate, which undergoes the endochondral sequence. This miniplate, or physis of the secondary ossification center as we prefer to call it, allows for outward growth of the articular and epiphyseal cartilage and for enlargement of the secondary ossification center. Labeling studies in the immature animal clearly demonstrate that articular cartilage chondrocytes and chondrocytes of this endochondral sequence, including bone and marrow cells at the outermost reaches of the secondary ossification center, can receive nutrition by diffusion from the articular surface (Fig. 17B). The bone and marrow of the secondary center also receive nutrition from the epiphyseal blood supply to that region (232). McKibbin and associates have also shown that, in the immature rabbit, fluid (and by inference nutrition) from the epiphyseal secondary ossification center blood supply can pass freely into joint cartilage (155). In the skeletally mature adult, however, transfer from bone to articular cartilage could not be demonstrated across the lowermost calcified cartilage layer (155). At skeletal maturity the inner layer of the articular cartilage calcifies and remains calcified through life, as noted previously. At this stage, synovial nutrients cannot pass through the calcified zone into underlying bone, and there is no vascular flow and thus no possibility of nutrition passing from epiphyseal bone through the calcified zone of articular cartilage to the upper three zones of cartilage. In the adult, articular cartilage receives nutrition only by diffusion from the synovial fluid, and the subchondral or epiphyseal bone receives nutrition only from the blood supply to the bone.

VIII. A X E S A L O N G W H I C H B O N E S

ARE PATTERNED Once the limb bud forms (Fig. 18A) there are three axes along which patterning of differentiation in a developing limb occur. These involve (1) the proximal-distal longitudinal axis, with those regions closer to the shoulder and hip developing in advance of those more distal, (2) the anteroposterior axis, which defines first digit side to fifth digit side structures, and (3) the dorsoventral axis, which defines the differentiation of extensor compartment from flexor com-

53

partment structures. The apical ectodermal ridge (AER) directs proximodistal growth, the zone of polarizing activity (ZPA) directs anteroposterior patterning, and the dorsal nonridge ectoderm directs dorsoventral growth.

A. Signaling Regions That Affect the Patterns of Bone Development Four major signaling regions have been identified in developing limbs, which appear to control patterning and morphogenesis (347) (Fig. 18B). (1) A continuous ridge of elevated epithelial cells, called the apical ectodermal ridge (AER), soon forms running anterior to posterior along the distal outer tip and distal inner borders of each developing limb bud. At the tip of the limb bud this ridge is markedly thickened due to an increased number of closely packed epithelial cells induced by the underlying mesenchyme. This structure stimulates growth and differentiation of the immediately proximal mesenchymal limb bud cells by inductive means. It serves to outline limb development along the proximaldistal axis. Ridge formation depends on signals from the mesenchyme, but, once established, there are reciprocal ridge-mesenchyme (or epithelial-mesenchymal) interactions as development proceeds. Surgical removal of the ridge in experimental investigations leads to decreased cell proliferation in the underlying mesenchyme and proximal-distal growth alteration, such that the more proximal humeral and femoral segments appear normal (because they had already been patterned) whereas the more distal segments are missing. (2) The AER mediates growth at the distal tip of the limb bud immediately adjacent to it from a zone of undifferentiated cells called the progress zone (PZ). As development proceeds, proximal to distal, damage to or removal of the AER immediately halts development. Early removals lead to proximal truncations and later removals lead to more distal truncations. The types of structures formed, however, are governed by the mesenchymal cells themselves. (3) Development along the anteroposterior axis is mediated from an area in the posterior mesenchyme referred to as the zone of polarizing activity (ZPA), which is not distinctive histologically but is defined on the basis of chick embryo transplantation experiments in which damage to the posterior caudal area induced changes in wing digit development. (4) Development along the dorsoventral axis is mediated by dorsal nonridge ectoderm Each of the controlling regions AER, PZ, ZPA and dorsal non-ridge ectoderm appear to be interdependent. Both the AER and the ZPA initially were defined by transplant experiments, but presently molecular localization is occurring. The primary molecular signaling components of regions of the limb bud are shown in Fig. 18C.

B. Models of Tissue Patterning Johnson and Tabin have reviewed the mechanisms of threedimensional formation of organisms via the process referred

$4

CHAPTER

1 ~

Developmental Bone Biology

FIGURE 18 Earlylimb bud characteristicsare outlined. (A) Histologic section of a limb bud of a rabbit in which the mesenchymal cells are uniform in appearance and densely packed. The apical ectodermal ridge is shown at right. (B) The developing axes and signaling regions of the early limb bud are illustrated. The zone of polarizing activity and the progress zone cannot be differentiatedby light microscopy but require immunocytochemistryand in situ hybridization demonstration of specific molecules concentrated there. (C) Primary molecularsignaling components of regions of limb bud. [Part C reprinted from Cohn, M. J., and Tickle, C. (1996). Trends Genet. 12:253-257, copyright 1996, with permission from Elsevier Science.]

to as pattern formation (180). They indicated that the general features of the body initially were produced along the rostral-caudal axis and that during subsequent development more specific regional differentiation occurred in semiautonomous fashion in areas referred to as secondary fields. Pattern formation in secondary fields such as a developing limb then occurred in four basic stages: (1) cells that make up the field are defined; (2) specific signaling centers are established within the field to provide positional information;

(3) the positional information is recorded on a cell by cell basis; and (4) cells differentiate in response to additional cues according to the encoded positional information. The molecular basis for these changes is being determined rapidly, and much study currently is underway to assess how the molecules relate to each other in the developmental cascade. Extensive research continues to determine how the pattern of tissues is synthesized (211). Several theories of pattern formation have been devised. These are becoming increas-

SECTION VIII ~ Axes along Which Bones Are Patterned ingly amenable to experimental verification by molecular markers using techniques of in situ hybridization and transgenic mouse or avian models. A superb monograph by Held (149) titled "Models for Embryonic Periodicity" summarizes the many theories of pattern formation into three conceptual classes: position-dependent, rearrangement, and cell lineage. The concept of pattern formation clearly is crucial to understanding both normal skeletal development and abnormal development in a wide array of disorders, from the numerous skeletal dysplasias to mild abnormalities such as hemiatrophy. There clearly are situations where pattem formation has been imperfect. Development proceeds from the limb bud stage, in which one sees only a uniform mass of undifferentiated mesenchymal cells, to full limb development by a combination of mechanisms referred to as differentiation, wherein one cell has generated many different types of cells and pattern formation in which the spatial control of differentiation occurs. Relatively simple mathematical models can be generated to help understand patterning mechanisms. These can be reduced to the three classes referred to earlier based on commitment of embryonic cells to different pathways of development long before they show any tissue differentiation, such that the cells are described as determined for different fates and possessing distinct states of determination. This commitment has allowed them to be determined for synthesizing different tissues and programmed in regard to their position in the limb. In the words of Held, "given the notion of states, the problem of pattem formation can be reduced to simple mathematical terms." Although the "causal relationships define distinct classes of m e c h a n i s m s . . , actual developmental pathways typically employ multiple strategies." Many of these models are amenable to investigation by transplanting cells from one position to another. If a transplanted cell changes its fate by adopting a fate appropriate for its new position, then its position was causing its state. If the cell moves back to its original position following transplantation, then its state was causing its position. If the cell changes neither fate nor moves back, then the third type of mechanism would be indicated. When p is used to designate the position of a cell and s is its state of determination or differentiation, then any pattern can be represented as a set of ordered pairs (p, s). The general problem thus becomes: "what causes the correlation of particular values of p and s?" Whenever two entities are correlated in nature, either one causes the other or both are caused by a third force. For p and s, there are three possibilities: (1) p produces s or the position of a cell causes it to adopt a particular state (positiondependent class); (2) s produces p the state of a cell causes it to adopt a particular position (rearrangement class); and (3) X produces p and s or some third agent (X) causes the correlation of positions and states. An example of X is cell lineage because a mother cell can divide directionally (causing the p of each daughter) and bestow instructions (s) asymmetrically.

55

Held then further defines the three classes of pattern formation models with additional subclasses. (1) The positiondependent class (p producing s) involves (a) the positional information subclass, as determined by gradient models, polar coordinate models, or the progress zone model; (b) the prepattem subclass, for which there are physical force models, reaction-diffusion models and induction models; (c) the determination wave subclass, for which there are chemical wave models (Belousov-Zhabotinsky reaction), the sequential induction model, the clock and wave-front model, and inhibitory field and competence wave models; and (d) the Darwinian subclass, with cell death and state change models. (2) The second class is the rearrangement class (s producing p) with adhesion, repulsion, interdigitation, and chemotaxis models. (3) The final class is a cell lineage class (X produces p, s) with quantal mitosis, stem cell, and cortical inheritance models. In summary, cell differentiation to form specific structural patterns occurs because either a cell's position causes its state, its state causes its position, or both are caused by a third agent. In the position-dependent class, the position of each cell relative to field boundaries or neighboring cells dictates its state of differentiation. In the rearrangement class, each cell adopts a state, perhaps randomly; the states then cause cells to move until they reach particular locations. In the cell lineage class, the cells divide asymmetrically according to rigid pedigree rules, placing each daughter into a definite position and assigning it a particular state. The various models and categories and their distinguishing features as defined by Held are summarized in Table V.

C. Positional Information Mechanisms Positional information models postulate a position-dependent assignment of differentiated states, whereby cells are informed of their positions and this information causes them to select particular states according to determined rules. Wolpert has indicated that cells acquire a positional value with respect to boundaries and then interpret this in terms of a program determined by their genes and developmental history (375). Cell-to-cell interactions can then specify position. He feels that positioned fields are small, being less than 0.5 mm long, with the times needed to specify position on the order of hours. The various models differ only in how they specify positional information. The mechanical gradient model is most common in this regard being mediated by long-range diffusible molecular signals. Each cell records the concentration of a particular chemical, which has an origin or source and a terminal point or sink. The polar coordinate model is also one defining positional information and assumes that cells assess their position by observing coordinates of their immediate neighbors, presumably by direct cell surface contact (41). In this model, the position of limb cells is specified by one coordinate corresponding to the proximal-distal axis of the limb and another corresponding to the position on one of a concentric series of circles at

56

CHAPTER I ~

Developmental Bone Biology TABLE V

Models of Tissue Patterning Mechanisms a

Distinguishing features

Model or category I. Position-dependent (p ~ s) class A. Positional information subclass

The position of each cell (relative to field boundaries or neighboring cells) dictates its state of differentiation Cells know where they are via coordinates that they "interpret" as particular states of differentiation; the coordinate systems allow the patterns to "regulate"

Gradient model

The coordinate system is established (independent of growth) to be a scalar variable with fixed boundary values

Polar coordinate model

A "shortest route rule" or "smoothing rule" fills in missing coordinates by intercalary growth

Progress zone model

Coordinates are assigned temporally as cells exit a growth zone

B. Prepattern subclass

Cells assume particular states of determination due to mechanical or chemical signals within the cell layer or by induction from an adjacent cell layer Identical signals are used for identical elements Patterns do not regulate (unless ad hoc assumptions are added)

Physical force models

Deformations arise at periodic intervals within a tissue layer, causing cells to adopt particular states of determination about a certain threshold of stress or strain

Reaction-diffusion models

Chemicals that have different diffusion rates react, causing an initially uniform chemical distribution to peak at "wavelength" intervals Above a certain threshold concentratin, cells adopt a particular state of determination

Induction model C. Determination wave subclass

Periodically arranged cells in one layer induce states of determination in the cells of an apposed layer States of determination are specified within a zone that traverses an array of cells

Chemical wave models

Traveling (or standing) waves in the concentration of a diffusible molecule (or precipitate) arise through chemical reactions

Sequential induction model

Each cell induces a neighboring cell to adopt a particular state

Clock and wave front model

Cells oscillate between two states and cease oscillating when a wave front reaches them

Inhibitory field and competence wave model

Cells are not "competent" to differentiate before a wave front reaches them Cells that can adopt a "preferred" state do so and inhibit neighboring cells from doing so

D. Darwinian subclass

Each cells adopts a state (perhaps randomly) and then examines the states of its neighbors; if it matches a neighbor, then it takes action to correct this "error"

Cell death models

Homotypic matches are eliminated by having one of the matched cells die

State change models

Homotypic matches are eliminated by having one of the matched cells change its state Each cell adopts a state (perhaps randomly); the states then cause cells to move until they reach particular locations

II. Rearrangement (s ~ p) class Adhesion models

The final location of a cell is determined by its ability to adhere to a target cell(s)

Repulsion models

Each cell moves as far away as possible from cells of its own kind

Interdigitation models

Stripes of unlike cells interdigitate

Chemotaxis models

Dispersed cells aggregate by mutual attraction

III. Cell lineage (X ~ p, s) class

Cells divide asymmetrically (according to rigid pedigree rules), placing each daughter in a definite position and assigning it a particular state

Quantal mitosis model

All cells undergo an asymmetric and polarized "quanta" mitosis, which assigns left daughters one state and fight daughters another

Stem cell model

A cell cyclically changes its state as it divides, causing the states of its daughters to alternate in space as it oscillates in time

Cortical inheritance model

A periodic pattern of molecules is created in the cortical layer of a cell, and each daughter differentiates according to the molecules it inherits

aDerived from Held (149), with permission.

SECTION VIII 9 Axes along Which Bones Are Patterned

different proximal-distal levels. The molecular basis of this system is beginning to be outlined because genes that specify the circumferential coordinate have been found in Drosophila embryo, with different genes expressed in varying local segments (41). Other genes are expressed at varying proximal-distal regions. The final mechanism in this class is the progress zone model, in which the cells at a particular region of the limb bud acquire positional information via temporal information in which they can measure time and switch off their specific function when they exit a zone.

D. Prepattern Mechanisms Prepattern mechanisms define cell differentiation patterns on the basis of mechanical or chemical signals within cell layers. Identical signals thus will induce identical structures. Several models have also been defined within this subclass of developmental mechanisms. Physical force models postulate that the physical properties of cells and matrices can produce local deformations in response to internal or external forces, and these distortions could promote the development of structure. Perhaps a more accepted model today is the reaction-diffusion model defined by Alan Turing based on pattern differentiation between chemicals. Chemically reactive molecules that diffuse at different rates cause the concentration of their product to peak at wavelength intervals, which induce periodic or segmented structures. Many of these reaction-diffusion models are somewhat indeterminate in that the final configuration of the pattern cannot be predicted exactly from starting positions unlike positional information, which is strictly deterministic, by which is meant they yield identical patterns from case to case. Such patterns are defined as epigenetic in that they are not invariably the same but relate to a cascade of formative principles, by which each region is specifically dependent on that that preceded it. Even the most subtle difference, therefore, can lead the developmental pattern into slightly different areas.

E. Determination Wave Mechanisms From 1920 to 1954, the reigning paradigm in developmental biology was that of the determination wave mechanism in which a propagating signal or substance spread from an "organized center" to control the fate of cells in its domain. Subsequently, the prepattern (1954-1969) and positional information (1969-present) schools gained prominence. Many models of definition within this subclass were defined. The progress of morphogenesis sequentially and along varying axes led to the development of several wave mechanism theories. Among the theories used to support this mechanism were those of chemical waves. An example of this was the Belousov-Zhabotinsky reaction in which the concentration of a chemical oscillates in time and space and produces patterns similar to those seen in structural organisms. Intricate patterns can be defined by employing contact-mediated cel-

57

lular communication instead of diffusible chemicals. The sequential induction model is an example in which the notochord induces formation of the central nervous system. Variations on this theme include waves of developmental signals, which induce adjacent groups of cells to oscillate between active and inhibited states. Even Darwinian models have been defined utilizing cell death and state change models. The Darwinian concept implies that, even within individual embryos, selection among competing groups of cells leads to longevity of some and programmed cell death of others. Currently there is great interest in the phenomenon of programmed cell death or apoptosis, which plays a major role in developmental states. A clear example of cell death in normal limb development relates to joint formation in which the interzone cells die after formation of the model of each developing bone to subsequently allow for joint cavitation. Some consider the death of terminal hypertrophic zone chondrocytes to represent another example of apoptosis in skeletal development. State change models imply the ability of cells to react to changes in adjacent cells by subtle differentiation to help maintain or recreate their previous state.

F. Rearrangement Mechanisms Rearrangement mechanisms refer to the concept of cell and tissue movements that characterize many early developmental states. Cell movement is particularly well-defined in central nervous system development. A cell's state of determination or differentiation causes it to assume a specific position relative to other cells by propelling it in a particular direction. This developmental subclass makes use of the "self-assembly" concept, for example, in mediating the supracellular neuroarchitecture in which specific cell axons link only with specific neurons based on specific chemical reactions. Both adhesion and repulsion models have been used to account for cell movement within an embryonic region.

G. Cell Lineage Mechanisms Cell lineage mechanisms were postulated for some time but have been almost completely supplanted by newer mechanisms. They implied that a cell's positions and states are assigned via strict pedigree rules with no involvement of intercellular communication. Cell lineage plays a major role in later tissue differentiation but participates little in embryonic morphogenetic patterning. Much of developmental biology until recent decades was based on descriptive embryology on the one hand and experimental embryology on the other in which certain regions at certain stages were either removed from a developing embryo or transferred to atypical positions with development then followed carefully. With the explosion of information concerning the specific molecules present within the de-

58

CHAPTER 1 ~

Developmental Bone Biolo~ty

veloping systems and recognition of patterning genes or homeobox genes, developmental studies have become much more specific. Utilization of in situ hybridization techniques and the generation of transgenic animals with specific molecules subtracted at the varying stages of development have allowed for more specific analysis. The fact that development is hierarchical is self-evident. What is unclear, even today, however, is whether the pattern is totally deterministic, which would allow development to be reduced to an elucidation of which specific molecules at each stage of development perform which activities, such as synthesizing certain substances, acting as enzymes, or regulating genes. On the other hand, much of development conceivably could be epigenetic in which only the early prepatterns are rigidly determined, after which each subsequent step is a combination of gene synthesis and automatic self-assembly based on the physical presence of certain molecules. These questions thus span the entire spectrum from gene to ultrastructure, genotype to phenotype, or linear DNA to three-dimensional morphology. In the epigenetic approaches, there are no genes that directly specify the entire sequence of events. There are only those that give approximate direction. The rest of the formation is based on epigenetic or self-assembly mechanisms. Thus, it is the interactions among many genes that would determine the outcome of any dynamic process. Many have suggested that the detailed structure of multicellular organisms occurs on the basis of many intermediate levels of interaction, each with its own emergent properties such that "the edifice is virtually entirely epigenetic." This idea was expressed by Driesch over a century ago in which he indicated that development started with a few ordered reactions but that each reaction created new structures by interaction, which were enabled by acting back upon the original ones to provide new differences and so on. "With each effect, immediately a new cause is provided and the possibility of a new specific action." Opposed to this are the strictly deterministic approaches such as the homeobox gene synthesis patterns. The homeobox genes are ordered along their chromosomes and allow specific anatomic regions to appear. They have not only a spatial and regional preference but also a temporal one. At the basis of much modern biology is the feeling that there may be a finite number of elemental patterning strategies.

IX. G E N E A N D M O L E C U L A R

CONTROLS

OF LIMB DEVELOPMENT Findings have outlined the gene and molecular mechanisms associated specifically with the early stages of limb embryogenesis (65, 84, 95, 96, 104, 150, 156, 179, 198, 201,233, 236, 241,295, 299, 316, 338, 342, 347). Specific genes and molecules controlling limb development and comprising the extracellular matrix are listed in table VIA. Four specific regions of the developing limb play major directing roles in morphogenesis. [See section VillA (Figs. 18B and 18C).]

A. Apical Ectodermal Ridge The apical ectodermal ridge (AER) is the thickened rim of epithelium present at the tip of each limb bud (Fig. 18A). It consists of pseudo-stratified elongated cells that are closely packed and linked by extensive gap junctions. AER maintains cell proliferation in the underlying mesenchyme progress zone, and there are constant reciprocal interactions between ridge and mesenchymal tissues. The AER mediates proximodistal outgrowth of the limb bud and controls the deposition of undifferentiated cells just proximal to it in the progress zone. FGF-8 is expressed throughout the AER early in limb development, whereas FGF-4 later is detected concentrated in the AER posteriorly. FGF-4 is expressed most prominently in the AER and appears to play the major role in proximodistal outgrowth and patterning.

B. Progress Zone The progress zone is at the distal region of the limb bud in which undifferentiated mesenchymal cells actively divide. Cell division in the progress zone appears to be timed accurately, and it is felt by some that the length of time cells spend in the progress zone specifies the structures they will form.

C. Polarizing Region The zone of polarizing activity (ZPA) is not distinct histologically but its controlling features were identified on the basis of transplant experiments. In the limb bud, the highest polarization activity is posterior near the tip of the bud just proximal to the progress zone. The polarizing activity is confined to the mesenchyme (Fig. 18B). Sonic hedgehog (Shh) appears to be the signal from the posterior polarizing mesenchyme involved in anteroposterior patterning, but BMP-2 also is synthesized there.

D. Dorsal Nonridge Ectoderm The signal necessary from the dorsal nonridge ectoderm to trigger dorsoventral development appears to come from expression of Wnt 7a molecules. There is clearly mutual dependence between signaling systems (180, 243,266, 379). The expression of Shh in posterior mesenchyme maintains FGF-4 expression in the posterior apical ridge, and Wnt 7a expression in dorsal ectoderm together with FGF-4 maintains Shh expression in posterior mesenchyme. Current investigation is involved extensively with assessing the developmental cascades to define how the various transcriptional regulators and signaling proteins induce each other and eventually direct cell positioning and the synthesis of extracellular molecules. Each of the factors listed previously induces patterns of Hox gene expression throughout the limb bud.

SECTION IX ~ Gene and Molecular Controls of Limb Development TABLE VIA Gene

Protein

ADA

Adenosine deaminase

ALPL

Alkaline phosphatase liver, bone, kidney type

ANXA5

Annexin A5; calcium and phospholipid binding protein, endonexin 2, placental protein 4, anchorin CII Arylsulfatase E

ARSE BMP1

BMP2

BMP3

BMP4

Bone morphogenetic protein 1; also known as procollagen C-proteinase Bone morphogenetic protein 2, a member of transforming growth factor-[3 superfamily Bone morphogenetic protein 3, a member of transforming growth factor-[3 superfamily Bone morphogenetic protein 4

BMP5

Bone morphogenetic protein 5, a member of transforming growth factor-[3 superfamily

BMP6

Bone morphogenetic protein 6, a member of transforming growth factor-[3 superfamily

BMP7

Bone morphogenetic protein 7. a member of transforming growth factor-13 superfamily, osteogenic protein 1 Bone morphogenetic protein 8, osteogenic protein 2 Bone morphogenetic protein receptor, type 1A

BMP8 BMPR1A

CA2

Carbonic anhydrase II

CASR

Calcium sensing receptor

59

Skeletal Gene D a t a b a s e ~

Cellular function

Involved in T-cell and B-cell immune function Linked to the outer membrane, enzyme acts physiologically as a lipidanchored phosphoethanolamine and pyridoxal 5'-phosphate ectophosphatase Collagen receptor of chondrocytes predominantly expressed in chondrocytes of growth plate, down-regulated in adult growth cartilage Involved in bone and skeletal development Induce formation of ectopic cartilage and bone in vivo

Disease

Adenosine deaminase deficiency Hypophosphatasia, murine (Alp 1- / - ) seizures but normal skeletal development

X-linked recessive chondrodysplasia punctata (brachytelephalangic type)

Induce formation of ectopic cartilage and bone in vivo Induce formation of ectopic cartilage and bone in vivo Involved in bone induction and tooth development Induce formation of ectopic cartilage and bone in vivo

? Fibrodysplasia ossificans progressiva Murine: mutations in short-ear locus association with homozygous deletions of BMP5 coding regions result in viable and fertile short ear mice with specific skeletal defects

Induce differentiation of osteoblast and chondroblast lineage cells from uncommitted mesenchymal precursors; BMP6 produced by osteoblasts is increased in the presence of estrogen and is thought to mediate the skeletal effects of estrogen Induce formation of ectopic cartilage and bone in vivo

Induce formation of ectopic cartilage and bone in vivo Cell surface receptor with predicted serine-threonine kinase activity, may bind to ligands from the transforming growth factor-[3 supergene family Involved in osteoclast function, bone remodeling and resorption Parathyroid hormone sensitive, involved in calcium homeostasis

Osteopetrosis with renal tubular acidosis and cerebral calcification Neonatal severe hyperparathyroidism; familial hypocalciuric hypercalcemia; autosomal dominant hypocalcemia; murine (Casr+/-) familial hypocalciuric hypercalcemia; murine ( C a s r - / - ) neonatal severe hypercalcemia (continues)

6O

CHAPTER I ~ Developmental Bone Biology TABLE VIA (continued) Gene

Protein

Cellular function

Disease

CBFA1

Core binding factor et subunit 1

Osteoblast-specific transcription factor

Cleidocranial dysplasia; murine Cbfa + / cleidocranial dysplasia; murine Cbfa - / perinatal lethal with deficient ossification of skeleton

CBP1

Collagen binding protein 1; collagen 1

CBP2

Collagen binding protein 2; collagen 2

CD36

Type 1 collagen receptor, thrombospondin receptor; CD 36 antigen CD 36 antigen-like 1; type 1 collagen receptor-like 1; thrombospondin receptorlike 1; scavenger receptor class B type 1 (SRB 1) CD 36 antigen-like 2; type 2 collagen receptor-like 1; thrombospondin receptorlike 2 Cartilage-derived morphogenetic protein 1

Bind specifically to type I and type IV collagen and gelatin; may be involved in the biosynthetic pathway of collagen Bind specifically to type I and type IV collagen and gelatin; may be involved in the biosynthetic pathway of collagen Bind to collagen and thrombospondin; involved in platelet collagen adhesion High-density lipoprotein receptor

CD36L1

CD36L2

CDMP1/GDF5

CKTSF1B 1

Cysteine knot superfamily 1; BMP antagonist 1; also known as gremlin

COLIA1

Type I collagen ct 1 chain

COLIA2

Type I collagen (x2 chain

COL2A1

Type II collagen ct 1 chain

COL3A1

Type III collagen (x1 chain

Platelet glycoprotein IV deficiency

Lysosomal integral membrane protein II

Involved in skeletal morphogenesis, i.e., chondrocytic differentiation and growth of long bones Block BMP signaling by binding to BMPs, preventing them from binding to their receptors

Major collagen of skin, tendons, and bone Major collagen of skin, tendons, and bone Collagen of cartilage and vitreous

Fetal collagen, expressed throughout embryogenesis essential for normal fibrillogenesis of collagen 1 in cardiovascular system and other organs

Greb chondrodysplasia; HunterThompson dysplasia; brachydactyly type C; murine brachypodism Murine: limb deformity mutation (ld) disrupts Shh-FGf4 feedback loop; in mouse embryos mesenchymal expression of gremlin is lost in limb buds; grafting gremlin expressing cells into ld mutant limb buds rescued FGF4 expression and restored the Shh-FGF4 feedback loop Osteogenesis imperfecta; type VIIA EhlersDanlos syndrome; osteoporosis Osteogenesis imperfecta; type VIIB EhlersDanlos syndrome Achondrogenesis type II; hypochondrogenesis; Kniest dysplasia; spondyloepiphyseal dysplasia congenita; Wagner syndrome; spondyloepimetaphyseal dysplasia; Strudwick type and Namaqualand type spondyloepimetaphyseal dysplasia; osteoarthritis with mild chondrodystrophy; transgenic C o l 2 a l - / - mice perinatal lethal with absent enchondral ossification Ehlers-Danlos syndrome type III; EhlersDanlos syndrome type IV; murine Col3al-/-Ehlers-Danlos type IV; aortic aneurysm

(continues)

SECTION IX ~ Gene and Molecular Controls of Limb Development

61

TABLE VIA (continued) Gene

Protein

Cellular function

COL4A1

Type IV collagen e~1 chain

COL4A2

Type IV collagen oL2 chain

COL4A3

Type IV collagen o~3 chain

COL4A4

Type IV collagen c~4 chain

COL4A5

Type IV collagen e~5 chain

Collagen of basement membrane, associated with laminin, entactin, and heparan sulfate proteoglycans to form basement membranes that separate epithelium from connective tissue Collagen of basement membrane, associated with laminin, enacticn, and heparan sulfate proteoglycans to form basement membranes that separate epithelium from connective tissue Associated with lamin, entactin, and heparan sulfate proteoglycans to form basement membrane separating epithelium from connective tissue Associated with lamin, entactin, and heparan sulfate proteoglycans to form basement membrane separating epithelium from connective tissue Collagen of basement membrane

COL4A6

Type IV collagen c~6 chain

Collagen of basement membrane

COL5A1

Type V collagen oL1 chain

COL5A2

Type V collagen oL2 chain

Minor collagen involved in regulation of fibrillogenesis of dermis rather than bone Minor collagen involved in regulation of type I and type II collagen fibrils; collagen of fetal membrane

COL6A1

Type VI collagen c~1 chain

COL6A2 COL 7A1

Type VI collagen o~2 chain Type VI collagen c~3 chain type VII collagen oL1 chain

COL8A1

Type VIII collagen oL1 chain

COL8A2

Type VIII collagen oL2 chain

COL9A2

Type IX collagen c~2 chain

COL9A3

Type IX collagen c~3 chain

COL6A3

Anchoring structural protein involved in maintaining the integrity of muscle fiber Anchoring structural protein Act as a cell binding protein Main constituent in anchoring fibrils

Endothelial cell collagen expressed in eye, skin, and calvarium; major component of Descemet membrane Major component of Descemet membrane Cartilage-specific fibril-associated collagen Fibril-associated collagen with interrupted triple helices expressed in cartilage

Disease

Autosomal recessive Alport syndrome type I; murine knockout Co14a3 Alport syndrome; induction of Goodpasture syndrome Autosomal recessive Alport syndrome type II; autosomal dominant benign familial hematuria-q37 X-linked recessive Alport syndrome; canine X-linked hereditary nephropathy X-linked Alport syndrome with diffuse leiomyomatosis contiguous gene deletion syndrome Ehlers-Danlos syndrome type I; EhlersDanlos syndrome type II Ehlers-Danlos syndrome type I; EhlersDanlos syndrome type II; murine Co15a2-/- spinal deformities, skin, and eye abnormalities Bethlem myopathy; murine C o l 6 a l - / Bethlem myopathy Bethlem myopathy Bethlem myopathy Autosomal dominant dystrophic epidermolysis bullosa; autosomal recessive dystrophic epidermolysis bullosa; dystrophic epidermolysis bullosa Bart syndrome type: transient bullous dermolysis of newborn

Multiple epiphyseal dysplasia type II intervertebral disk disease Multiple epiphyseal dysplasia; multiple epiphyseal dysplasia with myopathy

(continues)

CHAPTER 1 ~ Developmental Bone Biology

62

TABLE VIA (continued) Gene

Protein

Cellular function

COLI OA1 COLllA1

Type X collagen ot1 chain Type XI collagen (x1 chain

Minor collagen of cartilage Minor collagen of cartilage; fibrillar collagen

COL11A2

Type XI collagen ix2 chain

Minor collagen of cartilage; fibrillar collagen

COL12A1

type XII collagen et 1 chain

COL13A1

Type XIII collagen (x1 chain

COL14A1

Type XIV collagen etl chain

COL15A1

Type XV collagen (x1 chain

COL16A1

Type XVI collagen et 1 chain

COL17A1

Type XVII collagen (x1 chain, bullous pemphigus antigen 2 (BPAG2 or BP 180) Type XVIII collagen (x1 chain, also contains endostatin

Fibril-associated collagen with interrupted triple helices (FACIT) Short chain collagen with some features found in genes for fibrillar collagen and some unique features Fibril-associated collagen with interrupted triple helices (FACIT); extracellular matrix protein confined to dense and soft connective tissues; associated with mature collagen fibrils May be involved in the adherence of basement membrane to underlying connective tissue stroma Fibril-associated collagen with interrupted triple helices (FACIT); associated with type I and type II collagen fibrils and is involved in interaction of these fibrils with other matrix components Collagen found in stratified epithelium, maintains adhesion between epidermis and dermis Extracellular matrix proteins that have multiple triple-helix domains and interruptions Fibril-associated collagen with interrupted triple helices; produce adhesion to fibrils and provide interaction with other matrix component; expressed in cartilage Bind collagen

COLI 8A1

COL19A1

Type XIX collagen oL1 chain

COLLAR COL4A3BP

Type I collage, (x1 receptor Collagen type IV, ct3 (Goodpasture antigen binding protein) Type XII collagen ct l-like Collagen-like tail subunit of end plate acetylcholinesterase

COL12A1L COLQ

COMP

Cartilage oligomeric matrix protein

Anchor acetylcholinesterase to the underlying basal lamina, acetylcholinesterase-associated collagen Involved in calcium binding

Disease

Schmid metaphyseal chondrodysplasia Stickler syndrome; Marshall syndrome; murine Col 11 al - / - perinatal lethal with abnormalities in cartilage of limbs, ribs, mandible, and trachea Otospondylomegaepiphyseal dysplasia (OSMED); Weissenbacher-Zweymuller syndrome; Stickler syndrome without ocular anomaly; nonsyndromic heating loss

Generalized atrophic benign epidermolysis bullosa; nonlethal variant of junctional epidermolysis bullosa

Congenital myasthenic syndrome with end plate acetylcholinesterase deficiency

Pseudoachondroplasia; multiple epiphyseal dysplasia (Fairbanks and Ribbing types) (continues)

SECTION IX ~ Gene and Molecular Controls of Limb Development

63

TABLE VIA (continued) Protein

Gene CTSK

Cathepsin K

DCN

Decorin

DDR2

EOMES

Discoidin domain receptor family member 2 Diastrophic dysplasia sulfate transporter Delta (8)-delta (7) sterol isomerase emopamil binding protein Eomesodermin

EXT1

Exostosin 1

EXT2

Exostosin 2

EXT3 FBN1

Gene not cloned yet Fibrillin 1

FBN2

Fibrillin 2

FCN1

Ficolin 1

FCN2

Ficolin 2

FCN3

Ficolin 3-Hakata antigen

FGFR1

Fibroblast growth factor receptor 1 Fibroblast growth factor receptor 2

DTDST/DTD EBP

FGFR2

FGFR3

Fibroblast growth factor receptor 3

Cellular function

Disease

Expressed in osteoclasts involved in bone remodeling and resorption Small collagen binding proteoglycan of the extracellular matrix, affects the rate of fibril formation, also binds to fibronectin and TGF-13; when expressed ectopically docorin suppresses tumor cell growth by activating the epidermal growth factor receptor Tyrosine kinases activated by collagen

Pycnodysostosis; knockout Ctsk mice pycnodysostosis

Undersulfation of proteoglycan in cartilage matrix Involved in cholesterol biosynthesis

Diastrophic dysplasia; achondrogenesis type IB; atelosteogenesis type II X-linked dominant Conradi-HianermannHapple chondrodysplasia punctata; murine male 'tattered' (Td) X-linked Murine-targeted disruption of Eomes results in mice deficient in Eomes and arrest at the blastocyst stage Multiple exostoses type I; Langer-Giedion microdeletion syndrome; chondrosarcoma (loss of heterozygosity) Multiple exostoses type II; chondrosarcoma (loss of heterozygosity) Multiple exostoses type III Marfan syndrome; Shprintzen-Goldberg craniosynostosis syndrome

Essential for mesoderm formation and trophoblast development Involved in alteration of synthesis and display of cell surface heparan sulfate glycosaminoglycan Involved in regulation of bone growth

Major constituent of extracellular microfibrils found in elastic and nonelastic connective tissue Direct the assembly of elastic fibers during early embryogenesis Containing collagen-fibrinogen domain Lectin containing collagenfibrinogen Containing collagen-fibrinogen domain Involved in tyrosine kinase activation and signal transduction Involved in tyrosine kinase activation and signal transduction

Involved in tyrosine kinase activation and signal transduction; inhibits osteogenesis

Congenital contractural arachnodactyly (Beal syndrome)

Pfeiffer syndrome Pfeiffer syndrome; Apert syndrome; Crouzon syndrome; Jackson-Weiss syndrome; Beare-Stevenson cutis gyrata syndrome; Saethre-Chotzen syndrome Achondroplasia; hypochondroplasia; thanatophoric dysplasia type I and type II; Crouzon syndrome with acanthosis nigricans; Muenke nonsyndromic coronal craniosynostosis; severe achondroplasia, developmental delay, acanthosis nigricans syndrome; SaethreChotzen syndrome (continues)

64

CHAPTER I ~ Developmental Bone Biology TABLE VIA (continued) Gene

Protein

FN1

Fibronectin 1

FMOD

Fibromodulin

FUCA1

ct-L-Fucosidase 1

GDFIO

Growth differentiation factor 10, BMP 3b-precursor, bone inducing protein, member of transforming growth factor[3 superfamily Member of GLI-Kruppel family, a zinc finger transcription factor [3-Galactosidase 1

GL13

GM1/GLB1

Cellular function

Bind cell surfaces and collagen, fibrin, heparin DNA, and actin; involved in chell adhesion, cell motility, opsonization, wound healing, and maintenance of cell shape Involvement in the assembly of extracellular matrix by virtue of ability to bind to interact with type I and type II collagen fibrils Enzyme degrades fucosidase, the accumulation results in lysosomal storage Induce enchondral bone formation

Ehlers-Danlos type X may be associated with fibronectin deficiency

Involved in vertebrate development of limbs and other tissues

Postaxial polydactyly type IA; Greig cephalopolysyndactyly; Pallister-Hall syndrome GM1 gangliosidosis; Morquio type B syndrome Pseudohypoparathyroidism; McCuneAlbright polyostotic fibrous dysplasia

Enzyme degrades mucopolysaccharides

Guanine nucleotide binding protein, o~ stimulating activity polypeptide 1 N-Acetylglucosamine 6-sulfate sulfatase [3-Glucuronidase

Involved in regulating activity of parathyroid hormone sensitive adenylate cyclase Enzyme degrades heparan and keratan sulfate Enzyme degrades mucopolysaccharides

Homeobox A1; member of a family of transcription factors Homeobox A2; member of a family of transcription factors

Determine embryonic cell fate

HOXA3

Homeobox A3; member of a family of transcription factors

Determine embryonic cell fate

HOXA 4

Homeobox A4; member of a family of transcription factors Homeobox A5; member of a family of transcription factors

Determine embryonic cell fate

GNAS1

GNS GUSB

HOXA1

HOXA2

HOXA5

Disease

Determine embryonic cell fate

Fucosidosis (dysostosis multiplex)

Sanfilippo syndrome D or mucopolysaccharidosis type III Sly syndrome or mucopolysaccharidosis type VII; canine G u s b - / - S l y syndrome

Murine H o x a - / - embryonic mice show defective axonal path-finding, resulting in loss of cochlear nuclei and enlargement of lateral part of cerebellum Murine homozygous null H o x a 3 - / embryos show complete absence of thymus and alteration of hyoid cartilage; complementation with hoxd3 protein restores thymus and corrects alteration of hyoid cartilage, suggesting that Hoxa3 and Hoxd3 can carry out identical biological functions

Determine embryonic cell fate

(continues)

SECTION IX ~ Gene and Molecular Controls of Limb Development

65

TABLE VIA (continued) Gene

Protein

Cellular function

Disease

HOXA6

Homeobox A6; member of a family of transcription factors

Determine embryonic cell fate

HOXA 7

Homeobox A7; member of a family of transcription factors

Determine embryonic cell fate

HOXA9

Homeobox A9; member of a family of transcription factors

Determine embryonic cell fate

HOXAI O

Homeobox A10; member of a family of transcription factors

Determine embryonic cell fate

Murine-targeted disruption of Hoxal0 results in homozygous H o x a l 0 - / male and female mice with lumbar vertebral abnormalities and severe fertility defects

HOXA11

Homeobox A11; member of a family of transcription factors

Determine embryonic cell fate; pattern the posterior region of the vertebrate embryo and the appendicular skeleton

Murine-targeted disruption of Hoxal 1 and Hoxdl 1 results in double mutants with abnormal phenotype not apparent in mice homozygous for individual mutations; these double mutants have virtual absence of radius and ulna of the forelimbs, homeotic transformation of the axial skeleton, and kidney defects; these anomalies suggest that paralogous Hox genes function together to specify limb outgrowth and patterning along the proximodistal axis

HOXA13

Homeobox A13; member of a family of transcription factors

Determine embryonic cell fate

Mutation of Hoxal3 results in the autosomal dominant hand-foot-uterus syndrome characterized by hand and foot dysplasia, partial duplication of the female genital tract; murine mutation of Hoxal3 results in hypodactyly due to arrest of digital arch formation

HOXB1

Homeobox B 1

Determine embryonic cell fate

HOXB2

Homeobox B2

Determine embryonic cell fate; have a determinant role in the body plan organization with Hoxa2; control dorsoventral patterning of neuronal development in the rostral hindbrain

HOXB3 HOXB4

Homeobox B3 Homeobox B4

Determine embryonic cell fate Determine embryonic cell fate

HOXB5

Homeobox B5

Determine embryonic cell fate; involved in the positioning of mouse upper limb bud

Transgenic mice generated to ectopically expressed Hoxa7 died shortly after birth with multiple craniofacial anomalies, such as cleft palate, open eyes at birth, nonfused pinna; phenotype seen in that of human retinoic acid embryopathy Several patients with myeloid leukemia were noted to have translocation t(7;11), resulting in genomic fusion of nucleoprotein gene NUP98 with Hoxa9 and suggesting that NUP98 and Hoxa9 may be involved in myeloid differentiation

(continues)

66

C H A P T E R 1 ~ Developmental Bone Biology

TABLE VIA (continued) Gene

Protein

Cellular function

Disease

Transgenic mice generated with a gain of function mutation of Hoxb6 result in early postnatal lethality and craniofacial skeletal perturbations at birth, including open eyes, micrognathia, microtia, skull bone deformities and structural and positional alterations in the vertebral columns; complete or partial absence of the supraoccipital bone and malformation of the exo-occipital and basioccipital bones are also found

HOXB6

Homeobox B6

Determine embryonic cell fate; specify positional identity along the anteriorposterior axis

HOXB7

Homeobox B7

Determine embryonic cell fate

HOXB8

Homeobox B8

Determine embryonic cell fate; involved in establishing of anteroposterior polarity of anterior mouse limb; controlled Shh through a negative feedback loop

HOXB9 HOXB13

Homeobox B9

Determine embryonic cell fate

Homeobox B 13

Determine embryonic cell fate; maintains collinearity

HOXC4

Homeobox C4

Determine embryonic cell fate; involved in mouse anterior and posterior limb patterning

HOXC5

Homeobox C5

Determine embryonic cell fate; involved in patterning of mouse anterior limb; a target for regulation by retinoic acid and HOX homeoproteins

HOXC6

Homeobox C6

HOXC8

Homeobox C8

HOXC9

Homeobox C9

HOXCIO

Homeobox C 10

HOXC11

Homeobox C 11

Determine embryonic cell fate; involved in mouse anterior and posterior limb patterning Determine embryonic cell fate; involved in mouse anterior and posterior limb patterning and innervation of the limb; may be involved in chondrocytic differentiation Determine embryonic cell fate; involved in mouse posterior limb patterning Determine embryonic cell fate; involved in mouse posterior limb patterning Determine embryonic cell fate; involved in mouse posterior limb patterning

HOXC12

Homeobox C 12

Determine embryonic cell fate

HOXC13

Homeobox C 13

Determine embryonic cell fate; may have a function common to hair, nail, and filiform papillae of tongue

Transgenic mice with overexpression of a Hoxc8 transgenic exhibit cartilage defects with accumulation of proliferating chondrocytes and reduced maturation

Murine-targeted disruption resulting in homozygous deficient Hoxc 1 3 - / - mice with defects in hair, nail, and tongue, the most striking effect being brittle hair resulting in alopecia

(continues)

SECTION IX ~ Gene and Molecular Controls of Limb Development

67

TABLE VIA (continued) Gene

Protein

Cellular function

Disease

HOXD1

Homeobox D 1

Determine embryonic cell fate

Heterozygous deletion of the Hoxd cluster (Hoxd3-Hoxdl 3) in two human patients results in developmental defects of A-P limb and genitalia with oligodactyly and penoscrotal hypoplasia

HOXD3

Homeobox D3

Determine embryonic cell fate; involved in the regulation of cell adhesion processes; required to set up physiological constriction along the previously un-subdivided gut mesoderm

Murine: homozygous null H o x a 3 - / embryos show complete absence of thymus and alteration of hyoid cartilage; complementation with Hoxd3 protein restores thymus and corrects alteration of hyoid cartilage, suggesting that Hoxa3 and Hoxd3 can carry out identical biological functions

HOXD4

Homeobox D4

Determine embryonic cell fate; required to set up physiological constriction along the previously un-subdivided gut mesoderm

Murine: construction of minicomplex with all Hoxd genes deleted except for Hoxdl and Hodx3, the latter of which was functionally impaired, results in mice homozygous for minicomplex with retarded development, death within 2 weeks, absence of ileocecal valve and aberrant pylorus of stomach; in the absence of Hoxd function, mice lack sphincters

HOXD8

Homeobox D8

Determine embryonic cell fate; expressed in genitourinary tract and may play a continuing role in adult genitourinary tract function; required to set up physiological constrictions along the previously un-subdivided gut mesoderm

Murine: construction of minicomplex with all Hoxd genes deleted except for Hoxdl and Hoxd3, the latter of which was functionally impaired, results in mice homozygous for the minicomplex with retarded development, death within 2 weeks, absence of ileocecal valve, and aberrant pylorus of stomach; in the absence of Hoxd function, mice lack sphincters

HOXD9

Homeobox D9

Determine embryonic cell fate; required to set up physiological constriction along the previously un-subdivided gut mesoderm

HOXD 10

Homeobox D 10

Determine embryonic cell fate, required to set up physiological constriction along the previously un-subdivided gut mesoderm

Murine: construction of minicomplex with all Hoxd genes deleted except for Hoxdl and Hoxd3, the latter of which was functionally impaired, results in mice homozygous for the minicomplex with retarded development, death within 2 weeks, absence of ileocecal valve, and aberrant pylorus of stomach; in the absence of Hoxd function, mice lack sphincters Murine: construction of minicomplex with all Hoxd genes deleted except for Hoxdl and Hoxd3, the latter of which was functionally impaired, results in mice homozygous for the minicomplex with retarded development, death within 2 weeks, absence of ileocecal valve, and aberrant pylorus of stomach; in the absence of Hoxd function, mice lack sphincters (continues)

68

CHAPTER 1 ~ Developmental Bone Biology TABLE VIA (continued) Gene

Protein

Cellular function

Disease

HOXD11

Homeobox D 11

Determine embryonic cell fate; required to set up physiological constriction along the previously un-subdivided gut mesoderm

Murine: construction of minicomplex with all Hoxd genes deleted except for Hoxdl and Hoxd3, the latter of which was functionally impaired, results in mice homozygous for the minicomplex with retarded development, death within 2 weeks, absence of ileocecal valve, and aberrant pylorus of stomach; in the absence of Hoxd function, mice lack sphincters

HOXD 12

Homeobox D 12

Determine embryonic cell fate; interacting with Hoxdl 3 during mouse limb development; required to set up physiological constriction along the previously un-subdivided gut mesoderm

HOXD13

Homeobox D 13

Involved in regulating patterning during limb development; may play a role in reproductive physiology; may be involved in sphincter formation

HOX11

Homeobox 11, orphan homeobox gene located outside the HOX clusters, encodes a DNA binding nuclear transcription factor

Photo-oncogene involved in tumorigenesis of T-cell lymphomaleukemia; required for survival of spleen by maintaining splenic precursors during organogenesis

HOX11 L1

Homeobox 1 l-like 1

HSPG2

Heparan sulfate proteoglycan 2, perlecan

Required for proper positional specification and differentation of cell fate of enteric neurons Interact with extracellular matrix proteins, growth factors, and receptors and influence cellular signaling; essential for cartilage and cephalic development

Murine: construction of minicomplex with all Hoxd, the latter of which was functionally impaired, results in mice homozygous for the minicomplex with retarded development, death within 2 weeks, absence of ileocecal valve, and aberrant pylorus of stomach; in the absence of Hoxd function, mice lack sphincter Synpolydactyly; murine: S P D H - / severe synpolydactyly with malformation of all 4 feet including polydactyly, syndactyly, and brachydactyly, also lack preputial glands In human patients with T-cell leukemia and chromosomal translocation involving 10q24 the proto-oncogene TCL3, also known as HOX11, becomes activated following the arrangement; murine: by targeted disruption, homozygous HOX11 null embryos have spleen formation commencing normally up to a certain stage of embryogenesis and then undergo complete and rapid resorption Murine: Enx (Hox l l L 1 ) - / - mice develop neuronal intestinal dysplasia and megacolon Murine: disruption of perlecan results in 40% of H S P G 2 - / - mice that died in embryonic day 10.5 with defective cephalic development; 60% died shortly after birth with skeletal dysplasia resembling thanatophoric dysplasia type I; only 6% H S P G 2 - / - mice had exencephaly and skeletal dysplasia

IDA/IDUA

oL-L-Iduronidase

Enzyme degrades mucopolysaccharides

Hurler syndrome (dysostosis multiplex); Scheie syndrome (dysostosis multiplex)

IDS

Iduronate 2-sulfatase

Enzyme degrades mucopolysaccharides

IGF1

Insulin-like growth factor 1, somatomedin C

Prenatal and postnatal growth regulation

Hunter syndrome Growth retardation with sensorineural deafness and mental retardation (homozygous intragenic deletion); parents short stature (homozygous intragenic deletion); Knockout murine I G F - / - profound embryonic and postnatal growth retardation with neurological defects (continues)

SECTION IX ~ Gene and Molecular Controls o f Limb D e v e l o p m e n t

69

TABLE VIA (continued) Protein

Gene

Cellular function

Disease

Overexpression (loss of imprinting) of IGF 2 in Wilm tumor; Beckwith-Wiedemann syndrome; targeted disruption of IGF2 gene in mice by line resulted in growthdeficient heterozygous progeny; homozygous mutants were indistinguishable from heterozygous affected

IGF2

Insulin-like growth factor 2, somatomedin A

Modulation of growth hormone action, stimulation of insulin action, and involvement of development and growth

ITGA1

Integrin oL1

Cell surface receptor; dimerizes with the [3 subunit to form a collagen and laminin receptor

ITGA2

Integrin oL2, CD49B; et2 subunit of very late activating protein receptor (VLA-2)

Cell surface adhesion receptor; dimerizes with the [3 subunit to form a collagen and laminin receptor

ITGA3

Integrin oL2, CD49C; oL3 subunit of very late activating protein receptor (VLA-3) Integrin ~t7, clustering of integrin and Hox genes implies parallel evolution of these gene families

Cell surface adhesion receptor; a receptor for fibronectin, laminin, and collagen

ITGA 10

Integrin et 10

Bind to type II collagen; expressed in chondrocytes of articular cartilage

ITGAM

Integrin oLM

Dimerize with ITGB2 in vitronectin, fibrinogen, von Willebrand factor, thrombospondin, fibronectin, osteopontin, and collagen

ITGAV

Integrin o~V,vitronectin receptor, cell surface adhesion receptor

ITGB1

Integrin [31, member of a family of cell surface receptors

Dimerize with ITGB3 in vitronectin, fibrinogen, von Willebrand factor, thrombospondin, fibronectin, osteopontin, bone sialoprotein, and collagen; therapeutic targets for inhibition of angiogenesis and osteoporosis Associate with oL1 or oL6 to form a laminin receptor, with oL2 to form a collagen receptor, with ct5 to form a fibronectin receptor

LAR1 LHX2

Gene not cloned yet LIM homeobox protein 2

LOX

Lysyl oxidase

ITGA7

Specific cellular receptor for the basement membrane proteins laminin-1, -2, and-4; provide indispensable linkage between muscle fibers and extracellular matrix; clustering of integrin and Hox genes implies parallel evolution of these gene families

Putative transcription factor with two LIM and one Hox domain; may be involved in the development of leukemia by formation of chimeric proteins resulting from chromosomal translocation Extracellular Cu-dependent enzymes involved in the cross-linking and maturation of collagen and elastin

Autosomal recessive congenital myopathy; murine: homozygous null I T A G 7 - / mice viable and fertile but develop progressive muscular dystrophy

Murine: ablation of ITGAV resulted in o~V null mice with 80% dead during embryogenesis but appeared normal and 20% alive with intracerebral and intestinal hemorrhages and cleft palates

Autosomal dominant Larsen syndrome Elevated levels of Lhx2 found in patients with T-cell leukemias

Menkes syndrome, X-linked cutis laxa, Ehlers-Danlos V syndrome associated with lysyl oxidase deficiency (continues)

70

CHAPTER

1 ~

Developmental Bone Biology TABLE VIA (continued)

Gene

Protein

LUM

Lumican; keratan sulfate proteoglycan

LMX1B

LIM homeobox transcription factor 113 Mothers against decapentaplegic; Drosophila homologue 1

MADH1

MADH4

Mothers against decapentaplegic; Drosophila homologue 4

MADH5

Mothers against decapentaplegic; Drosophila homologue 5

MADH6

Mothers against decapentaplegic; Drosophila homologue 6

MADH9

MA N2 B 1

Mothers against decapentaplegic; Drosophila homologue 9 oL-Mannosidase B

MAP3K7/ MAPKKK

Mitogen-activated protein kinase 7

MATN1

Matrillin 1, cartilage matrix protein

Cellular function

Play a role in the regulation of collagen assembly of fibrils; expressed at high levels in adult articular chondrocytes Dorsal-ventral patterning of vertebrate limb Involved in the BMP signaling pathway, interacting with the activated receptors, undergoing phosphorylation, translocating to nucleus after binding to MADH4 Critical mediator of TGF[3 and BMP signaling pathways; putative tumor suppressor gene mutated or deleted in pancreatic and colorectal carcinoma

Involved in the BMP signaling pathway, interacting with the activated receptors, undergoing phosphorylation, translocating to nucleus after binding to MADH4; involved in the signaling pathway by which TGF-[3 inhibits the proliferation of human hematopoietic progenitor cells Involved in the TGF-[3 and BMP signaling pathway, expressed in vascular endothelium

Involved in the TGF-[3 and BMP signaling pathway; expressed in vascular endothelium Enzyme degrades mannosidase; accumulation results in lysosomal storage Participate in the regulation of transcription by TGF-[3; involved in the BMP signaling through binding with AP13 Cartilage matrix protein is a major component of the extracellular matrix of nonarticular cartilage, it binds to collagen

Disease

Nail patella dysplasia; murine: L M X l b - / nail patella syndrome

Pancreatic carcinoma; colorectal carcinoma; juvenile polyposis coli; head and neck squamous carcinoma; murine: homozygous MADH4 mice died before embryonic day 7.5; mutant embryos have reduced size, failure to gastrulate or express a mesodermal marker and show abnormal visceral endoderm development, heterozygotes show no abnormality; murine: compound heterozygotes for MADH4 and APC can be generated by meiotic recombination; simple heterozygotes for APC develop intestinal polyps; compound heterozygotes for APC and MADH4 develop polyps, which then undergo malignant transformation

Murine: targeted disruption results in M A D H - / - mutant mice with multiple cardiovascular abnormalities such as hyperplasia of heart valves and outflow tract abnormalities

oL-Mannosidosis (dysostosis multiplex)

Rattus: rats immunized with cartilage matrix protein develop polychondritis with erosion of nonarticular cartilage

(continues)

SECTION IX 9 Gene and Molecular Controls of Limb Development

71

TABLE VIA (continued) Gene MMP2

MMP8

Protein

Matrix metalloproteinase 2; gelatinase A; 72 kDa, type IV collagenase; 72-kDa gelatinase Matrix metalloproteinase 8, neutrophil collagenase

MMP9

Matrix metalloproteinase 9, gelatinase B, 92 kDa, gelatinase, 92-kDa type IV collagenase

MMPIO

Matrix metalloproteinase 10, stromelysin 2 Matrix metalloproteinase 13, collagenase 3

MMP 13

Cellular function

Cleave type IV collagen, the major structural component of basement membranes; catalyzes extracellular matrix degradation Cleavage of interstitial collagen in the triple-helical domain; cleaves collagen type I more so than type III Cleave collagen in extracellular matrix; key regulator of growth plate angiogenesis and apoptosis of hypertrophic chondrocytes

Weak cleavage action on collagen types II, IV, and V Degrade fibrillar collagen; probable involvement in the pathophysiology of human osteoarthritis cartilage Regulate early development of limbs, heart, and teeth

MSX1/HOX7

Homeobox 7

MSX2

MSH (Drosophila) homeobox homologue 2

Involved in tissue interaction and implicated in vertebrate craniofacial development

NFYA

Nuclear transcription factor Y,

Stimulate the transcription of various genes by recognizing and binding to the CCAAT motif in promoters of of collagen, albumin, and [3-actin genes Stimulate the transcription of various genes by recognizing and binding to the CCAAT motif in promoters of collagen, albumin, and [3-actin genes Basement membrane protein tightly associated with laminin; also binds to collagen Bind and inactive members of the growth factor-[3 superfamily, signaling proteins such as the BMPs, and are essential for proper skeletal development Involved in the processing of TGF-[3 precursor to the mature form, cleave propeptide of collagen type V, oL1

O~

NFYB

Nuclear transcription factor Y,

NID

Nidogen, enactin

NOG

Noggin

PACE

Paired basic amino acid cleaving enzyme, furin, membrane-associated receptor protein Gene not cloned yet Gene not cloned yet Peptidase D, prolidase, imidodipetidase

PDB1 PDB2 PEPD

Split immunodipeptides, play an important role in collagen metabolism because of the high levels of immunoacids in collagen

Disease

Murine: homozygous M m p 9 - / - mice with null mutation exhibit delay and abnormal pattern of skeletal growth plate, vascularization, and ossification; enlargement of growth plate eventually undergoes remodeling and produces an axial skeleton of normal appearance

Autosomal dominant hypodontia; WolfHirschhom syndrome (gene deleted); murine M s x l - / - cleft palate, facial and dental abnormalities Craniosynostosis, Boston type; MSX transgenic mice perinatal lethal with premature suture closure and ectopic cranial bone

Proximal symphalangism; multiple synostoses syndrome; murine N O G - / stillborn with multiple defects and bony fusion of appendicular skeleton

Paget disease of bone type I Paget disease of bone type II Autosomal recessive peptidase deficiency

(continues)

CHAPTER 1 ~ Developmental Bone Biology

72

TABLE VIA (continued) Gene

Protein

Cellular function

Disease

Zellweger syndrome; infantile Refsum disease Zellweger syndrome Zellweger syndrome; neonatal adrenoleukodystrophy; murine P x r l - / perinatal lethal with Zellweger syndrome Zellweger syndrome Rhizomelic chondrodysplasia punctata type I X-linked hypophosphatemic rickets; murine (HYP) X-linked hypophosphatemic rickets; murine (GYR) X-linked hypophosphatemic rickets and inner ear abnormalities microdeletion syndrome

PEX1

Peroxin 1

PEX2/PXMP3 PEX5/PXR1

Peroxin 2 Peroxin 5

Involved in peroxisomal matrix protein import Involved in peroxisomal assembly Encode a receptor for proteins with type 1 peroxisomal targeting signal (PTS 1) and also mediates PTS2targeted protein import

PEX6/PAF2 PEX7

Peroxin 6 Peroxin 7

Involved in peroxisomal assembly Encode a peroxisomal PTS2 receptor

PHEX

Phosphate regulating protein with homologies to endopeptidases on X chromosome

Involved in bone development

PLOD2

Procollagen lysine, 2oxoglutarate, 5-dioxygenase 2; lysine hydroxylase 2

PLOD3

Procollagen lysine, 2oxoglutarate, 5-dioxygenase 3; lysine hydroxylase 3

PRTN3

Proteinase 3, serine proteinase, Wegener granulomatosis autoantigen Parathyroid hormone receptor or parathyroid hormonerelated protein receptor

Form hydroxylysine residues in XaaLys-Gly sequences in collagen; serve as sites for attachment for carbohydrate units and are essential for stability of intermolecular collagen cross-links Form hydroxylysine residues in XaaLys-Gly sequences in collagen; serve as sites for attachment for carbohydrate units and are essential for stability of intermolecular collagen cross-links Cleave elastin, fibronectin, laminin, vitronectin, and collagen types I, III, and IV Involved in calcium homeostasis with hypercalcemia and suppression of PTH levels and exaggerated loss of cortical bone Involved in transcription expression of HOX genes Involved in transcription expression of HOX genes Expressed in fetal cartilage

PTHRP

SCML1 SCML2 SEDL

Sex comb on midleg (Drosophila-like 1) Sex comb on midleg (Drosophila-like 2) Sedlin

SMOH

Smoothed (Drosophila homologue)

SOX9

SRY box 9

A signaling component of the sonic hedgehog patched complex; function as an oncogene in basal cell carcinoma activated by somatic mutation in sporadic nevoid basal cell carcinoma syndrome and primitive neuroectodermal tumors targeting numerous genes (HOX, BMP) Regulates COL2A1 expression and implicates abnormal regulation of COL2A1 during chondrogenesis

Associated with Wegener granulomatosis hamster; tracheal insufflation of Prtn3 causes emphysema Murk Jansen metaphyseal dysplasia

Spondylopeiphyseal dysplasia tarda, X-linked Mutation of SMOH results in sporadic basal cell carcinoma; murine: transgenic mice overexpressing SMOH develop skin lesions resembling basal cell carcinoma

Campomelic dysplasia

(continues)

SECTION IX ~ Gene and Molecular Controls o f Limb D e v e l o p m e n t

73

TABLE VIA (continued) Gene

Protein

Cellular function

Disease

SPAR C

Secreted protein, acidic, cysteine-rich; osteonectin

Bone-specific phosphoprotein; binds selectively to hydroxypatite and to collagen fibrils at distinct sites; osteonectin accounts for unique property of bone collagen to undergo calcification

TBX1

T box 1, T named after the T or brachyury gene in mouse, member of a family of transcription factors

Involved in the regulation of developmental processes

TBX2

T box 2, transcription factor

Involved in the regulation of developmental processes

TBX3

Tbox 3

Involved in posterior elements of limb development

TBX4

T box 4, transcription factor

Expressed throughout developing hindlimb buds

TBX5

T box 5

Involved in heart and anterior elements of limb development

Holt-Oram syndrome

TBX6

T box 6, transcription factor

Involved in paraxial mesoderm and somite formation; in its absence, cells destined to form posterior somites differentiate along a neuronal pathway

Murine: creation of a Tbx6 mutation resulted in irregular somites being formed in the neck region but differentiated along neuronal pathway forming neural tube-like structures that flank the axial neural tube

TBX10

T box 10, transcription factor

TBX15

T box 15, transcription factor

Involved in the ~egulation of developmental processes Involved in the regulation of developmental processes in mouse; expressed in craniofacial region and developing limbs

TBX18

T box 18, transcription factor

Involved in the regulation of developmental processes

TBX19

T box 19, transcription factor

Involved in the regulation of developmental processes

THBS1

Thrombospondin 1

THBS2

Thrombospondin 2

Glycoprotein that mediates cell to cell and cell to matrix interactions; can bind to fibrinogen, fibronectin, laminin, and type V collagen Glycoprotein that mediates cell to cell and cell to matrix interactions; can bind to fibrinogen, fibronectin, laminin, and type V collagen

THBS3

Thrombospondin 3

Glycoprotein that mediates cell to cell and cell to matrix interactions; can bind to fibrinogen, fibronectin, laminin, and type V collagen

THBS4

Thrombospondin 4

Glycoprotein that mediates cell to cell and cell to matrix interactions; can bind to fibrinogen, fibronectin, laminin, and type V collagen

May be one of the human genes deleted in DiGeorge contiguous gene deletion syndrome

Ulnar-mammary syndrome

Murine: targeted disruption resulting in deficient mice with mild variable lordosis of spine at birth and lung abnormalities with pneumonia Murine: targeted disruption resulting in homozygous deficient mice with connective abnormalities such as fragile skin with reduced tensile strength, flexibility of tail, increased bone density and cortical thickness of long bones, abnormal bleeding time

(continues)

74

CHAPTER 1 ~ Developmental Bone Bioioyy

TABLE VIA (continued) Gene TRPS1 TWIST

VDR

Protein Gene not cloned yet Basic helix-loop-helix transcription factor, homologue of Drosophila TWIST Vitamin D hormone receptor

Cellular function

Disease

Involved in craniofacial and limb development; may function as an upstream regulator of FGFRs

Trichorhinophalangeal syndrome type 1 Saethre-Chotzen syndrome; murine TWIST + / - Saethre-Chotzen syndrome

Involved in the synthesis of osteocalcin the most abundant noncollagenous protein in bone

Vitamin D resistant rickets with end organ unresponsiveness to 1,25dihydroxycholecalciferol

aAdapted from Ho NC, Jia L, Driscoll CC, Gutter EM, Francomano CA (2000). A skeletal gene database. J. Bone Miner. Res. 15:2095-2122; with permission of the American Society for Bone and Mineral Research.

E. Overview of Gene Controls of Limb Development A large number of molecules are expressed in relation to the developing limbs serving as growth factors, transcriptional regulators, and signaling molecules. Detailed investigations are now beginning to reveal the temporal and spatial aspects of molecular expression indicating the role of each in the developmental cascade and showing in particular interrelationships between the various signaling regions in general and specific molecules in particular. Some of the major factors are listed in Table VIB. Many genes encoding regulatory molecules are expressed in the AER. The genes encode two types of protein: transcriptional regulators and signaling proteins. The Hox gene family represents an example of how transcription factors control development. Lmx-1 and engrailed-1 are other transcription factors. FGF-4 is particularly important, playing a role in signaling between the AER and underlying mesenchyme. The molecules involved include dlx, Msx-1 and Msx-2, engrailed-1 (en), bone proteins BMP-2 and BMP-4, retinoic acid, retinoic acid receptor- [3, fibroblast growth factors-2 and FGF-4, and FGF receptor-1. Retinoic acid is a major signaling molecule or morphogen in the zone of polarizing activity. Mesenchymal cells in the progress zone express several genes encoding regulatory proteins such as Msx-1 and Msx-2, Evx-1, Wnt-5A, and AP2. Mxs-1 and Mxs-2, originally known as Hox-7 and -8, are two related homeobox containing genes that are widely expressed in vertebrate embryos in particular where epithelialmesenchymal interactions occur. Msx-1 is highly concentrated in the distal mesenchyme in developing limbs with weak expression in the AER whereas Msx-2 is expressed weakly in mesenchyme but strongly in AER. The mesenchymal cells provide the signals that initiate limb bud development but the molecular trigger that stimulates AER formation is not yet known. There is rapid proliferation of mesenchymal cells, which induce the overlying ectodermal cells to form the AER. Early mesodermal induction pre-limb bud stage is the subject of intensive study with

members of each of the FGF, TGF-[3, and wingless/int-1 (Wnt) families implicated at varying stages and regions. The mesenchymal cells are primed to respond to morphogens by specific genes referred to as homeobox genes. These genes are responsible for initiating regional development by establishing patterns encoding positional information. They do so by producing transcriptional factors, which are small proteins that activate genes bound to them. Processes such as cell proliferation, cell differentiation, and cell death are regulated by homeobox genes, although these appear to be overlapping functions such that combinations of homeobox genes act on several groups of cells.

F. H o m e o b o x Genes Thirty-eight Hox genes with a homeobox sequence have been identified in mammals and are present in 4 clusters on 4 chromosomes (121,200, 201). The homeobox genes directing limb development are organized into 4 clusters termed Hox a, b, c, and d in current terminology (hox 1, hox 2, hox 3, and hox 4 prior to the 1990s). Each complex has 9-11 genes spaced over 100-150 kb. Sequence comparisons between the 4 clusters indicate that they evolved from a single cluster of genes. Hox genes within each cluster have direct homologues in the other 3 clusters. There are 13 sets of related homologues called paralog groups. The entire Hox gene family generally is displayed in a 4 (Hox a, b, c, d) by 13 (paralogs) array, with the horizontal axis representing physical linkage and chromosomal position and the vertical axis representing gene similarity. A gene is represented by a letter (chromosome location) and a number (relation to paralogs) (317) (Fig. 19). Each of the 38 Hox genes that are organized in 4 different chromosomal complexes are organized in the same way with the genes in each cluster all oriented in the 5' to 3' direction of transcription. There is also remarkable conservation of gene similarity between Drosophila and vertebrate species with the Drosophila complexes referred to as HOM-C complexes. It is the Hox genes that begin to explain the molecular

SECTION IX

|174 Group:

1.

9

Gene and Molecular Controls of Limb Development

Dimt:tions o( transcription ol" ANT-C and BX-C genes

2.

3.

4.

5.

6.

7.

IE IJ

ID 1.4

IC i.3

II 12

IA I.I

8.

9.

10.

11.

Ill I.?

IH IJI

U 1.9

12.

13.

Hox A (Hox I )

Human IF .~lou~ I..6

IK I.!1

IJ 1.10

Hox B IHox 21

|174174174174

Human -~1

2H

-~11

~F

.~A

211

)1~

3D 3.4

J4L" 3,1

.~

2D

:It

3A ,I,I

3,B 3.2

.111 JdJ

.1H ~k?

.Mr

.1(;

4s 4.3

4C 4.4

,liD U

4F U

4H 4.7

41 4.8

Hox C (Hox 31 Humln .~4ouse Hox D (Hox 4)

Human ~ .Mouse 4.1

4A 4.1

4B 4.2

3' ,11

S' Direction o( mmscrilxm o( Hon series

FIGURE 19 The Hox gene family is outlined. [Reprintedfrom Scott, M. P. (1992). Cell 71:551-553, copyright 1992, with permission from Elsevier Science.] basis for structural similarity throughout the vertebrate kingdom. Although some genes have been lost, it is evident that those that have persisted have a strong degree of structural necessity. There is also a strong correlation between the physical order of genes along the chromosome and their expression along the anteroposterior axis of the embryo. The Hox genes tend to be expressed in specific regions in an ordered array of spatially restricted domains such as the neural tube, neural crest, limbs, and hind brain segments. There is also a relationship based on the time of appearance of expression during embryogenesis. Retinoic acid (RA) induces Hox gene expression along with expression of other developmental molecules. The genes at the extreme 3' ends of the clusters are activated the earliest, have the most anterior boundaries of expression, and display the highest sensitivity to RA. As one progresses in a 5' direction, each subsequent gene has a later, more posterior, and reduced RA response pattern of expression. Hox genes are actively expressed in the polarizing region in which they appear to convey patterning information. Characteristic of patterning along the limb axis are the proximodistal morphogenetic progression, a temporal sequence in the activation of Hox genes, a partially overlapping pattern of Hox expression, especially posteriorly, and concentration in the polarizing region. There are relatively few naturally occurring mutations in mouse or human Hox, genes and their function has been inferred from their position in the developmental sequence. The large number of genes and the overlapping of adjacent areas have led to difficulty in assessing specific functions. The role played by Hox genes in skeletal development is complex. There is substantial overlap in expression domains among the members of a paralogous group, and the loss of a single gene can be overcome by functions of adjacent genes. The use of transgenic mutations has led to better definition

75

of specific Hox gene roles. With single gene transformations, however, the abnormal findings often are relatively minimal. This may be because there is compensation by other Hox genes. It is felt also that there is a requirement for several genes to combine in patterning a structure such that alteration of a single component may not be sufficient to generate a complete transformation. The Hox genes are expressed in specific patterns during morphogenesis of the upper and lower limbs (84). Studies in the chick limb bud show that evolving patterns of gene expression represent temporal and spatial overlap of several distinctly regulated expression domains. Hoxd gene expression is expressed sequentially, for example, in the development of the chick limb bud. In the early phases, Hoxd-9 and Hoxd-lO genes are expressed across the AP extent of the developing bud. The next phase of limb outgrowth involves activation of Hoxd-ll and Hoxd-12 genes in progressively restricted domains in the posterior half of the bud, and finally Hoxd-13 is activated later on. It is the sonic hedgehog (Shh) gene that is involved in the regulation of the Hoxd gene expression in the distal mesenchyme. Hoxd-9 and Hoxd-lO appear proximally before sonic is expressed. Abnormal expression of any of these genes can lead to structural abnormalities. When Hoxd-ll expression was altered in limb buds, many changes were noted including a first metatarsal with a deltoid bone shape, abnormal phalangeal structure, and metatarsals abnormal in length. Many of the structural limb abnormalities seen in the human thus can be secondary to abnormalities in Hox gene expression. It is also becoming evident that gene activity has a temporal sequence that can vary in the sense that the gene may be active early, quiet down, and then reactivate later. All of the Hox genes have specific domains of expression along the anterior-posterior or primary axis of the embryo. They are also expressed in more restricted domains along the secondary axes of the embryo such as the limb buds. The paralogous genes are not expressed in similar domains in the limb bud. Among areas of common expression, the Hoxd genes are centered in the ZPA at the posterior and distal tip of the developing limb bud, whereas the Hoxa genes are expressed in restricted domains along the proximal-distal axis of the limb but are not polarized along the AP axis of the late limb bud. With increasing studies of Hox expression patterns, it appears that the boundaries of the various genes are not as specific as once suspected. The product of the Shh gene, expressed in the ZPA, along with various FGFs produced in the apical ectodermal ridge are responsible for initiating and possibly regulating Hox gene expression. Signals from the AER and ZPA are coordinated to regulate Hox gene expression in the developing limb bud. Much less is known about Hoxb and Hoxc genes during limb development. Tickle and Eichele (347) have listed the genes involved in early vertebrate limb development as reproduced in Table VIB. They also note that hundreds of other genes both known and unknown are involved.

TABLE VIB

G e n e s Expressed in Early V e r t e b r a e Limb Buds a-~

(i) Genes Expressed in Limb Mesenchyme Transcription factors encoded by Hox complexes Hoxa-lO, Hoxa-11, Hoxa-13

Progressively more 5' genes restricted to more distal mesenchyme

Hoxd-9, Hoxd-lO, Hoxd-11 Hoxb-5

Progressively more 5' genes are restricted to more posterodistal Hoxd-12, Hoxd-13 mesenchyme Expressed anteriorly in forelimbs

Hoxe-6*

Expressed anteriorly in forelimbs

Other transcription factors Posterodistal mesenchyme Restricted to distal mesenchyme Msx-1 Anterior and distal mesenchyme Msx-2 Distal mesenchyme AP-2 Retinoid receptors and retinoid binding proteins RARoL[3 Expressed at low levels throughout mesenchyme Evx-1

RAR[3 RARe/

Throughout mesenchyme but proximally enriched Expressed at low levels through mesenchyme

RXRct[3 CRABP-I CRABP-II CRABP*

Ubiquitous Distal mesenchyme Dorsally enriched, higher levels proximally Anterior to posterior protein concentration gradient with anterior high point

Putative intercellular signaling molecules Wnt 5a Abundant in distal mesenchyme, low levels proximal mesenchyme Sonic hedgehog Discrete posterior domain associated with polarizing region BMP-2 Posterior mesenchyme, approximately coextensive with polarizing region BMP-4 Anterior and posterior mesenchymal domains FGF-2* Mesenchyme beneath ectoderm and apical ectodermal ridge (ii) Genes Expressed in Apical Ectodermal Ridge Transcription factors Dix Engrailed Msx-1 Msx-2 RAR[3 Putative intercellular signaling molecules FGF-2* Enriched posteriorly FGF-4 BMP-2 BMP-4 Wnt 5a (iii) Genes Expressed in Limb Ectoderm Transcription factors Engrailed

Enriched ventrally

Putative intercellular signaling molecules FGF-2* Dorsally enriched BMP-2 Wnt 5a

Enriched ventrally

Wnt 7a

Dorsal ectoderm only

aThis is a partial list of the best categorized molecules; a comprehensive list would consist of several hundred entries. bExpression data based on in situ hybridization analysis except for gene products marked by an asterisk, which indicates that expression pattern has been determined by immunohistochemistry. CDerived from Tickle and Eichele (347).

SECTION iX ~ Gene and Molecular Controls of Limb Development

G. Specific Details Concerning Gene and Molecular Controls on Limb Development and the Hox Gene Network The Hox genes appear to be involved in the establishment of pattems in coding positional information. The vertebrate homeobox containing Hox genes appear as candidates for pattem formation genes. As development proceeds, cells in the progress zone leave it as proliferation there continues and pushes the tip of the limb bud more distally and laterally. The cells are instructed as to their positional identities, however, while they are within the progress zone. The proximaldistal progression is seen with proximal structures at hip and shoulder developing before those more distally. The information given to the proximal cells thus is different from those more distal both in terms of quality and timing. In studies of the Hox gene clusters, those genes located at the 3' extremities of the complexes are expressed earlier and for more anterior positions than genes located at the 5' position, which are expressed later and in more restricted posterior areas. Genes appear from the 3' to the 5' regions. The ordering of the genes and their subsequent expression reflect the temporal sequence of their activation during limb bud outgrowth. The Hox expression domains also are related to establishment of the midline truncal structures. Hox gene expression in the limb as it develops reflects the clustered organization. In limbs the genes located at more 5' positions are expressed in successfully more posterior-distal areas at early stages. Expression also varies along the rostrocaudal axis. The area in the limb where Hoxd expression domains all seem to overlap is the zone of polarizing activity. Cells of the progress zone will also express various combinations of Hox proteins at different times or positions within the zone. The Hox genes thus are felt to provide patterning information or positional information, which is imparted to each cell as it leaves the zone and is then maintained during the next stages of limb development. Each Hox domain overlaps with the one adjacent to it such that there is not rigid direction of one part of one bone by one gene. Hox gene expression also leads to characteristic previously recognized patterning systems: (1) the patterning system operates in a craniocaudal or proximal-distal temporal progression; (2) the temporal sequence is observed in the activation of Hox genes; (3) a partially overlapping pattern of Hox expression domains is generated with an increasing overlap in the posterodistal areas; and (4) a particular area such as the polarizing region appears to control Hox gene expression. The Hox genes appear to impart information rather than respond to it. The progressive activation of more and more 5'-located Hox genes leads to the subsequent addition of segmented structures, such as the blastema for the limb bones.

H. Expression of Hoxa and Hoxd Genes during Normal Limb Development The earliest Hoxd gene expression was the uniform activation of Hoxd-9 and Hoxd-lO along the entire anterior-

77

posterior extent of the early limb bud. Subsequently, Hoxd-11, Hoxd-12, and Hoxd-13 were activated sequentially at the posterior border of the limb bud. Hoxa activation proceeded from Hoxa-9 and Hoxa-lO through Hoxa-11 and Hoxa-13. With the exception of Hoxa-13, the Hoxa genes appear to be activated uniformly along the extent of the limb bud. The study assessed 23 different Hox genes representing all 4 clusters. This included all of the members of the Hoxa and Hoxd clusters previously reported, all members of the Hoxc cluster from paralog 4 through 11, and one member of the Hoxb cluster. Hoxc expression was different from previous descriptions of other genes. Its most prominent domain is a wedge-shaped zone restricted to the AP region of the limb. It was not oriented around a known signaling center, such as the Hoxd genes around the ZPA or the Hoxa genes around the AER. This detailed study allows for assessment of the dynamic expression patterns of the Hoxa and Hoxd genes during limb development. It was felt that there were three independently regulated phases of gene expression, seen most clearly with the Hoxd genes. During phase 1, Hoxd-9 and Hoxd-lO are expressed uniformly throughout the early mesoderm without apparent anterior-posterior bias whereas Hoxd-11 through Hoxd-13 are not expressed. During phase 2, expression of Hoxd-9 through Hoxd-13 is initiated sequentially at the posterior-distal margin of the limb. With phase 3, there is a sequel initiation of transcription of Hoxd-13 through Hoxd-lO in a temporal sequence inverted from that observed during phase 2. The timing and positioning of Hox gene expression correlate well with the previously defined proximal-distal segment development from classical embryologic studies. During specification of the upper arm-leg, they observed the nonpolar expression of Hoxd-9, Hoxd-lO, Hoxa-9, and Hoxa-lO. During specification of the lower arm and leg, sequential activation and posteriorly polarized expression of Hoxd-9 through Hoxd-13 and uniform expression of Hoxa-11 occurred. Specification of the hand-foot region begins with phase 3 Hox gene expression, with Hoxa-13 transcription activated at the posterior border of the limb bud followed by Hoxd-13 and subsequently Hoxd-12, Hoxd-11, and Hoxd-lO. Shh can influence the expression of Hox genes in the limb. The initial phase of limb outgrowth and Hox gene expression occur prior to the onset of Shh expression. The onset of phase 2 expression, however, coincides with the onset of Shh expression, and Shh also appears to drive phase 3 Hox gene expression. The authors (241) conclude that "overwhelming experimental evidence demonstrates a causal link between Hox gene expression and morphology... Hox gene expression in the limb bud affects both the condensation of skeletal precursors in the limb bud and the subsequent growth and elongation of these elements." They note, however, that there is no obvious correlation between Hox gene expression and specific detailed morphology in the hand and foot.

78

CHAPTER 1 ~ Developmental Bone Biolofy

Genes located at 3' parts are expressed earlier and from more anterior positions, whereas genes at 5' positions are expressed later and in more restricted posterior areas. The Hox genes are expressed in a cascade fashion along the anteroposterior body axis. Their expression in the developing limb bud is initiated at a time when the pattern is being specified. Hox genes are felt to encode positional information. Hoxa and Hoxd genes are expressed in limb bud undifferentiated mesenchymal cells and then are restricted to precartilage areas and perichondrium. When their expression domains are superimposed they prefigure cartilage limb models. It has been suggested that the Hoxa genes are primarily responsible for subdividing the limb along the proximal-distal axis, whereas the Hoxd gene exerts influence on the anteroposterior axis. The AER directs limb outgrowth by maintaining undifferentiated proliferating mesenchyme immediately adjacent to it in the progress zone. As the cells leave the P, their fate in a proximal-distal sense has been determined. Removal of the AER early leads to proximal limb truncations whereas later removal involves more distal or digital truncations.

I. Expression of Hoxc Genes in the Chick Limb Bud Nelson et al. (241) undertook cloning of all Hox genes expressed in the chick limb budto note their relative position in development. Their findings are reviewed next. This group describes the expression patterns of Hoxc genes in limb development for the first time. The studies were based on in situ hybridization using both whole mount and sectioned embryos at varying stages. The Hoxc clusters are expressed in the anterior-proximal portion of either wing or leg or both. The most 3' members of the cluster, Hoxc-4 and Hoxc-5, are expressed only in the wing; Hoxc-6 and Hoxc-8 are in both the wing and leg. The more 5' members of the cluster, Hoxc-9, Hoxc-lO, and Hoxc-11, are expressed only in the leg. In contrast to the changing temporal and spatial patterns of expression of the Hoxa and Hoxd genes in limb development, the Hoxc group in the anterior and proximal region maintains the same relative domains as the limb grows.

J. Expression of Hoxb Genes There was only limited expression of Hoxb genes, and the only Hoxb gene expression in the chick limb was Hoxb-9, which specifically expresses in the hind limb in the anterior portion of the developing upper and lower leg. There was also some expression adjacent to the AER.

K. Signaling Molecules along Developing Limb Axes 1. RETINOIDS Retinoic acid (RA) was the first molecule identified that profoundly affected the anteroposterior body axis and also

the pattern of the developing limb (342). Retinoic acid is known to induce tlox gene expression, including Hoxd-11, Hoxd-12, and Hoxd-13 along with FGF-4, BMP-2, and Shh gene expression. The retinoids are small hydrophobic molecules that bind and activate nuclear receptors, which are ligand-dependent transcription factors that control the expression of target genes. Retinoic acid is present in chick and mouse limb buds in particular in the zone of polarizing activity. It is suspected that the concentration of retinoic acid varies in gradient fashion to provide positional information for cell deposition patterns. When RA is applied locally to the anterior margin of a limb bud, it induces extra digit formation. The chief observations implicating retinoic acid as a local chemical mediator in normal limb development are that it occurs in a limb bud, it is enriched in the ZPA, and the concentration of retinoic acid required for digit induction is in the same range as the endogenous retinoic acid concentration (150). Retinoids have been identified as molecules that can mimic the effects of ZPA grafts on limb patterning. Several RA receptors have been identified in the developing limb. The analysis of retinoid function provides molecular insights into the role of diffusible signals in the control of limb patterning. RA is distributed unevenly across the AP axis of the limb with a concentration greater in the posterior region, including the ZPA. It appears that secretion of the RA from the ZPA passes into the rest of the limb, producing a concentration gradient that can help to define cell position and specify the anteroposterior pattern. Retinoic acid is involved not only in limb bud differentiation but also later in the developmental sequence as a major regulator of chondrocyte maturation and matrix mineralization (177). Tissue culture studies with chondrocytes established that 3-week-old cultures treated with retinoic acid rapidly increased expression of the alkaline phosphatase, osteonectin, and osteopontin genes. There is also calcium accumulation in the RA-treated cultures in younger chondrocytes that were cultured. The effect of retinoic acid was different; it did not produce the previously described findings, but rather greatly promoted cell proliferation. RA thus appears to react throughout many stages of limb development with specific effects on each. Once cartilage has been formed, the RA appears both to induce expression of late maturation genes and to activate mineralization of the cartilage matrix.

2. SONICHEDGEHOGAND INDIANHEDGEHOG Sonic hedgehog (Shh), a secreted protein, was isolated by Riddle et al. (295) in 1993 in the chick and also by Kraus et al. (198) and Echelard et al. (95). Indian hedgehog (Ihh) was isolated by Lanske et al. (209) in 1996. It is expressed temporally slightly after Shh and regulates chondrogenic differentiation, but both Shh and Ihh have similar activities. The hedgehog proteins are a family of secreted signaling molecules that help initiate cell patterning in early embryogenesis and cartilage formation in skeletal development

SECTION IX ~ Gene and Molecular Controls of Limb Development (360). Shh is a vertebrate homologue of the Drosophila segment polarity gene, which localizes in the ZPA in chick wing bud, the posterior mesenchyme of mouse limb buds, as well as in Xenopus and zebra fish (104). It has been suggested that the Shh appears to pattern the anteroposterior limb axis. Like all the other signaling molecules, Shh is expressed in many regions of the early embryo long before limb bud formation is initiated (96). Its expression (in mouse, rat, and chick) coincides with formation of the embryonic floor plate dorsal to the notochord that induces differentiation of neural cells of the central nervous system, including the ventral spinal cord. When seen in developing limbs, Shh is expressed in a cluster of posterior mesenchymal cells, as noted previously. The temporal and spatial pattern of Shh expression suggests a close association between the gene and the organizing activity within the zone of polarizing activity (ZPA), thus serving as the molecular basis of the ZPA. Shh can induce extra digit formation in the anterior margin of a limb bud. Currently it is unclear whether Shh acts as a morphogen itself or exclusively induces the expression of other signaling molecules secondarily. Shh induces FGF-4 in the AER, the local expression of Hoxd-13 (needed for limb patterning), and also local expression of BMP-2. It has also been found that retinoic acid induces expression of Shh, which then activates Hox d - l l expression. Indian hedgehog (Ihh) is expressed in the prehypertrophic chondrocytes of cartilage in which it appears to regulate the rate of hypertrophic differentiation (305). It appears to work by inducing the expression of a second signal involving PTHrP (the parathyroid hormone-related protein), which then signals to its receptor in the prehypertrophic chondrocytes. Ihh has been seen to be expressed in the developing cartilage elements. Vortkamp et al. (359), using in situ hybridization in varying stages of chick embryos studied by whole mount preparation, detected Ihh in the endoderm of the developing midgut and in the lung and cartilage of the developing long bones of the limbs. It was not expressed early in development, whereas sonic hedgehog was expressed in the posterior mesenchyme suggesting that the two had different roles and that lhh acted later in the developmental sequence in forming skeletal elements. In companion studies, it was noted that type X collagen, as known previously, was expressed only in hypertrophic cells whereas Ihh was expressed in the transitional region between proliferating hypertrophic chondrocytes and was actually excluded from hypertrophic cells. Ihh expression, therefore, preceded hypertrophic state. BMP6 overlapped both the Ihh and the type X collagen expression domains at each stage. It was felt that lhh was expressed at a specific critical stage of endochondral bone formation. They proposed that the Ihh acts normally on the perichondrium to initiate a negative feedback loop that regulates the hypertrophic differentiation process. When lhh expressing cells underwent the final differentiation steps to become hypertrophic chondrocytes, they turned off the expression of lhh thereby allowing more cells

79

to commit to the differentiation pathway. Vortkamp et al. felt that the negative feedback loop initiated by lhh was mediated by the perichondrium and in particular by the parathyroid hormone-related protein expressed there. They concluded that both Ihh and PTHrP repressed hypertrophic cartilage differentiation and also that Ihh can induce PTHrP.

3. FIBROBLASTGROWTH FACTORS (FGFs) The fibroblast growth factors (FGFs) are a group of molecules that play a major role in normal developmental and physiological processes (78, 226). The FGF gene family currently comprises nine members, all of which are structurally related and generally encode proteins with a molecular mass of 20-30 kDa. FGF-1 and FGF-2 were purified more than a decade ago and have been studied extensively since. FGFs are potent mitogens with cells of mesodermal, neuroepidermal, ectodermal, and endodermal derivatives. Their normal physiological activities in vivo include embryonic and fetal development, neovascularization, and response to wounding. Some FGFs are oncogenes, and FGF-4 is one of the most frequently identified oncogenes in cell transformation. Although nine FGF genes have been identified, different isoforms can be generated among certain of these molecules, which would indeed underlie additional functions, and further diversity occurs by a multitude of posttranslational modifications. FGFs transduce signals to the cytoplasm through a family of transmembrane receptor tyrosine kinases referred to as the FGF receptors (FGFRs), and four mammalian FGFR genes are known. Of importance to limb development is the finding, by in situ hybridization studies, that FGFs are widely expressed during development. During the development of the vertebrate limb bud FGF-4 has been found to be synthesized by the apical ectodermal ridge, primarily in its posterior half, which itself has been implicated in outgrowth of the developing limb by maintaining proliferation of underlying cells in the progress zone (243, 244). FGF-4 can substitute for the AER in maintaining the polarizing region and can provide almost all of the signals necessary for the complete outgrowth and patterning of the chick limb. FGFs have been shown not only to induce limb bud formation but also to help maintain the proliferation of limb bud mesenchyme cells. It has been shown that beads soaked in FGF-l, FGF-2, or FGF-4 placed in the presumptive flank of chick embryos induce formation of ectopic limb buds, which can then develop into complete limbs. These results suggest that local production of an FGF may initiate limb development. Cohn et al. (66) interpreted Iheir experimental data to indicate that, under the influence of ~ , lateral plate mesoderm cells continue to proliferate and that the FGF led to local activation of sonic hedgehog in cells with potential polarizing activity, thus establishing a polarizing region. A progress zone was then established and further maintained by the activation of Hoxd genes. Cells in tlie progress zones produced a signal that induced formation of a formal AER

80

CHAPTER 1

9

Developmental Bone Biology

in the overlying epithelium. The FGF thus served to establish a limb bud not only with a polarizing region but also with a progress zone and an AER. Newly induced AER then produced FGF-4, which maintained the limb bud and led to the outgrowth and subsequent normal patterning seen. More recent studies have indicated that it is FGF-8 that is responsible for the induction, initiation, and maintenance of chick limb development (71). FGF-8 gene expression occurs in the ectoderm overlying the prospective limb forming territories, after which continuing FGF-8 secretion initiates limb bud formation. This is done by promoting the expression of sonic hedgehog (Shh) in the lateral plate mesoderm. Many studies have indicated that signaling molecules from the AER regulate limb development by their influence on cells at the distal tip of the limb bud mesoderm, referred to as the progress zone. The progress zone cells are influenced by signals from the zone of polarizing activity (ZPA) region of mesoderm at the posterior margin of the limb bud and also by signals from the ectoderm coming from Shh and Wnt-7a. Evidence indicates that FGF-8 functions to induce limb formation and plays a role in the initiation of limb bud outgrowth and the establishment of subsequent limb development, including maintaining the outgrowth and pattern formation in established limb buds. FGF activity is necessary for the initiation of Shh expression at the limb bud posterior margin.

4. WNT 7A The Wnt genes encode secreted proteins that associate with cell surface and extracellular matrix and have been implicated in many developmental processes, including the regulation of cell fate in pattern formation. Wnt 7a was noted to be expressed in the dorsal ectoderm and subsequent studies showed dorsal to ventral transformations of cell fate, indicating that Wnt 7a is a dorsalizing signal (266). It was also noted to act not only in the dorsal-ventral plane but also to have some role regulating patterning along the anteriorposterior axis. Subsequent studies showed that all three axes (dorsal-ventral, proximal-distal, and anteroposterior) are intimately linked by the respective signals Wnt 7a, FGF-4, and Shh during limb development (243). 5. TRANSFORMINGGROWTH FACTORS-~ (TGF-~) Another important group of secreted signaling molecules relating to development in general and that of the skeletal system in particular is the transforming growth factor-f3 (TGF-f3) family of peptide growth factors (50, 156). The transforming growth factor-f3 superfamily consists of as many as 25 genetically related polypeptide growth factors (191). They are often subdivided into four groups: (1) TGFJ3 family (TGF-i3 1-5); (2) activin-inhibin family; (3) BMP/ Dpp/Vgl family; (4) BMPs 2-8 and the Mulleran inhibiting substance. More recent studies indicate that five distinct TGF-[3 proteins (1-5) have been characterized from multiple vertebrate sources. The TGF-~ family consists of four dis-

tinct proteins: TGF-131,-2, -3, and-5, with TGF-~34 subclassifted under the TGF-131 group. TGF-f3 localization preceded a marked increase in type II collagen mRNA expression in transitional chondrocytes, suggesting a role for TGF-f3 in the induction and synthesis of extracellular matrix. TGF-13 has a role in the coupling of new bone formation to bone and cartilage resorption during skeletal development. Bone itself represents the most abundant source of TGF-f3 in the body. It is present and regulative of bone formation, chondrocyte, and osteoblast proliferation and has also been shown to be produced by growth plate chondrocytes. It also plays a role in inhibiting osteoclast formation and activity.

6. BONE MORPHOGENETICPROTEINS (BMPs) BMPs are signaling molecules known to induce cartilage and bone differentiation but also to play an important role in early limb patterning and even earlier in generalized embryogenesis (21, 18 l, 286, 287, 347). They are part of the larger transforming growth factor-13 family. High levels of BMP-4 expression are seen in the embryo in the posterior primitive streak and in ventral mesoderm around the posterior gut and umbilical blood vessels as well as in the body well. B MP-4 thus plays a role specifying posterior and ventral mesoderm (181). Expression of BMP-4 subsequently is seen in many tissues undergoing mesenchymal-epithelial interactions. In the limb itself, there is expression of B MP-4 throughout the mesenchyme with subsequent strong expression seen in the AER. The BMPs are members of the TGF-beta family and now comprise at least 8 members. BMP-2 and BMP-4 are expressed early in limb bud formation and are present both in the ZPA and the AER. BMPs initiate cartilage and bone formation in a sequential cascade. Cartilage and bone differentiation during the endochondral sequence involves a series of steps including initiation, promotion, maintenance, modeling, and termination events. The various signaling factors are defined at the molecular level with much focus on the BMPs. Regulatory signals by the several growth factors precede and control the deposition of the structural macromolecules of the extracellular matrix, such as the collagens, proteoglycans, and other glycoproteins. The BMPs belong to a large and continually expanding transforming growth factor-~3 superfamily, as noted earlier. The BMP family itself, which now includes subtypes, has three distinct subfamilies crucial for bone development; BMP-2, BMP-4, and BMP-5 through BMP-8. Defects in bone morphogenetic proteins are being shown to induce abnormalities of skeletal morphogenesis (190). Coordinate expression of TGF-f31, -f32,-133, and -134 has been demonstrated in chick embryo chondrocytes in vivo. In a more recent study, TGF-~31, -2, and -3 have been localized to transitional and hypertrophic chondrocytes and osteoblasts, cells that are all involved in the calcification of extracellular matrix (346). The appearance of TGF-~3 in growth plate chondrocytes appears to coincide with changes in the morphology of the cells as they pass from the proliferating

SECTION IX ~ Gene and Molecular Controls of Limb Development to the hypertrophic zone. In the growth plate, TGF-[3 functions as a stimulator of proteoglycan synthesis. Other studies have shown that TGF-[31 prevents terminal differentiation of epiphyseal chondrocytes into hypertrophic cells (14). In general, the TGF-[3 molecules stimulate the formation of extracellular matrix in connective tissues by increasing the transcription of genes and coding of many extracellular matrix proteins. The largest amounts of the TGF-[3 peptides have been found in bone and cartilage matrix. Studies have shown that TGF-[31 and TGF-[32 promote the chondrogenic differentiation of chick limb mesenchymal cells in culture. Chick limb mesenchymal cells express mRNA for chick and TGF-[31, -2, and -3 during cartilage differentiation in vitro (298).

7. PARATHYROIDHORMONE (PTH), PARATHYROID HORMONE RECEPTORPROTEIN (PTHRP), AND PTH/PTHRP RECEPTOR Each of parathyroid hormones, parathyroid hormone receptor protein (PTHrP), and the PTH/PTHrP receptor play a role in modulating skeletal development (209, 359). These important autocrine-paracrine factors have been identified in the growth plate chondrocytes. The PTH and PTHrP also act mitogenically on the growth plate chondrocytes acting through the PTH/PTHrP receptor. It is also felt that PTHrP inhibits type X collagen expression. It is the Indian hedgehog (Ihh) growth factor that induces secretion of the PTHrP by the growth plate chondrocytes. The PTHrP stimulates proliferation of chondrocytes and prevents their early hypertrophy in the growth plate. The effects of Ihh and PTHrP on chondrocyte differentiation are mediated by the PTH/PTHrP receptor. 8. INSULIN-LIKEGROWTHFACTOR (IGF) The insulin-like growth factors are regulatory molecules that also play important roles in skeletal development. IGF-1 stimulates longitudinal skeletal growth during the fetal period and throughout postnatal growth until skeletal maturation. It functions by mediating growth hormone actions on the growth plate and independently has the capacity to stimulate both chondrocyte proliferation and matrix synthesis. IGF-2 appears to be active in fetal life but its postnatal role is uncertain. There is a high degree of interaction between the IGFs and FGFs. The action of IGF-1 is such that it mediates completely the effects of growth hormone, which does not act directly on the cells. The activity involves effects on the IGF-1 receptor. The effects of growth hormone are mediated by their ability to produce growth factors such as the insulin-like growth factors. Receptors for both IGF-1 and IGF-2 have been found in the growth plate. The thyroid hormones also exert their effects on growth and in particular on the growth plate via mediation of the IGFs. Thyroxine (T4) and triiodothyronine (T3) act on the proliferative and upper hypertrophic zone chondrocytes by increasing DNA synthesis in cells and also increasing cell maturation, proteoglycan and collagen synthesis, and alka-

81

line phosphatase activity. This effect on cartilage, however, is mediated by the insulin-like growth factors. The effect appears to be mediated almost exclusively via IGF-1. IGF-1 is present in the growth plate in which it mediates the action of growth hormone. 9. VITAMIN D Vitamin D is a steroid made in the skin by the action of sunlight (157). It is biologically inert and must undergo two successive hydroxylations in the liver to 25-dihydroxyvitamin D and then in the kidney to the biologically active form, 1,25dihydroxyvitamin D [1,25(OH)2D]. The renal production of 1,25(OH)2D is closely regulated by serum calcium levels through the action of parathyroid hormone (PTH) and phosphorus. The major biologic function of vitamin D is to maintain circulating levels of calcium in the normal range. This is done by two mechanisms, one involving the efficiency of the small intestine to absorb dietary calcium and the other to allow for mobilization of calcium from bone when dietary calcium alone is inadequate to maintain normal levels. The 1,25(OH)2D induces monocytic stem cells in the bone marrow to differentiate into osteoclasts, which then perform the resorptive activity. Several different metabolites of vitamin D have been identified, including 24,25(OH)2D, but it is 1,25(OH)2D that is believed to be the major if not exclusive active agent for the biologic effects of vitamin D on calcium and bone metabolism. Vitamin D is important for bone mineralization although it does not appear to participate actively in the process. Instead it promotes the mineralization of osteoid and physeal cartilage by maintaining the extracellular calcium and phosphorus concentrations in the normal range, which then results in the deposition of calcium hydroxyapatite into the matrices. However, vitamin D has been found to play a more active role in growth plate cartilage development. All target tissues for vitamin D contain a vitamin D receptor (VDR) for 1,25(OH)2D. Such receptors also have been found in growth plate cartilage for both the 1,25 and 24,25 metabolites.

L. Matrix Metalloproteinases (MMPs) and Tissue Inhibitors of Matrix Metalloproteinases (TIMPs) Gross and histologic studies have clearly defined that the processes of synthesis and resorption are both involved in long bone growth and development. Over the past two decades the molecules involved in the remodeling processes of the extracellular matrix have been and are still being identified. The proteolytic enzymes involved in the extracellular matrix degradation are grouped into a family referred to as the matrix metalloproteinases (MMPs), which play a major role in the resorption of collagen and other macromolecules in normal prenatal and postnatal development and also in pathological disorders such as malignant tumor invasion, joint destruction in rheumatoid arthritis, and resorption of

82

CHAPTER 1 ~

Developmental Bone Biology TABLE VII Matrix Metalloproteinases a.b

MMP-1 MMP-2 MMP-3 MMP-7 MMP-8 MMP-9 MMP-10

Collagenase 1, fibroblast collagenase, interstitial collagenase Gelatinase A (72-kDa gelatinase) Stromelysin 1 Matrilysin Collagenase 2 (neutrophil collagenase) Gelatinase B (92-kDa gelatinase) Stromelysin 2

MMP- 11 MMP- 12 MMP- 13 MMP- 17 MMP- 18 MMP-19 MMP-20

Stromelysin3 Macrophageelastase Collagenase3 (rat osteoblast collagenase) MT4-MMP Collagenase4 Enamelysin

aMMP-4, -5, and -6 are no longer used. bDerived from Ann NY Acad Sci xix, 1999.

periodontal structures in dental disease. The MMPs are members of a subfamily of proteinases that contain zinc and show enzymatic proteolytic activity outside the cell (30, 127, 372). The initial MMP identified was collagenase, which is now referred to as MMP-1. Four major families of MMPs are the collagenases, gelatinases, stromelysins (including matrilysin), and the membrane type MMPs (Table VII). The metalloproteinases are essential for a breakdown of the extracellular matrix in normal development in which their activity is strictly regulated by an additional set of molecules referred to as tissue inhibitors of metalloproteinases (TIMPs). Imbalances in functions of the MMPs and/or their control by the TIMPs are associated with such pathological processes as tumor metastasis, rheumatoid arthritis, and osteoarthritic cartilage in which there is excessive tissue destruction. There are three major groups of the TIMP family (TIMP-1, TIMP-2, and TIMP-3). There is considerable specificity for tissue components for the various MMPs, although there is not an absolute tissue or molecule specificity. The collagenases are the only members of the MMP family that cleave fibrillar collagens, showing activity against types I, II, and III collagen. The collagenase family includes MMP-1, MMP-8, and MMP-13. The stromelysins can degrade many extracellular proteins, including the proteoglycans (MMP-3, MMP-7, MMP- 10, and MMP- 11). Gelatinases are affected in degrading type X and type XI collagen along with some of the other collagens. The more recently defined membrane type MMPs also have proteoglycan degradative activity. Gelatinase B (MMP-9) is limited to osteoclasts.

X. CHEMISTRY OF THE EXTRA CELLULAR MATRIX The three primary molecular constituents of epiphyseal cartilage are collagens, proteoglycans, and noncollagenous proteins, most of which are glycoproteins and phosphoproteins. The connective tissues such as cartilage, bone, tendon, ligament, and fascia are composites in a material sense of insoluble fibers and soluble polymers. The principal fibers are

collagen and elastin, whereas the principal polymers are proteoglycans and glycoproteins. Tissues such as tendon, which must withstand large tensional forces, are rich in collagen, whereas those like cartilage, which are subject to compressive forces, contain high levels of proteoglycans. Type II collagen and the proteoglycan aggrecan make up 90% of the organic cartilage matrix.

A. Collagen At present, 19 genetically distinct types of collagen encoded by at least 34 genes have been defined (344) (Table VIIIA). Approximately 80% of the total body collagen is types I and II. The collagen specific for cartilage, including the epiphyseal growth plate, is type II collagen (40). Several of the defined minor collagens, so named because they are present in relatively small amounts, have been localized to cartilage in particular types IX, X, and XI. Types VI, XII, and XIV also have a small cartilage presence. Type X collagen is localized in the matrix of the hypertrophic zone and has been hypothesized to play a major role in allowing for mineralization of the cartilage matrix, which occurs at that specific level. The collagen molecule is composed of three polypeptide chains, which are defined as oLchains and are assembled into a triple helix with a coiled-coil conformation (97, 259, 260, 282, 283) (Fig. 20A). In the major fibrillar collagens, including types I, II, and III, each c~ chain is composed of approximately 1055 amino acids. In type I collagen, there are two identical oL1 chains and one distinct a2 chain assembled into the triple helix. Type II collagen is composed of three oL1 (2) chains. The primary structure of the protein is a repeating sequence of Gly-X-Y triplets with glycine (Gly) present in every third position. Glycine is the smallest amino acid and its position is crucial for appropriate folding of the molecule. Glycine thus occupies a restricted space where the three helical a chains come together in the center of the triple helix. The most common amino acid in the X position is proline and the most common in the Y position is hydroxyproline. In collagen from mammals, approximately 100 of the X and

SECTION X ~ Chemistry of the Extracellular Matrix

83

TABLE VIIIA Collagen Types and the Location of Their Genes on Human Chromosomes a Type I

Gene COLIA1 COLIA2

II III IV

COL2A1 COL3A1 COL4A1 COL4A2 COL4A3 COL4A4 COL4A5 COL4A6

V

COL5A1 COL5A2

Chromosome

Expression

17q21.3-q22 7q21.3-q2 12q13-q14 2q24.3-q31 13q34 13q34 2q35-q37 2q35-q37 Xq22 Xq22 9q34.2-q34.3 2q24.3-q31

Most connective tissues, including bone, tendon, ligaments, skin (dermis) Cartilage, vitreous humor Extensible connective tissues, e.g., skin, lung, vascular system Basement membranes

Tissues containing collagen I, quantitatively minor component

COL5A3

VI

COL6A1 COL6A2 COL6A3

VII VIII

COL7A1 COL8A1 COL8A2

IX

COL9A1 COL9A2

21q22.3 21 q22.3 2q37 3p21 3q12-q13.1 lp32.3-p34.3 6q12-q14 lp32

Most connective tissues

Anchoring fibrils Many tissues, especially endothelium Cartilage (tissues containing collagen II)

COL9A3

Hypertrophic cartilage Cartilage (tissues containing collagen II)

COL12A1

6q21-q22 lp21 6p21.2 12q13-q14

COL12A1

6

COL13A1

10q22

Tissues containing collagen I Many tissues Tissues containing collagen I Many tissues Many tissues Skin hemidesmosomes Many tissues, especially liver and kidney Rhabdomyosarcoma cells

X

COLIOA1

XI

COLllA1 COL11A2

XII XIII XIV XV XVI XVII XVIII XIX

COL14A1 COL15A1 COL16A1 COL17A1 COL18A1 COL19A1

9q21-22 lp34-35 10q34-35 21q22.3 6q12-q14

aFrom Prockop and Kivirikko (281).

100 of the Y positions are proline and hydroxyproline, respectively. They are rigid amino acids that serve to limit rotation of the polypeptide backbone and thus contribute to the stability of the triple helix. The enzyme prolyl-4-hydroxylase converts approximately 100 proline residues in each polypeptide chain to 4-hydroxyproline. Ascorbic acid is a required cofactor for the reaction. Type II collagen is composed of three oL1 (2) chains. The fibril forming collagens are synthesized first as larger precursor molecules called procollagens. Each chain originally is synthesized intracellularly as a longer component

involving both C-terminal propeptides and N-terminal propeptides referred to as pro-oL chains (Fig. 20B). The pro-oL chains undergo proteolysis with severing of either propeptide following assembly, folding, and secretion of the procollagen molecule. The pro-oL chain N and C propeptides at the amino- and carboxy-terminal ends of the molecule, respectively, are connected to the central triple-helical collagen domain by short sequences, which are themselves nontriple-helical, that contain cleavage sites for the proteinases that process procollagen to collagen in the extracellular matrix. A folding into the triple helix occurs by spontaneous

84

CHAPTER I A

*

Developmental Bone Biology

N- PROPEPTIDE

SIGNAL

a - chain

C- PROPE P T IDE

-,,,1055 aa

"~, 250 aa

PEPTIDE

N-TERMINAL PROPEPTIDE

/

i

i

C-TERMINAL PROPEPTIDE

COLLAGEN MOLECULE

I•

GIc

(Mon)n

f'~ ~i,G IcNac l

i

,

.o

,

' I

To j'", t i

/

" \, Non-triple-Helical Domain Globular Domain /

Non-triple-Helical

I

,

s

s

" s

, ,, a

'I Domain

Non-triple- Helical Domain

Triple- Helical Domain

C ..= w A w ~ G a I - G I c GaI-O

w

Products of one, two or three genes

1

Gk~ 9 G~ ,,.. . uM

- X.V~A'xgH

(Man)n GIc Nac

Trimerization initiated via carboxy-terminal domains

J. t

o

amino-terminal domain

carboxy-terminal domain

FIGURE 20 Collagen formation is shown. (A) The components of a characteristic fibrillar collagen molecule are shown [reprinted from Olsen, B. R. (1991). In "Cell Biology of Extracellular Matrix," (E. D. Hay, ed.), pp. 177-220, Plenum Press, with permission; reproduced with permission from Prockop and Guzman, "Collagen Diseases and the Biosynthesis of Collagen," Hospital Practice 1977 12(12):61. 9 1977, The McGraw-Hill Companies, Inc. Illustration by Bunji Tagawa.]. (B) The intracellular (left) and extracellular (right) steps in collagen synthesis are shown. [From Prockop (1992), New Engl. J. Med. 326:540-546, with permission. Copyright 9 1992 Massachusetts Medical Society. All rights reserved.] (C) Mechanism of posttranslational assembly of fibrillar collagen is shown. [Reprinted from Francomano (1995), Nature 9:6-8, with permission.] (D) Electron micrograph of type I collagen from osteoid.

self-assembly in a specific zipperlike fashion from the carboxy (C) terminal toward the amino (N) terminal end of the molecule (Fig. 20C). As noted previously, the stability of the triple helix greatly depends on the presence of glycine resi-

dues at every third position in each of the three pro-et chains. Once the pro-or chains have been synthesized, significant posttranslational changes occur that are integral to normal collagen formation (97, 259, 283).

SECTION X ~ C h e m i s t r y o f t h e Extracellular Matrix

F I G U R E 20 (continued) The fibrils are 70-120 nm wide and show the typical cross-banding. (E) Type IX and type XI collagen are closely related in position to the much larger type II collagen molecule. [Reprinted from Olsen, B. R. (1995). Curr. Op. Cell Biol. 7:720-727, copyright 1995, with permission from Elsevier Science.] (F) Electron micrograph of cartilage shows narrower, dispersed type II fibrils of less than 20-nm diameter with no cross-banding seen.

85

86

CHAPTER 1 9 Developmental Bone Biology

TABLE VIIIB Steps in Collagen Synthesis" Intracellular

Extracellular

1. Gene selection, transcription, mRNA processing, translation 2. Posttranslational processing Hydroxylation of proline and hydroxyproline residues to 4-hydroxyproline, 3-hydroxyproline, and hydroxylysine Glycosylation of hydroxylysine residues (a) Galactosyl hydroxylysine (b) Glycosyl galactosyl hydroxylysine Assembly of pro-or chains, glycosylation of propeptides, disulfide bonding between pro-c~ chains Folding into triple helix Secretion from cell to extracellular region 3. Cleavage of procollagen extension peptides at C- and N-terminal ends of molecule to leave collage 4. Self-assembly of fibrils 5. Intermolecular cross-linking; conversion of lysine and hydroxylysine side chains to aldehydes and reactions of aldehydes to form covalent cross-links

aAdapted from Eyre (97), Olsen (259), and Prockop et al. (282, 283).

Intracellular steps in procollagen synthesis include the following: (1) hydroxylation of the proline residues to hydroxyproline and that of lysine residues to hydroxylysine; (2) glycosylation of hydroxylysine residues to form galactosylhydroxylysine and glucosylgalactosylhydroxylysine; (3) addition of a mannose-rich oligosaccharide to one or both propeptidases; (4) association of C-terminal propeptides; and (5) disulfide bonding to enhance chain association. Each of the preceding posttranslational changes occurs in the intracellular position with subsequent extracellular changes (97). In the amino-terminal propeptides of type I procollagen, cysteine is present and forms intrachain disulfide bonds, whereas in the carboxy-terminal propeptides, the cysteine is involved in both intrachain and interchain disulfide bonds. The propeptides account for one-third of the bulk of the procollagen molecule. They serve to direct assembly of the triple helix. After the C-propeptides have associated, the triple helix begins to form in zipperlike fashion toward the N-terminus. The protein, which is assembled in the rough endoplasmic reticulum, then passes through the Golgi complex before leaving the cell. The extracellular conversion of procollagen to collagen requires two enzymes, a procollagen amino protease that removes the amino propeptides and a procollagen carboxy protease that removes the carboxy propeptides. Once cleavage of the procollagen ends occurs, the collagen molecules remain free in the extracellular matrix and then spontaneously assemble into fibrils (Fig. 20B). Extracellular processing involves (1) conversion of procollagen to collagen and (2) cross-linking, which is mediated through lysine and hydroxylysine residues (3, 97, 98, 192, 282-284). Lysyl oxidase allows for conversion of small lysyl and hydroxylysyl residues of collagen to reactive aldehydes. Two major

types of cross-link then form. The first are referred to as intramolecular cross-links that join oL chains of the same molecule and are formed by aldol condensation of two of the aldehydes. Intermolecular cross-links involve condensation between an aldehyde derived from lysine, hydroxylysine, or glycosylated hydroxylysine and the amino group of a second lysine, hydroxylysine, or glycosylated hydroxylysine. They are Schiff bases. The main reducible cross-link is dihydroxylysinonorleucine. The major cross-links in skeletal tissues are the three hydroxypyridinium cross-links derived from three hydroxylysyl residues (98). They are particularly abundant in adult bone and cartilage. The dihydroxylysinonorleucine cross-links progressively decrease with tissue maturation, whereas the hydroxypyridinium molecules increase with time (Table VIIIB). After secretion from the cell, the procollagen molecules are enzymatically cleaved to collagen and the collagen then self-assembles into fibrils. The triple-helical collagen molecule is approximately 300 nm long and 1.5 nm in diameter. Fibrils are formed by the lateral and longitudinal association of the triple-helical molecules to each other. Each type I collagen molecule relates to the adjacent molecule in a quarterstagger array, such that the composite fibril appears as a striated pattern by electron microscopy (Fig. 20D). The crossstriations represent surface staining of the cylindrical fibril, which in cross section is composed of several hundred individual collagen molecules. The hole zone is roughly 67 nm wide. The longitudinal staggering of the molecules involves slightly less than one-quarter of the length of the molecule and leaves a "hole" between the end of one triple helix and the beginning of the next. It is widely felt by some that this hole zone in type I collagen of bone provides a site for the deposition of hydroxyapatite crystals in bone formation (130).

SECTION X ~ Chemistry of the Extracellular Matrix The three polypeptide chains of the collagen molecule are called et chains, with these then coiled into a left-handed helix with about three amino acids per turn. Each of the three helical chains together are then twisted around each other into a fight-handed superhelix. These actions lead some to refer to collagen as a "coiled coil." In the mammalian collagens, approximately two-thirds of the X and Y positions are occupied by a variety of amino acids, which help provide stability during the next higher level of hierarchical organization at the fibrillar level. In summary, synthesis of collagen has both intracellular and extracellular components (281-283) (Fig. 20B and Table VIIIB). The procollagen in the molecule is assembled and then secreted from the intracellular region, whereas in the extracellular environment the procollagen is converted to collagen and then incorporated into stable, cross-linked collagen fibrils. The first act of synthesis involves transcription and then translation of the collagen as the amino acids line up in the polypeptide sequence characteristic of the protein. Initially there is a signal sequence at the amino-terminal end, which for the pro-et chain is long, containing about 100 amino acid residues. Posttranslational changes are extensive, involving both the propeptide domains and the collagen itself in the pro-et chain. The signal sequences are removed as the amino-terminal ends of the pro-a chains enter the rough endoplasmic reticulum (RER). Hydroxylation then occurs within the RER, forming hydroxyproline and hydroxylysine. These actions are mediated partially by enzymes, including prolylhydroxylase and lysylhydroxylase. The hydroxylases act only on nonhelical substrates and not on collagen, which has reached the helical conformation. The next posttranslational change relates to glycosylation. Sugar residues are added to hydroxylysyl residues. Enzymes involved here are galactosyl transferase and glucosyl transferase. The first adds galactose to the hydroxylysyl residues and the second adds glucose to the galactosylhydroxylysyl residues. Cleavage from procollagen to collagen occurs with mediation by two enzymes, procollagen amino protease to remove the amino propeptides and procollagen carboxy protease to remove the carboxy propeptides. Cross-linking then occurs in the extracellular domain.

B. Collagen Groups The collagens are divided into differing groups on the basis of slightly differing molecular constituents (260, 344, 355), which lead to different types of polymeric structures or differing structural features. These include the following. (1) Collagens capable of forming linear fibrous structures are types I, II, III, V, and XI. Their main function is to resist tensile stresses on tissues. (2) FACIT collagens. Collagen molecules found on the surface of other collagens are referred to as FACIT collagens (fibril-associated collagens with interrupted triple helices) and include types IX, XII, XIV XVI, and XIX.

87

They are associated with collagens but do not form fibrils by themselves. (3) Network forming collagens. Collagen types IV, VIII, and X are involved in the formation of sheets or protein membranes having the ability to form regular hexagonal lattice structures. Type IV is seen in basement membranes, type VII in Descemet's membrane, and type X in the hypertrophic zone cartilage matrix (although there is no evidence that type X actually assumes a membrane conformation. (4) Beaded filament forming collagen. Type VI collagen forms beaded filaments that appear to be independent and are present in many extracellular matrices, including cartilage. (5) Collagen of anchoring fibrils. Type VII collagen forms anchoring fibrils present primarily in skin linking epithelial basement membranes to the underlying stroma. (6) Collagens with a transmembrane domain. Types XIII and XVII have both a cytoplasmic and an extracellular domain. Individual collagen types rarely are found in isolation in extracellular matrices and usually are grouped together. The cartilage-specific collagens are types II, IX, X, and XI, along with the more ubiquitous collagen type VI and the most likely types XII and XIV. Although the structures of the various collagens vary and many are hybrid molecules, whereas the collagen domain may only be small in comparison to the noncollagenous part of the molecule, certain characteristics are shared by all collagens. These include the triple helix in which three left-handed helices, ct chains, twist around each other to form a fight-handed suprahelix. Each et chain contains glycine at every third residue, and approximately 10%-12% of each of the remaining X and Y residues of the repeating sequence Gly-X-Y are proline and hydroxyproline, respectively. Hydroxyproline is essential for the formation of hydrogen bonds that stabilize the helix. The fibrillar collagens are synthesized as large precursor procollagen forms with noncollagenous propeptides at the N- and C-termini. The propeptides are completely removed in type I, II, and III collagens by specific proteinases before and during self-assembly and cross-linking of the triple helix. Collagen organization in cartilage fibrils shows evidence of structural intermixture of collagens involving primarily type II but also definable amounts of types IX and XI, forming what is referred to as a heterotypic fibril (34, 114, 234, 378). Type II collagen represents about 80% of the cartilage fibril with the remaining 20% distributed between types IX and XI (Fig. 20E). Type XI fibrils are buried within the type II fibrils, indicating a firm intermixture. It is felt that type XI collagen initially forms a core filament and that type II collagen is deposited around it. When the diameter reaches about 17 nm, the structure is such that the perimeter of the fibril serves as a binding site for type IX collagen, which is thus deposited on the surface. Attachment of type IX collagen prevents further growth in the diameter of the fibril and also limits adherence of adjacent fibrils. In summary, a core of quarter-staggered molecules of the fibrous collagen type XI is coated by molecules of the fibrous type II collagen. When

88

CHAPTER 1 ~ Developmental Bone Biology

a certain diameter is reached, which appears to be in the area of 17 nm, type IX collagen is attached to the fibril surface. There are hydroxylysine-derived cross-links between type IX collagen and type II collagen. In addition, the type IX molecules are attached to chondroitin sulfate chains of variable length. The or2 (IX) chain carries a covalently attached chondroitin sulfate chain, making this collagen a proteoglycan as well. The surface type IX plus the proteoglycan helps to prevent adjacent fibrils from aggregating. The collagen fibrils of cartilage therefore are much thinner than those of bone, averaging about 20 mm in width (Fig. 20F). The dispersed state of the fibrils prevents formation of a thicker fibril with quarter-stagger array, which explains why cartilage collagen by ultrastructure is not cross-banded. Many of the more recently defined collagens are in effect complex molecules in which the collagen is only one component. Type XVIII collagen is both a collagen and a proteoglycan with definition of heparan sulfate side chains. The type XVIII collagen has typical features of a collagen with sensitivity of its core protein to collagenase digestion, but it also has the characteristics of a proteoglycan particularly heparan sulfate because it has long heparitinase-sensitive carbohydrate chains and a highly negative net charge (141). Collagen XVIII is abundant in the basal laminae and after collagen type IV is the second most common collagen in the basal laminae. At the C-terminal globular domain the collagen XVIII molecule contains the angiogenesis inhibitor endostatin (308). This segment of the type XVIII collagen molecule has been shown experimentally to inhibit endothelial cell proliferation and to suppress angiogenesis. Currently it is the focus of much investigation in terms of therapeutic potential to minimize tumor growth.

C. Detailed Review of Specific Cartilage Collagens 1. COLLAGEN TYPE II This is the major fibrous collagen of cartilage and is also present in the vitreous humor of the eye and the intervertebral discs. It represents 80-90% of the collagen content of the cartilage matrix. It is a homotrimer of three type II ct chains [oL1(II)]3. It is processed by N- and C-proteinases in the same fashion as procollagen type I. It is closely associated with type XI collagen, which accounts for less than 5% of the cartilage collagen. Type II collagen has a higher hydroxylysine:lysine ratio and higher levels of glycosylated hydroxylysine than collagen types I and III. The degree of glycosylation also contributes to the thinness of the type II collagen fibril in relation to type I.

2. COLLAGEN TYPE XI Collagen type XI is also a fibrous collagen. It is a heterotrimer of three chains [oL1 (XI)oL2(XI)oL3(XI)]. Type XI collagen has been shown to form heterotypic fibrils with

type II, and their retention of the N-terminal noncollagenous domain is felt to restrict the overall diameter of these fibrils (Fig. 20E).

3. ThE FACIT COLLAGENS: TYPES IX, XII, AND XIV Collagen fibers in cartilage form a stable meshwork to counteract the swelling pressure generated by the hydrated proteoglycan aggregates. This is done by interaction of the collagen type II fibers with other matrix components such as type IX collagen (99, 203) (Fig. 20E). Type IX collagen is found along the surface of type II collagen fibrils. It is composed of three distinct polypeptide chains [OLI(IX)oL2(IX)et3(IX)] assembled in a 1:1:1 ratio. The type IX collagen molecules have a long triple-helical arm running along the fibril and a short triple-helical arm extending into the perifibrillar space where it terminates with a globular domain and four noncollagenous domains, one of which has a glycosaminoglycan side chain. Type IX is present throughout cartilage tissue and thus is secreted by all chondrocytes. Type IX is associated with a glycosaminoglycan (GAG) side chain consisting of either chondroitin sulfate or dermatan sulfate (230). The more recently described types XII and XIV collagen are similar to type IX collagen and are felt to associate on the surface with type I collagen containing fibrils (368). The presence of glycosaminoglycans has also been connected to types XII and XIV collagen. Type XII collagen has been localized to the perichondrium at the articular surface and around cartilage canals, whereas type XIV collagen is found throughout the matrix. 4. TYPE X COLLAGEN Type X collagen is a short chain collagen that is expressed exclusively by hypertrophic chondrocytes at sites of endochondral bone formation (147, 259, 260, 276, 312, 313). Its triple-helical domain is capable of forming a hexagonal latticelike structure. It is a homotrimeric molecule [ctl (X)]3. It is seen initially in the developing embryonic long bones in the central part of the diaphysis (where chondrocyte hypertrophy initially occurs), whereas type II collagen is seen earlier and throughout the cartilage model. Type X then is expressed farther toward each end of the developing bone as cartilage maturation and hypertrophy move in those directions. Eventually it is concentrated in the physes in the matrix adjacent to the hypertrophic cells. The amino acid sequence, gene structure, and molecular organization closely resemble those of type VIII. 5. TYPE VI COLLAGEN Type VI collagen forms beaded microfibrils, which occur in cartilage and also many other tissues. It is most commonly present as a heterotrimer [ct I(VI)oL2(VI)oL3(VI)]. Its function is still unknown, but its interactive properties suggest that it may have a bridging role between cells and their extracellular

SECTION X ~ Chemistry of the Extracellular Matrix

89

Protein Core

Aii Keratan Sulfate

Chondroitin : . ~ _ , , / S u l f a t e Chains

Chains \

Protein Core

Chondroitin Sulfate Chains

Keratan Sulfate Chains

L ee

".;,.Link Protein

e

Hyal Acid

Chondroitin Sulfate Rich Region

Keratan Sulfate Rich Region

_ ~

and Hyaluronic Acid Binding Region

CS

B Aggrecan

N

Ks,,,,.l!lll},))llii),)lli))}))ll),))l))li)))))},))l))))), c CS.'DS

Decorin

N

C .

CS,'DS Biglycan

N

Fibromodulin

N

c-x2(IX)

N

i

- C

~jKS

i ~

C

C I

CS/DS

~ I

100 nrn

F I G U R E 21 (Ai, Aii) The basic structure of the proteoglycan molecule is shown. [Reprinted from Buckwalter, J. (1983). Clin. Orthop. Rel. Res. 172:207-231 9 Lippincott Williams & Wilkins, with permission.] (B) The major proteoglycan molecules of cartilage are shown. [Reprinted from Roughley, P. J., and Lee, E. R. (1994). Microscopy Res. Tech. 28:385-397, copyright 9 1994. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.]

matrix. The fibril diameters in cartilage are in the 17- to 20-nm range.

D. Proteoglycans The other major group of molecular constituents in the connective tissues and the group that plays a major role in all cartilage tissue is the proteoglycans (PG) (42, 45-47, 238, 240, 274, 275, 301,302, 326, 357). They are a diverse family of molecules made of a core protein to which is attached one or more glycosaminoglycan (GAG) side chains (Fig. 21A i, ii). The following basic molecules comprise the GAG group in cartilage: chondroitin sulfate (CS), dermatan sulfate (DS), heparan sulfate (HS), keratan sulfate (KS), and hyaluronic acid (HA) (Fig. 21B). Each of these GAGs is sulfated except for hyaluronic acid. Only the sulfated GAGs become part of a proteoglycan. Combinations of these primary molecules

form five specific proteoglycan macromolecules of hyaline cartilage: aggrecan, decorin, biglycan, fibromodulin, and type IX collagen. Aggrecan is the largest in size and most abundant by weight. The term aggregating refers to the ability of single proteoglycans to interact noncovalently with hyaluronic acid to form large proteoglycan aggregates. The aggregating proteoglycan of hyaline cartilage is aggrecan. This large aggregating CS-PG is composed of about 85% CS, 6% KS, and 7% protein and accounts for 10% of the dry weight of cartilage. The primary role of aggrecan is to provide cartilage with its osmotic properties: to swell and hydrate the collagen fibrillar framework, giving cartilage its ability to resist compressive loads. The term nonaggregating proteoglycan refers to all molecules that do not interact specifically with hyaluronic acid. This family of molecules interacts with collagen molecules. The following are in this category. Biglycan (small CS-DS

90

CHAPTER I 9

Developmental Bone Bioloyy TABLE IXA

Properties of C a r t i l a g e

Proteolycans ".b

Proteoglycan type

GAG typec

GAG no.

Protein size a

Chromosome location

Gene size (kb)

Exon no.

Aggrecan Decorin Biglycan Fibromodulin Type IX collagen

CS and KS CS-DS CS-DS LS CS-DS

> 100 1 2 4 1

2297 329 331 357 677

15 12 X

>50 >38 8

15 8 8 32

aDerived from Roughley, Eunicer, and Lee (302). bAll data refer to human proteoglycan except that for fibromodulin, which is bovine, and oL2(IX),which is chick. CAbbrevations used: CS, chondroitin sulfate; DS, dermatan sulfate; GAG, glycosaminoglycan;KS, keratan sulfate. dprotein size is represented as the number of amino acid residues present in the secreted proteoglycan.

proteoglycan) is the predominant small proteoglycan of cartilage, possessing two dermatan sulfate chains and localized in the pericellular matrix. Decorin is a small CS-DS proteoglycan with one dermatan sulfate chain. Fibromodulin (small CS-KS proteoglycan) bears many keratan sulfate chains and is present in many tissues, including cartilage where it binds to type II collagen. Type IX collagen is considered to be a proteoglycan as well because one of its chains bears a glycosaminoglycan chain. Type IX is present on the outer surface of type II collagen. Aggrecan synthesis and breakdown in the growth plate at varying depths and different stages of development have been assessed in the distal tibia of fetal calf tissue (326). The rates of aggrecan synthesis and turnover were highest in the resting-proliferative zone compared to either the upper or lower hypertrophic zones. Aggrecan gene expression in the cells of the resting-proliferative zone and the upper hypertrophic zones was similar, but the levels were reduced in the deepest cells of the lower hypertrophic zone. Approximately 90% of the newly synthesized proteoglycan and the total proteoglycan population were able to aggregate and the monomers were relatively large. Aggrecan is a complex macromolecule consisting of an extended core, which contains several distinct domains: N-terminal G-1 domain adjacent to the G-2 domain, keratan sulfate-rich region, and an extended chondroitin sulfate-rich region and a C-terminal G-3 domain. The G-1 domain binds to hyaluronate stabilized by a separate link protein such that many proteoglycans can bind to a chain of hyaluronate to mobilize themselves in a dense collagenous network. Proteoglycan synthesis, studied by [35S]-sulfate incorporation into the glycosaminoglycan chains, showed heterogeneity of synthesis through the depth of the tissue. The major proteoglycan in the growth plate is aggrecan, which is also found in high concentrations in articular cartilage. Other proteoglycans present in the growth plate in smaller amounts include biglycan, decorin, and type IX collagen (28). Assessment comparing the upper resting proliferative zones and the

lower hypertrophic zones shows that the cell volume per total tissue volume increases by 5- to 10-fold and the matrix volume per cell also increases by up to 3-fold. In the lower hypertrophic zone, the extracellular matrix may occupy only 10% of the total tissue volume compared with over 60% in the upper zones. Thus, there is an overall loss of extracellular matrix (ECM) and proteoglycan. The composition of the ECM also changes with depth. Aggrecan is distributed throughout the growth plate. Proteoglycan concentration in the ECM increases toward the lower hypertrophic zone, but the total amount of the proteoglycan per tissue volume decreases. There is little variability in the size of the proteoglycan monomer with depth. The rate of proteoglycan synthesis per cell is reduced dramatically only in the lower hypertrophic zone. The mRNA levels of type II collagen and osteonectin are also reduced in the hypertrophic zone compared to chondrocytes in the resting zone, but there are high levels of type X collagen and hypertrophic chondrocytes associated with the onset of mineralization. The proteoglycans are summarized in Tables IXA and IXB. E. G l y c o p r o t e i n s a n d N o n c o l l a g e n o u s P r o t e i n s

A large number of proteins other than collagens and the proteoglycans are present although in much smaller amounts in the extracellular matrix (11,344). These generally are referred to as noncollagenous proteins, a term that reflects their highly variable chemical nature (148). These molecules interact with cells, with other macromolecules of the extracellular matrix, and with the mineral phase of bone and calcifying cartilage. Many of these proteins are phosphorylated, leading them to be described as phosphoproteins, and many are glycosylated and thus referred to as glycoproteins. Many subsequently have been given names on the basis of their presumed function. One of the most abundant noncollagenous proteins of bone is osteonectin, a phosphorylated glycoprotein. The principal bone-associated noncollagenous proteins are anionic. Several extracellular glycoproteins have been

SECTION X ~

TABLE IXB Name

Chemistry o f the Extracellular Matrix

Proteoglycans of Extracellular M a t r i x "'b

Prior n a m e s

GAG

Decorin

PG-11, PG-S2, PG-40

DC/CS

45K (36,383)

Biglycan Fibromodulin Lumiscan Pg-Lb Aggrecan

PG-1, PG-S!

DS KS KS DS CS, KS

45K (37,983) 50K (42,200) 50K (38,640) 43K (35,854) 225-250K

Versican Neurocan Perlecan Type IX collagen

PG-400

CS CS HS CS

400K (262,744) 150K (136,000) 400K (396K) 320K

PG-H

PG-Lt

91

Core size ~

Comments

Small PG of fibrous tissue, leucine repeats Leucinerepeats Leucinerepeats KSPG of cornea, leucine repeats Leucinerepeats Large PG of cartilage, aggregates with hyaluronic acid Fibroblastlarge PG Largeaggregating PG of brain HSPG of basement membrane Three ct chains

aDerived by Vogel (357). bAbbreviations: GAG, glycosaminoglycans;DS, dermatan sulfate; CS, chondroitin sulfate; KS, keratan sulfate; HD, heparan sulfate; PG, proteoglycan. CApproximate size determined by SDS/PAGEafter enzymatic digestion of GAG chains. Figure in parentheses is MW.

isolated, and these molecules have been shown to interact both with cells and also with other macromolecules and even the inorganic phases of bone. Posttranslational modifications include phosphorylation of serine and threonine. Groups of glycoproteins in cartilage include cartilage matrix protein (CMP), cartilage oligomeric matrix protein, link protein, and matrix GLA protein. Cartilage matrix protein has been found to localize primarily at the lower part of the proliferating zone of the endochondral sequence between the cells of the upper proliferating zone and the hypertrophic zone cells. It has been suggested that CMP is a marker for postmitotic chondrocytes that will shortly progress to the hypertrophic stage (64). It is present in articular cartilage. Cartilage oligomeric matrix protein is a glycoprotein found in cartilage particularly during periods of chondrogenesis. It is localized preferentially in the pericellular territorial matrix. Other glycoproteins seen in cartilage but more prominent in bone and other noncartilage tissues include the following: "adhesive glycoproteins" interacting with cells, such as fibronectin, vitronectin, laminin, thrombospondin, von Willebrand factor, and fibrinogen; "skeletal tissue associated glycoproteins," such as bone sialoprotein, osteocalcin, osteopontin, osteonectin, osteogenin (361), and other phosphoproteins such as phosphoserine and phosphothreonine; and the "elastin associated glycoproteins," fibrillin and MAGP. Both bone sialoprotein and osteopontin have been defined in hypertrophic chondrocytes and the adjacent mineralized cartilage matrix. The fixed negative charge of many of the bone-associated molecules is due to the fact that many are rich in acidic amino acids such as aspartic and glutamic, some contain stretches of consecutive aspartic acid residues

(osteopontin), and others contain stretches of consecutive glutamic acid residues (bone sialoprotein). Other molecules contain ~/-carboxyglutamic acid residues (osteocalcin and matrix GLA protein). Other posttranslational changes include sulfation of tyrosine residues (bone sialoprotein). Many of these molecules have been implicated in calcification and calcium ion binding, although it has been difficult to specifically pinpoint function. The noncollagenous ~/-carboxyglutamic acid containing proteins (bone GLA or osteocalcin and matrix GLA protein) have been associated with the calcification process. Bone sialoprotein has a restricted tissue distribution found primarily in bone and mineralized connective tissues. It is expressed by osteoblasts at high levels. The molecule is extensively glycosylated; about one-half of its serine residues are phosphorylated and it contains extended sequences of acidic amino acids particularly glutamic acid, glycine, and aspartic acid. Matrix GLA protein is a vitamin K-dependent protein initially isolated from bone matrix but subsequently found also in cartilage. Bone GLA protein, also known as -y-carboxyglutamic acid containing protein or osteocalcin, is a low-molecular-weight protein found primarily in bones. It contains three "y-carboxy glutamic acid residues, providing the molecule with calcium binding properties. Osteocalcin binds tightly to hydroxyapatite and may play a role in regulating crystal growth. Its synthesis is also vitamin K-dependent. Osteocalcin is the most abundant noncollagenous protein in mineralized bone matrix. It is a small protein containing 49 amino acids. Three amino acids are ~/-carboxylated via mediation of vitamin K, a property that allows for its close relationship to hydroxyapatite. It is abundant in bone tis-

92

CHAPTER I ~ Developmental Bone Biology

sue making up approximately 20% of all noncollagenous protein, but its precise function is still uncertain. Bone sialoprotein 1 (BSP1) is now referred to most commonly as osteopontin (167, 171). It is expressed early in bone development at high levels at sites of bone remodeling and is also seen to bind to hydroxyapatite. Its finding in relation to bone remodeling is associated with an apparent increased attachment to osteoclasts. Bone sialoprotein (BSP) is synthesized by osteoblasts and is a glycosylated phosphoprotein rich in sialic acid.

F. Cell Surface Proteoglycans Proteoglycans have been found as abundant molecules of the cell surface where they play a major role in morphogenesis (27, 197, 319). Virtually all epithelial cells express cell surface proteoglycans. The major transmembrane proteoglycans of the cell surface are syndecans, which contain both chondroitin sulfate and heparan sulfate. Syndecan 3 is a member of the family of heparan sulfate proteoglycans, which function as extracellular matrix receptors mediating the interaction of cells with extracellular components and as signaling molecules that control cell shape, adhesion, proliferation, and differentiation. Syndecan 3 responds to AER signals, mediates cell-matrix and cell-cell interactions involved in the onset of chondrogenesis, and also plays a role in regulating epiphyseal chondrocyte proliferation during endochondral ossification.

G. Temporal and Spatial Changes in Specific Molecular Expression within the Endochondral Sequence Virtually all structural changes are or presumably will be found to be associated with changes in molecular synthesis by the participating cells. With increasing sophistication of molecular identification, the specific cascade of changes is being clarified. Indian hedgehog (Ihh) is secreted by prehypertrophic chondrocytes and also induces another factor from the perichondrium, parathyroid hormone-related protein (PTHrP). A number of changes have been identified specifically during the chondrocyte hypertrophy phase of the endochondral sequence. These include the acquisition of collagen type X synthesis capability by the hypertrophic chondrocytes, the loss of synthetic capacity for types II and IX, a decrease in some proteoglycan synthesis, and an increase in the activity of metalloproteinases as well as alkaline phosphatase. Studies on cultures of hypertrophic tibial chondrocytes helped characterize as many as 18 up-regulated genes during chondrocyte hypertrophy (247). The genes identified included translational and transcriptional regulatory factors, ribosomal proteins, the enzymes transglutaminase and glycogen phosphorylase, type X collagen, and the carbohydrate binding protein galactin. Temporal and spatial differences in expression have also been defined for the proteoglycan core protein, aggrecan, and cartilage proteoglycan

link protein (238). These were expressed in the same regions and were confined to chondrocytes of the developing skeleton and other cartilaginous structures. The highest expression was found in the lower proliferative and upper hypertrophic zones of the physeal regions, whereas the resting zones showed less expression. In addition, cartilage that was calcifying and thus close to the osteochondral junction showed no expression. The phenomenon of loss of expression in the area of calcification was found throughout the subsequent stages of skeletal development. It is felt by some that the large aggregating proteoglycan of cartilage inhibits mineralization and thus may well prevent calcification in the proliferative and upper hypertrophic zones. The final stages of endochondral ossification are associated with matrix metalloprotein expression in relation to cartilage resorption, with MMP-9 and MMP-13 seen and with angiogenic regulators such as vascular endothelial growth factor (VEGF) in relation to vascular invasion bringing in osteoprogenitor cells (103). Studies have been done to assess the synthesis of type IX collagen during skeletal formation. The type IX collagen began to accumulate at the onset of overt chondrogenesis during the early condensation phase of the process in which mesenchymal cells became closely packed prior to depositing the cartilage matrix. The type IX synthesis coincided with the production of cartilage proteoglycan core protein and type II collagen accumulation. Expression and localization of the proteoglycans biglycan and decorin were also assessed in human tissues (28). Tissue sampling involved femur, tibia, and humerus. Although both were found in developing cartilage, there was a marked difference in their location in the developing epiphyseal regions. Biglycan core protein was localized to an outer cap of prospective articular cartilage at each epiphyseal end, whereas decorin was not detected at these sites but was found within the more deeply located resting cartilage in which staining for biglycan was weak. The cartilage matrix remained unstained for both around the vascular canals, which grew in from the perichondrium to the epiphyseal cartilage regions. Within the physeal regions, there was a highly segregated staining pattern in the upper proliferative zone in which biglycan was restricted to the territorial capsules of the chondrocytes and decorin was restricted to the interterritorial matrix. Staining for both was virtually absent in the lower proliferative, hypertrophic, and mineralizing zones of the growth plates. High levels of biglycan were detected in the region in the forming growth plates and in perichondrium-derived mesenchymal inside vascular canals. Low levels of decorin were found in articular cartilage, with a very high level detected near a rim of cartilage at the peripheral subperichondral locations and in chondrocytes arranged around vascular canals. In growth plates, low levels of biglycan and lower undetectable levels of decorin were observed in proliferating cartilage and high levels in hypertrophic chondrocytes. The biglycan gene was expressed at

SECTION XI ~ M i n e r a l i z a t i o n

high levels in preosteogenic cells both in the periosteum and in morphologically undifferentiated mesenchymal cells and vascular canals thought to be recruited for development of secondary ossification. Decorin was expressed maximally at the sites of appositional growth in the subperichondrium, whereas as biglycan expression predominated at the site of formation of growth plates. Specific localization of the gene for cartilage matrix protein (CMP) has also been found within the physis. Chen et al. (64) identified that it was the postproliferative chondrocytes that make up the zone between the zones of proliferation and hypertrophy that specifically transcribe the gene for CMP. This is referred to as the zone of maturation. CMP translation products were present in the matrix surrounding the nonproliferative chondrocytes of both the zones of maturation and hypertrophy such that CMP is a marker for postmitotic chondrocytes. This provided further indication that chondrocytes in each zone reside in an extracellular matrix with a unique macromolecular composition. The results were thought to be compatible with the demonstration of distinct switches at the proliferative-maturation transition and at the maturation-hypertrophy transition during chondrocyte differentiation. The chondrocytes themselves synthesize new matrix molecules to modify their preexisting microenvironment as differentiation progresses. Type X collagen synthesis is specifically related to the hypertrophic chondrocyte, which can be shown in tissue culture to stop type X collagen synthesis, resume proliferation, and reinitiate aggrecan synthesis when they leave the hypertrophic zone. The data are consistent with "a high degree of plasticity in the chondrocyte differentiation program." The expression of collagens I, II, X, and XI along with aggrecan synthesis by bovine growth plate chondrocytes in situ and by human fetal tissue have also been defined (306, 307).

XI. M I N E R A L I Z A T I O N Mineralization is an integral part of the terminal stages of the endochondral sequence and of the formation of bone in which the newly synthesized osteoid is mineralized in both endochondral and intramembranous sequences. Calcification of the cartilage cores in the hypertrophic cell region of the endochondral sequence is an essential step in endochondral ossification at physes and secondary ossification centers. The calcification occurs in the longitudinal cartilage columns immediately adjacent to the final three or four hypertrophic cells at the lowest part of the physis. The region calcified, which is actually in the central portion of the longitudinal septae, is referred to as the interterritorial matrix. The transverse septae rarely are calcified. The calcified septae persist in the metaphysis as cartilage cores and serve as a scaffold on which the metaphyseal bone is synthesized. Vascular invasion from the metaphyseal side stops abruptly after enter-

93

ing the lowest of the hypertrophic cell lacunae, passing only two or three cells upward. The vessels bring undifferentiated mesenchymal cells with them, and these cells soon differentiate to osteoblasts and synthesize osteoid on the calcified cartilage cores. The exquisite control and patterning of mineralization in relation to the endochondral sequence in the lowermost parts of the hypertrophic zone have long been the focus of study in terms of causation. Of note is the fact that the mineralization is concentrated in the longitudinal septae of the hypertrophic zone. Studies have identified molecules specific to the physeal cartilage and even some specific to the hypertrophic zone, and these molecules have been proposed to be involved in the mineralization front. The overall cartilage matrix is composed primarily of type II collagen and proteoglycans, particularly aggrecan, but other molecules have been identified, including types IX and XI collagen, osteonectin, osteocalcin, osteopontin, and bone sialoprotein. In addition, type X collagen is localized specifically to the hypertrophic zone matrix and is not found either in cartilage, which does not calcify, or in intramembranous bone (313). The cartilage matrix in the calcifying region has matrix vesicles containing alkaline phosphatase. These are involved in the deposition of mineral although the exact mechanism is still uncertain (6, 130, 206). Release of inorganic phosphate as a result of the activity of alkaline phosphatase can lead to the displacement of proteoglycan-bound calcium and its precipitation. The C-propeptide of type II collagen becomes concentrated in the mineralizing sites. The synthesis of type II collagen and the C-propeptide together with alkaline phosphatase is regulated by vitamin D metabolites. The association of alkaline phosphatase with calcification was first reported in the 1920s (101). Studies by Poole et al. indicated that, when calcification starts in the lower hypertrophic zone, it is not initiated within the matrix vesicles and indeed most calcification occurs in focal sites in which there are no detectable vesicles (275). They are still quite important, however, because they represent the location in which alkaline phosphatase with the enzyme associated and responsible for calcification is present. Much interest has centered therefore on type X collagen and its possible relationship to mineralization. No clear answer, however, has been provided. Many have proposed a positive relationship of type X collagen to mineralization because type X collagen appears in the mineralizing region at the appropriate time and is not present elsewhere. Schmid et al. (313) point out that type X collagen has been isolated from hypertrophic cartilage before mineralization, from the mineralizing front, and from regions of calcified cartilage. Synthesis occurs by hypertrophic chondrocytes in the upper regions of the hypertrophic zone before mineral is detectable histologically. Type X collagen persists in the cartilage cores after their mineralization and after bone has been deposited upon them. Gerstenfeld and Landis (125) showed an increase

94

CHAPTER 1 ~

Developmental Bone Biolo~ty

in type X collagen by chondrocytes cultured in the presence of 13-glycerophosphate. On the other hand, Poole and Pidoux (276) have suggested that type X collagen does not enhance mineralization but rather inhibits it. Their ultrastructural studies of type X collagen could not identify its association near early mineral foci or in association with matrix vesicles. They suggested that a coating of type X collagen on the cartilage fibrils actually directed early mineral formation away from those fibills to interfibrillar sites. In close analysis type X collagen appears to be deposited preferentially in the immediate pericellular chondrocyte region in which mineralization does not occur. This finding is consistent with the observation that the initiation of calcification does not occur in the transverse septae, which are in the immediate pericellular region because of their narrowness but rather occurs in the longitudinal interterfitorial septae of the growth plate, which are relatively farther from the cell. Attention has also been directed to the proteoglycans and their relationship to mineralization primarily because they are so extensively present in the physeal areas. Buckwalter has defined the two theories of the role of aggrecans and aggregates in physeal mineralization (42, 46). The first theory is that aggrecans-aggregates bind the calcium and inhibit the formation of mineral clusters in the matrix, whereas proteoglycanases prepare the hypertrophic zone matrix for mineralization by degrading the aggrecans. The second theory also proposes that the aggrecans-aggregates bind calcium but that they act as sites of mineral formation in the hypertrophic zone by increasing the concentration of calcium. The current impression is that they inhibit mineral growth throughout the uppermost parts of the physis and that an alteration in their structure subsequently serves to enhance mineral formation in the lower hypertrophic zone. The calcium-phosphorus (CAP) mineral phase has been studied extensively in relation to the cartilage mineralization of the epiphyseal growth plate (42, 46, 294). In a study of the calcified cartilage of the epiphyseal growth plate of young calves, X-ray diffraction of the samples revealed very poorly crystalline apatite (294). Fourier transform infrared spectroscopy and 31P nuclear magnetic resonance spectroscopy revealed significant amounts of nonapatitic phosphate ions. Their concentration increased during the early stages of mineralization but then decreased as the mineral content rose. The initial studies characterized the calcium phosphate mineral phase as a very poorly crystal ion: immature calcium phosphate apatite that was rich in labile nonapatitic phosphate ions with a low concentration of carbonide ions compared with bone mineral of the same animal. These studies were performed on the very young and mostly newly deposited mineral. It is felt that the maturation process of the mineral phase of calcified cartilage appears to proceed as it does in bone mineral, but some distinctive differences were found in the concentration of ions and in the temporal changes that occurred in the mineral phase. One of the main characteristics

of the mineral phase of calcified cartilage was the relatively large concentration of nonapatitic ions. Other important characteristics involved its low carbonide content. Its difference from mature bone mineral could be based on the fact that the mineral in calcified cartilage was less mature than that in bone due to its more rapid turnover. At all stages of calcification in the cartilage, including the earliest phase, the only crystalline solid detected by X-ray diffraction was apatite, with no evidence of a nonapatitic solid phase found. There is no evidence that the very earliest solid phase of CaP was amorphous. In the final stage of mineralization the characteristics of the mineral phase in cartilage tended to more closely approximate those of bone mineral. The appropriate development of mineralization of the endochondral sequence is dependent on circulating levels of calcium and phosphorus. These in turn are uniquely determined by functions of the parathyroid hormone (PTH) and vitamin D. Disorders of vitamin D metabolism in particular are associated with major abnormalities of the endochondral sequence, comprising the disorder referred to as rickets. Vitamin D of exogenous or skin origin accumulates in the liver and is converted to the circulating form 25-dihydroxyvitamin D [25(OH)2D], which is further activated in the kidney within the renal tubular cell mitochondria to the active metabolite 1,25-dihydroxyvitamin D [1,25(OH)2D] or calcitriol. This metabolite passes from the kidney via the blood stream to the intestine, bone, elsewhere in the kidney, and other organs in which it is physiologically active. It stimulates absorption of the intestinal calcium and phosphate as well as the mobilization of calcium from bone so as to normalize the serum calcium concentration. The kidney also produces a second metabolite, 24,25-dihydroxyvitamin D [24,25(OH)zD]. The concentrations of calcium and phosphorus that form the mineralized portions of the skeleton are closely regulated by the vitamin D-PTH system, whose purpose is to maintain the concentration of these minerals in a relatively narrow range in the extracellular fluid. Mineralization of bone tissue involves the deposition of hydroxyapatite (HA) crystals of calcium phosphate in the organic matrix composed of type I collagen and a variety of noncollagenous proteins, including proteoglycans, glycoproteins, phosphoproteins, and GLA containing proteins. Although the initial mineral crystals are deposited with the collagen fibrils, there is fairly widespread acceptance that collagen itself is neither the promotor nor the inhibitor of mineralization. The primary molecular factors that appear to be involved in the mineralization process include alkaline phosphatase, proteoglycans, fibronectin, thrombospondin, and the primary noncollagenous proteins bone sialoprotein, osteopontin, osteonectin, and osteocalcin (matrix GLA protein). There are still differing schools of thought as to the initial nucleation site in relation to the collagen and to the specific molecules that trigger that nucleation. In the view of some it is the hole region within the collagen that serves as the initial focus of mineralization based on the presence of

SECTION XI 9 M i n e r a l i z a t i o n

phosphoproteins and the physical mechanical gap that allows the nucleation to occur (130). Others feel that deposition is not site-specific other than being related to the collagen fibrils and that initial nucleation is involved with the association of matrix vesicles, which are small membranous structures derived from both hypertrophic chondrocytes and osteoblasts (6). Much effort has been made over the past three decades both to identify the noncollagenous molecules of bone and to determine their temporal and spatial representation in relation to the bone formation process, specifically mineralization both in the hypertrophic chondrocyte matrix and in the early osteoid of bone. Alkaline phosphatase is a specific product of osteoblasts and is expressed at high levels during bone development (163). Osteoblasts actively synthesizing extracellular matrix are always strongly positive for alkaline phosphatase activity whereas older mature osteocytes are not. Alkaline phosphatase activity is found within the matrix vesicles in which it is felt that initial hydroxyapatite formation occurs (6). Several groups have defined the fact that fibronectin is one of the earliest and most widespread of the extracellular matrix proteins, whose expression occurs with the formation of preosteoblasts prior to calcification and which is seen in diminished amounts after mineralization has occurred. Thrombospondin is yet another glycoprotein found to bind to calcium, hydroxypatite, and osteonectin. Although studies vary from different reports, most appear to be in agreement that bone sialoprotein and alkaline phosphatase appear earliest following fibronectin in the developing mineralization cycle, with osteonectin and osteocalcin formed relatively later. Cowles et al. concluded that fibronectin is one of the earliest matrix proteins expressed, which is rapidly followed by type I collagen, bone sialoprotein, and alkaline phosphatase with osteocalcin, osteonectin, and osteopontin having synthesis coinciding with mineralization (70). Roach has also indicated that osteopontin and bone sialoprotein are localized ahead of the mineralization front, suggesting that both proteins are necessary for the initiation of bone mineralization (296). An additional study by Hultenby et al. showed that BSP was expressed to a great extent in the osteoid synthesized by invading osteoblasts during the endochondral synthesis of bone on calcified cartilage (167). Osteopontin on the other hand was most pronounced at sites different from those of BSP, with the largest amount observed in cells close to the metaphyseal-diaphyseal border in which osteoclastic bone resorption was particularly active (51, 231). Although these two proteins occur early in the synthesis of bone, they appear to have different roles with bone sialoprotein present during the initial phases of bone formation, whereas osteopontin was particularly enriched at sites of osteoclast bone resorption (296). The other phosphoproteins, osteocalcin and osteonectin, generally are not present in areas of initial crystal formation but clearly are found in areas of fully mineralized matrix. It has been postulated, therefore, that they are more

95

important for controlling the size and speed of crystal formation rather than initiating the crystals. Osteocalcin has also been found to recruit osteoclast precursors to resorption sites. The initiation of calcification begins with the formation of individual small CaP crystals within collagen fibrils and the whole zone regions. When studies were made, including the expression of osteopontin and osteocalcin in growth plate cartilage, both proteins first appeared related to calcified cartilage in the hypertrophic zone with both molecules found concentrated at the periphery of the calcified cartilage at the future bone-calcified cartilage interface (231). Both were also found throughout the mineralized bone matrix. Landis has done extensive studies of mineralization and organic matrix interaction in particular using the mineralizing leg tendons of the turkey (207). This model is analogous to the mineralization of bone but is simpler to study, and thus specific events of mineral-matrix interaction can be visualized. The final event in mineralization of vertebrate tissues is deposition of a calcium phosphate salt, hydroxyapatite, associated with the organic matrix composed primarily of collagen. The crystals are shaped as thin plates, referred to as platelets. The size of individual crystals shows that they vary from 30 to 45 nm, whereas crystal thickness is uniform at 4-6 nm. Crystals eventually fuse to form larger platelets. Extensive study has attempted to determine the initial site of crystal deposition. It has been postulated that the hole zone within the collagen fibril is the initial site of renucleation, and much of the study revolves around the size and shape of the hole region and the size and shape of the individual crystal. In a three-dimensional high-voltage electron microscopic tomography and graphic image reconstruction study by Landis et al., it was shown that the earliest mineral deposition is within or along the characteristic banding pattern of the collagen. Some crystals appear to traverse adjacent hole and overlap zones but much of the material is located principally within collagen hole zones. The three-dimensional study confirmed that the mineral is platelet or tablet-shaped. Basically all of the individual crystals lie parallel to one another. The authors concluded that the mineral consisted of irregularly shaped thin platelets rather than rod-shaped components. Width and length were variable but the crystals appeared limited in thickness. The crystals appeared preferentially in the hole zones of collagen but also in overlap zones. Eventually crystal size outgrows the dimensions of a collagen hole zone but may be accommodated by channels or grooves in collagen formed by adjacent hole zones in register. The crystal platelets have their long axes nearly parallel to one another and to the long axes of the collagen fibrils with which they are associated. As to the site of nucleation, many crystals appear preferentially in the neighborhood of the collagen hole zones, but other crystals spatially separate from those in the hole zones are present in the collagen overlap zones. Nucleation events occur at different, spatially discrete sites and regions of collagen. This observation by Landis and colleagues is interpreted to argue

96

CHAPTER 1 ~ Developmental Bone Biology

against the influence on collagen mineral interaction of other matrix components such as extracellular or matrix vesicles.

XlI. EPIPHYSEAL GROWTH A. Physeal Chondrocyte Metabolism Brighton and colleagues focused on the chondrocyte metabolism as well as structure in the physeal regions (35, 36). The zone of proliferation was characterized by high oxygen content, high glycogen storage within chondrocytes, and high levels of mitochondrial ATP production. As the physeal sequence progressed, cells in the hypertrophic zone had a low oxygen tension, progressive consumption and lowering of glycogen levels until depletion at the lowest levels, and a switching of mitochondrial metabolism with cessation of ATP formation. The matrix of the hypertrophic zone was increasingly laden with calcium.

B. Studies of Cell Proliferation in Physeal Cartilage Using Tritiated Thymidine Autoradiography The development of tritiated thymidine autoradiography and its application to the study of physeal chondrocytes allowed for the determination of individual cell contribution to physeal growth. The titrated thymidine autoradiography technique is an accurate indicator of cell proliferation patterns and serves as a way of assessing the dynamics of growth. Thymidine, a specific precursor of DNA, labels cells only in the process of DNA synthesis and allows those cells to be followed through division and early differentiation. Thymidine is taken up from the circulation quickly with labeling reaching a maximum in 1-2 hr (186, 187). Uptake during bone development previously has been demonstrated in the upper regions of the physis (187, 188, 218), in secondary ossification center and metaphyseal bone, in epiphyseal including articular chondrocytes (223,224), and in cells of the perichondrial ossification groove of Ranvier (320). Kember was one of the first to develop this technique in relation to physeal studies (186-188). His initial studies were performed in the proximal tibia of the rat. Studies were done from 1 hr to 7 days at eight time periods. Kember was able to note cell localization initially in the upper reaches of the proliferating zone of the growth plate, with subsequent labeling in progressively lower regions with concentration in the hypertrophic zone by 7 days. This study indicated not only the areas of rapid cell proliferation but also the rate of cell differentiation and passage to each of the chondrocyte cycles, including the hypertrophic region. The basic measurement with tritiated thymidine was the percentage of cells labeled in an animal sacrificed at 1 hr after injection because any of the cells in synthesis would have divided. The autoradiographic studies were a far more convenient and accu-

rate way of measuring the portion of cells in division than mitotic counting. The zone of active cell proliferation clearly was noted to be in the region of columnar or proliferating cells. He was able to identify the stem cell compartment of the germinal zone of the physis due to the fact that proliferation in this region and, thus, labeling was less than that immediately below it. The progression of the label through the various regions of the physis toward the lower hypertrophic zone with time was clearly illustrated. Whereas labeling with tritiated thymidine allowed for initial identification of proliferating cells and subsequent assessment of their passage through various sequences toward chondrocyte hypertrophy, the method did not allow for assessment of the contribution of hypertrophy itself to longitudinal growth.

C. Kinetics of Epiphyseal Growth The description of the epiphysis given so far has concentrated on molecular and supramolecular structure, but it is extremely important to document the dynamic growth phenomena because these represent the primary function of the epiphyseal region. The epiphyseal growth plate maintains its height at the same extent throughout the vast majority of the period of growth due to coordinated functions of matrix synthesis at the epiphyseal end and resorption at the metaphyseal end. The epiphysis must also increase in width, which it does by interstitial growth of the epiphyseal cartilage immediately above the growth plate and also by addition of chondrocytes from the zone of loosely packed cells within the groove of Ranvier. The narrowing by resorption in the metaphyseal region was described previously. Longitudinal growth is also partially controlled by the elastic forces exerted by the surrounding periosteum. More detailed studies have been done to outline the specific features of cell and matrix activity that underlie longitudinal growth. It has been determined that such growth is due to three specific factors that may vary from time to time during the cycle of physeal development and function. Growth, by which is meant an increase in length, is a function of (1) chondrocyte cell proliferation, (2) matrix synthesis, and (3) chondrocyte hypertrophy. In relation to changes in shape, Hunziker and Schenk (173) state that there is "a high degree of coordination between matrix remodeling and chondrocyte shape change." They feel that, during both acceleration and deceleration of linear growth, it is the changes in hypertrophic cell activities that appear to play an important regulatory role. It is not simply changes in proliferative activity alone that modulate longitudinal growth. Indeed the importance of cell hypertrophy in regulating growth rate lies in the fact that it is a much quicker mechanism than new cell production. In a careful study, they documented that a proliferating chondrocyte would need approximately 54 hr (cell cycle time) to duplicate its own volume whereas during hypertrophy the corresponding volume, increase would be achieved in a period as short as 5 hr. Hypertrophy was a far

SECTION Xll 9 Epiphyseal G r o w t h

more efficient mechanism for bringing about columnar linear growth than cell proliferation alone. Actual matrix synthesis is felt to play a role in effecting but not in regulating longitudinal growth (because it does not contribute directly to acceleration or deceleration of this process by column prolongation). Matrix increase is more related to retaining the biomechanical properties of growth plate cartilage and to integrating chondrocytes in a highly ordered fashion rather than to triggering growth itself. The duration of the hypertrophic phase at approximately 48 hr was found to remain constant irrespective of animal age or growth rate, indicating that changes in duration would not be the triggering growth factor but, rather, changes in cell shape and size. These findings were somewhat contradictory to previous assessments of longitudinal bone growth, which had been determined upon measurement of bulk parameters such as growth plate height, cell proliferation activity as determined by tritiated thymidine incorporation, or matrix production evaluated by autoradiography of matrix components. They concluded that "growth acceleration was achieved almost exclusively by cell shape modeling, namely increase in final cell height and a decrease in lateral diameter." Even during varying rates of growth, the cell proliferation rate in the longitudinal direction and net matrix production per cell remained unchanged. "Physiological increase in linear growth rate thus appears to be based principally upon a controlled structural modulation of the chondrocyte phenotype." Cell matrix production in particular "appears to play a subordinate role in regulating longitudinal bone growth rate. The duration of the hypertrophic cell activity phase remains constant at approximately 2 days under the various growth rate conditions."

D. Amount of Growth at Each Epiphyseal Plate It was recognized in the mid- 1800s that growth was not uniform at each epiphysis. Oilier undertook experimental studies to assess the contributions to growth of each end of the major long bones (256-258). He implanted a lead nail in the middle of a long bone in rabbits, chickens and lambs and then assessed their relative positions several weeks to months later. He concluded that the proximal end of the humerus grew far more extensively than the distal end but that it was the distal ends of the radius and ulna that grew far more extensively than their proximal ends. In the lower extremity, the relationships were reversed. In the femur, most of the growth was at the distal end, and in the tibia, along with the fibula, there was more growth at the proximal than at the distal end. He thus clearly understood and documented that growth was most extensive at the shoulder, wrist, and knee. He also indicated that it was the most actively growing epiphyses that fused latest. He indicated, however, that those epiphyses that provided more longitudinal growth did not do so because of the fact that they continued growing longer, but rather that they contributed more throughout the entire period of growth. He then identified the clinical importance

97

of these varying growth features in terms of resections for disease, which were commonly performed in those days. The obvious conclusion was that resections in the elbow area had relatively few consequences for longitudinal growth, whereas at the knee they would lead to quite extensive limb shortness if done during the active growing years. Similar major growth limitations would occur with early resections at shoulder and wrist. The contributions to growth of each plate in each of the long bones were established by Digby (79) in a brief publication in 1916 as follows: proximal humerus, 80%; distal humerus, 20%; proximal radius, 25%; distal radius, 75%; proximal ulna, 20%; distal ulna, 80%; proximal femur, 30%; distal femur, 70%; proximal tibia, 57%; distal tibia, 43%; proximal fibula, 60%; and distal fibula, 40%. It is widely felt that these relative amounts of growth persist at a uniform level throughout growth. More recently, Pritchett (279) has reported a detailed study of upper extremity long bone growth in 100 males and 100 females between 1 and 19 years of age and lower extremity long bone growth in 123 males and 121 females from age 7 years to skeletal maturity. His data indicate that the percentage of growth plate activity at each long bone epiphysis is not constant throughout the growth period. He used the location of the nutrient artery as the fixed point for measurement. He concluded in agreement with Digby that the proximal growth plate of the humerus contributed 80% of the growth, the distal plate 20%, the proximal radius 25%, and the distal radius 75%. He differed from Digby in attributing the proximal ulna contribution at only 15% and distal ulna at 85 %. The proximal humerus (40%) and distal radiusulna (40%) accounted for 80% of upper extremity growth with the elbow region (10%- 10%) only 20%. Pritchett was able to assess growth at various ages with the very interesting finding that growth plate activity was not proportionately constant throughout growth. In the humerus before the age of 2 years, less than 75% of growth occurs proximally, increasing to 85% at age 8 years and remaining constant at 90% after age 11 years. At the distal ulna, 85% of growth occurred, but 90% by age 5 years and 95% by age 8 years. At the distal radius, growth was 80% overall but 85% after age 5 years and 90% by age 8 years. In the lower extremity the overall growth contributions were the same as those noted by Digby. As in the upper extremity, growth plate activity was not constant throughout growth (280). Although approximately 70% of growth of the femur occurs distally in both males and females, the proportion in the distal femoral growth plate in girls varies from 60% at 7 years of age to 90% at age 14, whereas in boys the contribution of the distal femoral growth plate varies from 55% at 7 years of age to 90% at age 16. The overall contribution of the proximal tibia growth plate is approximately 57%. This varies from 50% at 7 years to 80% at age 14 in girls and from 50% at 7 years of age to 80% at age 16 in boys. The amount of growth remaining in the distal femur and proximal tibia in males and

98

CHAPTER I ~

Developmental Bone Biology

females has been documented in greatest detail by Green and Anderson (7, 8). Pritchett has also developed similar charts from his data. The amount of growth that can be expected to occur throughout childhood or what is currently remaining in the proximal femur and distal tibia can also be calculated from these charts by defining the distal femoral growth as 70%, leaving the proximal femoral as 30%. Similar calculations can be made for the distal tibia. Karrholm et al. (183) have calculated the longitudinal growth rate of the distal tibia and fibula in children from 9 years of age to skeletal maturity in 30 boys and 31 girls. The patients were studied with Roentgen stereophotogrammetric analysis (RSE) with examinations quite frequently at intervals of 1-12 months, with most patients observed at 3-6 month periods. In the distal tibia in boys and girls, the average growth rate decreased from a plateau of about 10 and 11 years, respectively. The decreasing growth rates were found somewhat later in the distal tibia than in the distal fibula in both boys and girls. The growth rates, however, were close to 0 at about the same time in distal tibia and fibula in both sexes. Growth charts were established to provide clinical guidelines. The group also calculated the average daily growth rate of differing skeletal maturity, measuring micrometer level values at distal fibula and distal tibia. Charts were produced for the calculated remaining growth in the distal fibula and distal tibia in girls from 8 years of age and in the distal fibula and distal tibia in boys from 9 years of age. The data were expressed in charts with mean values as well as one and two standard deviations above and below the mean. Several sets of growth data are available covering lengths of major long bones followed longitudinally in groups of normal patients (7, 8, 225,279, 280). These include the work of Maresh (225), who published length data for males and females separately for humerus, radius, ulna, femur, tibia, and fibula from 2 months to 18 years of age. Measurements were made between the epiphyseal plates up to and including 12 years of age and entire bones, including the epiphyses from 10 years of age upward. Green and Anderson published values for femur and tibia in males and females. Approximate estimates of growth per year in the distal femoral and proximal tibial growth plates have been described by other observers. (1) Pritchett (280): average 1.3 cm per year of growth from the distal femur (half that amount in the final 2 years of growth), and average 0.9 cm per year from the proximal tibia (half that amount in the final 2 years). (2) White and Stubbins (369): 3A" growth per year at distal femur and 88 per year at proximal tibia. Pritchett (280) has also summarized growth phenomena by reviewing concepts derived over decades of investigation going as far back as Oilier in the nineteenth century. (1) Growth contributions from each of the upper and lower extremity growth plates are not equal. One dominant growth plate makes the major contribution to length in any particular bone throughout growth: in the humerus the proximal

growth plate and in the forearm the distal radial and ulnar growth plates. In the lower extremity the major contributions are at the knee, involving the distal femoral and proximal tibial growth plates. (2) Growth is not constant at all ages. The activity of the growth plate varies with age, with the more active growth plate assuming greater importance with increasing age. (3) The nutrient foramen marking the entrance of the nutrient vessel into the cortex provides a fixed point in the long bone and growth of each end can be measured in relation to it. (4) The more active growth plate is thicker than the less active growth plate. (5) The more active growth plate closes later than the less active growth plate, and virtually all growth during the last 2 years before growth plate closure occurs at the more active growth plate. (6) The nutrient canal is directed away from the more active growth plate and toward the less active growth plate.

E. Growth Slowdown and Growth Arrest Lines (Harris Lines) Harris clearly demonstrated the presence of transverse radiodense lines across the metaphyses of growing children who had been subjected to either generalized illness or localized bone disorders from which a recovery had occurred. Harris was a British radiologist who clearly delineated the phenomenon although he was inaccurate in the histopathologic interpretation: he felt the disorder was due to the deposition of calcium within the physis itself. Since that time, studies of clinical material by Ogden (251) and experimental assessments by Siffert and Katz (327) have shown that the transverse metaphyseal line is bone and represents thickened, transversely interconnected, trabecular networks with the normal longitudinally oriented trabecular bone on either side. With the slowdown in physeal growth during the time of illness, the normal endochondral patterning is thrown off and bone accumulation at the outer reaches of the metaphysis is increased. The growth arrest line is parallel to the adjacent physis. With the slowdown or complete cessation of growth, the trabeculae become thickened and fuse with each other transversely. They contain central cores of cartilage but are primarily bone. When the child enters a recovery phase, physeal growth increases and the normal pattern of metaphyseal bone is resynthesized with the trabeculae oriented along the long axis of the bone. With increasing time between the period of growth slowdown and the time of reassessment, the physis grows away from the transverse metaphyseal line. Histologic studies show the bone thickening to be present throughout the transverse diameter of bone within the metaphysis and not to be concentrated as an inner thickened rim of the cortex. If there are recurrent episodes of poor health, then the patient will often demonstrate multiple growth arrest lines. These are seen most prominently at the distal femur and proximal tibia where growth is most rapid, and they are seen infrequently around the elbow where growth is extremely slow. They can also be seen at the prox-

SECTION Xlll ~ Responses to Mechanical Stresses

imal humerus and distal tibia where growth is relatively extensive. The growth arrest lines are seen with many types of childhood illness. On occasion they are seen, however, in the absence of serious disorders. Transverse lines related to one physeal area also frequently are seen following fractures of the adjacent bone. They can serve as valuable indicators of the state of physeal function because the distance between the physis and the growth arrest lines should be the same throughout the entire width of the bone. If there is partial or complete focal physeal arrest, early demonstration of this can be seen by angulation of the growth arrest line with diminished distance between physis and the line at the area of tethering. Whereas growth arrest lines are identified clearly on plain radiographs, they are seen with a much higher degree of resolution and in earlier states of formation on MR imaging.

XIIl. RESPONSES OF DEVELOPING BONES AND EPIPHYSES TO MECHANICAL STRESSES A. Normal Responses to Mechanical Factors The importance of mechanical or physical factors in stimulating bone development has been commented on since the early years of scientific investigation of growth phenomena. As early as 1815 Howship felt that "mechanical pressure" was the principal agent in bringing about progressive changes of structure in growing bone (165). King (1844) (189) observed: "that pressure and tension affect the evolution of all parts, scarcely requires proof at this time." In commenting on the replacement of the cartilage model of the developing bone by bone tissue itself, he indicated that "the precise site of this 'bony' deposit seems to be determined by a certain excessive degree of pressure or tension, as in the center of a cubicle bone or an epiphysis or in the middle of a parietal or cylindrical bone. The continual depositions which succeed seem also strictly determined by the directions of pressure or extension or of both." He felt that internal absorption would seem to depend on the removal of pressure from the center to the circumference. In terms of the structuring of the internal bone, he felt that "every bone has most to resist on its surface, and least internally. In proportion as the exterior is strained and excited, so is it nourished, so does it grow; while as the inner parts are the more removed from physical tensions, they are carried off by absorption." That much depends on tension seems corroborated by the final remark that "when tension is removed from the center, it becomes absorbed which explains the excavation of bone, the course of simple atrophy, and the modeling of definitive callous." The cause which fixes the precise spots of incipient fetal ossification are conceived specific. The event takes place in a solid nidus at a point where many convergent forces and pressures are concentrated. The continuance of ossification (being as it

99

were a columnar growth against gravity) follows a similar rule for it is a deposition where pressure is greatest; and whether we regard the order in which the nuclei of all the bones begin or the order of rapidity with which each one grows, the activity is evidence in dependence on the tension of the parts. The form of every bone and process of bone, and even the arrangement of every fiber of cancellus seems to me to be regulated by the above principle. It was the works of Hueter (166) and Volkmann (358), published separately in 1862, focusing on the effects of pressure on cartilage growth, and that of Wolff (374), focusing on bone structure in relation to extrinsic forces and functional patterns in his classic book in 1892, that firmly implanted the concept of mechanical or biophysical effects on bone and cartilage structure (Table X). In 1905 Parsons (268) defined "epiphyses which occur at the articular ends of long bones, by which the pressure is transmitted from bone to bone" as pressure epiphyses. He contrasted these with epiphyses into which tendons are inserted, which he referred to as "traction epiphyses" (267, 269). Traction epiphyses were noted by a projection from one end of a shaft of a long bone into which a tendon was inserted and that had a separate center of ossification. Haines (136) and Barnett and Lewis (17) felt that traction epiphyses were derived from preexisting sesamoids. As early as 1911, Schaffer (310) and Gebhardt (122) used photoelastic models to assess mechanical stresses and their relation to ossification in developing cartilaginous epiphyses. Similar studies subsequently were shown by Pauwels decades later in an effort to relate mechanical forces to bone development (270). These contributions have been reviewed by Carter and Wong (58, 59). Gebhardt concluded that the central region of the developing cartilaginous epiphysis contained the greatest accumulation of stress, which then led to the formation of the bony epiphyseal nucleus. Pauwel's studies were interpreted to indicate that Gebhardt was incorrect and concluded that the secondary ossification nucleus formed in areas of "pure and high hydrostatic pressure." More recent studies by Carter and co-workers also felt this conclusion to be inaccurate, primarily because of the use of incorrect assumptions for boundary and loading conditions. Their work will be reviewed later. Carey wrote a series of studies on the dynamics of musculoskeletal histogenesis in which he sought to relate the biophysical forces generated by growth phenomena to the structure of the tissues with which they were associated (5356). He indicated that "the biophysical aspect is as important as the biochemical one in the problem of bone origin." He was particularly interested in the effect of developing muscle on the developing skeleton. He was one of the first to attempt to assess growth from a dynamic viewpoint in which interaction of forces resulted in what he referred to as a transference of energy. The forces generated were major factors in histogenesis. He referred to his concept as the "growth motive force," which he defined as "any agency which tends to

100

CHAPTER 1 ~

Developmental Bone Biology TABLE X

Bone structure Wolff's law (1870, 1892)

Cartilage growth tissues Hueter- Volkmann "law" (1862)

" L a w s " o f Bone a n d Cartilage D e v e l o p m e n t

Extrinsic forces on a bone modify the internal trabecular architecture and the external shape "Every change in the form and function of the bones, or of their function alone, is followed by certain definite changes in their internal architecture, and equally definite secondary alterations of their external conformation, in accordance with mathematical laws." There is an inverse relationship between compressive forces along the long axis of epiphyseal growth and the rate of epiphyseal growth. Increased pressure on the concave side interferes with normal bone growth while on the convex side less than normal pressure leads to overgrowth. Pressure beyond the normal on growth cartilage leads to growth retardation.

produce a transfer of kinetic energy from an active to a less active group of cells and of potential energy from a less active to an active group, and a cellular field of differential growth until equilibrium is established." Although the work postulated mechanical effects of tissue subsets on each other, no experimental work was done and some of the conclusions, though theoretically attractive and convincingly presented, have never been validated experimentally. His work did not specifically discuss the types of mechanical sources, although it frequently mentioned their importance; it was strongly based on histologic and radiographic studies. Carey referred to active and less active zones in reference to the rate of cell division per millimeter of cross section as determined histologically. He sought to assess how the increase of cellular components and their transformation of these led to the perfection of form out of the relatively formless antecedents (by which he means undifferentiated mesenchymal cells). These were phenomena that demanded close analysis. Along with many investigators at that time, he considered cellular differentiation to be influenced by the environment in the sense that it was partly dependent upon an interaction of the developing parts before external form and internal structure were perfected. He referred to the relationship of differing tissue groups to each other in development, and he used the term "inductive" to imply that some changes were produced without cell-cell contact and "conductive" to imply that change was produced with contact. Carey thus felt that, through conduction of the developing skeletal and muscular tissues upon each other, the factor of force inherently was involved. As the skeleton underwent its most rapid growth during early embryogenesis, a tensional elongating or stretching action was bound to be exerted upon the surrounding and less actively growing continuous syncytial mesenchyme. Carey defined his theory of the relationship

between muscle and skeletal development in simple terms: (1) there is a force manifested by rapid skeletal growth; (2) this force exerts a tensional or stretching action upon the surrounding mesenchyme, influencing the first steps of myogenesis; and (3) the first differentiated muscles react upon the primordial blastemal skeleton, resulting in a definite series of changes. These are seen in the formation of the condensed cartilaginous skeleton and later, as the muscles become more developed and vigorous, serve as a physiological foundation in the formation of the osseous skeleton. Carey thus defined a degree of equilibrium of the musculature and skeleton during any stage of development established between opposing myogenic and skeletal forces. His theory indicated that "mechanically, skeletal and the related muscular tissues are inter-dependent, one relying upon the other for its initial and continued differentiation." Stated another way, he felt that force was exerted in certain regions of the embryo by the genesis of a rapidly dividing group of cells upon a less active or relatively passive group of cells. In turn, the relatively passive group reacted upon the former. This action-reaction response was shown by alterations in the rate of growth or by changes produced in the external form or internal structure of groups of cells. He derived the term "growth motive" and defined it as any agency that tends to produce a kinetic and potential energy in a cellular field of differential growth. In attempting to relate biomechanical features to changes in shape and form, Carey provided brief definitions. He defined any definite alteration in the form or dimensions of an elastic body as a strain, with the simplest strain being the linear one. Elongation was defined as the ratio of change in length to the initial length. A negative elongation or shortening was called compression; a positive elongation or lengthening was called tension. He then related skeletal de-

SECTION XIII ~ Responses t o Mechanical Stresses

velopment from the limb bud stage on to the associated myogenesis in relation to the reciprocal forces of each. The core of his argument is that mechanical forces play a major role in cell differentiation as the limb bud progresses from mesenchymal condensation, to the cartilage model of the developing bone, to formation of the primary ossification center with hypertrophy of cells in the cartilage mass, and finally to cortical bone formation. He postulates that centrally there is more cell activity within the developing cartilage mass than there is at the periphery. As the central core of the limb pushes forth more rapidly than the mesenchymal cells at the periphery, there is a tendency for the latter to be pulled out, stretched, or elongated by the former. "The traction force of the rapidly growing appendicular core exerted upon the surrounding mesenchyme is the internal stimulus of a correlated part, resulting in the elongation of the nuclei of the premuscular mass in the direction of the long axis of skeletal growth." As differential growth continues, the growth motive force becomes more and more in evidence, and there is a drawing out or stretching of the peripheral syncytial cytoplasm in the direction of skeletal growth. The myofibril is formed roughly parallel to the long axis of the developing bone. The formation of the embryonic skeletal muscles is felt to represent a definite reaction to the growth of the skeleton, and in tum these muscles "tend to restrict the growth of the skeleton in length. This is manifested by an increasing condensation of the skeletal core." Cartilage then forms within the skeletal model to provide greater stability to counteract the formation that would occur as the primitive muscles begin to contract. He thus postulates continual interplay between muscle and skeletal development. As the growth motive force of differential growth continues, the musculature becomes too vigorous even for the cartilage base, which then leads to supplanting of the cartilage model by the osseous skeleton. Carey produced a series of drawings relating to the cell and matrix deposition patterns of the developing bone and the associated muscle in an effort to show how the various strains helped modulate cell differentiation (Fig. 22). He concluded that there was a direct transference of kinetic energy from the more rapidly growing skeleton to the less actively growing primitive musculature and then a reactive transference of potential energy from the latter to the former, tending to an equilibrium state. When muscle differentiated to the state that it began functioning, there was a direct transference of kinetic energy from this tissue back to the growing skeleton, tending to retard or alter its motion or growth. The active and passive play of the muscles led first to a stiffening of the skeletal model by cartilage formation and then to further stiffening by bone formation. In regard to bone formation, he noted in particular a sub-periosteal osteogenetic and constricting cell zone at the center of a developing long bone, which quickly encircled the shaft. Formation of the osseous skeleton was part of a feedback mechanism based on the stimulus of the functionally active thigh muscles and

101

the stimulus of the restriction to growth at the ends of the rapidly elongating cartilage model due to passive muscular resistance. The interactive forces also explained in a mechanical or biophysical sense why the bony epiphyses or secondary ossification center formed. The centrifugal expansion of central confined epiphyseal cartilage cells was resisted by the centripetal effect of the peripheral muscles, ligaments and tendons together producing the "adequate compression of differential growth." The epiphyseal bone thus represents a terminal pressure system segmented from the diaphysis. Carey indicates that muscle differentiation in the thigh shortly after the formation of the cartilage model of the bone is not a coincidence but a mechanically mediated cellular response to the traction or tension to which the continuous peripheral mesenchyme is subjected during the rapid growth of the femur in length. Carey then relates the rapid growth of the skeletal model of individual bones to what happens between them at the joint interface. Resistance during femoral growth, for example, is met at first at the proximal end in the centers of opposing growth for the triradiate cartilage (ilium, ischium, and pubis) and at the distal end in the centers of opposed growth for the tibia. The regions of cell concentrations between opposing developing bones are referred to as the interzone where joints eventually will form. The appearance of the joints, however, is felt by Carey to be the mechanical result of the opposed growth of contiguous accelerating growth segments. Early resistance to expansion of the femur in two directions, by associated growth of the hip region and knee region, combined with the constraints of the peripheral muscle leads to normal rotation of the limb plus slight bending of the bone. He feels that formation of periosteum around the cartilage model is also mediated by mechanical events. The perichondrial fibrosis becomes modified into a periosteal membrane. The osteoblasts form a matrix that is mechanically situated to serve effectively as a cellular reaction to the great strain to which the bent femoral beam is subjected. The appositional growth of bone strengthens the femoral beam at its weakest part. The mechanical result of this cellular reaction of bone differentiation is the progressive formation of a more stable base for the application of muscular forces. The proximal femoral-acetabular hip joint is formed due to different rates of growth of the femur and the acetabular cartilages. The head of the femur is seen, relatively speaking, to advance farther and farther into the acetabulum formed by the ilium, ischium, and pubis. The force of longitudinal interstitial growth of the femur per square millimeter of cross section is calculated to be 12 times greater than that of the acetabulum. This fact, associated with muscular restrictions to longitudinal femoral growth and a femur rendered more stable by primary ossification center bone, leads to the structural elaboration of the hip joint. Joints are not hereditary structures alone but rather the mechanical resultants of opposing centers of accelerated growth.

102

C H A P T E R 1 9 D e v e l o p m e n t a l Bone Biolotty

A

B

9.:::-..:: ..

'.!.:.; !.i~.:i;))::.:::! 9:.-.-.- :. :: .:';' ..:.:.. . . . . . . .

/ . ,, . -,. .- .~,-.. 9 ~'

i::. .::. . .i::::. . . . .

. . . . . .

-..,

~',

-

~

,3,~,

--~

" - -~-__ :v .. ~ i k ~ ~ . ~

[ii r; i'll::,:,~;:: : ~:

..'...-: :... - . . . :. -...... -: .... .,. . . ~ " ~... ......

,,:

- - : . . ~ ~._~-~.-"-~..

"'"

~-:,."

<.'~"'~

:,'

(6"

t i ": .... . . . :" "i:, ~

"... : 9 :

"..'"

.l

1,4~"-"~_:.~ ~ : : :": ~ D . ('-

L__--.=--~-~:~,'/~::

~:::;',':" : . .~-=~_ -~-~~~:~-~ :::: 'v:: 9 ~---~-~.: --: ~----:~5~ ~ ~" .:.. ~, . . r.--~'.~.-'~'~-;~. ---"

~ii.!.!i::-,,'..r .~~.--.~ ~

~,-.

~l./,

~'x

!-:;" .': 1: ;"

,... , ,

..

~-

i

i 18

17 D

"

-

.

9 . . . .

.

.

.-.o

e(',~o.- . .."...... :.. ...-9 ,

.

/

9

~.

".."

"-

.

..:.-

Z n d

~..0.s

I

tI.D('.

,p

r.D.e

C~L

,4.N.

9 -

9

- ?.-~

"

.

.

9i

~

-

~

F I G U R E 22 Illustrations from the work of Carey demonstrate his efforts to correlate the embryologic development of bone and muscle in relation to the feedback effects of the two tissue types. The work also demonstrates at a theoretical level how the mechanical actions of one tissue affect the developmental pattern of the adjacent tissue. The sections are from the developing hindlimb of the embryonic pig. The developing pelvis in each part is shown at fight with "il" representing the ilium and "is" representing the ischium. The head of the femur (H) is shown at fight and the distal femur (F) and proximal tibia (T) are shown at left. Changes in tissue types are shown progressively from the 18-mm stage (number 17) to the 25-mm (18), 32-mm (19), and 50-mm (20) stages. The stress lines as conceived by Carey are outlined on the developing femur. The initials IC represent the growth plate, which he refers to as intermediate growing cartilage. At progressive time periods the diaphyseal cartilage, surrounding perichondrium, and early periosteum are converted to areas of bone formation. The letters TP outline the tensile periosteum on the anterior surface of the femur, and the letters CP outline the compressible periosteum on the posterior surface. As bone is formed on the dorsal or anterior surface, the orientation listed as TOL refers to the tensile osseous lamella, whereas on the posterior surface the letters COL refer to the compressor osseous lamella. As illustrated by Carey, the bone is formed initially on the anterior surface and subsequently on the posterior surface. Although Carey illustrates stress lines in the developing embryonic model of bone, his work remained theoretical with little to no published experimental work either by himself or others to verify his interpretations. [Derived from (53).]

SECTION Xlll 9 Responses to Mechanical Stresses Overall, limb development is a feature of intrinsic selfdifferentiation of theecell mass plus an extrinsic mechanical interaction due to differential growth. The idea is presented e that shaping of bones, even in the embryonic time, is "an immediate mechanical resultant of the force of its own interstitial growth and the intermediate resistance encountered to this growth." Both growth and resistance are inseparable and both play major roles in morphogenesis. The formation of a joint is a resultant along the lines of juncture between the zones of accelerated growth of neighboring segments in opposite dimensions. That segment presenting the greater force of growth possesses the convex element, whereas the segment with a lesser force is concave in relation to the moveable joint. Joint form is not solely an inherited pattern of skeletal segmentation, but rather occurs due to the mechanical resultants of compressive and shearing stresses of prior centers of accelerated growth opposing each other in the blastemal skeleton. The histogenesis and morphogenesis of the musculoskeletal regions therefore are dependent as much upon the mechanical factors extrinsic to the cells of the region as upon the intrinsic faculty of the tissue itself to grow. Each of the three structural forms of the skeleton, scleroblastemal, cartilaginous, and bony, represents advancing degrees of consolidation reacting to a gradient increase of pressure. Carey provided an excellent description of the development of the thigh of the pig encompassing both the bone and muscular tissues. This is excellent as an embryologic description itself, although he utilizes his theory of growth motive force to comment on the reasons for the differentiative patterns. Whereas the mechanical theory is attractive, the ability of organ culture to demonstrate entire bone growth in vitro and in the absence of limb musculature did much to eclipse interest in the work. In our opinion, however, much of value is provided that warrants renewed exploration with the far more sophisticated microtechnology now available.

B. Normal Relationship of Epiphyseai Plates to Compressive and Tensile Stresses It has long been accepted that the external shape of a bone and the disposition of the trabeculae within it are related to the stresses experienced during normal activity. Thomson (345) noted a correlation between the structure of the distal femur and the shape of the distal femoral physis. Smith (329) (1962) was one of the first to study biomechanical relationships to underlying structure of the epiphyseal plate feeling that physeal structure and orientation might also bear a similar relationship to stress (Figs. 23A-23E). He defined the three stresses that can occur within any object: a compressive stress, which causes a diminution of the dimension along which it acts; a tensile stress, which increases the dimension along which it acts; and a shear stress, which causes relative displacement along the intervening parallel planes. At any chosen point the compressive and tensile stresses were max-

103

imal in two orthogonal directions, which are designated as the principal stresses at that point. At that same point shear stress is 0 in the direction of the principal stresses and maximal at a 45 ~ angle to them. In an immature bone, stability of any epiphysis in relation to the diaphysis depended on the relationship of the plate to the internal stress patterns to which it was subjected. If the growth plate conformed through its extent to the direction of one of the principal stresses, no shear stress would develop between the epiphysis and the diaphysis and the relationship would be stable. If that relationship did not exist then forces exerted on a bone during activity would cause shear stress, which would tend to lead to displacement of one part on the other. Shearing stress would be less the more the matrices are arranged along the pressure and tension lines of the system. Smith performed studies in the proximal tibia and proximal and distal femur using the photoelastic method, which allowed him to construct the directions of the principal tensile and compressive stresses throughout the whole bone. 1. PROXIMAL TIBIA Stress analyses were performed at the proximal end of the immature tibia with assessments illustrated in both sagittal and coronal planes. The epiphyseal plate in the sagittal dimension demonstrates anterior and inferior displacement in relation to the developing tibial tubercle cartilage extension of the epiphysis. The axial compressive forces are at right angles to the growth plate and the tensile stresses are parallel to it. Smith noted that due to the retroversion present in the upper end of the tibia, during weight bearing the bone is subjected to a combination of axial compression and forward bending. Diagrams from the Smith article demonstrate how compressive stresses are at fight angles to the growth plate and how even the shape of the growth plate itself relates quite well to the orthogonal compressive and tensile stress lines. 2. DISTAL FEMUR On a coronal section, which demonstrates both the principal compressive and tensile stresses, the epiphyseal plate itself shows a clear relationship to the principal stresses similar to that observed in the proximal tibia. On both medial and lateral sides the growth plate lies parallel to the lines of tensile stress; centrally it conforms to the line of compressive stress, and only in the intervening regions is it subjected to shear stress. Similar patterns are seen on the sagittal sections. 3. PROXIMAL FEMUR

Knowledge of the shape of the developing cartilage within the femur as well as that of the adjacent metaphysis is very important to understanding stresses here. In the first few years of life there is a single curvilinear growth plate in relation to the adjacent developing trochanteric and head and neck cartilage masses. That part of the proximal surface of the metaphysis that is associated with the head and neck maintains a more or less constant form throughout development.

104

CHAPTER 1 9

Developmental Bone Biology

!i!! o.

F I G U R E 23 Compressive and tensile stresses in relation to the physes of developing bones as shown by Smith (329). (A) The disposition of the tensile and compressive stresses and the maximal shear stresses. (B) Stress analysis of proximal tibia in the standing position, sagittal plane. In the illustration, the patellar ligament is in black, the stippled line is the growth plate, arrows denote applied forces, the main compressive stresses are the solid lines, and the main tensile stresses are the dotted lines. The same format is used in the following parts. (C) Proximal tibial stress analysis, coronal plane. (D) Distal femoral stress analysis, standing state, sagittal plane. (E) Distal femoral stress analysis, standing coronal plane. [Parts A - E from Smith, J. W. (1962). J. Anat. 96: 58-78. Reprinted with the permission of Cambridge University Press.]

SECTION Xlll 9 Responses to Mechanical Stresses From the junction between the trochanteric and head-neck parts of the cartilage the proximal surface of the diaphysis runs medially and slightly downward in a gentle curve convex upward. The plate thus really is not straight as is indicated by plain radiographs. The common stress pattern indicative of both compressive and tensile stresses is markedly similar to the analysis of the adult bone from over 100 years ago. The principal compressive stresses are seen medially and the tensile stresses laterally. In its most medial area the growth plate, rather than persisting in a transverse plane, turns sharply downward to reach the articular margin. Here the physis is parallel to the compressive stress and again shear is minimal. Shear stress is present at the line of angulation of the physis, which tends to lead to displacement of the femoral head downward and medially on the neck in predisposed individuals. Smith's analyses showed the greater part of the epiphyseal plates to be oriented parallel to one or the other of the principal stresses that affect the region during normal activity. Within the proliferating columns of growth plate chondrocytes, however, there are no direct relationships to the stress pattern, but these rather are oriented in the direction of bone growth. Quite often the stress patterns have an orthogonal relationship to the direction of bone growth so that the cell columns lie at right angles to the plane of the epiphyseal plate. On occasion, however, the cell columns occupy an oblique position. If an epiphyseal plate therefore is oblique to the direction of bone growth, there is a zone of cartilage cells in which these structural elements are not aligned to the stress pattern of the part and thus are exposed to shear stress tending to displacement. In general, these studies show that the orientation of epiphyseal plates is in harmony with stress patterns of normal activity. This is because each plate tends to lie parallel to one or the other principal stress so that shear stress between the epiphysis and diaphysis is minimal. At both proximal and distal ends of the femur in particular the epiphyseal plate tends to lie at fight angles to the compressive stresses within the bone. Murray (239) points out that stresses within the epiphysis itself are multiplanar and that trabeculae of the secondary center develop a specific orientation only after growth plate closure and continuity has been established between the bone of the epiphysis and that of the adjacent metaphysis and diaphysis. The mechanical structure of the cartilage has been much less studied thaia that of bone, and most of those studies are performed relative to articular cartilage rather than to physeal cartilage. Cartilage is less resistant than bone but is more elastic and is much better able to withstand pressure than tension. Much of the ability to withstand tension is due to its surrounding periosteal-perichondrial sheath. It was recognized decades ago that the surrounding perichondrium is essential for the structural integrity of the physeal cartilage. The perichondrial sheath runs parallel to the long axis of the bone and as such its fibrils are tension-resistant and not pressure-resistant structures. The fibrillar peripheral

105

tissues thus responsible for resisting tension and the nonfibrillar components are pressure-resistant.

C. Mechanical Stresses and Their Differences and Effects on Skeletal Development The work of Carter and associates (57-60, 376) has defined a model that suggests that mechanical loading histories greatly influence bone morphogenesis, beginning from the embryonic stages of endochondral development and continuing throughout life. Carter and associates have developed theoretical models of new bone formation in general and of the endochondral ossification mechanism in particular on the basis of assessments of mechanical stresses and the use of finite element analysis. Integral to their theoretical basis is the belief that "the progression of endochondral bone formation in the diaphysis and metaphysis, the formation of the secondary ossification center, and remodeling of the new bone may all be influenced by the mechanical forces created from an increasingly active muscular system." Their feeling is that the entire internal structure of long bones is governed by loading or tissue stress histories to which the bones are exposed during growth and development. Their studies seek to emulate the influence of cyclic mechanical stresses on bone development to show how these would correlate with the endochondral ossification sequences. By using two-dimensional computer models, Carter et al. conclude that mechanical stresses caused by intermittent loads influence the formation of the endochondral ossification patterns. Both the ossification of the diaphysis and the formation of the secondary ossification center of the epiphysis are predicted by locating regions of high intermittent shear stresses. Their model interprets the stresses on the developing bone as being primarily hydrostatic or shear. Hydrostatic or dilatational stress causes a change in the volume of compressible materials on the basis of either compression, which decreases the volume, or tension, which increases it, whereas distortional stress causes no change in volume but causes changes in shapes (Figs. 24A and 24B). Carter et al. propose that endochondral ossification within the cartilage model is accelerated by cyclic shear stresses and slowed by cyclic compressive dilatational stresses. They recognize that the complete state of stress in any material is best described by specifying six stress components of three principal stresses and their orientation with respect to a reference frame, but it is often "more informative to summarize the state of stress in terms of a dilatational (hydrostatic) component and a shear component." They use the standard convention with compressive dilatational stress as negative and tensile dilatational stress as positive. They express their theory of the influence of mechanical stress on endochondral ossification by assuming that the application of intermittent shear stress or strain energy accelerates ossification whereas intermittent compressive hydrostatic stresses retard or arrest ossification. They utilize their data to calculate the

106

CHAPTER

1 9

Developmental Bone Biology

HYDROSTATIC STRESS COMPRESSION

TENSION

t

t SHEAR STRESS equals

9

N

t #

D

/9

FIGURE 24 Diagrammaticillustration of stresses on a developing bone that tend to influence certain types of synthetic patterns. Both hydrostatic (dilational) stresses and distortional (shear) stresses play a role. (A) Dilational stress causes change in the volume of compressible materials, and shear stress results in distortion of the material. The thick lines represent unloaded, underformed shape and thin lines represent loaded, deformed shape. [Reprinted from Carter and Orr (60), with permission.] (B) Hydrostatic pressure in mesenchymal tissue arises due to external forces (arrows) or from internal growth forces (arrows) caused by increases in tissue volume encountering resistance peripherally. [Reprinted from Pauwels, E (1980). "Biomechanics of the LocomotorApparatus," Fig. 11, p. 388, copyrightnotice of Springer-Verlag,with permission.]

distribution of a stress-dependent "osteogenic index" using a two-dimensional computer model of a cartilage diarthrodial joint with five loading conditions applied along the joint contact surface. Carter et al. conclude that regions of high osteogenic index correspond to the site and shape of the secondary ossification centers, whereas regions of low osteogenic index correspond to the site where articular cartilage and growth plate would form and persist. The osteogenic index is a mathematical parameter for these theories that summarizes the local mechanical stimulation to endochondral ossification over a time period as a weighted mathematical sum of the acceleratory effects of cyclic shear stress and the inhibitory effect of cyclic hydrostatic pressure. Finite element stress analyses of the femoral cartilage model indicate their theory to be consistent with features of bone morphogenesis, including the development of the primary ossification site, the tubular diaphysis with marrow cavity, metaphyseal and epiphyseal trabecular bone, location and geometry of the growth plate, the appearance and location of the secondary ossification center, and the presence and thickness distribution of articular cartilage. They are able to model each of seven stages of ossification passing from the cartilage model of the developing bone to the fully developed bone at the termination of skeletal growth with obliteration of the physis (376). Their staging involves the following: stage 1, all cartilage; stage 2, primary ossification front; stage 3, front migration; stage 4, further front migration; stage 5, secondary ossification center formation; stage 6, expansion of secondary ossification center; and stage 7, growth plate closure. Their theory indicates that cartilage degeneration and subsequent ossification are promoted by intermittently applied shear stresses, which are calculated to be highest at the center of the epiphysis where the secondary ossification center is formed. High shear stresses also are noted at the sides of the bottom edge of their model where the ossification groove of Ranvier is formed. Intermittently applied hydrostatic compression inhibits or prevents cartilage ossification and this was found at the periphery of the cartilage model over the joint surface. Their work indicates that mechanical stresses play a critical role in early skeletal growth and ossification in particular in relation to embryonic and early postnatal tissue differentiation. They again stress that "the reality of our model of long bone skeletal morphogenesis assumes that the muscular system has begun ~o exert mechanical stresses on the skeletal anlagen prior to or at the time at which endochondral ossification commences." It is interesting to note the correlation between the views of Carter and associates and those expressed by Carey over 50 years earlier. Much of Carey's work was disregarded when tissue culture and organ culture experiments began to reveal the development of cartilage models of the bone in vitro without muscle influences, although it is now apparent that the effects of muscle on bone development in vivo are real. In summary, Carter and associates show that the ossification of the diaphysis and the formation of the secondary

SECTION XIII 9 Responses to Mechanical Stresses ossification center can be predicted by locating regions of high intermittent shear stresses. They define stress-dependent osteogenic indices using a two-dimensional computer model of a cartilaginous joint. Regions of high osteogenic index correspond to the site and shape of the in vivo secondary ossification centers, and regions of low osteogenic index correspond to the sites of articular cartilage and growth plate formation. They further postulate that the entire internal structure of long bones, including the medullary cavity and trabecular morphology, are governed primarily by the loading histories to which the bones are exposed during their growth and development. Endochondral ossification within the cartilaginous anlagen is accelerated by shear stresses and slowed by cyclic compressive dilatational stresses. The finite element studies of Carter et al. suggest that (1) the ossific nucleus appears in an area of high shear (deviatoric) stresses, (2) the edge of the advancing ossification front of the periosteum (zone of Ranvier) also experiences high shear stresses, and (3) the joint surface, where articular cartilage forms, is exposed to high-magnitude hydrostatic compression. These findings support the theory that intermittently applied shear stresses promote endochondral ossification, whereas intermittently applied hydrostatic compression inhibits or prevents cartilage degeneration or ossification. Frost and Jee (116, 117) presented their theoretical model correlating biomechanical effects on endochondral ossification.

D. Responses of Physes to Abnormal Pressures~ Pathogenesis of Creation and Correction of Deformity in Developing Bones Many clinical and experimental observations have been made concerning the response of various epiphyses and physes to pressure. The observations essentially are phenomenological in nature and very few studies have been done pertaining to the mechanisms by which the responses occur or to quantitation of the specific intrinsic and extrinsic epiphyseal pressures in normal and abnormal states. The presumed relationship between epiphyses and pressure phenomena has been ingrained in assessments for a very long time. Delpech (77) (1828), in his early orthopedic classic De L'Orthomorphie, commented on developmental malformations of thebones during the growth period caused by abnormal pressure on the growth cartilages. During the growth period, unequal pressure on the growth cartilages led to a slowing of growth at areas of increased pressure and an exaggeration of growth where pressure was diminished. Stated otherwise, bone growth increases in regions of diminished pressure and decreases in regions of high pressure. In relation to growth cartilages, growth and development occur in an inverse relationship to the pressures exerted on the cartilage. The separate works of Hueter (166) (1862) and Volkmann (358) (1862) related the pathogenesis of deformity to abnormal pressure on the growth cartilages. What has become known in the English literature as the Hueter-Volkmann law

107

postulates that pressure beyond normal on growth cartilage leads to growth retardation. Stated another way, they noted an inverse relationship between compressive forces along the long axis of epiphyseal growth and the rate of epiphyseal growth. Increased pressure on the concave side interfered with and slowed normal bone growth, whereas on the convex side less than normal pressure led to overgrowth. Neither worker ever stated the premise as a specific law or performed any experimental investigation. The observation is evident in a clinical sense, for example, in relation to infantile and adolescent tibia vara, which are felt to be caused by increased weight in relation to physes that are placed in a disadvantageous varus situation. Although this concept is widely accepted, very little work has been performed to document the extent of the pressures that are normal on the one hand and sufficiently great to cause damage on the other. The absence of normal pressure is also known to alter epiphyseal development. A clinical example of this relates to dysplastic hips, particularly those that are completely dislocated. In these the femoral head is smaller and misshapen in relation to the normal side and there is a delayed appearance of the secondary ossification center. It is unclear whether the delayed appearance of the secondary center is due to the lack of appropriate pressure on the head, which would normally come from the acetabulum, or due to a relative hypovascularity as a result of stretching of the femoral vessels with the head in an inappropriate position. The work of Wolff (374) has played a major role in revealing how extrinsic forces on bone modify both the internal (by which is meant trabecular) architecture and the external shape of the bone (Fig. 25A). His law has been translated as follows: "Every change in the form and function of the bones, or of their function alone, is followed by certain definite changes in their internal architecture, and equally definite secondary alterations of their external conformation in accordance with mathematical laws." Wolff refers, however, to lamellar bone deposition and not to bone development; indeed, he felt that early bone formation from cartilage models was independent of mechanical stress. Remodeling of a long bone following an acquired angular deformity is a function of three features. An excellent experimental model showing this was described by Ryoppy and Karaharju (304) following hind leg fractures in 3-week-old rats. They demonstrated correction of angular deformity, documenting not only metaphyseal and diaphyseal corrective mechanisms but also asymmetric epiphyseal growth as a corrective mechanism. With mild to moderate increased pressure on the concave side of a deformity, there was actually a positive effect on the concave side physis leading to stimulation of growth compared to the opposite side of the physis and correction of deformity. These principles are outlined in Fig. 25B. (1) There is increased new bone formation by the inner layer of the periosteum on the concave side by the intramembranous mechanism of bone formation. This serves to decrease the concavity but has no effect on the angulation or obliquity of the epiphyseal growth plates or

108

C H A P T E R 1 9 Developmental Bone Biology

F I G U R E 25 (A) Illustrations of the trabecular orientations and cortical bone thickness based on Wolff's law are shown. With angular deformation the concavity in both (Ai) and (Aii) is at right and compression predominates. Note the thicker cortex in (ii) on the concave compression side. [Reprinted from Pauwels, E (1980). "Biomechanics of the Locomotor Apparatus," Figs. 34 and 3, pp. 242 and 230, copyright notice of Springer Verlag, with permission.] (B) There are three mechanisms by which a growing long bone can correct an angular deformity. Remodeling can occur by three different mechanisms at the epiphyseal, metaphyseal, and diaphyseal regions. The first involves new bone synthesis on the concave side of the deformity in the metaphyseal and diaphyseal regions. This is mediated exclusively by the intramembranous mechanism, acts primarily at the diaphyseal level, and does not correct epiphyseal deformity. Bone is synthesized preferentially along the concave aspect with minimal amounts only on the convexity. As the overall bone increases in length and width, however, bone synthesis is along appropriate lines and the smaller deformed model becomes encased in a straighter, larger model. Second, the metaphyseal regions are remodeled by sequential resorption particularly on the concave side by tissues at the periphery of the metaphysis, which normally perform that role. The greater the deformity in any concavity on the metaphyseal side, the greater the activation of the osteoclast metaphyseal remodeling mechanism. Third, the epiphyses themselves correct for angular growth by asymmetric physeal function. There is a 2+ growth function of the physis on the concave side of the deformity and a slower 1+ on the convex side. It still is not clear whether this increase in activity is manifested by increased thickness of the physis or simply by increased activity on the concave side, which does not lead to any change in the thickness of the physis due to the symmetry of synthetic and resorptive phenomena.

articular surfaces at either end. (2) Asymmetric growth of the epiphyseal plate: the physis on the concave side has an increased rate of growth compared to the physis on the convex side. This serves to correct the obliquity both of the epiphyseal growth plate and of the joint surface. (3) Metaphyseal remodeling occurs with more extensive resorption on the concave side than on the convex side. Three mechanisms thus are operative in correcting angular deformity in a growing long bone involving altered patterns of cell function in each of the diaphyseal, metaphyseal, and epiphyseal regions. The question of the ability of a growing bone to correct rotational or torsional deformities also has long been of clinical interest, but both clinical and experimental evidence of its occurrence has been less than definitive. There is little question that rotational changes in long bones occur with growth, and this p h e n o m e n o n is well-documented with the decrease in femoral anteversion at the proximal femur. It appears that a major reason many clinicians doubted the effective correction of malrotation came from assessments of

long bone fractures with relatively severe rotational deformity that did not correct with time. Thus, the potential ability of long bones to correct rotational malalignment that appears to exist would be limited, however, by the extent of the rotational deformity and the age in relation to the developmental sequence at which it occurred. A study by Husby et al. (176) indicated, however, that even in the postnatal period acquired rotational abnormalities could be corrected with growth. They produced inward or outward rotation of approximately 25 ~ in the right femur of kittens by midshaft osteotomy and plate fixation. The left femur served as the control, and the angle of anteversion (AV) of the femoral necks was measured in axial radiographs from preoperative time to 58 weeks postoperatively. They noted that the AV angle of the unoperated left femurs did not change with growth in length over a 58-week period, beginning at 13 weeks of age when the operation was performed. The immediate postoperative mean torsional difference in the externally rotated group was - 2 0 ~ and in the inwardly rotated group + 2 7 ~. Most of the torsional deformities regressed during the first

SECTION XIII ~ Responses to Mechanical Stresses 12 weeks postoperatively, and by 1 year postoperatively the mean torsional difference between matched right and left femurs was only - 1o in the externally rotated group and +4 ~ in the internally rotated group. They demonstrated that in all animals the rotational deformities of the operated femurs regressed concurrently with longitudinal femoral growth. They concluded that the regression took place in the femoral growth plates primarily due to helical growth in length in the distal physis to account for the correction, although proximal femoral growth plate growth in a dorsal direction might also contribute to the physiologic regression of the AV angle of the femoral neck. The growth effects of one bone on the other in two bone segments are also shown particularly in pathological conditions such as hereditary multiple exostoses. In this disorder there is a marked tendency for the ulna to be more deformed and shortened than the radius in the upper limb and for the fibula to be more deformed and shortened than the tibia in the lower limb. The shortness of the ulna and the fibula combined with their continuing close approximation and ligamentous binding to the radius and tibia, respectively, leads to the deformities of the adjacent bones caused by asymmetric physeal tethering. The physes and metaphyses of the relatively uninvolved longer side develop angular deformities with their curvature concave toward the shorter bone.

E. The Effects of Pressure on Epiphyseal Growth It is evident in a clinical sense that deformity in the area of the epiphyseal and metaphyseal region of a growing long bone can be either improved with time, which implies a corrective and regulatory response by the open physis, or worsened with time, which implies negative or inhibitory effects on physeal growth. There is little documented quantitative data indicating how much altered physeal pressure has a positive or stimulatory effect and how much has a negative or inhibitory effect. It is evident, however, from both clinical observations and investigational studies, that mild to moderate increases in pressure actually serve to stimulate the physeal growth asymmetrically and thus lead to correction, whereas markedly increased pressure inhibits the growth of the physis and worsens deformity. The latter phenomenon was recognized throughout the nineteenth century and led to the principles of growing bone deformation articulated by Delpech, King, and Hueter and Volkmann. Increasing study of bone deformity patterns over the past few decades, however, has led to the recognition of asymmetric stimulation of physeal growth, leading to a corrective mode if the abnormal increased pressures are mild to moderate. Pauwels pointed out this corrective mechanism (270) (Fig. 26). One of his examples illustrates straightening of the femoral neck to a normal range after an adduction or varus osteotomy early in childhood. In the normal anatomic alignment, uniform distribution of compressive stresses in epi-

990"~f - - 7

. J 80*

.///~

109

R

.-" "'-.# /

,

f

FIGURE 26 Proximalfemoral angular correction is illustrated after varus osteotomyof childhoodor after subtrochantericfracture, whichhealed with extensive varus. The medial physis comes under increased pressure but respondsto this by increasingits rate of synthesisin relation to the more lateral aspect of the physis, thus correcting deformity. [Reprinted from Pauwels, E (1980). "Biomechanics of the LocomotorApparatus," Fig. 2, p. 509, copyrightnotice of Springer-Verlag,with permission.]

physeal cartilage leads to uniform increases in growth across the proximal femoral physis. Varus osteotomy diminishes the neck-shaft angle and tilts the epiphyseal cartilage such that the compressive stresses are now graded to the medial border of the femoral neck. Clinical examples show that, after a varus osteotomy in a young child, the femoral neck straightens. Pauwels indicates that "this straightening results from stronger growth in length of the medial part of the epiphyseal cartilage." He thus articulates the effect of stimulation, indicating that "unequal distribution of pressure in the epiphyseal cartilage, provided it is not above a certain threshold, leads to an unequal growth in length. This causes an alteration in the shape of the bone which in turn leads to a reduction in the bending stressing of the latter. Reduction in the bending stressing means an economy of material. This is the essence of functional adaptation." Pauwels further comments on pressure phenomena and their relation to physeal function. He presents in detail a case of coxa vara secondary to untreated subtrochanteric fracture of infancy in which the varus deformity gradually straightens following several years of growth. His analysis indicated that stresses on the epiphyseal cartilage of the femoral head and neck were such that "physiologically the resultant compressive force is perpendicular to its c e n t e r . . , it is stressed throughout purely in compression." In the hip with coxa vara, the resultant compressive force under load was no longer perpendicular to the center of the epiphyseal cartilage but formed an angle of 70 ~ with it open laterally. Bending thus was added to the normal pure compressive stressing of the epiphyseal cartilage, and in the lateral half of the physis the pure compressive stresses were decreased with the addition of tensile stresses. In the medial half, the compressive stresses were more than doubled due to the additional bending.

110

C H A P T E R 1 ~ Developmental Bone Biology

This pathological distribution of the stresses is illustrated by Pauwels in the accompanying diagram. The metaphyseal bone density is distributed uniformly in the normal hip but markedly denser to the medial side in the coxa vara and abnormally stressed side. This deformity corrected with time because of "fundamental and progressive alterations which have developed because of stronger longitudinal growth on the compression side as a consequence of the additional bending stressing of the epiphyseal cartilage." Pauwels reached three conclusions: (1) In bending stressing combined with compression, the stresses on the compression side are always greater than on the tension side. On the compression side, it is the sum of the two types of stresses that act together, whereas on the tension side it is their difference. (2) Bone formation and~resorption are controlled by the magnitude of the stresses. (3) The epiphyseal cartilage normally is stressed in pure compression. The addition of bending stressing causes more growth on the compression side where the stress is greatest. This serves to eliminate the additional bending stress. Gelbke (124) addressed the issue of the influence of pressure and tension on physeal growth. He pointed out that the epiphyseal cartilage plates can bear, without any damage, a much higher degree of pressure than is effected under normal physiological conditions. He demonstrated that pressure on the physeal regions along the long axis of the bone led to retardation of physeal growth and ultimately, if retained, to full fusion. Negative effects were more marked with continuous pressure. His work concluded that "only an extreme increase in pressure leads to an inhibition of longitudinal growth and to a narrowing of the epiphyseal plate." This has also been borne out by the work of Blount. Blount and Gelbke show instances where a metal staple and circumferential metal wire, respectively, can both be broken by the forces of growth without growth inhibition occurring. Where pressure is great and growth inhibition occurs, the epiphyseal plate is narrowed. Maas (220) (1924) demonstrated that where the pressure was uniform and parallel to the long axis of growth, a shortening without angular deformity was seen. Where the pressure was asymmetric, angular deformity occurred concave to the side of increased pressure. With relatively milder pressures, the physis would actually widen in response. Not only must pressure be extremely high to have negative effects, but it also must be continuous. Similar experiments were performed by Haas ( 133), who demonstrated that a circumferential wire loop encompassing epiphyseal and diaphyseal bone also led to growth retardation while staples also led to growth diminution and angular deformity. Arkin and Katz (9) studied the effects of pressure on epiphyseal growth and noted that pressures applied to epiphyses could be classified into two main types, one parallel to the direction of epiphyseal growth and the other at a fight angle to this direction. The forces applied in any direction intermediate between these two subsequently could be resolved into components or vectors parallel to one or the other. The extreme force generated by physeal growth has received some documentation. Blount and Zeier found that the

forces of growth in a child could break staples that themselves were able to withstand 900 lb of distraction in mechanical testing (115). Franz (115) reviewed early results on physeal inhibition in a clinical setting using the epiphyseal stapling technique of Blount. Strobino et al. (336) assessed growth forces in a calf in which a transphyseal spring mechanism was inserted. They found no inhibition of growth in the upper tibial epiphysis, even when the restraining forces were as high as 400 lb. Arkin and Katz (9) also assessed the effects of pressure perpendicular to the axis of epiphyseal growth. This pressure served to change the direction of growth, and the amount of change of direction varied with the degree of pressure. They felt that pressures applied in any direction between the perpendicular and the parallel could be broken down into vector components and the proportion effect of each component on rate and direction of growth could be determined. When forces were applied parallel to the direction of growth and were the same throughout the epiphyseal plate, growth was inhibited uniformly. It was inhibited asymmetrically if the pressure was greater on one side than on the other. Pressure thus serves to modify both the rate and the direction of epiphyseal growth. Torsional stress applied to developing femurs leads to deviation, which demonstrated no deformation of the shafts of the bone but rather deformation at the epiphyseal and adjacent metaphyseal regions. That portion of bone laid down by the epiphyseal plate while it was under torsional stress was the region that was deviated. The torsional effect was greater the smaller the thickness of the bone. Any force exerted on the epiphyseal plate in a direction not parallel to the direction of growth will deflect growth away from the deforming force. The amount of deflection produced is a function of the direction of the force being directly proportional to the angle formed by the direction of growth and the direction of the deforming force up to 90 ~ Arkin and Katz also concluded that pressure does not inhibit growth by an all or none law. Whereas enormous forces are required to fully inhibit growth, much milder forces, particularly if they are applied asymmetrically, can inhibit or otherwise modify epiphyseal growth. They reached four conclusions that remain notable today. (1) When a growing epiphysis is subjected to a stress, the rate or direction of the growth of that epiphysis or both are modified so as to yield to that stress. (2) Pressures applied in directions parallel to the direction of epiphyseal growth inhibit the rate of such growth. Whereas considerable pressures are necessary to stop cartilaginous growth completely, slight or even intermittent pressures can slow or hinder it. (3) Pressures applied in directions perpendicular to the direction of epiphyseal growth deflect the direction of such growth, resulting in lateral or torsional displacement of the newly laid down bone. (4) The ease with which angular or torsional deformities may be produced in a growing bone varies inversely with its diameter. Tschantz et al. (352) induced mechanical overload of the distal ulna in the dog by resecting a segment of the adjacent

SECTION XIV ~ Radiographic Characteristics in Development of Major Long Bone Epiphyses radius, leading to lysis of the ulnar physis and chronic tilting of the epiphysis. The model allowed for assessment of the effects of asymmetric loading on physes by histologic and radiologic criteria. Histologic study showed development of a large transverse intraphyseal fissure starting on the convex side and traveling various distances toward the convex side. In most cases the fissure was within the columnar, hypertrophic, or adjacent metaphyseal regions, but on occasion it passed within the germinal cell zone. Epiphyseal tilt of 5~ ~ developed between the fourth and tenth postoperative days. In those instances in which the germinal zone was involved early transphyseal bone bridge formation occurred. Repair along with correction of the tilt followed in most dogs when fracture was below the germinal zone. Overload stopped when the radius healed, thus allowing the ulnar physeal cartilage to heal and resume normal growth.

XIV. R A D I O G R A P H I C C H A R A C T E R I S T I C S IN D E V E L O P M E N T O F M A J O R L O N G BONE EPIPHYSES

A. General Information The time of appearance of the primary and secondary ossification centers has been an important point for study in long bone development (15, 16, 222, 245, 343) (Fig. 27A-27F). The variable times of appearance were well-established in textbooks of anatomy and osteology in the nineteenth and early twentieth centuries before the discovery and widespread use of X rays. With the widespread use of the radiographs, however, more accurate delineation of the time of appearance of the various secondary ossification centers was possible because larger numbers of individuals could be studied. Sequential studies of the same individual were possible and assessments were done in normal individuals (52, 105, 129). Despite the advantages of the radiographic technique, the hope that there would be more precise timing of secondary ossification center development with a narrower range of variability for each epiphysis has been realized only partially. The reasons for this are clear because the times of appearance of ossification centers vary considerably not only between the two sexes but even within the same sex. The time of fusion of epiphyses to metaphyses has been even more difficult to document reproducibly because of the previously mentioned variability plus the differing criteria used in different studies to determine when fusion has occurred. Two of the most detailed radiographic longitudinal studies of the time of appearance of ossification centers are those by the Cleveland group assembled by T. Wingate Todd at Western Reserve University and the Brush Foundation (112, 113, 153) and by Flecker (105). The latter also assessed the time of fusion. Unfortunately, the data of Flecker are presented in such a way that they are most difficult to comprehend. Three papers from the Cleveland group were published in 1939 and 1940 detailing the fetal age assessment of cen-

111

ters of ossification by A. H. Hill (153), the appearance of centers of ossification from birth to 5 years by Francis and Werle (112), and the appearance of centers of ossification from 6 to 15 years by Francis (113). These papers are particularly valuable because of their careful attention to detail relating to both primary and secondary ossification centers and the inclusion of all long bone epiphyses, as well as carpal and tarsal bone centers. They published their detailed database documenting the entire range of appearance of the secondary center at each site from the age at first documented appearance of any secondary center to the age at which 100% of the study population had shown that appearance. These figures not only show the ranges of time of appearance of the centers but also allow for detailed assessment of the progression of appearance of the centers with time. The studies were performed at 3, 6, 9, and 12 months of age, then every 6 months to age 5, and then annually to 15 years of age. Another study assessing large numbers of patients and presenting data clearly is that of Hansman (1962) (143), showing both time of appearance and fusion of ossification centers. The data of Hansman are reproduced in Tables XI and XII. The method of data presentation varies from paper to paper such that close study is needed to make comparative assessments. Hansman provided the range of ages for the appearance of each particular center plus the median age at time of appearance. Important radiographic features of development of each of the major long bone epiphyses follow.

B. Proximal Humerus Two secondary ossification centers form in the proximal humeral epiphysis (248). The medial center forms first and generally is present in all at 2 months of age. Hill notes that the medial secondary ossification center of the proximal humerus is present at birth in 26% of individuals, with the earliest fetal appearance between the 28th and 31st weeks. Studies indicate its presence within 2 days after birth in 45% of individuals, in 78-81% by 2 months of age, in 86% by 3 months and in 100% by 6 months. The lateral center appears later at a median age of 9 months in girls (range: 2 months to 289 years) and 1 year 2 months in boys (range: 4 months to 489years; 96% by 2 years). The two centers fuse into one between the ages of 5 and 7 years. The external shape of the proximal humeral epiphysis is retained with little change from birth until skeletal maturation. The radiographs indicate, however, that the internal shape of the epiphyseal growth plate as indicated by the epiphysealmetaphyseal junction changes from an initial transverse or curvilinear orientation to that of an inverted V. The lateral secondary center frequently is referred to as that of the greater tubercle, but it appears to be a true secondary center of the epiphysis and should not be considered an apophysis. The medial center relates to the inclined plane of the medial aspect of the physis and the lateral center relates to the inclined plane on the lateral side of the physis. Radiographic definition of the time of appearance of the two centers shows

112

CHAPTER 1 9 Developmental Bone Biology

FIGURE 27 Illustrationsof developing long bones showing secondary ossification center development. [Illustrationsfrom Quain's Anatomy as reproduced from an article on bone developmentby Bardeen in Keibel and Mall (16)]. The times of occurrence of bone formation are listed in the text and in Tables XI and XII. (A) Humerus. (B) Radius. (C) Ulna. (D) Femur. (E) Tibia. (F) Fibula.

a fairly wide range because of the superimposition of one on the other, depending on the rotation of the arm at the time of radiograph. This also applies to accurate definition of the time of fusion of the two secondary centers. From birth to 3 years of age, the plate is slightly curved, concave toward the metaphysis. Once the two ossification centers of the proximal humerus have fused, anteroposterior radiographs demonstrate an inverted V shape to the plate. The epiphysis

interdigitates with the metaphysis in a troughlike fashion along the anteroposterior axis.

C. Distal Humerus The newborn has a completely cartilaginous epiphysis (228). The physis is mildly curvilinear but approaches the transverse plane in the anteroposterior projection. It is somewhat

SECTION XIV ~ Radiographic Characteristics in Development of Major Long Bone Epiphyses TABLE Xl

113

M e d i a n A g e s in Y e a r s a n d M o n t h s a n d R a n g e s in A g e a t W h i c h C e r t a i n O s s i f i c a t i o n Centers Appear ~

Center Median

Girls' range

No. of cases

Boys' range

No. of cases

3-1

1-6 to 6-0

51

Median

Metacarpal

I

1-11

0-6 to 3-0

72

II

1-4

0-6 to 2-0

71

1-10

0-6 to 3-0

53

III

1-4.5

0-6 to 2-6

74

2-0

0-6 to 3-6

53

IV

1-6

0-6 to 2-6

74

2-1.5

1-0 to 4-0

54

V

1-7

0-6 to 3-0

73

2-5

1-0 to 4-0

54

Proximal Phalanx

I

2-1.5

1-0 to 3-0

69

3-4.5

1-6 to 6-0

49

II

1-2

0-4 to 2-0

70

1-7.5

0-6 to 3-0

55

III

1-1.5

0-4 to 2-0

72

1-7

0-6 to 3-6

55

IV

1-2

0-4 to 2-0

71

1-8

0-6 to 3-0

54

V

1-5.5

0-6 to 2-6

73

2-1.5

1-0 to 4-0

54

II

1-8

0-6 to 3-0

64

2-6

1-0 to 4-0

49

III

1-7.5

0-6 to 2-6

69

2-4

0-6 to 4-0

49

Middle Phalanx

IV

1-7.5

0-6 to 2-6

69

2-5

1-0 to 4-0

52

V

2-3

0-6 to 4-0

65

3-7

1-6 to 5-6

50

I

1-5

0-4 to 2-6

63

1-10.5

0-6 to 4-0

50

II

2-5

0-6 to 3-0

58

3-7

1-0 to 6-0

45

III

1-11

0-6 to 3-0

65

2-7

1-0 to 4-0

48

IV

2-0

0-6 to 4-0

66

2-8

1-0 to 4-0

48

V

2-4

0-6 to 3-6

63

3-7

1-6 to 5-0

48

Humerus head

81% by 0-2

0 to 0-4

58

78% by 0-2

0 to 0-6

45

Greater tuberosity

0-9

0-2 to 2-6

68

1-2

0-4 to 4-6

53

Capitellum

0-7

0 to 1-6

69

0-10

0 to 2-6

56

Lateral epicondyle

9-10

7-6 to 12-0

70

12-5

9-6 to 15-6

61

Distal Phalanx

Trochlea

9-5

5-6 to 12-6

70

10-7

7-0 to 14-0

58

Medial epicondyle

3-10

2-0 to 6-6

64

7-1

4-6 to 10-0

60

Radius head

4-10

2-0 to 8-0

70

6-3

2-6 to 9-6

57

Olecranon

8-8

6-0 to 11-6

69

11-3

7-6 to 14-0

54

Distal radius

1.15

0-4 to 3-0

69

1-4

0-4 to 3-6

53

Distal ulna

6-0

3-6 to 9-0

75

7-5

5-0 to 10-0

62

Femur head

0-5

0 to 1-0

67

0-6

0-2 to 1-0

55

68

4-0

2-0 to 6-0

60

Greater trochanter

2-10

1-6 to 4-0

Distal femur

100% by 0-2

0 to 0-2

58

100% by 0-2

0 to 0-2

47

Proximal tibia

100% by 0-2

0 to 0-2

58

98% by 0-2

0 to 0-2

47

Proximal fibula

3-1.5

1-0 to 6-6

60

4-5

2-0 to 6-6

57

Distal tibia

0-5

0 to 1-0

68

0-5.5

0-2 to 1-0

56

Distal fibula

1-0.5

0-4 to 3-0

69

1-4

0-4 to 2-6

50

aFrom Hansman (143).

114

CHAPTER 1 ~

TABLE Xll

Developmental Bone Biology

M e d i a n A g e s in Years a n d M o n t h s a n d R a n g e s in A g e a t W h i c h C e r t a i n E p i p h y s e a i Ossification Centers Fuse with Their Metaphyses a

Center

Median

Girls' range

No. of cases

Median

Boys' range

No. of cases

Metacarpal I

14-1

11-6 to 16-0

31

16-4

14-0 to 18-6

32

II

14-6

11-6 to 17-0

31

16-6

14-6 to 18-6

31

III

14-6

11-6 to 17-0

31

16-6

14-6 to 18-6

31

IV

14-5

11-6 to 17-0

31

16-5

14-6 to 19-6

31

V

14-5

11-6 to 16-9

31

16-6

14-6 to 19-6

31

Proximal phalanx I

14-2

11-0 to 17-0

33

16-3

14-6 to 18-6

33

II

14-2

11-0 to 16-6

31

16-4.5

14-6 to 18-6

32

III

14-2

11-0 to 16-6

32

16-4

14-0 to 18-6

32

IV

14-3

11-0 to 16-6

33

16-6

14-6 to 18-6

32

V

14-2

11-0 to 16-0

32

16=3

14-6 to 18-6

33

II

14-4.5

11-0 to 17-0

33

16-5

14-6 to 19-6

30

III

14-4.5

11-0 to 17-0

31

16-6

14-6 to 19-6

31

IV

14-4

11-0 to 17-0

31

16-5

14-6 to 18-6

31

V

14-3

11-0 to 17-0

33

16-4

14-6 to 19-6

30

Middle phalanx

Distal phalanx I

13-8

10-6 to 15-6

36

15-11

13-6 to 18-0

35

II

13-7

10-6 to 16-0

37

15-10

13-6 to 19-6

35

III

13-7

10-6 to 16-0

37

16-0

13-6 to 18-0

34

IV

13-8

10-6 to 16-0

35

15-10.5

13-6 to 18-0

34

V

13-7

10-6 to 15-6

37

15-11

13-6 to 18-0

33

Humerus head

15-7

13-0 to 17-0

27

18-2

16-6 to 20-0

19

Greater tuberosity

4-1

2-0 to 7-6

71

5-6

3-0 to 8-6

63

Capitellum

12-5

9-6 to 14-0

62

15-2

13-6 to 17-6

58

Lateral epicondyle

12-8

9-6 to 15-0

57

15-4

13-6 to 18-0

56

Trochlea

12-4

9-6 to 14-0

58

15-1.5

13-0 to 18-0

58

Medial epicondyle

14-1

11-0 to 16-0

51

16-4

14-0 to 19-0

45

Radius head

13-6

10-6 to 16-0

57

16-2

14-0 to 19-0

47

Olecranon

12-8

10-0 to 14-6

58

15-4.5

13-6 to 18-0

56

Distal radius

15-10.5

13-0 to 17-0

28

18-0

16-0 to 20-0

21

Distal ulna

15-11

12-6 to 17-0

30

17-10.5

16-0 to 20-0

20

Femur head

14-2

11-0 to 16-6

52

16-3

14-0 to 19-0

49

Greater trochanter

13-11

11-6 to 16-0

50

15-11

14-0 to 19-0

47

Distal femur

14-9

12-0 to 17-0

46

16-7.5

14-0 to 19-0

47

Proximal tibia

14-10

12-0 to 17-0

44

16-11

14-6 to 19-6

42

Proximal fibula

15-2

12-0 to 17-0

38

17-2

15-0 to 20-0

37

Distal tibia

14-10

12-0 to 17-0

44

16-10.5

14-0 to 20-0

46

Distal fibula

14-10.5

12-0 to 17-0

42

16-10.5

15-0 to 20-0

45

"From Hansman (143).

more curved in its lateral projection. In both projections, the physeal concavity is toward the metaphysis. The metaphysis relates to the epiphysis as a cylinder with its long axis in the coronal plane. T h e lateral capitellar ossification center ap-

pears at a median age of 7 months in girls (range birth to 1 89 years; 100% by 2 years) and 10 months in boys (range: birth to 2 89 years; 100% by 2 years). The trochlea ossification center appears much later at a median age of 9 years 5 months

SECTION XIV ~ Radiographic Characteristics in Development of Major Long Bone Epiphyses

in girls (range: 5 89189 years; 100% by 12 years) and 10 years 7 months in boys (range: 7-14 years; 100% by 15 years). The trochlea usually ossifies from multiple foci that fuse gradually with time. The medial epicondyle ossification occurs at a median age 3 years 10 months in girls (range: 2-6 89 years; 95% by 5 years) and 7 years 1 month in boys (range: 489 to 10 years; 95% by 9 years). The lateral epicondyle occurs at a median age 9 years 10 months in girls (range: 7 89 years; 100% by age 12) and 12 years 5 months in boys (range: 9 89189 years; 100% by 15 years).

D. Proximal Radius The proximal radial epiphysis is completely intra-articular (229). The physis approaches the transverse plane in all projections. There is a slight concavity to the plate, which relates to the gentle dome-shaped metaphysis. The ossification center of the radial head appears at a median age of 4 years 10 months in girls (range 2-8 years; 89% by 5 years) and 6 years 3 months in boys (range: 2 89189 years).

E. Distal Radius The plate closely approaches the transverse plane in all projections, with slight concavity relating to a dome-shaped distal radial metaphysis (249). The median age of appearance of the distal radial secondary center in girls is 1 year 2 months of age (range: 4 months to 3 years; 99% by 2 years of age) and 1 year 4 months of age in boys (range: 4 months to 389 years; 91% by 289 years). The secondary center appears centrally initially but tends to expand or ossify laterally toward the styloid, giving it a more lateral position during much of the growth period.

F. Proximal Ulna The proximal ulnar secondary ossification center resides within the olecranon (229). This appears relatively late at a median age of 8 years 8 months in girls (range: 6-11 89years; 97% by 10 years of age) and 11 years 3 months in boys (range: 7 89 years; 99% by age 13).

G. Distal Ulna The distal ulnar physis lies along the transverse plain in all radiographic projections (249). The secondary center appears at a median age of 6 years in girls (range: 3 89 years of age) and 7 years 5 months in boys (range: 5-10 years; 100% by 10 years of age). The ossification center is central in the epiphysis in all radiographic projections and expands equally in all directions.

H. Proximal Femur The shape of the proximal femur undergoes the greatest degree of change during development of any epiphyseal region

115

in both external contours and internal bone-cartilage relationships (235, 265, 334, 339). Two major epiphyseal ossification centers develop at the proximal end of the femur: one in the femoral head and the other in the greater trochanter, along with an apophyseal center of the lesser trochanter (1). The median time of appearance of the proximal femoral capital secondary ossification center is 5 months of age in females (range birth to 1 year; 100% by 1 year) and 6 months of age in males (range 2 months to 1 year). It is interesting to observe the specific time of appearance of the proximal femoral capital femoral epiphysis secondary center appearance. In the study by Francis and Werle (112), in girls the femoral head center is present in 40% at 3 months age, 89% at 6 months of age, and 100% by 12 months. In boys, 24% have appeared at 3 months, 91% by 6 months and 100% by one year of age. The median time of initial appearance of the greater trochanter secondary center is 2 years 10 months of age in girls and 4 years in boys (range 2 to 6 years; 94% by 41/2 years). The lesser trochanter begins to ossify as an apophysis at an approximate median age of 889years (range: 6-12 years) and 1189 years in boys (range: 7-15 years). The upper end of the femur is composed initially of a solid mass of cartilage. The proximal femoral growth plate has a curvilinear shape and underlies the cartilage of the head, the cartilage of the greater trochanter, and the cartilage of the intervening notch region. Toward the end of the first year this curvilinear shape is changed as there is a relative overgrowth of the head-neck region in relation to that of the greater trochanter. The appearance of the metaphyseal bone thus assumes an inverted V shape with the medial limb approximately one-third of the length compared to two-thirds the length of the lateral limb underlying the notch region and the greater trochanter. At approximately 3-4 years of age the notch region is transformed from cartilage to fibrous tissue, thus setting up two separate epiphyseal regions. From 6 years onward, there is a gentle dome shape to the femoral neck metaphysis, with the concavity of the epiphyseal plate toward the neck. A study of trabecular orientation is particularly helpful in assessing the growth patterns of the proximal femur. At birth the metaphyseal trabeculae are splayed with no specific preferential orientation underneath the curvilinear aspect of the single growth plate from medial to lateral sides. By 14 months, when there is deepening of the notch between the medial femoral head and lateral greater trochanter regions, the orientation of the trabeculae is seen. By 5 years the adult type pattern is demonstrated. Additional changes in shape during development relate to the external structure of the bone. These frequently are combined with use of the term anteversion but really comprise changes in two planes. One involves a rotation of the entire proximal end of the femur in relation to the diaphysis in a direction such that the head-neck midline axis in relation to the coronal plain of the distal femoral condylar region is increased. The second angle, which can be appreciated by considering a lateral projection of the proximal end of the

116

CHAPTER 1 ~ Developmental Bone Biology

femur with the diaphysis, places the head-neck-trochanter region in a flexed position relative to the diaphysis. It is possible to conceive of an absence of rotation of the proximal end with a flexed posture as well as anteversion of the head and neck without true flexion. The internal architecture of the bone trabeculae of the proximal end of the femur is appreciated better in the adult but is shown quite clearly in the child and provides many clues in relation both to normal and abnormal development.

I. Distal Femur The first secondary ossification center to appear in human development is that of the distal femur. This invariably is present normally at birth in 100% of girls and boys. It is always a single ossification center, which widens in both the coronal and sagittal planes with growth. This secondary center has been noted as early as 28 weeks of gestation and generally is described as being present in all fetuses after 33 weeks (221). Sonographic identification of the distal femoral center increases between its earliest appearance at the 28th gestational week. Its normal diameter at birth has been listed at 5 mm. The peripheral regions of the secondary ossification center of the distal femur frequently are irregular in shape on postnatal radiographs, a finding that must not be considered pathologic. This rough fringelike margin is especially common between age 2 and 6 years. Serial anteroposterior and lateral radiographic projections document distal femoral epiphyseal development. The configuration of the epiphyseal region changes little from the fetal period to that at skeletal maturity. The growth plate of the distal femur has an irregular shape. From birth to skeletal maturity in the anteroposterior projection, there is a double-concave shape toward the metaphysis. From 8 years onward, the depth of the concavity is relatively greater. On the lateral projection from birth to approximately 7 years of age, the epiphyseal growth plate projects in a transverse plane; during the next few years a slight double-concave curve can be noted; and over the final few years, a U-shaped growth plate concave to the metaphysis is seen.

J. Proximal Tibia The proximal tibial epiphyseal ossification center is also present in virtually 100% of individuals at birth (252). This secondary ossification center in order of appearance is second only to that of the distal femur. It develops during the fetal period somewhat later than the distal femur and occasionally is absent for a couple of months in the human at birth. Even with slow development, it should be present in 100% of individuals regardless of sex by 3 months of age. The initial secondary ossification center is centrally placed and spherical.

At the anterior, midline, and inferior parts of the developing tibial epiphysis the cartilage is continuous toward the area that will become the tibial tubercle. From birth to approximately 2 years of age in the anteroposterior projection, the plate is slightly curved, concave toward the metaphysis. Over the next several years, the plate approaches the true transverse plane, but again toward the end of growth a slight double concavity is seen. In the lateral projection from birth to approximately 13 years in the male and approximately 10 years in the female, the plate appears level although it is continuous with the apophyseal plate anteriorly and curves inferiorly at a sharp angle. This becomes readily apparent when anterior ossification commences. The secondary ossification center in the tibial tuberosity forms late toward the end of the first decade (253). It forms initially in girls between 8 and 14 years of age, with 64% present by age 10 years, and in boys between 9 and 15 years of age, with 59% present by age 12 years. The tibial tubercle ossification center initially forms distally. With time, the two centers approach each other; the anterior and inferior portion of the initial primary ossification center is extended distally and the secondary ossification center of the tibial tubercle is extended proximally. The tuberosity physis closes from proximal to distal.

K. Distal Tibia The secondary ossification center of the distal tibia appears in both girls and boys around 5 months of age, with that in girls seen only a few weeks before that in boys (216, 250, 254). The median age of appearance in girls is 5 months (range: birth to 1 year; 93% by 6 months) and that in boys is 51A months (range: 2 months to 1 year; 85% by 6 months). Several developmental points are of interest. By 3 years of age the distal fibula physis is at the same level as the distal tibial articular surface, a relationship that is then maintained until skeletal maturity. The distal tibial growth plate is horizontal on radiographic projections until around 3 years of age. At this time the medial one-third of the growth line is characterized by an undulation in which a focal area of the physeal region is indented superiorly into the metaphyseal region before regaining its linear orientation. This conformation is maintained until skeletal maturity. By 5 years of age the medial part of the distal tibial secondary ossification center begins to extend inferiorly into the medial malleolar cartilage mass. By 10 years of age approximately 50% of the medial malleolar volume has ossified. In approximately 20% of normal children there is an independent ossification center at the tip of the medial malleolus, although most of these fuse to the main distal tibial ossification center during adolescence (277). Some descriptions of these extra centers within the medial malleolus have noted their presence registered as high as 47% in girls and 17% in boys (318). The average age of appearance of these accessory centers was 7.6 years in gifts and 8.7 years in boys. It is important not to diagnose these as fractures because they are almost always

SECTION XIV ~ Radiographic Characteristics in Development of Major Long Bone Epiphyses

brought to light in association with radiographs for foot and ankle injuries. The plate is transverse in the anteroposterior projection, except for the slight focal proximal displacement in the medial one-third. The lateral projection shows a gentle curvature of the plate convex toward the metaphysis. The distal tibial epiphysis relates to the metaphysis as a cylinder with the long axis in the coronal plane. Closure of the distal tibial physis is unique passing from the medial side, which closes earliest to the lateral side. Ogden indicates that this closure pattern can occur over a period as long as 1.5 years.

L. Proximal Fibula The median age at appearance of the proximal fibular secondary ossification center in girls is at 3 years 1 month of age (range: 1-6 89 years; 97% by 489 years) and in boys at 4 years 5 months of age (range: 2-6 89 years; 92% by 5 years). The plate lies in a transverse plane in all projections.

117

and physeal presence because the former invariably diminishes before actual physeal obliteration occurs. The sequence of union of epiphyses in the human from first to last as defined by Stevenson is distal humerus, proximal radius, proximal ulna, proximal fibula, distal tibia, distal fibula, proximal tibia, distal radius, distal ulna, and proximal humerus. Humphrey and Oilier noted that the end of the shaft where growth was greatest united last. Payton (272) demonstrated, in the pig, that some epiphyseal growth occurred at the side nearest the joint, by which is meant the area under the articular cartilage. The epiphyseal growth is greater at the end of the bone where diaphyseal growth is greater but the proportionate difference is less. In Table VI, the times of growth plate closure are listed. Of increasing importance in a clinical sense is recognition of the pattern of closure in each of the specific physes. There can be considerable variation in the time of physeal closure of the various epiphyses in the body, epiphyses at the proximal and distal ends of any particular bone, and even within any epiphysis itself.

M. Distal Fibula The plate lies in the transverse plane in all projections (250). The median age at appearance of the secondary ossification center in girls is 189 years (range: 4 months to 3 years; 100% by 18 months) and in boys is 1 year 4 months (range: 4 months to 289 years; 10% by 2 years). The distal fibular epiphysis may exhibit irregular ossification, although accessory centers are found in only 1%.

N. Relationship of Proximal and Distal Tibial and Fibular Physes The proximal tibial physis is always more proximal to the physis of the fibula. The proximal tibial-fibular physeal difference was determined to be 0.4 cm prior to the first year, with a gradual increase to approximately 2.0 cm at age 12 years (22). The distal fibular physis is always more distal in position than the distal tibial physis. The distal fibular growth plate lies at the level of the tibiotalar articular surface. The distal tibial-fibular physis difference was also 0.4 cm prior to the first year of age and increased to approximately 1.0 cm at the age of 12 years.

O. Time and Pattern of Physeal Closure The time of physeal closure at each epiphyseal region has also been a matter of study since the nineteenth century due to its clinical importance and due to the fact that it seemingly could be defined relatively easily. Whereas it is true that no further growth will occur once the physeal cartilage has been obliterated, physeal growth invariably has diminished to negligible amounts and then terminated long before there is radiologic evidence of complete physeal disappearance. It is important to make the distinction between physeal function

P. Distal Tibial Growth Plate Closure An example of asymmetric physeal closure with important clinical implications is that of the distal tibia. This invariably closes not uniformly but over a several-month period from the medial side to the lateral side (216). At a certain stage, therefore, the medial one-third to one-half of the physis is closed whereas the lateral segment is open. This leads to a characteristic type of fracture, the Tillaux fracture, in which the Salter Harris type III conformation is seen with the physeal transverse component along the lateral aspect and the fracture then cracking into the secondary ossification center and out through the articular cartilage. The variable pattern of closure also contributes to the characteristic triplane fracture of the distal tibial epiphysis in the latter few years of growth.

Q. Proximal Femoral, Proximal Tibial, Metatarsal, and Phalangeal Growth Plate Closure Dvonch and Bunch (93) have studied closure of 25 proximal femoral and 10 proximal tibial growth plates. They felt that the proximal femur physis initiated closure superiorly, with closure proceeding inferiorly in all instances. Haines et al. (139) also identified closure of the proximal femoral epiphysis beginning peripherally at its superolateral margin. In the proximal tibia, closure occurred initially along the anteromedial aspect of the tibia and tibial tubercle and then proceeded posteriorly. Union of the larger epiphyses of bones other than those Of the hands and feet begins at or near the periphery of the cartilaginous physes. The union of the phalangeal physes is different from that of the major long bones

118

CHAPTER 1 9 Developmental Bone Bioloyy

as it occurs initially centrally and then proceeds centrifugally toward the periphery.

XV. WHY EPIPHYSES FORM AND THE EVOLUTION OF EPIPHYSES An important feature underlying consideration of the epiphyseal regions and one that has been referred to intermittently throughout the last century concerns the basic question of why epiphyses form. The contribution of the end of the bones to longitudinal growth was apparent as early as the mid-1700s, beginning with the classic experiments of Hales (140) and Duhamel (88-92) (Table I). The reasons for occurrence of the secondary ossification centers have been unclear. Two mechanisms have been proposed for their presence although definitive experimental verification of these mechanisms really is not feasible. The two theories advanced for the formation of secondary ossification centers are (1) mechanical, in the sense that they are required for the appropriate distribution of support for the developing ends of the bone, and (2) nutritional-vascular, wherein the central regions of the epiphysis, farthest from the peripheral vasculature, are considered to be prone to calcification of the matrix and subsequent ossification. Observations frequently were made that the secondary ossification centers formed in the central regions of the epiphyses and in particular in those of larger animals. Those animals that are relatively large at birth and that walk in the first few hours to days after birth are characterized by far more advanced development of the secondary ossification centers, in particular of the knee region, than is seen in the human in whom walking does not begin until the second year of life. In ungulates including ox, bovine, and lamb epiphyses, the secondary ossification centers of the distal femur and proximal tibia are developed far more extensively at birth than those in the human. These features were felt to be consistent with the interpretation that the secondary ossification centers formed in regions of relatively poor or diminished nutrition because the epiphyseal cartilage received its nutrition either via diffusion of nutrients through the articular surface or via penetration of vascular canals that had their origin in the perichondrium. It was felt that the chondrocytes in the most central regions of the larger epiphyses tended to degenerate because of relatively diminished nutrition, which provided a signal for calcification of the adjacent matrix. The mineralization then provided a signal that allowed for vascular invasion by way of the cartilage canals at which time the endochondral ossification mechanism was activated. Parsons addressed the pressure or articular epiphyses at the ends of the developing long bones in 1905 and focused on the questions (1) "how these epiphyses come," (2) "what use they are," and (3) "why they come" (268). His use of the term epiphysis refers to the secondary ossification cen-

ters. He noted that they were not essential for growth because birds did not develop epiphyses except at the anterior part of the proximal tibiotarsus, which was really a traction epiphysis at the site of attachment of the patellar tendon. In terrestrial vertebrates other than birds the epiphyseal cartilage sometimes just remained cartilaginous, sometimes calcified only, and sometimes progressed to a true bony secondary center. He observed that "the larger the cartilaginous mass at the end of a long bone, the earlier will an epiphysis appear in it" and "that the first ossific deposit occurs in the very center of the cartilage, and nowhere near the perichondrium." Parsons indicated that the secondary ossification centers did not help in growth, which occurred almost exclusively on the diaphyseal side of the growth plate, did not serve as protection for the physeal region, and did not contribute meaningfully to the shaping of the articular end of the bone. Although admitting "that I have heard no satisfactory explanation of the good which these epiphyses do," he pointed out that in certain ungulates, such as ox, sheep, and lambs, the secondary ossification centers appeared long before birth and that a lamb 2 days old shows them as well-developed as a human of 12 or 13 years. He hypothesized that because these animals are able to walk almost as soon as they are born the secondary ossification centers might have represented an adaptation from "some mechanical need." He felt that "some mechanical cause will be found for the early appearance of these ungulate epiphyses as the subject comes to be further inquired into." He speculated that it would eventually come down to some answer that had both chemical and mechanical components. The triggering formative mechanism was not pressure per se, which he felt would be least in the center of cartilage region, but rather one of the nutrition. The secondary centers begin "as a degenerative process in the least vascular part of the cartilaginous end of the bone, that is to say the center of it, and one would expect that the larger the mass of cartilage the less well-nourished would the center of it be, and so the more liable to the deposit of lime salts." This observation was consistent with the initial presence of secondary centers in the central part of the epiphyseal cartilage and earlier in the larger epiphyses. Haines stressed that epiphyses in general are a compromise arrangement to allow for the simultaneous growth and function of bones (137). The concept of appositional growth is established clearly for bone tissue and necessitates the epiphyseal cartilage mechanism to allow for interstitial growth at both ends. The development of a flattened growth plate is as much a reflection of the entire long bone biophysical environment as it is of any mechanism of orientation. One can think of the bulky cartilage epiphyses at each end as being under pressure, being attached to each other by the continuous outer fibrous layer of the periosteum and essentially separated by a strut of cortical bone holding them apart in between. Thus, the epiphysis takes the pull not only of the muscles directly inserted into it but also of those usually described as being inserted into the shaft of the bone. It is

References argued that if the articular cartilage and the growth cartilages can be separated, then each can be arranged in the most advantageous structural position, the former for giving a joint surface of appropriate shape and orientation even if the articular surface is offset from the main axis of the shaft and the growth plate for directing the arrangement of trabeculae along the long axis. In land living, stress beating animals the weakness of this large mass of undifferentiated cartilage between the two layers would be apparent, a weakness overcome by the introduction of the secondary ossification centers of calcification or ossification. Haines wrote a detailed review of the evolution of epiphyses, stressing the structural changes in the various groups (137). Cartilaginous epiphyses are present in bony fish and there is a clear arrangement of periosteal and endochondral bone in the long bones. The structure in fish indicates that the mechanism of growth is the same as that in land animals and that fish have epiphyses. They differ in that they have no secondary centers of calcification or ossification and no physeal ordering of flattened or hypertrophic chondrocytes. Haines defines the most primitive tetrapod with secondary centers as sphendon, in which there is central cartilage calcification, and outlines development of formal bony centers in lizards and mammals. The evolutionary changes of cartilage canals for vascularity of epiphyseal cartilage and secondary centers of calcification and ossification are concerned with better nutrition, strengthening, and better functional positioning of component parts of the epiphyses.

References 1. Acheson RM (1957) The Oxford method of assessing skeletal maturity. Clin Orthop Rel Res 10:19-39. 2. Aizawa T, Kokubun S, Tanaka Y (1997) Apoptosis and proliferation of growth plate chondrocytes in rabbits. J Bone Joint Surg [Br] 79:483-486. 3. Akesson K (1995) Biochemical markers of bone turnover. Acta Orthop Scand 66:376-386. 4. Alexander CJ (1976) Effect of growth rate on the strength of the growth plate-shaft junction. Radiology 1:67-76. 5. Amamilo SC, Bader DL, Houghton GR (1985) The periosteum in growth plate failure. Clin Orthop Rel Res 194:293-305. 6. Anderson HC (1967) Electron microscopic studies of induced cartilage development and calcification. J Cell Bio135: 81-101. 7. Anderson M, Green WT, Messner MB (1963) Growth and predictions of growth in the lower extremities. J Bone Joint Surg 45A: 1-14. 8. Anderson M, Messner MB, Green WT (1964) Distribution of lengths of the normal femur and tibia in children from 1 to 18 years of age. J Bone Joint Surg 46A:1197-1202. 9. Arkin AM, Katz JF (1956) The effects of pressure on epiphyseal growth. J Bone Joint Surg 38A:1056-1076. 10. Arsenault AL (1987) Microvascular organization at the epiphyseal-metaphyseal junction of growing rats. J Bone Miner Res 2:143-149.

119

11. Ayad S, Boot-Handford RE Humphries MJ, Kadler KE, Shuttleworth CA (1994) The Extracellular Matrix Facts Book. San Diego: Academic Press. 12. Baer MJ, Ackerman JL (1966) A longitudinal vital staining method for the study of apposition in bone. In: Studies on the Anatomy and Function of Bone and Joints. pp. 81-92, New York: Springer-Verlag. 13. Baer KE, von (1828, 1837) Ueber Entwickelungsgeschichte der Thiere, Beobachtung und Reflexion I. Thiel, 1828; II. Thiel, 1837 Konigsberg, Borntrager. 14. Ballock T, Heydemann A, Wakefield LM, Flanders KC, Roberts A, Spom MB (1993) TGF-131 prevents hypertrophy of epiphyseal chondrocytes: Regulation of gene expression for cartilage matrix proteins and metalloproteases. Dev Biol 158: 414-429. 15. Bardeen CR (1905) Studies of the development of the human skeleton. The development of the skeleton of the posterior limb. Am J Anat 4:279-302. 16. Bardeen CR (1910) Development of the skeleton and of the connective tissues. In: Manual of Human Embryology, ed E Keibel, FP Mall. pp. 292-397, Philadelphia: JB Lippincott. 17. Barnett CH, Lewis OJ (1958) The evolution of some traction epiphyses in birds and mammals. J Anat 92:593-600. 18. Barnewolt CE, Shapiro F, Jaramillo D (1997) Normal gadolinium-enhanced MR images of the developing appendicular skeleton: Part 1. Cartilaginous epiphysis and physis. AJR 169:183-189. 19. Barreto C, Wilsman NJ (1994) Hypertrophic chondrocyte volume and growth rates in avian growth plates. Res Vet Sci 56:33-61. 20. Baur A (1857) Zur lehre vonder verknocherung des primordialen knorpels. Archiv Anat Physiol 347-353. 21. Baylink DJ, Finkelman RD (1993) Growth factors to stimulate bone formation. J Bone Miner Res 8:565-572. 22. Beals RK, Skyhar M (1984) Growth and development of the tibia, fibula, and ankle joint. Clin Orthop Rel Res 182: 289-292. 23. Becks H, Simpson ME, Evans HM (1945) Ossification at the proximal tibial epiphysis in the rat. Anat Rec 42:109-119. 24. Belchier J (1736) Of the bones of animals changed to a red color by aliment only. Phil Trans Roy Soc London 39: 287-288. 25. Benninghoff A (1925) Form und bau der gelenkknorpel in ihren beziehungen zur funktion. Zwieter teil: Der aufbau des gelenkknorpels in seinen beziehungen zur funktion. Zeit f Zellforsch u Mikroskop Anat 2:783-862. 26. Bemays A (1878) Die entwicklungsgeschichte des kniegelenkes des menschen, mit bemerkungen uber die gelenke. Zeit f Anat Entwicklung 4:403-446. 27. Bernfield M, Gotte M, Park PW, et al. (1999) Functions of cell surface heparan sulfate proteoglycans. Ann Rev Biochem 68:729-777. 28. Bianco P, Fisher LW, Yuong ME, Termine JD, Robey PG (1990) Expression and localization of the two small proteoglycans biglycan and decorin in developing human skeletal and nonskeletal tissues. J Histochem Cytochem 38:1549-1563. 29. Bidder A (1906) Osteobiologie. Arch f Mikroskop Anat 68: 137-213. 30. Birkedal-Hansen H (1995) Proteolytic remodeling of extracellular matrix. Curr Op Cell Biol 7:728-835.

120

CHAPTER 1 ~

Developmental Bone Biology

31. Brachet A (1893) Sur la resorption du cartilage. J Int Anat Physiol 10:393. 32. Brashear RH (1963) Epiphyseal avascular necrosis and its relation to longitudinal bone growth. J Bone Joint Surg 45A: 1423-1438. 33. Breur GJ, Turgai J, Vanenkevort BA, Farnum CE, Wilsman NJ (1994) Stereological and serial section analysis of chondrocytic enlargement in the proximal tibial growth plate of the rat. Anat Rec 239:255-268. 34. Brewton RG, Mayne R (1994) Heterotypic type II, IX fibrils: Comparison of vitreous and cartilage forms. In: Extracellular Matrix Assembly and Structure, eds. PD Yurchenco, DE Birk, RP Mecham. pp. 129-170. San Diego: Academic Press. 35. Brighton CT, Hunt RM (1978) The role of mitochondria in growth plate calcification as demonstrated in a rachitic model. J Bone Joint Surg 60A:630-639. 36. Brighton CT (1978) Structure and function of the growth plate. Clin Orthop Rel Res 136:22-32. 37. Broca (1852) Sur quelque points de l'anatomie pathologique du rachitisme. Bull Soc Anat Paris 27:141-167. 38. Broca (1852) Sur quelque points de l'anatomie pathologique du rachitisme. Bull Soc Anat Paris 27: 542-596. 39. Brookes M (1958) The vascularization of long bones in the human foetus. J Anat 92:261-267. 40. Bruckner P, van der Rest M (1994) Structure and function of cartilage collagens. Microscopy Res Tech 28:378-384. 41. Bryant PJ (1993) The polar coordinate model goes molecular. Science 259:471-472. 42. Buckwalter JA (1983) Proteoglycan structure in calcifying cartilage. Clin Orthop Rel Res 172:207-232. 43. Buckwalter JA, Mower D, Schafer J, Ungar R, Ginsberg B, Moore K (1985) Growth-plate-chondrocyte profiles and their orientation. J Bone Joint Surg 67A:942-955. 44. Buckwalter JA, Mower D, Ungar R, Schafer J, Ginsberg B (1986) Morphometric analysis of chondrocyte hypertrophy. J Bone Joint Surg 68A:243-255. 45. Buckwalter JA, Hunziker E, Rosenberg L, Coutts R, Adams M, Eyre D (1987) Articular cartilage: Composition and structure. In: Injury and Repair of the Musculoskeletal Soft Tissues, ed SL-Y Woo, JA Buckwalter. pp. 405-425, Park Ridge, IL: American Academy of Orthopaedic Surgeons. 46. Buckwalter JA, Rosenberg LC, Ungar R (1987) Changes in proteoglycan aggregates during cartilage mineralization. Calcif Tissue Int 41:228-236. 47. Buckwalter JA, Roughley PJ, Rosenberg LC (1994) Age-related changes in cartilage proteoglycans: Quantitative electron microscopic studies. Microscopy Res Tech 28:398-408. 48. Burkus JK, Ogden JA (1984) Development of the distal femoral epiphysis: A microscopic morphological investigation of the zone of Ranvier. J Pediatr Orthop 4:661-668. 49. Burkus JK, Ganey TM, Ogden JA (1993) Development ofthe cartilage canals and the secondary center of ossification in the distal chondroepiphysis of the prenatal human femur. Yale J Biol Med 66:193-202. 50. Burt DW, Paton IR (1992) Evolutionary origins of the transforming growth factor-b gene family. DNA Cell Biol 11: 497-510. 51. Butler WT (1989) The nature and significance of osteopontin. Conn Tiss Res 23:123-136.

52. Campt JD, Cilley EIL (1931) Diagrammatic chart showing time of appearance of the various centers of ossification and period of union. Am J Roent Rad Ther 26:905. 53. Carey EJ (1920) Studies in the dynamics of histogenesis. Growth motive force as a dynamic stimulus to the genesis of muscular and skeletal tissues. Anat Rec 19:199-235. 54. Carey EJ (1921) Studies in the dynamics of histogenesis. Am J Anat 29:93-115. 55. Carey EJ (1922) Direct observations on the transformation of the mesenchyme in the thigh of the pig embryo (sus scrofa), with especial reference to the genesis of the thigh muscles, of the knee and hip joints, and of the primary bone of the femur. J Morph 37:1-77. 56. Carey EJ (1929) Studies in the dynamics of histogenesis. XIV. Radiology 13:127-168. 57. Carter DR, Orr TE, Fyhrie DP, Schurman DJ (1987) Influences of mechanical stress on prenatal and postnatal skeletal development. Clin Orthop Rel Res 219:237-250. 58. Carter DR, Wong M (1988) Mechanical stresses and endochondral ossification in the chondroepiphysis. J Orthop Res 6:148-154. 59. Carter DR, Wong M (1990) Mechanical stresses in joint morphogenesis and maintenance. In: Biomechanics of Diarthrodial Joints, volume 2, eds VC Mow, A Ratcliffe, SL-Y Woo. pp. 155-174, New York: Springer-Verlag. 60. Carter DR, Orr TE (1992) Skeletal development and bone functional adaptation. J Bone Miner Res 7:389-395. 61. Carvell JE (1983) The relationship of the periosteum to angular deformities of long bones. Clin Orthop Rel Res 173: 262-274. 62. Chappard D, Alexandre C, Chol R, Riffat G (1983) Les canaux intrachondraux: Histogenese, anatomie et histophysiologic de la vascularisation cartilagineuse du foetus humain. Lyon Med 249:417-428. 63. Chappard D. Alexandre C, Riffat G (1986) Uncalcified cartilage resorption in human fetal cartilage canals. Tissue Cell 18:701-707. 64. Chen Q, Johnson DM, Haudenschild DR, Goetinck PF (1995) Progression and recapitulation of the chondrocyte differentiation program: Cartilage matrix protein is a marker for cartilage maturation. Dev Biol 172:293-306. 65. Cohn MJ, Tickle C (1996) Limbs: A model for pattern formation within the vertebrate body plan. Trends Genet 12: 253-257. 66. Cohn MJ, Izpisua-Belmonte JC, Abud H, Heath JK, Tickle C (1995) Fibroblast growth factors induce additional limb development from the flank of chick embryos. Cell 80:739-746. 67. Cole AA, Wezeman FH (1985) Perivascular cells in cartilage canals of the developing mouse epiphysis. Am J Anat 174: 119-129. 68. Cole AA, Wezeman FH (1987) Morphometric analysis of cartilage canals in the developing mouse epiphysis. Acta Anat 128:93-97. 69. Cowell HR, Hunziker EB, Rosenberg L (1987) The role of hypertrophic chondrocytes in endochondral ossification and in the development of the secondary centers of ossification. J Bone Joint Surg 69A:159-161. 70. Cowles EA, DeRome ME, Pastizzo G, Brailey LL, Gronowicz GA (1998) Mineralization and the expression of matrix

References

71.

72. 73.

74. 75.

76.

77. 78.

79. 80.

81. 82.

83.

84.

85. 86.

87.

88.

89. 90.

91.

proteins during in vivo bone development. Calcif Tissue Int 62:74-82. Crossley PH, Minowada G, MacArthur CA, Martin GR (1996) Roles for FGF8 in the induction, initiation, and maintenance of chick limb development. Cell 84:127-136. Czermak N (1888) Vergleichende studiam uber die entwicklung des knochen u knorpelgewebes. Anat Anzeig 3:470-480. Dale GG, Harris WR (1958) Prognosis of epiphyseal separation--An experimental study. J Bone Joint Surg 40B: 116-122. Delaporte F (1983) Theories of osteogenesis in the eighteenth century. J Hist Biol 16:343-360. Delgado-Baeza E, Gimenez-Ribotta M, Miralles-Flores C, Nieto-Chaguaceda A, Santos-Alvarez I (1991) Morphogenesis of cartilage canals: Experimental approach in the rat tibia. Acta Anat 142:132-137. Delgado-Baeza E, Gimenez-Ribotta M, Miralles-Flores C, Nieto-Chaguaceda A, Santos-Alvarez I (1991) Relationship between the cartilage canal and the perichondrium in the rat proximal tibial epiphysis. Acta Anat 141:31-35. Delpech JM (1828) De l'Orthomorphie, 2 volumes. Paris: Gabon. e Deng C, Wynshaw-Boris A, Zhou F, Kuo A, Leder P (1996) Fibroblast growth factor receptor 3 is a negative regulator of bone growth. Cell 84:911-921. Digby KH (1916) The measurement of diaphysial growth in proximal and distal directions. J Anat Physiol 50:187-188. Dimitriou CG, Kapetanos GA, Symeonides PP (1987) The effect of partial periosteal division on growth of the long bones. Clin Orthop Rel Res 236:265-269. Dobson J (1948) Robert Nesbitt. Pioneers of osteogeny. J Bone Joint Surg 30B:551-555. Dodds GS (1930) Row formation and other types of arrangement of cartilage cells in endochondral ossification. Anat Rec 46:385-399. Dodds GS (1932) Osteoclasts and cartilage removal in endochondral ossification of certain mammals. Am J Anat 50: 97-126. Duboule D (1992) The vertebrate limb: A model system to study the hox~om gene network during development and evolution. BioEssays 14:375-384. Dubreuil G, Baudrimont A (1950) Manuel Theorique et Pratique d'Histologie, 4th edition. Paris: Vigot Freres. Ducy P, Zhang Rui, Geoffrey V, Ridall AL, Karsenty G (1997) Osf2/Cbfal: A transcriptional activator of osteoblast differentiation. Cell 89:747-754. Ducy P, Schinke T, Karsenty G (2000) The osteoblast: A sophisticated fibroblast under central surveillance. Science 289: 1501-1504. Duhamel HL (1739) Sur une racine qui a la propriete de teindre en rouge les os des animaux vivants. Mem Acad Roy Sci 52:1-13. Duhamel HL (1741) Premiere memoire. Mem Acad Roy Sci 54:98. Duhamel HL (1742) Troisieme memoire. Sur le developpement et lacrue des os des animaux. Mem Acad Roy Sci 55: 354-370. Duhamel HL (1743) Quatrieme memoire sur les os. Mem Acad Roy Sci 56:87.

121

92. Duhamel HL (1743) Cinquieme memoire sur les os. Mem Acad Roy Sci 56:111-145. 93. Dvonch VM, Bunch WH (1983) Pattern of closure of the proximal femoral and tibial epiphyses in man. J Pediatr Orthop 3:498-501. 94. Dwek JR, Shapiro F, Laor T, Barnewolt CE, Jaramillo D (1997) Normal gadolinium-enhanced MR images of the developing appendicular skeleton: Part 2. Epiphyseal and metaphyseal marrow. AJR 169:191-196. 95. Echelard Y, Epstein DJ, St. Jacques B, Shen L, Mohler J, McMahon JA, McMahon AP (1993) Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 75:1417-1430. 96. Erlebacher A, Filvaroff EH, Gitelman SE, R Derynck (1995) Toward a molecular understanding of skeletal development. Cell 80:371-378. 97. Eyre DR (1980) Collagen: Molecular diversity in the body's protein scaffold. Science 207:1315-1322. 98. Eyre DR, Paz MA, Gallop PM (1984) Cross-linking in collagen and elastin. Ann Rev Biochem 53:717-748. 99. Eyre DR, Apon S, Wu JJ, Ericsson LH, Walsh KA (1987) Collagen type IX: Evidence for covalent linkages to type II collagen in cartilage. FEBS Lett 220:337-341. 100. Farnum CE, Wilsman NJ (1987) Morphologic stages of the terminal hypertrophic chondrocyte of growth plate cartilage. Anat Rec 219:221-232. 101. Fell HB, Robisson R (1929) The growth, development, and phosphatase activity of embryonic avian femora and limb buds cultivated in vitro. Biochem J 23:767-784. 102. Felts WJL (1954) The prenatal development of the human femur. Am J Anat 94:1-45. 103. Ferguson C, Alper E, Miclau T, Helms JA (1999) Does adult fracture repair recapitulate embryonic skeletal formation? Mech Dev 87:57-66. 104. Fietz MJ, Concordet J-P, Barbosa R, Johnson R, Krauss S, McMahon AP, Tabin C, Ingham PW (1994) The hedgehog gene family in Drosophila and vertebrate development. Development Sup 43-51. 105. Flecker H (1942) Time of appearance and fusion of ossification centers as observed by roentgenographic methods. Am J Roent 47:97-159. 106. Flourens MJ-P (1841) Nouvelles recherches concernant l'action de la garance sur les os. (Quatrieme memnoire). Comp Rend Hebdo Acad Sci 12:276-283. 107. Flourens MJ-P (1841) Recherches sur le developpement des os (Cinquieme memoire). Comp Rend Hebdo Acad Sci 13: 671-682. 108. Flourens MJ-P (1842) Recherches sur le developpement des os. (Septieme memoire). Comp Rend Acad Sci 15:875-883. 109. Flourens MJ-P (1847) Theorie experimentale de la formation des os. Paris: J-B Bailliere. 110. Flourens MJ-P (1861) Note sur le developpement des os en longuer. Comp Rend Acad Sci 62:186-189. 111. Floyd WE, Zaleske DJ, Schiller AL, Trahan C, Mankin HJ (1987). Vascular events associated with the appearance of the secondary center of ossification in the murine distal femoral epiphysis. J Bone Joint Surg 69A:185-190. 112. Francis CC, Werle PP (1939) The appearance of centers of ossification from birth to 5 years. Am J Phys Anthrop 24:273-299.

122

CHAPTER 1 ~ Developmental Bone Biolo~ty

113. Francis CC (1940) The appearance of centers of ossification from 6 to 15 years. Am J Phys Anthrop 27:127-138. 114. Francomano CA (1995) Key role for a minor collagen. Nature Genet 9:6-8. 115. Franz CH (1971) Epiphyseal stapling: A comprehensive review. Clin Orthop Rel Res 77:149-157. 116. Frost HM, Jee WSS (1994) Perspectives: Applications of a biomechanical model of the endochondral ossification mechanism. Anat Rec 240:447-455. 117. Frost HM, Jee WSS (1994) Perspectives: A vital biomechanical model of the endochondral ossification mechanism. Anat Rec 240:435-446. 118. Ganey TM, Love SM, Ogden JA (1992) Development ofvascularization in the chondroepiphysis of the rabbit. J Orthop Res 10:496-510. 119. Garces GL, Hermoso JAH (1991) Bone growth after periosteal stripping in rats. Int Orthop 15:49-52. 120. Gardner E, Gray DJ (1950) Prenatal development of the human hip joint. Am J Anat 87:163-211. 121. Gaunt SJ (1991) Expression patterns of mouse hox genes: Clues to an understanding of developmental and evolutionary strategies. BioEssays 13:505-513. 122. Gebhardt (1911) [In a discussion]. Verhandlungen Anat Ges 25:169-172. 123. Gegenbaum C (1864) Ueber die bildung des knochengewebes. Jena Zeit f Med Naturwissen 1:343-369. 124. Gelbke H (1951) The influence of pressure and tension on growing bone in experiments with animals. J Bone Joint Surg 33A:947-954. 125. Gerstenfeld LC, Landis WJ (1991) Gene expression and extracellular matrix ultrastructure of a mineralizing chondrocyte cell culture system. J Cell Biol 112: 501-513. 126. Gerstenfeld LC, Shapiro FD (1996) Expression of bonespecific genes by hypertrophic chondrocytes: Implications of the complex functions of the hypertrophic chondrocyte during endochondral bone development. J Cell Biochem 62:1-9. 127. Giambernardi TA, Grant GM, Taylor GP, Hay RJ, Maher VM, McCormick JJ, Klebe RJ (1997/1998) Overview of matrix metalloproteinase expression in cultured human cells. Matrix Biol 16:483-496. 128. Gibson G (1998) Active role of chondrocyte apoptosis in endochondral ossification. Microsc Res Tech 28:505-519. 129. Girdany BR, Golden R (1952) Centers of ossification of the skeleton. Am J Roent 68:922-924. 130. Glimcher MJ (1989) Mechanism of calcification: Role of collagen fibrils and collagen-phosphoprotein complexes in vitro and in vivo. Anat Rec 224:139-153. 131. Gray DJ, Gardner E (1969) The prenatal development of the human humerus. Am J Anat 124:431-446. 132. Haas SL (1917) The relation of the blood supply to the longitudinal growth of bone. Am J Orthop Surg 15:157-171 and 305-316. 133. Haas SL (1945) Retardation of bone growth by a wire loop. J Bone Joint Surg 27:25-36. 134. Haasbeek JF, Rang MC, Blackburn N (1995) Periosteal tether causing angular growth deformity: Report of two clinical cases and an experimental model. J Pediatr Orthop 15:677-681. 135. Haines RW (1933) Cartilage canals. J Anat 68:45-64. 136. Haines RW (1940) Note on the independence of sesamoid and epiphysial centres of ossification. J Anat 75:101-105.

137. Haines RW (1942) The evolution of epiphyses and of endochondral bone. Biol Rev 17:267-291. 138. Haines RW (1947) The development of joints. J Anat 81: 33-55. 139. Haines RW, Mohiuddin A, Okpa FI, Viega-Pires JA (1967) The sites of early epiphyseal union in the limb girdles and major long bones of man. J Anat 101:823-831. 140. Hales S (1727) Vegetable statics. Statistical essays. London: Innys and Woodward. 141. Halfter W, Dong S, Schurer B, Cole GJ (1998) Collagen XVIII is a basement membrane heparan sulfate proteoglycan. J Biol Chem 273:25404-25412. 142. Ham AW, Cormack DH (1979) Histology, 8th edition. Philadelphia: JB Lippincott. 143. Hansman CF (1962) Appearance and fusion of ossification centers in the human skeleton. Am J Roent 88:476-482. 144. Haraldsson S (1962) The vascular pattern of a growing and full grown human epiphysis. Acta Anat 48:156-167. 145. Havers C (1691) Osteologia Nova or Some New Observations of the Bones. London: S Smith. 146. Hayashi K (1992) Three-dimensional organization of the cartilage canalmA scanning electron microscopic study by vascular cast of the rabbit's femoral head. Jpn Orthop 66: 548-559. 147. Haynes JS (1990) Immunohistochemical localization of type X collagen in the proximal tibiotarsi of broiler chickens and turkeys. Anat Rec 227:307-313. 148. Heinegard D, Oldberg A (1989) Structure and biology of cartilage and bone matrix noncollagenous macromolecules. FASEB J 3:2042-2051 149. Held LI, Jr (1994) Models for Embryonic Periodicity. Basel: Karger. 150. Helms JA, Kim CH, Eichele G, Thaller C (1996) Retinoic acid signaling is required during early chick limb development. Development 122:1385-1394. 151. Hernandez JA, Serrano S, Marinoso ML, Aubia J, Lloreta J, Marrugat J, Diez A (1995) Bone growth and modeling changes induced by periosteal stripping in the rat. Clin Orthop Rel Res 320:211-219. 152. Hesser C (1926) Beitrag zur kenntnis der gelenkentwicklung beim menschen. Gegenbaurs morphologisches jahrbuch. Zeit f Anat Entwicklungs 55:489-567. 153. Hill AH (1939) Fetal age assessment by centers of ossification. Am J Phys Anthrop 24:251-272. 154. Hintzsche E (1928) Untersuchungen an stutzgeweben. 1. Ueber die bedentung der gefasskanale im knorpel nach befunden an distalen ende des menschlichen schenkelbeines. Zeit f Mikroskop Anat Forsch 12:62-126. 155. Hodge JA, McKibbin B (1969) The nutrition of mature and immature cartilage in rabbits. An autoradiographic study. J Bone Joint Surg 51B:140-147. 156. Hogan BLM, Blessing M, Winnier GE, Suzuki N, Jones CM (1994) Growth factors in development: The role of TGF-b related polypeptide signaling molecules in embryogenesis. Development Sup 53-60. 157. Holick MF (1966) Vitamin D: Photobiology, metabolism, mechanism of action, and clinical applications. In: Primer on the Metabolic Bone Diseases and Disorders of Mineral Metabolism, 3rd edition, ed MJ Favus. pp. 74-81, Philadelphia: Lippincott-Raven.

References

158. Holtrop ME (1967) The potencies of the epiphyseal cartilage in endochondral ossification. Koninkl nederl akademie van wetenschappen--Amsterdam. Proceedings, Series C, 70, No. 1, 21-28. 159. Holtrop ME (1967) Factors influencing the growth rate in endochondral ossification. Koninkl nederl akademie van wetenschappen--Amsterdam. Proceedings, Series C, 70, No. 1, 29-30. 160. Holtrop ME (1971) The ultrastructure of the hypertrophic chondrocyte. Isr J Med Sci 7:473-476. 161. Holtrop ME (1972) The ultrastructure of the epiphyseal plate. I. The flattened chondrocyte. Calif Tissue Res 9:131-139. 162. Holtrop ME (1972) The ultrastructure of the epiphyseal plate. II. The hypertrophic chondrocyte. Calcif Tissue Res 9: 140-151. 163. Hoshi K, Ejiri S, Ozawa H (2001) Localizational alterations of calcium phosphorus and calcification-related organics such as proteoglycans and alkaline phosphatase during bone calcification. J Bone Miner Res 16:289-298. 164. Houghton GR, Rooker GD (1981) The role of the periosteum in bone remodeling and deformity. In: Mechanical Factors and the Skeleton, IA Stokes. pp 128-132, London: John Libbey. 165. Howship J (1815) Formation of bone. Med-Chir Trans 6: 263-295. 166. Hueter C (1862) Anatomische studien an den extremitatengelenken neugeborener und erwachsener. Archiv Path Anat u Physiol u f Klin Med 25:572-599. 167. Hultenby K, Reinholt FP, Norgard M, Oldberg A, Wendel M, Heinegard D (1994) Distribution and synthesis of bone sialoprotein in metaphyseal bone of young rats show a distinctly different pattern from that of osteopontin. Eur J Cell Bio163: 230-239. 168. Hunt CD, Ollerich DA, Nielsen FH (1979) Morphology of the perforating cartilage canals in the proximal tibial growth plate of the chick. Anat Rec 194:143-158. 169. Hunter J (1841) On the growth of bones. In: Observations on Certain Parts of the Animal Oeconomy. ed R Owen. pp. 319-323, Philadelphia, Barrington: Has.well. 170. Hunter WL, Arsenault AL (1990). Vascular invasion of the epiphyseal growth plate: Analysis of metaphyseal capillary ultrastructure and growth dynamics. Anat Rec 227:223-231. 171. Hunter GK, Kyle CL, Goldberg HA (1994) Modulation of crystal formation by bone phosphoproteins: Structural specificity of the osteopontin-mediated inhibition of hydroxyapatite formation. Biochem J 300:723-728. 172. Hunziker EB, Schenk RK, Cruz-Orive LM (1987) Quantitation of chondrocyte performance in growth-plate cartilage during longitudinal bone growth. J Bone Joint Surg 69A: 162-173. 173. Hunziker EB, Schenk RK (1989) Physiological mechanisms adopted by chondrocytes in regulating longitudinal bone growth in rats. J Physio1414:55-71. 174. Hunziker EB (1994) Mechanism of longitudinal bone growth and its regulation by growth plate chondrocytes. Microscop Res Tech 28:505-519. 175. Hurrell DJ (1934) The vascularization of cartilage. J Anat 69: 47-61. 176. Husby OS, Sudmann B, Gjerdet NR, Hitland SU, Sudmann E (1987) Spontaneous correction of femoral torsion diaphyseal osteotomies studied in kittens. Acta Orthop Scand 58:113-116.

123

177. Iwamoto M, Yagami K, Shapiro IM, Leboy PS, Adams SL, Pacifici M (1994) Retinoic acid is a major regulator of chondrocyte maturation and matrix mineralization. Microscopy Res Tech 28:483-491. 178. Jaramillo D, Vilegas-Medina OL, Doty DK, Dwek JR, Ransil BJ, Mulkern RV, Shapiro F (1995) Gadolinium-enhanced MR imaging demonstrates abduction-caused hip ischemia and its reversal in piglets. Pediatr Radio125:578-587. 179. Jessell TM, Melton DA (1992) Diffusible factors in vertebrate embryonic induction. Cell 68:257-270. 180. Johnson RL, Tabin CJ (1997) Molecular models for vertebrate limb development. Cell 90:979-990. 181. Jones CM, Lyons KM, Hogan LM (1991) Involvement of bone morphogenetic protein-4 (BMP-4) and VGR 1 in morphogenesis and neurologenesis in the mouse. Development 111:531-542. 182. Kalayjian DB, Cooper RR (1972) Osteogenesis of the epiphysis: A light and electron microscopic study. Clin Orthop Rel Res 85:242-256. 183. Karrholm J, Hansson LI, Selvik G (1984) Longitudinal growth rate of the distal tibia and fibula in children. Clin Orthop Rel Res 191:121-128. 184. Kazzander J (1894) Ueber die entwickelung des kniegelenkes. Archiv Anat Entwick. 161-176. 185. Keith A (1919) Menders of the Maimed. London: Henry Froude, Oxford University Press. 186. Kember NF (1960) Cell division in endochondral ossification: A study of cell proliferation in rat bones by the method of tritiated thymidine autoradiography. J Bone Joint Surg 42B:824-839. 187. Kember NF (1971) Cell population kinetics of bone growth: The first 10 years of autoradiographic studies with tritiated thymidine. Clin Orthop Rel Res 76:213-230. 188. Kember NF (1978) Cell kinetics and the control of growth in long bones. Cell Tissue Kinet 11:477-485. 189. King TW (1844) On the progressive development of the osseous bodies of the vertebrae and sternum. London Med Gaz 34:73-77. 190. Kingsley DM, Bland AE, Grubber JM, Marker PC, Russell LB, Copeland NG, Jenkins NA (1992) The mouse short ear skeletal morphogenesis locus is associated with defects in a bone morphogenetic member of the TGF superfamily. Cell 71:399-410. 191. Kingsley DM (1994) The TGFI3 superfamily: New members, new receptors, and new genetic tests of function in different organisms. Genes Dev 8:133-146. 192. Knott L, Baily AJ (1998) Collagen cross-links in mineralizing tissues: A review of their chemistry, function, and clinical relevance. Bone 22:181-187. 193. Koelliker A (1850) Mikroskopische Anatomie oder Gewebelehre des Menschen. Leipzig: Verlag von Wilhelm Engelmann. 194. Koelliker A (1889) Handbuch der Gewebelehre des Menschen, volume 1. Leipzig: Verlag von Wilhelm Engelmann. 195. Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, et al. (1997) Targeted disruption of Cbfal results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell 89:756-764. 196. Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C, et al. (1999) OPGL is a key regulator of osteoclastogenesis,

124

197. 198.

199. 200. 201. 202.

203.

204. 205. 206.

207.

208.

209.

210. 211. 212. 213. 214. 215. 216.

217. 218.

CHAPTER

1 9

Developmental Bone Biology

lymphocyte development and lymph node organogenesis. Nature 397:315-323. Kosher RA (1998) Syndecan-3 in limb skeletal development. Microsc Res Tech 43:123-130. Kraus S, Concordet J-P, Ingham PW (1993) A functionally conserved homolog of the Drosophila segment polarity gene hh is expressed in tissues with polarizing activity in zebrafish embryos. Cell 75:1431-1444. Krstic RV (1985) General Histology of the Mammal. Berlin: Springer-Verlag. Krumlauf R (1992) Evolution of the vertebrate hox homeobox genes. BioEssays 14:245-252. Krumlauf R (1994) Hox genes in vertebrate development. Cell 78:191-201. Kugler JH, Tomlinson A, Wagstaff A, Ward SM (1979) The role of cartilage canals in the formation of secondary centres of ossification. J Anat 129:493-506. Kulyk WM, Coelho CND, Kosher RA (1991) Type IX collagen gene expression during limb cartilage differentiation. Matrix 11:282-288. Lacroix P (1947) Organizers and the growth of bones. J Bone Joint Surg 29:292-296. Lacroix P (1951) The Organization of Bones. London: J and A Churchill. Landis WJ, Paine MC, Hodgens KJ, Glimcher MJ (1986) Matrix vesicles in embryonic chick bone: Considerations of their identification, number, distribution, and possible effects on calcification of extracellular matrices. J Ultrastruc Res 95:142-163. Landis WJ, Song MJ, Leith A, McEwen L, McEwen BF (1993) Mineral and organic matrix interaction in normally calcifying tendon visualized in three dimensions by high voltage electron microscopic tomography and graphic image reconstruction. J Struc Biol 110:39-54. Langenskiold A, Videman T, Nevalainen T (1993) Vital staining indicating cell migration towards the periphery in the growth plate. Acta Orthop Scand 64:683-687. Lanske B, Karaplis AC, Lee K, Luz A, Vortkamp A, Pirro A, Karperien M, Defize LHK, Ho C, Mulligan RC, Abou-Samra A-B, Juppner H, Segre GV, Kronenberg HM (1996) PTH/ PTHrP receptor in early development and Indian hedgehogregulated bone growth. Science 273:663-666. Larsen WJ (1993) Human Embryology. New York: Churchill Livingstone. Lawrence PA (1990) Compartments in vertebrates? Nature 344:382-383. Leser E (1888) Ueber die histologischen vorgange au der ossificationsgrenze. Arch Klin Chir 37:511-521. Levene C (1964) The patterns of cartilage canals. J Anat 98: 515-538. Levschin L. Zur entwickelung des knochenge webes. Bull de l'Acad Imp St. Peters 17:9-13. Lieberkuhn (1883) Mikroscop Morph Thierkorp 249. Love SM, Ganey T, Ogden JA (1990) Postnatal epiphyseal development: The distal tibia and fibula. J Pediatr Orthop 10: 298-305. Loven C (1863) Studier och undersokningar ofver benvafnaden etc. Medicinsk Archiv Carolinsk Inst 1. Lufti AM (1970) Study of cell multiplication in the cartilaginous upper end of the tibia of the domestic fowl by tritiated thymidine autoradiography. Acta Anat 76:454-463.

219. Lufti AM (1970) Mode of growth, fate, and functions of cartilage canals. J Anat 106:135-145. 220. Maas H (1924) Ueber den einfluss pathologischer druck-und zugspannungen auf das knochenwachstum. Zeit f Orthop Chir 44:352-375. 221. Mahoney BS, Callen PW, Filly RA (1985) The distal femoral epiphyseal ossification center in the assessment of thirdtimester menstrual age: Sonographic identification and measurement. Radiology 155:201-204. 222. Mall FP (1906) On ossification centers in human embryos less than 100 days old. Am J Anat 5:433-458. 223. Mankin HJ (1962) Localization of tritiated thymidine in articular cartilage of rabbits. I. Growth in immature cartilage. J Bone Joint Surg 44A:682-688. 224. Mankin HJ (1963) Localization of tritiated thymidine in articular cartilage of rabbits. III. Mature articular cartilage. J Bone Joint Surg 45A:529-540. 225. Maresh MM (1955) Linear growth of long bones of extremities from infancy through adolescence. Am J Dis Child 89: 725-742. 226. Mason IJ (1994) The ins and outs of fibroblast growth factors. Cell 78:547-552. 227. Maximow AA, Bloom W (1943) A Textbook of Histology, 4th edition. Philadelphia: WB Saunders. 228. McCarthy SM, Ogden JA (1982) Radiology of postnatal skeletal development V. Distal humerus. Skelet Radiol 7:239-249. 229. McCarthy SM, Ogden MD (1982) Radiology of postnatal skeletal development VI. Elbow joint, proximal radius, and ulna. Skelet Radiol 9:17-26. 230. McCormick D, Van Der Rest M, Goodship J, Lozano G, Ninomiya Y, Olsen BR (1987) Structure of the glycosaminoglycan domain in the type IX collagen proteoglycan. Proc Natl Acad Sci USA 84:4044-4048. 231. McKee MD, Glimcher MJ, Nanci A (1992) High-resolution immunolocalization of osteopontin and osteocalcin in bone and cartilage during endochondral ossification in the chicken tibia. Anat Rec 234:479-492. 232. McKibbin B, Holdsworth FW (1966) The nutrition of immature joint cartilage in the lamb. J Bone Joint Surg 48B: 793-803. 233. Melton DA (1991) Pattern formation during animal development. Science 252:234-241. 234. Mendler M, Eich-Bender SG, Vaughan L, Winterhalter KH, Brucker P (1989) Cartilage contains mixed fibrils of collagen types II, IX, and XI. J Cell Biol 108:191-197. 235. Morgan JD, Somerville EW (1960) Normal and abnormal growth at the upper end of the femur. J Bone Joint Surg 42B: 264-272. 236. Morgan BA, Tabin C (1994) Hox genes and growth: Early and late roles in limb bud morphogenesis. Development Sup. 181-186. 237. Muller H (1858) Ueber die entwickelung der knochensubstanz nebst bemerkungen uber den bau rachitischer knochen. Zeit f Wissen Zool 9:149-233. 238. Mundlos S, Meyer R, Yamada Y, Zabel B (1991) Distribution of cartilage proteoglycan (aggrecan) core protein and link protein gene expression during human skeletal development. Matrix 11:339-346. 239. Murray PDF (1936) Bones. Cambridge: Cambridge University Press.

References 240. Nathanson MA (1994) Basic science and pathology in vitro proteoglycan synthesis in response to extracts of demineralized bone. Clin Orthop Rel Res 263-281. 241. Nelson CE, Morgan BA, Burke AC, Laufer E, DiMambro E, Murtaugh LC, Gonzales E, Tessarollo L, Parada LF, Tabin C (1996) Analysis of hox gene expression in the chick limb bud. Development 122:1449-1466. 242. Nesbitt R (1736) Human Osteogeny. London: T. Wood. 243. Niswander L (1996) Growth factor interaction in limb development. Ann NY Acad Sci 785:23-26. 244. Niswander L, Tickle C, Vogel A, Booth I, Martin GR (1993) FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb. Cell 75:579-587. 245. Noback CR, Robertson GG (1951) Sequences of appearance of ossification centers in the human skeleton during the first five prenatal months. Am J Anat 89:1-28. 246. Nonidez JF, Windle WF (1949) Textbook of Histology. New York: McGraw-Hill. 247. Nurminskaya M, Linsenmayer TF (1996) Identification and characterization of up-regulated genes during chondrocyte hypertrophy. Dev Dynam 206:260-271. 248. Ogden JA, Conlogue GJ, Jensen P (1978) Radiology of postnatal skeletal development: The proximal humerus. Skelet Radiol 2:153-160. 249. Ogden JA, Beall JK, Conlogue GJ, Light TR (1981) Radiology of postnatal skeletal development IV. Distal radius and ulna. Skelet Radiol 6:255-266. 250. Ogden JA, McCarthy SM (1983) Radiology of postnatal skeletal development VIII. Distal tibia and fibula. Skelet Radiol 10:209-220. 251. Ogden JA (1984) Growth slowdown and arrest lines. J Pediatr Orthop 4:409-415. 252. Ogden JA (1984) Radiology of postnatal skeletal development IX. Proximal tibia and fibula. Skelet Radiol 11:169-177. 253. Ogden JA (1984) Radiology of postnatal skeletal development X Patella and tibial tuberosity. Skelet Radiol 11: 246-257. 254. Ogden JA, Lee J (1990) Accessory ossification patterns and injuries of the malleoli. J Pediat Orthop 10:306-316. 255. O'Hara JP, III, Dommisse GF (1983) Extraosseous blood supply to the neonatal femoral head. Clin Orthop Rel Res 174: 293-287. 256. Oilier L (1861) De la part proportionelle qui revient a chaque extremite des os des membres dans leur accroisment en longuer. J Physiol de 1' Homme et des Animaux 4:87-102. 257. Oilier L (1861) De l'accroissement en longeur des os des membres et de la part proportionelle qu'y prennent leurs deux extremities. Comp Rend Hebdo Acad Sci 62:130-132. 258. Oilier L (1867) Traite experimental et clinique de la regeneration des os et de la production artificielle du tissu osseux. Paris: Masson. 259. Olsen BR (1991) Collagen biosynthesis. In: Cell Biology of Extracellular Matrix, 2nd edition, ed ED Hay. pp. 177-220, New York: Plenum Press. 260. Olsen BR (1995) New insights into the function of collagens from genetic analysis. Curr Op Cell Biol 7:720-727. 261. O'Rahilly R, Gardner E, Gray DJ (1959) The skeletal development of the hand. Clin Orthop Rel Res 13:42-50. 262. O'Rahilly R, Gardner E, Gray DJ (1960) The skeletal development of the foot. Clin Orthop Rel Res 16:7-14.

125

263. O'Rahilly R, Gardner E (1975) The timing and sequence of events in the development of the limbs in the human embryo. Anat Embryol 148:1-23. 264. O'Rahilly R (1979) Early human development and the chief sources of information on staged human embryos. Europ J Obstet Gynec Reprod Biol 9:273-280. 265. Osborne D, Effmann E, Broda K, Harrelson J (1980) The development of the upper end of the femur with special reference to its internal architecture. Pediatr Radiol 137: 71-76. 266. Parr BA, McMahon AP (1995) Dorsalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb. Nature 374:350-353. 267. Parsons FG (1904) Observations on traction epiphyses. J Anat Physiol 38:248-258. 268. Parsons FG (1905) On pressure epiphyses. J Anat Physio139: 402-412. 269. Parsons FG (1908) Further remarks on traction epiphyses. J Anat Physio142:388-396. 270. Pauwels F (1980) Biomechanics of the Locomotor Apparatus. pp. 324-328, 375-407, 508-513, New York: SpringerVerlag. 271. Payton CG (1932) The growth in length of the long bones in the madder-fed pig. J Anat 66:414-425. 272. Payton CG (1933) The growth of the epiphyses of the long bones in the madder-fed pig. J Anat 67:371-381. 273. Poland J (1898) Traumatic Separation of the Epiphyses. London: Smith, Elder. 274. Poole AR (1986) Proteoglycans in health and disease: Structures and functions. Biochem J 236:1-14. 275. Poole AR, Matsui Y, Hinek A, Lee ER (1989) Cartilage macromolecules and the calcification of cartilage matrix. Anat Rec 224:167-179. 276. Poole AR, Pidoux I (1989) Immunoelectron microscope studies of type X collagen in endochondral ossification. J Cell Biol 190:2547-2554. 277. Powell HDW (1961) Extra center of ossification for the medical malleolus in children: Incidence and significance. J Bone Joint Surg 43B:107-113. 278. Prenant A, Maillard L, Bouin P (1911) Traite d'Histologie, volume 2, Histologie et Anatomie Microscopique, eds A Prenant, P Bouin. Paris: Masson et Cie. 279. Pritchett JW (1991) Growth plate activity in the upper extremity. Clin Orthop Rel Res 268:235-242. 280. Pritchett JW (1992) Longitudinal growth and growth-plate activity in the lower extremity. Clin Orthop Rel Res 275: 274-279. 281. Prockop WJ, Kivirikko KI (1995) Collagens: Molecular biology, diseases, and potentials for therapy. Ann Rev Biochem 64:403-434. 282. Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA (1979) The biosynthesis of collagen and its disorders. New Engl J Med 301:13-23. 283. Prockop DJ, Kivirikko, KI, Tuderman L, Guzman NA (1979) The biosynthesis of collagen and its disorders. New Engl J Med 301:77-85. 284. Prockop DJ (1992) Mutations in collagen genes as a cause of connective-tissue diseases. New Engl J Med 326:540-546. 285. Ranvier LA (1889) Traite Technique d'Histologie, 2nd edition. Paris: Savy.

126

CHAPTER I ~

Developmental Bone Bioloyy

286. Reddi AH, Cunningham NS (1993) Initiation and promotion of bone differentiation by bone morphogenetic proteins. J Bone Miner Res 8:499-502. 287. Reddi AH (1994) Bone and cartilage differentiation. Curr Op Genet Dev 4:737-744. 288. Redler I, Mow VC, Zimny ML, Mansell J (1975) The ultrastructure and biomechanical significance of the tidemark of articular cartilage. Clin Orthop Rel Res 112:357-362. 289. Renaut J (1893) Traite d'Histologie Pratique, volume 1, Paris: Rueff et Cie. 290. Retterer E (1886) Sur le mode de developpement des cavites articulaires chez les mammiferes. CR Hebd Soc Biol Ser 8: 45-48. 291. Retterer E (1894) Sur le mode de formation des articulations. CR Hebd Soc Biol 1 Ser 10:862-865. 292. Retterer E (1900) Evolution du cartilage transitoire. J de 1' Anat Physiol 36:467-565. 293. Retterer E (1902) Ebauche squelettogene des membres et developpement des articulations. J de 1' Anat Physiol 38: 473-509, 580-623. 294. Rey C, Beshah K. Griffin R, Glimcher MJ (1991) Structural studies of the mineral phase of calcifying cartilage. J Bone Miner Res 6:515-525. 295. Riddle RD, Johnson RL, Laufer E, Tabin C (1993) Sonic hedgehog mediates the polarizing activity of the ZPA. Cell 75:1401-1416. 296. Roach HI (1994) Why does bone matrix contain noncollagenous proteins? The possible roles of osteocalcin, osteonectin, osteopontin, and bone sialoprotein in bone mineralization and resorption. Cell Biol Internat 18:617-628. 297. Roach HI, Erenpreisa J, Aigner T (1995) Osteogenic differentiation of hypertrophic chondrocytes involves asymmetric cell divisions and apoptosis. J Cell Biol 131: 483-494. 298. Roark EF, Greer K (1994) Transforming growth factor-b and bone morphogenetic protein-2 act by distinct mechanisms to promote chick limb cartilage differentiation in vitro. Dev Dynam 200:103-116. 299. Roberts DJ, Tabin C (1994) The genetics of human limb development. Am J Hum Genet 55:1-6. 300. Rosenberg L (1971) Chemical basis for the histological use of safranin O in the study of articular cartilage. J Bone Joint Surg 53A:69-82. 301. Rosenberg L (1989) Structure and function of proteoglycans. In: Arthritis and Allied Conditions. A Textbook of Rheumatology, 1l th edition, ed DJ McCarty. pp. 240-255, Philadelphia: Lea and Febiger. 302. Roughley PJ, Eunicer RL, Lee ER (1994) Cartilage proteoglycans: Structure and potential functions. Microscopy Res Tech 28:385-397. 303. Russell ES (1982) Form and Function. A Contribution to the History of Animal Morphology. Chicago: University of Chicago Press. 304. Ryoppy S, Karaharju EO (1974) Alteration of epiphyseal growth by an experimentally produced angular deformity. Acta Orthop Scand 45:490-498. 305. St-Jacques B, Hammerschmidt M, McMahon AP (1999) Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev 13:2072-2086.

306. Sandell LJ (1994) In situ expression of collagen and proteoglycan genes in notochord and during skeletal development and growth. Microscopy Res Tech 28:470-482. 307. Sandell LJ, Sugai JV, Trippel SB (1994) Expression of collagens I, II, X, and XI and aggrecan MRNAs by bovine growth plate chondrocytes in situ. J Orthop Res 12:1-14. 308. Sasaki T, Fukai N, Mann K, Gohring W, Olsen BR, Rimpl R (1998) Structure, function, and tissue forms of the C-terminal globular domain of collagen XVIII containing of the angiogenesis inhibitor endostatin. EMBO J 17:4249-4256. 309. Schafer EA (1893) General anatomy or histology. In: Quain's Elements of Anatomy, Volume I, part II, ed EA Schafer, GD Thane. London: Longmans, Green. 310. Schaffer J (1911) Trajektorielle strukturen im knorpel. Verhandlungen Anat Ges 25:162-168. 311. Schenk RK, Hunziker EB (1991) Growth plate: Histophysiology, cell and matrix turnover. In Rickets, Nestle Nutrition Workshop Series, volume 21, ed FH Glorieux. pp 63-76, New York: Raven Press. 312. Schmid TM, Linsenmayer TF (1985) Developmental acquisition of type X collagen in the embryonic chick tibiotarsus. Dev Biol 107:373-381. 313. Schmid TM, Cole AA, Chen O, Bonen DK, Luchene L, Linsenmayer TF (1994) Assembly of type X collagen by hypertrophic chondrocytes. In: Extracellular Matrix Assembly and Structure, eds PD Yurchenco, DE Birk, RP Mechan. pp. 171-206, San Diego: Academic Press. 314. Schoney L (1875) Uber den ossifications process be: vogeln und die neubildung von rothen blutkorperchen an die ossificationsgrenze. Arch f Mikroskop Anat 12:242-253. 315. Schulin K (1879) Ueber die entwickelung und weitere ausbildung der gelenke des menschlichen korpers. Archiv Anat Entwick 240-274. 316. Schwabe JWR, Rodriguez-Esteran C, Belmonte JCI (1998) Limbs are moving: Where are they going? Trends Genet 14: 229-235. 317. Scott MP (1992) Vertebrate homeobox gene nomenclature. Cell 71:551-553. 318. Selby S (1961) Separate centers of ossification of the tip of the internal malleolus. Am J Roent 86:496-501. 319. Selleck SB (2000) Proteoglycans and pattern formation. Sugar biochemistry meets developmental genetics. Trends Genet 16:206-212. 320. Shapiro F, Holtrop ME, Glimcher MJ (1977) Organization and cellular biology of the perichondrial ossification groove of Ranvier. J Bone Joint Surg 59A:703-723. 321. Shapiro F (1987) Epiphyseal disorders. N Eng J Med 317: 1702-1710. 322. Shapiro F (1992) Vertebral development in the chick embryo during days 3-19 of incubation. J Morph 213:317-333. 323. Shapiro F (1998) Epiphyseal and physeal cartilage vascularization: A light microscopic and autoradiographic study of cartilage canals in newborn and young postnatal rabbit bone. Anat Rec 252:140-148. 324. Shapiro F, Rivas R (1996) Structural stages in epiphyseal morphogenesis. J Bone Miner Res 11 (Supp 1):$300. 325. Shapiro E (1997) Variable conformation of GAP junctions linking bone cells: A transmission electron microscopic study of linear, stacked linear, curvilinear, oval, and annular junctions. Calcif Tissue Int 61:285-293.

References 326. Shapses SA, Sandell LJ, Ratcliffe A (1994) Differential rates of aggrecan synthesis and breakdown in different zones of the bovine growth plate. Matrix Biol 14:77-86. 327. Siffert RS, Katz JF (1983) Growth recovery zones. J Pediatr Orthop 3:196-201. 328. Skawina A, Litwin JA, Gorczyca J, Miodonski AJ (1994) Blood vessels in epiphyseal cartilage of human fetal femoral bone: A scanning electron microscopic study of corrosion casts. Anat Embryol 189:457-462. 329. Smith JW (1962) The relationship of epiphyseal plates to stress in some bones of the lower limb. J Anat 96:58-80. 330. Speer DP, Dahners L (1979) The collagenous architecture of articular cartilage. Clin Orthop Rel Res 139:267-275. 331. Speer DP (1982) Collagenous architecture of the growth plate and perichondrial ossification groove. J Bone Joint Surg 64A: 399-407. 332. Stieda L (1872) Die Bildung des Knochengewebes. Leipzig: Engelmann. 333. Stockwell RA (1971) The ultrastructure of cartilage canals and the surrounding cartilage in the sheep fetus. J Anat 109: 397-410. 334. Strayer LM (1943) The embryology of the human hip joint. Yale J Biol Med 16:13-26. 335. Strelzoff ZJ (1873) Ueber die histogenese der knochen. Cent f Med Wissen 274; Untersuchungen aus dem pathologischen institut zu Zurich. Herausgeg v. Eberth 1873. 336. Strobino LJ, French GO, Colonna PC (1952) The effect of increasing tensions on the growth of epiphyseal bone. Surg Gynec Obstet 95:694-700. 337. Stump CW (1925) The histogenesis of bone. J Anat 49: 136-154. 338. Tabin C (1995) The initiation of the limb bud: Growth factors, hox genes, and retinoids. Cell 80:671-674. 339. Taussig G, Delor MH, Masse P (1976) Les alterations de croissance de l'extremitie superieure du femur. Rev Chir Orthop 62:191-210. 340. Teitelbaum SL (2000) Bone resorption by osteoclasts. Science 289:1504-1508. 341. Testut L (1899) Traite d' Anatomie Humaine, volume 1. Paris: Octave Dion. 342. Thaller C, Eichele G (1987) Identification and spatial distribution of retinoids in the development chick limb bud. Nature 327:625-628. 343. Thane GD (1893) Osteology, volume II, part I of Quain's Elements of Anatomy, eds. EA Schafer, GD Thane. London: Longmans, Green. 344. Thomas JT, Ayad S, Grant ME (1994) Cartilage collagens: Strategies for the study of their organisation and expression in the extracellular matrix. Ann Rheum Dis 53:488-496. 345. Thomson MB (1902) The relation of structure and function as illustrated by the form of the lower epiphysial suture of the femur. J Anat Physiol 36:95-105. 346. Thorp BH, Anderson I, Jakowlew SB (1992) Transforming growth factor-[31, -132, and -133 in cartilage and bone cells during endochondral ossification in the chick. Development 114:907-911. 347. Tickle C, Eichele G (1994) Vertebrate limb development. Ann Rev Cell Biol 10:121-52. 348. Tomes J, De Morgan C (1853) Observations on the structure and development of bone. Phil Trans Roy Soc Lond 143:109-139.

127

349. Trueta J, Morgan JD (1960) The vascular contribution to osteogenesis. I. Studies by the injection method. J Bone Joint Surg 42B:97-109. 350. Trueta J, Little K (1960) The vascular contribution to osteogenesis. II. Studies with the electron microscope. J Bone Joint Surg 42B:367-376. 351. Trueta J, Amato VP (1960) The vascular contribution to osteogenesis. III. Changes in the growth cartilage caused by experimentally induced ischemia. J Bone Joint Surg 42B: 571-587. 352. Tschantz P, Taillard W, Ditesheim PJ (1977) Epiphyseal tilt produced by experimental overload. Clin Orthop Rel Res 123:271-279. 353. Tschistowitsch T (1897) Zur frage von der angebomen rachitis. Virch Arch f Path Anat u Phys u f Klin Med 148:140-177 and 209-233. 354. Uranossow G. Beitrage zur lehre vonder erwicklung des knochengewebes aus knorpel (quoted in E. Retterer, J Anat Physiol 36:467-565, 1900). 355. Van der Rest M, Garrone R (1991) Collagen family of proteins. FASEB J 5:2814-2823. 356. Virchow R (1860) Cellular Pathology, translated by F. Chance. London: John Churchill. 357. Vogel KG (1994) Glycosaminoglycans and proteoglycans. In: Extracellular Matrix Assembly and Structure, eds PD Yurchenco, DE Birk, RP Mecham. pp 243-279, San Diego: Academic Press. 358. Volkmann R (1862) Chirurgische erfahrungen uber knochenverbiegungen und knochenwachsthum. Arch f Path Anat Physiol u Klin Med 24:512-541. 359. Vortkamp A, Lee K, Lanske B, Segre GV, Kronenberg HM, Tabin CJ (1996) Regulation of rate of cartilage differentiation by Indian hedgehog and PTH-related protein. Science 273: 613-622. 360. Vortkamp A, Pathi S, Peretti GM, Caruso EM, Zaleske DJ, Tabin CJ (1998) Recapitulation of signals regulating embryonic bone formation during postnatal growth and in fracture repair. Mech Dev 71:65-76. 361. Vukicevic S, Paralkar VM, Cunningham NS, Gurkind JS, Reddi AH (1990) Autoradiographic localization of osteogenin binding sites in cartilage and bone during rat embryonic development. Dev Bio1209-214. 362. Waldeyer (1865) Ueber den ossifikationsprocess. Arch Mikroskop Anat 1:354-375. 363. Wang YC (1975) Intrachondral microcirculation and cartilage growth. Cartilage and bone development during the embryonic stage. Chin Med J 1:449-460. 364. Warwick WT, Wiles P (1934) The growth of periosteum in long bones. Br J Surg 22:169-174. 365. Watermann R (1964) Gefassdurchtrittsoffnungen und arthrose am oberschenkel. Z Orthrop 98:492-508. 366. Watermann R (1960) Gefassbedingle erhebungen und einsenkungen an den schichtgrenzen der kniegelenksnahen wachstumzonen. Z Orthop 93:565-579. 367. Watermann R (1966) Gefassversorgung der distalen femurepiphyse. Z Orthop 101:247-257. 368. Watt SL, Lunstrum GP, McDonough AM, Keene DR, Burgeson RE, Morris NP (1992) Characterization of collagen types XII and XIV from fetal bovine cartilage. J Biol Chem 267: 10093-20099.

128

CHAPTER 1 ~

Developmental Bone Biology

369. White JW, Stubbins SG, Jr (1944) Growth arrest for equalizing leg lengths. J Am Med Assoc 126:1146-1148. 370. Wilsman NJ, Van Sickle DC (1970) The relationship of cartilage canals to the initial osteogenesis of secondary centers of ossification. Anat Rec 168:381-392. 371. Wilsman NJ, Van Sickle DC (1972) Cartilage canals, their morphology and distribution. Anat Rec 173:79-94. 372. Woessner JF, Jr (1994) The family of matrix metalloproteinases. Ann NY Acad Sci 732:11-20. 373. Wolff CF (1759) Theoria Generations. Halle. 374. Wolff J (1892) Das Gesetz der Transformation der Knochen. Berlin: Hirschwald. 375. Wolpert L (1996) One hundred years of positional information. Trends Genet 12:359-362.

376. Wong M, Carter DR (1990) A theoretical model of endochondral ossification and bone architectural construction in long bone ontogeny. Anat Embryol 181:523-532. 377. Wozney JM, Rosen V, Celeste AJ, Mitsock LM, Whitters MJ, Kriz RW, Hewick RM, Wang EA (1988) Novel regulators of bone formation: Molecular clones and activities. Science 242: 1528-1534. 378. Wu J-J, Woods PE, Eyre DR (1992) Identification of cross linking sites in bovine cartilage types IX collagen reveals an antiparallel type II-type IX molecular relationship and type IX to type IX bonding. J Biol Chem 267:23007-23014. 379. Yang Y, Niswander L (1995) Interaction between the signaling molecules Wnt-7a and Shh during vertebrate limb development: Dorsal signals regulate anteroposterior patterning. Cell 80:939-947.

CHAPTER

2

lmagin9 Approaches for Epiphyseal Assessment* I. Introduction Technical Considerations III. Imaging Characteristics IV. Normal and Abnormal Growth and Ossification

V. Imaging of Abnormalities by Location: Disorders of the Lower Extremity

II.

VI.

I. I N T R O D U C T I O N

Imaging of Disorders by Location: Disorders of the Upper Extremity

become the primary imaging modality to evaluate developmental hip dysplasia in infants under 6 months of age. Sonography is based on the formation of an image from a reflected ultrasonic pulse. The sound waves are reflected at any sharp interface between two tissues. Sound travels rapidly through water without forming echoes; tissues with abundant water, such as cartilage, are anechoic or sonolucent. The bony cortex is a powerful reflector and is well-seen sonographically as an echogenic line. Unfortunately, because the waves are reflected, it is impossible to image the structures deep to the cortex such as the marrow. High-frequency transducers provide high spatial resolution but poor penetration. Although the cartilaginous epiphysis is hypoechoic, it contains numerous echogenic tubular epiphyseal vascular canals (76) (Figs. 1 and 2). These canals initially are parallel, but later radiate to the ossification center (3). Spectral, color, and power Doppler sonography demonstrate flow within the vessels of the canals (5) (Fig. 3). In the hip, cartilage of echogenicity similar to that of the epiphysis is detectable in the acetabulum. Sonographic detection of an increase in the width of the acetabular cartilage is an early sign of developmental dysplasia of the hip (71). The ossification center, initially a small area of increased echogenicity in the epiphysis, becomes evident weeks before it is visible radiographically (28). The physeal-metaphyseal junction is seen as a pallisade of echogenic small linear structures, almost converging to form an echogenic band (Fig. 1). Fibrocartilage, as in the acetabular and glenoid labra, is uniformly hyperechoic. Similar sonographic anatomy can be seen in other epiphyses, such as the shoulder, elbow, and knee (Fig. 2).

The study of normal and abnormal skeletal development was greatly enhanced by the discovery and clinical use of radiography. Plain radiographs were useful for determination of early development of the secondary ossification centers in infancy and early childhood and for determination of physeal closure at the time of skeletal maturity. The interpretation of images was based on observing the bone and inferring the cartilage that produced the bone. The advent of magnetic resonance imaging (MRI) and sonography has allowed direct evaluation of the cartilaginous, vascular, and fibrous structures that are involved in the formation and growth of bones. Scintigraphy has the capability of allowing a functional assessment of the growing skeleton. The purpose of this chapter is to review newer imaging modalities that continue to transform the perception and interpretation of epiphyseal development and growth. There has been an explosive increase in the capabilities of the imaging modalities. The understanding of the images has also increased dramatically, although, unfortunately, not at the same pace. Each of the newer techniques will be described in the subsequent sections, followed by a discussion of the application of these modalities in the primary anatomic regions of the skeleton. The use of these techniques will also be referenced in each of the subsequent chapters in relation to the specific deformities described. II. T E C H N I C A L C O N S I D E R A T I O N S

A. Ultrasonography 2. PRENATAL SONOGRAPHIC EVALUATION 1. POSTNATAL EVALUATION OF THE SKELETON

OF THE SKELETON

Sonography is very useful for imaging epiphyses, particularly for the developing femoral head (Figs. 1 and 2). It has

Prenatal skeletal sonography focuses on assessments of normal and abnormal growth. The length of the echogenic femoral shaft is compared to gestational age, head circumference, and abdominal circumference (64). Marked deviation

*This chapter was written by Dr. Diego Jaramillo.

129

130

CHAPTER 2

~ Ima~ting Approaches for Epiphyseal Assessment

F I G U R E 1 Normal ultrasound of the hip in a 6-week-old boy. (A) Coronal image of the hip shows the femoral head (f) as a relatively echolucent, speckled area. The hyaline cartilage of the acetabulum (straight arrow) is of similar echotexture. The fibrocartilaginous labrum and capsule are echogenic (curved arrow). (B) Transverse image of the hip. The sonolucent femoral head shows linear structures compatible with epiphyseal vessels (arrow). The femoral metaphysis (m) and ischial tuberosity (i) are well-seen.

from the normal femoral growth, particularly if the other measurements are not significantly abnormal, suggests a skeletal dysplasia. Features that add specificity to the diagnosis include the curved femurs in thanatophoric dysplasia, the lack of calvarial ossification in osteogenesis imperfecta and hypophosphatasia, the small thorax in asphyxiating thoracic dystrophy, and the absence of spinal ossification in achondrogenesis (1, 52). Prenatal sonography can detect epiphyses and ossification centers. The epiphyseal ossification center of the distal femur becomes apparent sonographically at 28-35 weeks (Fig. 4), followed 2-3 weeks later by the proximal tibial center. Detection of proximal humeral ossification, beginning at 38 weeks, is important because it usually is associated with pulmonary maturity (52). Although sonography has been the only important modality for prenatal evaluation of the skeleton, faster MR imaging sequences have allowed the visualization of epiphyseal development in utero (50).

F I G U R E 2 Normal ultrasound of the knee in a newborn. This patient had contralateral knee dislocation. The sagittal sonographic image shows normal epiphyseal and secondary ossification center (SOC) development in the distal femur (arrow) and tibia. The patella and the patellar tendon (at right) are well-demonstrated.

B. Magnetic Resonance Imaging Magnetic resonance imaging (MRI) is based on the application of radio waves to protons that have been previously aligned by a magnetic field. The radio signal is applied at the

SECTION I! ~ Technical C o n s i d e r a t i o n s

131

FIGURE 3 Normal vascularity of the proximal femoral head in a 3-month-old girl. The spectral, color, and power Doppler ultrasonography technique provides images in the coronal plane, which show that there is flow within numerousepiphysealvessels.The vessels have a parallel configuration.

frequency of precession of the protons. The protons resonate with the radio pulse and acquire energy, which shifts their orientation away from the direction of the field. Once the radio frequency pulse is discontinued, the protons relax back into alignment with the field and emit a radio wave signal that is interpreted by a computer to generate an image. Clinical imaging is based on the evaluation of signals from hydrogen protons. These can be linked to oxygen, as in water molecules, or to carbon, as in fat molecules. MR imaging evaluates multiple tissue parameters, but the primary ones are T1 and T2 relaxation. T1 relaxation reflects the increase in signal due to the return of the excited protons to a state of alignment with the magnetic field. T2 relaxation reflects the loss of signal due to dephasing of the protons with respect to each other once the exciting pulse is discontinued. These two processes occur at the same time, but im-

ages can be optimized to show contrast due to differences in T1 relaxation or T2 relaxation. These are T1- and T2weighted images, respectively. Water has a slow T1 and T2 relaxation. Its excited protons are slow to realign themselves with the main magnetic field. The slow increase in magnetization of water results in low signal intensity on Tl-weighted images. The loss of signal due to dephasing is also slow, and therefore the signal intensity remains high on T2-weighted images. Fat has a relatively rapid T1 and T2 relaxation. Therefore, the fast signal increase of fat results in high signal intensity on Tl-weighted images; the relatively fast signal loss results in lower signal intensity on T2-weighted images. The signal characteristics of cartilage are complex because of the binding of water to collagen and proteoglycans; this will be discussed later (33). Bone and fibrous tissues have very few mobile protons and thus have

132

CHAPTER 2 ~

Imaging Approaches for Epiphyseal Assessment C. Computed Tomography (CT)

FIGURE 4 Normal distal femoral epiphyseal development in a 36week-old fetus. Coronal sonographic image of the distal femoral epiphysis shows a well-developedossification center (arrow). The surrounding cartilaginous epiphysis is well-definedsonographically.

very low signal intensity on all sequences. Fibrocartilage has a high component of fibrous tissue and is also hypointense on all sequences. Gadolinium, the contrast material most frequently used in MR imaging, works by shortening T1 relaxation, that is, by making structures more intense on T 1-weighted images. In the immature skeleton, the administration of gadolinium resuits in visualization of numerous vascular canals in the epiphyseal cartilage (3). Additionally, the physis and the miniplate demonstrate enhancement, probably by diffusion. The hematopoietic marrow enhances more than the fatty marrow, as hematopoietic marrow is more richly vascularized (18). MR is useful to detect the structural integrity of cartilaginous structures, tendons, and ligaments. It can detect subtle marrow abnormalities such as infection, tumor, and trauma. MR allows visualization of the contour of unossified skeleton. This can be of great use in the evaluation of complex limb deformities such as hemimelias, proximal focal femoral deficiency, and complicated cases of developmental dysplasia of the hip. It can also differentiate among the various structures of the cartilage such as epiphyseal and physeal cartilage, and the zone of provisional calcification (ZPC) (Fig. 5). It can detect subtle physeal abnormalities before they result in growth disturbances (37, 38, 42, 43). Finally, MR can also be of great utility in demonstrating epiphyseal perfusion in children treated for developmental dysplasia of the hip (44) and in children with Legg-Calv6-Perthes disease (16).

CT displays various musculoskeletal structures with pixel values according to their X-ray attenuation coefficients. In the musculoskeletal system, only four densities are clearly definable: bone, fat, water, and soft tissue. CT is particularly useful to demonstrate the relationship between bony structures. Two- and three-dimensional multiplanar reconstructions are easily obtained (12). Spiral (or helical) CT, which allows the acquisition of a continuous set of data, is a modality with increasing application. It allows very fast acquisitions; the data acquisition of most musculoskeletal examinations can be completed in less than 1 min. It also reduces the radiation dose necessary for examinations optimized for two- or a three-dimensional reconstruction (12). Because spiral CT acquires a continuous set of data, it is possible to modify the reconstruction intervals and the location of the slice center as needed (74). The main disadvantage of spiral CT is that it produces blurring, particularly along the z axis. In some instances, such as the detection of subtle fractures, spiral CT appears to be less sensitive than conventional CT (51). Pediatric musculoskeletal CT is useful in the evaluation of trauma, the imaging of articular structures, particularly in the hip, and the evaluation of growth disorders. CT can show the relationship of bony fragments in complex fractures, which is facilitated by the ease of performing reconstructions (Fig. 6). It is more sensitive than MR imaging for the detection of intra-articular bony fragments, but it cannot detect cartilaginous and ligamentous abnormalities. CT is particularly useful in demonstrating the anatomy of complex distal tibial fractures such as the triplane and Tillaux injuries (20). CT is very useful in the preoperative and postoperative evaluation of hip disorders. In infants who have been placed in spica casts, CT demonstrates recurrent dislocation and detects obstacles to reduction. This can now be done with minimal exposure to radiation, by decreasing the technique to 20-40 mA (19). In adolescents and young adults with developmental dysplasia of the hip, CT with three-dimensional reconstructions shows the configuration and orientation of the femoral head and the acetabular architecture. CT can be used in the evaluation of growth arrest (54) by scanning in either the direct sagittal or direct coronal plane and constructing an axial map of the physis. In general, however, MR imaging provides better information than CT about the status of the physeal cartilage (63).

D. Scintigraphy Skeletal scintigraphy is based on the administration of one of various bone seeking radiopharmaceuticals. A technetium-99m (99mTc) labeled phosphate such as [99mTc]-methylene diphosphonate is administered intravenously, and the patient is imaged using a gamma camera. Imaging can be performed at the time of the injection (radionuclide angiography), im-

SECTION III 9 Imaging CharacterisOcs

133

FIGURE 5 Distalfemoral ossification irregularity in a 2-year-oldboy. (A) Coronal gradient-recalledecho image shows the contrast between hypointense (dark) bone and hyperintense (bright) cartilage. There is fragmentation of the medial aspect of the distal femoral ossification center (arrow). Notice the normal appearing proximaltibial ossificationcenter. (B) CoronalT2-weightedimage showsbetter definition between the zones of cartilage. The marrow of each one of the parts of the distal femur is normal. However, the overlying cartilage is inhomogeneous. Although this is presumed to be a normal variant, the signal characteristics suggest stress on the cartilage.

mediately after the injection (early imaging), at 2 - 4 hr after the radiotracer administration (72), and at 24 hr after the injection. The distribution of tracer in bone relates to blood flow, bone turnover, and growth. In the growing skeleton, the radiotracer is localized primarily in the primary spongiosa of the metaphysis (26) rather than in the physeal cartilage. The unossified epiphysis shows no radiotracer activity, but epiphyseal perfusion can be assessed after the development of epiphyseal ossification. Skeletal scintigraphy can aid in the diagnosis of epiphyseal ischemia (11) and in the assessment of physeal activity (26). In Legg-Calv6-Perthes disease, Conway has demonstrated that scintigraphy can be used at presentation to detect the lack of epiphyseal perfusion and during recovery to evaluate the pattern of revascularization. A femoral head that revascularizes adequately will show increasing activity in the area of the lateral column of the femoral head. A hip with poor revascularization will show extension of activity through the growth plate into the base of the epiphysis by new vessels coming from the metaphysis.

In children with various growth disturbances, scintigraphy can show focal decrease (as in bony bridges or after epiphysiodesis) or increase (as in Blount disease) in the activity of the physis (26). III. I M A G I N G C H A R A C T E R I S T I C S

A. Cartilage Hyaline cartilage and fibrocartilage have different appearances on sonography and MR imaging (33). Hyaline cartilage is sonolucent, and on MR imaging it behaves like water, with peculiarities in signal characteristics due to differences in the binding of water to macromolecules of the matrix. Fibrocartilage is markedly echogenic, and on MR images it is hypointense on all sequences, presumably because the majority of its protons are nonmobile. The imaging characteristics of epiphyseal cartilage, physeal cartilage, the zone of provisional calcification, the physis of the secondary center of ossification, the perichondrium, and articular cartilage have been presented in two recent reports (35, 36).

134

CHAPTER 2 ~ Imagin9 Approaches for Epiphyseal Assessment (the "magic angle") collagen fibers have the highest signal intensity. Cartilage is susceptible to loss of signal intensity by the "magnetization transfer effect" (75). If a radio frequency pulse stimulates bound protons, the effect can be transferred to free protons in the cartilage. The net effect is a decrease in the signal intensity of the cartilage, which appears to depend on its collagen concentration. All fast spin echo sequences normally result in decreased signal of the epiphyseal and physeal cartilage because they the exhibit magnetization transfer effect. Special magnetization transfer sequences can be used to decrease the signal intensity of the cartilage. These sequences have been used to facilitate the differentiation between hyaline cartilage and joint fluid or synovium (60). 2. PERICHONDRIUM

F I G U R E 6 Sixteen-year-old girl who suffered trauma to the left distal femur. Oblique three-dimensional reconstruction with surface rendering shows near-anatomic alignment of the fragments with only mild separation of the fragments at the articular surface. The Salter-Harris IV configuration of the fracture is well-seen.

1. HYALINE CARTILAGE

Hyaline cartilage transmits sound waves fairly uniformly. The echogenicity of cartilaginous structures such as articular cartilage (33) and epiphyseal cartilage (25) is low. On MR images, hyaline cartilage has a high water content and therefore tends to have a low signal intensity on Tl-weighted images and a high signal intensity on T2-weighted images. Water in cartilage, however, is bound to large proteoglycan and collagen molecules, which accelerate the relaxation of the protons of cartilage. Because of this, the signal intensity of cartilage is higher than that of water on Tl-weighted images and a lower than that of water on T2-weighted images. The signal characteristics of various zones of cartilage differ, presumably because of variations in their histochemical characteristics. For example, on T 1-weighted images, the physeal cartilage is of slightly higher signal intensity than the epiphyseal cartilage. On T2-weighted images, physeal and epiphyseal cartilage can be clearly differentiated; the signal intensity of the physeal cartilage is much greater than that of the epiphyseal cartilage (Fig. 7). The orientation of the collagen fibers also affects the signal characteristics. Tissue anisotropy decreases signal intensity of the cartilage when the collagen fibers are not distributed uniformly, such as in the reticular layer of the articular cartilage (29). There is variation in signal intensity depending on the angle between the collagen fibers and the main magnetic field. At 55 ~ with respect to the magnetic field

Understanding the perichondrial structures is important for the interpretation of plain radiographs and studies using the more advanced and sensitive imaging techniques (35). The perichondrial bony ring of Lacroix surrounds the farthest extent of the metaphysis, including part of the physis, and plays a major role in physeal-metaphyseal support and remodeling. The perichondrial bone collar, sometimes referred to as the collar of Laval-Jeantet (57) in the radiologic literature, can appear as a linear mass of bone in the juxtaphyseal and upper metaphyseal areas. This linear structure measures 1-3 mm in infancy and becomes less prominent thereafter. It is most apparent in the distal radius and should not be confused with a fracture. Ossification of the bony ring of Lacroix can also appear as small spicules in the physeal periphery and, again, should not be considered abnormal, particularly in the context of the investigation of child abuse. The perichondrial collar can be lost during rickets (58). The perichondrial structures appear to explain many imaging features of the metaphyseal fractures demonstrated in battered children (47). The plane of the metaphyseal fracture separation approaches the physis in the central part of the bone. Peripherally, the perichondrial structures provide support to the most immediate juxtametaphyseal metaphysis and divert the fracture shaftward toward the metaphysis. This results in the typical comer and bucket handle appearances of this injury (46). The perichondrium is firmly attached to the epiphyseal cartilage just beyond the physis, whereas the periosteum in childhood is only loosely attached to the bone. Younger children are particularly prone to subperiosteal extension of hematoma, tumor, or infection due to the laxity of the periosteal attachment. The perichondrium acts as a tethering point to the spread of subperiosteal disease, creating a typical angular appearance of sub-periosteal collections in the metaphyseal region. The vertex of the angle is the perichondrium, and the sides are the bone and the elevated periosteum, which is readily detectable on MR images as a line of decreased signal intensity.

SECTION IV ~ Normal and Abnormal G r o w t h and Ossification

135

FIGURE 7 Normal appearance of the epiphysis and changes in the epiphyseal cartilage probably related to the beginning of ambulation. (A) Sagittal T2-weighted image in a 13-month-old boy shows that the signal of the distal femoral epiphyseal cartilage (closed arrows) is relatively well-defined. The epiphyseal cartilage is nearly isointense with the ossification center. The physis is of higher signal intensity than the epiphyseal cartilage, and the meniscus (open arrow) is of very low signal intensity due to its fibrocartilaginous composition. (B) Sagittal T2-weighted image in a 25-month-old boy shows that there has been interval development of decreased signal intensity along the weight bearing portion of the bone (arrow).

B. Bone All bone is highly echogenic on sonograms because bone does not transmit sound waves. Bone has a high attenuation value on CT; the attenuation value correlates directly with the electron content and, thus, the calcium content of a pixel. Bone has very few mobile protons and therefore is hypointense on all MR images.

C. Bone Marrow The distribution of bone marrow is perhaps the most important determinant of the signal intensity of the osseous structures on MR images. Hematopoietic marrow contains 40% water and 40% fat; the content of fat increases with age (2). On T 1-weighted images, hematopoietic marrow is of lower signal intensity than muscle in the neonate, but of slightly greater signal intensity than muscle during childhood. Fatty marrow contains 80% fat and has a signal similar to that of the fat in subcutaneous tissues.

The marrow in the neonate is almost entirely hematopoietic. The transformation to fatty marrow occurs in a predictable pattern (14, 48, 53, 62, 67, 73, 78), beginning in the periphery and progressing toward the axial skeleton. In each bone, the transformation begins in the diaphysis and progresses toward the metaphyses. Most epiphyses on T1weighted images appear fatty beyond 6 months after the radiographic appearance of the ossification center (41). In the proximal humerus and proximal femur, the metaphyses demonstrate fatty changes only after 15 years of age and frequently remain hematopoietic into adulthood.

IV. NORMAL AND ABNORMAL GROWTH AND OSSIFICATION Several specialized zones of cartilage at the ends of growing bones are differentiated by MR imaging (35). On most pulse sequences, the physeal cartilage is a band of high signal intensity clearly separated from the hypointense zone of

136

CHAPTER 2 ~ lrnagin9 Approaches for Epiphyseal Assessment

FIGURE 8 Physealand epiphyseal abnormalityin a 5-year-oldgirl with Blountdisease. (A) Coronalgradient-recalledimage of the knees shows marked irregularity and widening of the medial physes of both proximal tibiae. The abnormality is significantly more severe on the left. However,there is no evidence of bridging. (B) Maximal intensity projectionof thin section spoiled gradient-recalled image provides an axial map of the physes (arrows). Note the irregularity due to the mamillaryprocesses. There is no evidence of a bridge, which would be seen as an area of decreased signal intensity. [Reprintedfrom (35), with permission.]

provisional calcification (ZPC). In infants and young children, the hyperintense physis can be differentiated from the less intense cartilaginous epiphysis on T2-weighted images. In older children, the high-signal-intensity physis lies between two hypointense platelike structures: the ZPC and the epiphyseal bone plate. The physis of the secondary ossification center surrounding the secondary center of ossification has signal characteristics similar to those of the main physis. The articular cartilage is also of similar signal, but on high-resolution imaging it has a complex multi-layered configuration. The physis of the secondary center as well as the cartilaginous vascular canals enhance markedly after the administration of gadolinium (3). Any insult to the growth plate can result in abnormalities of the physis, which usually are most conspicuous on MR images (63). In acute injuries, the signal from the abnormal physis can be focally or uniformly increased. Physeal fractures can disrupt the cartilage along the longitudinal axis of the bone, allowing communication between epiphyseal and metaphyseal vessels and ultimately the formation of a bony bridge. We have demonstrated the sequence of vascular invasion of the cartilage, followed by bone deposition and formation of a bony bridge using gadolinium-enhanced MR imaging in a rabbit model (43). In some physeal fracture separations, MR imaging can demonstrate a plane of separation between the layers of the cartilage. The plane of separation can involve the epiphyseal side of the physis, but usually affects the metaphyseal side of the physis, that is, the hypertrophic zone (39). Chronic physeal trauma can also result in physeal abnormality detectable by MR imaging. MR imaging of the wrist in gymnasts, for example, shows areas of widening of the distal radial physis, with cartilage extending into the metaphysis (68). Bony bridging can also occur in association with

the physeal widening. Similar abnormalities have been demonstrated in children with other overuse injuries, in children with chronic physeal trauma due to myelodysplasia, and in tibia vara (Blount) disease (Fig. 8) (17). Disruption of the physeal cartilage due to nontraumatic insults is also seen with MR imaging. Although the physis is a partial barrier to the spread of disease, metaphyseal osteomyelitis often involves the physis (Fig. 9). MR imaging demonstrates that the signal from the cartilage is interrupted in the region of infection (38). Gadolinium enhancement can show nonperfused fluid collections, which may require drainage. Ischemia, as occurs with Legg-Perthes disease (Fig. 10), or meningococcemia frequently result in bony bridging detectable by MR imaging or CT (Fig. 11). Benign tumors such as enchondromas, osteochondromas, and bone cysts can disrupt the physis and result in bridging (38). Malignant tumors often trespass the physeal barrier; MR imaging demonstrates that metaphyseal osteogenic sarcoma extends into the epiphysis in 80% of cases (55). Once a bridge develops across the physis, MR imaging can help in the demonstration of the size and location of the bony bridge and in the evaluation of the resulting growth arrest. We typically image bony bridges in the coronal and sagittal planes using a combination of gradient-recalled echo sequences and fat-suppressed proton density sequences (Fig. 12). These two sequences maximize contrast between physeal cartilage (which is of very high signal intensity) and bone (which is of very low signal intensity). Additionally, we create an axial map of the physis by generating a maximal intensity projection image from a three-dimensional fatsuppressed gradient echo sequence obtained in the axial plane (7) (Fig. 13). This map demonstrates the location of the bony bridge. A ratio between the area of the bridge and the area of the physis can be calculated in order to decide on the resectability of the bridge.

SECTION IV ~ Normal and Abnormal G r o w t h and Ossification

137

F I G U R E 9 Osteomyelitis with physeal involvement in a 9-year-old girl. (A) Frontal radiograph shows a lytic lesion in the proximal metaphysis of the tibia (arrow). It is uncertain whether it involves the physis. (B) Coronal post-gadolinium, fat suppressed, Tl-weighted image shows that the focus of osteomyelitis, which is centrally necrotic, is seated on the physis. There is inflammatory reaction involving both the epiphysis and the metaphysis. (C) Sagittal inversion recovery image shows edema of the entire marrow. There is interruption of the anterior aspect of the physis.

Analysis of the P a r k s - H a r r i s or growth recovery lines shows the growth disturbance and confirms the location of the physeal abnormality. A P a r k s - H a r r i s line is a disk of transverse bony trabeculae that develop whenever there is a temporary slowdown of growth. It occurs after most diseases and is almost always seen following immobilization for a fracture. It is useful to think of the P a r k s - H a r r i s line as an

indicator of the location of the physis at the time of the insult. In the absence of growth arrest, a Parks-Harris line is parallel to the physis, as the entire physis migrates away from the line at a uniform speed. A bony bridge will tether the physis, so that it cannot migrate away from the Parks-Harris line (56). The farther away normal physeal cartilage is from the bridge, the more normal the growth and the greater the

138

CHAPTER 2 ~

lma~tin~tApproaches ]or Epiphyseal Assessment

F I G U R E 10 Premature growth plate closure in a 10-year-old girl with Legg-Calvf-Perthes disease. (A) Frontal radiograph of the right hips shows evidence of flattening of the femoral head. The physis is wavy and slightly less well-seen than on the contralateral side, but definitely open. (B) Coronal fast spin echo T2-weighted image shows the discrepancy between the articular surface of the femur and the acetabulum with medial pooling of synovial fluid (arrow). (C) Coronal gradient echo image shows that the physis on the right side is open but markedly wavy. (D) Frontal radiograph obtained 3 years later shows that the right femoral physis has closed prematurely. Relatively greater trochanteric overgrowth is seen. The left remains open.

SECTION IV ~ Normal and Abnormal G r o w t h and Ossification

139

F I G U R E 11 Severe growth disturbance of the distal right tibial epiphysis secondary to meningococcemia. This is a 6-year-old boy who had suffered purpura fulminans as an infant. Both distal tibial growth areas were affected. (A) Frontal radiograph of the right ankle shows marked shortening of the distal right tibia with a suggestion of central growth arrest. The epiphysis is small and irregular. Note the irregularity of the talar dome. (B) Coronal CT scan of both ankles confirms the presence of central bone ridge growth arrest of the fight tibia and extensive irregularities of the epiphyses of both tibias and the right talar dome. The left talus is markedly fragmented and was almost nonexistent. (C) Oblique angiographic projection of the distal leg and ankle region shows that the anterior tibial supply is completely obliterated (arrow). The circulation to the foot is provided almost exclusively by the posterior tibial artery.

F I G U R E 12 Posttraumatic bony bridge in a 6-year-old child. (A, B) Frontal and lateral radiographs of the femur show evidence of lateral femoral growth arrest. The physis is indistinct laterally and centrally. (C) Two coronal gradient-recalled echo images show evidence of interruption of the high signal from the physis in the lateral aspect of the anterior distal femur (arrow). (D) Two coronal Tl-weighted images show interruption of the dark line corresponding to the zone of provisional calcification. The bony bridge con-

F I G U R E 13 Posttraumatic bony bridge in a 12-year-old boy. (A) Frontal radiograph at the time of injury shows a Salter III fracture of the distal tibia involving the medial malleolus. There is displacement of the medial fragment. The distal fibula shows evidence of a Salter I injury. (B) Sagittal fat-suppressed proton density image done postfracture shows that the anterior aspect of the medial tibial physis is obliterated by a bony bridge (arrow). The posterior aspect of the physis is open. (C) A more normal portion of the physis is shown for comparison. The sagittal image in the midphysis shows continuity of the bright signal, corresponding to the growth plate cartilage (arrows). (D) Axial spoiled gradient-recalled maximal intensity projection map of the physis shows an area of decreased signal intensity in the anterior and medial aspect of the distal tibial physis corresponding to the bridge (arrow).

F I G U R E 12 (continued) tains high-signal-intensity fatty marrow indicative of epiphyseal and metaphysal marrow communication. (E) Sagittal fat-suppressed proton density images show that the location of the bridge is relatively central. There is tethering of the growth plate in the vicinity of the bridge. The bridge is of inhomogeneous signal intensity. (F) Postsurgical frontal radiograph shows that the bridge has been resected.

142

CHAPTER 2 ~ Irnagin9 Approaches for Epiphyseal Assessment

separation between the Parks-Harris line and the physis. On MR images, particularly on Tl-weighted images, the ParksHarris line is seen clearly as a low-signal-intensity structure, which contrasts sharply with the adjacent hyperintense marrow (37). With growth arrest, the Parks-Harris line will be slanted with respect to the physis, and the two will converge at the point of physeal tethering.

V. I M A G I N G O F A B N O R M A L I T I E S BY LOCATION: D I S O R D E R S OF T H E LOWER EXTREMITY

A. Disorders of the Hip 1. PROXIMAL FOCAL FEMORAL DEFICIENCY (PFFD) In PFFD, imaging is performed to determine the extent of development of the proximal femur (32). The Aitken classification, based on the extent of radiographic abnormality of the proximal femur (45), fails to give an adequate idea of the status of the femoral head, the relationship of the femur and the acetabulum, and the degree of continuity between the proximal and distal segments. More recently, sonography has proved to be a useful way to assess the relationship of the hip structures (23). Magnetic resonance imaging also depicts the structures of the hip and shows the degree of involvement of the musculature, which is important in presurgical planning. Finally, MR imaging can provide valuable information regarding the status of the physis and the vascularity of the femur (49).

2. DEVELOPMENTALDYSPLASIA OF HIP (DDH) Sonography has allowed a clear visualization of the unossifted capital femoral epiphysis (22) and the contour of the cartilaginous acetabulum (Fig. 1). Dynamic imaging in the transverse plane evaluates the stability of the hip following stress maneuvers, such as the Barlow maneuver (27). More recently, the addition of Doppler sonography has proved useful to detect normal perfusion and changes in blood flow with varying degrees of abduction. There is considerable controversy regarding the indications for hip sonography and the degree to which it has influenced the outcome of children with hip dysplasia (30). A limited, mostly morphological approach can be used for screening. Coronal images are obtained to evaluate the degree of coverage of the femoral head (normally 50% or greater), the degree of inclination of the acetabulum with regard to the ilium (quantified by the oLangle), and the inclination of the fibrocartilaginous labrum with respect to the ilium ([3 angle). It is also useful to evaluate the presence or absence of the ossification center, as hips with an epiphyseal ossification center appear to be at less risk of developing avascular necrosis with abduction. If only infants at risk are examined, it is possible to perform a more complete examination, evaluating for instability in the transverse plane.

FIGURE 14 Re-dislocationof a hip after closed reduction in a 2month-old girl. An axial CT scan of the hips shows that the proximal metaphysis of the left femur is displaced posteriorly (open arrow). The proximal femoral epiphysis on the right is well-located and is completely cartilaginous (closed arrow).

CT can be performed to assess reduction in patients placed in spica casts (Fig. 14). CT shows whether the ossification center is present and demonstrates the relationship of the femoral head to the acetabulum and the degree of posterior or lateral subluxation of the femoral epiphysis. The cartilaginous head is isodense to the adjacent soft tissues; the position of the head can be inferred by following the contour of the most proximal metaphysis. The center of the femoral head normally is slightly posterior to a line drawn between the lucent triradiate cartilages. A modified Shenton's line, drawn between the most anterior proximal metaphysis and the anterior pubis, is also useful in assessing reduction; this line normally should be continuous (70). A pulvinar may be seen as a radiolucent structure in the depth of the acetabulum. The dislocated femoral head most often is located laterally and posteriorly. MR imaging is used primarily in complicated cases in which there are obstacles to reduction or growth disturbances (Figs. 15 and 16). In older patients, CT with threedimensional reconstruction and MR imaging are valuable for the evaluation of acetabular dysplasia, deformity of the femoral head, and coverage of the femoral head by the acetabulum (Fig. 17). There is increasing interest in using MR imaging to evaluate growth disturbances of the hip that are due to abductionrelated ischemia during the treatment of hip dysplasia. A study in experimental animals has shown that gadoliniumenhanced MR imaging can show abduction-related areas of ischemia in the unossified capital femoral epiphysis (44). Ischemia can result in dysfunction or bridging of the physis of the femoral head. When the femoral head is unossified, the signs of physeal abnormality include loss of the highsignal-intensity zone corresponding to the physis on T2weighted images, loss of the zone of provisional calcification on any pulse sequence, and abnormal location of the growth recovery lines. In the hip, growth recovery lines, known as

SECTION V ~ Disorders of the Lower Extremity

143

FIGURE 15 Growthdisturbance of the proximal femur following avascular necrosis during therapy for developmentaldysplasia of the hip in a 3-year-old girl. (A) Frontal radiograph shows that the ossification center of the femoral head is small and the acetabulum is irregular. The hip is well located. The O'Brien lines (arrows) are closer to the proximal physis than to the lesser trochanter, indicating proximal femoral growth slowdown. There is irregularityof the proximal femoral physis. (B) Coronal gradient echo image shows that even though the physis is open, it is markedlyirregular. There is cartilage extending into the metaphysis.

O'Brien's lines (56), usually become visible after immobilization in a spica cast. If the physis functions normally, the physis of the femoral head migrates away from the growth recovery line faster than that from the greater trochanter. In cases of ischemia to the proximal femoral physis, the growth recovery line will be closer to the femoral head than to the greater trochanter (Fig. 15). In older children, once the femoral head is almost completely ossified, physeal abnormalities usually are manifested as bony bridges with associated deformity (Fig. 16).

3. LEGG--CALVIE--PERTHES DISEASE (LCP) There is little agreement regarding the therapy for L e g g Calv6-Perthes disease. Hence, it is difficult to establish the role of imaging in this disorder. Plain radiographs can usually establish the diagnosis. If the radiographs are normal, a sonogram can detect articular fluid. A sonographically guided aspiration can exclude septic arthritis (77). On skeletal scintigraphy, ischemia is seen as an area of decreased tracer activity. At a later stage, scintigraphy can demonstrate revascularization, as was shown previously. MR imaging can be useful at several stages. At presentation, if the diagnosis is unclear, MR imaging can show a decrease in the signal intensity of the epiphyseal marrow on Tl-weighted images and increased signal intensity on T2-weighted images, consistent with osteonecrosis (Fig. 18) (8). MR can demonstrate causes of decreased containment such as effusion, synovial thickening, and hypertrophy of the epiphyseal

cartilage (65). Additionally, MR can exclude other diseases and narrow the differential diagnosis (34). MR, however, is most useful at a later stage to help define the need for surgery

FIGURE 16 Lateralgrowth arrest in a 5-year-oldgirl who had a history of hip dysplasia treated in an abduction spica cast. Coronal Tl-weighted image shows a valgus deformity.There is a bony bridge in the lateralaspect of the physis (arrow).

144

CHAPTER 2 ~

Imaging Approaches for Epiphyseal Assessrnenr

F I G U R E 1"7 Marked acetabular dysplasia in an 8-year-old girl. (A) Frontal radiograph of the hips shows bilateral dysplastic acetabula with abnormal ossification of the acetabular roof. There is decreased lateral coverage of the femoral head. (B) Coronal gradient-recalled echo image shows that the gap between the acetabulum and the femur is filled by unossified acetabular cartilage, which is of high signal intensity. The labrum (closed arrow) of the acetabulum can be well identified and confirms the lack of coverage of the lateral femoral head. (C) Three-dimensional reconstruction with surface rendering shows the relative lateral displacement of the right femoral head and the lack of coverage by the bony acetabulum.

and suggest the prognosis. Determination of containment and definition of the degree of flattening of the femoral head are achieved readily by MR. M R imaging can also establish physeal abnormalities such as physeal bridging, which is

seen in 60% of cases and strongly correlated with subsequent growth arrest (40) (Fig. 10). Finally, it is possible that gadolinium-enhanced M R imaging may help determine the phase of evolution of the disease and the degree of recovery

SECTION V ~ Disorders of the Lower Extremity

145

FIGURE 18 Epiphysealmarrow abnormality in a 7-year-old boy with Legg-Calv6-Perthes disease. Coronal T2-weighted image of the proximal femur shows marked increase in signal intensity in the marrow of the femoral head. The abnormality extends into a small portion of the metaphysis. There is lateral displacement of the femoral head and evidence of a joint effusion. The radiograph was normal.

of the head (16). Arthrography can also demonstrate containment, degree of loss of sphericity of the femoral head, and most importantly, how the relationship between the femoral head and the acetabulum changes with position. 4. SLIPPED CAPITAL FEMORAL EPIPHYSIS (SCFE) In patients with slippage of the femoral head, CT shows the degree of physeal separation and the direction and extent of epiphyseal displacement (Fig. 19) (66). Coronal, sagittal, and three-dimensional reconstructions reveal the extent of varus deformity as well as the relative contributions of the inferior and posterior epiphyseal displacement to the deformity (69). Additional slices at the level of the femoral condyles allow the measurement of femoral anteversion; children with retroversion are more prone to slippage (21). MR imaging can provide similar multiplanar information (Fig. 20), and can also demonstrate ischemia to the femoral head.

B. Disorders of the Knee 1. NORMAL VARIANTS In the knee, it is important to visualize the menisci and ligaments as well as the chondro-osseous structures. The best sequence for visualization of meniscal abnormalities continues to be spin echo, proton-density-weighted imaging. Gradient-recalled echo imaging is also very adequate and can be prescribed in a radial fashion. In children, the menisci contain a central vessel that enters the meniscus on its peripheral aspect and bisects the substance of the meniscus (79). The vessel tapers toward the tip of the meniscus. A1-

FIGURE 19 Bilateral slipped capital femoral epiphysis in a 15-yearold boy. (A) Axial CT scan of the hips shows that the proximal femoral epiphyses are displaced posteriorly. There is posterior bowing of the left femoral neck, indicating a greater chronicity. Both growth plates are irregular and there is physeal widening on the left. (B) Coronal reconstruction shows bilateral physeal widening and inferior epiphyseal displacement on the left.

though it can resemble a horizontal tear, it never reaches the articular surface. On MR imaging the distal femoral epiphyseal cartilage undergoes changes with age. In the infant who does not bear weight, the cartilaginous epiphysis is of relatively homogeneous signal intensity on T2-weighted images (Fig. 7). By early childhood, there is a distinct area of decreased signal intensity along the weight bearing region. Inhomogeneity of the posterior femoral condyle becomes more apparent with age, possibly representing a response to stress. Occasionally, the signal heterogeneity may affect the bone as well. These abnormalities resemble and may be precursors of osteochondritis dissecans. Irregularities of ossification at the chondroosseous junction, well-described radiographically, can also be very apparent by MR imaging. The physis of the distal femur becomes progressively more undulated with age. Physeal closure on MR imaging is

146

CHAPTER 2 ~

Imagin9 Approaches for Epiphyseal Assessment;

F I G U R E 20 Slipped capital femoral epiphysis in a 12-year-old boy. (A) Frontal radiograph shows inferior displacement of the epiphysis and widening and irregularity of the growth plate. The metaphyseal border is poorly defined, especially medially. (B) Coronal fat-suppressed T2-weighted image of the femur shows the area of physeal abnormality as a high-signal-intensity region (arrow). The inferior epiphyseal displacement is well-demonstrated. (C) Axial T2-weighted image demonstrates the posterior displacement of the femoral head.

seen as loss of signal intensity, which occurs centrally in the distal femur and proximal tibia. The hypointense bone of the zone of provisional calcification remains for several years after. 2. CONGENITAL ANOMALIES Congenital abnormalities of the knee are best studied with MR imaging (49), although sonography may suffice to diagnose a congenital dislocation. Epiphyseal underdevelop-

ment in conditions such as congenitally short femur and hemimelias can be difficult to appreciate radiographically. In patients with proximal focal femoral deficiency and congenitally short femur, the distal femoral epiphysis is fiat and there is no well-formed intercondylar notch. The cruciate ligaments usually are absent. In hemimelias, the unossified epiphysis on the side of the abnormality is malformed or absent. In these cases, MR imaging can demonstrate whether the remaining bone of the leg articulates with the femur and whether there

SECTION VI 9 Disorders of the Upper Extremity is a well-developed extensor mechanism (quadriceps muscle, patellar tendon) (61). 3. BLOUNT DISEASEmTIBIA VARA

In B lount disease, MR imaging is used to detect a bony bridge in the proximal tibia, particularly prior to performing a valgus osteotomy. MR imaging can also be performed to assess the status of the unossified epiphysis of the proximal tibia (17). MR imaging findings of Blount disease reflect stress injury to the entire medial aspect of the knee. The proximal tibial physis is irregular and deviated distally, with areas of focal physeal widening. Bony bridging can occur in the medial aspect of the tibial physis. In physes that are markedly curved, coronal or sagittal sections may falsely suggest bridging due to volume averaging with adjacent bone. Imaging in three planes thus is important in the evaluation of the proximal tibia. The signal intensity of the unossified medial tibial epiphysis is often increased on T2-weighted images. The contour of the unossified medial tibial plateau is always nearly normal. Abnormalities can exist in the distal femoral marrow and distal femoral physis. The medial meniscus typically is wide and has increased signal intensity. 4. LESIONS ASSOCIATED WITH CHRONIC REPETITIVE TRAUMA

Osteochondritis dissecans is most common in the lateral aspect of the medial femoral condyle. It is bilateral in onethird of the cases. MR imaging is useful to determine whether the overlying cartilage is intact. High signal intensity on T2-weighted images extending between the osteochondral fragment and the parent bone indicates instability, as it suggests that the overlying articular cartilage is disrupted (15). MR imaging is insensitive to the detection of loose intra-articular bodies. The extensor mechanism of the knee can have abnormalities that are related to traction on the cartilaginous tibial tubercle and the unossified patella. Increased weakness in these structures as they develop from fibrocartilage to hyaline cartilage is believed to account for these lesions. The patellar sleeve fracture, an avulsion of the cartilaginous inferior pole of the patella in early adolescence, can be completely inapparent on radiographs. MR imaging can show the extent of the cartilaginous injury (4). MR can also show the cartilaginous lesion in Osgood-Schlatter disease and in avulsion fractures of the unossified or partially ossified tibial tubercle. However, clinical examination and plain radiographs usually suffice for diagnosis. 5. OTHER DISORDERS

Epiphyseal osteomyelitis is seen most commonly in the distal femoral epiphysis; MR imaging shows that the focus of infection is usually at the chondro-osseous junction. There is typically increased prominence of the vascular canals in the vicinity of the lesion.

147

C. Disorders of the Distal Tibia and Fibula 1. NORMAL DEVELOPMENT AND NORMAL VARIANTS The relative level of the tibial and fibular physes, important as an indicator of normal growth, varies with age. At birth, the physes of the tibia and fibula are at almost the same level. Later, the fibular physis becomes situated more distally at the level of the talar dome (59). The asymmetrical pattern of closure of the distal tibial physis has a significant influence on radiographs, CT, and MR studies. The physis, initially fiat, develops an undulation on its medial and anterior aspect, the bump of Poland or "Kump's bump," located over the medial talar hump. On CT and MR images of older children, sagittal and coronal images can give the false appearance of closure due to volume averaging (10). Normal physeal closure resulting in loss of the normal cartilaginous signal on MR images begins in this region. 2. TRAUMA Prior to physeal closure, MR imaging is useful in the evaluation of complications of certain fractures of the growth plate. Physeal fractures of the tibia and fbula are common during puberty and early adolescence. MR imaging is useful in demonstrating bony bridges that are seen most commonly in the distal tibia medially in the area of Poland's hump. Imaging of the fractures that occur during physeal closure, the juvenile Tillaux fracture and the triplane fracture, is best done with CT (20). The goal of imaging in these cases is to characterize the often multiplanar fracture patterns and to demonstrate any separation of the bony fragments, particularly intra-articular components. We routinely obtain 1-mm-thick axial slices through the area of the fracture and generate two- and three-dimensional reconstructions. The two-dimensional coronal and sagittal images are usually the most helpful for evaluation of the injury.

VI. IMAGING OF DISORDERS BY LOCATION: DISORDERS OF THE UPPER EXTREMITY A. Shoulder Epiphyseal underdevelopment can occur in the humeral head as a result of neuromuscular dysfunction. Patients with Erb's palsy typically have a smaller humeral epiphysis and a poorly developed posterior glenoid. These conditions predispose one to posterior dislocation of the humeral head. The glenohumeral relationships can be demonstrated with CT (31). MR imaging demonstrates the cartilaginous glenoid to better advantage (24). The physis of the posterior glenoid becomes progressively deficient as the deformity increases. In traumatized neonates, the distinction between proximal humeral epiphyseal separation and shoulder dislocation can be difficult to establish. Sonography can differentiate between the two by showing a lack of continuity between

CHAPTER 2 ~ lrnaffin9 Approaches for Epiphyseal Assessment

148

the humeral epiphysis and the shaft in the case of a SalterHarris type I injury (9).

B. Elbow Imaging by sonography and MR has increased our understanding of elbow abnormalities, particularly in congenital and traumatic conditions. In the newborn and infant, dislocations of the elbow, particularly those of the radial head, can be seen sonographically. Sonography has the advantage of showing the structures in real time, so that the positional variation in articular relationships can be assessed. In older patients, or when the anomalies also involve the bones such as in radioulnar synostosis, MR imaging can provide a more global picture. Imaging of traumatic and posttraumatic deformities of the elbow is a growing application of MR (6). MR imaging can demonstrate the extension of a fracture in the unossified epiphysis. This can be of value in non-displaced or minimally displaced lateral condylar fractures, in which the extension into the joint can predispose one to instability and nonunion. Separations of unossified epiphyses or apophyses can be difficult to diagnose radiographically and can be seen clearly with MR imaging. In patients with posttraumatic deformity of the elbow, MR imaging is useful for evaluation of the integrity of the physis. In younger patients with mostly cartilaginous distal humeral epiphyses, an abnormal physis usually is irregular with loss of the contour of the zone of provisional calcification. In these patients, the physeal cartilage can be destroyed focally, and yet there is no evidence of bony bridging as the adjacent epiphysis is also cartilaginous. Older patients with posttraumatic abnormalities of the elbow sometimes benefit more from CT than from MR imaging, particularly if there is a question of associated intraarticular loose bodies.

C. Wrist Two distal radial physeal abnormalities have been studied with MR imaging. As was mentioned previously, gymnasts show areas of widening and occasionally of physeal bridging on gradient-recalled echo imaging. In some patients with Madelung deformity, MR imaging has shown bridging of the medial aspect of the physis (13).

References 1. Avni EF, Rypens F, Zappa M, Donner C, Reglmorter NV, Cohen E (1996) Antenatal diagnosis of short-limb dwarfism: Sonographic approach. Pediatr Radiol 26:171-178. 2. Ballon D, Jakubowski A, Gabrilove J, Graham MC, Zakowski M, Sheridan C, Koutcher JA (1991) In vivo measurements of bone marrow cellularity using volume-localized proton NMR spectroscopy. Magn Reson Med 19:85-95.

3. Barnewolt CE, Shapiro F, Jaramillo D (1997) Normal gadolinium-enhanced MR images of the developing appendicular skeleton: Part 1. Cartilaginous epiphysis and physis. AJR 169:183-189. 4. Bates GD, Hresko MT, Jaramillo D (1994) Patellar sleeve fracture: Demonstration with MR imaging. Radiology 193:825-827. 5. Bearcroft P, Berman L, Robinson A, Butler G (1996) Vascularity of the neonatal femoral head: In vivo demonstration with Power Doppler US. Radiology 200:209-211. 6. Beltran J, Rosenberg ZS, Kawelblum M, Montes L, Bergman AG, Strongwater A (1994) Pediatric elbow fractures: MR evaluation. Skeletal Radio123:277-281. 7. Borsa JJ, Peterson HA, Ehman RL (1996) MR imaging of physeal bars. Radiology 199:683-687. 8. Bos CFA, Bloem JL, Bloem RM (1991) Sequential magnetic resonance imaging in Perthes disease. J Bone Joint Surg 73B: 219-224. 9. Broker FHL, Burbach T (1990) Ultrasonic diagnosis of separation of the proximal humeral epiphysis in the newborn. J Bone Joint Surg 72-A:187-191. 10. Chung T, Jaramillo D (1995) Normal maturing distal tibia and fibula: Changes with age at MR imaging. Radiology 194: 227-232. 11. Conway J (1993) Scintigraphic classification of Legg-Calv6Perthes disease. Semin Nucl Med 23:274-295. 12. Conway W, Totty W, McEnery K (1996) CT and MR imaging of the hip. Radiology 198:297-307. 13. Cook PA, Yu JS, Wiand W, Lubbers L, II, AJC, Kean JR, Cook AJ (1996) Madelung deformity in skeletally immature patients: Morphologic assessment using radiography, CT, and MRI. J Comput Assist Tomogr 20:505-511. 14. Dawson K, Moore S, Rowland J (1992) Age-related marrow changes in the pelvis: MR and anatomic findings. Radiology 183:47-51. 15. De Smet A, Fisher DA, Graf BK, Lange RH (1990) Osteochondritis dissecans of the knee: Value of MR imaging in determining lesion stability and the presence of articular cartilage defects. AJR 155:549-553. 16. Ducou le Pointe H, Haddad S, Silberman B, Filipe G, Monroc M, Montagne JP (1994) Legg-Perthes-Calv6 disease: Staging by MRI using gadolinium. Pediatr Radiol 24:88-91. 17. Ducou le Pointe H, Mousselard H, Rudelli A, Montagne J-P, Filipe G (1995) Blount's disease: Magnetic resonance imaging. Pediatr Radiol 25:12-14. 18. Dwek JR, Shapiro F, Laor T, Barnewolt CE, Jaramillo D (1997) Normal gadolinium-enhanced MR images of the developing appendicular skeleton: Part 2. Epiphyseal and metaphyseal marrow. AJR 169:191-196. 19. Eggli KD, King SH, Boal DKB, Quoigue T (1994) Low-dose CT of developmental dysplasia of the hip after reduction: Diagnostic accuracy and dosimetry. AJR 163:1441-1443. 20. Feldman F, Singson R, Rosenberg Z, Berdon W, Amodio J, Abramson S (1987) Distal tibial triplane fractures: Diagnosis with CT. Radiology 164:429-435. 21. Gelberman RH, Cohen MS, Shaw BA, Kasser JR, Griffin PP, Wilkinson RH (1986) The association of femoral retroversion with slipped capital femoral epiphysis. J Bone Joint Surg 68A: 1000-1007. 22. Graf R (1987) Guide to Sonography of the Infant Hip. New York: Thieme.

References

23. Grissom LE, Harcke HT (1994) Sonography in congenital deficiency of the femur. J Pediatr Orthop 14:29-33. 24. Gudinchet F, Maeder P, Oberson JC, Schnyder P (1995) Magnetic resonance imaging of the shoulder in children with brachial plexus birth palsy. Pediatr Radiol 25:S125-S128. 25. Harcke H (1994) Hip and musculoskeletal US. In: Current Concepts: A Categorical Course in Pediatric Radiology, J. Seibert ed. pp. 119-124, Oak Brook, IL: Radiological Society of North America. 26. Harcke H, Mandell G (1993) Scintigraphic evaluation of the growth plate. Semin Nucl Med 23:266-273. 27. Harcke HT, Grissom LE (1990) Performing dynamic sonography of the infant hip. AJR 155:837-844. 28. Harcke HT, Lee MS, Sinning L, Clarke NMP, Borns PF, MacEwen GD (1986) Ossification center of the infant hip: Sonographic and radiographic correlation. AJR 147:317-321. 29. Henkelman RM, Stanisz GJ, Kim JK, Bronskill MJ (1994) Anisotropy of NMR properties of tissues. Magn Reson Med 32:592-601. 30. Hernandez RJ, Cosnell RG, Hensinger RN (1994) Ultrasound diagnosis of neonatal congenital dislocation of the hip. A decision analysis assessment. J Bone Joint Surg 76B:539-543. 31. Hernandez RJ, Dias L (1988) CT evaluation of the shoulder in children with Erb's palsy. Pediatr Radiol 18:333-336. 32. Hillman J, Mezgarzadeh M, Revesz G, Bonakdarpuor A, Clancy M, Betz R (1987) Proximal femoral focal deficiency: Radiologic analysis of 49 cases. Radiology 165:769-773. 33. Hodler J, Resnick D (1996) Current status of imaging of articular cartilage. Skeletal Radiol 25:703-709. 34. Hubbard AM, Dormans JP (1995) Evaluation of developmental dysplasia, Perthes disease, and neuromuscular dysplasia of the hip in children before and after surgery: An imaging update. AJR 164:1067-1073. 35. Jaramillo D, Shapiro F (1998) Growth cartilage: Normal appearance, variants and abnormalities. MRI Clin North Am 6: 455-471. 36. Jaramillo D, Connolly SA, Mulkern RV, Shapiro F (1998) Developing epiphysis: MR imaging characteristics and histologic correlation in the newborn lamb. Radiology 207:637-645. 37. Jaramillo D, Hoffer F, Shapiro F, Rand F (1990) MR imaging of fractures of the growth plate. AJR 155:1261-1265. 38. Jaramillo D, Hoffer FA (1992) Cartilaginous epiphysis and growth plate: Normal and abnormal MR imaging findings. AJR 158:1105-1111. 39. Jaramillo D, Kammen BE Shapiro F (2000) Cartilaginous path of physeal fracture-separations: Evaluation with MR imaging~ an experimental study with histologic correlation in rabbits. Radiology 215:504-511. 40. Jaramillo D, Kasser JR, Villegas-Medina OL, Gaary E, Zurakowski D (1995) Cartilaginous abnormalities and growth disturbances in Legg-Calvr-Perthes disease: Evaluation with MR imaging. Radiology 197:767-773. 41. Jaramillo D, Laor T, Hoffer FA, Zaleske DJ, Cleveland RH, Buchbinder BR, Egglin TK (1991) Epiphyseal marow in infancy: MR imaging. Radiology 180:809-812. 42. Jaramillo D, Laor T, Zaleske D (1993) Indirect trauma to the growth plate: Results of MR imaging after epiphyseal and metaphyseal injury in rabbits. Radiology 187:171-178. 43. Jaramillo D, Shapiro F, Hoffer FA, Winalski CS, Koskinen ME Frasso R (1990) Posttraumatic growth plate abnormalities: MR

44.

45. 46.

47.

48.

49.

50.

51.

52.

53.

54. 55.

56.

57.

58.

59.

60.

61.

62.

149

imaging of bony bridge formation in rabbits. Radiology 175: 767-773. Jaramillo D, Villegas-Medina OL, Doty DK, Dwek JR, Ransil B J, Mulkern RV, Shapiro F (1996) Gadolinium-enhanced MR imaging demonstrates abduction-caused hip ischemia and its reversal in piglets. AJR 166:879-887. Kalamchi A, Cowell HR, Kim KI (1985) Congenital deficiency of the femur. J Pediatr Orthop 5:129-134. Kleinman P, Marks S (1995) Relationship of the sub-periosteal collar to metaphyseal lesions in abused infants. J Bone Joint Surg 77A:1471-1476. Kleinman P, Marks S, Richmond J, Blackbourne B (1995) Inflicted skeletal injury: A postmortem radiologic-pathologic study in 31 infants. AJR 165:647-650. Kricun M (1985) Red-yellow marrow conversion: Its effect on the location of some solitary bone lesions. Skeletal Radiol 14: 10-19. Laor T, Jaramillo D, Hoffer FA, Kasser JR (1996) MR imaging in congenital lower limb deformities. Pediatr Radiol 26: 381-388. Levine D, Tatabu H, Gaa J, Atkinson M, Edelman R (1996) Fetal anatomy revealed with fast MR sequences. AJR 167: 905-908. Link TM, Meier N, Rummeny EJ, Garmann S, Steinbeck J, Schroder M, Oelerich M, Haussler MD (1996) Artificial spine fractures: Detection with helical and conventional CT. Radiology 198:515-519. Mahony B (1994) Ultrasound evaluation of the fetal musculoskeletal system. In: Ultrasonography in Obstetrics and Gynecology, P Callen ed. pp. 254-290, Philadelphia: Saunders. Moore S, Dawson K (1990) Red and yellow marrow in the femur: Age-related changes in appearance at MR imaging. Radiology 175:219-223. Murray K, Nixon G (1988) Epiphyseal growth plate evaluation with modified coronal CT. Radiology 166:263-265. Norton K, Hermann G, Abdelwahab I, Klein M, Granowetter L, Rabinowitz J (1991) Epiphyseal involvement in osteosarcoma. Radiology 180:813-816. O'Brien T, Millis MB, Griffin PP (1986) The early identification and classification of growth disturbances of the proximal end of the femur. J Bone Joint Surg 68A:970-980. Oestreich AE, Ahmad BS (1992) The periphysis and its effect on the metaphysis. I. Definition and normal radiographic pattern. Skeletal Radiol 21:283-286. Oestreich AE, Ahmad BS (1993) The periphysis and its effect on the metaphysis. II. Application to tickets and other abnormalities. Skeletal Radio122:115-119. Ogden JA, McCarthy SM (1983) Radiology of postnatal skeletal development. VIII. Distal tibia and fibula. Skeletal Radiol 10:209-220. Peterfy CG, Majumdar S, Lang P, van Dijke C, Sack K, Genant HK (1994) MR imaging of the arthritic knee: Improved discrimination of cartilage, synovium, and effusion with pulsed saturation transfer and fat-suppressed Tl-weighted sequences. Radiology 191:413-419. Pirani S, Beauchamp R, Li D, Sawatzky B (1991) Soft tissue anatomy of proximal femoral focal deficiency. J Pediatr Orthop 11:563-570. Ricci C, Cova M, Kang YS, Yang A, Rahmouni A, Scott WWJ, Zerhouni EA (1990) Normal age-related patterns of cellular

150

63. 64. 65.

66.

67.

68.

69.

70.

CHAPTER 2 ~

Imaging Approaches for Epiphyseal Assessment

and fatty bone marrow distribution in the axial skeleton: MR imaging studies. Radiology 177:83-88. Rogers L, Poznanski A (1994) Imaging of epiphyseal injuries. Radiology 191:297-308. Romero R, Athanassiadis A, Jeanty P (1989) Fetal skeletal anomalies. Radiol Clin North Am 28:75-99. Rush BH, Bramson RT, Ogden JA (1988) Legg-Calve,Perthes Disease: Detection of cartilaginous and synovial changes wth MR imaging. Radiology 167:473-476. Schlesinger AE, Hemandez RJ (1992) Diseases of the musculoskeletal system in children: Imaging with CT, sonography, and MR. Am J Roentgenol 158:729-741. Sebag G, Dubois J, Tabet M, Bonato A, Lallemand D (1993) Pediatric spinal bone marrow: assessment of normal age-related changes in the MRI appearance. Pediat Radiol 23:515-518. Shih C, Chang C, Penn I, Tiu C, Chang T, Wu J (1995) Chronically stressed wrists in adolescent gymnasts: MR imaging appearance. Radiology 195:855-859. Siegel D, Kasser J, Sponseller P, Gelberman R (1991) Slipped capital femoral epiphysis. A quantitative analysis of motion, gait, and femoral remodeling after in situ fixation. J Bone Joint Surg 73A:659-666. Smith BG, Kasser JR, Hey LA, Jaramillo D, Millis MB (1997) Postreduction computed tomography in developmental dislocation of the hip. Part I: Analysis of measurement reliability. J Pediatr Orthop. 17:626-630.

71. Soboleski DA, Babyn P (1993) Sonographic diagnosis of developmental dysplasia of the hip: Importance of increased thickness of acetabular cartilage. AJR 161:839-842. 72. Treves S, Connolly L, Kirkpatrick J, Packard A, Roach P, Jaramillo D. Bone. In: Pediatric Nuclear Medicine, 2nd edition, S. Treves, ed. pp. 233-301, New York: Springer-Verlag. 73. Vogler J, Murphy W (1988) Bone marrow imaging. Radiology 168:679. 74. White K (1996) Helical-spiral CT scanning: A pediatric radiology perspective. Pediatr Radiol 26:5-14. 75. Wolff SD, Balaban RS (1994) Magnetization transfer imaging: Practical aspects and clinical applications. Radiology 192: 593-599. 76. Yousefzadeh D, Ramilo J (1987) Normal hip in children: Correlation of US with anatomic and cryomicrotome sections. Radiology 165:647-655. 77. Zawin J, Hoffer F, Rand F, Teele R (1993) Joint effusion in children with an irritable hip: US diagnosis and aspiration. Radiology 187:459-463. 78. Zawin JK, Jaramillo D (1993) Conversion of bone marrow in the humerus, sternum, and clavicle: Changes with age on MR images. Radiology 188:159-164. 79. Zobel MS, Borrello JA, Siegel MJ, Stewart NR (1994) Pediatric knee MR imaging: Pattern of injuries in the immature sketon. Radiology 190:397-401.

CHAPTER

3

Developmental Dysplasia of the Hip I. II.

VIII.

Terminology Development of the HipmEmbryonic and Fetal Periods

IX.

III. Primary Etiologies of Hip Maldevelopment IV. Etiology and Pathoanatomy of Developmental Dysplasia of the Hip V. Epidemiology and Its Relation to Pathophysiology VI. Summary of Pathoanatomic and Pathophysiologic Findings and Discussion of Pathogenetic Sequences VII.

X. XI. XII.

Natural History of Hip Dislocations, Subluxations, and Dysplasia

Brief History of Treatment Approaches in Developmental Dysplasia of the Hip Imaging Techniques Used to Assess Hip Position Assessments of Hip Growth and Development Following Closed and Open Treatments Treatment Based on the State of the Underlying Pathoanatomy, Including Secondary Changes Avascular Necrosis as a Complication of Treatment of Developmental Dysplasia of the Hip

Until a few years ago, the term congenital dislocation of the hip (CDH) was used to describe the entity, although some used the term congenital dysplasia of the hip in an effort to encompass the entire spectrum of the disorder. Dunn defined congenital dislocation of the hip as an "anomaly of the hip joint, present at birth, in which the head of the femur is, or may be, partially or completely dislocated from the acetabulum" (61). The entity is now known increasingly by the term developmental dysplasia of the hip (DDH). The word developmental has replaced congenital (1) because it focuses on abnormalities in development that predispose one to the condition and that worsen in the absence of normal hip positioning and (2) because it is not definitely clear that all dysplastic hips are structurally abnormal and/or detectably so at the time of the initial postnatal examination. The term dysplasia itself is a vague and general one referring to a poorly defined disease process. Delayed and thus imperfect development of the acetabulum and the proximal femur is referred to as a dysplastic process. Acetabular dysplasia and proximal femoral dysplasia themselves are either primary disorders and/or disorders that occur secondary to growth in the presence of undetected and untreated developmental hip disease. Developmental dysplasia of the hip therefore encompasses a spectrum of hip abnormalities. These include the following: (1) an initial subluxatable or dislocatable hip in which the femoral head is located in a normal relation to the acetabulum in certain positions (generally flexion and abduction), but has a partial or complete loss of continuity in other positions. This situation can correct itself spontaneously

I. T E R M I N O L O G Y Developmental dysplasia of the hip is a general term referring to a spectrum of deformities, usually diagnosed in the neonatal period, in which the structural relationship of the proximal femur to the acetabulum is intermittently or continuously abnormal. The spectrum includes the following: (1) a subluxatable or dislocatable hip associated with capsular laxity in which the head of the femur moves partially or totally out of the acetabulum with extension and adduction and back into it with flexion and abduction; (2) a subluxated hip in which there is a partial but persisting loss of the normal relationship of the head of the femur to the acetabulum in extension with the head more lateral than normal in the acetabulum and the acetabulum more shallow than normal with its lateral roof angled outwardly and upwardly; and (3) a dislocated hip in which there is a complete and persisting loss of any femoral head-acetabular relationship, regardless of the position of the hip. Developmental dysplasia of the hip (DDH), as currently defined, is not associated with clinically evident connective tissue, neuromuscular, or other diseases. The single most important initial pathoanatomic change appears to be a capsular laxity, which renders the hip unstable at birth, with all subsequent abnormalities being secondary phenomena that develop an increasing variation from the norm the longer a hip is allowed to grow with any persisting malposition. The terminology used to describe this condition has always been variable and imprecise primarily due to the imperfect understanding of the pathoanatomy and the timing of its initial occurrence.

153

154

CHAPTER 3 ~ Developmental Dysplasia of the Hip

within a few days of birth or if untreated it can progress to persisting deformity. (2) A subluxation of the hip, which refers to a partial loss of continuity between the femoral head and acetabulum in which the abnormal relationship is present throughout the entire range of movement. (3) A dislocated hip in which there is a complete loss of continuity between joint surfaces at all times, regardless of the position of the hip. Terminological distinctions are not merely a semantic issue; imprecise use of terms implies imprecise understanding of the underlying pathoanatomy, which can lead to investigations and treatments that are not fully appropriate.

II. D E V E L O P M E N T O F T H E H I P ~ EMBRYONIC AND FETAL PERIODS A. General Aspects of Hip Development The embryonic period in the human refers to the first 8 weeks of development by which time each of the various organs, including the cartilaginous outlines of the long bones and vertebrae, has been formed. By the end of the embryonic period, the average embryo is of 3 cm crown-rump (C-R) length. The fetal period from 3 months of age to birth is associated with an increase in size and more subtle organ differentiation. Watanabe outlined hip development in 288 hips from 144 embryos and fetuses from 14 to 300 mm C - R length ending at 24 weeks gestation (340). The femoral head diameter around 11 weeks was 2 mm, at which time the joint space was formed and the head could be dislocated by cutting the capsule. The femoral head diameter by 24 weeks was around 8 mm. The diameter of the femoral head increases in size in a linear pattern and parallels the growth of the entire body. The femoral head is spherical at the beginning of development and remains so throughout growth. The neck shaft angle was 130~ during fetal development. Femoral anteversion averaged - 4 ~ from 10 to 15 weeks, 5~ from 15 to 20 weeks, and 11~ from 20 to 24 weeks, but there was a wide range of variability at these times of both positive and negative values. At birth femoral anteversion had increased to 35 ~ Watanabe's study found no examples of full dislocation, but there were 26 dysplastic hip joints characterized by "an overall hypoplasia of the entire hip joint with a shallow acetabulum." The femoral head was always stable with flexion and tended to subluxate with extension. The femoral head and acetabulum had reached infantile shape prior to joint space formation such that dislocation could not occur during the embryonic period. Strayer studied hip development from human embryos 6.5 to 237 mm crown-rump length (292). He also concluded, in agreement with other observers, that all elements of the hip joint differentiate in situ in one mass of blastema. The head of the femur is globular in shape at all times during its development, and the relative proportions of the blaste-

mal and early cartilage developmental segments of the pelvic bones entering into formation of the acetabulum are the same in early embryos as in fetal stages and postnatal life. The ligamentum teres develops in situ within the joint. Congenital dislocation of the hip cannot occur before the opening of the joint cavity. The synovial lining did not develop as a cellular ingrowth but rather from cells in situ as part of the original blastemal mass. The synovium forms along the line of cleavage, which appears between cells as the interzone tissues are liquefied. The acetabulum develops by growth and fusion of processes from the iliac, ischial, and pubic cartilages. Each of these develops around the femoral head and their fusion initially produces a shallow acetabulum. The portion of the acetabulum to which each pelvic cartilage contributes is approximately the same as those later furnished by the corresponding pelvic bones, being 2/5 ischium, 2/5 ilium, and 1/5 pubis. Each of the pelvic cartilages has a centrifugal growth pattern within the blastema. The region that will become the hip joint initially is composed of dense blastemal tissue referred to as the interzonal tissue while the embryo is growing from 20 to 30 mm in length. Cavity formation begins in the tissue between the cartilage of the acetabulum and the cartilage of the head of the femur. The region of the ligamentum teres is differentiated from the blastema of the joint. The interzone tissue other than the ligamentum teres becomes looser in texture with time and ultimately is re'orbed to leave the joint cavity. Other studies indicated that the greater trochanter was evident at 30 mm and the femoral neck and lesser trochanter at 34 mm (292). The hip joint cavity was described by Moser (1893) as appearing first in the lateral part of the joint at 34 mm; Haines described an initial cavity at 34 mm. The ligamentum teres developed in situ with Moser describing it as early as 20 mm and Strayer noting it at 23 mm. The glenoid labrum was noted at 30 mm as a transition with the outer region of the acetabular cartilage. Dimeglio et al. reviewed the prenatal aspects of hip development, stressing the unique interrelationship of the pelvis, femur, and associated muscles on normal structure (57). They particularly stressed three aspects of growth: (1) the enlargement and full development of the acetabulum; (2) the harmonious spherical enlargement of the femoral head and its secondary ossification center; and (3) the elongation of the femoral neck in the postnatal period. Detailed observations on the prenatal development of the human hip joint were provided by Gardner and Gray (91) in a study based on 52 human embryos and fetuses ranging in crown-rump length from 12 (6 weeks) to 370 mm (term) and by Andersen (6) in a study of 30 human embryos-fetuses from 20 (7.5 weeks) to 121 mm (16 weeks). Their observations are in good agreement and are combined herein.

1. ORIGINOF LIMB BUO The lower extremity limb bud is seen in embryos 3-4 mm in length as a small protuberance on the anterior and lateral

SECTION Ii 9 Development of the Hip aspect of the body wall at the level of the lumbar and first sacral segments. The specific tissue differentiation for each bone then follows from blastemal tissue or undifferentiated mesenchymal cells, to precartilage, to cartilage, and then to bone. The region of the future hip joint appears as a group of densely packed, undifferentiated cells in the form of a cone with an oblique base applied to the side of the body. The first appearance of the acetabulum is in the 14- to 15mm embryo as a line of cells of diminished density proximal to the head of the femur. This region initially was felt to represent an arc of 65-70 ~, which subsequently deepened to enclose a full half-circle of 180 ~ as the joint cavity itself formed. The interzone demonstrates increased cell density by 15-22 mm. Early differentiation of the ligamentum teres and periarticular capsular structures is noted around 23 mm. As development and growth proceed from 23 to 45 mm, the cartilage of the ilium grows out over the head of the femur with the labrum attached to its outer margin. Increases in the extent of the elements of the acetabulum are responsible for the relative lateral displacement of the labrum. The acetabulum is never fiat; from the earliest stages it extends, together with the labrum, beyond the midway point of the head and always has a concave shape. Differentiation of blastema in the innominate region begins in the ilium just above the acetabulum at the 15-mm stage. This most lateral region lags behind the shaft and head of the femur in differentiation at all stages. The three cartilage centers become vascularized separately and serve to outline the Y-cartilage. Chondrification radiates from the three centers of these regions. 2. GLENOID LABRUM AND TRANSVERSE ACETABULAR LIGAMENT The glenoid labrum is formed at the earliest stages of the formation of the acetabulum, as early as 19 mm, and appears histologically as a condensation of blastema at the cartilage periphery. By 25 mm it is clearly differentiated. It becomes vascularized at 61 mm. The transverse acetabular ligament also forms during this time period; the site of the transverse ligament of the hip joint is considered by Strayer to be the weakest point of the structure. By 28 mm a condensation for the transverse acetabular ligament is seen, and by 30-33 mm the ligament is well-defined. The superior labrum covers the widest diameter of the femoral head. The anteroinferior part of the acetabulum, which is known as the acetabular notch, is covered by the transverse acetabular ligament (196). This ligament is the support for the acetabular labrum as it crosses the notch. 3. JOINT CAPSULE AND SYNOVIUM In 12- to 15-mm embryos, avascular blastemal tissue in the region of the future joint is denser than that in the adjacent anlagen. This density is more pronounced at 17 mm with a definite interzone being present. The interzone is more definite by 20 mm, and it is possible to define a threelayered interzone, the middle layer of which is directly con-

155

tinuous with the mesenchymal tissues surrounding the joint except in those areas of capsular condensation. The outer layers of the interzones are directly continuous with the perichondrium of the femoral and acetabular anlagen. The capsule surrounding the joint is defined. Contained within it is a portion of the mesenchyme surrounding the joint that is structurally a part of the interzone. This intra-articular mesenchyme is the first indication of what will become synovial mesenchyme. The intermediate layer of the interzone is continuous with the blastemal synovial mesenchyma and both are vascularized. The three-layered arrangement of the interzone is still more pronounced at 22-25 mm. Early spaces appear to be forming within the middle layer at these stages. By 30-33 mm a clear cavity is present around the periphery of the joint. Even at the time of opening of the joint space, it is not possible morphologically to distinguish between the cells of the inner margin of the capsule that will eventually form the synovial membrane and the capsule itself. The first indication of the fibrous capsule is seen at 20 mm with a condensation appearing as a direct continuation of the perichondrium of the femur and pelvis. 4. JOINT CAVITY Joint cavity formation represents a programmed degenerative and mechanical process with no evidence of ingrowth of tissue from the outside to provide a lining for the joint. Early evidence of degeneration is seen at 23 mm, with increases in the intercellular spaces in the interzonal cells between the head of the femur, the ligamentum teres, and the acetabulum. At 36-42 mm spaces filled with fluid are formed. Andersen times formation of the joint cavity between 34 and 42 mm. Vascularization of the interzone is an integral part of joint cavitation. Joint cavitation begins in the central area of the joints and then moves toward the periphery. Cavity formation at the hip takes place as an annular rim, limited medially by the head of the femur and laterally by the glenoid labrum; in the middle of the developing joint cavity the ligamentum teres remains present. In later stages of cavitation, the space is enlarged centrally around the ligamentum teres and peripherally passing beyond the tip of the labrum and surrounding the head in its entirety and also the neck down to the capsular insertion. 5. RETINACULA OF WEITBRECHT The extension of the joint space down the neck of the femur leaves as intrascapular structures the perichondrium, the retinacula of Weitbrecht, and the ascending cervical vessels. The retinacula of Weitbrecht are intracapsular flattened band structures of the hip joint present on the interior of the capsule and passing toward the margin of the femoral head. The retinacula are synovial-covered capsular reflections or prolongations. Walmsley has provided a clear description (339). The blood vessels eventually supplying the proximal femur perforate the capsular attachment at the base of the neck and pass along the surface of the neck, entering the

156

CHAPTER 3 9

Developmental Dysplasia of the Hip

metaphysis of the neck and the epiphysis of the head through small foramina. "From the points where they perforate the capsule these vessels derive and carry inwards indefinite fibrous prolongation of the capsule wall which are covered over or are completely invested by reflections of synovial membrane. These elements constitute the retinacula of Weitbrecht." The fibrous prolongations terminate at varying distances from their origins, whereas the synovial reflections covering the vessels continue toward the cartilaginous margin of the head. The retinacula are reflections or continuations of the synovial membrane combined with fibrous sheath prolongations of the capsular wall, which carry within themselves the blood vessels of the head and neck. 6. LIGAMENTUM TERES

At 22 mm the first suggestion of the ligamentum teres is found. The ligamentum teres is present in 22- to 25-mm specimens as a region of greater cellularity but is not sharply demarcated from the neighboring interzone. There is never any evidence of a depression in the head to receive the ligamentum. The separation of the ligamentum teres to form a free mass within the joint occurs simultaneously with the opening of the remainder of the cavity, which is characterized by peripheral vascularization, degeneration, and splitting between the cells along its margin. The ligamentum teres is well-defined in the 30- to 33-mm fetuses. Blood vessels are first noted within the ligamentum teres at 60 mm. The ligamentum teres originates broadly from each side of the acetabular notch and from the transverse acetabular ligament. It is attached to a depression on the femoral head just below and posterior to the center of the head (196). 7. EXTRA-ARTICULAR LIGAMENTS The hip receives additional stabilization from its extraarticular ligaments. Anterior and superior support is derived from the iliofemoral ligament, referred to by some as the Yligament of Bigelow or in French literature as the ligament of Bertrand (Fig. 1A). Posteriorly, support comes from the ischiofemoral ligament, the lower border of which is an almost discrete thickening at the back of the neck referred to as the orbicular ligament or zone (Fig. 1B). 8. SKELETON

The outline of the forming skeleton of the hip joint largely consists of condensed blastemal tissue in the 12-mm embryo. Outlines of the acetabulum barely are indicated at 12 mm but are slightly more apparent at 14 and 15 mm. An increase in cartilage depth is noted particularly in the 30- to 33-mm stages as the pubic, iliac, and ischial cartilages clearly enter into formation of the acetabulum. Both trochanters are welldefined but the neck is quite short. By 49-50 mm a center of ossification is present in the ilium near the acetabulum and vessels have begun to penetrate the glenoid labrum and the

adjacent cartilage. The femoral head is beginning to be vascularized by vessels from the perichondrium of the neck. Much more extensive vascularization of the femoral head and neck is noted in 85- and 95-mm fetuses. The perichondrium is evident as a distinct condensed layer of cells enclosing not only the extracapsular part of the cartilages but also the intracapsular portion continuous with the chondrogenesis layers of the blastema. The interzone primarily is the blastemal disk between the cartilages but later is continuous at its periphery with the synovial mesenchyme. Later the interzonal and synovial mesenchyme tissue gives rise to the various intra-articular structures and articular cavities formed within it. The rim of cartilage tissue associated with the fibrous glenoid labrum is a constant finding in fetuses 140 mm and larger and is continuous with the cartilaginous glenoid labrum. 9. SUMMARY After 8-9 weeks of embryonic development, the general form of the hip joint resembles that of the adult. As early as 1878, Bernays concluded that the development of the knee and hip joints up until this time is determined genetically rather than being due to mechanical features (17). Most observers feel that the tissue cut off by the developing hip capsule persists as the synovial mesenchyme from which the intra-articular structures arise. This tissue is continuous with the interzone. Intra-articular structures arise in situ. The majority of studies also support the fact that the general form and major components of joints are present before cavity formation begins. Smaller irregular spaces are present at the periphery of the joint at 8-9 weeks of age, and these rapidly coalesce to form a singular articular cavity. The appearance of the joint cavity marks the beginning of the period designated by Bernays as the stage of completion. The formation of synovial tissue takes place soon afterward. Concavity of the acetabulum is apparent throughout the period of development from 8 weeks to term. The future acetabulum is indicated as early as 13-15 mm and the acetabular fossa by 22-25 mm. The femoral head and neck are present in 22- to 25-mm specimens and both trochanters are shaped by 28 mm. The interzone is seen between the anlagen of the femur and acetabulum. By 20 mm its middle portion becomes thinner and lighter in cell density and a three-layered arrangement is evident. The outer layers are continuous and serve as chondrogenous layers for the adjacent cartilage. The intermediate layer is continuous with the adjacent synovial mesenchyme, which becomes intra-articular in position with formation of the capsule and in which clefts appear by 2 2 25 mm. Cavities begin to appear at this time first in the synovial mesenchyme. Cellular condensations for the capsule make their appearance in some regions of the joint as early as 20 mm. Ligamentum teres and glenoid labrum appear as cellular condensations at 22-25 mm and the transverse acetabular ligament by 28 mm. All three of these

9

i ii i i iilil ililiiliiilii!!iii !iii!i! ~! iii i~! i ~iiliil~i~ii~iiiiiiiiiiiiiiiiiiiliiii!iiiiiilliiiili!iilii!?ili!iililiiiili!!i!!ii!i~ iiiii~iiiliiliii~i~ iiiiiii~i!iil !iiilii~iiii~ilil84184 i~i!~!iiiiiii~iliiiiiiiliiiii~iii~iii~!~i iiiiiii iiiiiiiiill !!ii!!!i i!ii!ii~i!i!~ii i~!iii ili!!ii!!!iiii~!iiiiiiii~i!~ ii!i!i~iii!i84184184 i/~84

i i/ili ~i~!~i~!~iii~ii~iii!~!~!ii~i~!ii~i!iii~iii~i~i~ii!iiii!i~iiii~iiii!ii!ii!~ii~!~/~!!~ii~!iii/~iii~ii iiiiiiil84 iii ~ i ! i i i~ i

Fibrous capsule Acetabular (Glenoid) Orbicular zone lbular fossa

Epiphyseal plates

'g o*o OO O 'o O*o'W-"Synovial pad of fat

PB

o

~o

eo o o o 0

Lig. of head of femur (Lig. teres femoris)

Acetabular labrum & transverse lig.

F I G U R E 1 The basic structures of the hip joint are outlined. (A) Anterior view of the hip joint demonstrates the iliofemoral ligament sometimes referred to as the Y ligament of Bigelow. It has an important structural role supporting the hip articulation and limits extension beyond the normal range. (B) A posterior view of the hip joint shows the ischiofemoral ligament and above the superior and posterior part of the iliofemoral ligament. The lower margin of the ischiofemoral ligament is an almost discrete structure referred to as the orbicular ligament by others. It is particularly evident on normal hip arthrograms. The synovial protrusion at the lower margin of the orbicular ligament (another arthrographic finding) is shown clearly. Note that the capsule and ligaments of the hip joint insert more distally anteriorly along the intertrochanteric line compared to their posterior insertion seen here, which leaves the most distal portion of the neck extracapsular. (C) At puberty, the depth of the acetabulum is increased further by three secondary ossification centers at the periphery of the acetabular cartilage. The os acetabuli (OA) is the epiphysis of the pubis and helps form the anterior wall of the acetabulum. The acetabular epiphysis (AE) is the epiphysis of the ilium and forms a major part of the superior wall of the acetabulum, whereas a third smaller epiphysis in the ischium also is formed. (D) Coronal section drawing illustrates the main features of the developing hip. Note in particular the lateral extension of the acetabulum by cartilaginous and fibrocartilaginous-glenoid labrum tissues. Note the insertion of the capsule laterally and superiorly above the acetabular labrum and cartilage onto the side of the ilium. This recess is a normal anatomic feature and is well-featured in a normal hip arthrogram. A similar attachment of the capsule inferiorly beyond the acetabular labrum is seen. Medially, the floor of the acetabulum is covered by fibrous fatty tissue, the synovial pad, and the origins of the ligamentum teres, leaving the articular cartilage to be present superiorly, posteriorly, and laterally. [Parts A, B, and D reprinted from J. C. B. Grant (1972), "An Atlas of Anatomy," 9 Lippincott Williams & Wilkins, with permission. Part C reprinted from (233), with permission.]

158

CHAPTER

3

~

Developmental Dysplasia of the Hip

structures arise in situ. Vascularization of epiphyseal cartilage begins by 50 mm.

B. Embryonic, Fetal, and Postnatal Development of the Femur Development of the normal shape of the proximal femur is very important clinically. There are frequent structural abnormalities at this area, and the hip is the site of a large number of congenital and acquired deformations. Felts studied 53 femurs from crown-rump (C-R) lengths of 31-485 mm (9th embryonic week to term plus 1 infant of 3 weeks) (74). The assessment of several developmental dimensions demonstrates the dynamic aspects of growth that involve not only increases in length and width but also changes in shape and angular position in particular at the proximal end. The joint cavities form at the hip at 30 mm C - R length and at the knee at 33-37 mm C - R length. He assessed 15 different growth parameters, the following of which are particularly important to an understanding of developmental dysplasia of the hip: (1) diameter of head, (2) inclination of head and neck angle with the determined longitudinal axis of the femur (head-neck shaft angle), and (3) torsion (anteversion or retroversion), involving the angle relating the proximal to the distal segments with the proximal angle along the head-neck axis and the distal along the posterior intercondylar plane. Felts measured the angular dimensions involving proximal femoral torsion of the head-neck shaft and inclination in a large series of femurs. Values for torsion of the adult femur, by which is meant the angular displacement of the head and neck anterior (anteversion) or posterior (retroversion) to the frontal plain as measured in relation to the bicondylar axis of the distal femur, have been reported extensively. Of greater importance than the actual degree of angulation is the change during development from embryonic to fetal to postnatal periods. Generally there is no anteversion initially in the embryonic and early fetal period; anteversion of the head and neck increases in the few months before birth, decreases appreciably during the first year of life, and continues to decrease until skeletal maturity. There is a marked tendency for diminution of the antitorsion or anteversion angle in the postnatal period. The femoral torsion is much more variable in the late prenatal and early postnatal period than in the adult. From the data presented in several series, femoral torsion has a perinatal value of 30-40 ~ with anteversion decreasing to 12~ in adulthood. Several studies have shown that the torsion actually is negative (retroversion) in many in the embryonic and early fetal stages. In studies measuring torsion with embryo lengths less than 29 mm C-R, values reported were - 1 0 ~ - 9 ~ - 6 ~ - 4 ~ - 9 ~ - 1 0 ~ - 4 ~ - 2 2 ~ and - 1 5 ~ Le Damany observed no torsion (0 ~ in the early fetal period with an increase toward term (164, 165). Von Lanz showed a mean anteversion of 10~ at 4 months fetal life, increasing uniformly to 34 ~ at birth and diminishing throughout child-

hood and adolescence to a mean value of 11 ~ at maturity (331). Torsion at any given fetal time shows a very high degree of variation between individuals, and its increase in the prenatal period is even greater than its postnatal decrease. Increase in torsion is characteristic of the late fetal prenatal period in all studies. Felts refers to series from 1879 forward with reasonably comparable adult data with mean numbers for anteversion presented of 11.67 ~ 11.33 ~ (fight), and 14.07 ~ (left), 11.63 ~ (fight) and 14.71 o (left), 11.76 ~ (fight) and 9.73 ~ (left), and a bilateral average of 11.23 ~ and 8.02 ~ (74). In virtually all studies the standard deviations are fairly wide, although the differences between fight and left are very small. In one of the largest series using a standardized technique, Fabry et al. documented anteversion radiologically in 864 hips in 432 normal children from 1 to 16 years of age (68). The average measurement was 31.13 ~ at 1 year of age with gradual diminution to 15.35 ~ at 16 years. Harris defined a normal range of anteversion from 35 ~ at birth decreasing to 11 ~ in adulthood (110). Hoaglund and Low studied Caucasian and Chinese proximal femoral anteversion in 294 adult cadavers (120). The average anteversion in 112 male Caucasian femurs was 7 ~ (range = 2-35~ in 31 female Caucasian femurs it was 10~ ( - 2 ~ to 25~ in 116 male Chinese femurs it was 14~ ( - 4 ~ to 36~ and in 35 female Chinese femurs it was 16~ (7-28~ Other mean adult values reported in the older literature are 15.3 ~ 14.3 ~ 11.6 ~ and 11.9 ~ Efforts to measure anteversion in patients are still fraught with problems of accuracy and interobserver variability. Plain radiographic techniques, computerized axial tomography, and clinical examination of hip internal and external rotation in extension all provide an indication of the extent of anteversion, but little appears to be gained clinically from precise radiographic documentation (293). A major question has always been where the torsion occurs between the proximal and distal femoral epiphyseal regions. Pitzen felt that torsion was not restricted to the proximal region of the fetal femur but was present throughout its length (228). Felts also was of the opinion that in the femur torsion is not a feature restricted to a localized area but is present throughout most of the shaft (74). This is different from the humerus in which torsion occurs at the junction of the proximal epiphysis and metaphysis. The head-neck shaft angle of inclination of the proximal femur as measured in the anteroposterior plane has also been studied extensively. The normal neck shaft angle also decreases with age, but the range of variability is considerable. Harris measured the average as 137 ~ at birth, 145 ~ at 1-2 years, 143 ~ at 2 - 4 years, 135 ~ at 4 - 6 years, 134 ~ at 6 - 8 years, 133 ~ at 8-12 years, and 120-125 ~ in adulthood (110). Von Lanz showed a similar pattern of change, with the highest mean angle of inclination of 145 ~ at 1-3 years of age followed by diminution to a mean of 126 ~ at skeletal maturity (331). The angle was 135 ~ at 4 months fetal, decreasing to the lowest prenatal value of 122 ~ at 8 months of

SECTION II 9 Development of the Hip fetal life. Adult values documented are similar to studies from the late 1800s to the mid-twentieth century, indicating values of 124 ~, 126 ~, 129.6 ~, and 126.4 ~. Humphrey published one of the earliest detailed studies in 1888 (127). The average angle in 30 adult femurs was 124 ~ (range = 113135~ in 15 additional femurs from individuals greater than 70 years of age it was 123.7 ~, and in 30 adult bones from Germany it was 128 ~. He also documented the higher, more valgus, neck shaft angle in patients with diminished to absent weight bearing capability and in fetuses. Hoaglund and Low noted adult angles of 135 ~ in Caucasian and Chinese femurs with no sex variance (120). Measurement of this angle in the fetus is difficult because of the variable amount of antitorsion, which tends to increase the angle in the anteroposterior plane, the relative shortness of the neck in the fetus compared with the postnatal period, and the difficulty of measurement because radiographic studies cannot be used due to the lack of bone in the femoral head. The proportion of growth at the ends of each of the major long bones is well-established for the postnatal period, with 30% proximally and 70% distally. Assessment in the embryonic and fetal periods is not as well-studied, but there appears to be a more nearly equal contribution of growth from each end of the femur.

C. Embryonic, Fetal, and Postnatal Development of the Acetabulum The acetabulum forms initially as a cartilaginous mass of tissue differentiated from the mesenchymal blastema. Laurensen studied the development of the acetabulum in the fetus using arthrograms, plain radiography, gross inspection, and histology in 14 fetuses from 14 weeks of age to full term (160). At 14 weeks, the acetabular roof is entirely cartilaginous and a well-formed labrum, referred to by Laurensen as a limbus, of characteristic shape projects laterally where it is separated from the joint capsule by the typical lateral recess. The zona orbicularis fits closely around the neck of the femur and the ligamentum teres is present, as are the acetabular fossa and transverse acetabular ligament. The bony roof of the acetabulum is beginning to form but is much less extensive than in the newborn. The acetabular bone becomes more prominent with time but the basic relationships remain unchanged. Lee et al. studied acetabular development from 6 to 20 weeks of gestation. Acetabular anteversion changed little in the early fetal period (169). The cartilaginous femoral head in the fetus is almost spherical and fits deeply into the acetabulum. There is no marked change in the relative size of the acetabulum and the femoral head with early fetal development. As development proceeds, however, Laurensen showed in the two oldest specimens, one at 300 mms C - R length and one at term, that there was a clear increase in the size of the femoral head in relation to that of the acetabulum. Measurements from the

159

90-mm period onward were essentially equal for the acetabulum and femoral head in relation to the depth of each, but at the latter two time periods the femoral head was relatively larger than the acetabulum at 11.0 to 8.5 mm and 10.0 to 8.5 mm. Le Damany felt that these differences served as a possible precondition for subluxation and dislocation in the newbom (163-167). The relative shallowness of the acetabular socket in late prenatal and early postnatal development has also been noted by Ralis and McKibbin and is one of the possible predisposing causes of neonatal hip instability (246). The more lateral parts of the acetabulum grow both by endochondral ossification and, at the inner and outer iliac cortices, by periosteal intramembranous bone that, similar to what is seen in the long bone, always forms slightly in advance of the contained endochondral bone. At birth, ossification of the acetabular roof has not yet completed its progress from the primary center of ossification at the greater sciatic notch. From that center, ossification spreads inferiorly toward the triradiate cartilage, anteriorly toward the anterior inferior iliac spine, and at later times laterally toward the limbus. In terms of the developing and advancing perichondrial bone in the pelvis, the advancing edges are the two lowermost edges of sheaths of bone, one on the inner and one on the outer surface of the ilium. With hip dysplasia, the lateral superior acetabular spur is less prominent than the medial spur and endochondral bone formation in the lateral part of the roof lags behind that in the medial part. Abnormal pressure of the laterally, proximally subluxed femoral head directly on the limbus and indirectly on the lateral spur or edge of perichondral bone and the adjacent acetabular cartilage retards cartilage development laterally and secondarily retards endochondral ossification in the lateral part of the acetabular roof and perichondrial ossification of the outer iliac wall. Diminished pressures on the acetabular cartilage associated with a completely dislocated head also disturb the normal developmental sequence laterally. Ponseti studied postnatal acetabular development by histologic and radiographic techniques in 10 normal, full-term infants and 3 children 7, 9, and 14 years of age (234). In infancy the cartilage of the acetabular socket is continuous medially with the triradiate cartilage. The acetabular cartilage forms the outer two-thirds of the acetabular cavity, whereas the ilium above the horizontal flange, the ischium below it, and parts of the triradiate cartilage form the medial wall of the acetabulum. The pubis actually is separated from the acetabular cavity by intervening cartilage. Fibroadipose tissue, referred to as the pulvinar, is interposed between the femoral head and the nonarticular depth of the acetabulum. The fibrocartilaginous labrum is at the peripheral margin of the acetabular cartilage and the joint capsule actually inserts several millimeters above the most peripheral rim of the labrum into the fibrous tissue coveting the outer surface of the acetabular cartilage. At the inner and outer margins of the ilium a characteristic perichondrial groove of Ranvier forms as the intramembranous periosteal bone extends slightly

160

CHAPTER 3 ~ Developmental Dysplasia of the Hip

beyond and covers the physis of the acetabular cartilage. In the postnatal period, bone formation gradually increases at the expense of acetabular cartilage. The acetabular cartilage preceding development of the bony acetabulum originally contains elements from the ilium, ischium, and pubis. At the medial depth of the acetabulum these three cartilage growth plates intersect to form the triradiate cartilage, which is composed of three linear components: one anterior and slanted superiorly, one posterior and oriented horizontally, and one oriented vertically (Fig. 1C). The medial wall of the acetabulum thus is formed by the ilium above the horizontal flange and the ischium below the horizontal flange plus parts of the triradiate cartilage itself. Interstitial growth within the triradiate cartilage allows for expansion of the hip socket during growth. At puberty the depth of the acetabulum is increased further by three secondary centers of ossification at the periphery of the cartilage (Fig. 1C). The os acetabuli, the epiphysis of the pubis, forms the anterior wall of the acetabulum; the acetabular epiphysis, the epiphysis of the ilium, forms a major part of the superior wall of the acetabulum; and a third small epiphysis in the ischium is also formed. The concavity of the acetabulum develops in response to the presence of the spherical femoral head. This is evident in cases of hip displacement in which acetabular development is correspondingly abnormal. It has also been shown in a more controlled environment by Harrison, who observed that following excision of the femoral heads in rats the socket failed to develop in terms of depth and area. The growth of the acetabulum involves both interstitial growth of the acetabular cartilage, appositional growth from the perichondrium at the periphery of the cartilage, and eventually intramembranous periosteal new bone formation at the acetabular margin, much as occurs in the developing cortex of a long bone (57). Walker studied histological development of 74 paired acetabulae from normal human fetuses from 12 to 42 weeks (335). At a gross morphological level the labrum is noted to contribute to a minimum of one-fifth of socket depth and often more. Histologic sections show the labrum to be increasingly fibrous, as distinct from fibrocartilaginous, the closer the fetus is to full term. Cartilage cells intermingle with fibroblasts primarily at the acetabular cartilage-labrum junction. Histologic sections from the developing superior quarter of the acetabulae show bone development beginning from the medial ischial side and the posterior areas adjacent to the sciatic notch. The bone of the superior roof develops first with that of the walls of the socket following. The posterior and medial bone development of the acetabular socket precedes the anterior and lateral bone development. Severin, quoting from Faber as well as his own observations, noted the following arthrographic criteria of a normal joint: (1) the labral thorn should lie under or possibly early on 1-2 mm above the horizontal " Y " line of Hilgenreiner; (2) the cartilaginous acetabulum should cover at least one-

half of the femoral head; (3) there should not be a great quantity of contrast medium in the bottom of the acetabulum; and (4) the shape of the head of the femur should be practically spherical (278). The limbus (labrum) lies lateral and slightly superior to the head of the femur. All of the articular structures identifiable on the arthrogram of the normal hip of the young infant, including the labrum and capsule, are present in the 14-week-old fetal hip as well. Damage to the triradiate cartilage, though extremely rare clinically, can occur and leads to development of a shallow acetabulum, a shortened hemipelvis, and lateral subluxation due to acetabular dysplasia (106, 230). It can occur with infection, trauma, and surgical damage. The bone and soft tissue components of the developing hip are shown in a coronal section drawing (Fig. 1D). Pertinent elements of hip formation are summarized in Table I.

III. P R I M A R Y E T I O L O G I E S OF HIP M A L D E V E L O P M E N T Many primary disorders can cause the hip to be subluxed or dislocated in utero, in the immediate postnatal period, in infancy, or in childhood. The effects of that displacement are dependent on the primary cause, the time during the developmental period that the displacement occurs, the age at which it is detected, and the effectiveness of treatment. In the newborn period, the distinction of an abnormal hip currently lies between (1) an idiopathic developmental dysplasia of the hip or (2) a teratologic hip dysplasia caused by an associated disorder that would abnormally affect hip development prior to birth. An idiopathic developmental dysplasia of the hip implies a perinatal hip displacement in an otherwise normal child presenting as a subluxatable or dislocatable hip due to an isolated hip capsular laxity with proximal femoral and acetabular changes occurring secondarily. A teratologic hip dysplasia implies the existence of developmental abnormalities other than capsular laxity, which originate in the embryonic or early fetal time period, generally are more severe leading to a subluxed or dislocated hip prior to birth, and are associated with evident neural, muscular, or connective tissue disorders. Recognizable teratologic abnormalities leading to hip dysplasia in the newborn are connective tissue disorders such as the skeletal dysplasias, joint laxity syndromes (Ehlers-Danlos), a wide array of dysmorphic syndromes most of which are chromosomal or genetic in origin, and spinal dysraphism syndromes, including meningomyelocele. Currently there appears to be an underappreciated spectrum of teratological disorders ranging from the obvious association of hip dislocation with gross structural abnormalities of other regions to what may well be localized hip region abnormalities of embryonic or early fetal derivation in an otherwise normal appearing child. A small percentage of dislocatable hips in the newborn are associated

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

TABLE I

Embryonic, Fetal, a n d Postnatal Features o f Hip D e v e l o p m e n t

Hip joint structures all form by differentiating

in situ from one mass of undifferentiated mesenchymal cells. The femoral head and acetabulum reach infantile shape prior to joint space cavitation such that dislocatioin cannot occur in the embryonic period.

Proximal Femur Anteversion

Neck shaft inclination

Early fetal period 0~ (neutral version) with many studies showing retroversion; middle-late fetal period with increasing anteversion to --~30-35~ at birth; postnatal fairly rapid decrease birth to 3 years of age, gradual decrease thereafter to --~10-12 ~ at skeletal maturity. Maximal at fetal stages of 150~ -v140-145 ~ at birth; progressive diminution postnatally to --~120-125 ~ at skeletal maturity.

Acetabulum Many studies (but not all) show late fetal growth of the femoral head to be relatively greater than that of the acetabulum causing slight acetabular shallowness at birth. Acetabulum forms from iliac, ischial, and pubic cartilage masses with triradiate cartilage in depths of acetabulum. Initial acetabular roof bone forms from the posterior, medial region of ilium adjacent to sciatic notch; endochondral ossification then proceeds in anterior, inferior, and eventually lateral directions. Radiolucent roof over femoral head medial to lateral: acetabular cartilage, fibrocartilaginous labrum, capsule. Acetabular obliquity (anteversion) remains unchanged during development--it ranges from 15 to 30~ but averages 20~

with unrecognized neuromuscular disorders, generally myopathies, which may not be diagnosed until several years have elapsed. There may well be an overlap between mild and currently undetected teratologic abnormalities localized to the proximal femur and acetabulum, which thus date to the embryonic and early fetal time periods and predispose one to perinatal subluxation and dislocation and which even at pathoanatomic assessment appears as a capsular laxity and not as the primary bone or cartilage abnormality. In addition, there may be associated soft tissue abnormalities in a teratologic sense that are present throughout the fetal period and render the hip nonresponsive to routine early therapy. The terms idiopathic and teratologic are somewhat imprecise but at present provide for a reasonable differentiation of cause and a good indicator of expected responses to basic treatments. Many neuromuscular abnormalities cause an imbalance in the hip musculature and the development of positional abnormalities postnatally over several months to years. The

161

cerebral palsies, spinal muscular atrophies, and myopathies often present with hips normally positioned at birth, which proceed to develop subluxation and dislocation due to muscle imbalance and delayed or abnormal gait patterns. Proximal muscle weakness, prolonged nonambulation due to neurological or neuromuscular disorders, and even limited ambulation with an inability to run all predispose a child to retention of the newborn and first year proximal femoral characteristics of increased anteversion and inclination angles. Both of these femoral structures predispose one to subluxation and dislocation tendencies. If adductor muscle tightness is also present, such tendencies increase. The structural development of the proximal femur--including particularly the normal diminution of anteversion and inclination anglesmis highly dependent on normal gait patterns. Similar findings are seen in some connective tissue disorders, such as Down syndrome in which ligamentous laxity is prominent.

IV. E T I O L O G Y A N D P A T H O A N A T O M Y OF DEVELOPMENTAL DYSPLASIA OF THE HIP In this section we will refer in detail to descriptions of pathologic studies and only briefly for reasons of historical perspective to theories of pathogenesis that have been proposed with little or no experimental or observational backing or subsequent verification. We will clarify as much as possible whether the observations made were representative of idiopathic developmental dysplasia of the hip, by which is meant a hip disorder in the absence of any other apparent abnormality, or teratologic dysplasia in which other abnormalities were found. Such distinctions often were not made by the original authors but they remain important. Many pathologic dissections that purport to describe the underlying pathoanatomy of idiopathic developmental dysplasia of the hip have been made on babies who were either stillborn, died in the newborn period, or lived for only a few months, such that some descriptions would appear to represent teratologic hips rather than an isolated idiopathic developmental dysplasia of the hip. Our incomplete knowledge of the etiology of hip dysplasia may indicate that this distinction is arbitrary but it is one that has persisted for some time.

A. Early Clinical-Pathoanatomic Descriptions 1. PALLETTA,1820 The earliest detailed pathoanatomic description of congenital dislocation of the hip is by Palletta of Milan, Italy (221). In descriptions published in 1783 and again in 1820, he referred to 5 cases. Some were observed in clinical patients but he also included assessment of a bilateral dislocation of the hips in a child who died at 14 days of age. Both femoral heads, which remained spherical, were situated above the acetabulum but had not yet been surrounded by

162

CHAPTER 3 9

Developmental Dysplasia of the Hip

any new socket. They were present in the area of the anterior inferior iliac spine. The acetabulae were filled completely with a fatty substance, the anterior portion of the socket was closed by the acetabular ligament, which had been turned on itself, and the joint capsule was much broader and looser than normal and moreover was stronger and very thick. The intra-articular ligament was longer than normal. Examination of the enlarged capsule and lengthened ligaments allowed one to appreciate the increased movements of the femoral head in several directions. Palletta's work established the existence of the hip dislocation in a child of only 14 days of age, which, based on the pathoanatomy, clearly was of congenital origin. Delpech referred to and quoted extensively from Palletta's work, which was written in Latin, in his book De l'Orthomorphie, published in 1828 (55). 2. DUPUYTREN,1826 One of the major early clinical descriptions of congenital dislocation of the hip was made by Dupuytren (62). He clearly defined the entity in terms recognizable today, discussed the possible underlying causes, distinguished it from other abnormalities of the hip, and commented on possible treatment methods. Up until that time it had not been widely recognized as a specific entity, and his work stands as an outstanding contribution to its understanding. He indicated that the disorder involved a transposition of the head of the femur from the acetabulum onto the external wing of the ilium, a transposition that one observed from birth and that appeared to be the result of a defect in the acetabulum, which was not as deep or as complete as normal, rather than being due to trauma or disease. The disorder belonged to the class in which dislocations of the femur were upward and outward. Some practitioners previously had been aware of two varieties of dislocation defined as being traumatic or pathologic. Dupuytren set out to describe the third type, which he called original or congenital dislocation to distinguish it from the other two. He defined the displacement of the head of the femur upward and outward, shortening of the involved limb, rising of the head of the femur onto the external iliac wing, prominence of the greater trochanter, shortening of almost all of the muscles of the superior part of the thigh toward the iliac crest where they formed around the head of the femur, the uncovering of the ischial tuberosity due to the displacement of the muscles, rotation of the limb internally, displacement or telescoping of the thigh up and down and inward and outward, a slanting or obliquity of the femur, which was greater the larger and older the patient, a sharp angle of the thigh with the pelvis, and a thinness of the entire limb in particular at its superior part. Dupuytren noted the limited movements of the involved extremity, especially those of abduction and rotation, all of which led to difficulty with standing, walking, and all forms of lower extremity exercises. The involved lower limb was atrophied in relation to the opposite unaffected side and to the trunk and upper

extremities. The pelvis was large and prominent. He noted the lumbar lordosis and the horizontal position of the pelvis on the femur due to the displacement. He described what would later be referred to by Trendelenburg as the waddling gait. Dupuytren distinguished the congenital dislocation from traumatic dislocations or dislocations due to disease, defined the congenital case as showing frequent bilaterality and a lack of swelling, abscess formation, fistulas, or evidence of scarring, and noted that people affected with congenital dislocation did not experience any discomfort in childhood of either the hip or the knee but rather experienced fatigue and numbness with walking. The clinical signs were due to elevation of the head of the femur to the iliac fossa and shortening and prominence of the muscles drawn toward the iliac crest. Dupuytren noted that if attention was called to the disorder early, at birth, several clinical signs were already present: widening of the hips, prominence of the greater trochanters, and obliquity of the femur. The tendency, however, was for patients to present for assessment only after they had begun walking when awkwardness of gait was noted. In many individuals, presentation for medical attention was not until 3 or 4 years of age. Basic understanding of the disorder was limited because there was little opportunity for pathologic study as the patients were otherwise well. Dupuytren did study a few instances, however, and reported on them. The muscles were always pulled up toward the iliac crest; some were remarkably developed but others were thin and atrophied. The hypertrophic muscles were those that continued their activities; others had their activities limited due to the change of position and were often so fibrotic and yellow that one could scarcely note any muscle tissue persisting. The superior part of the femur for the most part preserved its form, although on occasion the internal and anterior part of the head of the femur lost its roundness slightly primarily due to positioning against parts that were not particularly organized to receive it. The acetabulum of the iliac bone either was absent completely or offered only a trace of a small bony prominence, which was irregular and in which there was no trace of articular cartilage or a synovial capsule. It was filled with a fibrous border that generally was composed of resistant cellular tissue covered by muscles inserting onto the lesser trochanter. In two or three subjects that Dupuytren examined, the round ligament (ligamentum teres) was greatly lengthened, flattened superiorly, and worn away in certain areas due to pressure and rubbing of the head of the femur. The head was lodged into a cavity analogous to that developed in femoral head dislocations following trauma; the new cavity was superficial, almost bereft of a rim, and was situated in the external iliac fossa above and behind the acetabulum, a position that was proportionate to the shortening of the limb or, put another way, to the elevation of the head of the femur. In summary, the findings in these subjects were similar to those seen in cases of pathologic or traumatic dislocation, with the difference being that those Dupuytren examined had

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip findings that appeared to date from a more remote time and to have been so positioned from the earliest time of life. Dupuytren, considering the possible causes of displacement, listed: (1) displacement due to associated illness during the fetal time period contracted from the mother but that had nevertheless healed prior to birth; (2) the effect of trauma that had displaced the head of the femur from the acetabulum, after which the hip developed abnormally because the head of the femur was unable to function normally; and (3) due to maldevelopment of the acetabular socket as an evolutionary problem in particular because the socket was a somewhat complex union of three pieces of bone. He did not place any faith in the first theory of prenatal disease. The second possibility of a force that caused the head of the femur to displace from the cavity was felt to be credible. He felt that this was possible due to the position of the fetus in utero, which was one of marked flexion of the lower extremity that forced the head of the femur continually against the posterior and inferior capsule of the joint causing a continual strain. This strain was without an effect in healthy individuals but perhaps caused a problem in others not as wellconstituted whose tissues were less resistant. This relatively weakened region, therefore, would be susceptible to allow passage of the head of the femur out of the socket allowing a dislocation to happen. The final possibility of maldevelopment of the acetabulum was also feasible; such a possibility was related to embryologic and anatomic studies on embryo-fetal development, which indicated that the final regions of hip development were those of the joint cavities and in particular those in which several regions of the bone were required to unite such as in the acetabulum. It was known that the acetabulum was composed of three separate segments and that the formation of this cavity appeared to be one of the final aspects of hip bone development. The development of the cavity did not form in keeping with that of the femur, which thus was displaced to the external region of the ilium. Dupuytren indicates that "in the 3 preceding hypotheses, the displacement of the head of the femur was not only congenital; in each that we have come to examine it was original and dated from the first organization of the parts. It is a defect of original conformation, a defect in the organization of germs." 3. SEDILLOT,1835 Sedillot discussed dislocation of the femur and its presence on the external iliac fossa (274). He defined the formation of the new capsule in a dislocated hip, which he recognized as being continuous with the old such that the new articulation on the side wall of the iliac fossa was in continuity with the previous acetabulum. He commented on the production of the false acetabulum and indicated that the view that congenital dislocations were due to the absence of the acetabulum was incorrect. He felt that with his work he had established that "the most frequent cause of congenital displacement of the femur is the looseness and relaxation of

163

the ligaments which remained intact and allowed the great mobility of the thigh." Sedillot described the late pathoanatomy of two hip dislocations, detailing cadaver studies on a young woman with bilateral hip dislocation who was between 20 and 25 years of age and a woman of 35 years of age with a unilateral dislocation. In the first patient he noted that the capsular ligament had remained intact and that the original acetabular cavity, which no longer contained the femoral head, was triangular in shape and filled with synovial tissue. The round ligament was thin and flattened. The femoral head was smaller than normal and flattened in that region that corresponded to its relationship to the iliac bone. It had entirely lost its former spherical shape. The femoral neck was quite short. The appearance was felt to be indicative of a complete relaxation of the ligamentous apparatus, which thus played a causative role of extreme importance. The dislocation or at least the disposition to it was present at birth and thus was congenital, and many of the subsequent abnormalities were secondary due to the displacement of the femur. The appearance of the original acetabulum, which was smaller and shallower than normal, could easily be explained by the long duration of the dislocation and by the fact that any bony cavity would become obliterated when it did not contain the body that it was naturally destined to contain. Similar findings were present in Sedillot's second dissection. The original acetabulum was smaller and not as deep as normal. The round ligament was intact. The femoral head was atrophied particularly at its superior part, being somewhat conical in shape and flattened against the iliac region. The femoral neck was short. The external iliac fossa had formed a cavity that was quite deep to serve as a false articulation for the femoral head. Sedillot concluded his work with 12 observations, most of which appear accurate today: (1) dislocation was longstanding; (2) dissection revealed the possibility of reduction; (3) there was proportional atrophy of the head of the femur and the acetabulum; (4) the head was accompanied by formation of a new capsular ligament, which was continuous and united with the old capsule in a manner that produced a large capsule embracing both the old and the new articulations and leaving free communication between them; (5) the persistence of the round ligament (ligamentum teres); (6) development of a new fiber cartilage on the portion of the external iliac bone, which formed the depth of the new false articular cavity for the femoral head; (7) deposit of bone on the external wing of the ilium surrounding the head of the femur in such a manner as to represent a new acetabulum; (8) formation of a new joint between the lesser trochanter and the anterior surface of the ischium; (9) alterations in shape that were, however, not as deep as the original in relation to the pelvis; (10) femoral atrophy that primarily involved thinness of the bone, which was less marked distally than proximally; (11) the description of the best positions to lead to repositioning of the dislocated bone into its natural cavity; and (12) the immediate reestablishment of movement

164

CHAPTER 3 ~ Developmental Dysplasia of the Hip

or articulations following reduction. The work of Sedillot, even though it was based on studies of young adults, pointed out the key role played by capsular laxity and secondary deforming changes in the appearance of congenital dislocation of the hip. 4. CRUVEILHIER,1849 Cruveilhier, one of the preeminent professors of pathologic anatomy in Paris, wrote on congenital dislocation of the hip, summarizing the findings of Palletta and Dupuytren and discussing seven cases he had observed (52). Most of his cases plus most of those in the literature were bilateral, which tended to rule out any specific traumatic episode. The capsule invariably was enlarged without rupture, again leading to nontraumatic etiology; it remained enlarged and thickened in the congenital form of dislocation. There invariably were abnormalities in development of the acetabulum, although the original socket could always be identified. The varieties of acetabular change were great, however. In some instances it maintained its size and shape reasonably well such that it would have been able to receive the femoral head if reduction could have been brought about. In other cases, the socket was obstructed by the adjacent ligaments, and often it was filled with a fibro-fatty tissue. The femoral head also took on changes both of size in the sense that it was smaller than normal and of shape, with flattening of one part of the surface of the head often seen. The round ligament or the ligamentum teres invariably was longer than normal. Cruveilhier recognized that during fetal life individuals who subsequently were born with a congenital hip dislocation had presented a normal conformation of the affected joint with the acetabulum and femoral head in perfect relationship. Some cause served to remove the femoral head from the acetabulum, and due to the laws of bone development the empty acetabular socket became narrower, deformed, and filled with fat while the head of the femur also became deformed due to the abnormal and unequal pressures to which it was subjected in its new position. The major problem, therefore, was to determine the cause of the dislocation in the hip that appeared to be developing normally. After much discussion that focused on the reality of an enlarged capsule and round ligament, he felt that external force placed on the fetus via uterine compression in association with certain intrauterine positions of the hip and limited amniotic fluid all conspired to lead to the dislocation. The external pressure was not considered to be a violent, one-time phenomenon but rather something lasting over a considerable period of time in association with the other findings.

5. ROSER, 1864, 1879 Roser, as early as 1864, pointed out that congenital dislocation of the hip resulted from an abnormal adducted position of the legs during fetal life (256). He "based this belief on observations in children whose flail hips could be dislocated by adduction of the leg and then reduced again by

abduction." In a second article in 1879, he again made these points and pleaded with his obstetrical colleagues to actually perform this test on the newborn child (257). He suggested "that children no longer be allowed to reach the age of 2 years before their hip dislocations are diagnosed." Not only was Roser accurate in his impression that most if not all cases of congenital dislocation of the hip could be diagnosed by a newborn exam, but he also recommended a treatment, not widely adopted for several decades, that could bring about reduction and cure of the disorder. He indicated that "I believe that many, even most, of these cases would still be curable if the disorder were detected in the newborn and if the necessary abduction appliance were applied at once. I believe that with plaster boots held apart by a cross bar or cross board the object would be most simply obtainable."

6. VERNEUIL,1852, 1890 Verneuil was able to study a fetus that died of respiratory problems at term. It appeared physically normal but showed a unilateral dislocatable hip (328). The external appearance of the child was normal except for the region of the left hip, which showed an appearance consistent with dislocation. The dislocation described was pathologic in association with a septic arthritis. The associated muscles, however, were shorter and less developed, although they otherwise appeared normal. The joint had lost its normal form. The capsule was markedly distended superiorly and contained the head, which was displaced upward and posteriorly. The joint was filled with a seropurulent fluid and the joint lining was thickened and thrown into folds. The anterior part of the articular surface of the femoral head was no longer convex. The head repositioned itself into the acetabulum with attempted reduction. On opening the joint, the capsule was elongated particularly superiorly and anteriorly. The round ligament (ligamentum teres) was intact and the internal surface of the capsule had developed an irregular surface. The articular surfaces were fully covered with cartilage, although there was some loss of translucency. The femoral head was somewhat smaller than that on the opposite side and was somewhat flattened at the point where it pressed against the iliac fossa. The acetabulum had not lost its primary shape or regularity but was somewhat flattened at its superior margin. The displacement of the head was secondary to, and in this instance had occurred prior, to birth. It was quite easy to reduce the hip by simple traction, and it was at the earliest stages that one should treat the disorder. The case was presented to indicate that in rare instances congenital dislocation of the hip could occur, but that it was associated with pathological states and did not occur as an isolated phenomenon. a. Hip Dislocation Due to Infantile Paralysis Verneuil continued his studies of childhood hip dislocation throughout a long surgical career and summarized his work in 1890 (329). His primary theory of hip dislocation, initially propounded in 1866, was that the large majority of hip dislocations seen in clinical practice and referred to as congenital

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip really represented dislocation after birth secondary to infantile paralysis. Although he recognized that occasional cases of hip dislocation occurred secondary to pathological intraarticular disease (one case of which he reported in 1852), he considered them to be extremely rare. Vermeuil and his colleagues had been searching during the course of their pathoanatomic dissections in newborns for isolated congenital dislocations over a period of several decades and subsequently had never seen the lesion in question. The conclusion that Verneuil reached was that "really the luxation did not exist," and therefore the term congenital was inaccurate because dislocation essentially was never seen at the moment of birth but was produced afterward. He did not accept the viewpoint that congenital dislocation in most cases was due to defects of hip development and that the dislocation was present at the time of birth but offered no signs to allow it to be detected; the head or the acetabulum was misformed but still appropriately related until they lost that relation at the time walking began. Verneuil reasoned that if there were features predisposing to delay the luxation, they should be demonstrable in either the bones, the ligaments or the muscles at pathoanatomic assessment. True congenital or intrauterine dislocations directly observed at the moment of birth were extremely rare, whereas the large majority of dislocations became evident around the beginning of the 2nd year of life and then increasingly during the course of the 1st decade. The extreme rarity of the truly congenital dislocation diagnosable at birth was further evidenced by the inability of Verneuil and several colleagues to detect any at newborn pathoanatomical assessment over an extended period of time. Such disorders did exist but only with "extreme rarity." In his clinical practice he had seen over 300 cases of dislocations referred to by others as congenital. He did not accept the delayed appearance of a dislocation in a predisposed, but slightly abnormal hip. He made the not unreasonable conclusion that if the anatomist did not find any hip dislocation in the fetus at full term at dissection, perhaps it really did not exist. In a semantic sense, if the dislocation was not observed at birth and only was diagnosed later, it was not truly congenital in the accurate sense of the word. Although Verneuil even then was swimming against the tide, his argument was not without merit. It addressed several medical and legal concerns in terms of the time of diagnosis of the disorder, and in a sense it has been incorporated into the current terminology of developmental dysplasia of the hip in which the profession now recognizes, at least on a partially scientific basis, that the displacement may not be either present or even diagnosable at birth but only appears later with increased stresses. Verneuil indicated that many adopting the congenital view in their writings continually were referring to rare isolated cases from the literature rather than performing newer studies of their own. He clearly recognized that severe teratological dislocation occurred, including those of proximal femoral focal deficiency and multiple congenital anomalies in stillborn fetuses. He con-

165

cluded his work by commenting on the accuracy of the paralytic luxation etiology. 7. BRODHURST, 1876 Brodhurst noted that the vast majority of the congenital hip dislocations were upward and outward with the head lying on the dorsum of the ilium (25, 26). A marked female preponderance was noted with a 3:1 female:male ratio. He discounted the theory of spasmodic muscular retraction, which had been offered by Guerin and Carnochan in separate works. He felt that "the cause of congenital dislocation of the hip, as it usually presents itself, is a purely mechanical cause." The dislocation occurred with nonroutine or difficult labor and especially with breech presentations. In this position, the head of the femur must press against the posterior and interior portions of the capsule of the joint so that traction in this position (associated with birth) will readily cause the head of the bone to escape from a shallow acetabulum. The dislocation was "produced at birth" through downward force supplied to the thigh in "endeavoring to hasten the birth in breech presentations." Some very rare instances of congenital dislocation had occurred in which the head of the femur was misformed, the cavity of the acetabulum was developed imperfectly, and other deficiencies and abnormalities existed, a situation that refers to what we now call a teratological dislocation. He indicated that the latter were rare and were quite separate from the types of congenital hip dislocation that he was discussing. In the idiopathic variety that he was discussing, children were healthy, well-developed, and well-nourished. a. PathologicalAnatomy In the idiopathic congenital hip dislocation, the acetabulum at birth was never altered in shape or dimension and the head of the femur retained its normal appearance. Changes took place, however, secondarily with persistence of the dislocation both in the acetabulum and in the head of the femur as the cartilage wasted and the acetabulum became filled with cellulo-osseous material, while the head of the bone became somewhat irregular in shape and its cartilage became thin. The capsular ligament retained its integrity but became elongated, and the ligamentum teres was stretched and eventually became slender and finally gave way. The head of the bone came into direct contact with the ilium. A false articulation was formed that ultimately developed a new capsule, whereas "a cavity is formed to receive the head of the bone by deposition of osseous matter upon the ilium." Brodhurst again indicated that "when dislocation occurs without other abnormality both the acetabulum and the head of the thigh bone are usually perfect at birth." He indicated that treatment at birth should be relatively simple but rarely occurred because the diagnosis was not made at that time. By the time the diagnosis was made after months and years had elapsed, secondary changes had already taken place that tended to impede reduction. In the third edition of his book, Observations on Congenital Dislocation of the Hip, Brodhurst (1896) defined four

166

CHAPTER 3 9 Developmental Dysplasia of the Hip

types of congenital dislocation, with type I being what we refer to today as idiopathic or developmental dysplasia of the hip (26). He again commented on the mechanical causation in which the high incidence of breech presentation and the trauma associated with delivery displaced the femoral head. The dislocation commonly was overlooked at birth and came to clinical evidence only when the child began to stand and walk. A waddle was seen in association with the tilted pelvis and lumbar lordosis. He recognized that when dislocation is recognized early, the head of the femur may be immediately restored and retained in the acetabulum. When it remains displaced for several years changes take place, such as retraction of muscle, filling up of the acetabulum, and some flattening of the head in the bone. These changes were most marked in those cases in which displacement had not occurred in utero; for the parts are fully formed so that neither is the femoral head so much flattened nor is muscular retraction so great; the acetabulum also is found to be fully developed. In this series not only is development complete, but there is no other abnormality present. The second type of hip dislocation was "produced in utero," being what we would refer to as teratologic. Brodhurst felt, however, in agreement with much then current thought that such dislocations occurred as a result of accident, shock, and spasmodic action. Others had referred to this type of development as arrested both in the acetabulum and in the head of the femur, but he suggested that "it occurs in consequence of the parts being displaced in utero before the development is complete." In the third category, he placed dislocation resulting from inflammation and destruction of the joint, which we would recognize as septic dislocation. In the fourth class were placed malformations such as those associated with spina bifida or club foot that we would recognize as a teratologic dislocation. Brodhurst clearly pointed out that dislocations had many causes and that the pathoanatomy was bound to differ in various types. After several years of dislocation, "the acetabulum becomes more or less filled up with fatty or with fatty and fibrous matter with retraction of the trochanteric muscle so that even though the head of the femur may be reducible difficulty will be experienced in retaining it in position." He again contrasted the neonatal dislocation produced at birth in which the femoral and acetabular components were normal with dislocation that occurred in utero in which the acetabulum and femur will not be fully developed. In those dislocations in utero in which development is imperfect, it depended on the stage of development at which the parts had arrived when dislocation took place as to whether there was hope of normal restoration. 8. SUMMARY OF THEORIES OF CAUSATION IN THE NINETEENTH CENTURY: REEVES, 1885

Reeves summarized well the theories of causation of congenital dislocation or malposition of the hip that had evolved

over several decades (248). In reviewing the theories and relating them to his extensive clinical practice, he reached the conclusion that "different cases have different etiology." In the process he reviewed many descriptions that had been given. The observations made, for the most part, are recognizable as accurate today, and even at that time Reeves noted that many of the descriptions that the authors felt related to the primary cause of the disorder were in fact secondary deformations. Dupuytren and others observed that the hip capsule was too large, which allowed for displacement, a view commonly held today as being a primary etiologic component (62). Sedillot also regarded the looseness of the hip joint ligaments to be a primary causative factor (274). Abnormal position of the fetus in utero was felt by many to lead to an abnormality of development, which itself produced the hip dislocation deformity. Among those of this opinion were Dupuytren, Cruveilhier, and Roser. It remains well-recognized today that the breech position in particular has a relatively high incidence of associated hip dysplasia, although that position is still associated with the minority of cases. A direct mechanical force was considered by many to be causative, for example, the view of Brodhurst in which the trauma of delivery led to the physical displacement of the hip (25). Reeves and others, however, felt there was insufficient evidence for that. Neuromuscular abnormalities were felt to be causative in many instances. Guerin, Carnochin, and others ascribed the disorder to muscular retraction associated with pathological conditions of the nervous system. Although many recognized associations with neuromuscular disorders, it was widely felt that muscular tightness was secondary, which is the usual belief today. Many, including Verneuil, implicated a neurological paralysis with atrophy of the hip muscles as being a cause of most instances (329). Many physicians at that time recognized infantile paralysis as being associated with hip dysplasia but thought that dysplasia occurred gradually over several years following birth, such that congenital paralytic displacement rarely was seen. Another group of physicians attributed the malformation to primary developmental abnormalities of the hip region, some placing these in the acetabulum, some in the proximal femur, some in the capsule, and some in the entire hip complex. Reeves also felt that this was a fairly common source of the dysplasia. Certainly these descriptions can be recognized as essentially descriptive of what we refer to today as teratologic hip dysplasia. The increased incidence of hip dysplasia in females was recognized very early and has been a consistent observation ever since. 9. SAINTON,1893 Sainton wrote a two-part study of the anatomy of the childhood hip and the pathogenesis of congenital dislocation of the femur (260, 261). a. Anatomy of the Childhood Hip Anatomic studies were performed on more than 30 hip joints with the largest number in children between birth and 1 year of age (260).

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip The study involved assessment of the external form of the hip joint, measurement of the relative dimensions of the femoral head and acetabulum, and examination of the interior of the bone on decalcified sections to assess the development of ossification. At 2.5 months the embryonic head of the femur had formed its regular shape and was contained by the acetabular concavity in the iliac cartilage, but as yet there was no indication marking the femoral neck or separating the head from the shaft. By 3.5 months the shape of the articular surface resembled that of the postnatal period but the neck was still quite short. Around 3.5 months the greater trochanter began to be shaped. During this time period of early and initial hip joint development, the presumptive joint area separating the primitive head from the acetabulum was filled with tissue rather than being an empty space. The cellular substance in the intermediate zone later became subject to a process of resorption, which passed from the central parts of the joint toward the periphery. During intrauterine life, the hip cavity was sufficiently deep to be able to contain the convex head of the femur. The peripheral fibrocartilage labral rim of the acetabulum also was developing. Important observations in view of subsequent hip disorders were that prior to birth the acetabulum was deep enough and the neck almost completely absent throughout the embryonic period such that the hip articulation favored retention of the anatomic position. Significant changes had occurred, however, by the time of birth and in the immediate years thereafter. The normal acetabulum was so constructed as to hold the femur in place. Three elements of difference were noted between the proximal femur of the newborn and that of the adult: (1) the head was relatively much larger in the infant than in the adult; (2) the neck was longer in the adult than in the child; and (3) the greater trochanter was more prominent in the adult than in the child. b. Femoral-Acetabular Articulation Studies in the infant indicated that the femoral head was relatively larger than in the adult during the first year of life, whereas the acetabulum was relatively shallow. Three epiphyseal regions formed in the upper part of the femur, these being the secondary ossification center of the femoral head, that of the greater trochanter, and that of the lesser trochanter. It was already known that the femoral head epiphysis was entirely intrasynovial in the developing hip, which explained the frequency of infection from the shaft regions and the rapid invasion of the articular cavity in cases of epiphyseal osteitis. During the first year of life a single uniform growth plate underlies the head-neck region and the greater trochanter. By 1 year, however, an angle is formed between the growth plate underlying the greater trochanter and that underlying the head-neck region. By 3 years of age the neck is almost completely ossified and the head is quite large. The secondary center of the greater trochanter appears around 3 years of age. c. Acetabulum The relative amplitude of the acetabulum in the infant is not nearly as great as that in the adult and,

167

thus, is slightly less able structurally to maintain the head. The triradiate cartilages are well-developed. The depth of the cartilage was measured in the studies, and the part of the head completely contained in the acetabulum was defined by cutring the head-neck-acetabular junction. This enabled measurements of the diameter of the segment of the head within the acetabulum and the portion augmented by the cartilage rim. The relationship between the head and the acetabulum thus was established. In the very young subjects aged 4 - 6 weeks the acetabulum was not yet very deep. At 36 days a similar relationship persisted, as seen shortly after birth, and the acetabulum was not able to receive the major portion of the head. In summary, the femoral head-acetabulum relationship at the time of birth presented a cavity that predisposed it to some instability. Comments then were made on the development of the triradiate cartilage. Three cartilaginous branches separate the points of ossification, forming the ilium, the ischium, and the pubis. The cartilaginous regions are almost as wide as they are long. The capsule itself is well-formed from birth and offers considerable stability. In summary, (1) the anatomic neck of the femur is very short in the infant; (2) the diameter of the neck, however, is relatively large in relation to its appearance in the adult; (3) the diameter of the head is relatively larger in the infant than in the adult; and (4) the acetabulum is not as deep in the infant as in the adult and the head is particularly contained by the posterior part of the cavity. The anatomy was such that luxation of the femoral head, though difficult in the adult, was relatively easy to produce in the infant due to the structure of the joint.

d. Pathogenesis of Congenital Luxation of the Femur Sainton then summarized studies of the previous several decades and provided a pathoanatomic assessment of three cases of dislocation, two of which were in newborns with teratogenic hips and one in a girl 12 years of age with a longstanding dislocation (261). Traumatic Theory of Congenital Hip Dislocation: Two types of trauma were postulated to cause hip subluxation, these being intrauterine trauma and obstetrical trauma at birth. By the late nineteenth century few maintained intrauterine trauma as a cause of hip dislocation, although until a few decades previously it was a view widely held. Sainton himself placed no faith in that theory simply because it failed to conform to the known facts. The concept of obstetrical trauma, however, was more commonly held, relating the occurrence of dislocation in association with traction at the time of birth, particularly with breech presentations. The occurrence of hip dislocation was felt to be much greater, however, than the occurrence of either breech presentation or excessively traumatic deliveries. Even at this time, however, it was becoming apparent to many that traumatic deliveries generally were associated with either epiphyseal separations or actual limb fracture rather than true joint dislocations. Inflammatory-Pathologic Theories: Two commonly held theories of causation involved hydrarthrosis and sepsis. The

168

CHAPTER 3 ~

Developmental Dysplasia of the Hip

dislocation was felt to be secondary to either intrauterine or postnatal hip joint inflammation. This theory evolved because there undoubtedly were many cases of septic arthritis of the hip, which did dislocate. Malgaigne thought that most luxations, which he felt were rather rare, were secondary to joint fluid accumulation or joint infection. These arguments, however, were soon combated in relation to the clinical situation. Sainton objected to the idea of dislocation associated with intra-articular fluid because such fluid had never been seen in the pathoanatomic specimens. In addition, cases of intrauterine sepsis that were known to occur were never associated with congenital dislocation. Myogenic and Neurogenic Causes of Dislocation: Throughout the first half of nineteenth century, abnormalities of the nervous system were felt to be associated with congenital hip dislocation. Delpech of Montpellier in particular stressed the importance of muscle development in relation to bone formation and developed the theory that bony, ligamentous, and muscular malformations were indicative of pathological anatomy and thus also an indication of embryonic maldevelopment. One specific theory of abnormality was the theory of muscular contracture of Guerin (1841), who felt that virtually all orthopedic deformities were caused by contractures. He placed such abnormalities as scoliosis, clubfoot, torticollis, and dislocation of the hip into the general law of etiology of orthopedic deformities, which for him was almost universal. He was one of the early practitioners of tenotomy for the correction of such deformities. He also noted pathoanatomic findings that allowed one to establish a frequent relationship between certain malformations of the central nervous system and those of the joints, with congenital dislocation of the hip being particularly prominent. Sainton felt that this theory depended too much on the assessment of patients with teratologic abnormalities because most of the patients with the congenital hip disorder were otherwise normal. Any muscle tightness was felt by Sainton to be secondary in the large majority of cases rather than being the primary cause of the disorder. The other neuromuscular theory was of infantile paralysis as presented in 1866 by Verneuil. The theory of infantile paralysis as the cause of dislocation of the hip in the newborn was hotly debated at the time. Verneuil felt that previous practitioners had studied only old cases of dislocation, but that in reality the dislocation was present from the intrauterine period and was caused by a partial paralysis of muscles in the peritrochanteric region. The argument was not well-supported. Sainton indicated that there were clear examples in which infantile paralysis led to hip dislocation, but these occurred well after birth and did not represent the large majority of cases, a view still accurate today. Theory of Primary Developmental Malformations: Sainton supported this theory as underlying the primary cause of congenital hip dislocation based on his own pathoanatomic studies and those of others, particularly Grawitz (97). He felt that, if during the course of operative management of such

dislocated hips (which was becoming increasingly common at that time) one did not find abnormalities of the hip or found only alterations that were insignificant, then it would be necessary to search for other causes of the congenital dislocation. He stated, however, that in the disorder "articular abnormalities were to the contrary quite pronounced and capable in themselves of leading to the displacement of the femur onto the iliac bone." Pathoanatomic abnormalities had been discussed and were reviewed. The alterations of form varied with age. The acetabulum was narrower and not as deep as normal. The head of the femur was larger; in other cases it was smaller and almost conical in shape. The round ligament was lengthened when it existed but was often absent; the capsule was elongated, deformed, quite large, and capable of receiving the displaced head. The neck often was shorter than normal and its direction was changed to increased anteversion. In a word, each of the regions of the hip was modified to some extent, and because these alterations were primary it seemed unnecessary to search for other causes of displacement; it was necessary simply to indicate the embryologic causes of these different malformations. Objections to this theory could still be raised, however, and the questions commonly asked still concerned whether these changes were primary or secondary and whether they preceded or followed the dislocation. If they were only an epiphenomenon, they would lose a great part of their interest and one could really not attribute to them a primary pathogenic role. If they were indeed primary, then the question to be asked next concerned their cause and whether they alone were capable of producing the dislocations or at least rendering them imminent. One problem that then existed was that most of the pathoanatomic studies had been done on subjects a few years old and had rarely been done on the newborn. Many authors, including Sedillot, had clearly noted that the pathological anatomy was composed of two types of findings: those that were primary and truly congenital and others that were secondary and occurred later. It was felt that, in those cases assessed early, the primary lesions found did not appear particularly extensive either in the acetabulum or in the femoral head, such that most of the changes described were indeed secondary. There was information available even at that time that in the newborn hip with dislocation the acetabular cavity itself looked good and the femoral head had no modifications. The round ligament almost always was intact and was never lacking. e. Pathoanatomy Detailed analyses of dissections from two newborn infants with hip dislocation were given. The first case was a fetus that was born with many congenital malformations and died within an hour of birth. There was a flexion and adduction deformity of one hip along with bilateral clubfoot, cystic kidneys, and an appearance consistent with a decrease in amniotic fluids such that the fetus had been compressed during intrauterine life. One hip was normal and the other abnormal. The round ligament was intact on both sides but was longer on the dislocated side. The

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip acetabulum was more shallow on the dislocated side. The acetabular rim did not exist especially at the superior region on the involved side, and the head of the femur, which was much smaller than that on the normal side, no longer had a hemispherical shape. It was flattened on one side and rested its deformed and flattened segment on the superior rim of the acetabular cavity. The entire proximal region of the thigh on the subluxed side was atrophied. Measurements of the acetabulum were smaller on the involved side than on the normal side. The Y (triradiate) cartilage did not show any modification of its normal shape or size. The second observation was made in a female fetus born spontaneously following an apparently normal pregnancy but who died several hours after birth. One lower extremity was shorter than the other and was held in abduction and external rotation. There was also a bilateral clubfoot deformity. The opposite hip was normal. The acetabulum was scarcely seen and the femoral head was markedly atrophied. The round ligament was quite long, and a pseudo-capsule had formed on the outer wing of the iliac fossa into which the capsule and synovium were attached. The neck was markedly shorter on the involved side, and the head in relation to the shaft was implanted at essentially a fight angle (anteversion), which was also indicative of shortness of the neck. Sainton felt that the acetabular cavity did not even exist on the involved side, although where the head rested against the fossa a small new cavity was formed within which an articular capsule had also formed. The Y-cartilage of the base of the acetabulum also was studied and no modification of it was seen. These two studies were presented as proof that at the moment of birth very pronounced bony deformation was present both of the proximal femur and of the acetabulum. It was evident that these deformities would be increased later when the individual began to walk, but they would only represent an exaggeration of modifications of form that already existed at the time of birth. The final dissection was performed on a 12-year-old girl who had signs of bilateral congenital dislocation of the hips and who died during the process of therapeutic reductions. The femoral head was found displaced into the external iliac fossa. The capsule was thickened considerably by associated fibers except at its superior area where it rested against the head of the femur and the greater trochanter. The muscular insertions were present and normal. The position of the head was maintained in flexion and adduction in relation to the pelvis. In this particular case, there was total absence of the round ligament and the area of the femoral head where the ligament normally inserted was associated with a round depression. The head itself was somewhat irregular, being ovoid and flattened at the top. The articular cavity was divided into three portions involving the new joint superiorly, which contained the femoral head, a second part that had developed into a flattened region referred to as the pseudoacetabulum, and a third part that corresponded to the original acetabulum and that was triangular in shape and the smallest

169

of the three regions. The Y-cartilage normally was conserved, and Sainton felt that the theory that attributed congenital dislocation to premature fusion of the Y-cartilage was false. 10. KIRMISSON,1894 Kirmisson wrote extensively on the etiology and pathogenesis of congenital hip dislocation (147). The vast majority of disorders were noted to occur in females, and unilateral dislocations were more frequent than bilateral. He supported the existence of congenital dislocation of the hip evident at the time of birth. Two dissections were reported. In one of the teratologic cases the round ligament was present, but it was much longer than that on the opposite side and the acetabular cavity was much more shallow than normal. The rim of the acetabulum had no prominence in general and did not exist especially at the posterior part. The head of the femur was smaller than that on the opposite side and did not have the normal hemispheric shape, being flattened on one side. The femoral head was not positioned in the acetabulum, which was far too small to receive it; it rested instead with its deformed side on the superior part of the socket. In a second case the round ligament also was present but longer than normal. The capsule was inserted at the base of the fossa in a region in which there was neither a border nor a prominence of any acetabulum. Kirmisson commented that his study of both specimens left no doubt about the existence of the displacement. In the large majority of cases, however, hip displacement was not recognized by physicians or family either at the moment of birth or in the first few weeks of life. The luxation perhaps was about to appear but in effect did not yet exist. In reviewing eight cases under his care, Kirmisson noted that in only two was the deformity noticed at the time of birth; in the large majority the diagnosis was made only when the children began to walk. It was evident that the deformity, associated with the pathoanatomic abnormalities, existed at the joint from the moment of birth but was completed and magnified under the influence of walking. Paralysis or contracture of muscles was rejected as a general theory for causation of congenital dislocation, and the contractures appeared as secondary phenomena. The contractures led to hip flexion, adduction deformities, and tightness of the fascia lata. In the young infant with hip dislocation, the hip was very flexible and one could put it in any number of positions. Kirmisson negated the hypothesis of muscular paralysis, although he based this opinion on operative interventions done to perform open reduction of the hip in which it was noted that the muscles appeared clinically normal. The disorder represented a primitive malformation of the joint. Lorenz had noted the absence of the round ligament in 40 of 57 cases and noted that above the age of 5 years it was rarely seen (147). In some cases it was present and in others it was absent, such that one could not propose a theory of hip dislocation on that structure alone, and changes in it also

170

CHAPTER 3 9 Developmental Dysplasia ojr the Hip

were felt to be secondary. Kirmisson discussed the possibility that the dislocations were due to an arrest of development of solely the acetabular cavity as had been suggested by others. Examination of the two cases demonstrated abnormalities not only of the acetabulum but also of the proximal femur; "each of the elements constituting the hip joint articulation participated in the deformity." The major question was the origin of the malformation, but he was objective enough to conclude that "we are completely ignorant of the cause of this malformation." 11. DEVELOPING AWARENESS OF TERATOLOGICAL CONGENITAL HIP DISLOCATION: EXAMPLES ASSESSED IN THE FETUS AND NEWBORN

a. LePage and Grosse LePage and Grosse described the congenital dislocation of one hip in a child who died at 14 days of age (170). The infant, who was born at term, had shortening of one lower extremity and multiple congenital anomalies, including a cleft palate, facial asymmetry, ectopic testicle, and a large hernia. Both thighs were hyperflexed and rigidly held against the abdominal wall, with the legs flexed on the thighs. The fight femoral head clearly was dislocated and the limb on that side was shorter than that on the opposite side, with the left being normal. A detailed autopsy was performed. The pelvis was asymmetric, with the fight side smaller and less well-developed than the left. The ischium and sacrum on the involved side also were less prominent. There were no anomalies or irregularities of the musculature about the displaced fight hip. The articular capsule on the normal side inserted circumferentially about the acetabular cavity. On the involved side it was greatly stretched out, being large at either extremity opposite the acetabular cavity and the femoral head but narrow in its middle portions. The inferior part of the capsule was stretched out over the empty acetabular cavity. The superior, part which was very thick, completely covered and almost enveloped in its circumference the dislocated femoral head just under the anterior superior iliac spine. The normal acetabular cavity on the opposite side was deep and regular in shape, with a wellformed cartilaginous rim. It completely contained the femoral head and measured 11 mm in its anteroposterior diameter and 13 mm in vertical diameter. The round ligament (ligamentum teres) was 6 mm long and did not permit any displacement of the head from the acetabular cavity. On the fight side, the acetabular cavity was atrophied, not very deep, without a bony rim, and without a cartilaginous rim. It measured only 6 mm in anteroposterior diameter and 8 mm in vertical diameter. At its posterior superior part, the bony rim, which normally forms a support against which the femoral head rests, was flattened by the friction of the dislocated head. The displacement of the head outside the acetabular cavity also was made possible by the abnormal length of the ligamentum teres, which measured 10 mm and inserted at the bottom of the acetabular cavity. There was no premature ossification of the triradiate cartilage.

The normal femoral head was hemispheric in shape with a diameter of 13 mm. The femoral neck was well-formed. The distance between the base of the greater trochanter and the top of the head was 17 mm. On the involved right side the entire proximal portion of the femur was atrophied involving the head, neck, and greater trochanter. The femoral head was deformed, conical in shape, and smaller, measuring only 9 mm in diameter. The femoral neck scarcely was seen to exist and the difference from the base of the greater trochanter to the top of the head was only 11 mm. The entire femur on the right was thinner than that on the left, although essentially it was of the same length and also presented a curvature convex to the lateral side. An illustration of both proximal femurs and the pelvis, once the capsule had been removed and the head displaced, showed the atrophied femoral head on the involved side and the empty acetabular cavity on the involved side covered by a membrane consisting of the stretched capsule, whose most superior part had been stretched and distended by the displaced femoral head resting up against the ilium. The ligamentum teres was thin and stretched; there was a depression just above the acetabulum where the femoral head had been resting. LePage and Grosse felt that the dislocation of the hip resulted from an arrest of development that was present not only at birth but had been present for some time during the fetal state. The diagnosis was made in the newborn and was markedly different from most instances of congenital dislocation of the hip, which tended to be made only when the infant began to walk and was noted to have an abnormal gait. They felt that, at the time of birth, there was a considerable difference between the case they were describing, which we would now refer to as a teratologic dislocation, and the more common displacement because the latter rarely was diagnosed in the newborn period. They felt that the common congenital dislocation of the hip was not truly dislocated at birth. At birth there was not a displacement of the femoral head but rather an anomaly of articulation, which consisted of the fact that the femoral head was not fitted deeply into the cavity but rather placed opposite an immature model of the cavity. It was only under the influence of walking that the position of the head changed and rose to position itself adjacent to the iliac fossa. The displacement was based on the early articular malformation but really only occurred and magnified itself with gait. LePage and Grosse indicated that, when the displacement was recognized at birth, it was almost always due to a difference in the position, length, or size between the two lower extremities that allowed the diagnosis to be made. In the case they were describing, the displaced hip was diagnosed on the basis of the shortening of the involved side and the multiple malformations present that led to a very careful examination. Because congenital displacement rarely was diagnosed at birth, it was even more rare to examine the skeleton of a newborn who had this malformation. The case of Barr and Lamotte was referenced, in which a similar congenital dislocation was associated with an

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip arrest of development of the acetabular cavity, the iliac bone, the femoral head, and the entire superior part of the femur. LePage and Grosse also commented that a stoppage of development was not necessarily required for a dislocation; it was sufficient if there was a disproportion in the dimensions between the elements comprising the hip. An acetabular cavity that had slightly atrophied or was smaller in size than expected then could be incapable of receiving and maintaining the femoral head for its normal development. Many observers felt that the smallness of the acetabular cavity was due to premature fusion of the triradiate cartilage, but the studies of LePage and Grosse, Grawitz, Kirmisson, Broca, and Lorenz were not able to define this. In addition, Sainton and Delanglade always found preservation of the triradiate cartilage, such that there was no evidence for its abnormality. They felt that it would be very rare for there to be significant atrophy of the acetabulum in the presence of an open triradiate cartilage to allow for inappropriate fit with the adjacent femoral head. The authors felt that an arrest of development was the cause of the dislocation. Trauma previously had been listed as a cause of the displaced hip, but there was no trauma in their case based on the details of delivery, which occurred for a very small infant over the course of only 10 min to a mother who had four previous pregnancies. In addition, the capsule was intact at autopsy without evidence of tearing or bleeding. They also referred to multiple experiments in which individuals experimentally attempted to produce dislocation of the hip in autopsy specimens by manipulation and managed only to produce either fractures of the proximal femur or proximal femoral growth plate fracture-separations. b. Cautru Cautru described a case of congenital dislocation of the hip in a patient with multiple congenital malformations, who died several hours after birth (38). There was shortening of the left lower extremity and the hip region was held in abduction and external rotation. There were bilateral clubfoot and a complete anterior dislocation of each radius at the elbow. The fight hip was normal but the left clearly was subluxed with a small atrophied or underdeveloped femoral head and a poorly developed acetabulum. c. Kirmisson Kirmisson reported on a dislocation of the hip in a stillborn fetus (148). There were asymmetries of the pelvic region, including the sacrum and ilium, and considerable atrophy on the affected side. There was a posterior subluxation of the femoral head, and the acetabular cavity on the affected side pointed almost completely forward and also was markedly atrophied. The femoral head, rather than being fully contained in the acetabulum, was straddling the posterior ridge of the acetabulum. The femoral head had a vertical posterior depression against the point where it rested against the acetabular ridge. At that area the acetabular rim was flattened. The anterior and middle regions of the acetabular cavity were empty and did not relate to the femoral head. The capsular attachments were such that the femoral head remained within the capsule, but the capsule was notably

171

elongated. The involved femoral head quite obviously was smaller in each of its dimensions. Vertically it measured 1.3 cm compared to 1.6 cm on the opposite normal side, the transverse diameter was 1.4 cm in comparison with the normal 1.7 cm, and the anteroposterior diameter was 1.3 cm compared to the normal 1.6 cm. The ligamentum teres was only slightly elongated and the surrounding musculature was normal. Each of the elements of the hip region was affected; the lesions did not involve simply the femoral head but rather all structures constituting the joint. d. Potocki Potocki reported a detailed analysis of a congenital dislocation of the hip in 1905 (236) from a child stillborn at approximately 7.5 months. The thigh was shortened on the involved side and there was an associated clubfoot. The head was larger than normal. There was also a flexion and adduction contracture of the left hip with limited movements into extension and abduction. The involved thigh was flexed and rotated internally. Based on the appearance of the limb, a diagnosis of congenital dislocation of the hip was made and confirmed radiographically. The musculature of the hip, thigh, and leg on the involved side was less bulky than normal. The muscular insertions were normal, however, except for the pyramidal muscle, which was attached to the superior aspect of the hip joint capsule. There were some structural abnormalities of the sciatic nerve. There was an arrest of development of the pelvis quite pronounced on the affected side. A large part of the external iliac fossa on the affected side was filled with the articular capsule of the hip. The capsule was thin in its anterior and superior part, but the ischio- and pubofemoral ligaments were very thick. Once the capsule was incised, there was a spacious articular cavity for the femoral head. The cavity was divided into two parts by the rim of the acetabulum. The superior part had for its edge the insertion of articular capsule opposite the external face of the ilium and below the acetabular rim. At this level, however, the iliac cartilage was not covered by articular cartilage. The true acetabular cavity was empty and poorly developed. The rim of the acetabulum was missing except in its most inferior portion. The depths of the acetabulum were almost completely occupied by the ligamentum teres. The triradiate cartilage persisted and was of normal large size. The empty acetabular cavity was more elongated along the longitudinal dimension; its transverse diameter was 8 mm, whereas its vertical was only 4 mm. On the normal side, the transverse diameter of the articular cavity was 11 mm and its height was 13 mm. The femoral head was not hemispherical in shape but rather flattened, particularly that part that was relating to the iliac bone. The head also was smaller than normal in terms of height and diameter. The ligamentum teres was much longer than normal as well as thinner. On the normal side, the round ligament (ligamentum teres) was thick and short and did not permit any displacement of the femoral head from the acetabular cavity. The neck of the femur on the involved side was shortened.

172

CHAPTER 3 9 Developmental Dysplasia of the Hip

In summary, each of the bony elements entering into the hip joint was notably smaller and deformed, whereas the articular capsule and round ligament were markedly elongated and the rim of the acetabulum was scarcely protruding (i.e., markedly flattened). The author denied trauma as the cause, showing no difficulty during pregnancy or delivery, which was vertex in nature. There were neither contractures or paralysis of muscles. The most common cause was felt to be an arrest of development affecting each of the elements of the hip joint. The development had not been affected from the earliest stages, but rather was due to an illness in midpregnancy. 12. CLARKE,1896 Clarke presented a series of illustrations from a case of bilateral congenital dislocation of the hip, which was stillborn at full term and in which bilateral hip dissections had been performed (43). Both hips were fully flexed with the knees in the extended position. Abnormalities of development of the acetabulae are shown and compared with the appearance in the normal. Malposition of the femur in relation to the acetabulum and the tightened capsule was noted. The joint deformity was reduced, after opening of the capsule, by downward pressure on the femur and internal rotation. The misshapen proximal femur and pelvis also were illustrated. 13. I~wTLWV, 1900 Keetley also remarked on the controversy as to whether or not the disorder was truly congenital and recognized that, although the preconditions might exist at the time of birth, the actual dislocation appeared to occur afterward (145). His text had a detailed section on the etiology of congenital dislocation of the hip. Grawitz had studied seven patients with teratologic hip dislocations, of whom five had bilateral and two had unilateral dislocations. All had multiple congenital anomalies including spina bifida, clubfoot, clubhand, and scoliosis. The development of dislocation of the hip in association with infantile paralysis also occurred. Those individuals described were not in the idiopathic developmental dysplasia category. The Y-shaped cartilage was relatively undeveloped and the acetabulum disproportionately small in relation to the head of the femur. Among the reported causes of hip dislocation, which were hypothetical without experimental or pathoanatomic evidence, were position in utero (Dupuytren), deficiency of liquor amnii (Roser), intrauterine injuries secondary to trauma to the mother's abdomen, and spasmodic fetal muscular action. The position in utero argument was not felt to be compelling because almost every fetus is placed with its thighs hyperflexed. Another hypothesis related trauma to the fetus during birth in particular with breech presentations. Keetley argued that, although breech presentations have a higher incidence of hip dislocation, their absolute numbers were still small in relation to the number of cases of hip dislocation. Other hypotheses related the dislocation to congenital absence of an acetabular rim

(observations made by Lockwood and Grawitz), relaxation of the ligaments of the joint (Sedillot), and disease of the fetal hip joint. Keetley hypothesized, in agreement with Sedillot, that "so called congenital dislocation of the hip may sometimes or even often be due to abnormal laxity of the ligamentous structure of the joint, the dislocation of necessity occurring long after birth." In support of this latter theory he commented that the ligamentum teres generally is intact, but at the same time observations noted laxity of the capsule and external ligaments of the joint. He strongly supported that etiology in a child in which dislocation was discovered a long time after birth and in which there had been no symptoms, no history of infantile paralysis or injury, and no other associated congenital deformity. Even at this relatively early date all data pointed to the strong preponderance of females, with nearly 9 patients out of 10 being female. The dislocation was unilateral only slightly more commonly than bilateral in a proportion of less than 3 to 2. Keetley summarized the then current information well and reached a conclusion, which regrettably still exists today, that "with regard to the majority of the cases classed together by surgeons under the name of congenital dislocation of the hip it is easy to suggest many theories of their origin and difficult if not impossible at present to prove one." The pathological changes found depended on the age of the patient and the variable causes of the deformity. (1) Newborn: The head and neck of the femur were smaller than normal and also altered in shape, sometimes being short or round and sometimes long and conical. The acetabulum "was always small" even in proportion to the diminished size of the head of the femur. It was narrow and oval. Fat occupied its cavity, and the deficient depressed posterior margin was encroached upon by the cartilage surface of the new acetabulum. The ligaments and the joint capsule were stretched but not torn; the ligamentum teres was unusually long, thin, and flat. The joint capsule enclosed both the old and new acetabulum. The pelvis also was somewhat misshapen. In the teratologic cases described the cartilage was intact, the joint capsule and the round ligament were lengthened, and luxation could be reduced although it quickly redisplaced. There was defective development of the Y-shaped cartilage of the acetabulum. (2) Older children: Once the child began to walk the changes in the hip region became more marked. The acetabulum was narrower, smaller, and shallower and it began to assume a three-cornered shape, filled with fat, and no longer able to receive the femoral head even when reduction was performed. The capsule and ligamentum teres were longer. A regular false joint formed on the dorsum of the ilia. In many patients operated after several years, the original acetabulum scarcely could be noted. Keetley indicated that for "practical purposes the acetabulure gradually becomes obliterated." A pathoanatomic study from a full-term stillborn fetus demonstrated the hyperflexed position of the lower extremities. The entire innominate bone was deformed, and the acetabu-

SECTION IV ~ Etiology and P a t h o a n a t o m y of Developmental Dysplasia of the Hip

173

FIGURE 2 Illustrations from the work of Le Damany show normal features of developmental hip anatomy and his theories as to how subtle abnormalities of femoral and acetabular development in particular in combination predispose to dislocation when the child assumes the upright posture with hip extension. (A) This illustration shows changing proximal femoral anteversionduring development [derived from (163,167)]. Part (A) at left illustrates neutral version in the early fetal stages. The distal femur is shown by the dotted lines and the proximal femur is outlined in continuous lines. In each of A, B, and C the femoral head is at the middle part of the image, the trochanter at the left, and the distal femoral condyle at tight. The axis of the proximal femoral head and midneck is illustrated by the line from the trochanter through the midneck and head. There is no version in A, anteversion as great as 40~ during the fetal period in B, and diminution of the anteversion to approximately 10~ at skeletal maturation. (B) The combination of changes predisposing to subluxation and dislocation are shown [derived from (165)]. A normal alignment at left shows normal degrees of anterior acetabular opening and proximal femoral anteversion to allow the head to be seated within the acetabulum. At right there is increased anterior opening of the acetabulum (increased acetabular obliquity) in association with increased proximal femoral anteversion. These two features can lead to relative instability particularly when the hip assumes the uptight posture.

lum was oval and much smaller than in a normal specimen. The joint displacement prior to opening of the capsule was shown as was the reduction possible once the capsule had been freed and the femur pulled downward and rotated inward. The femoral head was displaced superior and posterior to the acetabulum, there was proximal femoral antiversion, and the ligamentum teres was elongated and flattened. 14. LE DAMANY,1904 Le Damany wrote a series of articles on the pathogenesis of congenital dislocation of the hip that remain of value today, particularly his discussion of normal hip development and diagrams illustrating the possible relationship of subtle proximal femur and acetabular abnormalities to the disorder (163-167). He reviewed the evolutionary development of the hip in humans and other species. The human changed hip position from flexion in utero to extension postnatally; this fact, if combined with particular variations of acetabular and proximal femoral development, could lead to hip dislocation. During the final stages of intrauterine growth, the relatively large size of the fetus and in particular of the femur in relation to other species caused the hip to assume the hyperflexed position and relatively increased pressures to be exerted to the hip region of the femur from uterine pressure against the knee. These two changes led to altered relationships of the proximal femur and acetabulum in comparison to those present in the embryonic and early fetal stages. Le Damany noted that in the human fetus the acetabulum was hemispheric during the first two-thirds of fetal life but that during the last trimester its depth in relation to its width gradually diminished. He felt that early on the depth was one-half that of the diameter but that at birth it was only

two-fifths. In adult life it increased further to three-fifths. Therefore, a relative lack of depth of the acetabulum occurred around the time of birth, and this had negative implications in relation to hip stability. The proximal femur also increasingly underwent rotational changes during the latter stages of intrauterine growth. It went from neutral anteversion (0 ~ at 4 months to anteversion as high as 3 5 - 4 0 ~ at birth, followed by diminution to approximately 10 ~ by the time of skeletal maturation (Fig. 2A). To Le Damany these changes were examples of extrinsic pressure affecting growth of the relatively pliable cartilage model of the bones. When the changes in both the acetabulum and the proximal femur were exaggerated slightly, the possibility of dislocation was increased greatly. The proximal femoral and acetabular abnormalities or, more accurately, angular-rotational excesses therefore were maximal at term. He also noted that no anterior obliquity or tilt of the acetabulum occurred in any of the species other than humans. Normal anterior obliquity of the acetabulum did not vary with developmental age, although it had a relatively wide range of values in the normal between 15 ~ and 30 ~. Le Damany then established a quantitative index that when applied to cadaver studies of available specimens with CDH, led to an indication as to what amount of proximal femoral-acetabular deformity led to dislocation. He simply added the degree of femoral anteversion to the degree of anterior acetabular obliquity. Femoral anteversion in the normal varied between 30 ~ and 50 ~ whereas acetabular obliquity varied between 15 ~ and 30 ~. The most extreme angulations therefore would be 50 ~ (femoral) plus 30 ~ (acetabulum) leading to an index of 80 ~ that would favor dislocation, whereas the safest values would be 30 ~ (femoral) plus

174

CHAPTER 3 9

Developmental Dysplasia of the Hip

15~ (acetabular) leading to an index of 45 ~ In the adult the normal index would be in the range of 32 ~, involving the proximal femur anteversion of 12~ and the acetabular anteversion of 20 ~. It was those infants with an index greater than 60 ~ that were subject to dislocation (Fig. 2B). If all of the constituent elements of the hip were within the normal range, the subtle deformities were capable of correction, often spontaneousl3~, but if the additive features led to a sufficient degree of deformity, then a dislocation would result. These relatively subtle developmental variations were an increased femoral anteversion accompanied by an increased anterior opening obliquity of the acetabulum. These were reciprocal occurrences such that each finding could be in the normal range, but when they were both somewhat more marked the situation was set for a possible dislocation. He further stressed that the definitive dislocation itself would not occur until several months after birth, at which time hip function in the extended and then uptight position began to have its most negative effect. Positioning of the human hip into extension further worsened the stability. He made a series of comparative hip studies throughout other species, pointing out that dislocation was virtually unheard of in any other species and that the reason for this was primarily the maintenance of the flexed position of the hip in all but the human. In Le Damany's outline of normal hip development, embryo, neonatal, and adult femur specimens were assessed. Proximal femoral anteversion was noted to begin only during the second half of pregnancy. There was essentially no anteversion of the proximal femur during the embryonic phase and he listed it at 0 ~ By birth it had increased to an average of 40 ~ (ranging from 30 to 50~ whereas by the time of adulthood it had decreased to a mean of 12~ In association with this was the fact that the acetabulum was most shallow relatively in relation to the position and size of the femoral head in the newborn period, which also predisposed to subluxation and dislocation. There was also slight anterior obliquity with the acetabulum opening not only to the side but anteriorly, which worsened stability. The secondary changes in a dislocated hip had been described previously, but he felt that the initial subtle pathoanatomic reasons had not been appreciated. Le Damany divided the causes of congenital hip dislocation into two basic categories: teratologic and anthropologic, the latter referring to what we now term idiopathic DDH. In a detailed presentation he divided congenital hip dislocations into group A, which were relatively rare and included those of teratologic, traumatic, and pathologic origin, and group B, which were relatively frequent and referred to as anthropologic luxations. The traumatic dislocations were not caused by direct extrinsic damage to the fetus but occasionally were produced by difficulties in birth, although even then most hip displacements at birth were felt to be fracture-separations. Among the pathologic disorders were those involving neuromuscular or infectious disorders of the hip, and the

teratologic abnormalities referred to those associated with spina bifida or true abnormalities of development with many systems malformed. The term anthropologic dislocation was used to refer to the combination of abnormalities that Le Damany himself had described and thus appeared as idiopathic congenital or developmental dislocation of the hips because the patients otherwise were normal. He described these as being congenital because the predisposing variations developed during the latter stages of intrauterine life, although the actual dislocation occurred after birth. The abnormalities involved extensive anteversion of the proximal femur and increased anterior obliquity of the acetabulum beyond a stable range once the postnatal posture of hip extension was assumed. Both of these were due to the same mechanism of increased force on the hip region due to the relatively large size of the human fetus, the length of the femur, and the external uterine pressure on the relatively soft developing bones. Postnatally, the position of the human hip into extension in association with the femoral and acetabular variations led to the dislocation. He stressed that in the anthropologic type the patient was normal otherwise, whereas in the rarer intrauterine disorders many patients had other abnormalities. The anthropologic types resulted from an exaggeration of the normal variations or imperfections of the human hip and were extremely frequent. The possibility of cure, however, was quite high because both femoral and acetabular tendencies were to correct to a normal range with development. The displacement thus is not truly congenital, but predisposition for the displacement shortly after birth occurred due to late intrauterine events. Le Damany was quite aware of the existence of the dislocatable hip at birth, which he felt was quite frequent and which generally went on to a spontaneous recovery without treatment in a period of time varying from several days to several months. He even described the clinical maneuver that would sublux or dislocate the hip and the reverse maneuver that would allow it to relocate. When the hip was placed in a position of flexion and adduction and a slight force was applied to the knee along with a force from within outward on the thigh, the head would displace over the posterior border of the acetabulum. Replacement into the acetabulum would occur with a reverse flexion-abduction maneuver. Le Damany also indicated that the principles of therapy were quite simple because the dislocation in the anthropologic type was due to only a slight deviation from the norm. He referred to the proposed treatment as being almost completely geometrical and mechanical and noted that those under 3 years of age should be readily curable. The first principle simply was to place the articular surfaces in exact coaptation such that the new acetabulum would be formed exactly in the site of the existing but slightly shallow one. The pressure of the head would hollow out the cavity. "If we wish to reform an articulation it is not to immobilization that we ought to have recourse but to motion." He went on

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

to indicate that "function makes the organ." Le Damany noted that, at the time of birth, the deformations that provoked the congenital subluxation, torsion of the femur and anterior obliquity of the acetabulum, had not yet produced any deformation in either the head or the acetabulum. It is only with time that the secondary changes occurred as the hip remained subluxed or dislocated. These changes were well-recognized and included atrophy of the empty acetabulum, further deformation of the head of the femur and of the neck, proximal femoral anteversion, which did not correct from its neonatal level, enlargement of the capsule with or without the hourglass narrowing, lengthening and thickening of the round ligament (ligamentum teres) followed by its thinning and eventual rupture, lengthening of certain muscles and shortening of certain others, and shortening and atrophy of the involved limb. The anatomic abnormalities on which Le Damany commented therefore preceded and prepared for the dislocation, such that they were necessary for its occurrence but were not the sole cause of it. The positional abnormalities were at their maximum in the newborn. There was little awareness of his anthropologic variant because newborn pathoanatomic assessments, which were extremely rare, did not note any abnormalities because the shape of the head was normal, the shape of the acetabulum was normal, and the dislocation had not yet occurred. Only if the detailed studies that he had described were performed would one notice the fact that the predisposing features for the dislocation were present. The dislocation rarely was present at birth but occurred only after birth due to the extended position of the hip and the onset of weight bearing. Although Le Damany's work was mainly theoretical, in one article he did present pathoanatomic descriptions of four hips varying in age from 3 to 18 months in which the abnormality he described was defined. Three of these he felt would have corrected spontaneously and one had gone on to subluxation. 15. BENNETT,1908 Bennett discussed congenital dislocation of the hip, indicating that birth injury was not a contributing factor but that abnormal laxity of ligamentous structures of the joint was causative (15). Bennett combined previous theories, feeling that in a child who is otherwise normal the dislocation was a "pure accident, dependent on the position in utero which makes it possible, and the laxity of tissues around the joint which makes it probable. For the accomplishment of the deformity, it needs but a slight movement in the required direction.., either in or outside the uterus." He agreed with Dupuytren's hypothesis in which the extremely flexed uterine position in a child with lax tissues served to allow the head to slip onto the posterior acetabular rim, which as a result of the pressure exerted did not grow at the same rate as the remainder of the circumference. As a result, at birth the head was lying on an atrophied posterior rim, and it was

175

just chance whether the early postnatal movements placed it in or out of the socket. a. Pathology The capsule necessarily is stretched with the head displaced from the acetabulum. With walking the stretching is worsened and the capsule becomes thickened. As it stretches across the acetabulum it becomes adherent at the rim so that the acetabulum appears obliterated, being covered by thick fibrous tissues reaching from rim to rim. The adductor muscles are shortened secondarily as is the iliopsoas. The glutei are not shortened initially and the muscles that pass from the pelvis to the greater trochanter (obturators, gemelli, etc.) are lengthened. The hip ligaments adjust to the deformity and shorten along with the adductor muscles and hamstring muscles and thus become factors opposing reduction. Those that are attached to the lesser trochanter are particularly involved. Bony changes involve a gradual loss of the hemispherical character of the head of the femur, which tends to become flattened, and an alteration in the shape and depth of the acetabulum, which tends to become less circular and more triangular in shape with the tissues at the bottom filling it. Anteversion of the femoral head is seen and the neck shaft angle is increased. The usual position of displacement of the femoral head is upward and backward. 16. ADDITIONAL PATHOANATOMIC STUDIES Large numbers of pathoanatomic studies in the early decades of the twentieth century originated from central Europe due to both the high prevalence of the disorder in those countries and the high level of medical investigation. The increasing use of open reduction allowed for better appreciation of the pathoanatomy. a. L u d l o f f Ludloff performed one of the early studies of acetabular abnormalities in association with congenital dislocation of the hip (182). He made assessments of four individuals in relation to hip abnormalities prior to birth, shortly after birth, in adolescence, and in adulthood. Additional detailed reports from the literature were added. He concluded that the depth of the acetabular socket without exception was decreased significantly in all cases from the embryonic time onward. The width of the upper half of the acetabulum always was more seriously affected (diminished) than the lower half, which led to a triangular shape of the acetabulum. The reduction in depth increased during development of the deformity along with flattening of the periphery of the acetabulum and inversion of the limbus. The inversion of the limbus was the main reason for the flattening of the rim of the acetabulum. When the limbus was inverted, it played a major role in causing the protrusion of the head from the acetabular socket. Several illustrations from his own work and the work of others showed the marked underdevelopment of the acetabulum in the dysplastic hip compared with the normal. The shallow and flattened acetabulum was illustrated along with the widened and enlarged capsule, the anteversion of the proximal femur, and the elongation of the

176

CHAPTER 3 ~ Developmental Dysplasia of the Hip

ligamentum teres. The dysplastic acetabulum took on a triangular shape with its base toward the obturator foramen and its apex superior. Cross-sectional drawings of the acetabulum showed both its diminished width and depth as well as its loss of sphericity, with the floor tending to a flattened position. b. Deutschlander and Loeffler Deutschlander clearly outlined the pathoanatomy of the complete dislocation in the process of describing the limitations of closed reduction (56). Illustrations from his work show the primary difficulties with reduction due to the shallow and dysplastic acetabulum, the presence of fibrous fatty tissue within it, and the hypertrophied capsule covering the inferior half of the entrance to the socket. He also clearly demonstrated the dangers of capsular interposition in association with closed reduction. Similar illustrations of the various tissue interpositions limiting the effectiveness of closed reduction were shown by Loeffler (178). c. Werndorf Werndorf described the pathologic anatomy from a child who died having surgery on the contralateral hip (348). Findings included a socket reduced in size and flattened, a triangular shape to the acetabulum, the upper acetabular roof completely absent, and the socket base so thickened that barely a fingertip could be inserted. The ligamentum teres was lengthened and broadened considerably and completely filled the socket. The proximal end of the femur was retarded in growth with flattening of the head and considerable anteversion of the femoral neck. The flattening involved the medial and posterior parts of the head. The head was still contained within the capsule, which was lengthened, tubular in shape, widened at its upper part, narrowed centrally at the isthmus, and thickened considerably. Essentially all of the muscles passing from the pelvis to the femur were shortened, particularly the adductor group. The muscles, however, passing from the pelvis to the greater trochanter tended to be lengthened. d. Lance The secondary changes in hip structures in untreated CDH were also assessed by Lance (154). The femoral head was small and conical in shape, being flattened on one side. The secondary ossification center was late to appear and was not central along the long axis of the neck. The neck tended to be short and thickened and also rotated into increased anteversion. This frequently was in the range of 40 ~ 60 ~ at a time when it should have been diminishing to the 15~ ~ range. The lesser trochanter tended to be markedly hypertrophied due to changes in weight bearing in relation to its attached muscles. The acetabular changes were similar to those described by Ludloff. The head always rested within the capsule, which, however, was enlarged allowing displacement to occur. In some cases the capsule remained quite spacious, whereas in others it took on an hourglass shape with a narrow central isthmus below the head but above the original acetabulum. At times, the capsule was interposed between the head and the side wall of the ilium, whereas at other times it always remained superior to the head. There

were also examples of attachment of the capsule either to the external wall of the ilium or to the head and neck of the femur in many instances.

B. Later Clinical-Pathoanatomic Descriptions 1. FAIRBANK,1930 Fairbank discussed the pathoanatomy of congenital dislocation of the hip in detail, basing his report on 50 open surgical procedures for CDH, 46 dislocated hips from the Dupuytren museum in Paris, and an extensive literature review (70). His study is particularly strong on the late pathoanatomic changes seen in the adult years. The ultimate structural changes in congenital hip dislocation that had gone untreated or in those diagnosed late are well-described. Fairbank attributed the primary pathology even in the affected fetus to the "poor development of the upper margin of the acetabulum." a. Acetabular Findings The acetabular development is markedly abnormal when not accompanied by the presence of the femoral head in its normal position. With complete dislocation, the acetabulum becomes triangular in shape with its base toward the obturator foramen and the apex pointing upward and backward. This triangulation is considered to be the result of continued growth of the anterosuperior and posterior boundaries of the socket unchecked by the pressure of the femoral head. The margins of the acetabulum usually are straight and sharp and the cavity retains some depth, even though the floor is more or less flat. The cavity is filled by cartilage and fibrous fatty tissue. The bones of the adjacent pelvis also are abnormal: the obturator foramen is more triangular than normal, the pubic angle is increased, the ilium is shorter and broader than normal, and the anterior border is prolonged in the vertical direction. The anterior-inferior iliac spine is twisted to conform to this outline. A false acetabulum forms on the outer border of the ilium. As a rule, it is larger than the femoral head, which rests in or against it, and the disproportion suggests considerable mobility of the femur in both the anteroposterior and vertical directions. In some cases, the edges are well-developed and occasionally a deep hemispherical cup with a polished eburnated floor is seen. The appearance of the false acetabulum varied greatly. In 38 hips, Fairbank noted that it was only a shallow depression with little or no margin in nearly onehalf (17); in 9 there was no sign of a false joint, whereas in 9 others a well-marked socket with lipped margins and an ebumated floor was present. b. Femoral Changes The head is smaller than normal, even in the young child with dislocation, and the secondary ossification center is late in its appearance. The head becomes flattened by pressure against the ilium on its inner and posterior aspect. In adults, there is often erosion and pitting of the cartilage surface to actual complete disappearance of it. The shape of the head also is variable in several specimens but almost always small and imperfectly shaped. Antever-

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip sion is quite common in the patients with congenital hip dislocation. Whitman noted the normal angle of anteversion at 30 ~ at birth with gradual reduction to 10-15 ~, whereas if the hip was dislocated this reduction did not take place. In the study by Farrell et al., based on radiograms in 336 cases, in nearly one-half the angle was over 20 ~, and in these about one-half gave an angle of 20-50 ~ whereas in the remainder the angle was over 50 ~. c. Capsule As the head migrated upward into its dislocated position, it carried in front of it a dome of the capsule, which blended with the periosteum above and behind the acetabulum. Where this fusion occurred in the floor of the false acetabulum, the two were transformed into fibrocartilage. The capsule, although initially lax and thin, becomes thickened with time in particular in those regions in which it has a weight bearing function; it could be as much as onethird of an inch in thickness in a child of 13 years. With complete dislocation, the thickened capsule rides upward and becomes tightened and thus serves as a barrier against easy reduction. The joint cavity develops an hourglass shape. The isthmus between the true and false joints is accentuated by the altered position of the psoas tendon and by the thickened capsule. This well-developed isthmus occurs with time and generally is seen only after 3 years of age. d. Muscle The adductor muscles are always shorter than normal. The iliopsoas tendon plays a major role with displacement. When the femur is displaced posteriorly and upward the pelvis tilted, and the lordosis marked, this tendon takes a practically horizontal course upon leaving the pelvis. It is quite stressed and generally causes a deep groove in adult specimens below the anterior-inferior spine. Most felt that the tendon had to be divided as part of any open reduction. The gluteal muscles tend to be unaltered in length because the trochanter is displaced outward as well as up. The horizontal muscles, the obturators, gemelli, and quadratus, are lengthened and the direction of their fibers is altered. They no longer run horizontally but pass upward and backward to reach the trochanter. Support of the hip in the dislocated position is due almost exclusively to the soft tissues with the false acetabulum having relatively little support function. The work is shared by the capsule and the muscle. The capsular strain is taken by the thickened bands passing from pelvis to femur in front and below and particularly by the capsular sling arching over the neck. The muscles that assist the capsule and prevent further stretching of this sling are the iliopsoas in front and the obturator group behind, with the gluteus minimus helping to some extent. The abductor muscles in the dislocated hip act under decided mechanical disadvantages. Fairbank concluded that the older the patient at the time of reduction, the greater the chance of an imperfect anatomical result. Manipulative reduction is best "at an early age," defined at that time as before the age of 3 years and better still before the age of 2 years. He was convinced that closed reduction would be appropriate in the younger patients as

177

many of the capsular changes noted by early proponents of open reduction were relatively late secondary changes. The details of open reduction have been defined in England by Burghard, whose operation included dividing the psoas and enlarging the isthmus. Fairbank felt, however, that it was rarely necessary before the age of 4 years. The antetorsion frequently was corrected by osteotomy, but Fairbank felt that if the hips were reduced early, before the fourth year, the antetorsion would correct with resumption of gait over time.

2. LEVF.UF, 1947 Leveuf made a clear distinction between primary congenital subluxation of the hip and primary congenital dislocation of the hip, basing his opinion on multiple arthrographic studies as well as findings from operative interventions (175). The primary subluxation presented some anatomical characteristics that were distinctly different from those of a dislocation and he considered that the two disorders, therefore, were distinct rather than representing a situation in which a subluxed hip that worsened became a dislocated hip. In a subluxation the limbus is forced upward and outward, whereas in a dislocation the limbus is displaced downward and inward toward the acetabulum. This is particularly wellillustrated by a drawing of the two disorders. Specific differences were noted in the structure of the acetabulum, the head of the femur, the capsule, and the neck of the femur in each disorder. In subluxation the cartilaginous roof and the limbus, which were forced by the head of the femur against the external iliac wall, and the iliopsoas muscle always appeared atrophied, whereas in a true dislocation the cartilage roof and the limbus were forced toward the acetabulum, which led to the limbus being hypertrophied and interposed between the head and the acetabulum during a closed reduction. The acetabulum in the dislocation "generally retains a satisfactory depth with the well developed group." In subluxation the head of the femur was deformed very early, being enlarged and widened transversely as well as flattened at its superointernal pole. In dislocation on the other hand the head retained a regular contour for a long time. The joint capsule in a subluxation was enlarged but never interposed between the femoral head and the acetabulum, and the round ligamentum teres was "practically always absent." In dislocation, on the other hand, the capsule generally was interposed between the head and the acetabulum, and the round ligament was present in only about one-third of the cases. In dislocation there was a clear "interposition of soft parts" (the limbus, the round ligament, and the lower fold of the capsule) where they constituted an obstacle to reduction, whereas such interposition of soft parts never existed in a subluxation. The final region assessed was the neck of the femur, which in subiuxation often was associated with a valgus position as high as 150-155 ~ (normal 130~ Anteversion also was quite common in subluxation. On the other hand, in dislocation there was no valgus of the neck and anteversion was rare. Leveuf

178

CHAPTER 3 ~

Developmental Dysplasia of the Hip

further reinforced his opinion that the two disorders were distinct entities, indicating that "in our opinion not one well established fact can prove that a subluxation, showing the characteristics which have been described, can become a luxation." This radical distinction has not been widely accepted by the orthopedic community, but his description of the pathoanatomic findings in the two variants for the most part appears very accurate (Figure 3Fi).

3. BADGLEY, 1949 Badgley reviewed two main theories of etiology that had been prominent for many years, one indicating the lesion to be the result of a primary germinal fault and the other indicating that it was due to a mechanical defect of development of environmental origin (11). Those who felt that the developmental abnormality was a primary failure of proper formation of the acetabulum had little to no direct evidence for this, and the theory primarily was proposed and supported by orthopedic surgeons and radiologists. Badgley felt that acetabular abnormality that was clearly a part of the congenital dislocation hip syndrome was a secondary developmental problem and not a primary one. Mechanical factors clearly were involved in some instances of congenital hip dislocation. Badgley referred to the report by Tridon that 121 cases of congenital dislocation of the knee had been associated with 20 instances of congenital dislocation of the hip. Although this was indicative of a mechanical etiology, it did not substantiate the mechanical theory for dislocation of the hip except in this unusual circumstance. The hyperextended knee seemed to be analogous to the high incidence of DDH in breech malposition. He also felt that the mechanical concept of Le Damany seemed plausible but did not adequately account for all features. The high incidence of CDH in those with breech presentation also is considered to represent an example of mechanical effects on development. Badgley reviewed the embryologic principles of development, concentrating on the importance of "perfect timing" for the development of the constituent parts. Early embryologic development is intrinsic to the specific part, but ultimately, when the gross skeletal model was being refined and perfected, the importance of extrinsic factors increased. The timing of various events and the necessity for "perfect timing" during the various phases of growth of the hip are essential. Rotation of the limb buds is an important feature of embryonic development. The limb buds and extremities undergo rotational changes during development to the extent that they ultimately twist around their longitudinal axes and rotate through an angle of approximately 90 ~. The alteration of position of the limb buds starts prior to the separation of the components of the hip joint, and "this postural change of the limb bud prior to motion in the hip joint may be a definite factor in the production of the inclination of the neck of the femur." Most postural change, however, occurs after commencement of the joint cavitation after the 30-mm stage of fetal development. The femoral region must rotate inter-

nally approximately 90 ~ at the hip joint as part of the normal developmental sequence. Adaptive changes in the acetabulum and the upper end of the femur are necessitated by the rotation phenomenon as well as the development of the oblique position of the acetabulum. The inclination of the acetabulum is important, with 30-40 ~ of forward inclination and 60 ~ of downward inclination. Dega, in a review of 100 fetal skeletons, showed the angle of forward inclination of the acetabulum to be 29.5 ~ and the downward inclination in relation to the transverse plane to be 62.8 ~ These changes in acetabular development also were pointed out by Le Damany, who commented on the tilting of the iliac bone of the sacrum in the human, obliquity of the acetabulum, and anteversion of the head and neck of the femur associated with femoral torsion. If the sum of the obliquity of the acetabulum and the anteversion of the neck of the femur was greater than 60 ~ dislocation occurred. Dega also pointed out that the femoral head and acetabulum developed in close correlation but that perfect adaptation of the component parts was present in the intrauterine position of flexion. Badgley postulated that the fault did not lie in a hereditary failure of one part, but, in the embryonic and early fetal stages, conceivably there could be an interference in the orderly time development of reciprocal parts after the formation of the joint cavity. Extrinsic factors were more likely to be involved than hereditary genetic factors in both the acetabular structures and the femur, with changes occurring on the basis of a secondary adaptive fault from an alteration in the normal timing of development. Badgley then concentrated on the changes at the upper end of the femur, which involved increased anteversion of the head and neck region. It was inappropriate to ignore the changes in the femur or to call them secondary changes when they were as marked a part of the deformity as the acetabulum. He felt that both acetabular and proximal femoral changes were reciprocal faults secondary to a developmental error. They related to interference of the time of development of the intrinsic pattern producing environmental extrinsic faulty development and leading to the structural abnormality. Anteversion occurred primarily in the diaphysis, with the head and neck in normal relation with the trochanters although anteverted in relation to the shaft. The radiologic evidence of apparent coxa valga at least partially is due to increased femoral anteversion as can be shown by taking a radiograph with the femur in internal rotation, at which time the angle of inclination will be found in most to approximate the normal. His concept of congenital dysplasia of the hip was that, through a developmental fault, the acetabulum has failed to deepen and the head and neck of the femur have become anteverted. The anteversion tends to rotate the cartilaginous head forward and laterally so that the glenoid labrum and acetabulum cover less of the head than usual. The adaptation of the head and acetabulum continues dynamically to require growth changes, altering the intrinsic mosaic pattern, and manifesting in subluxation or acetabular dysplasia. Rotation of the

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

limb buds may be an important early factor in the abnormal development. Interference with the orderly timing of rotation could produce a failure of the intrinsic mosaic design. The altered environment then would produce adaptive features in all structures of the hip joint and not simply a primary change in the acetabulum alone. The loss of the normal dynamic reciprocal relationship of the component parts of the hip joint during the stage of rotational adjustment of the limb buds may produce the secondary adaptive changes that lead to acetabular dysplasia or congenital dislocation. The known embryological development of the hip joint thus is opposed to the theory of a primary inherited failure of development of a portion of the acetabulum alone. 4. HOWORTH AND ASSOCIATES Howorth noted that displacement initially was lateral and that if present in utero the displacement was posterior. After birth, with extension of the hip and especially with weight bearing, displacement tended to be upward and anterior. The only pathology seen in every case was an elongated and lax capsule. All of the bony changes were felt to be secondary involving both the acetabulum and the proximal femur. He found no examples of "hourglass" constriction even though the capsule was elongated. The capsule was pulled from below across the opening of the acetabulum, but it really was the transverse acetabular ligament that allowed for the hourglass appearance. The amount of elongation depended upon the degree of displacement. Howorth commented on the inversion of the labrum into the acetabulum with full displacement. His illustrations of progressive changes with increased subluxation and eventual dislocation were quite consistent with those presented previously. Although Howorth and Massie did little pathoanatomic study themselves, their extensive clinical experience and writings were based on efforts to relate to the underlying pathoanatomy (125, 126, 187-190). They articulated the prominent view that congenital dislocation of the hip was due to a "pathologic relaxation of the joint capsule" and that "all other pathologic changes develop subsequently as the result of simple mechanical stresses." That viewpoint was widely adopted in midcentury and has remained the cornerstone of most management philosophies of congenital or developmental dislocation of the hip. Bone and cartilage changes involving acetabular dysplasia, increased proximal femoral anteversion, and delayed appearance of the proximal capital femoral secondary ossification center all were felt to be secondary mechanical sequelae of a failure of the femoral head to be appropriately positioned in the acetabulum. The fact that development of these structures occurred normally once the head was relocated definitively supported the impression that they were secondary and not primary phenomena. Howorth consistently maintained that the constant and essential pathoanatomic feature and the primary anatomic cause of displacement of the hip of the fetus or infant was elongation of the capsule. Subluxation and dislocation rep-

179

resented the same disorder, with dislocation simply being a more severe variant. The disorder was not embryonic in nature but occurred in the late fetal or immediate postnatal period. Any deformities begin to regress as soon as the hip is completely reduced such that, if complete reduction is maintained in a young infant, complete correction will occur. Bony changes in the acetabulum and femoral head were secondary to the displacement and not primary causes of it. Howorth and Smith described 72 cases of congenital dislocation of the hip treated by open reduction, with 16 under 3 years of age, 39 from 3 to 5 years of age, 14 from 6 to 9 years of age, and 3 at 10 years old or more (123). Complete dislocation was seen in 93%. There were no posterior dislocations in this group; all were anterosuperior dislocations. Common findings involved proximal femoral anteversion greater than 60 ~ in 20% and greater than 45 ~ in 58%, although 15 hips had an anteversion of less than 30 ~ The ligamentum teres was absent in 5, ruptured in 2, thinned in 7, and elongated and often thicker in the rest. In 56 hips, the capsule was found pulled up inferiorly with the transverse ligament across the lower portion of the acetabulum as the capsule followed the displacement of the head. In 1935, 39 additional patients were reported with Farrell involving 49 hips (72). At surgery, 9 of the hips were dislocated high posterosuperiorly and 40 anterosuperiorly. The empty acetabulum tended to fill inferiorly with hypertrophied tissue and ligamentum teres. The ligamentum teres was absent in 16, ruptured in 3, and separated in 2. The capsule was elongated and pulled up inferiorly but not constricted superiorly. The pathology appeared consistent with primary elongation or stretching of the capsule of the hip rather than dysplasia of the acetabulum or anteversion of the femur.

C. Most Recent Clinical-Pathoanatomic Descriptions with an Emphasis on Early Capsular Laxity 1. SCAGLIETTI AND CALANDRIELLO, 1962 Scaglietti and Calandriello provided a detailed description of the pathoanatomy of the congenitally dislocated hip in their description of correction by open reduction (267). Their material was based on 182 operative procedures in 162 patients in which 48% had the open procedure as a primary initial treatment. They divided the obstacles to reduction into two major groups, extra-articular and intra-articular. a. Extra-articular Obstacles The two major extra-articular obstacles are the relatively shortened gluteus medius and iliopsoas muscle groups. In long-standing cases of dislocation, particularly when the patient has become ambulatory, the gluteus medius muscle is shortened and must be released from its iliac crest origins in association with correction. The iliopsoas produces great problems because it normally passes over the anterior part of the hip joint capsule to insert into the lesser trochanter. When the head of the femur is not in the acetabulum, the iliopsoas tendon is stretched tightly and

180

CHAPTER 3 ~ Developmental Dysplasia of the Hip

crushes the capsule anteriorly against the mouth of the acetabular cavity. As the femoral head is displaced further, the pressure of the iliopsoas tendon against the capsule increases and helps cause the central capsular narrowing referred to as the isthmus, leading to the hourglass deformation. At times adhesions are formed between the capsule and the psoas tendon. b. Intra-articular Obstacles Capsule: The capsule is thickened and greatly enlarged due to the hip displacement that has occurred. In particular the upper part of the capsule bulges upward and outward as it functions as the main support structure for the displaced head. In many patients the capsule becomes adherent to the lateral wall of the ilium. Though normally the capsule reaches to the base of the neck, in some cases it becomes attached to the periphery of the head and the adjacent neck by fibrous adhesions. This finding is referred to as the "pericephalic insertion." As noted previously, the enlarged capsule is not shapeless but often has a central hourglass constriction that prevents passage of the dislocated head into the true acetabulum with closed reduction maneuvers. Limbus: Inversion of the labrum between the femoral head and acetabular cartilages can occur, although the structure interposed is not just the labrum but rather a double fold of thickened capsule and the enclosed labrum, which we refer to as the limbus. The interposed structure therefore is much thicker and bigger than the labrum alone. During the subluxation phase of any hip displacement the capsule often becomes adherent to the outer surface of the labrum, such that when total head dislocation occurs the inverted structure is both capsular and labral together (forming the so-called inverted limbus). Ligamentum Teres: The ligamentum teres is seen on some occasions but not others. When present it is thickened and elongated in some, whereas in others it is elongated, thin, and atrophic. On occasion it appears to have been absent congenitally. Head and Neck of the Femur: The commonly described anteversion of the femoral neck is seen in most instances, although it alone is not sufficient to prevent reduction of the head into the acetabulum. Acetabulurn: The two pathoanatomic aspects to the acetabulum include the presence of an enlarged ligamentum teres with fibrous and fatty tissue and a deficient cartilage coveting the roof of the acetabulum. 2. STANISAVLJEVIC,1964 Stanisavljevic wrote a monograph on congenital hip pathology in the newborn comparing both normal and dysplastic hips (287). Three hundred newborn hips were dissected (150 infants) with congenital pathology detected in 12 hips (all female). Four infants had bilateral and 4 unilateral congenital pathology involving 5 complete dislocations, 4 subluxations, and 3 dysplastic hips.

a. Normal Hip Structure Normal studies after all hip musculature had been removed but with the articular capsule and ligamentum teres intact revealed that it was impossible to dislocate or subluxate the femoral head from the acetabulum. The "click" that is often noted in normal hip exams was caused not by femoral head subluxation or dislocation but rather by an interposition of an unusually large ligamentum teres or by sliding of the iliopsoas tendon over the enlarged iliopectineal bursa. Anteversion of the femoral head-neck was between 25 ~ and 35 ~ and the average neck shaft angle was 140-145 ~. Asymmetric thigh folds often were seen in children with hips that were normal. b. Congenital Hip Subluxation A positive diagnosis of hip subluxation was made in four hips. Case 1: Bilateral findings. A full-term child died at 6 hr. No other congenital deformities were seen. The findings were indicative of prenatal congenital superoposterior subluxation. The iliopsoas tendon was hypertrophic. The capsule was loose. The acetabulum was of abnormal shape and filled with abundant pulvinar. There was not a normal limbus on the superoposterior aspect of the acetabulum, and in this location there was a well-developed sulcus through which the femoral head could luxate. The limbus was misshapen almost completely circumferentially and was found to be fused on its outer aspect with the internal side of the capsule, with no space present between the capsule and the limbus as in a normal hip. The ligamentum teres was hypertrophied. The femoral head was misshapen. The neck of the femur was shorter than normal. The neck shaft angle was 145 ~ and the anteversion was 65 ~. Case 2: Stillborn female with no other congenital anomalies. The tendon of iliopsoas muscle was hypertrophic and the capsule was loose. The femoral head was subluxated superoposteriorly and the misshapen and shallow acetabulum was filled with abundant pulvinar. There was no normal acetabular limbus on the superior-posterior aspect. The ligamentum teres was longer and thicker than normal. There was a well-developed sulcus on the superoposterior aspect of the acetabulum through which the head could subluxate. The femoral head was misshapen and smaller than normal. Case 3: Full term stillborn female. The iliopsoas tendon was hypertrophic and the capsule loose. The acetabulum was of abnormal shape and filled with abundant pulvinar. The limbus was not normal on the superoposterior aspect, showing a sulcus through which the femoral head could luxate. The limbus was misshapen circumferentially and found to be fused with the internal side of the capsule, with no space between limbus and capsule. The ligamentum teres was hypertrophied. The femoral head was misshapen, the neck of the femur was shorter than normal, the neck shaft angle was 140 ~ and anteversion was 48 ~ The findings were indicative of a prenatal congenital superoposterior subluxation.

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip The most common prenatal congenital hip pathology in the examples of subluxation was a defect of the superoposterior region of the acetabulum, which allowed the femoral head to sublux. The bone of the acetabulum in this region also was less developed and thus dysplastic. Studies of ranges of motion indicated that with the hip flexed 90 ~ or beyond and abducted 70-75 ~, the femoral head was wellseated in the acetabulum leaving the defect in the superoposterior aspect free from any pressure by the femoral head. c. Congenital Hip Dislocation Congenital hip dislocation was detected in 5 hips of 4 babies among the 300 hips studied. Each of these would appear to fall into the teratologic category. Case 1: Stillborn female at the 8th month of fetal life. No other congenital abnormalities were found. There was an unusually thick iliopsoas tendon. The acetabulum was very small, shallow, deformed, and filled with abundant pulvinar, which explained the easy "telescoping" of the femoral head noted on clinical examination. The ligamentum teres was longer and thicker than normal. The edges of the acetabulum were fiat and an intact acetabular limbus could not be detected. The capsule was larger than normal, thick at the superoposterior region, and within the acetabulum anteriorly and inferiorly obliterating a portion of it. On the posterosuperior aspect of the acetabular edge a sulcus was found that corresponded to the size of the ligamentum teres, which was causing pressure. The head of the femur was spherical but smaller than the opposite side. The femoral neck was very short and the head was retroverted 10~ The findings represented a congenital superoposterior dislocation of the left hip. Case 2: Stillborn female at 5 - 6 months. The right iliopsoas tendon was thicker than the left. Relocation of the femoral head was not possible by clinical manipulation. The entrance of the acetabulum was small because of the interposition of an enfolded acetabular limbus circumferentially. There was a high attachment of the capsule on the superior and superoposterior region of the ilium; no adhesions between the capsule and the external surface of the ilium were present. The capsule was larger and thicker than normal. The head of the femur was smaller although spherical and the femoral neck was short. There was no anteversion or retroversion. The findings represented a prenatal congenital dislocation. Case 3: Stillborn female at 4 - 5 months. No other congenital abnormalities were seen, and the iliopsoas tendon was of normal size. It was not possible to reduce the hip with the capsule intact. The entrance into the acetabulum was restricted almost entirely by the limbus, which left only a small opening for the ligamentum teres. The superior capsule was attached to the external surface of the ilium, forming a new acetabulum in which the deformed and flattened femoral head was found. The femoral neck was absent. The opposite hip had the exact same findings. Case 4: Female child died 12 hr after delivery. There was hypertrophy of the iliopsoas muscle and tendon. The capsule

181

was thick and tight anteriorly, and the femoral head was immediately under the capsule anteriorly and was dislocated anteriorly. The anterior-inferior portion of the limbus was compressed and folded into the acetabulum and the remaining limbus inverted. Pulvinar tissue was abundant. The femoral head and neck were of normal shape. The neck shaft angle was 145 ~ and anteversion was 55 ~ d. Dysplasia of the Hip Dysplasia of the hip was found in two instances without subluxation or dislocation. Following removal of muscle and tendon in one case, there was an increase of motion of the femoral head in the acetabulum but it was not possible to sublux or dislocate it. In another case the acetabulum was abnormal in shape but the femoral head engaged the acetabulum well with the thigh flexed to 80 ~ and beyond. The limbus was compressed and folded upward on the superoposterior portion of the acetabulum and compressed and folded peripherally in the anterior and anteroinferior part of the acetabulum. The femoral head was of a normal spherical shape. 3. SALTER,1968 Salter reviewed the pathogenesis of congenital hip dislocation stressing the importance of capsular laxity in initiating the displacement, with bony changes of acetabular and proximal femoral dysplasia being secondary and thus reversible with early relocation of the joint (263). His hypothesis of the sequence of events in the etiology and pathogenesis of congenital dislocation of the hip remains active today.

4. DUNN, 1969 Dunn reported a landmark series of several studies of normal and congenital dislocated hips assessed at necropsy (59). Twenty-two normal hip joints from fetuses with gestational ages ranging from 13 to 40 weeks were dissected postmortem, and he reported little change observed in the relative depth of the acetabulum or in the general morphology of the joint. "In no case was it possible to provoke subluxation either before dissection or following exposure of the joint capsule. Even after division of the capsule the femoral head remained snugly within the closely fitting limbus of the acetabulum, unless considerable force was applied to the leg with the femur in full adduction and external rotation." In a second group, 23 joints were dissected whose clinical examination had revealed hip instability (61). There were 15 infants with 8 bilateral cases and 7 unilateral with fight and left joints almost equally affected. All infants had died during labor or shortly thereafter with gestational ages from 27 to 44 weeks. Breech presentation had occurred in 8 cases. The variable pathologic changes were divided into 3 subgroups of grades 1, 2, and 3. Ogden has also adopted this classification approach (215). a. CDH Grade 1 There were seven examples of this type of deformity referred to as dislocatable hip. The head of the

182

CHAPTER 3 ~ Developmental Dysplasia of the Hip TABLE II

P a t h o g e n e s i s o f D e v e l o p m e n t a l D y s p l a s i a o f t h e Hip

A. Prenatal-Intrauterine Factors 1. Epidemiologic 4-8:1 Female:male incidence Breech position (especially after 30-34 weeks) Genu recurvatum-knee in extension Oligohydramniosa First child a Twins a Family history (genetic component), 1-5% 2. Structural Capsular laxity (hormonal, genetic) Irregularities of labrum

Often two or more features that individually are at upper range of normal but combined predispose to instability

Increased acetabular anteversion Increased proximal femoral anteversion

B. Postnatal Factors Sudden passive extension of hips at birth and immediately postnatally predisposes to subluxation and dislocation. Hips develop in utero in flexed position and are most stable in flexion. Infants positioned postnatally with hips extended and adducted are most prone to instability. C. Temporal Factors Prenatal onset If hip is positioned imperfectly in third trimester, structural changes at birth are relatively marked; the earlier the malposition occurs the worse the secondary changes because growth has occurred in utero in an abnormal position. Perinatal onset If hip is well-positioned throughout the third trimester, hip structure at birth is normal, or so minimally abnormal it is grossly undetectable, other than the capsular laxity that makes it dislocatable. Teratologic hip: The more severe variants of DDH are often referred to as teratologic. This term is used somewhat imprecisely but includes two general situations: (1) the abnormal hip occurs with other structural changes implying primarily a mesenchymal cell defect or (2) the hip development is normal until malposition occurs prenatally in the third trimester with secondary changes occurring in utero. The latter may well be the more common variant. The two variants can be defined as mesenchymal teratologic hip dysplasia and prenatal teratologic hip dysplasia. D. Postnatal Responses to Hip Instability 1. Spontaneous stabilization Spontaneous stabilization of dislocatable hips (without treatment) in normal position occurs in about 50% of cases. Usually within the first week, occasionally in the first 3-4 weeks. 2. Dislocatable hips with delay in diagnosis. Excellent results with early treatment. The later the time of diagnosis, the greater the secondar-y changes, the more complicated and prolonged the treatment, and in some the less excellent the results. aMinimize chances of spontaneous version.

femur was located normally within the acetabulum, but dislocation over the posterior or posterosuperior lip of the acetabulum was possible with relatively gentle backward pressure on the head of the femur with thighs flexed and adducted. Dunn felt that the crucial pathology "appeared to lie in the limbus itself," which was unstable being stretched and slightly everted in the posterosuperior aspect giving the

acetabulum an elliptical outline instead of the normal circular one. Dislocation was only partial in these cases with further displacement restrained by the capsule and the ligamentum teres. b. CDH Grade 2 There were four examples. The limbus was more everted, particularly at its posterosuperior margin, the capsule more stretched, and the ligamentum teres further

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip lengthened. Instability was marked and partial or complete dislocation usually was present at rest. The acetabulum usually was shallower than normal and the head of the femur frequently had lost some of its sphericity and was reduced in size. c. C D H Grade 3 There were 12 hips in this group. In each case the head of the femur was dislocated upward and backward, and the limbus particularly in its posterosuperior aspect was compressed and inverted into the joint so that it formed a partial floor of a false acetabulum. The ligamentum teres emerged through a crescentic gap bounded by its free margin. The acetabulum invariably was shallow and partially developed, and the head of the femur was smaller than normal and less spherical. Some of the infants otherwise were normal, although many had associated malformation of the neuromuscular system or urinary tract. Dunn felt that the whole spectrum of abnormality is a single pathological entity. He defined congenital dislocation of the hip as an "anomaly of the hip joint, present at birth, in which of the head of the femur is, or may be, partially or completely dislocated from the acetabulum." Dunn later reported additional observations as his series increased to 48 hips that were dislocated or dislocatable at birth that subsequently had been dissected. The studies were from 31 infants who died within a few days of birth with 16 bilateral and 15 unilateral cases of CDH. The gestational ages ranged from 27 to 44 weeks, and fully 16 of the infants (half of the cases) had presented breech. Clinical examination at birth had revealed instability of the hip in every case and this was confirmed under direct vision after exposure of the joint capsule. Dunn indicated that at least 1% of all newborn infants in Great Britain had congenital dislocation of the hip as determined by a careful neonatal examination. Of these, 85% were present in infants who otherwise were normal and 15% in infants who had many other malformations. In those infants who were formed normally, 90% had a CDH of grade 1 and only 10% had grades 2 and 3; there was an overall perinatal mortality of 5%. In the 15% that were malformed, only 50% had grade 1 CDH with 50% being grades 2 and 3 and perinatal mortality as high as 70%. Dunn used the term malformations to refer to abnormalities forming during the embryonic period that therefore were teratologic, whereas the term deformations was used to refer to deformities arising after the embryonic period in a normally formed part and considered secondary to extrinsic intrauterine pressure. The large majority of CDH cases, therefore, represented postural deformity imposed on an otherwise normal hip during the late periods of intrauterine position. Congenital dislocation of the hip frequently was found with other postural deformities, including torticollis, deformities of the skull, face, and mandible, and clubfoot. There was a high degree of association with breech presentation, first pregnancies, and oligohydramnios. In a large prospective study performed

183

over several years, Dunn noted that 56% of infants with CDH were first-born, whereas 50% had presented by the breech position. In the patients with CDH alone, there was a 4:1 female:male ratio. 5. MCKIBBIN, 1970; RALIS AND MCKIBBIN, 1973 McKibbin described findings in a child with bilateral CDH who died shortly after birth (194). Studies were compared with hip dissections in 15 intact pelves obtained from full-term infants who died in the first 2 weeks of life from causes unrelated to the musculoskeletal system. The child with the CDH was born at term from a breech position. The infant had been in the extended leg position with fully flexed hips and fully extended knees. The child died of cerebral hemorrhage with no other congenital malformations seen. The only significant pathoanatomic finding in both hips was excessive laxity of the capsule. Indeed at initial assessment with the muscles intact, it was possible to reduce the dislocation either by abduction and flexion or by abduction, extension, and medial rotation maneuvers. When the hip was flexed it actually was relatively loose and dislocated. Once the capsule had been removed it was noted that there was increased length of the ligamentum teres and also increased fibrous fatty tissue in the acetabular base (pulvinar). The acetabular labrum, however, was normal as were the acetabulum and proximal femur both to gross inspection and to measurements of angulation, which placed them in the normal range. McKibbin thus felt that the dislocation in this instance clearly began in utero. The sequence of events was initiated by primary laxity of the capsule, allowing the flexed hip to dislocate irrespective of the bony conformations, and the acetabular and proximal femoral changes were secondary. He interpreted the concept of acetabular dysplasia to be entirely secondary to the abnormal position of the head in relation both to this case and to other readings and interpretations of the literature. Ralis and McKibbin sought to assess the possible changes in the size and shape relationship of the femoral head to the acetabulum from embryo to midchildhood ages (246). They studied quantitatively 44 hip joints from 11.5 weeks of embryonic age to 11 years. Measurements of the acetabulum at each age included the greatest diameter and depth of the femoral head and the greatest diameter and height and the percentage cover of the femoral head. They confirmed the oft-disputed opinions of Sainton (260, 261) and Le Damany (163-167) that the human acetabulum was shallowest at the time of birth compared with both fetal and later postnatal findings. The perinatal period thus was the time when the risk of dislocation was greatest and dependence on soft tissue support was highest. In three embryos, the acetabulum was extremely deep-set and almost totally enclosed the femoral head. As growth continued its shape began to change, and as fetal age increased the acetabulum became increasingly shallow until at birth it represented as little as one-third of a

184

CHAPTER 3 9 Developmental Dysplasia of the Hip

complete sphere. After birth, the cavity steadily deepened again with growth. The femoral head shape also changed, with the least stability observed in the newborn period. In the embryo, femoral head shape was globular, coming closest to representing a complete sphere (about 80% of a complete sphere) than at any other time. As birth approached, it was closer in shape to a hemisphere, whereas after birth return to a globular shape partially recurred. The proportion of the head contained within the acetabulum gradually diminished with fetal growth, reaching a minimum around the time of birth but increasing again postnatally. These findings were felt to predispose to dislocation in the perinatal period because a joint consisting of more than half of a sphere inherently was stable, whereas a more shallow hip with only hemispherical (or less) relationships was more dependent on soft tissue support.

6. MILGRAM AND TACHDJIAN, 1976 Milgram and Tachdjian reported on a 10-month-old patient who died and who had multiple congenital anomalies (199). The right hip was dislocated and had been untreated. The hip disorder was considered to be teratologic. The severe changes, in particular in the acetabulum, at l0 months of age suggested an early intrauterine dislocation with subsequent malformation of hip development. The study showed a lax redundant capsule, an elongated large ligamentum teres, anterosuperior dislocation of the hip, hypoplasia of the true acetabulum, which was filled with fat and fibrous tissue, and an abnormal limbus. The fibrous limbus did not appear to represent an inverted labrum with adherent capsule; rather, no distinct tissue planes were present and the fibrous limbus appeared to emerge from the floor of the false acetabulum and project over the rim of the true acetabulum. Its appearance "appeared to have been locally induced by the presence of the dislocated femoral head." Laurenson also reported a similar finding in a 26-week-old malformed human fetus. a. Milgram, 1976 Milgram reported a case of bilateral dislocated hips in a 74-year-old male who had had no treatment for the disorder and also "never had pain referable to the hips and thighs," even though he was employed as a security guard (198). The left hip joint was disarticulated at postmortem assessment. The femoral head was oval and flattened medially but had a thin layer of fibrocartilage covering the articular surface. There were no degenerative changes of exposed subchondral bone or osteophytes. A thick fibrous capsule completely surrounded the femoral head separating it from the pelvis. There was no false acetabulum and the ligament teres was absent. The atrophic acetabular fossa was filled with fibrous tissue. The right hip joint was assessed following longitudinal sectioning. On this side also there was no bony contact between the femoral head and the ilium. The elongated thick capsule provided support to the femoral head. As on the opposite side there was no ligamentum teres,

the acetabulum was filled with fibrous tissues and was very shallow, and the femoral head, though slightly flattened medially and somewhat smaller in size, showed no degenerative arthritis. 7. WALKER, 1980-1983; WALKER AND GOLDSMITH, 1981 Walker and Goldsmith performed a series of studies in human fetuses from 12 to 42 weeks of age to document the developing structure of the proximal femur and acetabulum (337). Both hip joints (280) from 140 fetuses were assessed following elective abortion (62.2%), stillbirth (23.7%), and death during the perinatal period (14.1%). In an effort to concentrate on normal development and subtle variability, the hips had to demonstrate normal hip joint morphology by classic criteria and no displacement of the femoral head with the acetabulum. The joints were dissected, morphology was inspected, and measurements were taken of the depth and diameter of the acetabulum, the diameter of the femoral head, the length and width of the ligamentum teres, the neck shaft angle, and the anteversion of the proximal femur. Multivariate analysis showed no significant differences between males and females or between right and left sides. The acetabular depth was the slowest growing hip variable, and acetabular indices of less than 50% indicated a shallow socket at term. There was a strong relationship between the size of the femoral head and the acetabular diameter, but in many joints the femoral head diameter exceeded that of the acetabulum. Findings were interpreted to indicate that the soft tissue structures about the joint of necessity played an important role in neonatal joint stability. Except for the neck shaft angle, the means for all variables studied increased steadily with time, with the strongest increase between 12 and 20 weeks of age. Acetabular depth showed the slowest growth in the period studied, and a consistent linear growth trend was apparent in only the acetabular and femoral head diameters. Maximum values for proximal femoral antetorsion were not observed at term but rather at 32 weeks. In a number of hips the femoral head diameter exceeded the acetabular diameter such that deep seating of the head within the socket was not possible. In younger fetuses, after cutting the capsule some force was required to displace the head from the socket, but in older fetuses division of the capsule produced immediate subluxation or dislocation of the head from the socket. Socket coverage of the femoral head was increased in the flexed position in utero and any movement of the femur out of this position decreased socket coverage of the head. One observation not previously appreciated by many was that in 56% of the femurs the lesser trochanter was more prominent than the greater. The data were then interpreted in relation to congenital hip disease. Although there is a clear clinical female preponderance of hip dysplasia, the study showed no difference in fetal development of the acetabulum or femur of the

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

hip joint either between males and females or between right and left sides. This information was considered to support indirectly the hypothesis that the preponderance of DDH in females was due to a greater influence of maternal sex hormones on female fetuses, allowing for capsular laxity and hip displacement. As early as 1905 Le Damany had felt that the acetabular socket was shallower at term than at any other period of fetal life. This appeared to be true in the present study as well in the sense that, whereas femoral head and acetabular diameter increased by more than 4-fold in the period studied, increasing depth of the acetabulum was less than 4-fold. Because of the relative shallowness of the acetabulum, soft tissue structures of necessity played an important role in the stability of the hip joint in older fetuses and in the newborn. Many joints demonstrated a position of maximum fit or congruency in which there was maximal coverage of the femoral head by the socket; this corresponded to the normal position in utero. This observation was made most frequently in third-trimester fetuses. The neck shaft angle measurements were somewhat lower than those published previously based on radiographs and also there was no apparent change in the angle with age. The problem with radiographic measurements is the change with rotation in which internal rotation increases the angle and external rotation decreases it. It thus appears that the greatest increase in this angle occurs during early fetal life, and the values of 125 ~ were similar to those noted in the adult. Femoral torsion changes were seen, with the mean value of femoral torsion at birth around 35 ~. Many had noted neutral values up to 24 weeks of development, with Le Damany and Watanabe in particular feeling that antetorsion develops almost exclusively in the second half of pregnancy. With an increase in the amount of positive torsion there was a change in the structure of the proximal femur, which led to the lesser trochanter becoming directed more medially. The opinion was given that torsion appeared to take place in the femoral shaft and not specifically at the head-neck area. The torsion values were distinctly lower at term in this study than those reported from studies using radiographic measurements, and increases were noted to occur in the 12- to 18-week period. The mean adult value of 11.2 ~ was exceeded in the study by a relatively small 62% of femurs. With a high correlation shown between the depth of the acetabulum and acetabular and femoral head diameters, such data could be useful clinically. If so the current most acceptable method of obtaining that data would be by ultrasound. Walker documented a considerable degree of morphological variability in the human fetal hip joint in his study of 280 hips and 140 normal fetuses (334). Sixty-five of 92 hips in 46 fetuses between the ages of 12 weeks and term showed structural variants, although they were neither subluxated or dislocated and showed no other statistically significant morphological differences from normal joints. Variants observed included flattening (14) or rounding of the rim of the labrum,

185

localized dips in the labrum (20), folding of the labrum (6), capsular folds (4), extension of the pulvinar pad between the joint surfaces (6), and kinking of the ligamentum teres (7). Most fetuses with variants showed more than one variant. The vast majority of these were localized to the anterosuperior quadrant and increased with fetal age. The variant hips formed 55% of the fetuses older than 28 weeks and only 23% of those younger than 28 weeks. These findings are supportive of the belief that these variants potentially would be associated with congenital-developmental hip disease. The structural changes of hips thus would follow a continuum from normal hips to variants of dysplastic-subluxateddislocated hips. There was rounding of the rim of the labrum with the extent of the rounding being variable, ranging from the entire circumference to only one quadrant. Flattening of the rim also was seen. The flattening allowed some inward fold and overhang of the labrum on the articular surface. This led to a slight ledge instead of the normal smooth transition at the junction between the inner margin of the labrum and the articular cartilage of the joint surface. A fold of the capsule projected into the acetabulum in four hips. The fold was always in the anterosuperior quadrant of the acetabulum and was associated with flattening or rounding of the labrum. In six hips there was a thin relatively avascular sheet of fibrous tissue that was connected with the pulvinar centrally in the acetabular psoas, extended over the articular cartilage, and had a free peripheral head. The tissue thus was interposed between the socket and the head of the femur and was not connected to the synovial membrane. On occasion, kinks were seen in the ligamentum teres sometimes at the acetabular origin and sometimes at the femoral insertion. On many occasions several of the previously listed mild variations were present in a single hip. It was not possible to say whether these would have led to dislocatable hips at birth. Clinical examination of the intact specimens, however, indicated stability, but obvious problems with clinical interpretation are still present. Several variants, some of which might well be associated with a dysplastic hip eventually, were well-documented in a large series. Walker also performed a study on 12 abnormal fetal hip joints that were not described as variant hips because they had more frequent or more complex aberrant features, the impression of dysplasia grossly or actual malalignment between the femoral head and acetabulum at dissection (336). Walker found that 10 of the 12 abnormal hip joints had several dimensions less than the range for their comparative normal age group by 2 standard deviations or greater. Two general features characterized the findings. One was that the neck shaft angle rarely was in the abnormal range, indicating that in fetuses and neonates abnormality of the femoral angles alone was a poor indicator of hip dysplasia. The major irregular finding in virtually every case related to the soft tissue component of the acetabular rim, with flattening

186

CHAPTER 3 9 Developmental Dysplasia o]r the Hip

particularly of the anterior and superior rim and abnormalities of the labrum, which was described as indistinct, flattened, folded inward, rounded, or associated with a dip rather than being spherical. Increased amounts of fibrous fatty tissue also were seen in the depths of the acetabulum. 8. PONSETI, 1978; IPPOLITO ETAL., 1980 Ponseti studied the acetabulum in six infants who died soon after birth and had unilateral hip dysplasia (233). Findings of a thickened capsule were similar to many previous descriptions. In most there was a cartilaginous ridge in the acetabulum that separated the hip socket into two sections. These appeared to be what some might refer to as a teratologic hip because of both the severity of the findings and the neonatal deaths. In three of the hips the ridge was formed exclusively by a bulge of acetabular cartilage as defined histologically. The labrum thus was not interposed but was everted and plastered against the capsule. The surface of this acetabular cartilage was covered by a thin layer of fibrous tissue. In the other three hips the more commonly described finding was noted, in that there was an inverted labrum that covered the bulge of acetabular cartilage. In each instance, however, it was noted that the more medial part of the acetabular complex was both anatomically and histologically normal; this encompassed the triradiate cartilage and the adjacent acetabular cartilage and bone. It thus is evident from multiple assessments that the bulk of the acetabular dysplasia in association with C D H - D D H is occurring laterally due to abnormal pressure from a subluxed head or no pressure at all from a fully dislocated head. Ippolito et al. performed histologic, histochemical, and ultrastructural studies on hip joint capsule and ligamentum teres from 9 patients (12 hips) with CDH from 2 months to 4 years 2 months of age (131). They concluded that the slight changes seen were secondary to mechanical stresses on the tissues caused by the dislocation rather than representing primary pathological abnormalities. The collagen fiber bundles were thicker than in the normal capsule but appeared of normal shape ultrastructurally. In the ligamentum teres the elastic fibers were thicker and more numerous in the dislocated hips. 9. SOMERVILLE,1982 Somerville mentioned the causes of dislocation of the hip in infancy as resulting from (1) the effects of muscle imbalance, (2) congenital contractures, (3) intrauterine compression, (4) true teratological abnormalities, or (5) factors that are suspected but still unproved (285). Only in the last group, in which the large majority of patients were concentrated, was the dislocation an entity unto itself that warranted being called true "congenital dislocation of the hip." Somerville discounted the theory of primary acetabular dysplasia as being a common cause of CDH due to the fact that 60-70% of all cases spontaneously stabilized in the first week of life and the large majority of the others responded with cure to simple splinting over a several-week period. Capsular laxity was

felt to be a major cause of the disorder. Capsular laxity can be generalized as in connective tissue disorders such as the Ehlers-Danlos syndrome. Laxity could develop in a part of a capsule only while the rest remains normal or even contracted. This is the common type with CDH. A hormonal relationship with capsular laxity has been suggested and was attractive theoretically, fitting many of the facts of the disorder, although no definite proof of such a relationship has yet been defined. Somerville's major contention is that displacement occurs as a result of stretching of a previously weakened capsule and that it is a part of the capsule that is stretched rather than the whole. In a large majority of instances "the only abnormality is the capsular laxity" with all other changes secondary. All subluxations initially are anterior and anterosuperior, and by definition dislocations, which only a more severe variant of subluxation, also initially are anterior. Subluxation and dislocation are part of the same condition, not two different conditions. Displacement occurs when a newborn hip is placed suddenly into extension, in particular if proximal femoral anteversion or lateral rotation of the hip is associated. Any extension of the hips in the newborn child, forceful or not, potentially is dangerous. Displacement thus occurs after birth when the hips are extended for the first time. Posterior positioning develops with time in the untreated as capsular enlargement greatens and the child begins to walk but is rarely seen under 4 years of age. Somerville clearly indicates the difficulties that occur when the limbus is inverted into the joint. He uses the term limbus to define the normal fibrocartilaginous outer rim of the acetabulum. Eversion of the limbus and subluxation take place slowly, whereas inversion of the limbus takes place rapidly and may be established within a few months. The inverted limbus prevents the head from entering the acetabulum fully. Somerville also indicates that the neck shaft angle at the proximal end of the femur at birth is normal to only minimally altered. The main femoral deformity is the angle of anteversion, which is increased generally to as much as 45 or 50 ~ and may be increased as high as 90 ~. The hips with greater degrees of femoral anteversion are more at risk for dysplasia and dislocation. He records the usual angle of anteversion at birth as being 25-35 ~ The pathogenesis of the disorder thus is a mechanical one characterized by capsular laxity and femoral anteversion. If capsular laxity does not spontaneously correct itself in the first few weeks of life or if splinting in flexion is not done, the disorder proceeds to a hip subluxation and then to anterior dislocation, which eventually over a few years is converted to a posterior dislocation. Somerville continually stresses the importance of the position of the limbus, which in subluxation is not inverted and does not cause an obstruction to reduction but in dislocation is inverted and causes an obstruction to concentric reduction. Arthrography is important in assessing the position of the limbus. With a complete dislocation the limbus is turned into the joint; the posterior part of the limbus is always inverted.

SECTION IV ~ Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

Pooling of the dye in the joint indicates that the femoral head is not in that particular area and that the obstruction must be at the periphery of the acetabulum where dye does not appear. It has been suggested by Severin (278, 279) that, if the head is kept in relationship to the acetabulum, even with an inverted limbus the head will wear out the limbus or force it into the appropriate position. Somerville disagreed, feeling that a false reduction would lead to damage to the associated tissues and almost certainly to premature cartilage deterioration. Renshaw (252) also has shown that soft tissue interposition between femoral head and acetabulum that is accepted following closed reduction represents an inadequate reduction that can lead to poor results. An inverted limbus thus must be removed or repositioned to allow for a concentric reduction. Somerville himself favored excision of the limbus; at hip arthrotomy "usually the inverted limbus will be seen quite easily, looking rather like a medial meniscus." The inverted limbus then is excised, after which the hip reduces uneventfully when the leg is rotated internally. This closes the gap in the capsule so that no capsular suture even is needed. The patient is placed in hip spica in the internally rotated position for 1 month. In many instances a proximal femoral osteotomy is performed to correct anteversion with some degree of varus also built into the correction. Figures 3A-3H show many examples of the pathoanatomy of congenital-developmental hip dysplasia from works of several authors. Table II summarizes pathogenetic factors in developmental dysplasia of the hip.

D. Multifactorial Causes of DDH Involving Late Stage Structural Modifications of the Hip, Mesenchymal Tissue Abnormalities, and Intrauterine Mechanical Stresses Due to Positioning 1. OVERVIEW Increasingly it is considered that DDH occurs in relation to several associated factors because multiple pathoanatomic dissections along with clinical, radiographic, arthrographic, and sonographic studies almost invariably fail to reveal marked structural abnormalities in the perinatal period. There is widespread acceptance of the fact that there are two prime factors in the mechanism of late intrauterine and perinatal displacement of the hip: hip capsule laxity and mechanical stresses placed on the fetal hip joint. Some observers still consider that subtle primary hip abnormalities, often described as variations of normal development, in combination predispose one to hip dysplasia. These involve acetabular labrum abnormalities, a more shallow and anteverted acetabulum, and increased proximal femoral anteversion. Wilkinson (354-357) and Seringe et al. (275, 276) have studied and summarized well the multifactorial considerations. Both feel that the mechanical factors are most important and that DDH is associated with a characteristic position in utero involving hyperflexion of the hip, with adduction and exter-

187

nal rotation causing an abnormal pressure on the greater trochanter leading to an expulsion of the head upward and backward. Wilkinson further demonstrates that in effect it is the hyperextended position of the knees that places further increased stress on the fetal hip joint. Seringe et al. sought to explain how the mechanical intrauterine forces on the hip could account for the tendency to capsular stretching and subsequent instability regardless of the position of the fetus in relation to the maternal pelvis. They felt that the knee hyperextension theories of Wilkinson and others related only to the breech position. Their theory thus was considered to be equally valid in full breech presentations whether the knees were flexed or extended and in the cephalic presentations in which the knees were almost always flexed. They identified three dislocating postures, each of which was characterized by the damaging role of lateral rotation of the lower limb as indicated by Wilkinson. The first position had the knees extended and the lower extremities in lateral rotation, the second had the knees semiflexed but also in lateral rotation, and the third had the knees in full flexion and in contact with each other in neutral rotation but with an excessive femoral anteversion, which is equivalent to lateral rotation in terms of the effect of the proximal femur on the acetabulum and capsule. The laterally rotated position in some way was caused by fetal-maternal mechanical compression. In these situations, which are most common with breech presentations, there is failure of the leg folding mechanism. In cephalic or vertex presentations only 1% of babies have the knees extended. Much of the information on the multifactorial influences on CDH-DDH has been drawn from epidemiologic studies on large numbers of patients. These findings and their relationship to the pathophysiology of the disorder are presented in more detail in Section V. Seringe et al. have reviewed both the intrinsic and extrinsic causes. 2. INTRINSIC CAUSES These include a primary acetabular dysplasia, subtle imperfections of the anatomic structure of the labrum, increased anteversion of the femoral head and neck, and capsular laxity. Primary acetabular dysplasia is accepted only rarely as a primary causative mechanism at the present time because the acetabulum is normal or virtually normal in almost all hip dysplasia patients at birth, and, more importantly, concentric and maintained reduction of the femoral head in the first year of life leads to rapid normalization of acetabular development. A primary dysplasia of the acetabulum can be considered to exist only where abnormal development persists, even with clear concentric reduction. Proximal femoral anteversion and capsular laxity are considered to be predisposing causes to DDH. Seringe et al. considered capsular laxity to be secondary to displacement of the femoral head and not the primary cause of the dislocation because their studies defined not a diffuse laxity initially but rather "an elongation of the posterosuperior part of the capsule" (275, 276). The capsular laxity may be due to

F I G U R E 3 Several examples of pathoanatomy of the hip in long-standing congenital hip dislocation are shown. (A) The hip on the right is normal and that on the left abnormal [derived from (171)]. On the right the normal acetabulum and the ligamentum teres (b) are shown. On the left note the dysplastic hip with the small and misshapen acetabulum (b), spacious false acetabulum (d) superior to the original acetabulum and indented into the lateral iliac wall, the elongated ligamentum teres (c), and the misshapen head-neck region (a). (B) The normal hip is shown at left and the abnormal dislocated hip at right [from (182)]. The acetabulum on the abnormal side is smaller than that on the normal side, is triangular in shape, and shows an enlarged capsule (pointer) with flattening of the acetabular rim and a more spacious joint. (C) A dislocated hip is seen with elongated ligamentum teres, stretched capsule, small and relatively shallow original acetabulum, and more spacious pseudo-acetabulum above and laterally [from (182)]. (D) An illustration of Deutschlander (1910) clearly delineates understanding of the pathoanatomy of the hip dislocation [from (56)]. The capsule (g) is enlarged and elongated and provides most of the superior support for the displaced femoral head. The acetabulum is shallow and oblique and filled with fibrous fatty material. The capsule and transverse acetabular ligament (e) are pulled across the opening of the acetabulum by the displacement. (E) Developmental changes in the acetabulum in which the hip remains dislocated were appreciated in early pathoanatomic studies of CDH [from (167)]. A cross section of the acetabulum in the normal childhood hip is illustrated at right and one from a dislocated hip at left. Note that the acetabulum is shorter, is shallower, and has rounded margins on the involved side. (Fi) Many studies appreciated the structural differences between a hip subluxation (right) and dislocation (left) [reprinted from (174), with permission]. Leveuf felt that each entity was separate in the sense that the subluxed hip did not become a dislocated hip, although most others before and since felt that worsening subluxation would lead to dislocation. In the subluxed hip (right), the acetabulum is shallow and oblique, but the labrum remains on top of the upwardly displaced femoral head and is plastered against the capsule, whereas in the dislocated state at left the labrum and adjacent capsule are inverted into the joint and the femoral head is displaced even further and supported only by the

SECTION IV 9 Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

189

F I G U R E 3 (continued) capsule superiorly. (Fii) The more commonly accepted view illustrated here is that with displacement the hip passes from a normal anatomic arrangement (left) to a subluxed state (center) to a fully dislocated state (right) [reprinted from (267), with permission]. Crucial to the normal anatomic alignment is the radiolucent roof of the acetabulum, which in the normal state is composed of the acetabular cartilage plus the triangular fibrocartilaginous labrum and the adjacent capsule, which attaches beyond the labrum into the side wall of the acetabular bone. With subluxation the head moves laterally and upward but the labrum remains on top of the head, although it tends to become flattened. In the fully dislocated state the head is displaced onto the side wall of the ilium, and both the labrum and part of the capsule are inverted to lie between the acetabular cartilage and the femoral head cartilage. The inverted labrum and capsule together are referred to as the limbus. (G) Wilkinson illustrated the differentiation of a reducible hip displacement from an irreducible hip displacement position of the limbus postreduction [reprinted from (356), with permission]. At left, the head is fully and concentrically reduced into the acetabulum, and the posterior fibrocartilaginous labrum and thickened redundant capsule are displaced to their normal position. At fight, the hip is deemed irreducible or imperfectly reduced because even though the head is moved into reasonably good position with relationship to the acetabulum, there is inversion of the labrum and thickened capsule (the limbus) preventing fully concentric femoral head-acetabular conformation. (H) A teratologic dysplastic neonatal hip is shown [reprinted from Milgram, J.W. (1976). Clin. Orthop. Rel. Res. 119:107-111, 9 Lippincott Williams & Wilkins, with permission]. The displaced femoral head is at right and the dysplastic acetabulum at left. Note the elongated ligamentum teres, increased fibrous fatty tissue in the depths of the acetabulum, and a flattened malpositioned capsule and labrum interposed between the acetabular cartilage and the femoral head cartilage.

a genetic connective tissue abnormality or to a perinatal hormonal imbalance in which circulating "relaxing hormones" from the mother are superadded to the hormones of the child, leading to the tendency to capsular laxity in the latter. This theory is attractive theoretically and correlates well with the much higher female incidence of DDH in the range of 4-8:1. It was proposed several decades ago, but no truly definitive demonstration has been made. 3. E X T R I N S I C CAUSES The extrinsic or mechanical causes of DDH refer to intrauterine forces on the fetal hip and help to explain the association of DDH with many epidemiologic features. The postural-mechanical cause of the disorder is supported by the high incidence of DDH and breech positions, high incidence with genu recurvatum, torticollis, and foot deformities, its association with first births (when the uterine and abdominal muscles are relatively tight), and its association with oligohydramnios in which there is diminished intrauter-

ine fluid limiting spontaneous version in the third trimester, twin pregnancies, and excessive birth weight. The mechanical predisposition can also be defined as being due to intrauterine factors, factors at the moment of birth at which time the hip is relatively forcibly hyperextended, and factors in the immediate postnatal period in which those groups that position the infants with the hips flexed and abducted tend to have a much lower incidence of long-range problems than those in which the child is placed with the hips extended and adducted. The latter two factors are not considered to produce hip instability or dislocation but rather to diminish the likelihood of spontaneous stabilization occurring. Seringe et al. defined a unifying theory for the pathogenesis of CDH-DDH. Dislocation of the hip was produced mechanically by a combination of two factors. The first involved the position of the proximal femur in which the head was not directed toward the depths of the acetabulum but rather toward the rim of the cavity and the capsule; this was the dislocating posture. The second feature required a force

190

CHAPTER 3 9

Developmental Dysplasia of the Hip

pushing on the proximal femur to lead to its displacement; these forces involved pressure on the greater trochanter and also relative overactivity of certain muscle groups with the predisposed dislocating posture primarily involving the psoas, adductor, and hamstring muscles. The dislocating posture involved excessive flexion with a certain degree of lateral rotation or an excess of proximal femoral anteversion and some adduction. In summary, it was pressure on the greater trochanter of the femur in lateral rotation or with excessive anteversion that served to drive the femoral head above and behind the acetabulum. In the case of breech presentation the pressure on the greater trochanter arose from contact with the narrow upper part of the paternal pelvis, whereas in cephalic presentations contact arose from the maternal lumbar spine, which helped explain the greater frequency of left unilateral dislocations because the fetal spine is more often on the left. The dislocation thus develops at the end of fetal life or even during the period of labor under the primary influence of the mechanical factors. At birth the dislocated hip is freed from intrauterine constraints and two possible pathways of development can occur. If the instability persists, the dislocation is perpetuated and becomes progressively more irreducible leading to a congenital dislocation of the hip. The other pathway involves spontaneous improvement, which if complete leads to a normal hip and if incomplete can leave the patient with a subluxation or residual acetabular dysplasia. The majority of cases labeled as teratologic really were just straightforward congenital dislocations that appeared structurally worse because they had occurred earlier in the intrauterine period such that the secondary changes were more marked at the time of birth. E. E x p e r i m e n t a l R e p r o d u c t i o n o f Hip Dislocation

A few experimental studies have (1) effectively reproduced the mechanism of dislocation as seen in breech malpositions by holding the knee in extension and (2) shown the secondary adaptive changes of the acetabulum in those situations in which the femoral head was not present in the acetabulum during early development, as occurred due to capsular laxity in causing DDH. The concept of capsular laxity, defined first by Sedillot (274), also was accepted by Laurent (157) and Langenskiold et al. (158). Experimental dislocation of otherwise normal hips in experimental animals was followed shortly by dysplastic changes similar to those seen in humans. The work in rabbits (158) and puppies (284) lent support to the belief that stretching of the soft parts of the joint was the primary cause. 1. PRODUCTION OF SUBLUXATION AND DISLOCATION OF THE HIP BY MAINTENANCE OF THE KNEE IN THE EXTENDED POSITION IN YOUNG RABBITS

Michelsson and Langenskiold were able to produce subluxation and dislocation of the hip in rabbits less than 3 weeks

old by placing the knees in extension in plastic tubes and maintaining that position for several weeks (197). The hip was in normal position and the animal was allowed to move about. In another group of animals the knee was held in extension but the hamstring muscles were released surgically either proximally or distally. In the group with the knee immobilized in extension, 83 of 87 hips developed some abnormality involving hip dysplasia, subluxation, or complete dislocation. In another group of 7 rabbits in which the hamstrings were cut, no hip changes were noted. In 8 rabbits both hind limbs were immobilized in the same way. On one side the hamstrings were cut leading to no hip changes, whereas on the other they were left intact leading to dislocation in 4 and subluxation in 4. Investigation clearly demonstrated that in almost all young rabbits less than 3 weeks of age it was possible to cause subluxation or dislocation of the hip, along with most of the secondary changes characteristic of human CDH, by simply immobilizing the knee in extension for 3-4 weeks. In these experiments the hip was allowed to move freely and the animals were noted to keep the hip in a normal position of flexion. Dislocation occurred despite the fact that the hip was free to move and was in a normal position. They concluded that "the only obvious pathological factor in the region of the hip provoked by immobilization of the knee in extension was increased tension in the hamstring muscles." That conclusion was supported further by a finding that, when the hamstring muscles were cut, immobilization of the knee in extension was not followed by any positional changes. They concluded that this well reproduced the mechanism of dislocation in the human breech position in which hip dysplasia was far more common than in children born in the vertex position with the knees flexed. 2. DEVELOPMENTAL CHANGES IN THE ACETABULUM FOLLOWING EXPERIMENTAL DISPLACEMENT OF THE FEMORAL HEAD DURING EARLY GROWTH

It has been unclear whether the developmental acetabular and proximal femoral changes in CDH-DDH are primary abnormalities leading to displacement or whether they occur secondary to the displacement due to growth in an abnormal position. Experimental studies have addressed the matter by removing the femoral head surgically from the acetabulum in early life and following development of the bony segments with time. Smith et al. performed an experimental displacement of the hip in young puppies, dislocating the fight hip surgically in each of 22 animals under anesthesia while leaving the contralateral hip in place (284). Study then was performed from 4 to 8 weeks following dislocation. As a consequence of experimental dislocation of the hip in puppies 3-5 weeks of age the following changes were observed: (1) acetabular dysplasia as early as 4 weeks following dislocation; (2) progressive dysplasia to the point of an unrecognizable acetabulum at the time of maturity; (3) normal acetabular development following experimental dislocation and imme-

SECTION IV 9 Etiology and Pathoanatomy of Developmental Dysplasia of the Hip

O)

a

b

.Or)

(b)

(g)

191

(r

(h)

F I G U R E 4 Illustrations from the work of Harrison, who induced femoral head dislocation in the young rat and studied secondary developmental changes. (A) The shape of the normal acetabulum is shown at left (a) and that of the acetabulum, which developed in the absence of a femoral head, is shown at right (b). The abnormal acetabulum is much smaller and shallower than the normal and is misshapen with an oval appearance compared to the normal circular shape. (B) Cross sections of the acetabulum in two experimental groups are shown. The changes are quite similar to those outlined in human developmental dysplasia (Fig. 3E). Parts a above and f below show normal acetabulae. The figures to the right in each instance (b, c, g, and h) are abnormal. There is decreased width of the acetabulum, decreased depth, and a general misshapen appearance. [Parts A and B reprinted from Harrison, T. J. (1961). J. Anat. 95: 12-24, with the permission of Cambridge University Press.]

diate relocation; (4) pronounced changes in the head and neck of the femur in the dislocated state; and (5) no abnormalities in the head and neck of the dislocated hips in which the femurs were relocated. They thus felt that it was the displacement of the otherwise normal femur in relation to the acetabulum that led to the secondary developmental changes. Similar findings in the rabbit were produced by Langenskiold et al. (158). The fight hips of rabbits 1-5 days of age were dislocated under anesthesia by manipulation. Findings then were followed in 101 animals by radiographs at varying intervals. Histologic sections also were made. In the study, 29 of the 101 hips spontaneously reduced and subsequently developed in normal fashion. In those that remained dislocated, many of the features of CDH in the human were seen. These included dysplasia of the acetabulum, dysplasia of the head of the femur, anteversion of the femoral neck, capsular distortion with formation of an isthmus and limbus, and acetabular dysplasia without dislocation. Histological analysis of 23 animals demonstrated changes in the acetabular roof, limbus, and femoral head similar to those seen in human CDH. Their interpretation, similar to that of Smith et al., was

that the changes in the femur and acetabulum were secondary to the displacement, which itself could be described as due to some abnormality of the soft tissues. Histologic sections of the displaced femur and acetabulum clearly show that the acetabular dysplasia occurs almost exclusively due to growth abnormalities lateral to the triradiate cartilage, either due to pressure of the femoral head against the lateral acetabular cartilage or due to complete displacement of the head leaving the cartilage to grow without any mechanical stimulus. Harrison performed a classic study demonstrating the influence of the femoral head on pelvic and acetabular growth and shape in the rat (112). He performed several surgical procedures leading to unilateral and bilateral excision of the femoral head, dislocation of the hip, and amputation of the entire lower extremity, removing the femur completely and leaving the acetabulum to develop in the absence of mechanical pressures. Acetabular development was grossly and histologically abnormal in each of these groups (Figs. 4A and 4B). The acetabulum was narrower, shallower, and smaller on the operated side. The rim of the acetabulum gradually lost its sharp edge and became blunt and ovoid with the long

192

CHAPTER 3 ~

Developmental D[tsplasia of the Hip

axis running dorsoventrally. Histologic exam showed the capsule to invert into the socket along with the rim of articular cartilage. The acetabulum became filled with fat and sealed laterally with a layer of fibrous tissue. Histologic sections showed a particular maldevelopment of the lateral acetabular cartilage with the triradiate cartilage being relatively unaffected. Even histologically the iliac, ischial, and pubic growth cartilage plates within the triradiate cartilage in the medial wall of the acetabulum were unaffected in either thickness or histological structure. Subtle variations were seen in each of the different approaches. In addition, there were general pelvic growth abnormalities, although the acetabulum was affected most severely. The major disturbances of acetabular development included failure of the socket to develop in area and depth, blunting irregularity of the acetabular margin, and atrophy and degeneration of the articular cartilage. It was the acetabular cartilage of the lateral regions that was affected, with the triradiate growth cartilages medially remaining histologically normal and for the most part supporting total length growth of the innominate bone. He also concluded that "it is more likely that the dysplasia found in congenital dislocation of the hip is the consequence rather than the cause of the dislocation." The experiment showed yet another example of alterations in the pattern of mechanical stress, leading to growth abnormalities due to absence of the pressure of the femoral head against the acetabular cartilage. Absence of the femoral head from the socket resulted in marked acetabular dysplasia and pelvic asymmetry. 3. BREECH MALPOSITION AND HORMONAL LAXITY CAUSING HIP DISLOCATION IN YOUNG RABBITS

Wilkinson reproduced the features of hip dysplasia in 6to 8-week-old rabbits by splinting one of the hind limbs in full extension to reproduce the breech posture and by giving the animals estrone and then progesterone to induce joint laxity hormonally (354). Assessments then were done after 6 weeks of splintage and hormone therapy. Both male and female rabbits were used. Dislocation was produced preferentially in the females, who had the combination of the hormones and the knee extension splinting. Dislocation also was more common with lateral femoral rotation breech malposition along with the presence of hormonal joint laxity. Subsequent anatomic studies showed soft tissue and bone deformities in the dislocated regions quite consistent with those seen in the human disorder. With lateral femoral rotation there was inversion of the posterior capsule of the hip along with swelling and fibrosis. Wilkinson felt these were analogous to the limbus. The ligamentum teres often was thickened but never ruptured. When the knee extension and hormonal joint laxity were performed in isolation, no dislocations occurred. All displacements occurred when both were acting, indicating that both were essential to the dislocating mechanism. Medial and lateral rotation of the hip was dictated by the mode of application of the knee hyperextension splinting.

Hormones induced joint laxity. One of the primary epidemiologic findings in all studies with congenital hip disease is the high degree of female preponderance. This generally has been considered to be associated with the increased circulating female hormones present in the female fetus, both acquired from the mother and present in the child. During pregnancy the hormonal level rises in the mother, allowing for progressive softening and lengthening of the matemal pelvic ligaments and producing a pliancy of the birth canal that makes labor easier. During the second trimester, placental estrone and progesterone pass into the fetal circulation. These hormones produce minor degrees of fetal laxity by their direct action on the developing ligaments. In the female fetus they stimulate the immature uterus to produce relaxin. Andren and Borglin indicated that diminished hepatic function in children with congenital dislocation of the hip allowed circulating levels to be even higher and thus increasing the hormonal laxity (7). Delgado-Baeza et al. also produced pelvic deformity in 2-week-old rats by experimental methods involving immobilization of the knee in extension, hormonal alterations, surgical dislocation by capsular release and section of the ligamentum teres, and resection of the femoral head (54). 4. MECHANICAL INDUCTION OF HIP DEFORMATION AND DISLOCATION IN VITRO

Hjelmstedt and Asplund have studied human hip stability by mechanical means, inducing displacement in vitro in infant autopsy specimens (119). Loading of hips at 45 ~ flexion induced deformation and dislocation similar to that seen clinically with capsular stretching and no gross structural damage; in addition, periods of time as short as 3 hr sufficed to lead to displacement. Their obvious conclusion was that hip displacement could be induced in vitro solely by moderate mechanical forces in a short period of time.

V. E P I D E M I O L O G Y

AND ITS RELATION TO PATHOPHYSIOLOGY

Certain epidemiologic considerations are important in understanding the underlying bases of congenital-developmental hip abnormalities.

A. Sex Incidence There is a markedly higher incidence of the disorder in females compared to males. In major series reported over the past several decades, females account for roughly 80% of the cases, a 4:1 ratio. In a large study from New York City, the female:male incidence of unstable hips was 4.15:1 (8). Many series show an even higher female predominance: Putti, 5.7:1 (in 1879 cases) (240); Farrell et al., 5.7:1 (in 310 cases) (73): Lempicki et al., 6.6:1 (in 1010 cases) (171); and Grill et al., 8:1 female:male predominance in 2636 dysplastic hip patients (100). Somerville described typical congeni-

SECTION V ~ Epidemiology and Its Relation to Pathophysiology

tal hip dislocation to be 8 times as common in girls as in boys (285).

B. Incidence and Side of Hip Instability The incidence of hip instability is quite high and generally varies between 1% and 3% of all newborns, depending on the region of the world where the assessments are made and the inclusion criteria used. A large study of 23,408 patients in New York City from 1966 to 1972 showed an incidence of 1.33% involving dislocatable or dislocated hips (8). These showed 82% to be classified as dislocatable and 18% as dislocated. The incidence of instability is quite reasonably wellknown based on large-scale neonatal screening studies. Howorth summarized 11 studies reported from 1950 to 1975 examining 200,000 newborn infants with an instability rate of 0.9% (126). In 5 major studies done by investigators with a committed interest in diagnosing hip dysplasia in the newborn the incidence was 1.4% (of 105,375 infants). It is widely estimated that 1 of each 100 newborns will show some hip instability, with 50-60% of those stabilizing spontaneously without treatment within 1 week of birth (60). The criteria used by differing authors and in particular in relation to large screening programs can make comparative studies difficult, but the trends and tendencies shown by these numbers appear to be accurate. The left hip is affected more commonly than the fight in most series, with bilateral involvement intermediate. Putti (240) showed 39% bilateral, Farrell et al. 27% (73), and Coleman 38% (45).

C. Effects of Intrauterine Environment Several mechanical features of the intrauterine environment have been shown to be related to congenital hip problems. There are many who feel that these mechanical effects are primary causes of DDH. Among the most common of these effects are breech presentation, oligohydramnios, an increased incidence in first-born children due to the relative tightness of the uterine and abdominal muscles, and twin pregnancies. Breech presentations of first-born children are most affected with hip instability (10%).

1. BREECHPOSITION Congenital-developmental dislocation of the hip has been recognized since the late nineteenth century to have an increased incidence in breech presentations (25, 26). Studies of CDH-DDH have documented 15-25% as occurring with breech presentations. The incidence of unstable hips in breech presentations was noted to be 6.35 times that in vertex presentations (8). Their large study noted breech presentations to be present in 4% of all pregnancies (with vertex presentations at 96%), whereas Vartan (327) documented a 2.2% breech incidence, Barlow (12) a 4.4% incidence, and Wilkinson a 2.6% incidence (356). Suzuki and Yamamuro studied the hips in 6559 newborn infants to assess mechanical factors of presentation in rela-

193

tion to congenital hip deformity (297). The incidence of CDH was 1.1%; 0.7% with cephalic presentation, 2% in footling presentation, and 20% in single-breech (knees extended) position. CDH was also found in 6 out of 7 with congenital genu recurvatum (hyperextended knees). These observations are consistent with the fact that a fetal position with the hip flexed and the knee extended predisposes one to the development of CDH-DDH. Suzuki and Yamamuro later illustrated this mechanism by ultrasound, demonstrating femoral head movement in 5 patients with hip dislocation 3-4 months of age. With the hip joint flexed the dislocated head was posterior to the acetabulum and displaced even further posteriorly when the knee was extended (299). The actions of the hamstrings aided in the displacement with this positioning. Wilkinon studied breech malposition and its relation to CDH (34, 254, 355, 357). The term breech position simply refers to the lie of the fetus in utero, whereas the causative association of hip dysplasia was with true breech malposition, which involved hip flexion and knee extension. The most common intrauterine posture at birth is the vertex or cephalic posture, which is established by 30-34 weeks of gestation. In the vertex posture the legs are folded with the hips and knees both flexed. If leg folding never takes place, the primary breech posture is common. The breech posture with knees hyperextended is common during the second trimester, but once the fetal spine, hips, and knees flex spontaneous version usually converts the breech to a cephalic presentation. This conversion is favored by low uterine muscle tone of the multigravida along with increased intrauterine fluid. Vartan, in a classic study of fetal positioning in utero, observed that the breech position was common until the 30to 34-week period of gestation, being present in 25% of pregnancies (327). Spontaneous version occurred in 60% of these in association with knee flexion. When therapeutic version was added to the spontaneous corrections, only 8.5% of 1000 infants who at one time were in breech position were formal breech deliveries. In the entire series of 3875 patients the eventual incidence of breech delivery was 2.2%. Vartan also noted that maintenance of the breech posture was associated with extended knees and diminished amniotic fluid, both of which limited spontaneous and even therapeutic version to the cephalic position. In the primigravida muscle tone is high and fluid relatively less. In frank breech births, Wilkinson noted breech malposition with the knees extended and the legs internally rotated or the knees semiflexed and legs externally rotated. Persistence of these postures beyond the 28th week of gestation constitutes breech malposition, and after the 32nd week they usually remain unchanged until birth. Vartan noted that prior to the 30th week the breech posture was present in 1 out of 4 but that spontaneous version occurred with knee flexion. The majority of cephalic or vertex presentations show the hips and knees to be fully flexed and the thighs adducted and internally rotated, and only 1% have the knees extended. First-born breech deliveries usually have the hips flexed and the knees relatively

194

CHAPTER 3 ~ Developmental Dysplasia o f the Hip

extended with the lower extremities externally rotated. At a later stage of breech malposition there is still locked external rotation of the lower extremities, but some flexion of the knees is present. The high percentage of extended knee postures in breech positions and in those with DDH sustains the theory that failure of the leg folding mechanism is one of the chief mechanical causes of breech birth. Evidence of delayed leg folding of some extent was present in 80% of newborn infants with hip displacement. Wilkinson noted that in 174 patients with CDH in his hospital there was clinical evidence of breech malposition in at least 25% (356). A total of 60% of patients with CDH were first-born. In 20 consecutive cases of neonatal CDH, all of the newborns had displayed the locked-breech birth posture. Breech position was noted by Barlow to have a high correlation with CDH-DDH. In an examination of 8814 normal babies, only 4.4% were breech presentations. Of 139 with hip abnormalities, however, 17.3% had breech presentations (12). Calculations from his data indicate that approximately 6% of babies with breech presentations have hip dysplasia, whereas the data of Artz et al. (8) showed 7% of breech presentations with unstable hips. In summary, the overall incidence of human birth in a breech position is 2-4%, whereas anywhere from 15 to 25% of children with CDH are born from the breech position. The knees are extended in utero in 56 to 75% of children born in a breech position, whereas only 3% of children born in a cephalic-vertex position have their legs extended in utero. In a newborn infant born in a breech position, the frequency of congenital displacement of the hip is in the range of 4-7 %.

2. BIRTHORDER The incidence of hip dysplasia in most series is shown to be higher in first-born infants. In the study by Artz et al. of infants with unstable hips, 63% were first-born, 21% second, 8 % third, 4% fourth, and 4% fifth (8). The birth order affects the intrauterine environment because muscular forces on the fetus are greatest in the first pregnancy due to both tighter uterine and abdominal musculature. Dunn noted 58% of patients with CDH to be first-born (60).

D. Extrauterine Postnatal Environment The positioning of infants shortly after birth can vary from society to society, and in those groups in which infants tend to be positioned primarily with the lower extremities and hips abducted the incidence of subsequent hip problems is low, and in those in which the lower extremities are maintained for relatively long periods of time in the adducted and extended position the incidence of hip problems is increased (263).

E. Genetic Considerations Developmental dysplasia of the hip is not a classic hereditary disorder in the sense that clear recessive or dominant patterns

of inheritance have not been documented. There appears, however, to be a slightly increased likelihood of the disorder occurring in those in which a clear family history of CDH has been documented. The relatively few studies that have been done in this regard and the imprecise use of terminology and diagnostic criteria make it difficult to interpret studies published, but a slight predisposition to familial occurrence appears to be real. Stalder and Jani calculated the genetic risk after a boy with CDH was approximately 10% for the next child, after a girl with CDH the risk was 3%, and when one of the parents had CDH the risk for the first child was 5% (286). The familial incidence is postulated further to be due to joint laxity syndrome, which itself predisposes to the capsular hip joint laxity allowing displacement to occur. Dunn reported a strong family history of 3% (60).

F. Ethnic Considerations Congenital hip problems are extremely common throughout European societies and in groups in North and South America originally of European descent. Once again, one must note the definition used in various studies to allow for the CDH diagnosis and the rigorousness of the examination criteria. There are differences even within the European community. It has long been recognized that there is a particularly high incidence of congenital or developmental hip disorders in central Europe. Tonnis et al. (321) reported that with ultrasound screening of newborns as many as 2.7% were found with pathologic hips. The incidence has been reported as quite high in those from Lapland and northern Scandinavia and in North American Indians, although the extrauterine environment in which the children are placed may play a major role in this regard. Several studies from a few decades ago also indicated a markedly decreased incidence among Chinese and Black children.

G. Spontaneous Stabilization of Hips without Treatment Barlow performed detailed studies documenting instability in newborn clinical examinations of hips followed by a tendency to spontaneous stabilization in normal position in as many as 60% without treatment over a short period of time (13). By using clinical assessment only, he reported on examinations of 19,625 children in which 357 abnormally lax hips were foundm168 frankly dislocated and 189 dislocatable. Considerable spontaneous stabilization occurred, however, as determined by the fact that in 931 babies examined on the day of birth and 1197 babies examined at the age of 1 week there were far more unstable hips in the former group than in the latter. The percentage of abnormal hips was 2.3% in the first group and only 1% in the second group. By extrapolating these figures to larger groups, he predicted that, if one were to find 100 abnormal hips on the day of birth, there would be only 40 abnormal hips 1 week later such that 60 of the original 100 would have undergone spontaneous recov-

SECTION VI ~ Pathoanatomic and Pathophysiologic Findings

ery. Spontaneous recovery could occur up to the age of about 6 weeks, but most of the stabilization occurred quite early. The occurrence of spontaneous stabilization led to some differing practice approaches in different centers. In some centers all patients with unstable hips at birth are reassessed clinically at 1 week. If the patient has undergone stabilization to clinical assessment, then no further treatment is indicated. If the dislocatable state persists, then treatment is instituted. Use of the Pavlik harness has made treatment much easier, and increasing medical-legal concerns also have led many to treat all hips that are dislocatable even at birth. The widespread use of ultrasound has further refined neonatal assessment, and many are guided by the sonographic indices as well. There are definite complications noted with the Pavlik harness and other forms of abduction splinting, such that it still would appear to be medically acceptable to reexamine an unstable hip clinically at 1 week of age, at which time an ultrasound exam can also be performed. Treatment then would be guided by clinical exam and ultrasound at 1 week rather than by the finding of dislocatability on the day of birth. The tendency of many unstable hips to stabilize spontaneously was commented upon in some detail by Le Damany in 1914 (167). What he referred to as the "subluxable hip of the newborn" was frequent in France. He clearly described the dislocation and relocation maneuver in which one could detect the head coming out of the socket and then again a reentrance of the head into the acetabular cavity. He noted that "this difficulty is almost spontaneously corrected in a length of time varying from several days to several months," that it was more frequent in girls, and that it was due to a relative diminution of the acetabulum. If the torsion of the femur was not too great recovery would take place, but if the femur had excess anteversion then the subluxable condition would be the first stage of a congenital dislocation. The clinical maneuver, widely known presently as the Barlow maneuver, was described. Dislocation occurred when the hip was placed in the flexion-adduction position, following which a slight force applied to the knee combined with a force from within outward on the thigh caused the head to half-emerge above the posterior border of the acetabulum. Displacement of the thigh toward a flexion-abduction position allowed the head to reenter the cavity. Lance also commented on the tendency for many unstable newborn hips to heal spontaneously without treatment in the normal position (154).

VI. S U M M A R Y O F P A T H O A N A T O M I C AND P A T H O P H Y S I O L O G I C FINDINGS AND DISCUSSION OF PATHoGENETIC SEQUENCES

A. Overview Structural variations from the normal, actual deformities, and predisposing epidemiologic features leading to femoral

195

head-acetabular malposition and their time of onset each play a crucial role in defining the extent of hip malformation at birth. It appears likely that idiopathic developmental dysplasia of the hip is associated with certain predisposing epidemiologic features plus a variety of subtle structural abnormalities, some of which are tolerated without leading to instability, whereas others combine to lead to an unstable joint. It also is likely that the ultimate severity of the disorder is dependent on the time of displacement, with those occurring in utero predisposing to greater secondary change than those occurring during labor or postnatally with extension of the hip. The earlier in utero the imperfect relationship is established the worse the secondary changes, and the longer that relationship persists postnatally the further those changes worsen. Extrinsic mechanical theories of causation and epidemiologic characteristics appear to play a major, although probably not exclusive, role in predisposing one to CDHDDH. Chief among these are female infant, breech presentation, "dislocating posture" of hyperflexion, lateral extremity rotation and femoral anteversion, genu recurvatum, oligohydramnios, first birth (primigravida), excessive infant weight, twinning, genetic considerations, and (closely allied)ethnic group. The subsequent presentation follows the outlines of Salter's approach to the pathogenesis of deformity in CDH (DDH) (263).

B. Capsular Laxity The primary pathoanatomic defect in developmental dysplasia of the hip is considered widely to be capsular laxity. This diminishes the restraint on the femoral head-acetabular relationship and allows the femoral head to sublux or dislocate laterally, posteriorly, and then superiorly in relation to the acetabulum. The capsular laxity in some descriptions is present throughout the hip capsule, but others have localized it initially to the posterosuperior region. With persistence of the dislocated position the capsule enlarges universally, hypertrophies, and may undergo isthmic narrowing just above the rim of the original acetabulum. This narrowing is partly due to extrinsic capsular pressure by the iliopsoas tendon, which most changes its normal relationship as its insertion is pulled laterally and then superiorly by the displaced proximal femur and its lesser trochanteric insertion. Many dissections of newborn hips with this condition show both the acetabulum and the proximal femur to appear to be structurally normal. Assessments at finer levels of resolution have, however, attributed subtle changes to underlie a tendency to displacement. Among these are the observations of Le Damany (163-167) and others on (1) slight degrees of proximal femoral anteversion combined with (2) increased anterolateral acetabular obliquity, (3) a relatively shallow newborn acetabulum, and Walker's (334) observance of (4) morphological changes in the labrum. In general, however, the acetabulum is shaped and positioned normally, the labrum is shaped and positioned normally, and the proximal femur is normal in size and shape and in particular does not show

196

CHAPTER 3 9

DevelopmentalDysplasia of the Hip

increased anteversion in relation to the contralateral normal side. Most of the changes that develop in the acetabulum and proximal femur postnatally appear to be secondary to their growth in an unstable mechanical environment in which the femoral head is not seated continually within the acetabulum due to capsular laxity.

C. Acetabular Dysplasia As the femoral head develops in an imperfect relationship to the acetabulum, bone and cartilage development in both bones is retarded and tissue deposition is abnormal. Acetabular dysplasia develops in which the upward slope of the lateral acetabulum persists and evens worsens, whereas in the normal hip the acetabular index diminishes and the acetabulum deepens. It is important to think of the developing hip in a three-dimensional sense, especially because the various descriptive terms and radiologic indicators almost exclusively refer to a two-dimensional representation. Acetabular dysplasia, which is measured on plain radiographs as the acetabular index, defines an upward slope of the bony acetabulum, but the primary problem is a shallowness and anterolateral tilting of the cartilaginous acetabular socket. With the altered pressure applied by the proximal femur against the lateral acetabulum, cartilage development is slowed and bone deposition is slowed secondarily as well.

D. Proximal Femoral Dysplasia Abnormal development also is evident in the proximal femur with delayed appearance of the secondary ossification center and failure of postnatal diminution of the proximal femoral anteversion, which normally is most extensive in the later stages of the third trimester just prior to birth. If the capsular abnormality leads to altered hip stabilization in a subluxed position, then the abnormal development is relatively slight and the secondary ossification center on the involved side is present but somewhat smaller than that on the normal side. In completely dislocated hips, the secondary ossification center often can be delayed in appearance for several months compared to the normal side. The second aspect of developmental irregularity in the subluxed or dislocated proximal femur relates to persisting femoral anteversion. Proximal femoral anteversion is as great as 30-35 ~ in the newborn and decreases at a fairly constant rate to 10-15 ~ by early adolescence. If the hip is allowed to develop in the dislocated position, femoral anteversion persists and may even increase. It is not unusual, therefore, to find 30 ~ or 40 ~ of anteversion in an 18-month-old child with a unilateral dislocated hip. The clinical correlate of this pathoanatomic development is increased internal rotation needed to seat the femoral head into the acetabulum with reduction maneuvers.

E. Adductor Muscle Tightness As hip dislocation persists, the head remains lateral, posterior, and somewhat proximal to the acetabulum. The soft tissues react in a characteristic way. The adductor muscles tighten as early as 2 months of age, an occurrence that tends to hold the hip in the subluxed or dislocated position. With greater displacement and at later time periods, the iliopsoas and gluteus muscles are relatively shortened as well.

F. Soft Tissue Deformation The medial capsule stretches and elongates to accommodate the displaced femoral head. As the head migrates laterally and superiorly the capsule is pulled across the inferior segment of the acetabulum. The ligamentum teres remains attached to the femoral head but hypertrophies in both length and width. In long-standing complete dislocations it is absent, having either ruptured or, in the opinion of some, been congenitally absent. This structure, normally present in the fovea, must now traverse a greater distance and comes to fill much of the dysplastic acetabulum. Significant changes occur in what is referred to as the radiolucent roof of the hip. This term refers to the lateral regions of the acetabular cartilage, the labrum, and the capsule. The position of the labrum is crucial to an understanding of the pathoanatomy of developmental dysplasia of the hip, and the assessment of labral position is important to determine in considering the advisability of nonoperative versus operative treatment. As the femoral head moves progressively farther away from its normal position in the acetabulum, it pushes the labrum outward and upward. In the subluxation phase, the head pushes up against the labrum but the labrum is still on top of and thus supporting the head. With even further displacement the labrum is pushed against the capsule by the femoral head but still maintains its relation to the head. As the head completely displaces from the acetabulum, the labrum no longer lies on top of the head but rather slips between the cartilage head and the acetabular cartilage, carrying with it part of the capsule that is adherent to its outer surface. The combination of the inverted labrum plus the associated flattened capsule is referred to as the limbus. The limbus therefore is not a normal anatomical structure but rather a pathological structure composed of two normal anatomical structures in an abnormal position. At this stage the femoral head is supported only by the capsule, which reacts by increasing its thickness becoming markedly hypertrophic. As the femoral head tides up against the capsule outside the acetabulum, some stability is provided eventually as the head rests against the outer wing of the ilium and forms what is referred to as a pseudo-acetabulum. By this stage of the developing pathoanatomic picture, it is not possible to relocate the femoral head into the acetabulum by closed reduction maneuvers. Adductor tightness limits the range of abduction, and the acetabulum has become filled

SECTION VI ~ Pathoanatomic and Pathophysiologic Findings

with soft tissue structures that can no longer be pushed away to accommodate the femoral head. These include the inferomedial capsule, which has been stretched across the inferior margin of the acetabulum as it is attached to the lesser trochanter, the ligamentum teres, which has lengthened and hypertrophied as it remains attached to the displaced femoral head, fibrous fatty tissue in the depths of the acetabulum occupying the space where the femoral head should have been, and the inverted labrum-capsule or limbus, which further blocks normal positioning of the femoral head. False reductions can be achieved by closed means, and the adequacy of any reduction after the first few months of life should be checked by imaging other than plain radiographs. With mild to moderate complete dislocations undergoing closed reduction the head usually sweeps the inferior capsule away from the joint surfaces, although the inverted limbus prevents anatomic reduction superiorly. In severe dislocations treated after walking has begun, as was generally the case several decades ago, the inferior capsule also can be inverted with closed reduction maneuvers.

G. Idiopathic and Teratologic Dysplasia The terms idiopathic and teratologic in developmental dysplasia of the hip imply a clear-cut differentiation between two types of hip dysplasia, but in reality although the terms have some validity they are far from being definitive. In clear-cut examples of either type, we can be comfortable that an accurate impression is left. In the straightforward idiopathic hip dysplasia, there appears to be sufficient evidence to conclude that capsular laxity is the offending occurrence (setting aside what causes the laxity), that the proximal femur and acetabulum are structurally normal initially, and that, if diagnosis is made shortly after birth and the child is placed in a Pavlik harness or similar apparatus in which the hips remain flexed and slightly abducted, normal hip development will follow once the capsular tightening has occurred. Indeed, 50% of dislocatable hips in the newborn will stabilize spontaneously in normal position without treatment within a week. At the other end of the spectrum are infants that are either stillborn or born with extensive connective tissue abnormalities involving the skeletal system, who have displaced hips in which stable reduction cannot be performed by closed means even at birth and in which either autopsy or direct examination at surgery shows clear structural developmental abnormalities of the proximal femur and acetabulum. There are many patients, however, who fall between these two extremes such that the pathogenesis of the hip dysplasia is unclear. For example, there are normal appearing patients except for the dislocatable hip who undergo a seemingly uneventful closed reduction, but who are found after several weeks to months to still have either a dislocatable or a dislocated hip, which subsequently is amenable to repair only by surgery. On the other hand, there are patients with clear connective tissue or neuromuscular

197

abnormalities, such as Ehlers-Danlos disease, skeletal dysplasia, myopathies, or even arthrogryposis, who undergo uneventful closed reduction of the hip and Pavlik harness or hip spica treatment and soon develop a stable hip that is radiographically normal, such that the teratologic etiology does not preclude a good result obtained by straightforward means. Therefore, there are two general possibilities in relation to the underlying pathogenesis of developmental hip dysplasia. The simplest concept indicates that the differences seen in several patients, and in particular those who appear otherwise normal, are related to the time at which displacement occurs. If there is malposition due to capsular laxity during, for example, the final 2 months of intrauterine life, then the secondary changes would be more extensive leading to the failure of treatment with simple closed reduction and flexion-abduction splinting. One thus is not dealing with a teratologic hip but rather with a relatively straightforward DDH that occurred several weeks prior to birth and was not amenable to diagnosis or treatment. The second occurrence is that simple capsular laxity alone cannot be implicated necessarily as the sole etiologic agent in all cases, such that other developmental abnormalities affecting the acetabulum, proximal femur, and associated soft tissue structures can be implicated and thus lead to the need for more complicated treatment protocols. In an effort to assess those individuals who fall between the two major extremes of idiopathic and teratologic, the use of imaging techniques plays a major role. These are done increasingly frequently postnatally in relation to ultrasound primarily but also to MRI and CT scanning.

H. Worsening of Secondary Changes with Time Secondary changes develop in the postnatal period and are increasingly more marked the longer the hip remains subluxed or dislocated. A dislocated hip shows diminished abduction with a prominent adductor muscle-tendon band felt on abduction. This adaptive change can be seen as early as 6-8 weeks of age, becomes progressively tighter, and ultimately may prevent the performance of relocation by simple clinical manipulations. Radiologic correlates increasingly demonstrate acetabular dysplasia with an increased acetabular index, which is an indication of diminished development of the bone part of the acetabulum, delayed appearance and a smaller sized secondary ossification center of the femoral head, and one that is placed more laterally than that on the normal side. The position of the secondary ossification center is either close to or on the other side of Perkin's line, which is a vertical line dropped from the outer bone margin of the acetabulum. There is a break in Shenton's line which is a continuous curvilinear relationship involving the inferior neck of the femur and the inferior surface of the superior pubic ramus at the obturator foramen. The longer the femoral head remains outside the acetabulum, the more

198

CHAPTER 3 ~

Developmental Dysplasia o1r the Hip

misshapen it becomes and the less uniform the fit between the two when reduction finally is achieved. Table II summarizes the pathogenesis of developmental dysplasia of the hip.

VII. N A T U R A L H I S T O R Y O F H I P DISLOCATIONS, SUBLUXATIONS, AND DYSPLASIA

A. Natural History of Complete Dislocations There have been a limited number of studies over the past few decades following patients with untreated unilateral or bilateral dislocated hips well into their adult years. In a complete dislocation, the opinion appears well-grounded that pain is a relatively infrequent occurrence, in particular in the first three decades of life. Patients are sometimes described as doing well with an untreated complete dislocation, although these references appear to refer to the minimal to absent discomfort and the fact that the patients can walk independently and function in society. The instability and waddling gait would appear, however, to preclude running and athletic pursuits. These long-term studies, however, defined certain occurrences clearly. Discomfort can be present in the lower back with the associated lumbar lordosis and in the hip on the dislocated side. Generally, however, the onset of discomfort is in the midfourth to midfifth decades. An excellent study by Wedge and Wasylenko reported on 32 patients with 42 complete dislocations between 16 and 86 years of age (343). When a modified Harris hip score was applied to their assessment, 41% of the patients had a good rating, 14% were fair, and fully 45% had poor clinical ratings. The presence or absence of a false acetabulum played a major role in influencing the clinical fate of patients. Wedge and Wasylenko demonstrated only a 24% chance of a good clinical result with a well-developed false acetabulum, whereas in those with a moderately developed or absent false acetabulum the clinical rating improved to 52%. The presence of a false acetabulum, which by definition would also imply imperfect positioning of the head in relation to that acetabulum, would lead to wear and tear changes on the cartilage and subsequent discomfort. It appears better from a clinical sense to have no false acetabulum because the cartilage degeneration and bone eburnation phenomena would not occur, and it is these latter two occurrences that lead to discomfort. Some have felt that, in patients with bilateral dislocations, the lower back pain is greater due to the more marked compensatory hyperlordosis, whereas in patients with complete unilateral dislocations discomfort relates to limb length inequality, with ipsilateral knee deformity and pain, scoliosis, and more marked gait disturbances. The flexion-adduction hip deformity on the involved side is compensated by a valgus knee, which in itself leads to increased knee stresses and discomfort.

F I G U R E 5 An adult with osteoarthritis of the hip following imperfectly treated developmental dysplasia of the hip in childhood. Note the persisting acetabular dysplasia, lateral acetabular subchondral bone sclerosis, the relatively spacious acetabulum, and the femoral head riding laterally and upward. There is joint space narrowing and early osteophyte formation.

B. Natural History of Dysplasia and Subluxation The worst position in terms of development of symptomatic discomfort is the subluxed position because the patients remain ambulatory from the early years of life but the wear and tear on the femoral head cartilage and outer acetabulum is great. In those with dysplastic hips with the relatively mildest degree of acetabular malformation, symptoms of degenerative joint disease can occur but generally in mid to late adult life (Fig. 5). The fact that subluxation of the hip is the position that primarily leads to the development of degenerative joint disease and clinical symptomatology also was stressed by Wedge and Wasylenko, who noted that only 42% of the 38 subluxed or dislocated hips had a good clinical rating (343). The degree of subluxation also appeared to correlate roughly with the symptoms. The most severe subluxations led to onset of symptoms during the second decade of life; those with moderate subluxation presented during the third and fourth decades, and those with minimal subluxation had symptoms even later.

C. Osteoarthritis in Adult Life Following Childhood CDH It is now widely recognized that any result less than anatomic restoration of the hip predisposes one to osteoarthritis in mid to late adult life. This observation took some time to evolve into clinical certainty, however. Over the past few decades the diagnosis of primary osteoarthritis has been

SECTION VIIi ~ Treatment Approaches in Developmental Dysplasia of the Hip made less frequently based on a more careful assessment of the adult radiograph, which increasingly is interpreted to indicate the sequelae of previously unrecognized childhood hip disease. Morville (1936) showed excellent awareness of the fact that osteoarthritis of the adult hip frequently and indeed almost always was preceded by structural hip disorders of childhood (209). He analyzed 100 cases of osteoarthritis of the adult hip (referred to in those days as arthritis deformans). In only 16 of the 100 had the arthritis developed in an anatomically normal hip joint, and even in 12 of these an associated cause was noted to be either arthritis of childhood, trauma, or infection. The remaining 84 cases showed anatomically abnormal conditions. He divided these into two types with 38% representing congenital hip subluxation and 46% representing subluxations due to acquired hip disorders, the two most common of which were Legg-Perthes disease and slipped capital femoral epiphysis. In those arthritis cases secondary to congenital hip dysplasia, the acetabulum was steep and the femoral head was subluxated upward and outward, which was noted radiographically by interruption of Shenton's line. The deformity undoubtedly had developed on the basis of the congenital fiat acetabulum, although it was not until the age of about 30-50 years that the arthritic changes began. In the other group of predisposing hip disorders, the acetabulum was broader than normal but not steep leading to primarily a lateral subluxation of the head. He indicated that "it is undoubtedly the deformity--the subluxation-that is primary but for years giving no symptoms until the arthritic changes occur in a secondary fashion." Because so many of the arthritic conditions in the adult hip developed on the basis of childhood hip disorders, there would be great value in improving treatment of the latter. In relation to congenital hip dysplasia, he felt that the indications for treatment toward an anatomically normal joint should be enhanced. Wiberg also was one of the early investigators to document definitively that a subluxation persisting from childhood would, in mid adult life, frequently lead to osteoarthritis of the hip (350). He presented 18 cases with X rays from 15 to 30 years following treatment of a childhood congenital subluxation of the hip who evolved into osteoarthritic deformities with typical changes in the femoral head and acetabulum. Murray assessed 200 radiographs of osteoarthritis several years ago and determined that only 35% were truly idiopathic with 65% of cases being due to underlying childhood hip disorders, the main ones being a persisting acetabular dysplasia in 25% of the cases and 40% showing a tilt deformity felt to be consistent with minimal slipped capital femoral epiphysis (211). A study by Gade in Norway in 123 operated cases of hip osteoarthritis had shown an almost 50% incidence of acetabular dysplasia in adult osteoarthritis, with only 24% being primary (85). These observations strongly support the need to diagnose a hip dysplasia disorder as early as possible. The most desir-

199

able result is an anatomic restoration of femoral-acetabular position to preclude subsequent degenerative changes.

VIII. BRIEF HISTORY OF TREATMENT APPROACHES IN DEVELOPMENTAL DYSPLASIA OF THE HIP The four principal aims of treatment for a developmentalcongenital dysplastic hip are straightforward. They include the following: (1) reducing the hip anatomically so that the femoral head is seated normally in the acetabulum without the interposition of soft tissues such as the inverted labrum or capsule. This can be achieved by either closed or open methods. (2) Maintenance of the reduction for a sufficient period of time to allow tissue reconstitution to stabilize the hip such that loss of position does not occur when the hip is no longer immobilized. (3) Development of a structurally normal proximal femur and acetabulum and a snug capsule. If these do not develop spontaneously with growth following relocation, surgical methods involving proximal femoral derotation and varus osteotomy, acetabular augmentation, or redirection procedures or capsulorrhaphy may be needed. (4) Avoidance of avascular necrosis of the femoral head.

A. Gradual Development of Reasonably Effective Closed and Open Treatments 1. EVOLUTION OF TREATMENTAPPROACHESIN 1800S

AND EARLY 1900S Congenital dislocation of the hip was placed on a firm scientific grounding by the works of Palletta (221) in 1820 and Dupuytren (62) in 1826. For the next 50 years, little headway was made in successful relocation and maintenance of relocation for the disorder. Good results were difficult to obtain due to the relatively late ages at diagnosis, usually after walking had begun, and the imperfect means of immobilization. Pravaz is widely credited with successfully reducing congenital dislocation of the hip by closed means in 1838 using strong recumbent traction in extension (239). The nature and length of the treatment, however, were scarcely acceptable, because it involved the hip being kept in extension for as long as 8-10 months and, more importantly, never was shown convincingly to lead to retained position after being discontinued. Paci (220) in Italy (1888) and Lorenz (180) in Austria (1895) are credited with developing closed reduction into an effective treatment. Lorenz attempted to reduce the hip fully, whereas Paci appeared to improve its position without succeeding in bringing about a true reduction and stabilization using gentle traction. Lorenz stabilized the closed reduction position with hip spica immobilization followed by abduction splinting, which at the time revolutionized the therapy for CDH. One-stage reductions under anesthesia, using considerable force, were done in the Lorenz method. The patients generally were 3-10 years of age;

200

CHAPTER 3 ~ Developmental Dysplasia of the Hip

each of the flexor, adductor, abductor, and extensor muscle groups were removed forcibly from their origins or stretched so that the limb was flaillike before the actual reduction was able to occur. Although spica casting in the Lorenz position involving flexion of 90 ~ and full abduction of 90 ~ of the hips subsequently has been shown to lead to a high incidence of avascular necrosis, his approach is considered of great importance in the treatment profile for CDH because he was able to reduce the head into the socket and keep it there sufficiently long that ultimate stability ensued. Lorenz published his major work in 1895 and continued to practice, write, and speak on it for the next quarter century, publishing a lengthy treatise on CDH in 1920 (181). He stressed not only the reduction of the hip but the maintenance of reduction until stability followed. He actually developed his approaches to closed reduction based on an extensive experience with open reduction, switching to the closed means as a primary approach after several postoperative deaths and the occurrence of considerable rigidity and contractures. Bennett, in 1908, wrote that Lorenz was reporting approximately 50% "good anatomic results" in follow-up of 680 hips. Ironically, for a disorder that can now be treated by closed means in the large majority of cases, it was the operative approaches that first gained prominence. Around 1870 operative intervention was performed more frequently, and in 1888 Poggi succeeded in replacing the congenitally dislocated femoral head in a 12-year-old female by open reduction followed by 50 days of traction in extension and gradual reambulation with crutches (231). The pathoanatomy was well-described, including the enlarged thickened capsule, the absent ligamentum teres, the narrow capsular isthmus overlying the small acetabulum, and a small, deformed nonspherical head. Following enlargement of the capsular foramen with longitudinal incisions, the true acetabulum was uncovered and found to be filled with fibrous tissue and the remnant of the ligamentum teres. "I deepened the acetabulum, reshaped the deformed head giving it its proper contour, and by means of well directed traction and incisions in the capsular ligaments, I was able with only little difficulty to replace the head into the newly deepened acetabulum." Closure was associated with removal of excess capsule and its repair. Hoffa (121, 122) in Germany and Lorenz (181) in Austria both played major roles in developing techniques for operative open reduction for CDH. The method of Hoffa also involved curettage of the acetabulum to deepen it and extensive release of the surrounding muscles. The Lorenz procedure was less involved because he left the muscles intact and concentrated on the capsule and joint reduction. Many good results were reported. Burghard was one of the proponents of open reduction in England, describing approaches similar to those used today (32). Ludloff used the inferomedial hip approach for open reduction (183). The complications of surgery of that era, including sepsis, were steep and closed reduction was used increasingly.

2. EARLY REVIEWS OF TREATMENT RESULTS Results improved gradually in a series of reports collected over 20- to 25-year intervals. Around the first decade of this century detailed reports began to be published, although the criteria for assessing results of course were variable. Kirmisson gave the number of "cured cases" as 11.3%, Hoffa 19.8%, Lange 45.5%, Lorenz 63.3%, and Waldenstrom 64%. It was after these reports that more careful assessment began to be performed because there was widespread recognition that anatomic healing to complete normalcy was rare, even though functional exams in childhood were reasonably good. A report by Bradford et al. on 144 cases treated at the Children's Hospital, Boston, showed that the two primary methods of treatment involved open reduction and closed manipulative reduction (24). The earliest open procedures, which utilized the Hoffa technique in which the acetabulum was deepened with a curette, were ineffective. Subsequent utilization of the Lorenz open approach led to much improved results. The muscles of the hip region were spared, the capsule was freed and any constrictions were divided, and the capsule then was repaired following reduction of the femoral head. As the techniques evolved reduction was improved by dividing specific tightened structures around the hip such as the adductor muscles, repairing the enlarged capsule to provide support for the reduced head, and in some instances performing a proximal femoral derotation osteotomy to correct femoral anteversion, particularly when it was 60 ~ or greater. Many cases were treated with manipulative reduction using the Paci-Lorenz method. It was recognized that inversion of the distorted capsule into the acetabulum in front of the reduced head could take place during closed reduction, a finding that led many to perform the open procedure. Unsatisfactory cases were considered to be due to imperfect reduction with folding of the capsule in front of the head of the femur, uncorrected persisting femoral anteversion, and defects of the acetabulum. They recognized that "for successful treatment of cases of congenital dislocation of the hip after reduction it is necessary that the capsule should fit closely and not loosely around the head." Lorenz subsequently indicated that the best results were obtained when reduction was done between the ages of 2 and 3 years. He reported on a large number of hips, 1057, in which good to excellent results were obtained in 57% of unilateral cases and 53% of bilateral cases (278). Criteria for good results involved the prevention of upward slipping of the femoral head by an appropriately shaped acetabular roof, free mobility of the joint, and excellent muscle function. Lange reported a massive series involving 2200 reduced hips over a period from 1904 to 1925 (156). Results improved in the later time periods with an indication of 22% good results from 1904 to 1914, which increased to 63.7% between 1915 and 1925. The definition of anatomic healing was somewhat vague by today's standards but involved good function, good mobility of the joint, and X rays showing no dislocation,

SECTION VIII ~ Treatment Approaches in Developmental Dysplasia of the Hip

subluxation, or serious deformities of the femoral head and neck region. Putti published a report on 523 cases involving closed treatment from 1899 to 1927 (243). His detailed categorization graded anatomic and functional results from 0 to 10. He showed good anatomic results (grades 9 or 10) in 34% and good functional results in 40%. There also were improvements during the later time periods such that in the latter group, from 1921 to 1927, good anatomic results (marks of 9 and 10) were seen in 37% of bilateral cases and 59% of unilateral cases, whereas the highest grade functional results were seen in 40% of bilateral cases and 68% of unilateral cases. Indications for open reduction became clearer with Deutschlander, who wrote extensively beginning in 1908, indicating that surgery remained warranted when closed reduction was unsuccessful and in particular when there were hindrances to reduction such as a long ligamentum teres or capsular interposition. Many surgeons continued to resort to open reduction when closed reduction failed. Galloway (88) was one of the earliest in North America to favor the method, whereas Farrell and Howorth (72) described 122 procedures from their unit in 1935. The forceful measures of reduction plus the extreme positions of hip immobilization in plaster spicas led to severe sequelae attributable to avascular necrosis. At fault in the latter regard were the Lorenz position of maximal 90 ~ abduction and flexion of 90 ~ and the Lange position characterized by full abduction in extension and marked internal rotation. Once the closed and open methods had at least enabled the head to be repositioned in the socket, attention then was directed (1) to achieving a structurally and functionally normal hip, which involved anatomic reduction of the femoral head into the acetabulum without capsular or soft tissue interposition and normal structural shapes of the two bones, and (2) to preventing avascular necrosis as a complication of therapy. 3. MID-TWENTIETH CENTURY OVERVIEW OF RESULTS a. L e v e u f Leveuf summarized the long-term results that were being achieved with closed and open reductions of congenital hip dislocation (172, 173). His interpretation constantly stressed the importance of understanding and demonstrating the underlying pathoanatomic features. b. Closed Reductions Leveuf performed an extensive review of the results following closed reduction by analyzing 602 cases from major clinics throughout France with a 10to 40-year follow-up (172). He also extensively reviewed the literature on closed reductions from other countries. He stressed the importance of assessing not only the functional result but also the anatomic result and clearly documented that the long-term result was in fact dependent on the anatomic reconstitution of the hip because the functional results during childhood and adolescence almost invariably were good. He clearly demonstrated that with the increasing pas-

201

sage of time many results originally graded as good and excellent became fair to poor, with progressive development of arthritis in the affected hips. His analysis did not attribute the better results in the more recent cases to improvements of technique but indicated that less than perfect anatomic reconstitution early on invariably led to worsening over the succeeding decades. This finding was strengthened by results from several clinics in which the same technique by the same individuals had been used over several decades. In one clinic Leveuf noted that good results from 10 to 14 years were seen in 58% of the patients, from 15 to 20 years this had diminished to 30% of the patients, and from 21 to 32 years this had further diminished to 22% of the patients. Other groups with comparable statistics showed a diminution of good results from 48% at 10-27 years, to 31% at 1530 years, to 25% at 15-40 years. The long-term prognosis was dependent not on the short-term clinical profile but rather on the anatomic profile as revealed radiographically. It was only the good and excellent results anatomically that led to good and excellent long-term results. It was not the technique of reduction that played the major role in determining the quality of the result but rather the anatomic reorientation that was achieved. The portion of good long-term results was highest in the groups that were treated at the youngest ages, and unilateral dislocations had better results than bilateral. In assessing the fair and poor results Leveuf noted two categories of findings. One involved hips that had remained subluxed in which femoral-acetabular alignment was less than perfect, and in the other group there were major modifications of the shape and structure of the femoral head and acetabulum as a result of the treatment, findings that are interpreted today as secondary to avascular necrosis of the femoral head. Subluxation was noted radiographically by interruptions of the femoral neck-obturator line (Shenton's line in the English literature), valgus positioning of the head and neck in relation to the shaft, persisting femoral anteversion, and obliquity of the lateral aspect of the acetabulum. c. Open Reductions The results of open reduction for treatment of congenital hip dislocation were improved slightly from the 25% rate with closed reduction, but analysis indicated only a 40% range with good long-term results (173). Poor long-term results were due to failure to correct the valgus deformity and anteversion of the head and neck region. The slightly improved results over those from closed reduction were due to the removal of interposed limbus and capsular tissue. In some instances imperfect handling of these interposed tissues still constituted a problem. Leveuf pointed out that in a theoretical sense the approach taken by Zahradnicek of Czechoslovakia was required for marked improvement of results, which involved not only open reduction with the removal of interposed soft tissues (limbus and capsule) but also simultaneous correction by osteotomy of the valgus and anteversion deformities of the proximal femur by resecting a trapezoidal segment at the base of the neck with a

202

CHAPTER 3 ~ Developmental Dysplasia of the Hip

wedge based posteriorly to correct the anteversion and medially to correct the valgus (173,362). These procedures also led to a certain degree of shortening, which eased the reduction. Although the principles of this approach were excellent, in reality there were many problems because with the techniques of that era there was some mortality. d. Summary of Treatment Approaches Leveuf and Bertrand summarized approaches to treatment, outlining many principles that are still valid (174). Treatment should not be defined rigidly as either closed or open, as was the case in many clinics; the approach that best suited an individual patient was the desired one. It was essential to make a precise anatomic diagnosis of the status of the dysplastic hip prior to deciding whether closed reduction or operative intervention was the desired treatment. Consideration of the age of the patient played a major role because the deformity worsened with time particularly with the beginning of weight bearing because many of the changes were secondary and not primary abnormalities. Because many patients were seen after walking had begun, between 18 months and 2 years, it was important to determine whether the dislocation was reducible, whether some deformations had been induced secondarily following a trauma of reduction, and whether after treatment certain malformations compromised the eventual stability persisted, these being a dysplastic acetabular roof and valgus and anteversion of the proximal femur. In a young infant difficulty in obtaining closed reduction was unusual. False reduction could be present, however, and increased with the age of the patient due to the interposition of the limbus or the capsule or both between the femoral head and the acetabular cartilage. Due to the technique of arthrography, however, they felt that false reduction should be recognized and not accepted. The second group of concerns related to circulatory problems following reduction, which led to AVN and subsequent growth abnormalities even if the femoral head was reduced successfully. They recognized that these growth problems were secondary to the various reduction maneuvers because these changes did not exist in dislocations that had never undergone any treatment. They did not accept, however, that these were exclusively cases of AVN because they reported that such findings were seen infrequently to the point of not existing with open reductions. The third group of problems occurred with imperfect long-term development in those cases in which acetabular dysplasia or proximal femoral rotational changes had not been corrected. The value of early treatment was recognized. Early diagnosis by clinical assessment and radiography was possible and the acetabular index of Hilgenreiner was a particularly sensitive, early indicator of acetabular dysplasia (116, 149). The value of arthrography was stressed in efforts to note interposed tissues that compromised good end results. They noted that Putti had treated 777 cases of early dysplasia by splinting with the hips in abduction and slight flexion, and a review of 478 of these cases indicated 92% excellent results.

Even at that time (1941), however, they recognized that many of these hips would have been corrected to a normal range spontaneously. Only those hips that were shown to be truly dislocated warranted treatment. e. Gill; Bost et al. Fan"ell and Howorth Gill published a detailed review in 1948 of 105 congenitally dislocated hips, which required either closed or open reduction for treatment (92). Of the 105 hips treated by closed reduction, initially 52 subsequently required open reduction for appropriate positioning. Results were divided into perfect, excellent, satisfactory, and failure. Only 25% of dislocated hips could be expected to become perfect or excellent after closed reduction, with an increase to 35% when the dislocation was reduced in the first 3 years of life. Another 15% became satisfactory for a varying number of years, with an increase to 20% when treatment occurred before the end of the third year. Failure, however, was noted in 60% of all primarily reduced dislocations and 45% of those reduced in the first 3 years of life. Subsequent open reduction and femoral and acetabular procedures were of value but clearly represented failures of closed treatment. Gill also noted that the earlier the treatment began the better the results. In those treated under 2 years of age there was a 36% rate of perfect and excellent results, which diminished to 34% in those treated in the third year, 21% in those treated in the fourth year, and 15% in those treated beginning over 4 years of age. Bost et al. also reported a detailed assessment of 112 hip dislocations treated by variable methods in 1948 (23). Excellent or nearly excellent anatomical results were obtained in 55.4%, with 60.7% of the functional results excellent. They noted a direct relation between early treatment and improved results. Farrell and Howorth, reviewing over 600 cases from the first third of the 1900s, reported 42% "successful" closed reductions and 77% "successful" open reductions (72). Howorth described the open reduction technique (124). Steindler reported only 11% "anatomically perfect hips" from his institution in 1950 (289). He also documented the progressive deterioration of results with time due to failure to achieve anatomic restoration of joint structure with initial treatment. In 114 hips 1-5 years postreduction, good results were 70.6%, fair 14.3%, and poor 15%; at 10-20 years the corresponding percentages were 52.6%, 17.4%, and 30%.

f Treatments for Hip Dislocations after Walking Had Begun Most patients by the middle of the twentieth century were still presenting for definitive treatment with diagnosis made between 18 months and 2 years of age or later. Closed reduction was attempted first, and it was becoming possible to assess the quality of the reduction by arthrography. Reduction was felt to be unsatisfactory if there was either interposition of the limbus or interposition of the capsule between the femoral head and acetabular articular surfaces. Diagnosis of these occurrences by arthrography then mandated open reduction. If bony malformation occurred it then would be corrected at a second stage. Although some performed acetabular shelf procedures at that date, they did not feel that

SECTION VIII ~ Treatment Approaches in Developmental Dysplasia of the Hip

was necessary if the femoral head could be deeply seated in the acetabulum by osteotomy. The final group of patients constituted those who had been walking for a few years but still had a dislocated hip. Radiographic and arthrographic assessment was essential. Preoperative traction was necessary to help reduce the marked displacement of the head. Because of difficulties with morbidity and mortality in one-stage procedures they recommended two procedures several months apart: open reduction was performed in the first and osteotomy was used to correct bony malformation in the second.

B. Progressively Earlier Diagnosis and Treatment of Congenital Dislocation The advisability of early diagnosis and early treatment also was advocated by Roser (257) in 1879 and Hilgenreiner (116) in 1925 along with many others, although little response to this occurred among the orthopedic and surgical professions for several decades. 1. HILGENREINER Hilgenreiner in 1925 reviewed the approaches to congenital hip dislocation, concentrating on the timing of treatment because at that time most patients were still a few years old when treatment was performed (116). Because the condition often could be diagnosed in the newborn he felt that theoretically treatment in a very young infant, though difficult, was desirable. He concluded that "with the development of methods for early diagnosis and consequently the possibility of very early recognition of hip disease on one hand, and with the development of the retention splint on the other, there is nothing to prevent the universal early treatment of congenital dislocation of the hip, even in infants." He concluded that every congenital hip dislocation can be treated as soon as it is recognized and that the earlier the reduction is done the simpler the method of reduction, such that in the infant it can be accomplished even without anesthesia. Good results depended mainly on improved development of the acetabulum and the femoral head, and with early treatment delayed complications could be avoided. The abduction splint was best for retaining the reduced femoral head in the acetabulum in the infant, and even early operation, if needed, was desirable compared with the long-term complications when the hip remained dislocated. Results that seemingly were good at the termination of treatment often degenerated during the later years of growth. Hilgenreiner commented on the increasing recognition of avascular necrosis (AVN) and also on the recognition that the harshness of the reduction, particularly in patients older than 2-3 years of age, was the cause of the AVN. He indicated that "today it is far more important that the age limit be moved downward." Many continued to wait once diagnosis was made until even the third or fourth year of life, but Hilgenreiner noted that "this remarkable attitude is hard to understand because from the beginning, it was quite evident

"'~

'

203

/t z

FIGURE 6 An illustration from the classic work of Hilgenreiner (116) delineating radiographic changes in dysplastichips. The acetabularindex (et) is increased on the dysplastic left side and the H (h) distance is decreased on the dysplastic side. The transverse line describedby Hilgenreinerlinking the two triradiate cartilages now carries his name.

that this lesion is similar to a traumatic dislocation. The lesion does not improve by waiting and delay is not harmless. The loss of contact between the acetabulum and head causes secondary effects on the capsule, the ligamentum teres, and the hip musculature changing the shape of the bones and making it more difficult to reduce the hip and to maintain the reduction therefore unfavorably affecting the end result." One of the problems of early treatment in the first year of life was difficulty with the hip spica casts. Hilgenreiner reviewed the works of several authors who had attempted to promote the earliest possible treatment time. Bade was quoted by Hilgenreiner as stating in 1908 that "if every doctor would carefully examine the hip joints of the newborn, and if he had the slightest doubt consult an orthopaedic surgeon or get x-rays in order to clarify the issue, then it is probable that many cases of congenital limping would not develop and could be avoided even before the infant takes its first step." Walther, Joachinsthal, and Loeffler also concluded that those treated early, especially before walking began, had the best results (116). Vulpius and Engelmann also recommended that treatment should begin at the time of diagnosis and was markedly preferable before walking began (116). Hilgenreiner noted the relative ease of reduction in the early months of life, even without anesthesia. He commented on the typical "noise" that could be appreciated with reduction. Asymmetry of the inguinal, adductor, and gluteal thigh folds was noted with dislocation, but he also stressed that the formation of infant thigh folds could vary greatly in relation to both number and symmetry "even when the hips are normal." He also described the subtle changes in the hip radiograph in the early months of life, with hip dislocation even prior to formation of the femoral head secondary ossification center. Particularly important were variations in the slope of the acetabulum, findings described by the angle of the acetabulum (Fig. 6). He defined the acetabular angle and felt that in the normal infant the angle usually was less than 20 ~. He described the " Y " line, subsequently referred to as the

204

CHAPTER 3 ~ Developmental Dysplasia of the Hip

line of Hilgenreiner, in the following way: draw a horizontal line connecting the two Y-shaped cartilages. Determine the acetabular angle, which is the angle formed by the bony acetabulum with the previously mentioned horizontal line. After the first year of life, the secondary ossification center of the femoral head was below the horizontal line and medial to a vertical line drawn from the outer edge of the bony acetabulum perpendicular to the vertical line, as described by Perkins and Ombredanne. He developed a splint that held the hips in abduction but was much less cumbersome than the plaster spica. 2. PUTTI Treatment was begun as early as 4 months of age (242). Although surgeons since the late 1800s had occasionally recommended early treatment for congenital dislocation of the hip, by which was meant during the first year of life, it really was the experience and influence of the Italian surgeon Putti (as well as Hilgenreiner) that placed the concept on a firm footing. Closed reduction in the older patient had improved results somewhat but there were still many imperfect results. The approach to be followed, he suggested, was one of "lowering of that age limit." At that time, the earliest age of treatment was considered to be 2 years, although Putti contended there was no theoretical or practical reason to forbid commencing treatment before that age. When treatment was begun in the first few months of life, he was able to abduct the hips gently without anesthesia so that the head repositioned itself into the acetabulum. Therapy then required maintaining this position for a few months in order to secure permanent reduction. He developed an abduction cushion for that purpose. The average duration of treatment was from 8 to 12 months. The treatment was felt to be ideal, however, in the sense that there was no anesthesia and no manipulation, which even then was recognized to lead to osteochondritis (avascular necrosis). Because rigid immobilization in plaster hip spica casts was avoided, there was also a marked diminution of atrophy of muscles and joint rigidity. He concluded that the best way to improve the results of treatment was to lower the age limit at which treatment began. Long-term studies were demonstrating that, if a hip was reduced incompletely, the results were always less than ideal such that "no complete or permanent restoration of function occurs apart from perfect anatomical reduction." The younger the patient, the easier the reduction and the less the likelihood of soft tissue interposition. He clearly articulated the principle that one must "abolish" the age limit and begin treatment at the very moment the deformity is observed, even if that be on the day of birth. He even went so far as to make the radical statement concerning the value of submitting every newborn child to a routine X-ray examination of the hips. He was able to treat large numbers of patients effectively during the first year of life; in one series the average age was 4 months. In 119 cases treated by the abduction method from 34 days to 16 months of age (average age = 4 months), complete cure was achieved in 113.

3. ORTOLANI

Ortolani of Italy furthered emphasis on the great value of early diagnosis and early treatment in congenital hip dysplasia (219). He stressed the value of newborn examination of the hips based on his feeling that congenital dysplasia developed in utero and was present at birth. The hip instability was diagnosable clinically by the "click" sign in which the femoral head was dislocated with adduction and relocated with abduction maneuvers. This sign came to be referred to widely as the Ortolani sign and was considered by Ortolani and others to be seen most classically within the first 2 months of life. The sign was best noted in cases of mild to moderate hip dysplasia. Those with severely dysplastic hips often were not reducible and the clicking sensation was not detectable. Ortolani felt however that the vast majority of cases of congenital hip dysplasia were mild to moderate and thus detectable in the early weeks of life. 4. VON ROSEN AND BARLOW

Separate reports by Von Rosen (332) and Barlow (12) in 1962 stressed the value of routine neonatal hip examinations encompassing all newborns in detecting hip instability and allowing for careful observation and appropriate early treatment of those with instability. It shortly became accepted that the hip examination was valuable as a routine part of the neonatal assessment, an approach that has minimized considerably the late initial detection of the dysplastic hip. Barlow stressed that the most essential finding was the dislocatable-relocatable hip as determined by the clinical maneuvers. Limited abduction was not of value in the newborn because it was a later finding representing a secondary change. Asymmetry of the thigh folds also was of no specific value because less than half of the dislocated hips had asymmetric folds and the large majority of children with asymmetric folds had normal hips. Although many individual surgeons had recognized that it was possible to detect congenital instability of the hip in the neonatal period and that early examination and treatment of hip displacement almost certainly would lead to improved results and minimization of the occurrence of the secondary adaptive changes, it was not until the 1950s that widespread examination of newborns was studied and adopted as common practice. Large centers in Sweden and Great Britain were particularly prominent in this regard. Simple hip abduction splinting for several weeks led to a cure in the large majority of cases. Howorth, in a report in 1977, summarized several of these studies reported from 1950 to 1975 and indicated that 2010 displacements had been reported in 155,255 examinations (126). At present, it is widely accepted that hip examination specifically checking for instability should be an integral part of the neonatal pediatric examination. Procedural questions involve whether the procedure should be done by an orthopedic surgeon and whether ultrasound studies should be performed routinely in all children or only where clinical concerns have been raised. Plain radiographs do not have sufficient resolution to effectively

SECTION VIII ~ Treatment Approaches in Developmental Dysplasia of the Hip diagnose most cases of infantile developmental dysplasia of the hip. Secondary ossification centers of the proximal femur have not formed, the acetabular index is too variable, and there is no dynamic component to the method. Occasional hip dislocations can be missed even with careful clinical examination screening programs (203). Mitchell reports 4 dislocated hips diagnosed after walking began that had been either not recognized or not recognizable in a total of 31,961 newborn examinations between 1962 and 1968. The incidence of unstable (126) or luxated (100) hips detected early was 0.7%.

C. Assessments of Congenital Hip Dislocation Treated by Closed Reduction 1. RADIOGRAPHIC CLASSIFICATION SYSTEM (SEVERIN)

Severin greatly enhanced the detailed assessment of resuits by developing a clinical-radiographic categorization of the long-term results (278). He related the effects of treatment in childhood years to the likelihood of a predisposition to osteoarthritis in mid to late adult life if perfect anatomic repositioning was not achieved. He performed one of the earliest and most detailed studies of the results of closed reduction treatment of CDH and developed a categorization system based on the radiographic appearance, which relied heavily on the determination of the CE (center-edge) angle of Wiberg (Fig. 7). He ultimately reviewed treatment in 330 patients with 448 cases of congenital dislocation of the hip. Most patients had been treated in the 1920s and 1930s when the best time for treatment was considered to be between 2 and 3.5 years of age, even if diagnosis had been made prior to that time. As a prelude to his study, he performed additional studies on the CE angle of Wiberg, stressing in particular values between 6 and 13 years of age to compare with Wiberg's studies, which assessed men and women between 20 and 35 years of age. a. Measurement of the CE Angle Severin defined measurement of the CE angle as follows: "If one draws a line between the center of the head and the outside edge of the acetabulum and another through the center of the head parallel with the longitudinal axis of the body (at fight angles to the " Y " line of Hilgenreiner) the 2 lines will normally form an angle, the CE angle." In a normal hip the CE angle is positive, but with maldevelopment of the acetabular roof, upward and outward displacement of the head, or some other deformity, the center of the head moves closer to and perhaps beyond the lateral edge of the acetabular roof. The CE angle diminishes and when lateral to the edge of the bony roof it becomes negative. Wiberg's studies combined data from males and females and assessed values between 20 and 35 years of age. He defined an angle in that age group of less than 20 ~ as abnormal, 20-25 ~ as uncertain, and greater than 25 ~ as normal. Severin performed studies between the ages of 6 and 17 years involving 200 normal hips in 100 subjects, 52 females and 48 males. The values for children

I

205

E

Head/Neck Shaft Angle

F I G U R E 7 Measurement of the center-edge (CE) angle of Wiberg plays an important role in assessing results in the treatment of CDH-DDH. It is most valuable in adolescent and adult assessments, but can be used as early as 6 years of age.

14-17 years of age were similar to those of Wiberg in adults from 20 to 35 years of age, such that Wiberg's standards for the CE angle apply down to and including the age of 14 years. Values for the developing hip were represented by those between 6 and 13 years of age (136 hips); a CE angle less than 15 ~ was abnormal, between 15 and 19~ uncertain, and greater than 20 ~ normal. In summary, CE values greater than 20 ~ are normal for those between 6 and 13 years of age and values greater than 25 ~ are normal for those older than 14 years of age. Abnormal values are less than 15 ~ for those 13 years of age and less than 20 ~ for those 14 years of age or older. b. Severin Classification for Radiographic Assessment of Long-Term Results Severin performed a long-term study on 330 patients involving 448 hips that had been treated by closed reduction and who at the end of treatment were considered to be "primarily successful results" by the treating physicians (278). His follow-up series involved 306 patients with 417 involved hips. Of these, 266 were female and 40 male, a 6.7:1 female:male ratio. He categorized the end result into six groups. The functional results tended to deteriorate as the patient grew older because of the development of secondary changes in the joint. He indicated that "only the anatomically cured cases can reckon with freedom from future trouble." Those in group 1 were considered to be normal, but even here he noted that "there is always something to distinguish the hip with CDH from a normal one, even though some of the findings are sufficiently mild that the patient was still considered normal." Group 1. Anatomically well-developed hip joints with a spherical femoral head and a normal CE angle. Group 1 was subdivided into groups 1a and lb. 1a: CE angle of more than 19 ~ for ages 6-13 and more than 25 ~ for ages 14 years and up. lb: CE angle of 15-19 ~ for ages 6-13 years and 20-25 ~ for ages 14 years and up. Group 2. Hips exhibit a distinct but moderate deformity of the femoral head or neck or acetabulum but have a

206

CHAPTER 3 ~

Developmental Oysplasia of the Hip

satisfactory structure with otherwise normal conditions in the joint overall. This group is also subdivided into group 2a with normal CE values for age and group 2b with uncertain CE values for age. Group 3. Dysplastic hip joints with a maldeveloped acetabular roof but no subluxation. The CE angles are less than 15 ~ in those 6-13 years old and less than 20 ~ in those 14 years of age and up. Group 4. Subluxation. The femoral head is displaced upward and laterally in relation to the acetabulum and there is an unmistakable break in Shenton's line such that the hip is subluxated. The acetabulum is always more or less dysplastic and the femoral head often is deformed. Two subclassifications are present. 4a: Moderate subluxation with the CE angle still positive or equal to 0 ~ 4b: Severe subluxation with a negative CE angle. To be considered as a subluxation, the acetabulum must be a direct continuation of the original acetabulum. Group 5. The femoral head articulates with a secondary acetabulum developed at the margin of the original acetabular roof. This secondary acetabulum is separated from the original joint cavity and does not represent a continuation of it but rather a remolding phenomenon. Group 6. Complete re-dislocation of the femoral head. Deformities of the femoral head and neck occur in all of the groups except group 1. The deformities are moderate in groups 2 and 3, but from groups 4 to 6 they are of major degree and generally are combined with shortening and thickening of the neck. In the long-term study in which patients were categorized from group 1 to group 6, there were 417 hips assessed with the following results: group 1 (well-developed hip joints), 4.6%; group 2 (moderate deformation of the femoral head or neck or acetabulum in an otherwise well-developed joint), 7.7%; group 3 (dysplasia without subluxation), 8.6%; group 4 (subluxation), 47%, of these, group 4a with slight subluxation involved 17.7% and group 4b with severe subluxation involved 29.3%; group 5 (femoral head is in a secondary acetabulum in the upper part of the original acetabulum), 13.9%; and group 6 (re-dislocation), 18.2%. The prognosis was better for unilateral than for bilateral cases, and the later the hips were reduced the more deformities were seen in the femoral neck and femoral head. The late results for the joint as a whole were best if the reduction was done early, which in the then current framework meant before the age of 1 year. Femoral head and neck deformities invariably were seen simultaneously. 2. CLOSED REDUCTION IN THE EARLY WEEKS OF LIFE: PAVLIK HARNESS

Over the past few decades, treatment by closed reduction combined with earlier diagnosis, preferably in the newborn period, has become the desired and widely accepted approach. The Pavlik harness is desirable (222). Ease of

management of the child is improved compared to a hip spica, and hip motion in the reduced range is vastly more physiological than rigid immobilization. In a review of 3611 dislocated-dysplastic hips in 2636 patients less than 11 months of age using the Pavlik harness, healing rates were determined depending on the age treatment began and the degree of displacement based on the Tonnis radiographic and Graf sonographic classifications. The average age at initial treatment was 4.1 months (range = 2 days to 11 months). Results were better the lower the degree of displacement defined by the Tonnis 1-4 grading criteria. The healing rate at follow-upmaverage 4.5 years (range = 1-9 years)mwas Tonnis grade 1, 95.4%; grade 2, 92.3%; and Tonnis grade 3, 52%. (Few grade 4 hips were treated with the harness.) The authors concluded that the Pavlik harness could be the preferred primary treatment from the neonatal period to 6-7 months of age. The rate of AVN was 2.4% (higher in Tonnis 2 - 4 patients). A review of current approaches to closed and open reduction of congenital dislocation of the hip by Gabuzda and Renshaw summarizes present opinion well (84).

D. Treatment by Open Reduction By the middle third of the twentieth century the indications for open reduction were becoming much clearer, and it increasingly was resorted to when treatment was performed after 1 year of age. The use of arthrography made the indications for open reduction much more specific because lack of a fully concentric reduction of the femoral head into the depths of the acetabulum, which generally meant interposition of the limbus (labrum plus capsule) following closed reduction, was soon recognized by most to mandate the open intervention. There were some, such as Severin, who felt that hip spica immobilization could be continued even with inversion of the limbus with the expectation that pressure of the head against the acetabulum with time would wear away the inverted tissue (279). Most felt, however, that this either was not true or that if it succeeded it would still have left damage to the articular cartilage surfaces, which in the long run would not be ideal. In the opinion of Scaglietti and Calandriello, open reduction is resorted to in three situations: (1) in teratologic dislocations in which both the acetabular dysplasia and the displacement are so marked that they are considered to have originated during the embryonic or early fetal stages and are not amenable to closed treatment; (2) in those with increased age and dysplasia in whom the combination of clinical and radiographic findings indicate little to no likelihood of successful closed reduction; and (3) in all children over 3 years of age (267). The open reduction addressed each of the extraarticular and intra-articular obstacles to reduction described previously in the section on pathoanatomy. In Scaglietti and Calandriello's review of operative procedures, which ranged from those under 1 year of age to those between 4 and 5 years

SECTION VIII 9 Treatment Approaches in Developmental Dysplasia of the Hip of age, various findings were assessed. In the 162 children operated, only 11% were under 1 year of age, whereas there were 32% between 1 and 2 years, 29% between 2 and 3 years, 15% between 3 and 4 years, and 13% between 4 and 5 years. Shortening of the gluteus medius muscle was noted as a formal obstacle preventing reduction in only 3.2%, although the actual frequency probably was greater because this muscle attachment was freed from the iliac crest of necessity for operative approach for open reduction. Tightness of the iliopsoas muscle presenting as a real obstacle to reduction was seen over 50% of the time, and they almost invariably performed a " Z " elongation of the musculotendinous region. This eased reduction of the femoral head and partially relieved the isthmic constriction of the capsule. Pericephalic insertion of the capsule was seen in one-third of the patients and was often a sequel to previous attempts at closed reduction and hip spica immobilization. Occasionally, however, it was found in primary reductions. Both internal and external capsular adhesions were seen on occasion and were always felt to be acquired following closed reduction procedures. The inverted limbus-capsule was observed in 35%. In those less than 2 years of age it was rarely seen, and the femoral head slipped easily over the labrum with surgical attention not required. In those over 2 years of age who were walking, it appeared that the inverted capsule and limbus usually were hypertrophic. The ligamentum teres appeared congenitally absent in 20% of the hips. In 31% there were remnants of the ligamentum teres, usually as an atrophic fragment with fibrous fatty tissue in the depths of the acetabulum. In 49.2% the ligament was intact with conditions sometimes being thin and elongated, whereas in others it was long and flattened. It was always removed because it obstructed complete reduction. The head of the femur was deformed in 21%, usually being pear-shaped from flattening against the ilium. Some anteversion of the neck of the femur was almost always evident as determined by the degree of internal rotation necessary to reduce the head completely into the acetabulum. The anteversion rarely if ever, however, prevented the centering of the head and derotation osteotomy was needed at primary reduction in only 1.6% or 3 hips. Shallowness of the acetabulum was due mainly to obliquity of the roof and was always present to a varying extent. On only 11 occasions (6%) was acetabular reconstruction carried out at the same time as reduction. The depths of the acetabulum were always filled with a mass of fibrous fatty tissue usually containing ligamentum teres, and this tissue was always removed. In only 9 hips (5%) was the head too large for the acetabulum or so misshapen that the fit was inappropriate. In summary, Scaglietti and Calandriello concluded that in only 12% of hips was there only a single obstacle to reduction. The most frequently found abnormal structures, usually in combination, were a tight psoas muscle, hypertrophic ligamentum teres, fibrous fatty tissue in the depths of the acetabulum, a pericephalic insertion of the capsule, and an inverted limbus. Bony procedures rarely were done as

207

part of the primary procedure. When indicated due to continuing imperfect development, they were performed as secondary procedures involving either a proximal femoral varus-derotation osteotomy or a reconstruction of the acetabular roof. Their acetabuloplasty consisted of a curved osteotomy about 1 cm above the upper rim of the acetabulum into the iliac wall, levering down of the bone to cover the anterolateral part of the head, and maintenance of position with bone graft. In 187 hips in 162 patients treated surgically during the years 1947-1959, 72% of the patients were operated before the age of 3 years. The best results were obtained when the patients were operated on at an early age, which they defined as up to 3 years. Beyond that the percentage of good results decreased greatly. Their system of assessment observed 68.3% favorable results in 171 hips treated by open reduction. Open reduction was resorted to more frequently by more surgeons, and at this time a major consideration involved what to do with the inverted limbus (fold of labrum plus capsule). Some, such as Somerville (285), recommended excising the limbus, but most, including Salter (262) and Hall (105), strongly recommended saving the tissue, which in reality is an integral part of the normal structure of the hip, by freeing it and replacing it in its normal position supporting the head as a continuation of the acetabular cartilage. Open reduction was accompanied by excision of the ligamentum teres, removal of the fibrous fatty tissue in the depths of the acetabulum, opening of the inverted limbus so that it would be everted, and freed from its previous position allowing the articular cartilage of the head to relate directly to that of the acetabulum, with the repositioned labrum serving as a superior support for the head. It was essential to free the transverse acetabular ligament, which invariably was stretched and tightened and served as a major block to reduction of the head. Capsular repair (capsulorrhaphy) is an integral part of the open reduction procedure. Renshaw also stressed the need for concentric reduction, removing any interposed tissues between the femoral head and acetabular cartilages, by open reduction, if necessary (252). Mau et al. reported good to excellent results with the Ludloff anteromedial approach (192). In those operated beyond 18 months of age, the anteversion of the proximal femur and persisting acetabular dysplasia are taken into consideration. Here again opinions differ as to the approaches to be taken. Some feel that both of these areas of bone-cartilage deformation will correct once the head is appropriately repositioned and stabilized into the acetabulum. Others feel that a safer approach is to correct the bony deformities as well. Salter (262) strongly recommends the innominate osteotomy, whereas Pemberton (223, 224) used the pericapsular iliac osteotomy after 18 months of age. Some preferred to correct the femur with a derotation-varus osteotomy and allow the acetabular correction to occur spontaneously once function had improved. The timing of such procedures will be discussed in greater detail next.

208

CHAPTER 3 ~

DevelopmentalDysplasia of the Hip

E. Acetabular Corrective Procedures for Treatment of Hip Dysplasia 1. OVERVIEW OF DEVELOPMENT OF ACETABULAR PROCEDURES: THREE BASIC APPROACHES

Salter has clarified the differing principles underlying the three surgical approaches designed to improve acetabular dysplasia (265). (1) Acetabuloplasty refers to an incomplete osteotomy of the pelvis, which levers the roof of the acetabulum downward with its new position maintained by a bone graft. The variants of this procedure used most commonly are the Pemberton pericapsular osteotomy and the Mittelmeier type osteotomy. (2) An extracapsular shelf operation refers to an operation in which the existing acetabular roof is extended outside the fibrous capsule of the joint to serve as a more effective buttress for the femoral head. (3) The final procedure is a complete pelvic osteotomy extending into the sciatic notch, which allows the acetabulum to be redirected. Widely practiced variants of this technique are the innominate osteotomy of Salter (262), the medial displacement osteotomy of Chiari (151, 165), and the triple innominate osteotomy described initially by Steel (288). The various acetabular procedures and the effectiveness of the results have been reviewed by Tonnis (313-315,320). Deficiency of the lateral acetabulum has long been recognized as one of the associated features of a CDH, in particular one that has been left unreduced for more than a few months. Much attention has been directed to assessment and management of the acetabular deformity. As early as 1892, Konig performed an acetabular shelf procedure to increase the depth of the acetabulum and its lateral coverage (150). Many variations of this technical approach were proposed over the next 60 years. The shelf procedure does not redirect the acetabulum but augments it by providing a bone buttress superiorly and laterally. Until the early 1960s, major technical variations involved improved methods of containment utilizing shelf procedures to provide a buttress effect, increasing the lateral bulk of the acetabulum and thus increasing femoral head coverage and decreasing the CE angle. A major innovation was described by Salter in 1961 with development of the innominate osteotomy in which the entire acetabulum was tilted in an anterior and lateral direction to improve femoral head coverage (262). He utilized the procedure as early as 18 months of age. Around the same time, there were other modifications of acetabular procedures described by Mittelmeir (204, 205) in Germany and Pemberton (223, 224) in the United States, which involved transverse osteotomies just above the capsule and levering down and forward of the acetabulum followed by interposition of bone graft to maintain position and stability. Each of these procedures has had wide usage in varying parts of the orthopedic community. 2. ACETABuLAR PROCEDURES a. Innominate Osteotomy (Salter) The Salter innominate osteotomy can provide excellent femoral head coverage

up to the age of 6 or 7 years (262, 265). A pelvic osteotomy is performed and the entire acetabulum is redirected to improve both lateral and anterior coverage (Fig. 8A). Rotation occurs through the cartilaginous symphysis pubis and only one transverse cut in the pelvic bone is needed. A triangular wedge of bone from the anterior superior iliac spine region is removed and inserted into the opened gap to stabilize the pelvis and maintain the corrected position. Salter stressed that "the basic abnormality responsible for instability of the reduced congenital dislocation is the abnormal direction in which the entire acetabulum faces." Instead of facing downward it is directed anterolaterally such that the femoral head is covered inadequately anteriorly when the hip is extended and laterally when it is adducted. The principle of innominate osteotomy is redirection of the entire acetabulum to enhance stability in the upright functional position. The procedure has been widely although not universally adopted and serves as a major improvement in the treatment of CDH. The prerequisites for acetabular redirection, as defined by Salter and Dubos, include the following: (1) the necessity of bringing the head of the femur to a level opposite the acetabulum; (2) release of contractures of the adductor and iliopsoas muscles; (3) the complete concentric reduction of the femoral head within the true acetabulum whether by closed or open reduction; (4) congruity of the hip joint; (5) a good preoperative range of motion; and (6) a preferred age of intervention between 18 months and 6 years (266). Below 18 months of age, acetabular operations rarely are necessary because repositioning of the head into the acetabulum by closed or open means generally will allow for spontaneous acetabular correction to a normal range. By the 18-month age, normal bone development of the acetabulum and femoral head is no longer assured even with prolonged retention of the hip in the reduced position in the opinion of some, including Schwartz. Hall also has stressed that the first and most essential prerequisite in association with performance of the innominate osteotomy is complete concentric reduction of the femoral head in the true acetabulum (105). The innominate osteotomy does not change the shape or the capacity of the acetabulum but changes its direction so that the head of the femur is covered adequately in the standing position. Additional procedures have been described to overcome the fact that rotation of the acetabulum at the symphysis pubis is more difficult due to diminished flexibility after 6 or 7 years of age. Osteotomies have been devised in which the ilium and the superior and inferior pubic rami are cut to allow a freely floating pelvic segment to be tilted into appropriate position (the Steel osteotomy). Sutherland and Greenfield performed an osteotomy just lateral to the symphysis pubis and removed a small piece of bone to allow for easier rotation of the acetabulum and some medial displacement (295). Another innovation was the acetabular "dial procedure," done at skeletal maturity, which involves cutting the bone around the entire acetabulum close to the articular

S E C T I O N VIII ~ T r e a t m e n t

Approaches

in D e v e l o p m e n t a l

D y s p l a s i a o f t h e Hip

F I G U R E 8 (Ai) The principle of the innominate osteotomy is illustrated. (a) A lateral view of the pelvis shows the line of osteotomy above the joint. (b) Postosteotomy the site has been opened to displace the acetabulum anteriorly and laterally to improve coverage of the head. The bone graft from the iliac crest is inserted as shown and held with two Kirschner wires. (c) Anteroposterior view postosteotomy. (Aii) An example of the effectiveness of the innominate osteotomy in correcting acetabular dysplasia is shown. (Aiia) Anteroposterior X ray of the pelvis shows a normal hip on the right and acetabular dysplasia with lateral subluxation of the proximal femur on the left. The secondary ossification center on the right is within the lower inner quadrant, whereas on the left it is positioned laterally in the lower outer quadrant. The acetabular index measures 20 ~ on the right and 36 ~ on the left. The X ray was performed at 6 months of age. (Aiib) Anteroposterior radiograph at 2 years of age shows persisting lateral subluxation of the left femur with acetabular dysplasia. The secondary ossification center now is considerably smaller on the left than on the right. (Aiic) Innominate osteotomy was performed and stabilized with two K-wires. The distal fragment was displaced somewhat laterally and tilted anteriorly. (Aiid) Anteroposterior X ray at 3 years shows full containment of the left femoral head, increased size of the secondary ossification center, and correction of the acetabular dysplasia. (Aiie) Anteroposterior radiograph at 8 years of age shows a normal hip on the left with excellent acetabular development and an intact Shenton's line. (Aiif) Lateral hip film at 8 years of age highlights normal hip on the left. (B) The most surgically direct approach to persisting acetabular dysplasia beyond 2 years of age involves both acetabular and proximal femoral surgical corrections. In a combined procedure favored by Tonnis and many others particularly in Europe the

209

CHAPTER 3 " Developmental Dysplasia of the Hip

210

B

[3

C

FIGURE 8 (continued) acetabulumis levered downwardwith the correction held by a bone graft (an acetabuloplasty),and the proximalfemur is redirected into the depths of the acetabulum by a combined derotationand varus osteotomy.[Part B reprinted from Fritsch, E. W. (1996). Clin. Orthop. Rel. Res. 323:215-225, 9 LippincottWilliams& Wilkins,withpermission.]

cartilage and rotating the bony acetabulum to a better position but not interrupting the medial rim of the pelvis. At skeletal maturity acetabular dysplasia is treated by a periacetabular osteotomy as described by Ganz (90, 201). b. Pericapsular Osteotomy of Ilium (Pemberton) The Pemberton osteotomy not only redirects the sloping acetabular roof but also diminishes the capacity of the acetabulum (222, 223). This is essentially a hinge-type procedure. Pemberton described a pericapsular osteotomy of the ilium to correct directly the acetabular dysplasia and CDH. His rationale for the procedure was the desirability of rotating the acetabulum forward and laterally to close the anterior defect by shifting the more posterior and medial portions of the socket over the femoral head to provide a good weight bearing surface. The procedure was designed to make the

acetabulum somewhat smaller without undue acetabular distortion to stabilize the hip. The acetabular roof is rotated around the femoral head because one of the rationales for the procedure is the fact that the femoral head is small relative to the acetabulum in CDH-DDH. The procedure is completed posteriorly through the posterior limb of the triradiate cartilage. There is sufficient plasticity in the horizontal arms of the triradiate cartilage to permit the downward, anterior, and lateral displacement of the roof of the acetabulum and to prevent fracture or excessive distortion of the acetabular joint surface. The osteotomy is done separately through both tables, inner and outer, of the ilium. Pemberton and others have stressed that the osteotomy must not go into the sciatic notch because a fracture into the notch would greatly destabilize the acetabular segment. The earliest recommended time of intervention is 18 months with an upper age limit of approximately 8-10 years, dependent on the presence and plasticity of the triradiate cartilage. The best results occurred in those under 4 years of age because the procedure is easier to perform and acetabular remodeling is greater than at later ages. A high level of acceptable results also was achieved in the group from 4 to 7 years of age, whereas those between 8 and 12 years had a higher proportion of fair results. Pemberton was in agreement with Salter and others that the tendency for acetabular correction based on femoral repositioning was limited by 18 months of age and that operation as early as 18 months and preferably between 18 months and 4 years of age on the acetabulum itself produced excellent results. He felt that "the defect in the acetabulum allowing the head to come out was not one of a relatively shallow acetabulum but was rather an abnormal direction with the acetabular surface directed forward and laterally as well as being comparatively shallow." The operation is based on the principle that "since the triradiate cartilage is the only flexible structure in which the size and shape of the acetabulum could be changed it was determined that this would be used as a hinge. If the iliac portion of the acetabular roof was detached, the iliopubic and ileo-ischial limbs of the triradiate cartilage could be used as a variable hinge to wrap this iliac portion around the femoral head either anteriorly, laterally, or a combination of both directions." Once the acetabulum is repositioned it is locked into position with a bone graft from the iliac crest. Early reviews by both Coleman (46) and McKay (193) were favorable. They stressed the prerequisites of the pericapsular iliac osteotomy: (1) the hip must be concentrically reduced or reducible at open operation; (2) the hip must have a satisfactory range of motion; and (3) congruity of the femoral head and acetabulum must exist. The incomplete pericapsular osteotomy rotates the anterior and superior portion of the acetabulum forward, laterally, and downward, utilizing the triradiate cartilage as its fulcrum. This is in comparison to the complete innominate osteotomy, which rotates the entire acetabulum with the site of rotation through the pubic symphysis. Salter and others were concerned that the incom-

SECTION VIII ~ Treatment Approaches in Developmental Dysplasia of the Hip plete pericapsular osteotomy of Pemberton, which hinged at the triradiate cartilage medially, of necessity would produce an angulation of the acetabular articular surface with altered internal congruity. c. O t h e r A c e t a b u l o p l a s t i e s Other acetabular osteotomies have been developed in Europe by Lance (153), Dega (53), Mittelmeir et al. (204, 205), and Chiari. Schulze et al. commented favorably on the Dega procedure but were tending to switch to the Pemberton technique (271). The Mittelmeir procedure, acetabuloplasty combined with varus-derotation osteotomy (in which there must be complete femoral head reduction into the acetabulum prior to the acetabuloplasty), is also favored by Tonnis (315-318) (Fig. 8B). Tonnis' modification involved osteotomy of the acetabulum under fluoroscopic control "5 mm" from the roof, directing the osteotome to the most posterior point of the triradiate cartilage. Only a small bony bridge was left intact at the innermost posterior wall directly above the triradiate cartilage. The acetabular roof was levered down and held in position by a wedge-shaped bone graft. These acetabular procedures, often combined with proximal femoral varus derotation osteotomies, are most effective in the first 6 years of life. The Chiari osteotomy, passing obliquely upward from the lateral margin of the acetabulum through the inner iliac wall, is designed to medialize the femoral head and pelvis leaving the superior iliac segment to support the head with the capsule interposed. It is used infrequently today especially in North American centers. Teot et al. have experimentally attempted a biological augmentation acetabuloplasty with a vascularized pedicle graft transferred from the iliac crest, utilizing both the growth cartilage of the crest plus some adjacent iliac bone (304). Another experimental approach to acetabular correction was the application of physeal distraction with an external fixator to the triradiate cartilage in immature dogs 2.5 to 4 months of age. The procedure was designed to enlarge the acetabulum in both depth and width (1). 3. RESULTS OF ACETABULAR PROCEDURES Two subsequent reviews by independent groups have supported the value of the Pemberton procedure. Eyre-Brook et al. reported on 37 procedures and favored this type of acetabular procedure when a direct approach to the acetabulum was warranted (67). A large majority of their procedures, which were either isolated or combined with a primary open reduction, were performed between 1 and 3.5 years of age. One of the concerns with this procedure had always been with the fact that correction was obtained through the triradiate cartilage with the possible dual complications of premature fusion and subsequent growth irregularity of the acetabulum and creation of an incongruous shape of the acetabulum, which subsequently would not remodel. Eyre-Brook et al. specifically studied the triradiate cartilage in follow-up radiographs and felt that early closure or other abnormality was not seen. They commented that

211

the hinge occurred due to a fracture-separation at the triradiate cartilage with movement occurring from the fractureseparation either laterally into the acetabulum or medially toward the pelvis. This would be best demonstrated by CT scanning or MR imaging, although we are not aware that such studies have been performed. Faciszewski et al. reviewed 52 procedures with an average follow-up of 10 years (69). The average age of the patients in their series was 4 years with a range between 3 and 10 years. The upper age limit was 10 years because the triradiate cartilage was relatively thin after that age. They identified no radiographic evidence of premature arrest of the triradiate cartilage or of acetabular chondrolysis and concluded that the pericapsular osteotomy was a safe and effective procedure for the treatment of residual acetabular dysplasia. Plaster et al. did report premature arrest of the triradiate cartilage after a Gill acetabuloplasty at 14 months of age, but they attributed it to the placement of bone graft across the physis itself (230). Acetabular dysplasia worsened with time. Results by Salter and colleagues in Toronto showed an extremely favorable response to the innominate osteotomy procedure from the initial report in 1961 onward (262, 263, 265, 266). Many other centers achieved excellent results with this procedure as well, although some had difficulty. Salter and colleagues have stressed the need to adhere stringently to the prerequisites and indications for the procedure as well as close attention to the technical detail. Salter and Dubos, using open reduction and innominate osteotomy for complete dislocation, noted 94% excellent or good results with treatment from 1.5 to 4 years of age, which diminished to 57% excellent-good at the age of 4-10 years (266). With osteotomy alone for subluxation, excellent results were achieved in 94% of 16 hips at 1.5 to 4 years of age and 58% excellent and 33% good in 12 hips 4-16 years of age. Roth et al. had an 85% excellent result rate with primary open reduction-innominate osteotomy in 65 dislocated hips in children 1.5-4 years of age at initial treatment and a 92% excellent rate in 12 hips with osteotomy alone for subluxation at 1.5-4 years (258). Salter, in his initial report in 1961, described results in 25 hips between the ages of 18 months and 6 years (262). Ninety-two percent of those patients had excellent (80%) or good (12%) results by the Severin classification. Heine and Felske-Adler studied 43 hips treated with the innominate osteotomy and open reduction under 4 years of age and found the hips to have normal or slightly abnormal acetabular indices, although when assessed in terms of the CE angle and other hip indices the results were somewhat disappointing with only 34% considered good or fair (114). Some of the less than excellent results were due to the relatively high rate of preoperative AVN and due to the fact that many patients had had other procedures. Blamoutier and Carlioz studied the results of their Salter procedures in 43 hips with an average follow-up of 10 years (21). Overall results were satisfactory in 60% of cases with much better results in those under 5 years of age at surgery. Many of the

212

CHAPTER 3 " Developmental Dysplasia o]c the Hip

failures were related to technical difficulties, although imperfect results seemed to follow some cases in which no problem had been encountered. They eliminated patients who had previously had AVN as a result of earlier treatment. The average age at operation was 3 years 10 months, with a range between 18 months and 10 years. Complications were infrequent and no case of AVN attributable to the procedure was seen. According to the classification of Severin, 83.3% of the hips were classified as an excellent group 1 result. With a more rigorous classification, however, based on several radiographic indices, the number of normal hips diminished to 60.4%. The problematic results involved decreased ranges of motion, persisting dysplasia of the acetabulum, and excessive coverage of the hip. When patients were followed for 10 years or more, many reports of less than perfect results were seen. Mader et al., assessing 20 cases, did not observe any with fully normal sphericity and concentricity (185). Morel studied 23 cases and found 60.8% with good results (206). Fournet-Fayard et al. noted 76.7% good and excellent results (81). Blamoutier and Carlioz concluded that the Salter osteotomy was useful over the long term for correction of residual acetabular dysplasia, particularly if the operation was done before 5 years of age. The operation was technically demanding, however, and if performed imperfectly was unlikely to lead to a good result. Even with the procedure, however, a considerable number of patients could be seen with persisting dysplasia several years afterward. In a review by Waters et al., 29 hips operated between the ages of 18 months and 5 years were assessed for CDH (341). All were felt to have good or excellent radiographic results using the Severin classification. Three patients required reoperation for failed initial procedures. The average length of follow-up was 9 years 3 months. The immediate postoperative mean acetabular index was 12~ which was the same as the long-term index. The center edge angle increased immediately following surgery to a mean of 36 ~ and the long-term result had a mean of 39 ~ In the Severin radiographic classification, 25 of 29 hips were class I, 3 were class II, and only 1 was class IV. McKay reported 73% of 26 patients between the ages of 18 months and 6 years with Severin class I or II radiographic results after innominate osteotomy (193). Crellin reviewed 25 hips in 21 patients treated by innominate osteotomy between 14 months and 5 years 3 months, with 72% having Severin grade I results and 24% Severin grade II (50). Gallien et al. assessed 43 hips treated between the ages of 14 months and 4 years with 68% having good or excellent results (87). They reported a 5% incidence of AVN. Barrett et al. assessed 42 hips treated between 18 months and 4 years 2 months by either innominate osteotomy alone or combined with open reduction and noted 88% with either grade I (62%) or grade II (26%) Severin gradings (14). Hansson et al. reported on 83 procedures with overall radiographic results good or excellent in only 41% and fair or poor in 59% (108). The study reviewed the first 15 years of experience in

the entire country with the procedure in which 26 surgeons were involved. Their best results were obtained in hips with subluxation that had not been treated previously or treated only with closed reduction. The poorest results were obtained in hips with residual subluxation or dislocation after previous operation. The results in this Swedish series, however, were best when the procedure was performed before the age of 5 years. Windhager et al. reviewed 63 innominate osteotomies done at a mean age of 4.1 years (range = 18 months to 18 years) and followed for a mean of 15.7 years (359). The failure rate was 29%; 25% in those operated before 4 years of age and 41% in those operated at a later age. Improvement with time helped some in the younger group but not in the older group. Those with mild or moderate dysplasia presurgery did better than those with severe pathologic dysplasia.

F. Proximal Femoral Osteotomies As the patient increases in age, frequently it is necessary to perform either acetabular or femoral osteotomies and occasionally both. When growth occurs with the femoral head not well-seated in the acetabulum, the proximal femur tends to persist with a coxa valga and anteverted structure. Deeper seating of the femoral head is improved by proximal femoral osteotomy; much improvement can occur with derotational osteotomy alone, although on occasion an element of varus correction is added. The improvement in overall hip structure has been noted frequently even when only one component of the bony deformity, the acetabular dysplasia or the proximal femoral anteversion-valgus deformity, is corrected. In those between 2 and 4 years of age, surgical correction of acetabular dysplasia often with open reduction and capsulorrhaphy will lead to improvement of the femoral structure without additional intervention. Other schools of orthopedics note improved acetabular depth following proximal femoral derotation-varus osteotomy alone. A femoral shortening procedure also plays a role in the surgical treatment of complete hip dislocation, especially in association with open reduction and acetabular-femoral osteotomy in a patient 3 years of age or greater, but it can be used in those who are even younger. That decision is made at the time of surgery based on the ease of reduction. The frequency of avascular necrosis is substantially lower in the older age group when femoral shortening accompanies the reduction. The procedure is designed to reduce the tightness of structures about the hip and thus prevent excess stretching and tightening of the associated vasculature. Karadimas et al. studied growth of the proximal femur after varus-derotation osteotomy in the treatment of CDH (142). The neck shaft angle was noted to improve from varus to the normal range with time. In their patients, the average age at reduction and osteotomy was 2.3 years. Patients were divided into three groups to assess the subsequent growth response to the angle of varus osteotomy: those corrected to

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

less than 100 ~ those between 100 and 110 ~ and those to greater than 110 ~ The optimal varus correction was to between 100 and 110 ~ because the highest percentage of patients remodeled when left with this angulation after the procedure. If varus was greater with correction to an angle of inclination less than 100 ~ fewer hips regained a normal neck shaft angle. Osteotomy was done in three groups: those with osteotomy alone, those with excision of the limbus along with osteotomy, and those requiring formal open reduction and osteotomy. In each of the three groups at the mean age of 2 years, the initial varus correction was between 100 and 110 ~ and in each the angle increased with time. In the osteotomy alone group, the neck shaft angle gradually increased to approximately 122 ~ at age 8 years, after which it remained stable to the termination of the study at ages 16 and 17 years. Similar patterns were found in the other groups, with ages 8-10 years representing the time to which correction occurred, after which there was either no change or slight diminution in the angle. Chiunard and Logan concluded that the ideal neck shaft angulation after varus osteotomy was 90-100 ~ although Karadimas et al. felt that slightly less varusization between 100 and 110 ~ was better. Mau also noted spontaneous straightening of the femoral neck following varus osteotomy in CDH treatment. The indications for femoral intervention are being narrowed gradually. Bialik and Benyamini express a relatively common view based on their patients assessed between 1976 and 1990. When femoral osteotomy was performed, they generally opted for both derotation and varus correction. The average age for patients having this procedure was 46 months (range = 18-120 months).

G. Combined Acetabular and Proximal Femoral Osteotomies Fritsch et al. studied femoral and acetabular remodeling after proximal femoral intertrochanteric varus-derotation osteotomy done in combination with an acetabuloplasty with the wedge in-lay technique for CDH. In earlier studies, Mittelmeir reported that only 40 of 155 hips developed sufficient correction of the acetabular dysplasia with an acetabular angle below 20 ~ after femoral osteotomy alone in early childhood. The dual procedure was recommended for use between 2 and 5 years of age. They performed an intertrochanteric varus osteotomy and used the medial-based bone wedge to insert into the acetabulum to help stabilize roof deflection. They aimed for a slight primary overcorrection of the neck shaft angle to 110 ~ with correction of the anteversion of the femoral neck to plus 10~ A small compression blade plate was used for stabilization. Their report assessed 101 patients. The preoperative average of the neck shaft angle was 143 ~ (104-168 ~ with 69 hips showing a severe valgus deformity greater than 140 ~ Postoperatively the average of 112 ~ was noted (94-130~ and this increased to a mean of 129 ~ at the time of the latest assessment (90-158~

213

The average varus correction was 32 ~ The postoperative spontaneous increase in the neck shaft angle was 17.9 ~ indicating an average percentage increase of 61%. Acetabular correction maintained itself over the several years (mean = 9 years) postsurgery. In the 101 hips the mean preoperative acetabular angle was 33.8 ~ with postoperative correction to 19.2 ~ and angle at follow-up 17.9 ~. The center-edge angle also maintained itself long-term: preoperative 8.9 ~, correction to 24.8 ~ and maintenance at 24.8 ~ at 9 years. Bernbeck initially reported great expectations that the proximal femoral osteotomy alone would have a positive effect on acetabular development (18). Blockey (22) and Kasser et al. (144) found that intertrochanteric osteotomy alone was sufficient to treat residual acetabular dysplasia in the second and third years of life and even in older children with only mild deformity. Many others including Mittelmeir (204), however, felt that proximal femoral osteotomy did not improve an imperfect acetabular roof sufficiently on its own. Fritsch et al. felt that one of the main advantages of the varus-derotation osteotomy was to improve the centering of the hip such that the biomechanical stability could be improved remarkably by the varusization (82). There was some reciprocal improvement on the acetabular side in their opinion, but not enough for the varus procedure to stand on its own. The varus procedure reduced the intra-articular pressure, which was increased by the associated acetabular repositioning. They presented some evidence to indicate that, although many cases of AVN of the femoral head were related directly to open reduction, the rate of postoperative pressure deformities seemed to be higher when acetabular osteotomy of various types was not combined with either varus or shortening femoral procedures. The proximal femoral varus osteotomy also minimizes the anteversion that is an integral part of the overall hip correction. Although there was spontaneous increase in the neck shaft angle postsurgery, this tended to place the value at maturity into the normal range. Fritsch et al. noted very few instances of pathological recurrent valgus. They concluded that "doing a varus derotation osteotomy in combination with a correction of the acetabulum has no disadvantages and only beneficial effects regarding the shape of the acetabulum and the femoral head and neck development." Reichelt and Hansen felt that the best results were seen with a head-neck shaft angle of 115 ~ (range = 110-125 ~ (249).

IX. I M A G I N G T E C H N I Q U E S U S E D T O ASSESS HIP POSITION The plain radiograph remains helpful in assessing development of the hip throughout childhood. Other imaging modalities take precedence, however, in assessment of developmental dysplasia of the hip in particular at extremely important times during diagnostic and treatment activities. Ultrasound has become the primary mode of assessment in

214

CHAPTER 3 ~

Developmental Dysplasia of the Hip

the newborn and in relation to the effectiveness of treatment particularly during the first 6 months of life. CT scanning is virtually essential to document the effectiveness of closed or open reduction when the child is immobilized in hip spica because plain radiographs in the anteroposterior projection can be totally misleading. The CT image in cross section defines the position of the head in relation to the acetabulum as well as acetabular depth. Three-dimensional CT reconstructions help direct corrective acetabular osteotomies. MR imaging can be used to assess femoral head vascularity immediately following closed or open reduction.

A. Plain Radiographic Indices Many studies have been published on the normal values of several indices as determined by plain radiographs throughout the childhood years. The most valuable are the acetabular index of Hilgenreiner (116), the center-edge angle of Wiberg (350), and the time of appearance of the femoral capital ossification center (291,361). The plain radiograph is used infrequently today in the first few months for assessment of hip dysplasia because the ultrasound has proven to be of such great value (93-95). Even prior to sonography plain radiographs in the newborn period more often than not were misleading concerning the presence of hip position and stability. The absence of the secondary ossification centers at birth and a tendency for most hips to relocate in the abducted position often lead to the opinion that no abnormality was present. This often resulted in considerable delay prior to diagnosis. Although the acetabular index often was elevated in DDH, the range of normal and abnormal values frequently overlapped such that a definitive diagnosis could not always be made in this fashion. Figure 9A illustrates some of the reference points used on the anteroposterior radiograph. Figure 9B points out the underlying anatomy and its relation to the pathogenesis of acetabular dysplasia. The radiolucent tissues can be revealed by imaging modalities beyond the plain radiograph.

1. MEDIAL AND SUPERIOR GAP MEASUREMENTS Bertol et al. used the anteroposterior radiograph to good advantage to assess a couple of indices that, if not definitive, strongly pointed to the presence of a poorly positioned hip (19). In their unit abduction radiographs were not obtained. On the anteroposterior film the medial gap that measured the separation between the proximal femur and the pelvic wall was increased significantly in cases with dislocation. A medial gap greater than 5.0 mm was indicative of femoral head displacement. A superior gap represented the distance between the most superior portion of the proximal femur and Hilgenreiner's horizontal line through the triradiate cartilages. The normal superior gap was 9 mm and the medial gap 4.5 mm. With lateral displacement the medial gap widened, and with superior displacement the superior gap narrowed. In normal patients the medial gap was

4.1 + 1.1 mm and the superior gap was 9.5 + 0.8 mm. Values for medial gap abnormalities in unilateral left and right CDH cases were 5.9, 5.5, and 4.8 mm, whereas in bilateral cases they averaged 6.2 mm. The superior gap diminished in unilateral cases to values between 7.9 and 8.6 mm, with further displacement in the more severe bilateral cases of 7.0-7.2 mm. 2. TONNIS PLAIN RADIOGRAPHIC CLASSIFICATION Tonnis has defined a plain radiographic classification for assessing the degree or stage of developmental dysplasia of the hip (315). This classification is dependent on the presence of the secondary ossification center. Grade 0 is a normal hip; grade 1, ossification center of the proximal femoral capital epiphysis is medial to Perkins' line; grade 2, ossification center is lateral to Perkins's line but below the superolateral bone margin of the acetabulum; grade 3, ossification center is at the level of the superolateral margin of the acetabulum; and grade 4, ossification center is above the superolateral margin. Perkins originally pointed out the appropriate normal position of the femoral head ossification center on a plain anteroposterior radiograph (225). The bony head was, up until 4 years of age, always beneath a line joining the innermost parts of the ilium at the upper part of the triradiate cartilage and medial to a vertical line from the outer margin of the bony acetabulum described by Perkins as the anterior inferior iliac spine. These two lines enabled one to gauge whether a hip was fully reduced or persisted with subluxation or dislocation. The vertical line is known as "Perkins' line."

3. ACETABULAR INDEX (ACETABULAR ANGLE) Several studies have documented the normal range of values for the acetabular index (AI) with growth (5,273). In material derived from over 2000 radiographs from birth to 7 years of age, the mean AI in girls was 30 ~ at birth to 1 month of age, with diminution to 23 ~ at 5 - 6 months, 20 ~ at 2-3 years, and 15~ at 5-7 years (361). It has been widely recognized that the index varies with even slight alterations in positioning and radiographic technique. Work from the 1920s and 1930s, however, established the AI as the first sensitive radiographic indicator of an early hip dysplasia because it was observable from birth and several months before the normal appearance (and abnormal delayed appearance) of the secondary ossification center of the head of the femur (116, 241, 242, 149). The abnormal AI was due to the increased slope of the bony acetabulum, which reflected delayed development of lateral acetabular bone (deepening the acetabulum and diminishing the slope), which in turn reflected development of lateral acetabular cartilage. Subluxation of the head led to dysplasia by increased pressure against the lateral acetabular cartilage, slowing its growth, and complete dislocation of the head removed the major stimulus to growth (Fig. 9B). Hilgenreiner used radiographs initially to assess developmental retardation of the acetabulum and proximal

F I G U R E 9 This figure illustrates some of the plain radiographic indices used to assess hip position and development. (A) The anteroposterior film is used for measurements. The lines of Hilgenreiner and Perkins and the relation of the secondary ossification center of the femoral head to them are the major radiographic determinants of hip development. Up until 4 years of age the secondary ossification center is contained almost completely within the lower inner quadrant in the normal (1). Any movement laterally (2) or superiorly (3) indicates subluxation to dislocation. The acetabular index is a major determinant of the degree of acetabular development (Fig. 6). With lateral subluxation the head would appear either partially or totally in quadrant 2; with increased subluxation and proximal displacement it would appear in quadrant 3, and where there is complete dislocation and primarily anterior or posterior positioning the femoral head would appear in quadrant 4 overlapping acetabular bone. (B) When looking at a plain radiograph of the developing hip it is important to recognize that the radiolucent acetabular cartilage, fibrocartilaginous labrum, and capsule are not visualized (top) (Bi). The radiolucent roof is composed of three structures visible by arthrography or ultrasonogram (labrum, acetabular, cartilage, capsule). The pathogenesis of acetabular dysplasia is illustrated below (Bii). Although the acetabular index measures the bony conformation of the acetabulum, its underdevelopment is secondary to imperfect development of acetabular cartilage and imperfect positioning of the labrum and thus represents a secondary indirect change. In the dysplastic hip with subluxation and dislocation, the secondary ossification center is smaller in size, more laterally positioned, and often irregular in shape and density. The bony acetabulum slants obliquely upward at a greater angle, and the outer margin (open arrow) usually is rounded compared with the normal sharp comer (solid arrow). (C) The acetabular index invariably decreases toward the normal once the head is relocated in the acetabulum and maintained there. All hips with an acetabular index greater than 30 ~ after 6 months of age are considered pathologic. The chart illustrates the relatively rapid diminution of the acetabular index to the normal range once treatment is begun in the first 8 months of life. Treatment involving closed or open reduction alone started after that age still leads to improvement of the acetabular index but does so at a much slower rate. Treatment by reduction alone after 18 months of age may not lead to completely normal acetabular reconstitution. The index at the start of treatment is indicated by the solid circle, and the eventual result with closed treatment is shown by the open circle below. (Di) Plain radiographic characteristics on an anteroposterior film of a normal hip on the right and a dislocated left hip. Note the normal acetabular index on the right, the larger secondary ossification center present in the lower inner or first quadrant, the intact Shenton's line (arrow),

F I G U R E 9 (continued) and the bony arc of continuity between the outer ilium and the lateral femoral neck. On the displaced left side the secondary ossification center is smaller and is laterally and proximally displaced to lie in the upper outer or third quadrant. The acetabular index is increased markedly due to the dysplastic acetabulum, and both Shenton's line and the outer iliac-lateral neck arc are displaced. (Dii) A normal hip is seen with the acetabular index within normal range bilaterally, and each secondary ossification center developed to the same size and within the first quadrant. (E) Several examples of hip dysplasia are seen. These outline abnormalities in the acetabular index, absent or smaller secondary ossification centers on the affected side, lateral and proximal displacement of the femoral head in relation to normal positioning, and imperfect arcs involving Shenton's line and the lateral iliac-cervical line (El). A newborn with left hip dislocated and bony acetabular roof already dysplastic (greater obliquity, diminished subchondral bone). (Eli) Right hip dislocation--secondary ossification center smaller than opposite normal side, lateral and superior displacement of proximal femur, acetabular dysplasia, rounded outer "corner" of bony acetabulum (open arrow) on right dysplastic side and squared-pointed "corner" on normal left side (solid arrow). (Eiii) Dislocated left hip with absent proximal femoral secondary ossification center (markedly delayed appearance). Acetabular dysplasia. (Eiv) Left hip dislocation. Secondary ossification center present but smaller or dislocated side. Severe acetabular dysplasia. (Ev) Bilateral hip dislocation. Bilateral acetabular dysplasia. (Evi) Dislocated right hip. Note difference in subchondral acetabular bone. (Evil) Bilateral hip dysplasia with dislocation seen toward end of first decade. Severe acetabular dysplasia. [Part A reprinted from Quinn, R. H., et al. (1994). J. Pediatr. Orthop. 14:636-642, 9 Lippincott Williams & Wilkins, with permission. Part C reprinted from (4), with permission. Part Di reprinted from Tonnis, D. (1987). "Congenital Dysplasia and Dislocation of the Hip," Fig. 9.15a, p. 110, 9 Springer Verlag, with permission.]

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

femur in hypoplasia and clearly illustrated the acetabular angle or index (116). Putti showed the value of early hip radiographs as an invaluable aid to the diagnosis of dysplastic hip disease (241,242). Kleinberg and Lieberman further stressed the value of the acetabular angle in assessing hip development (149). They were among the first to document definitively the value of the acetabular angle in differentiating between normal and early dysplastic hips. They studied 23 normal infants from 1 to 7 days of age, and the radiographs noted an average acetabular angle of 27.5 ~ (range = 25-32~ In a second group of 20 normal children in age varying between 11 and 24 months, the acetabular index in 40 hips was 20 ~ (range = 18-25~ They next studied 35 congenitally dislocated hips in children from 12 to 36 months of age, some of whom had undergone treatment. The acetabular index was much higher at 37.5 ~ (range = 28-49~ They concluded that, in an infant having a high acetabular index above 30 ~ a dislocation of the affected hip was likely. They also commented on the value of early treatment, indicating that any infant with a high acetabular index should be placed in an apparatus that would lower the limbs into marked abduction, allowing for femoral head-acetabular growth conformity and clinical stability. An acetabular angle exceeding 30 ~ was considered pathological by the large majority of observers even in the newborn and was considered to indicate acetabular dysplasia, subluxation, or true dislocation (159). Virtually all studies, however, show some infants in the newborn to 1-month-old bracket with an AI a few degrees above 30 ~, which is understandable because the 30 ~ number is often an average value. The detailed study by Coleman of 1956 hip radiographs in normal and clinically abnormal Navajo children from birth to 3 months of age documents the number of studies with AI greater than 30 ~ in those clinically normal (45). The 30 ~ plus number is most worrisome if it persists after 3 months of age. Almby and Lonerholm concluded, after a detailed study of the acetabular index in normal and abnormal hips, that "after the first 6 months of life an acetabular angle exceeding 30 ~ is invariably pathologic (5)" (Fig. 9E). Due to the wide range of variability in the first few months of life, the acetabular angle was not always definitive in the early time periods. The acetabular angle invariably was higher on the unstable hip side than on the stable hip side. Before treatment the angle of the unstable hip was greater than on the stable side in all unilateral cases, with the difference of 10 ~ or more in 8 of 22 infants diagnosed before the age of 7 months and more than 10~ in all 6 unilateral cases diagnosed after 7 months. The acetabular index tended not to change much in the untreated unstable hips, with the mean value at 1-2 months of age of 38.2 ~ and that at 18 months and greater of 43.2 ~. Between 1 and 18 months of age the average values ranged between 36.8 and 41.6 ~ The acetabular index also was shown to be quite valuable in assessing responses to therapy. After treatment the acetabular angles were reduced in all cases. In the cases treated conservatively, the angles in the

217

previously unstable hips diminished and even reached the normal values except in the age groups over 16 months. The study included measurements of the same radiograph by different observers, and on average the measurement of the acetabular index did not exceed 2 ~ (Fig. 9C). The upper range of normal in the newborn was felt--in separate reports--to be 30 ~ (159). Hilgenreiner himself came to regard the " h " distance as a more reliable sign of hip malformation because it was less subject to change with positioning variations. The great sensitivity of the acetabular index to differentiate normal from dysplastic hips also was shown in a report by Dyson et al. of 160 infants less than 1 year of age with congenital dislocation of the hip or a milder acetabular dysplasia (63). Normal values in 68 hips were 27 ~ in 41 hips with acetabular dysplasia 32.3 ~ and in 47 hips requiring reduction 37.3 ~ Scoles et al. documented the acetabular index in a study of serial radiographs of normal children at 3, 6, 9, 12, 18, and 24 months of age (273). There were 100 children studied, 50 males and 50 females. The mean acetabular index in gifts at 3 months of age was 25 ~ with a progressive decrease to 18~ at 24 months. In males the 3-month mean was 22 ~ decreasing to 19 ~ at 24 months. Harris defined the normal acetabular index as 30 ~ at birth, 25 ~ from 1 to 3 years of age, and 10 ~ from 8 years of age to adult years (110). The acetabular angle (acetabular index) was almost always slightly greater in females than males. At 1 month of age the average was 26.6 ~ in girls (range between 20 and 35 ~) with an average of 24.6 ~ in boys (18-33~ The average values decreased slowly with age. The mean in females at 12 months was 23.1 ~ and at 5 years was 20.9 ~. In males the mean at 12 months was 24.3 ~ and at 5 years was 17.8 ~ In a study by Tonnis and Brunken reported by Almby et al., the mean acetabular angle was 30.3 ~ at 2 months of age, 27.1 ~ at 4 months, 23.7 ~ at 6 months, 22.1 ~ at 8 months, 21.8 ~ at 1 year of age, 18.8 ~ at 2 years of age, 15.6 ~ at 3 years of age, and 15.5 ~ at 5 years of age (4, 312). Coleman also confirmed the normal diminution of AI with growth, even in the first few months of life, and the finding that the angle on average was 2-3 ~ higher in females than in males (45). 4. TIME OF APPEARANCE OF THE SECONDARY OSSIFICATION CENTER

The secondary ossification center of the femoral head was present in 20% of all children at 3 months, in 80% at 6 months, in 96% at 9 months, and in 100% at 12 months. Mackenzie found the secondary center to be present in 33% of normal children at 3 months of age and in 89% at 6 months (184). In radiographs taken 200 days after birth (6 months 3 weeks), the proximal femoral capital epiphysis will be present in 90%. Both bony centers normally are present in all individuals at 1 year of age. The time of appearance of the center thus varies greatly in the normal, and on occasion the time of appearance may vary in the two hips of the same child. An ossification center was found in only 1 femoral head in 7 of

218

CHAPTER 3 9 Developmental Dysplasia of the Hip

172 normal infants (4%) 1-5 months of age by Pettersson and Theander. Ossification generally occurs a few weeks earlier in girls than in boys. In a study by Pettersson and Theander (227) of 455 normal infants, the mean age at onset of ossification was 4.0 months in girls and 4.6 months in boys. Ossification did not occur before 1 month of age. The center was present in all girls by 7 months and in all boys by 9 months. A large radiographic study of 2208 hip joints in normal infants was performed by Yamamuro and Chene from 3 days to 60 months of age, with approximately half of the infants younger than 8 months of age (361). By 1 month of age only 0.2% had an ossification center, with 2% at 2 months, 20% at 3 months, 50% at 5 months, 75% at 6 months, and 100% at 10 months. In this study ossification also was earlier in girls than in boys; the nucleus was seen in 50% of girls at 4 months of age and in 50% of boys only at 6 months of age. Height and width measurements from separate studies are shown in the figures. The secondary center generally forms as a single gradually enlarging mass. It is initially punctate and successively round, oval, and hemispherical with width greater than height. The secondary center in the normal child almost always appears simultaneously on both sides. In the study by Yamamuro and Chene, unilateral appearance at the initial stage was seen in only 7 of 2208 radiographs (361). Delayed appearance occurs in dislocated femoral heads, in markedly subluxated femoral heads, or in heads suffering from AVN with treatment. (It also frequently is delayed in many of the skeletal dysplasias.) In the study by Pettersson and Theander the secondary ossification center in the dislocated hip showed a marked tendency to appear later than on the contralateral normal side, to be smaller than that of the normal side, and, after treatment, to be more irregular in appearance than that on the normal side (228). Thus, it is evident that dislocation by itself retards development of the secondary ossification center and that treatment also can slightly retard local development. In many cases of DDH with treatment the center forms eccentrically. Rungee and Reinker studied 17 cases of DDH having arthrograms to assess the clinical significance of ossific nucleus eccentricity (259). Eccentric positioning was noted in 11 of 17, but there was no relation to the cartilaginous femoral head sphericity. It was also seen in 20% of controls. Although observation for it is warranted, it is not indicative of problems by itself; it is seen from normal controis but may also be indicative of vascularization problems. Bertol et al. noted that the average age of appearance of the secondary center on the normal side of a unilateral CDH was 4 months compared to 5 months on the opposite dislocated side (19). The use of relatively complex calculations still enables "three-dimensional" information to be calculated from plain anteroposterior radiographs (113). Figures 9D and 9E show plain radiographic findings in cases of CDH-DDH.

B. Arthrography in Assessing Hip Position and Anatomy Arthrography remains a valuable technique in assessing hip disorders in the young due to the fact that it can outline the shape and position of the cartilaginous femoral head and in particular demonstrate its relationship to adjacent soft tissues such as the acetabular cartilage, labrum, and capsule. In addition, it allows for a dynamic study in which the stability of the femoral head can be assessed in several positions in relation to the acetabulum. Although the procedure is invasive and almost always is performed under general anesthesia, in many instances it is done at the time of hip reduction and thus provides useful information concerning the adequacy of closed reduction and the possible need for open reduction. Arthrography has been in use for almost 90 years, but it was the work of Leveuf (175), Severin (277, 278), and Wiberg (351) that placed it into a meaningful clinical framework particularly for CDH assessments. Severin indicated that the first detailed arthrogram was described in 1908 by injecting air into a postmortem specimen, giving the socalled air arthrogram (278). Broach and Goldhamer injected a solution of potassium iodide and were able to identify both the limbus and the zona orbicularis. The technique was used first clinically by Sievers (1927) (281) and Bronner (1927) (27) using Iodopen and air for injection, respectively. Due to imperfect contrast media, however, these authors failed to develop the technique into a useful clinical tool. Severin's work in particular was helpful in that regard, and by the end of 1940 he had collected about 250 arthrographic assessments of hip joints in 100 patients from 3 months to 8 years of age, although the vast majority were under 4 years of age. Faber also utilized the technique in the 1930s for CDH to good advantage. Crawford and Carothers, describing their technique in the immature hip, recommended a series of 8 radiographs post-dye injection with the hips in neutral, Lauenstein, abduction-internal rotation, neutral push-pull, neutral pull-push, neutral weight bearing, and true crosstable lateral positions (48). They assessed the femoral head sphericity in each view, looked for a uniformly thin layer of dye in the medial joint space, assessed the position of the head in relation to the triradiate cartilage, investigated whether as in the normal at least half of the spherical femoral head was clasped by the cartilaginous socket, and looked for the thorn-shaped limbus and the ligamentum teres, which in actual fact should not be well-visualized in a normal hip. They favored a medial approach for injection because the major pathological processes to be demonstrated are at the superior and lateral parts of the joint. Should any dye be extravasated at the time of injection, the structures obscured would not be the most important. Severin used an anterolateral approach for injection just lateral to the femoral artery and about 2 cm below Pouparrs ligament.

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

The specific anatomy of the developing hip can be clarified in both the normal and the abnormal by the arthrographic procedure (Figs. 10A- 10D). The cartilaginous surface of the femoral head is well-outlined as is the acetabulum, involving not only the bone of the acetabular roof, which is seen on plain radiographs, but also the acetabular cartilage and the terminal circumferential projection of fibrocartilage referred to as the labrum. The outer border of the labrum tapers to a fine point, which lies free on the surface of the femoral head allowing dye to be interposed between its lateral outer surface and the capsule and between its inner surface and the femoral head articular cartilage. The labrum is free except at the inferomedial region where it continues into the rounded transverse ligament. Inferiorly, the capsule is attached around the base of the neck of the femur. Slightly lateral to the acetabular edge, a thick ring of ligament, the zona orbicularis, passes around the acetabulum outside the joint capsule and can be seen indirectly by the arthrogram as the joint capsule is narrowed showing relatively little dye at its region and a clear-cut notch in the capsule just below the inferomedial border of the acetabulum with outpouching of the capsule both above and below. The ligamentum teres attaches to the bottom of the acetabular fossa and then to the femoral head at its anteromedial central portion in an area referred to as the fovea capitis. Severin identifies the normal appearance of the hip arthrogram in a developing child (Fig. 10B) and those characteristic features seen with dislocation (Fig. 10C). Those features most essential to assessment of a normal hip involve (1) the sphericity of the femoral head cartilage and its relationship to the acetabulum, (2) the appropriate positioning and shape of the cartilage of the acetabulum and its fibrocartilaginous terminal and peripheral rim of tissue referred to as the labrum, and (3) the passage of dye onto either side of the labrum such that the labrum itself is defined clearly as lying on the superolateral aspect of the femoral head articular cartilage. One can also note, in specimens in which the appropriate amount of dye has been injected and in which there is no obliteration from a medial installation, that the transverse ligament is a depression just below the medial inferior border of the acetabulum and the zona orbicularis, which is present just above the base of the neck insertion of the capsule. Hips that are displaced are characterized by either upward positioning of the limbus or interposition of the limbus between the femoral head and acetabular cartilage, by the slightly to markedly lateral displacement of the femoral head in relation to the acetabulum, and, with considerable displacement, by the presence of the ligamentum teres passing from the deepest part of the acetabulum to its insertion onto the femoral head. In the normal, the articular surface of the femoral head must appear spherical and a little more than half of the femoral head must be present and placed concentrically within the cartilage acetabulum. Mitchell was one of the earlier practitioners of the arthrographic technique for CDH, describing the use of the proce-

219

dure in 200 instances (202). Arthrography was done on all displaced hips treated at his institution from 1954 to 1961. He favored injection via the superior and anterior approach. He placed major importance on the assessment of the labrum, whether it was positioned above the femoral head or inverted, and if on the outer margin of the head whether there was subluxation of the head (Fig. 10A). He divided the hips into those with primary instability, although there was location of the head in the acetabulum, partial displacement in which the labrum was positioned normally but displaced upward, and complete displacement in which the dislocation could be subdivided into tight or loose variants. In the loose dislocation, the femoral head is markedly displaced and a diagnosis of dislocation with interposition of soft tissue is never in doubt. In those hips with primary instability with the hips lying relaxed in lateral rotation, the labrum is not well-outlined because of the tendency of the femoral head to push up against the labrum and not allow the dye to interpose between labrum and capsule. When the hips are abducted, the femoral heads pass more deeply into the acetabulum and the outline of the labrum is seen clearly. With partial displacement the same situation applies except that the femoral head tides upward and outward to a greater extent, pushing the labrum with it, although the labrum itself is not interposed. Once again, when the femur is abducted, the head slides back into the socket and the labrum and its rose thorn configuration is revealed. With complete displacement, the labrum is interposed between the head and the acetabular cartilage and in no position is the rose thorn configuration seen. Because arthrography was used in many centers through the better part of three decades in large numbers of cases, many reports have followed. The statements by Somerville and Scott from 1957, however, remain valid today: "In all arthrographs, there are many shadows to be seen and there is consequently a tendency to read too much into them. There are 2 principal features to be looked for: the position and shape of the limbus and pooling of the contrast medium on the floor of the acetabulum" (177). Liu et al., in a review of 102 arthrograms, assessed 35 in detail and concluded that the medial pooling and morphology of the acetabular limbus were the most helpful diagnostic criteria (177). Forlin et al. in an exhaustive review of their cases, defined eight types of hip morphology from normal to dislocated (80). Most of the important structures were best observed on a neutral view, with frog lateral and abduction internal-rotation views being the most useful in determining the depth of reduction and appropriate degree of coverage. Arthrography also was performed extensively in patients at the Newington Children's Hospital with review in relation to preliminary traction by Quinn et al. (244). They used an arthrographic system proposed by Tonnis, which grades the dislocation on the basis of the position of the femoral head with the hip in neutral position in relation to the four quadrants defined by Hilgenreiner's and Perkins' lines. These def-

220

CHAPTER 3 ~

Developmental Dysplasia of the Hip

F I G U R E 10 Hip arthrograms continue to provide much useful information in the assessment of nonosseous tissue both prior to and after treatment of CDH-DDH. (A) Bilateral hip arthrogram shows a normally positioned femoral head on the right and a completely dislocated femoral head on the left. On the right the acetabular index is normal. The arrow laterally points to the normal appearance of the fibrocartilaginous labrum as indicated by the small linear line of dye, which outlines the triangular fibrocartilaginous labrum. The

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

initions are similar to the less involved scheme of Mitchell, with grade 1 showing the femoral head to be displaced laterally by no more than two-thirds of its width relative to the superior rim of the acetabulum with the labrum everted and still covering the femoral head and progressing to grade 4b in which the femoral head is dislocated completely laterally and superiorly and the labrum is inverted into the acetabulum, which would obstruct any attempt at closed reduction. A study by Ishii et al. in 42 patients with CDH demonstrated good correlation between the arthrograms and the operative findings of a large ligamentum teres, hourglass constriction of the capsule, and a prominent transverse acetabular ligament (132). The number of hips, however, found at operation to have an inverted labrum was smaller than had been predicted from the arthrogram. Tanaka et al. performed arthrography is 228 hips, using the technique to assess concentricity of reduction and the nature and type of soft tissue interposition (302). Findings were graded into types 1, 2, 3, 4a, and 4b and decisions on whether to intervene surgically were dependent on findings, although only type 4b cases generally underwent operative procedures

C. Ultrasonography in the Diagnosis of Newborn Developmental Dysplasia of the Hip Ultrasonography since the early 1980s has developed into an extremely useful technique for the assessment of developmental dysplasia of the hip particularly in the newborn period and early months of life. In most centers it has completely replaced conventional plain radiography in the first 6-9 months of life (270, 315). Its value diminishes with increasing size of the secondary ossification center. It is particularly valuable in the developing hip in which early diagnosis is extremely beneficial to the long-term results of treatment, and plain

221

radiographs are of limited usefulness because bone does not begin to appear in the secondary ossification center of the femoral head until 3-4 months of age and the bony portions of the acetabulum are relatively small in relation to the cartilaginous and fibrocartilaginous parts. Graf of Austria has been particularly instrumental in recognizing the value of ultrasonography in the assessment of the neonatal hip and in developing the technique into a very useful clinical tool in relation to hip dysplasia (93-95, 315). Ultrasonography allows for detailed examination of the infant hip in a noninvasive fashion and without radiation exposure. It can provide imaging of the cartilaginous structures and can differentiate them both from bone, fibrocartilage (which is the primary constituent of the labrum), capsular tissue, and fat. Because the soft radiolucent tissues are the crucial determinants of the status of a developing hip, particularly in the first 3-4 months of life, the technique has proven itself to be invaluable and is now essential for hip assessment. Nichols et al. showed excellent correlation between two newborn cadaver hips assessed by ultrasound followed by anatomic dissection (212). In addition, it has allowed for a dynamic appreciation of the stability of the hip because ultrasonography can be performed when the hip is clinically manipulated. Figure 11 illustrates the value of the ultrasound in comparison with a normal plain radiograph, an arthrogram, and histologic sections of the developing hip. The hyaline cartilage of the femoral head is homogeneous in structure and thus gives a poor or almost no echo. The radiolucent soft tissue roof supporting the head of the femur is particularly well-demonstrated by ultrasonography; by this we refer to the hyaline cartilage of the acetabulum, the fibrocartilaginous labrum, and the capsule, each of which provides support to the head. These structures also can be defined by arthrograms, although this procedure requires an invasive technique and

F I G U R E 10 (continued) dye laterally has tracked into the normal recess between the outer aspect of the labrum and the hip capsule, which inserts above the labrum onto the side wall of the acetabulum. This is an essential component of a hip arthrogram, indicating that the labrum is not only on top of the head but that its relationship to the capsule also is normal. The obviously dislocated head on the opposite side is evident as is the intracapsular communication (white dye) between the displaced head and its redundant capsule and the small persisting dysplastic acetabulum. (B) The normal hip arthrogram is illustrated. Correlation with Figs. 1A, 1B, and 1D is helpful. The dye is shown as the dense black accumulation around the femoral head. C indicates the cartilaginous acetabulum and its peripheral fibrocartilaginous labrum. A refers to the indentation created by the orbicular ligament. This is tightest posteriorly and serves as the inferior margin of the ischiofemoral ligament. B indicates the inferior protuberant margin of the synovium just beneath the orbicular ligament. This synovial protrusion can also be seen in the gross anatomic drawing of the hip region in Fig. 1B. The arrow with the circle at its base indicates the transverse acetabular ligament. There also are synovial outpouchings immediately medial and inferolateral to this ligament. Note also at the superolateral aspect of the joint that the dye passes between the outer margin of the distal tip of the labrum and the adjacent capsule. Some refer to this as the thistle sign. (C) This illustrates a normal right hip arthrogram and an abnormal hip arthrogram with a dislocated left hip. Note the normal position of the fibrocartilaginous labrum at the superolateral margin of the femoral head on the fight and its inverted posture on the left. The characteristic thistle sign is seen in relation to the right (arrow), whereas the inverted labrum with the absent thistle sign is seen at left (arrow with square box). The elongated ligamentum teres is shown at left (H). (Di) A normal hip arthrogram is shown at left. The labrum is intact and the thistle sign is normal with tracking of dye lateral and superior to the fibrocartilaginous labrum. The orbicular ligament is seen clearly where dye penetration is lacking and the synovium is outlined above and below it. The indentation of the transverse acetabular ligament is normal as are the synovial outpouchings just above and below it. The head is deeply seated in the acetabulum. (Dii) The femoral head is completely dislocated from the acetabulum. The dye communicates from the enlarged capsule through the capsular isthmus into the small acetabulum. Inversion of the labrum is indicated by the absence of dye between the articular surface of the femoral head and that of the acetabulum. [Part A reprinted from (202), with permission. Parts B and C reprinted from (278), with permission.]

222

CHAPTER 3 9

DevelopmentalDysplasia of t h e Hip

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

some form of anesthetic in infants and children. The acetabular cartilage also is homogeneous and thus quiet in a sonar sense. The ultrasound examination of the hip enables one to determine the position of the cartilaginous femoral head in relation to the acetabulum, the state of development of the depth of the acetabulum, the earliest stages of appearance of the secondary ossification center, and, in a dynamic sense, the stability of the hip when it is manipulated in various positions. The value of the technique in the infant hip is due to the fact that the ultrasonic wave either does not penetrate or penetrates only slightly into bone, whereas such waves penetrate quite well into muscle, connective tissue, hyaline cartilage, and fibrocartilage. In ultrasound studies with appropriate resolution, Graf reports the demonstration of an extensive number of anatomic features (315). Although the large majority of these are not particularly essential to a clinical assessment, they show the extreme value of the technique in outlining the underlying anatomy. Among the structures that can be identified readily are the cartilaginous head of the femur, the cartilaginous greater trochanter, the bone of the femoral neck, the bone-cartilage junction underlying the single physis at the developing end of the femur, and the early formation of the secondary ossification center of the femoral head. In the acetabulum, the components identifiable are the lateral bony convexity of the acetabulum, the cartilaginous acetabulum, the fibrocartilaginous labrum, the transverse ligament of the acetabulum, the triradiate Y-cartilage, the joint capsule, the overlying muscle layers, and particularly, the iliopsoas muscle. The hyaline cartilage of the acetabulum and femoral head in particular is anechoic or nonechogenic due to its unique and uniform structure involving a homogeneous matrix with a high water content and relatively few cells. Indeed, the cartilage is outlined only indirectly by the high-echo structure of the surrounding tissues. There is no specific delineation of the cartilaginous femoral head where it lies against the acetabular cartilage in the depth of the acetabulum because both cartilage tissues are nonechogenic. Farther toward the medial side of the acetabulum, however, in the area of the acetabular fossa, there is good differentiation of the femoral head from the fat and fibrous tissue in the

223

depths of the acetabular fossa. The lateral margins of the acetabulum are differentiated in the region in which the hyaline cartilage ends and the fibrocartilaginous labrum begins. In a 1984 report, Graf indicated that 3500 infant hips had been investigated sonographically in his unit (95). Development of the bony acetabulum was a particularly sensitive indicator in relation to the adjacent acetabular cartilage and the position of the head. This observation has been made for several decades, and the acetabular dysplasia measured by an increased acetabular index on plain radiographs has long been used to indicate imperfect seating of the femoral head and the development of a dysplastic hip. Graf derived two reference lines and two specific angles when the image was the equivalent of an anteroposterior assessment by radiographic projections. The basic measuring line, which was involved in assessment of the two angles, runs along the outer bone of the ilium. The second line drawn is referred to as the inclination line and runs from the most prominent convex point of the bony acetabulum (or the most lateral part of the bony acetabulum) along the inner aspect of the lateral most articular cartilage and down through the tip of the labrum. The angle between these two lines is referred to as the [3 angle or inclination angle and characterizes the extent of the bone supplementing the lateral most bony roof of the acetabulum. The third line is referred to as the line of the acetabular roof and it passes from the lower edge of the ilium bone medially to the top of the osseous convexity laterally. This line, therefore, essentially is a reproduction of that drawn to define the acetabular index in a plain radiograph. The et angle characterizes the extent of the obliquity of the acetabulum or in a sense the osseous convexity in the terminology of Graf. The greater the angle, the deeper and more normally developed the bony acetabulum; and conversely, the smaller the angle, the greater the degree of acetabular dysplasia. On the basis of his numerous studies, Graf established the ot angle of 60 ~ or greater as fully normal and any angle less than 43 ~ as indicative of a high degree of dysplasia. Graf subsequently defined four types of hip in the sonographic classification (Figs. 12A-12D). Type 1 is normal. Type 2 shows a dysplasia of the acetabulum. Many of these

FIGURE 11 Correlationbetween histologic, plain radiographic, ultrasonographic, and arthrographic findings is shown. (A) Histologic appearance of the newborn hip is seen. The acetabular bone and cartilage are seen as is the fibrocartilaginous labrum, the capsule, and the lateral recess between the two. (B) The hip structure is outlined clearly by the artist's drawing. (1) Acetabular bone, (2) acetabular cartilage, (3) fibrocartilaginous labrum, (4) capsule, and (5) lateral recess between the acetabular cartilage and labrum and the capsular attachment. Components 2, 3, and 4 of the hip roof (acetabular cartilage, fibrocartilaginous labrum, and capsule) are radiolucent. (C) Plain hip radiograph showing the bony acetabulum and the developing secondary ossification center with other structures all radiolucent. (D) Structures that can be deciphered on a well-performedhip ultrasound: (1) secondary ossification center; (2) fibrocartilaginous labrum; (3) lateral acetabular cartilage; (4) acetabular bone; (5) junction of hip capsule and outer margin of iliac bone; (6) hip joint capsule; (7) cartilaginous femoral head; (8) hip abductor muscles; and (9) lateral most extent of cartilaginous femoral head. (E) A normal hip arthrogram provides information similar to the ultrasound but requires an invasive procedure. Note the dye passing on the outer side of the labrum (arrow) between the labrum and the capsular insertion at the junction of the labrum and the acetabular cartilage. This appearance outlining the two borders of the labrum by dye is referred to as the thistle sign. (F) Normal hip ultrasound with secondary ossification center present. [Parts B and D reprinted from Graf, R. (1983). J. Pediatr. Orthop. 3:354-359, 9 LippincottWilliams & Wilkins, with permission.]

FIGURE 12 The four-stage ultrasonographic classification of hip dysplasia of Graf is shown in this figure. Part (A) shows type 1 and the A and B variants; (B), type 2; (C), type 3; and (D), type 4. Note the progressive diminution of the oLangle with increasing grade of hip abnormality. [Parts A - D reprinted from Tonnis, D. (1987). "Congenital Dysplasia and Dislocation of the Hip," Figs. 14.24a, 14.25a, 14.26a, and 14.28a (pp. 198-202), 9 Springer Verlag, with permission.] (E) Summary of interpretation of oLvalues (derived from references 66, 94, 95, 315, 321 ).

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

hips are only immature and will continue to develop in a normal fashion. In their early studies including infants up to 3 months of age, in 90% of type 2 hip joints there was no subsequent pathology demonstrable. Graf considers the acetabulum to be premature in terms of the development of depth of its roof and makes the point that it is analogous to the delayed appearance of the secondary center of the femoral head at only the 3- or 4-month period. He thus subdivides the type 2 categorization into type 2a which is a type 2 finding prior to the third month of life, and type 2b, which refers to the type 2 categorization persisting after the third month. In the type 2b case, there is delayed ossification and the likelihood is higher that a true dysplasia is developing. In the type 3 hip, the femoral head has begun to sublux laterally, although it is almost always contained partially within the acetabulum. The nonosseous structures of the acetabular roof yield to the increasing pressure and, although the labrum is in its normal position, it is pushed upward and outward and the bone of the acetabulum is delayed in its development such that the lateral osseous convexity is not as sharp but is more rounded and less well-defined laterally. In this situation, the [3 angle will increase and the a angle will decrease. In the subtype 3a, the acetabular cartilage lateral to the ilium bone is pushed upward, although it remains structurally intact and low echo in nature. In subtype 3b, lateral subluxation is more marked and thus there is increased pressure on the acetabular cartilage such that it begins to be transformed histologically as well. The sonographically low-echo structure vanishes as the cartilage takes on a more fibrocartilaginous constitution and thus the cartilaginous convexity becomes somewhat echo-dense. In the type 4 categorization, there is complete dislocation of the femoral head, which is visible in the soft tissue, while the acetabulum is empty. Figure 12E shows the range of values in types I-IV. Graf defines the usefulness of the sonogram to be concentrated in the period from the newborn era to the 10th month of life. Following 10 months of age, the bony tissues are sufficiently prominent that the ultrasound wave has a small area of penetration and is less able to define the relevant structures and positions. The sonogram has clear treatment implications in relation to the position and the underlying pathoanatomy. The type 1 hip is normal. The type 2a hip under 3 months of age is still physiologic, whereas the type 2b hip older than 3 months is indicative of a developing dysplastic situation. Repeat assessments are needed to determine whether there is any deformation of the acetabular cartilage convexity indicative of further subluxation or impending dislocation. When subluxation has occurred laterally, the cartilaginous acetabular roof is under increased pressure, which diminishes its development, and this is also reflected in the underdevelopment of the acetabular roof bone laterally. The echoic nature of the cartilaginous roof will determine whether changes are just beginning such as in type 3a or have already lead to fibrocartilaginous tissue transformation in the 3b category.

225

Several clinical examples of ultrasound findings in relation to treatment are shown in Fig. 13. Exner performed ultrasonography on 615 newborn children to determine standardization and findings (66). A type 1 hip, which is defined as normal or mature, has an et angle equal to or greater than 60 ~ Both hips were type 1 in 90% of the boys and 80% of the girls. Type 2a (physiologically immature hip) with an t~ angle equal to or less than 59 ~ (a less developed hip) was found in 9% of the boys and 17% of the girls. Type 2b, defined as a critical hip with the oL angle equal to or less than 49 ~ was found in 1% of the boys and 2% of the girls. Type 3 is a severely dysplastic hip with the ct angle equal to or less than 43 ~ and was found in one boy and one girl, whereas the type 4 dislocated hip with a medially displaced labrum was not seen. He concluded that, after the age of 3 months, all children who had not developed a type 1 hip could be considered to have some form of dysplasia requiting further investigation and treatment. Children less than 3 months of age frequently are seen with a less well-developed acetabulum and a smaller oLangle who spontaneously developed type 1 hips and are referred to as physiologically immature in relation to hip development. After 3 months of age, Exner felt that hips with an et angle of less than 50 ~ are dysplastic and need treatment, whereas those between 50 and 55 ~ are followed until they reach type 1 maturity. A study by Szoke et al. confirmed the findings of Graf and showed the prognostic value of the classification (301). The majority of newborn hips were Graf type 2a with physiological immaturity. Type 2b, 3, and 4 hips all required therapy. Falliner and Hassenpflug showed that routine ultrasound did not appear to increase the number of clinicallysignificant dysplastic hips needing treatment, but it did contribute to earlier diagnosis and the need for fewer operative reductions (71). Hangen et al. showed the value of assessing the effectiveness of Pavlik harness treatment by ultrasonography while the harness was on (107). The technique greatly minimized the number of radiographs and gave an earlier indication of treatment failure or success. Terjesen et al. also estimated the percent coverage of the femoral head by the bony acetabulum in the ultrasound studies (306). This measurement has been used increasingly particularly in North America to determine the depth of the acetabulum in relation to the femoral head. The measurement is referred to as the bony rim percentage (BRP) and the lower normal limit in the initial study was approximately 50% in hip joints from 156 children from 2 months to 2 years of age. The bony rim percentage (BRP) was calculated as a ~ x 100. At a mean age of 4 months, the mean B RP was 64 and at 8 months it was 62. The range of values was such that the authors felt that any measurement below 50 was abnormal, indicative of imperfect coverage. This measurement can be made up to approximately 8-10 months of age

F I G U R E 13 Several examples of ultrasound findings in clinical cases are reviewed. (A) Examples of normal hip ultrasounds are shown from a (Ai) newborn, (Aii) 2-week-old, (Aiii) 2-month-old, and (Aiv) 4-month-old. At 4 months of age, the secondary ossification center is seen clearly. The images are taken from different patients. The et angle in (1) is 50 ~ (2) 52 ~ (3) 50 ~ and (4) 62 ~ (B) A series of ultrasound studies show normal development following diagnosis of bilateral dislocatable hips at birth in an otherwise normal child. Treatment began with continual Pavlik harness use beginning at 6 days of age when the first ultrasound study was performed. (Bi) The right hip is located but the acetabulum is shallow and only slightly less than one-third of the femoral head is covered. The et angle is 35 ~ (Bii) The left hip also reveals a shallow acetabulum with less than one-third coverage of the femoral head although the head is located. The et angle measures 38 ~ (Biii) Right hip at 22 days shows improved coverage of the femoral head with the ct angle increased to 48 ~ (Biv) Left hip shows the et angle increased to 50 ~ (By) Right hip ultrasound at 1.5 months shows increased development with the et angle now at 55 ~ and initial and relatively early appearance of the secondary ossification center. There is still only approximately 40% coverage of the head, however. (Bvi) The left hip at 1.5 months shows improved coverage with the ct angle at 56 ~ and early appearance of the secondary ossification center. (Bvii) The right hip at 3 months shows excellent maturation with the ct angle at 58 ~

F I G U R E 13 (continued) enlarged secondary ossification center, and slightly greater than 50% coverage of the femoral head. (Bviii) The left hip at 3 months shows the oLangle at 60 ~ increased secondary ossification center, and a slightly greater than 50% coverage of the femoral head. (Bix) Anteroposterior radiograph at 3 months of age shows both femoral heads to be well-located in the acetabulum

228

CHAPTER

3 ~

Developmental Dysplasia o]r the Hip

F I G U R E 13 (continued) in the frog lateral position. The secondary ossification centers are both present and equal in size. Acetabular development is good. Weaning from the Pavlik harness began at this stage. The acetabular index is 30 ~ on the right and 26 ~ on the left. (Ci) A dislocatable right hip was diagnosed at 3 months of age. The ultrasound on the right shows an extremely shallow acetabulum with an oL angle of only 30 ~ no coverage of the femoral head, and a markedly laterally displaced secondary ossification center.

SECTION IX ~ Imaging Techniques Used to Assess Hip Position

229

F I G U R E 13 (continued) (Cii) The normal left hip shows an oLangle of 52 ~ and slightly greater than 50% coverage of the femoral head based on the appearance and location of the secondary ossification center (D) A boy with bilateral dislocatable hips at birth was assessed. Bilateral hip ultrasound at 10 days showed both heads to be located. The acetabulae were shallow, however, based on the ot angles and the Pavlik harness treatment was performed. (Di) Right hip sonogram shows the femoral head to be located but in a very shallow acetabulum with one-third femoral head coverage and the oLangle 40 ~ (Dii) The left hip at 10 days shows an extremely shallow acetabulum with the c~ angle at 38 ~ and femoral head coverage only one-fourth of its full extent. (Dill) Right hip sonogram at 2 months of age shows markedly improved coverage to 50%, early development of the secondary ossification center centrally, and the a angle increased to 48 ~ (Div) At 2 months shows almost 50% coverage of the femoral head, early development centrally of the secondary ossification center, and deepening of the acetabulum with the ot angle increased to 54~ Treatment with the Pavlik harness was continued until 3 months of age when weaning began. (Dv) Anteroposterior X ray at 4 months of age shows early development of both secondary ossification centers with excellent positioning and improved development of both acetabulae. (Dvi) Anteroposterior radiograph at 16 months of age shows normal hip development. Both secondary ossification centers are well-developed and well-positioned. The acetabular index is 28~ bilaterally. Note the uniform chondral bone formation of both acetabulae. (E) Male patient with bilateral developmental dysplasia of the hips. Both hips were dislocatable at birth and were shown to be dislocated by ultrasonogram at 3 days of age. (Ei) Right hip sonogram shows the femoral head to be widely dislocated. The oLangle, however, is relatively well-developed at 50~ (Eli) The left hip sonogram also shows complete lateral dislo_eation of the femoral head, with the c~ angle at 44 ~ The patient was placed in a Pavlik harness and repeat sonogram at 8 days of age showed relocation of the right femoral head but persisting wide dislocation of the left. (Eiii) Right hip with the femoral head located and acetabular development allowing for an ot angle of 40 ~ (Eiv) The left femoral head remains dislocated in the Pavlik harness. The oLangle is 40 ~ The patient was allowed to proceed to full right hip stabilization in the Pavlik harness. (Ev) Anteroposterior X ray at 4 months of age shows a relocated right femoral head with excellent acetabular development compared to the opposite side. The left hip remains dislocated and acetabular development is extremely oblique and shallow. (Evi) Anteroposterior radiograph at 4.5 months of age shows excellent position of the right hip on the frog lateral and a dislocation on the left. At 4.5 months of age the patient had a closed reduction of the left hip in association with a percutaneous adductor tenotomy. The patient was immobilized in a bilateral hip spica. (Evil) CT scan immediately following cast change at 6 months showed excellent relocation. The secondary ossification center was evident on the previously relocated right hip, whereas that on the left had not yet appeared. Hip spica immobilization was continued to 11 months of age at which time a night only cast use was continued for an additional 2 months. Anteroposterior radiographs at 8 months (Eviii) showed improvement on the left. (Eix) Note the appearance of the secondary ossification center in the 11-month film. The hip remained stable following discontinuation of the immobilization at 13 months. (Ex) Radiograph at 1 year 3 months of age shows improved development on the left. (Exi and Exii) Excellent development continued with normal hip radiographs seen at 5 years of age in both AP (Exi) and frog lateral projections (Exii).

due to the r e l a t i v e i n c r e a s e in b o n e d e n s i t y after that. T h e r e

T e r j e s e n et al. also r e v i e w e d their e x p e r i e n c e w i t h ul-

w a s e x c e l l e n t c o r r e l a t i o n b e t w e e n r a d i o g r a p h i c and ultra-

t r a s o u n d in c o n g e n i t a l hip d y s p l a s i a in c h i l d r e n o l d e r than

s o u n d findings, w h i c h s u p p o r t e d the i n c r e a s e d use o f ultras o u n d p a r t i c u l a r l y in the first f e w m o n t h s o f life. A l t h o u g h there w a s r e l a t i v e l y w i d e a c c e p t a n c e o f the use o f the e~ a n g l e

age group, a l t h o u g h the r e l a t i v e l y large size o f the s e c o n d a r y ossification c e n t e r o f the f e m o r a l h e a d limits the assess-

as defined b y Graf, the r e f e r e n c e p o i n t s w e r e not a l w a y s c l e a r and s o m e a u t h o r s h a d f o u n d the m e a s u r e m e n t difficult

2 years o f age (305). U s e f u l data can be o b t a i n e d e v e n in this

ments. In addition, plain r a d i o g r a p h y after 2 years o f age is m o s t helpful.

or impractical. T h e [3 a n g l e is n o t u s e d w i d e l y as a clinical

M e l z e r a n d W u l k e r p o i n t e d out s o m e p o t e n t i a l errors in

tool in m o s t N o r t h A m e r i c a n centers. T e r j e s e n et al. s t r o n g l y

hip j o i n t s o n o g r a p h y (195). T h e s e i n c l u d e d u n s a t i s f a c t o r y

s u p p o r t e d the v a l u e o f the h e a d c o v e r a g e index, w h i c h t h e y

i m a g e quality, n e g l e c t e d s t a n d a r d i z e d p l a n e s o f e x a m i n a t i o n ,

r e f e r r e d to as the b o n y r i m p e r c e n t a g e as that w a s i n d i c a t i v e

i n a d e q u a t e c o n s i d e r a t i o n o f f o r m variants o f the b o n y acetab-

o f the f e m o r a l h e a d c o v e r a g e b y the b o n y r o o f o f the acetab-

ulum, e r r o n e o u s p l a c e m e n t o f m e a s u r i n g lines, u n c e r t a i n t y

ulum. A similar i n d e x w a s d e v e l o p e d b y M o r i n et al., alt h o u g h slightly different l a n d m a r k s w e r e used.

as to the r a n g e o f n o r m a l values for the e~ angle, insufficient f o l l o w - u p after the initial e x a m i n a t i o n , and p e r f o r m a n c e

230

CHAPTER 3 ~ Developmental Dysplasia o f the Hip

of only single-view studies. Of particular importance is the reproducibility of oL and 13 angle determination. In the view of some the particular value of the ultrasound is the morphologic appearance of the femoral head in relation to the acetabulum and the adjacent soft tissue structures rather than excessive reference to the numerical values. Of great value as well is the real-time finding, indicating the stability of the hip to manipulation. Shuler et al. reported their extensive experience, which confirmed most of the points previously described. They also commented on the great importance of equipment with high resolution and observer experience. Not all observers recommend routine ultrasonography for all newborns. Castelein et al. performed a correlative study of 614 newborn hips assessed by clinical examination and ultrasonography within 48 hr of birth (35). The ultrasound was considered normal when the oLangle was 60 ~ or greater and no instability was detected in the frontal or transverse planes. Of the 614 hips, 82 (13.4%) had ultrasound abnormalities despite a normal clinical examination. However, 79 of these improved spontaneously without treatment to a normal clinical-sonographic range within 12 weeks. Only 3 hips (0.5%) developed definite hip dysplasia, 0.5% of the entire group and 3.7% of the sonographically abnormal group. A longer term study from the same center showed the same proportional changes, with 4 dysplastic hips eventually persisting after 100 sonograms in clinically normal hips (36). Clarke showed complete 100% correlation between a positive Barlow or Ortolani test and sonographic abnormalities (44). On several occasions bilateral hip abnormality was found sonographically when only unilateral involvement was detected clinically.

D. Long-Term Studies of Sonographic Indices in Normal and Abnormal Hip Development As experience and data increase in relation to ultrasound studies of hip development, the normal developmental values are becoming better defined as are those associated with dysplastic findings. Tschauner et al. have developed tables assessing the maturation curve of the oLangle within the first year of life (325). These values better define the spontaneous maturation of the physiologically immature hip joints, which refer to the 2a categorization in those less than 3 months of age. The mean value of the o~ angle crosses the 60 ~ border line at about the age of 2 months and reaches the 64 ~ mature level at about 4 months. The 64 ~ level is more or less unchanged until the end of first year of life. Wagner et al. have utilized ultrasound to assess fetal hip development (333). Their study proceeded from the 14th to the 40th week of gestation. From the 20th week of gestation, they were able to note the bony structures and begin their assessment of the acetabular configuration. The soft tissue and cartilaginous acetabular components were recognizable from the 21 st week.

Harcke and colleagues have developed extensive experience in North America in relation to the ultrasound technique for hip dysplasia (109, 208). Many along with Harcke favor the dynamic assessment of the hip in multiple positions modeled on the Barlow and Ortolani maneuvers used in the clinical examinations. The dynamic approach can illustrate the instability of the hip. The technique developed by Graf is referred to by some as a morphological approach because it relies to a greater extent on static measurements and classifies the hip of an infant into four types, with some subtypes, ranging from a normal joint to severe dislocation based on the degree of angular changes documented. Many of the values derived during the first decade of ultrasonography have been reviewed by Tonnis et al. in reference to the Graf system (321). With increasing fine-tuning of the documentation, Tonnis described the Graf l a, lb, 2a, 2b, 2c, 2d, 3a, 3b, and 4 categorizations. A detailed classification chart deriving from Graf and also utilizing the work of Tonnis and others is reproduced. In their review of several studies, the results of newborn screening between 1985 and 1987 are presented. This review encompasses 5174 hips: la, 1.2%; lb, 66.1%; 2a, 30%; 2b, 0; 2c, 1.2%, 2d, 0.9%, 3a, 0.5%; 4, 0.04%.

E. CT Scan to Assess Hip Structure Computerized tomography (CT) became an extremely valuable technique for the assessment of developmental hip dysplasia almost immediately upon its availability for clinical use. Each of the coronal, transverse axial, and sagittal orientations can play a valuable role in assessing hip structure. Single images in different planes are quite instructive, and three-dimensional reconstructions are invaluable in assessing overall hip structure (10, 75, 155, 168). CT scanning is used in two major ways in relation to developmental dysplasia of the hip. The first involves postreduction assessments in hip spica casts to define the position of the femoral head in relation to the acetabulum (226, 283, 311). The transverse axial plane serves to show the relationship of the head to the acetabulum with the patient in hip spica (Fig. 14). If the secondary ossification center is present the analysis is quite simple, and even prior to its appearance the relationship of the neck to the triradiate cartilage is determined readily. Plain radiographs in the anteroposterior projection through a hip spica have a high degree of inaccuracy, for example, when the hip is posteriorly dislocated exclusively, and the use of the CT scan postreduction now is virtually mandatory to assess position in those centers in which it is available. Smith et al. have shown the value of assessing femoral head position postreduction by CT scanning (283). They also showed the ability of the CT study to quantify the amount of abduction in the spica cast and found a much higher incidence of AVN of the femoral head with abduction greater than 55 ~ The second use of the CT technique prior to osteotomy is in preoperative assessment of acetabular or femoral deformity and the relation of the femoral head to the acetab-

SECTION IX 9 Imaging Techniques Used to Assess Hip Position

FIGURE 14 CT scans clearly demonstratethe positionof the proximal femur in relationto the acetabulumfollowingclosedreductionand hip spica immobilization. The study is greatly superior to plain radiographs and accurately defines the position of the proximalfemur in relation to the acetabulum. They are particularly accurate when both secondary ossification centers are present but still are of great value even when they are not.

ulum. Three-dimensional reconstruction is used in particular for preoperative planning. It is particularly helpful in determining anterior and posterior head coverage. Examples of CT assessments of the hip in developmental dysplasia are shown here and were discussed in more detail in Chapter 2. Normal acetabular angles in the axial plane were defined by multiple CT scans in children from 6 months to 17 years of age (345). The two most valuable indices are the axial acetabular index and the acetabular anteversion. CT scans of 170 normal hips were done. The axial acetabular index diminished progressively with development, showing increasing depth of the bony acetabulum. The index at 6 and 12 months was 131.4 ~ decreasing to 119 ~ at 4 years, 119 ~ at 9 years, and 93 ~ at skeletal maturity. Acetabular anteversion on the other hand was relatively linear and unchanged with development, being 12.4 ~ at 6 months and 13.5 ~ at maturity. Bony acetabular development is more prominent and occurs earlier in the posterior than in the anterior acetabulum. At no time was more than 50% of the femoral head covered anteriorly, whereas about 50% of the head is covered posteriorly by 11-12 years. The triradiate cartilage fuses by 11-13 years and is slightly earlier in girls, at which time the spherical configuration of the acetabulum has been reached (345). Jacquemier et al. noted similar findings for acetabular anteversion in children (135). They performed 143 CT images on children from 1 to 15 years of age, none of whom showed any hip pathology. The mean anteversion value was 12.78 ~ and it remained constant during growth from 1 to 15 years of age. There were only four measurements done in newborns and they showed a mean acetabular anteversion angle of 10.5 ~. The group thus postulated that from birth there is a short increase in the acetabular anteversion angle that stabilizes at 1 year of age and remains constant until age 15 years. Adult values from the studies of Reikeras et al.

231

gave a mean CT value of 17 ___ 6 ~ (250). The authors felt that the slight increase in angle after 15 years of age was not a growth phenomenon changing orientation, but rather was due to ossification of the peripheral acetabular cartilage to its complete extent. Other adult values were 16.5 ~ by Visser et al. (330) and 15.5 ~ by Terver et al. (307). Murphy et al. utilized CT scanning to define the threedimensional geometry of the hip joint in both normal and dysplastic situations (210). They analyzed 49 normal hip joints and 20 dysplastic hip joints. In the normal acetabulum the structure was found to be nearly a full hemisphere, which was anteverted 20 ~ and abducted 53 ~ The dysplastic acetabulum did not tend to be anterolaterally maldirected but rather globally dysplastic. There was, however, considerable variability from patient to patient, showing the advantages of CT reconstruction prior to corrective acetabular surgery. In a companion study, Millis and Murphy showed the potential value by using three-dimensional hip reconstruction to better plan acetabular and femoral redirection (200).

F. MR Imaging to Assess Position and Vascularity of the Femoral Head Postreduction Magnetic resonance imaging can also be used to great advantage in assessments of developmental dysplasia of the hip. Its main value, particularly in distinction to CT scanning, is its ability to assess soft tissue structure including vascularity. It is also able to define bony outlines at the same time. In most centers CT scanning is used to assess reduction in spica postprocedure and to perform three-dimensional reconstructions of femoral and acetabular anatomy prior to bony correction by osteotomy. MR imaging is particularly useful during the first year of life and in more complicated cases in which questions arise as to soft tissue position and vascularity in association with treatment (310). MR imaging is particularly valuable in assessing the radiolucent regions of the acetabulum both before and after closed or open reduction. It is able to assess not only the bony acetabulum but in particular the cartilaginous rim and the fibrocartilaginous labrum as well as the capsule. Greenhill et al. documented the loss of acetabular sphericity in association with the dysplastic hips (98). Suzuki utilized three-dimensional reconstructions of MR images to define the orientation of the acetabulum (296). Kashiwagi et al. also used the MR imaging technique to assess the acetabular rim in patients with DDH from 8 days to 6 months of age (143). Their assessment concentrated on the acetabular cartilage and in particular on the fibrocartilaginous labrum. They divided their patients into three groups according to the shape of the acetabular rim. The position of the femoral head in relation to the acetabulum also was readily interpreted. In group 1 the acetabular rim was sharp, in group 2 it was rounded, and in group 3 it was inverted. The position of the femoral head was classified further in terms of its type of displacement, with type A showing the femoral head displaced posteriorly

232

CHAPTER 3 ~

DevelopmentalDysplasia of the Hip

but in contact with the inner wall of the acetabulum. In type B, the femoral head was in contact only with the posterior margin of the socket with its center anterior to the edge of the acetabulum, and in type C, the femoral head was displaced outside the socket with its center posterior to the acetabular edge. Studies in our unit have demonstrated clearly the ability to assess the vascularity of the femoral head in the normal state and its diminution during treatment with particularly extreme positions of reduction and immobilization (136, 137). Assessments both in the piglet and in human patients with DDH being treated with reduction and hip spica casting have shown the value of this technique in immediately determining vascularity. This will be reviewed in Section XII on AVN.

X. ASSESSMENTS OF HIP GROWTH AND DEVELOPMENT FOLLOWING CLOSED AND OPEN TREATMENTS A. Growth and Development of the Hip Following Closed Reduction in Early Infancy Wientroub et al. studied developmental hip indices in 164 normal children from ages 3 months to 5 years (353) and in 43 normal adolescents and adults and compared the data to hip development in children who underwent closed reduction of CDH and were followed from ages 3 to 60 months and longer (352). Treatment was started for those with CDH between 3 and 6 months of age. No hips had surgical treatment. There were 44 affected hips studied. The study compared values for the acetabular angle, CE angle of Wiberg, comprehensive quotient as defined by Heyman and Hemdon, and Shenton's line. The normal mean acetabular angle in those 3 - 6 months of age was 20.9 ~. This decreased continually to the 12 ~ range by 48 months. In the normal adolescent and adult group, between 10 and 26 years of age, the mean was in the 13 ~ range. The center-edge angle increased with time from 20.87 ~ in the 3- to 6-month group to 30 ~ at 60 months and almost 36 ~ in the adult group. There was excellent correction of values in the abnormal hips successfully treated by closed reduction and cast immobilization. The mean acetabular angle in the 3- to 6-month group was markedly increased initially at almost 38 ~. With time this diminished and at the 60-month time frame had a mean of 18 ~ This was still slightly greater than the mean in the normal group at this time, which was 13 ~. The mean center-edge angle at 3 - 6 months was only 9 ~ but had increased to 25 ~ at 60 months, although this too was slightly less than the normal, which at this age graded at almost 30 ~. There was continuation of improvement after 5 years of age (60 months), with the acetabular angle in those greater than 61 months diminishing to 15~ and the CE angle increasing slightly to 26 ~ The CE angle becomes a reliable measurement in children more than 5 years of age. Shenton's line was not particularly

valuable in quantitating responses to CDH therapy. The acetabular angle was the most valuable measurement in assessing growth response to treatment of CDH. They concluded that there was a good response to closed reduction and cast immobilization for CDH in the early months of life, and the process of deepening the acetabulum and increasing the congruity of the hip continued even after the age of 5 years. Chen et al. defined those hips with unilateral DDH with poorer results as showing relatively increased center-head distance discrepancy values (39). This index on anteroposterior pelvic radiographs measures the distance from the center of the femoral head to the center line of the body drawn from the midpoint of the sacrum through the symphysis pubis to document lateral subluxation of the femoral head. A1binana et al. assessed associated pelvic changes in DDH diagnosed after 4 months of age, noting how the entire hemipelvis showed secondary changes with femoral head malposition (2). Yet another series of pelvic growth responses to the type of femoral head-acetabular relationship can be found in the acetabular teardrop conformation (140).

B. Acetabular Development Following Hip Reduction by Closed, Open, or Varus Osteotomy Treatments Harris et al. have shown the capability of the acetabulum to remodel to a normal range (acetabular index) if hip congruity is obtained before 4 years of age (111). A study of 85 hips with complete dislocations who were more than 1 year of age at the time of treatment was described. An excellent result (normal acetabulum) had an acetabular angle of 21~ or less, good was 22-24 ~, fair 25-27 ~, and poor greater than 27 ~. Normal and good acetabulae were defined as satisfactory results and fair and poor acetabulae as unsatisfactory. In the entire series there were 80% satisfactory and 20% unsatisfactory results. The percentages were maintained in those that were admitted at age 2 years or over and those admitted at age 3 years and over. Provided that congruity was obtained and maintained without deformation of the femoral head, acetabular intervention should not be necessary if the patient began treatment under the age of 4 years or if congruity was obtained in a functional position by 4.5 years of age. Others favoring the occurrence of acetabular correction with growth are Ralis and McKibbin (246) and Brougham et al. (28), who feel that such correction to a normal range will be seen following closed or open reduction up to 4 - 5 years of age. The latter demonstrated that failure to obtain concentric reduction or its loss by lateral migration of the femoral head leads to persisting acetabular dysplasia. The average age at which the acetabulum stopped showing improved development was 5 years (but with a range of 17 months to 8 years). When acetabular obliquity did not improve by a yearly change in the acetabular index of 5 ~ response was considered to be diminishing. After that age acetabular dysplasia would re-

SECTION X ~ Assessments of Hip Growth and Development

main even with concentric reduction. Harris and Brougham et al. felt that no meaningful change in acetabular angular parameters occurred after 8 years of age in the normal and that correction of dysplastic acetabulae was only minimal between 8 and 22 years of age. Lempicki et al. also favored a process of allowing remodeling to correct residual acetabular dysplasia after concentric reduction (171). Prior to 3 years of age bone surgery was not needed, there was spontaneous remodeling between 3 and 5 years of age, and only after 7 years was little remodeling expected. Massie and Howorth, in an assessment of 58 hips treated by open reduction after the age of 1 year and followed to adult life, demonstrated that adequate reduction obtained and maintained could result in a normal acetabular response up to the age of 8 years, but that in particular the results up to 3 years of age were excellent (189). Over 3 years of age, acetabular augmentation procedures were indicated progressively. Similar excellent responses of slight femoral head asymmetry occurred following appropriate reduction up to the age of 5 or 6 years in those instances in which avascular necrosis did not occur. The feeling was widespread as expressed by Massie in 1956 that "there is now sufficient evidence to state categorically that if the diagnosis is made within the first year of life and treatment is instituted promptly, a near normal hip should result consistently (190)." Kasser et al. also supported the ability of the acetabulum to respond on its own to changes in position of the femoral head up to 8 years of age (144). They performed proximal femoral varus and derotation osteotomy to treat persistent acetabular dysplasia in congenital hip dysplasia. The operation was performed in 44 hips, and results were assessed based on 3 groups according to age at operation. A rapid return to a valgus femoral neck shaft angle by remodeling was not a cause of failure in any patient. Results invariably were good in patients less than 4 years of age at the time of operation. The acetabular index corrected nicely in this group. Acetabular correction was also seen in most patients operated between 4 and 8 years of age, although 4 of 13 in this group showed persistent dysplasia. The results diminished in predictability as the patients approached the age of 8 years. In the final group there was no benefit from isolated femoral osteotomy in 10 of 11 hips in patients older than 8 years at the time of surgery. The matter of acetabular remodeling capability after closed or open reduction of the femoral head remains contentious. Rejholec and Stryhal concluded that varus osteotomy alone from age 2 years on almost always fails to lead to spontaneous full correction of acetabular dysplasia and usually is followed by recurrent valgus of the proximal femur (251). They interpreted the findings as suggestive of primary acetabular growth deficiency. Both Salter and Pemberton felt that little correction would occur with femoral head reduction alone after 18 months of age (223, 262). Schwartz followed acetabular development radiologically after closed-

233

open reduction in 50 congenitally dislocated hips for a mean of 12 years (272). In 90% of hips, maximum development of the acetabulum (acetabular angle) occurred within 2 years of reduction. He concluded that after closed reduction a potential for acetabular development existed but only for a certain period. There was a distinct difference in acetabular development in hips reduced before and after 20 months of age; this represented the critical age defining acetabular reconstitution, with those less than 20 months always showing excellent remodeling potential. A similar conclusion was reached by Almby et al. in their study of acetabular growth potential following the treatment of 27 unstable hips by closed reduction. There was an impressive reduction of the acetabular angle following successful closed reduction (4) (Fig. 9C). The younger the patient, the quicker the acetabular reconstitution. Even in the normal hips decline in the angle was greatest in the first year of life. Following reduction the acetabular angle decreased at a rate of 3.2 ~ per month (range = 2.2-5.7 ~ when treatment was started before 14 months but at a rate of only 1.1 ~ (range = 0.6-1.5 ~) when treatment was started after 14 months. They showed that treatment beginning after 14-16 months led to slower, less complete responses and that, in agreement with Schwartz and Salter, the upper limit for definite successful response of the acetabular component was in the 16- to 20month range. Cherney and Westin also studied acetabular responses following reduction of CDH by following 105 hips for an average of 8 years postreduction (40). They also noted an early acetabular response to repositioning primarily during the first year in those up to 3 years of age. The prereduction acetabular index was a more reliable predictor of the need for future acetabuloplasty than the age at which reduction occurred. When the acetabular index was 29 ~ or less, 17% eventually required acetabular surgery; 30-37 ~ had a 29% surgery rate, and greater than 37 ~ had a 60% rate. The ability of the dysplastic acetabulum to respond by spontaneous growth alone to a normal anatomic state following concentric reduction of the head has been narrowed by clinical experience to the time from 18 months to 4 years.

C. Acetabular Development after Removal of the Limbus in Infancy O'Hara studied 61 patients treated in early childhood for CDH in 31 of whom excision of the limbus had been performed in (217). Major deleterious effects were noted at follow-up in adolescents and young adults who had had the limbus excised compared with others having open or closed reduction in whom the limbus was retained. In normal hips, the lateral acetabular epiphysis appeared to contribute between 7 ~ and 12~ to the center-edge angle with a mean of approximately 10~ All 31 patients subjected to removal of the limbus failed to develop a lateral acetabular epiphysis,

234

CHAPTER 3 9 Developmental Dysplasia of the Hip

although it was evident on the contralateral normal side in all patients by 13 years of age. Thus, there was a meaningful anterolateral acetabular defect in association with removal of the limbus. The lateral acetabular epiphysis was evident in all 30 patients in the group in which the limbus was preserved. O'Hara concluded that "retention of the limbus was the only factor predictive of its appearance." Somerville, who was the original proponent of excising the limbus, felt that there would be no long-term abnormal development of the acetabulum, but the studies by O'Hara and Sherlock et al. demonstrated developmental problems. O'Hara concluded that, whereas it is necessary to remove the interposed limbus to effect a concentric reduction, it is actually the tight medial capsular isthmus that blocks reduction. Although the limbus often is interposed, it should be retained by everting it because retention of this anatomic structure not only will contribute to post-relocation stability of the hip but is essential to normal acetabular development with time. He concluded that "excision of limbus is thus avoidable, undesirable, and unnecessary."

radiograph signs were relatively early indications of growth problems.

F. Proximal Femoral Growth Following Femoral Osteotomy Proximal femoral growth following varus-derotation osteotomy tends to show an increase to more valgus but with the value almost always in the normal range. This pattem was documented in five studies. Fritsch et aL reported preoperative/ postoperative/follow-up values after intertrochanteric varusderotation osteotomy at 143~176 ~ (82). Other studies were similar: 134~176 ~, 152~176 ~, 150~ 112~ ~, and 134~176 ~ (82).

XI. T R E A T M E N T B A S E D O N T H E S T A T E OF THE UNDERLYING PATHOANATOMY, INCLUDING SECONDARY CHANGES

A. General Overview D. Acetabular Growth Following Acetabular Surgery Acetabular growth following acetabular surgery is wellmaintained. Studies of acetabuloplasties show the maintenance of correction several years postsurgery. In most studies the acetabular angle decreases further toward the end of growth as ossification of the lateral acetabular apophyseal centers occurs. Three long-term studies after the Salter innominate osteotomy showed preoperative/postoperative/ long-term values of 34.2~176 ~, 30.50/20.90/20.2 ~, and 37~176 ~ (82). Six long-term studies of other acetabuloplasties confirmed the same findings: Dega procedure, 34.3~ 22.2~176 Mittlemeir procedure, 33.8~176176 and four Pemberton procedure studies (82).

E. Growth Disturbance Lines in Proximal Femur: O'Brien O'Brien recognized growth disturbance lines (Harris lines) in plain radiographs of the proximal femur after operative procedures for CDH (213). They were seen regularly in children between 6 months and 4 years of age and normally formed an L-shaped line with the shorter line underlying the medial neck physis and the longer line the greater trochanteric physis. Physeal growth adjacent to the medial line normally was twice as great as that adjacent to the lateral. Any growth damage to the medial (head-neck) physis manifested as alterations in the position or linearity of the shorter line. O'Brien et al. then showed such growth disturbance with partial or complete growth cessation in 33 of 68 patients undergoing CDH treatment (214). Two patterns were noted: one at the junction of the medial and lateral portions of the physis and the other in the medial area alone. These plain

Developmental dysplasia of the hip demonstrates well the interplay between the primary pathoanatomic features of the condition and the negative effects on growth, which are secondary sequelae of a failure to either recognize or adequately treat the primary situation. The longer the proximal femur is related to the socket imperfectly, the more extensive the interventions required to correct the primary and secondary defects and the less optimal the final results (Table III). As a general rule, the earlier treatment begins in a developmental dysplasia of the hip, the simpler, quicker, and more effective it is. The one concern about treatment for the completely dislocated hip initiated in the first 3 or 4 months of life is the slightly increased risk of avascular necrosis. Examination of the hips is now a uniform part of the newbom physical examination. Bilateral hip ultrasound is performed as part of the routine newbom assessment for all babies in some geographical areas and should be performed with any concem about abnormality on the physical examination. Predisposing factors for a DDH are breech presentation, oligohydramnios, first birth, twinning, and a positive family history. In hips that are clearly dislocatable, most centers begin treatment in the first few days of life, but in recognition of the well-demonstrated fact that some hips will stabilize spontaneously in the first week of life without treatment, other centers defer management until 1-2 weeks of age. Many treatment regimens are available, but a general consensus concerning the treatment of developmental dysplasia of the hip is present within the orthopedic community. There is extensive use of the Pavlik harness or variants of it, hip spica casting, and surgery involving open reduction, pelvic osteotomy, and proximal femoral osteotomy. Considerable variability remains in methods of assessing anatomy and pathoanatomy both of the hip at the time of discovery of the

SECTION Xl ~ Treatment Based on the State o f the Underlyino Pathoanatomy

TABLE III .

.

.

.

.

.

.

.

.

.

Increasino Pathologic Changes and Need for More Extensive Treatment the Older the Patient is before Diagnosis or before Full Reduction Is Achieved

.

Age range at diagnosis

ii

iii

Pathoanatomic findings

Treatment profile

Newborn-6 weeks

Capsular laxity

6 weeks +

Contracture of hip muscles 9 adductors Additional contracture of hip muscles 9 iliopsoas, gluteus minimus and medius Lengthening and hypertrophy of ligamentum teres

12 weeks +

Birth+

Fibrous fatty tissue in depths of acetabulum (pulvinar) Acetabular dysplasia

6-8 months +

Subluxation: acetabular dysplasia with more spacious acetabulum Dislocation: acetabular dysplasia with smaller, shallower, misshapen acetabulum

Birth+

Proximal femoral dysplasia 9 proximal femoral anteversion

4 months +

8-10 months +

9 misshapen smaller femoral head 9 delayed formation secondary ossification center of femoral head Interposition of labrum; usually not inverted postreduction before 6 months unless teratologic Narrowing of capsular isthmus

2-4.5 years +

Inferior acetabulum covered by transverse acetabular ligament, capsule Acetabulum filled with thickened ligamentum teres, excess fibrous fatty (pulvinar) tissue Pseudo-acetabulum

6-8 months +

235

abnormality and in the timing and type of the specific treatment used. In the following sections, we outline approaches to management concentrating on consideration of the underlying pathoanatomy as a guide.

B. Diagnosis Made in the N e w b o r n Period Diagnosis should be made in the newborn or early postnatal period. The Barlow maneuver performed with the hips in

Responds to Pavlik harness applied initially between birth and 12 weeks May need prereduction traction, percutaneous adductor release

Closed reduction with or without prereduction traction, percutaneous adductor release; hip spica

Responds with spontaneous correction to Pavlik harness, 0-12 weeks; hip spica following closed reduction, 12 weeks + Responds to concentric reduction 9 hip spica (used for several months); 9 acetabular surgery: 18 months (earliest); 18 months to 4 years (increased intervention); 4.5 years+ all require acetabular surgery Spontaneous correction with closed reduction 0-9 months or open reduction 10-18 months Proximal femoral varus derotation osteotomy 2 years Corrects spontaneously with reduction that is maintained Open reduction, labral repositioning if inverted postreduction (any age) Open reduction and capsulorrhaphy if it prevents reduction Open reduction Open reduction Open reduction, acetabuloplasty, proximal femoral osteotomy (varus, derotation, shortening)

90 ~ of flexion indicates whether the hips are stable or dislocatable. If the hip is dislocatable, adduction in association with gentle lateral and posterior pressure against the lesser trochanter will move the femoral head out of the socket over the rim of the acetabulum; the abduction maneuver, characterized by pressure against the greater trochanter in an anterior and medial direction, will serve to relocate the hip. The other positive clinical sign in a unilateral dislocation is referred to as the Galeazzi sign in which the knee on the

236

CHAPTER 3 ~

DevelopmentalDysplasia of the Hip

dislocated side is lower than that on the normal side with the hips and knees flexed and the feet flat on the examining table. Abduction generally is full even with dislocatable hips in the newborn period. Asymmetric thigh folds are not helpful diagnostically; they are present in many with normal hips and absent from some with abnormal hips. The large majority of dislocatable hips can be assessed clinically, but there is a subset of infants in whom dislocatability is not demonstrated or demonstrable clinically. The value of ultrasound is great in such situations. In the fully normal hip, upon ultrasound examination the femoral head is stable within the acetabulum through all ranges of motion, the head is 50% or more covered by the acetabulum, and the oL angle is 60 ~ or greater. The latter two indices in a well-located but immature hip may take several weeks to normalize fully. Plain radiographs in anteroposterior and frog lateral positions are of limited value in the newborn period even with both standardization of technique and considerable interpretive experience. If diagnosis is made within the first 4 - 6 weeks, application of a Pavlik harness followed by position check at selected intervals by ultrasound provides excellent results in the large majority of cases. The dislocatable hip is reduced with flexion and moderate abduction and the head slips under the labrum. The maneuver and harness application require no sedation or anesthesia and adductor releases are not needed. Some use the Pavlik harness initially up to 3 months of age and some even to 6 months. It is not the particular age that is the guide to treatment but the safe attainment of the fully reduced position. Beyond 3 months of age, patient size and strength gradually decrease the effectiveness of the harness. The Pavlik harness treatment often is referred to as a "functional" treatment because hip motion is maintained, albeit in a narrowed range, allowing for muscle function, intra-articular joint motion and synovial fluid nutrition and lubrication. Evidence exists that 50% of those hips that are dislocatable at birth will stabilize spontaneously in a normal position within the first week of life without specific intervention. Many feel that it is permissible not to treat but rather only to observe patients in this time period. Others are concerned that some of these hips might stabilize in a subluxed rather than fully relocated position and consider it reasonable to treat all dislocatable hips to be more assured of a better longterm result. It also is necessary to recognize that avascular necrosis can occur even with the use of a Pavlik harness and a clear treatment philosophy is needed. If observation alone is chosen, it is essential to assess the child at 1 week for clinical stability and confirm position and development sonographically. The initial treatment principle involves holding the hip in the reduced position for a sufficient period of time to allow the capsule to tighten to a normal level and thus stabilize the femoral head in its appropriate relationship to the acetabulum when immobilization is discontinued. In a straightforward dislocatable hip, there generally is no tight-

ness of the hip adductors. It has been found empirically that 2-3 months of treatment will, in the large majority of cases, allow the capsule to tighten adequately such that a normal hip will develop. Some treat for only 6 weeks. It is important to recognize that treatment protocols for developmental dysplasia of the hip are almost exclusively empiric. Both clinical assessment and imaging in relation to pathoanatomic concems should guide the treatment protocol at all stages. In general, in developmental dysplasia of the hip, earlier diagnosis leads to better long-term results and requires much less extensive intervention. In the newbom phase, conservative methods are used to hold the hip in the flexed and moderately abducted position using a Pavlik harness. These allow the child some hip movement but only within a restricted range. It is flexion beyond 100 ~ that is most effective in reducing hip dislocation. Abduction, particularly extremes of abduction, is less effective in achieving reduction and carries a very high risk of causing avascular necrosis. The Pavlik harness should hold the hips in 120 ~ of flexion or more and also in much less than full abduction. The harness is applied to not allow extension of the hips beyond the 120 ~ of flexion, but it does allow for knee and ankle flexion and extension. The hips also move in their limited range, and these ranges of joint motion are physiologic and a marked improvement over the rigid hip spica. The Pavlik harness treatment must be used on a complete 24-hr per day basis for dislocatable and dislocated hips. If it is removed for bathing, position can be lost and not necessarily regained when the hamess is reapplied. In most instances, flexion and slight abduction reduces the head with no interposition of the limbus. Ultrasound examination in harness is essential during the period of treatment to confirm good reduction of the hip. Examination after 2-3 months of treatment assesses for two key factors: (1) clinical assessment to demonstrate stability of the hip with dislocatability no longer possible and (2) normal ultrasound examination. Plain radiographs can also help confirm relocation and bone development, in particular in relation to the acetabular index, and the presence, location, and size of the secondary ossification center of the femoral head after 3 - 4 months of age. The longer the hip has been dislocated prior to closed reduction, the greater the secondary changes and thus the more complicated and the greater the length of treatment required.

C. Diagnosis Made at 3 Months of Age If a dislocatable hip is diagnosed at 3 months of age or older, hip spica treatment tends to be used because the Pavlik harness is less reliable in this age group. The children are older and can kick more vigorously, and in addition at 3 months some of the earlier secondary changes can be present and a more cautious approach is warranted. That being said, treatment must be individualized and some use the Pavlik harness

SECTION XI ~ Treatment Based on the State of the Underlying Pathoanatomy as an initial treatment up to 6 months of age, which is the age limit actually referred to by Pavlik himself. An empiric rule of thumb is that any hip should remain immobilized in the reduced position for the same amount of time that it had gone undetected and untreated after birth. That this is only an approximation is indicated further by the fact that it is often unclear at what stage the dislocation occurred because some clearly are intrauterine in the third trimester. The earliest clear-cut secondary change to clinical assessment is tightness of the hip adductor muscles on the affected side. Tightness to abduction in untreated DDH generally is noticeable between 2 and 3 months of age. Opinion is divided as to whether this adductor tightness itself requires specific intervention. There is good evidence that a 1- to 2-week period of prereduction or presplinting traction in hip flexion with gradually increasing, but never full, abduction helps lessen avascular necrosis. Percutaneous adductor release often is performed with or without traction prior to casting or Pavlik harness treatment. Several studies have documented the markedly diminished incidence of avascular necrosis of the femoral head in association with precast traction, percutaneous adductor release, and utilization of a hip spica cast position, which favors flexion well beyond 90 ~ but abduction only within the stable range, rather than previous regimens, which frequently immobilized the hips in flexion or extension but full abduction. Of the three parameters designed to minimize AVN of the femoral headmprereduction traction, percutaneous adductor release, and hip immobilization in a physiologic positionmit is agreed by all that position is most crucial in preventing AVN; forced abduction or abduction approaching 90 ~ is the main culprit. If tightness to full abduction is found, the treatment protocol is expanded to loosen the tight adductors prior to closed reduction and immobilization to minimize the risk of avascular necrosis. The patient can be placed in Bryant's traction for 1-2 weeks prior to undergoing closed reduction and immobilization either in a hip spica cast or (in some centers) in a Pavlik harness. Studies have suggested that it is not the traction alone that reduces the incidence of avascular necrosis, but no studies have shown it to be deleterious. Closed reduction generally is done under general anesthesia. Hip position should be checked either by ultrasound or by hip arthrogram at the time of reduction to rule out limbus inversion. CT is used to assess position postreduction in a hip spica cast. The possible use of MR imaging based on studies we have done on piglets and humans can assess both femoral head-acetabular position and head vascularity in patients at risk for AVN.

D. Diagnosis Made at 6 Months of Age If diagnosis is made at 6 months of age or beyond, a much more prolonged period of hip spica immobilization is needed, provided that closed reduction without interposition of the labrum is achieved. Considerable variability exists between

237'

different Centers and even between orthopedic surgeons in the same centers concerning management programs for dislocated hips initially diagnosed at 6 months of age or older. The treatment should be guided by pathoanatomic demonstration (pre- and postreduction) not by programmatic guidelines. If a hip dislocation has not been detected until the 6- to 12-month range, questions must now be raised as to whether closed reduction is possible. Generally speaking, the adductor muscle mass is demonstrably tight at this stage and prereduction traction plus percutaneous adductor release is warranted. With the child anesthetized, one can get a good sense of the position of the hip with a flexion, abduction, and internal rotation maneuver. Plain radiographs are not definitive in assessing the effectiveness of closed reduction, and additional approaches to confirm the completeness of reduction are needed. Generally speaking, in most patients under 3 months of age one can rest reasonably assured that a closed reduction will bring the head into the appropriate relationship with the acetabulum with the labrum in its normal position on top of the head rather than being inverted. Ultrasonography defines the quality of the reduction and it appears wise to demonstrate the position of the labrum post-reduction. Ultrasound is progressively less effective with the appearance and increased size of the secondary ossification center. An arthrogram in which 1-2 cm 3 of Renograffin dye are injected into the hip joint under fluoroscopic control then is helpful. If the labrum is in a normal position, a triangular thistlelike appearance is seen above the superolateral part of the femoral head on arthrography because the capsule does not articulate with the tip of the labrum but rather with the base of the labrum at the fibrocartilage junction. If the labrum has become inverted along with part of the capsule, referred to as the limbus, this characteristic relationship is not seen on the arthrogram and raises the question as to whether surgical intervention is warranted. Some previously felt that the hip could be immobilized by hip spica cast in this position and that pressure of the femoral head against the limbus would cause it to atrophy such that a normal head-acetabulum cartilage relationship eventually will be established. A more common belief currently is that an interposed limbus invariably leaves a ridge in the femoral head and predisposes one to midlife osteoarthritis, a concept of pathoanatomy that leads to surgical intervention to reposition the inverted limbus. Previously there was a divergence of opinion on what to do with the inverted limbus at surgery, some such as Somerville feeling that it should be removed surgically with others feeling that it represents a normal anatomic structure, namely, the labrum, and that it should be repositioned on top of the femoral head rather than excised. The latter view is now held almost universally. If surgical intervention is chosen in the 6- to 12-month period, attention is paid not only to the limbus but also to the capsule. A most important reason for intervening, other than to reduce the hip

238

CHAPTER 3 ~

Developmental Dysplasia of the Hip

concentrically and reposition the labrum, is to perform a capsulorrhaphy to tighten the capsule and stabilize the femoral head-acetabular relationship structurally rather than waiting for it to occur on the basis of immobilization over a several-month period in a hip spica cast.

E. Diagnosis Made at 12 Months of Age If diagnosis is not made until 12 months of age, open reduction to reposition the inverted limbus and perform capsulorrhaphy almost always is required. Hip spica immobilization follows. If closed reduction is done at this age it is essential to assess for limbus inversion by arthrogram.

F. Diagnosis Made at 18 Months of Age If the condition has gone undetected until 18 months of age, the secondary bone and cartilage changes will be extensive involving acetabular dysplasia and femoral anteversion, and the soft tissue changes also will be marked with adductor muscle tightening, transverse acetabular ligament and inferior capsule thickening and migration across the acetabular inlet, fibrous fatty tissue in the acetabulum, hypertrophy of the ligamentum teres, an inverted limbus, and an enlarged and thickened capsule. In this age group, therefore, although closed reduction may be attempted, it must be assessed by arthrogram to determine that the head is fully seated in the acetabulum without interposition of the limbus. The period of immobilization also is lengthy as it must be continued to allow the acetabulum and the proximal femur to develop such that they have a normal or close to normal conformation when the immobilization is released. The most detailed imaging assessments are needed in this group to assess the presence, absence, and extent of soft tissue and bony malformations both before reduction and after reduction during the period of immobilization. Differing general approaches have been used. The most aggressive surgical approaches to bone deformity begin at 18 months in the view of Schwartz, Salter, Pemberton, and others, who considered that full acetabular remodeling cannot be expected in all with closed or open reduction alone and that acetabular surgery to correct the dysplasia should be done along with reduction. The most common approach for those not comfortable waiting for acetabular response to repositioning is open reduction with innominate osteotomy, but some centers prefer the Pemberton procedure. The least aggressive surgical approaches to bone deformity rely on spontaneous acetabular remodeling following closed or open reduction up to 4-4.5 years of age. Between these ages, the older the patient the more likely the need for acetabular surgery. The next decision, once reduction has been achieved, concerns whether acetabular or proximal femoral bony surgery is done. If a hip is subluxed at this age, there are similar decisions needed regarding bone surgery. There is good evidence that if one bony component of the hip is corrected

surgically along with concentric reduction, the other component will correct to a normal range with growth. Some follow a programmatic approach, doing one procedure or the other. The most desirable approach for many is to correct surgically the particular component most needing correction and to correct both if necessary. The approach of Tonnis and also Mittelmeir et al. (both based on extensive experience) recommends both acetabular and proximal femoral correction along with open reduction as early as 2 years of age. In summary, with prolonged immobilization following closed or open reduction up to 4 years of age, the remodeling potential of the acetabulum and proximal femur is sufficiently great in the opinion of some that with persistence a normal relationship will be established. Some feel that after 18 months of age this remodeling will not correct all hips and may leave many with only a partially corrected subluxed hip. Bone surgery is used increasingly the older the patient, with virtually none recommending reduction alone at 4-4.5 years of age. The period of immobilization can be reduced dramatically by correcting the structural relationships by osteotomy. Open reduction is followed by pelvic osteotomy to bring the acetabular relationship quickly toward the normal position. The innominate osteotomy developed by Salter tilts the acetabulum laterally and forward such that, once healing has occurred, a virtually normal relationship has been established. The pericapsular osteotomy of Pemberton or acetabuloplasty procedures levering down the acetabular roof and stabilizing with a bone graft (used commonly in central Europe) directly correct the acetabular dysplasia. The advocates of acetabular osteotomy contend that, following associated open reduction with capsulorrhaphy, the femur speeds up its natural development such that femoral varus-derotation osteotomy is rarely necessary. An alternative approach, again following open reduction, is to perform a varus-derotation osteotomy of the proximal femur to deeply seat the femoral head in the acetabulum. Good evidence has accumulated that, once the head is deeply within the acetabulum and anteversion is corrected, acetabular development is hastened and a normal relationship is established with time such that acetabular osteotomy need not be done. Once again varus-derotation osteotomy is advocated by some between 2 and 4.5 years of age and by more after this age. Generally it is essential for persisting coxa valga and severe anteversion.

G. Diagnosis Made between 18 Months and 4.5 Years of Age Once a patient is between the ages of 2 and 4.5 years, the necessity for some surgical intervention is agreed to virtually by all. The surgical necessity is for open reduction and attention to bone deformity dependent on the degree of deformity. Some opt for relatively prolonged postreduction casting followed by splinting as noted earlier, others use pelvic osteotomy to hasten the development of a normal acetabulum,

SECTION Xli ~ Avascular Necrosis as a Complication of Treatment some favor proximal femoral osteotomy to correct anteversion and coxa valga, and some correct acetabulum and femur both by osteotomy after 2-3 years of age. The older the patient with the hip either untreated or treated imperfectly, the more extensive the surgical procedures that must be considered. One can readily appreciate the increasing extent of intervention required with each passing year of either undetected or imperfectly corrected proximal femoral-acetabular relationships.

239

range. Precise surgery is required because of the limited time remaining for growth and remodeling. Acetabular depth must be reestablished because a sloping acetabulum invariably leads to long-term degenerative changes. In this age group, the Steel innominate osteotomy or a Pemberton osteotomy can be used, but increasingly the periacetabular procedure of Ganz is favored. The Chiari and shelf procedures have been done but are not particularly physiologic or attractive in otherwise normal ambulatory patients. Proximal femoral varus-derotation osteotomy might be needed.

H. Imperfect Hip Structure after 5 Years of Age The treatment of imperfect hip structure after 5 years of age remains an area with considerable variation of opinion. Imperfect hip structure persisting after the age of 5 years results either from previously undiagnosed hip dysplasia or subluxation or as a result of treatment for a congenital hip disorder that has left the individual not fully corrected and thus with persisting abnormality. There is excellent documentation of the fact that these hips will become symptomatic in adolescence, early adult life, or mid-adult life with the timing of the discomfort reasonably correlated with the extent of the abnormality; those with the mildest of abnormalities develop the symptoms at later time periods. It is in this time frame, however, that the approaches vary the greatest. Dimeglio and Pous outlined surgical treatment in those with congenitally dislocated hips from 5 to 10 years of age (58). The approach that seemed to be the best involved iliopsoas and adductor tenotomy, open reduction, capsulorrhaphy, femoral shortening with varusization and derotation as needed, and innominate osteotomy. A similar outline was provided by Heinrich et al. who listed nonoperative and operative procedures available for dysplasia, subluxation, and dislocation from walking age to skeletal maturity (115).

I. Preadolescent-Adolescent Hip Dysplasia Catterall has outlined several principles of approach to this group based on a series of abnormalities noted at examination under anesthetic with arthrography (37). The types of movement are defined as stable concentric movement, stable eccentric movement, reducible subluxation, irreducible subluxation, unstable lateral segment, and unstable movement with hinge abduction. The unstable lateral segment refers to a tear of the acetabular labrum, which can best be demonstrated dynamically by arthrographic and fluoroscopic movements. This dynamic test also can be supplemented by studies on the three-dimensional structure of the acetabulum in relation to the nature of the dysplastic involvement and of the proximal femur in relation to the shape of the femoral head, length of the femoral neck, and degree of head-neck shaft inclination and proximal femoral anteversion. After the age of 8 years, the potential for acetabular remodeling is minimal to nonexistent and the existence of a structural defect implies the need to correct it surgically to an anatomically normal

XII. A V A S C U L A R N E C R O S I S AS A COMPLICATION OF TREATMENT OF DEVELOPMENTAL DYSPLASIA OF THE H I P There are two major problems that can occur in relation to treatment of congenital-developmental dysplasia of the hip. One is the failure of complete anatomic restoration, which predisposes one to eventual osteoarthritis. The other is avascular necrosis, which leads to varying degrees of disordered proximal femoral growth. In this section, we detail knowledge of the proximal femoral blood supply, ways in which it can be damaged during treatment, the sequelae of that damage, and new methods of diagnosing avascularity within hours so that prompt responses can minimize or eliminate the complication.

A. Blood Supply of the Proximal Femur The blood supply of the proximal femur has been welldefined and plays an extremely important role at all ages in relation to hip disorders (42, 51, 216, 218, 323, 324, 326, 338, 360). The proximal femoral capital epiphysis is completely intracapsular and the blood supply is somewhat tenuous.

1. GENERAL PATTERN OF BLOOD SUPPLY The blood supply of the proximal femur has its origin from either the femoral or the profunda femora artery. Two vessels that ultimately will supply the femoral head and neck then pass from these arteries circumferentially around the femur, where they are referred to as the medial femoral circumflex and lateral femoral circumflex arteries. The medial circumflex artery is the thicker and longer of the two and passes posterior to the femur at the level of the base of the neck. The lateral circumflex vessel is anterior and tends to thin and become incomplete medially. On occasion, a complete circumferential arc is formed by these two vessels. A series of vessels referred to as the ascending cervical arteries then passes from both the posterior and the anterior circumflex vessels in a proximal direction. These vessels initially are external to the capsule of the neck but soon pass through the capsule and run along the surface of the femoral neck bone, where they are contained in a loose connective tissue

240

CHAPTER 3 ~ Developmental Dysplasia ojr the Hip

matrix referred to as the retinaculum of Weitbrecht. The vessels then pass over the physis and enter the cartilage of the developing femoral head in a thin rim between the periphery of the articular cartilage and the physeal cartilage. The greatest concentration of vessels entering the femoral head is at the superolateral aspect of the head-neck junction. They then ramify from lateral to medial within the femoral head; the vessels are referred to as the lateral epiphyseal vessels. They are concentrated at the posterolateral region of the head and neck junction. The lateral epiphyseal vessels, which were clearly defined by Trueta as well as others, supply fully four-fifths of the bone and cartilage of the femoral head. Some medial epiphyseal vessels traverse the space between the articular cartilage and physis medially. Most of the vessels on the medial side, however, which also are derived from the ascending cervical arteries, pass into the metaphyseal region to supply the femoral neck bone. A small blood supply to the femoral head comes from the ligamentum teres, but this at best supplies one-fifth of the femoral head adjacent to the ligamentum and cannot be relied upon if damage to the other major vessel system has occurred. Indeed, when the ligamentum teres is cut during the process of open reduction of a dislocated hip, there is either no bleeding or only a trace of bleeding. The pattern of blood supply remains the same through childhood and even after growth plate fusion. Although merging of the epiphyseal and metaphyseal vasculature occurs, the overall pattern remains the same throughout the remainder of adult life. 2. ADDITIONAL DETAILS OF BLOOD SUPPLY Many studies have been made on the blood supply of the head and neck of the femur. The basic outline described herein is recognized by all, as is the fact that there are slight variations in many individuals. Studies continue to be made primarily to determine the extent of contribution of the various vessels and subtle but possibly meaningful changes that occur during time periods from the late fetal stage to skeletal maturation.

a. Origin of the Deep and Circumflex Femoral Arteries Early studies defined the origin of the vessels supplying the proximal femur. These derived initially from the femoral artery itself, although most of the circumflex vessels took their origin from the profunda femora artery. Williams et al. performed extensive studies of this region of the vasculature and also reviewed the previous studies (358) (Fig. 15A). Their total review eventually reported on 1576 dissections. They defined 7 types of patterns of vessel origin, but 91% of the patients studied had patterns in 3 groups only. When all patients were combined, the type 4 pattern was seen in 898 of 1576 or 57%, the type 2 pattern in 309 of 1576 or 20%, and the type 3 pattern in 224 of 1576 or 14%. In the type 2 pattern the medial femoral circumflex artery took its origin directly from the femoral artery with the lateral femoral circumflex artery originating from the profunda femora, and in type 3 the reverse was seen with the lateral femoral circum-

flex artery originating directly from the femoral artery and the medial branch originating from the profunda femora artery. The type 1 pattern was seen in 61 patients or 3.9% and the type 5 pattern in 50 patients or 3%.

b. Vascular Contribution from the Ligamentum Teres Virtually all authors now agree that there is no meaningful contribution to viability of the cartilage or bone of the head of the femur coming from the vessels within the ligamentum teres. The pattern of its contribution, however, changes at varying ages and also shows considerable variability in different individuals. Wolcott was unable to demonstrate blood vessels entering the femoral head in specimens from infants and children under 10 years of age by injection techniques of opaque materials or by histological serial sections (360). Walmsley showed that the vessels of the ligamentum did not supply the secondary ossification center in two children aged 2 and 6 years (338). He also referred to the work of several other authors indicating that, in the adult, vessels within the ligamentum teres provided only a minimal amount of blood to the femoral head. Tucker showed that the artery of the ligamentum, which he referred to as the foveolar artery, arose from either the obturator or the medial femoral circumflex artery (326). In children up to the age of 13 years the contributions to vascularity of the head were quite minimal. In 8 of 24 specimens the artery penetrated the fovea and supplied the cartilage of the head with occasional contribution to the bony center. The vessels, however, were very small. In the other 16 specimens the blood supply exclusively appeared to involve the fibrous tissue of the ligament and only its attachment to the head. In adults, however, there was increasing size and depth of penetration of the vessels, and they supplied the immediate regional area of the bony head in 14 of 20 specimens. The fact that these vessels had a minimal effect in the child and a slightly increasing effect in the adult was noted by others. Lauritzen also found the artery of the ligamentum teres to have little if any importance to the epiphyseal blood supply in children (161). Trueta found vessels from the ligamentum teres in the late fetal period and at birth, but they were not constant and entered the cartilage only to a minimal extent (324). Even this small contribution then rapidly diminished such that from shortly after birth to 4 years of age the vessels of the ligamentum teres did not contribute to the nourishment of the head, and indeed it was not until 8 or 9 years of age that the vessels did contribute something to the blood supply of the head.

c. Lateral and Medial Circumflex Arteries and Ascending Cervical (Retinacular) Arteries The major and for the most part exclusive blood supply of the femoral head comes from the lateral and medial circumflex arteries from below, which pass proximally through the capsule and then run on the surface of the femoral neck in the retinaculum. They are referred to by various reports as the ascending cervical arteries or as either the capsular or the retinacular vessels. Walmsley recognized that vessels supplying the femoral head and its osseous center were derived from the medial circumflex

A L

M

M

L

D .V

1

L

la

M

L

1r

lb

M

L

2

3

L L

De

4

M

5

6

F I G U R E 15 (A) Most of the blood supply of the femoral head is derived from the medial and lateral circumflex femoral vessels. The origin of these vessels, however, from the femoral artery and profunda femora artery is variable. The pattems of formation were reviewed extensively by Williams et al. (B) Anterior view of the vasculature of the hip region. (C) Posterior view of the vascular supply of hip region. (D) A composite picture of vascularity going from the femoral artery to the retinacular cervical neck vessels is shown. (E) View of the proximal femoral blood supply in transverse section showing the lateral and medial circumflex femoral arteries, the

Arterial brancnes to superior luteal art r _ .,,POSTERIOR ASCENDING g e y ~_.~~ /yCERVICAL ARTERIES Capsule \\/.J~.....,~/

E

MEDIAL ASCENDING ~ ~ ) ~ ~ 7 - ~ ~ ~ ~ .,~~~-"~..,XX~g#/

CERVICAL ARTERIES ~

9

,

)I Fernoraii<', :. r~#J,l~//

!i:i!!::abt~i::: ~ermCor~fll::terY~

Femoral artery/

/ "~i '#' ,

,

~

/

/

Profundus femoral artery'

/

~

,

'".~ ' ~ ',~ i i i'i i ) i N . G i/ ~ ~ '.~ _

~ ~ ~ 'ANTERIOR /1 II II I ASCENDING / \ I I 1/ CERVICAL ARTERIES ~ I~ /~i / ': v/

~

Lateral c,rcumflex femoral artery

Fi

~I

f

x~rcaO;htant

i ~'-"~--~""

! I

/ / T k

Medial Circumflex Femoral Artery (pos

I

/ / / /

I

/ I

I

[ I /

/

Capsule

i

Femoral Artery

Ascending Cervical Retinacular Vessels

a posteror b medial c anterior d lateral

Profunda Femoris Artery

Origin from

Major Vessels

Lateral Circumflex Femoral Artery (anterior)

Extracapsular Vessels Around Base of Neck

Fii

Vessels

Ascending Cervical ~ , ~ Retinacular Vessels ~

(on neck, withincapsule, subsy

a b r d

Lateral epiphyseal Medial epiphyseal Inferior metaphyseal Superior metaphyseal ~ ~ ~ Retinacular

/~. ~-'~,.~ ~'~,-'~ ~ ~

-%

RetinacularAscending Cervical Vessels (on surfaceof neck)

Vessels on Surface of Neck

Intraosseous- Intracartilaginous Vessels

FIGURE 15 (continued) branching medial, posterior, anterior, and lateral ascending cervical arteries, their penetration through the hip joint capsule, and their position as retinacular vessels on the femoral neck. (F) Entire sequence of blood supply from origin from major vessels to intraosseous, intracartilaginousposition is shown. [Part A from Williams, G. D., et al. (1934). Anat. Record 60:189-196, copyright 9 1934, reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc. Parts B and C reprinted from Grant, J. C. B. (1972). "An Atlas of Anatomy," 9 Lippincott Williams & Wilkins, with permission. Part D reprinted from Tonnis, D. (1987). "Congenital Dysplasia and Dislocation of the Hip," 9 Springer Verlag, with permission. Part E reprinted from (42), with permission.]

SECTION XII ~ Avascular Necrosis as a Complication of Treatment branch of the profunda femoris artery and then from those entering the neck of the femur at the articular margin of the head (338). Experimental work as reported by Wolcott also indicated that no marked change occurred in animals in which the ligamentum teres alone had been severed (360). Wolcott was of the opinion that the capsular arteries were the only essential source of blood supply to the ossifying center in the head of the femur. He also showed that it was the vessels from the medial circumflex artery in particular that provided the major blood supply to the head and neck regions. Both Wolcott and Tucker observed that the vessels of the ligamentum teres increased in size with age, particularly beginning in the second decade of life. Tucker studied 44 femurs from birth to 77 years of age (326). He utilized vascular injection studies with barium sulfate followed by radiography after decalcification. He also noted three series of vessels supplying the upper end of the femur, these being the nutrient artery of the shaft, the retinacular or capsular arteries, and the artery of the ligamentum teres. He also noted that the vessels do not run in the capsule but rather pass through it at the base of the neck and lie within the retinaculum on the surface of the neck. The nutrient artery sent its terminal vessels into the neck to supply the metaphyseal bone of that region. The nutrient vessels anastomose with the ascending cervical branches of the retinacular arteries once the latter had passed into the bone. The retinacular vessels arose from the medial and lateral femoral circumflex arteries, but there also was significant extracapsular anastomosis in the region of the trochanteric fossa to which the inferior gluteal, profunda femorus, obturator, and circumflex arteries contributed. He stressed that the circumflex arteries were superficial to the distal part of the fibrous capsule and that they did not run within its substance but rather passed through it at its base only. There were three main groups of retinacular arteries, these being the posterosuperior, the posteroinferior, and the anterior. The first two groups were branches of the medial femoral circumflex artery and they ran along the upper and lower borders of the neck of the femur. The anterior was the smallest and least constant and its vessels were branches of the lateral femoral circumflex artery. In their cervical course, the retinacular vessels that lie loosely under the synovial membrane supplied many of the branches of the femoral neck, which anastomosed internally with the nutrient artery of the shaft. Tucker noted that the posterosuperior group of retinacular vessels ran along the surface of the neck and crossed the physis on its peripheral surface, and only then did they enter into the cartilage of the femoral head to run toward the central portions. The posteroinferior and anterior vessels, however, often passed through the peripheral comers of the growth plate. This pattern maintained itself with the adult. With obliteration of the physis each of the three vessel systems freely anastomosed with each other, these being the nutrient vessels from the neck, the epiphyseal vessels within the head, and even the vessels from the ligamentum teres.

243

Chung studied 150 specimens from the 6th fetal month to 15 years of age (42). He noted the anastomotic extracapsular ring formed by the medial and lateral femoral circumflex arteries and observed the subsynovial intra-articular ring similar to that noted by Tucker and Hunter. The primary contribution to the circumferential extracapsular arterial ring was from the medial femoral circumflex artery, which tended to supply the medial, posterior, and lateral components. The lateral femoral circumflex artery compromised the anterior portion of the ring. The arteries passed through the capsule and then gave rise to the ascending cervical vessels. Chung referred to these ascending cervical arteries by position, referring to them as anterior, lateral, posterior, and medial ascending cervical arteries. Only the anterior branches were from the lateral circumflex. The ascending cervical arteries traverse the capsule at the base of the femoral neck. The numerous epiphyseal and metaphyseal branches of the lateral ascending cervical arteries supplied the greatest volume of the femoral head and neck. These vessels tended, however, to arise from a single arterial stem at the posterior trochanteric fossa. Others refer to the ascending cervical arteries as the retinacular vessels. Once they have passed through the capsule, the vessels then pass beneath the synovium and lie within the retinacular tissue on the surface of the neck. The four ascending cervical arterial groups (anterior, medial, posterior, and lateral) formed a subsynovial anastomotic ring on the surface of the neck at the margin of the articular cartilage, the ring defined by Hunter as the circulus articuli vasculosus. Often the anterior portions were incomplete. Chung documented 26 complete tings, 25 incomplete anterior tings, 13 incomplete posterior tings, and 49 incomplete combined tings. O'Hara and Dommisse also supported considerable contribution to the femoral head from the lateral as well as medial circumflex arteries at birth (218). Their injection, however, was much more proximal, and they were able to determine that branches of the inferior gluteal artery in 17 of 18 specimens contributed a major part of the blood supply to the femoral head by anastomosing laterally with the anterior and posterior circumflex arteries. d. Intracartilaginous-lntraosseous Vessels Trueta and Harrison utilized the barium sulfate injection technique to study the vascular supply of the head and neck in the adult (323). The vascular patterns established during the growth phase were not replaced at maturity but persisted, with the only change being anastomosis between the epiphyseal and metaphyseal circulations within the bone once the physis had been obliterated. They developed the terminology referring to the vessels within in the bone as the lateral and medial epiphyseal vessels based on their sites of entry and the superior and inferior metaphyseal arteries. The lateral epiphyseal and both groups of metaphyseal arteries arose from the medial femoral circumflex artery, whereas the medial epiphyseal artery in the adult was a continuation of the artery of the ligamentum teres, which itself came from the acetabular

244

CHAPTER 3 "

DevelopmentalDysplasia of the Hip

branch of the obturator artery. The proportionate contributions of each of the vessels were assessed. The lateral epiphyseal arteries predominated in the epiphysis and the inferior metaphyseal arteries predominated in the metaphysis. The lateral epiphyseal arteries supplied at least four-fifth of the epiphysis in most, whereas the inferior metaphyseal vessels were found to supply about two-thirds of the metaphyseal tissue adjacent to the femoral head. The lateral epiphyseal vessels entered the head primarily at the superior and posterosuperior regions usually through 2-6 entry points. The ascending cervical vessels, when they passed over the neck region, gave off 2-4 superior metaphyseal arterial branches. The inferior metaphyseal arteries tended to enter the bone close to the inferior margin of the articular cartilage. The metaphyseal vessels were not quite as well-defined as the epiphyseal vessels at their point of origin and they tended to ramify on the surface of the neck before passing into the bone. The metaphyseal arteries, by their frequent interconnecting anastomoses in the subsynovial tissues, conform to the pattern described by Hunter that he named the circulus articuli vasculosus, such that on the neck of the femur the circle was prominent although it tended to be somewhat deficient anteriorly. Lauritzen also confirmed that the deep branch of the medial circumflex artery terminated as the lateral arteries to the epiphysis in childhood (161). Trueta then studied the slight variations in the normal vascular anatomy of the head and neck region during growth in 46 specimens from late fetal life to 17 years of age (324). He divided the vascular studies into several time flames involving birth, the infantile phase from 4 months to 4 years, the intermediate phase from 4 to 7 years, the preadolescent phase around 9 and 10 years of age, and adolescence. In particular, he noted vessels from the medial side of the metaphysis passing through the physeal region of the epiphyseal growth plate into the cartilage of the femoral head. Most of the head, however, was supplied by the lateral epiphyseal vessels entering from the region of the trochanteric notch and tending to pass horizontally toward the center of the head. A characteristic of the intra-epiphyseal vessels was that every artery nearing its termination was found to break into a number of precapillaries and capillaries, which then again joined a single large vein that ran back in close contact with the artery. In the late fetal period, vessels from the ligamentum teres did pass into the cartilage of the femoral head, distributing themselves in an area similar to one found in later life. The three vascular systems of the femoral head in the late fetal period and at birth, however, did not communicate with each other. By 4 months of age, the vascular pattern had changed. There was complete disappearance of the penetrating vessels from the ligamentum teres. The secondary ossification center, which tended to appear around 4 months of age, thus was supplied almost exclusively by the lateral epiphyseal vessels with occasional supply from the inferior metaphyseal vessels as well. The vessels of the lateral epiphyseal circulation supplied the secondary ossification center and also the adjacent

cartilage. They originated from the posterosuperior aspect of the femoral neck and represented the termination of the vessels coming from the medial femoral circumflex arteries. The intermediate vascular pattern was characterized by blood supply to the bony epiphysis from only one source, that being the lateral epiphyseal vessels. In the preadolescent period from the age of 7 years upward reinstitution of vascular supply from the ligamentum teres into the adjacent periphery of the secondary ossification center began to be seen. The growth plate itself remained an isolating barrier between the femoral epiphysis bone and that of the neck. There were no real changes into the adolescent period until physeal obliteration, at which time there was union between vessels and bone of the secondary ossification center of the epiphysis and the metaphyseal neck. The importance of the vessels from the ligamentum teres increasingly was evident in a localized region during the adolescent period. Even from 4 months to 4 years, Trueta demonstrated blood flow from the medial metaphyseal vessels crossing the area to be occupied by the growth plate and passing into the head. After about 8 or 9 years of age, the vessels to the ligamentum teres in a sense reactivated to supply some blood to the femoral head. Lagrange and Dunoyer also demonstrated the vascular supply to the developing femoral head with high-quality arteriograms, showing the superolateral ascending cervical arteries and the lateral epiphyseal vessels clearly (152). e. Changing Vascular Patterns in the Early Years Ogden concentrated on vascular supply in the early years of life, studying 36 hip joints ranging in age from the 7th fetal month to 3 years (216). Though he noted the same pattern as others, he felt that initially the proximal femoral epiphysis including growth plate was supplied approximately equally by the lateral circumflex and medial circumflex arterial branches. Subsequent vascular development was characterized by regression of the lateral circumflex system and increased relative development of the medial or posterior circumflex system. During the first year of life, he noted the diffuse canalicular vascular network, which was primarily end-arterial within the cartilage, and its epiphysis and also documented several arterial vessels passing through the growth plate from metaphysis to epiphysis. The lateral circumflex artery supplied the anterolateral growth plate, the majority of the greater trochanter, and the anteromedial part of the femoral head. The medial circumflex artery tended to supply the posteromedial part of the head, the posterior growth plate, and the posterior part of the greater trochanter. The artery of the ligamentum teres supplied only a small medial area of the femoral head. By 3 years of age, the entire capital femoral epiphysis and growth plate were supplied by the medial circumflex artery through two major retinacular vessel systems, the posterosuperior and posteroinferior. The lateral circumflex vessels primarily supplied the greater trochanter and a small area of the anteromedial metaphysis but really little to no part of the capital femoral epiphysis. Ogden pointed out that, once epiphyseal vascularity was associated with development of the secondary ossification center, anas-

SECTION XII ~ Avascular Necrosis as a Complication of Treatment

245

F I G U R E 16 (A) Cartilage canals are seen (arrows) in the normal femoral head of a child who died a few weeks postnatally of nonorthopedic disease. Note the large lateral epiphyseal vessel entering the head from the superolateral region of the neck (curved arrow). (B) Higher power view of the cartilage canal in piglet femoral head. Arterioles, venules, and sinusoids are embedded in a fibrous connective tissue. [Part A reprinted from Barnewolt et al. (1997). Am. J. Roentgenol. 169:183-189, with permission. Part B reprinted from Jaramillo et al. (1996). Am. J. Roentgenol. 166:879-887, with permission.]

tomoses between the epiphyseal vessels were established. The secondary center thus ended the system of cartilage vessel independence. By 15-18 months of age no vessels were observed crossing the growth plate. Ogden noted, in agreement with Tucker, that the principal blood supply to the femoral head eventually was derived from two vascular systems coursing along the posterosuperior and posteroinferior aspects of the developing femoral neck and that indeed both of these were continuations of the medial femoral circumflex artery. There tended to be a change in pattern only in that, from a multiple, small vessel supply at birth, these two posterior systems evolved over approximately 18-24 months such that each system entered the head in a characteristic localized region. At birth, vessels penetrate the chondral epiphysis every few millimeters along the intertrochanteric notch, but with time more focal entry points are found. Most of the initial anterior epiphyseal and physeal contributions in the lateral circumflex artery regress, and thus there is a relative and absolute increase in elongation and thickness of the branches of the medial circumflex artery along the posterior femoral neck. Ogden noted that both the posterosuperior and posteroinferior vessels appear to have significant vascular roles. Of importance also was the development of anastomoses through the secondary ossification center. The blood supply of the proximal femur at progressively higher levels of detail is shown in Figs. 15B-15G.

B. Epiphyseal Blood Supply: Cartilage Canals Fetal and postnatal epiphyseal cartilage has a blood supply within its substance that is present within cartilage canals. These are seen initially in the human at 3 months in the developing fetus in the distal femoral epiphysis, where they arrive from the posterior surface. By 7 months all the larger cartilage epiphyses in the human fetus have vascular canals.

The cartilage canals subsequently will course through the epiphyseal cartilage, although they are never present within the articular cartilage and are scanty to absent in the physeal cartilage. Each canal contains an arteriole, a venule, and a capillary or sinusoidal plexus lying in a matrix of connective tissue (Fig. 16). These canals show conspicuous branching patterns in the cartilage, but they are independent of each other and do not formally anastomose with adjacent vessels. The epiphyseal vessels have a 2-fold function: (1) to provide nutrition to the developing epiphyseal cartilage and (2) to provide a source of preosteoblasts and osteoblasts for development of the secondary ossification center. The nutritive function to the cartilage is an important one and antedates the appearance of the secondary ossification center by several months. The vessels within the cartilage canals are derived from vessels initially associated with the perichondrium (280). As the cartilage mass enlarges, vessels that were perichondrial in position persist and become embedded within the cartilage matrix of the epiphysis. The vessels also position themselves within the cartilage by an active process associated with chondrolysis. Although the extraosseous and extracartilaginous blood supply of the proximal femur is well-known and tends to be relatively standard from patient to patient, the cartilage canals within the epiphyseal cartilage are random in appearance with no specific pattern identified from patient to patient based on serial section studies. The cartilage canals also, for the most part, are independent of each other with no tendency to form anastomoses. Haines has described six specific patterns of the cartilage canals in the human fetus based on serial section reconstructions (103). The patterns were simple unbranched canals, simple branched canals, doubleand multiple-rooted canals, tunnel canals (traversing a peripheral segment of cartilage from edge to edge), divided and rejoining canals, and transphyseal communicating canals. He noted that anastomoses did not occur between separate

246

CHAPTER 3 ~ Developmental Dysplasia of the Hip

canals. The canals thus can be considered as end vessels, which has implications in relation to possible avascular necrosis because end vessels and the tissue regions they supply are more susceptible to damage because additional feeder vessels are not available to establish new routes of flow and nutrition. The independent nature of the individual cartilage canals has been noted by Hintsche (117), Haines (103, 104), and Hurrell (128). The articular cartilage regions of the epiphysis are never traversed by cartilage canals. There is considerable transphyseal vessel passage, however, with metaphyseal marrow vessels passing through the physeal cartilage and into the epiphyseal cartilage adjacent to the upper regions of the growth plate (216, 280). These are considered by most authors to be present on the basis of growth of the developing of epiphyseal region with the cartilage vessel then encompassed by the physeal area, but evidence has been presented based on tritiated thymidine autoradiographs that active cartilage canal proliferation occurs. They tend to be relatively transient and in the human are described infrequently after birth. One of the regions, however, where they tend to persist is the proximal femoral head-neck area. In certain animals, namely, the lamb, transphyseal vessels are a common feature of all epiphyses even several months after birth. Many of them contain blood cells on histologic section. The secondary ossification centers invariably form within the central region of each epiphyseal cartilage. The reasons for this positioning have been thought to be due to diminished nutrition centrally or mechanical causes. Haines points out that the secondary centers of ossification appear where cartilage canals are scanty or absent (avascular lamina) based on observations from his reconstructions in the proximal cat femur and humerus at 4 days and distal human femur at birth. At these times mineralization of the central cartilage matrix is occurring followed by invasion of vascularity from the adjacent cartilage canals with an early differentiation of undifferentiated mesenchymal cells to preosteoblasts and osteoblast progenitors, which then synthesize bone on the calcified cartilage cores. Details of the blood supply to the proximal femur, including that contained within the cartilage canals, are essential to an understanding of avascular necrosis, one of the major complications of the treatment of DDH. The patterns of growth abnormalities that occur at the proximal femur when part of the growth plate (usually the cervical part) is damaged while the trochanteric portion continues to grow have been illustrated well by Morgan and Somerville (207), Taussig et al. (303), Siffert (282), and Tonnis (315).

C. Recognition of the Problem of Avascular Necrosis as a Complication of Treatment for Developmental Dysplasia of the Hip Treatment of dislocation of the hip in the latter part of the nineteenth and early twentieth centuries concentrated on

methods to reposition the head of the femur in the acetabulum and to maintain it there until sufficient stability had been obtained to allow for its retention post-immobilization. Methods involved both closed reduction and operative open reduction procedures followed by hip spica casting. Although there was slow but steady improvement in the incidence of relocation of the hips, it gradually became evident that there were many negative growth sequelae following both open and closed reduction. With the widespread use of radiographs and the clinical and pathologic recognition of the Legg-Perthes disorder, it became increasingly evident that the negative sequelae were due to avascular necrosis (AVN) of the femoral head occurring secondary to treatment. The AVN also was recognized as being iatrogenic because the particular findings characteristic of it were not seen in those hips that remained even completely dislocated in the absence of treatment. Reports during the 1930s and 1940s spoke of the high incidence of avascular changes, similar to those seen in Legg-Perthes, following treatment of congenital dislocation of the hip. The incidence of avascular necrosis was very high in clinical reports; one series estimated it to be as high as 73% (363) of all hips, with many reports in the 25-50% range. It was also noted in some instances to involve the non-dislocated contralateral hip, which of necessity also had been immobilized by hip spica treatments. Massie in 1951 documented a high incidence in his own institution of 35% in 89 hips overall with a 30% incidence in open reductions (188). In terms of causation, Massie commented on varying degrees of trauma in the reduction and plaster fixation of a dislocated hip after the age of 1 year. Long-term follow-up study documented patients into adulthood, showing how the resulting deformities predisposed the hips to the development of degenerative arthritis. A subsequent review continued to show a high 42% incidence of AVN. Massie's long-term study of the consequences of avascular necrosis complicating treatment of congenital hip dislocation indicated that 35% of patients treated for CDH developed definite vascular changes, with another 20% having partial involvement, and that in 72% of the hips involved by such changes results in the adult were unsatisfactory. Judet reported an incidence of 51% AVN (41 of 79 cases) with closed reduction using the Lorenz technique and 22% (11 of 50 cases) following open reduction and postoperative immobilization using much less abduction (138). Buchanan et al. reviewed 125 children treated at a single institution from 1945 to 1976 by various means with a 36% incidence of AVN (30). Ponseti reviewed results in 1944 and indicated a 46.2% rate of avascular necrosis, with radiographic and clinical changes in 80 of 173 hips (232). Variability in the degree of damage and in the long-term results was noted, although the earlier the radiographic evidence appeared the worse the prognosis. In over one-half of the cases, however, the ultimate damage in terms of femoral head shape and corresponding abnormality of the acetabulum was severe. In

SECTION XII ~ Avascular Necrosis as a Complication of Treatment the study reported by Esteve, patients treated with congenital dislocation of the hip in whom reduction was attempted by closed treatments between 1920 and 1950 showed an incidence of avascular necrosis of 52% (65). The rate of avascular necrosis following initial manipulation was extremely high at 69% of 64 hips, whereas a gentler method subsequently adopted in which reduction was obtained gradually on a specially constructed frame diminished the rate to 38% in 77 hips. Patients from the same institution, primarily including the group reviewed by Esteve but adding a few additional cases, were reviewed by Lima et al. In that intensive review of 184 hips, 90 or 48.9% showed some degree of pathological development of the femoral head consistent with an AVN event (176). The frequency of AVN was much higher in those cases treated initially by manipulation (68.4%) than in cases that initially were more gently reduced on a frame over a longer time period (37.5%). As in most papers, there was a much higher incidence of AVN in the more severe groups with incomplete or defective reductions. There was a substantial reduction of the incidence of AVN (referred to as osteochondritis) after slow but progressive reduction on an abduction frame. Immobilization of the hip in extreme rotation also had negative effects, with the implication that the stretching of the capsule in forced internal rotation and abduction obstructed the lateral epiphyseal vessels. Other high incidences of avascular necrosis were reported by Leveuf (23%) (175), Gage and Winter (33%) (86), Westin e t al. (24%) (349), Mardam-Bey and MacEwen (29%) (186), Powell et al. (32%) (238), Gregosiewicz and Wosko (21%) (99), Lempicki et al. 19% (171), Bost et al. (52%) (23), McKenzie et al. (25%) (184), and Hilgenreiner 33% (116). Pous et al. have reviewed the mechanisms causing AVN and methods used to minimize it (237).

D. Efforts to Understand and Treat the Causes of Avascular Necrosis 1. PRINCIPLES OF TREATMENT TO DIMINISH AVN

With recognition of the problem efforts were made to determine its cause and diminish its occurrence. Three management features were prominent in efforts to minimize the trauma to the vascularity of the proximal femur: (1) Prereduction traction. A several-week period of traction to slowly stretch the tightened hip muscles and improve the relationship of the femoral head to the acetabulum preceded a gentle closed reduction. (2) Reduction was accompanied by a percutaneous adductor tenotomy particularly in those instances in which traction had not allowed for a full release of the tightened medial structures. (3) Elimination of extreme positions of hip immobilization postreduction because these positions in hip spica or other devices were themselves detrimental to the vascularity. The two most commonly used positions from the early decades of the century were both implicated in causing damaging effects on the vascularity. One was the Lange position in which following reduction

247'

the hips were held in full extension, marked abduction, and marked internal rotation, and the other was the Lorenz position in which following reduction the hips were held in 90 ~ flexion and 90 ~ abduction. 2. CLINICAL AND RESEARCH STUDIES INTO AVN

One of the earliest reports of the necessity for less traumatic treatment of CDH to reduce treatment complications was that of Crego and Schwartzmann, who rejected the use of forceful manipulation and reduction along with prolonged plaster immobilization in the Lorenz "frog" position (49). They described their subsequent treatment of patients treated before 7 years of age with generally good to excellent results and no evidence of avascular necrosis (described as "Pertheslike findings"). Four patients were treated initially between 1 and 2 years of age, 70 between the ages of 2 and 6 years, and 4 between 6 and 7 years of age. Initial treatment involved a gentle closed reduction in which the manipulation was clearly described as being performed "with no more force than is required to demonstrate telescoping, the head of the femur is brought within the acetabular fossa." If that simple manipulation was ineffective, then either skin or skeletal traction was used for 2-4 weeks preliminary to reduction. Forceful manipulation under anesthesia was never used. In addition, once reduction had been obtained the head of the femur was held in the acetabulum by casting in internal rotation and abduction of the thigh but without using the extreme frog position. When this method did not lead to normal or near normal results, subsequent treatment involved a derotation femoral osteotomy and an acetabuloplasty. Studies by Salter et al. (264) and by Gage and Winter (86) showed the clinical benefits of changes in closed reduction management with both concluding that a program of prereduction traction, percutaneous release of the tightened adductor muscles, gentle reduction, and avoidance of the extreme Lorenz or Lange positions reduced the incidence of avascular necrosis. The Salter study showed a decrease in AVN in three consecutive 5-year periods from 30% of 66 hips to 15% of 97 hips to 5% following adoption of the "human" position of immobilization. Gage and Winter documented a diminution of AVN from 34.8% during the first 5 years of their study period to 4.5% in the last 5 years. The position of immobilization favored by Salter on the basis of clinical and experimental studies is referred to as the "human position," in which the hips are flexed greater than 100 ~ but with abduction much less than 90 ~ and limited to a safe zone (263, 264). This position for hip spica immobilization also was described by Fettweis in 1968, who referred to it as the sitting-squatting position (76, 77). The safe zone for position of abduction is determined as being between complete abduction in which the head is well-located in the acetabulum and adduction in which the head dislocates posteriorly. The position of appropriate reduction and immobilization is midway between these two points. This position allows for seating of the femoral head in the acetabulum

248

CHAPTER 3 ~

DevelopmentalDysplasia of the Hip

without approaching the extremes of abduction, which compress the blood supply. The stability of reduction was noted to be more dependent on hip flexion greater than 100 ~ than on extremes of abduction. Salter and colleagues also undertook a series of studies in the neonatal pig hip, which was felt to be closest to the human in terms of structure and blood supply (264). They assessed piglets 3 weeks of age, studying 20 pigs in each of 5 categories. In group I an adduction contracture was produced in the hip joints of infant pigs by maintaining them in the crosslegged position by an elastic tape for a period of 3 weeks. By the end of this period the adducted muscles were tight and an adduction contracture had been produced. Angiography was performed after the animals were sacrificed and there was no evidence of disturbance of the vascular pattern in any of the 20 femoral heads, indicating that the adduction contracture alone was not problematic for vascularity. In group II plaster of Paris hip spica casts were applied with the infant pigs under general anesthesia and the hip joints were placed into 80 ~ abduction, a position comparable to the frog position used in the human, and maintained for 2 weeks. Subsequent micro-opaque angiography revealed some degree of avascular pattern disturbance in 30 of 40 femoral heads, although the disturbance was considered to be slight. In group III the adduction contracture was produced followed by hip spica immobilization with the hips flexed at 90 ~ and abducted fully at 70 ~ This resulted in a severe disturbance of the vascular pattern in all 40 femoral heads on subsequent angiography. Salter et al. felt that the vascular block was in the vessels within the cartilage of the femoral head. In group IV adduction contractures were produced and released by adductor myotomy prior to hip spica immobilization in abduction, which led to no disturbance of the vascular pattern in any of the femoral heads. In group V, which was a long-term study, the adduction contractures were produced followed by casting in the maximally abducted position, but after immobilization for 2 weeks the animals were allowed to survive for 3-10 weeks. Radiographic studies only were done in this latter group, which revealed a "head within a head" finding that was interpreted to indicate that osteonecrosis had occurred at the time of the abduction casting procedure. This study led to the recommendation in clinical practice for continuous traction before closed reduction, subcutaneous adductor tenotomy at the time of reduction, and immobilization of the reduced hip in a position that did not compress the femoral head rigidly against the acetabulum. Salter et al. introduced the "human" position of immobilization, which involved marked hip flexion beyond 100 ~ with less than complete abduction. This position was also described by Fettweis (76). In a companion clinical study they felt that they had diminished the incidence of avascular necrosis by using these principles from 30% to 5%. The deleterious effects on femoral head circulation of hip positioning in maximal abduction and maximal internal rotation also

were shown experimentally by others. Law et al. used radioactive microspheres to assess femoral head vascularity in puppies (162). Maximal abduction to 90 ~ markedly decreased flow to the capital femoral epiphysis. Submaximal abduction and a position of 120 ~ flexion, 40 ~ abduction did not reduce blood. Another series of experiments by Schoenecker et al. in puppies using the hydrogen washout technique also showed that forced hip abduction and maximal internal rotation drastically reduced femoral head circulation, whereas immobilization in flexion resulted in the highest rate of femoral head blood flow (268, 269). Gage and Winter performed a landmark study critically reviewing 20 years of experience at their hospital (86). They assessed avascular necrosis as a complication of closed reduction of congenital hip dislocation between 1948 and 1967. Only completely dislocated hips were assessed. The incidence of AVN increased with the increasing age of the child at the time of initiation of therapy. In those within the first year of life, the incidence was 25% overall, second year 29.6%, third year 34.5%, and fourth or greater years 52.6%. Traction prior to reduction was helpful. In those who had no traction, the incidence of AVN, which generally was equally spread between partial and total, was 66.6%, but when traction was applied from 0 to 10 days, the incidence diminished dramatically to 28.6%. There really was no ehange in the incidence of AVN when traction was extended either from 10 to 20 days or for greater than 20 days. The position of immobilization also was noted to have an influence on the incidence of AVN. Both the Lorenz and Lange positions were deleterious in a large number of instances. During four consecutive 5-year periods studied, results improved on the basis of increasing utilization of prereduction traction, gentle closed reduction, and the less extreme positions of immobilization. In the first 5-year period from 1948 to 1952, the incidence of partial or total AVN was 47.8%, with subsequent decreases to 35.6%, 22.5%, and 17.4%. When total avascular necrosis was considered, the incidence declined from the first 5-year period rate of 34.8% to only 4.5% in the last 5-year period. The incidence of partial necrosis remained essentially the same across the four periods. The major deleterious effects of the Lorenz position of immobilization in 90 ~ hip flexion and 90 ~ hip abduction with rigid immobilization in a hip spica cast also were illustrated by Allen, who reviewed hips that were treated because of a diagnosis of hip dysplasia without actual dislocation based on the early assessment of hip radiographs with an increased acetabular index only (3). Once the importance of early diagnosis became accepted, large numbers of hips in newborn children were radiographed in some centers and the diagnosis of dysplasia was given to all those with more than 30 ~ measured for the acetabular angle. Allen was able to study 150 cases with a diagnosis of hip dysplasia from his institution. Seventy-seven cases were treated with the abduction Frejka pillow with none showing any evidence of avascular necrosis; 50 underwent no treatment and also showed no

SECTION Xll 9 Avascular Necrosis as a Complication o f Treatment

subsequent evidence of avascular necrosis, whereas 20 were treated in the cast with 14 showing avascular necrosis and only 6 being normal. Three patients were treated in an abduction apparatus, one of whom developed avascular necrosis. In some instances the opposite normal hip developed the AVN because it also had been immobilized in the hip spica cast. This work not only stressed the negative effects of the full flexion-full abduction position of casting but also showed the need to not overtreat by rigid methods what were essentially stable and only slightly dysplastic hips. Determination of the precise degree of abduction that is safe for blood flow and yet adequate for treatment of the dysplastic hip is difficult and in fact may not be completely attainable in all cases. Ramsey et al. have defined three zones of abduction in the infant hip (247). Mild abduction, in the "zone of re-dislocation," may result in failure of treatment. Excessive abduction, in the "zone of ischemia," interferes with vascular perfusion of the femoral head. Between these two zones lies the "safe zone," in which therapy is successful and not complicated by ischemia. The extent of the "safe zone" varies between infants. Another example of diminution in the rates of AVN by careful attention to the problem and utilization of different approaches is reported by Tonnis. He reported a rate of 92% AVN (when even the mildest forms were included) from 1945 to 1959 when the closed manual reduction and hip spica mobilization treatment of Lorenz had been used (313). A subsequent change in their clinic used the Pavlik harness and overhead traction, allowing for spontaneous reduction followed by manual reduction if that did not occur; in the period from 1956 to 1966 the percentage of necrosis with complete dislocations was still high but had diminished to 66%. From 1970 to 1973 manual reduction no longer was used and the method of reduction involved overhead traction, adductor tenotomy, and use of the Pavlik harness, which diminished the percentage of necrosis further to 32%. Other treatment modifications followed with more judicious use of open reduction, which did not have as high a rate of necrosis, and change in the position of immobilization postreduction to the "human position" described independently by Salter and by Fettweis in Germany at the same time. 3. AVASCULAR NECROSIS AND THE AGE AT

CLOSED REDUCTION An interesting observation initially made by Weiner et al. was that AVN in association with closed reduction even with the use of traction had a relatively high incidence in the first 3 months of life, a relatively low incidence between 4 and 12 months of age, and then a correspondingly higher incidence after 12 months of age (344). They noted an incidence of AVN of 14% from 0 to 3 months, 8% from 3 to 6 months, 6% from 6 to 12 months, 22% from 12 to 18 months, 44% from 18 to 24 months, and 53% from 24 to 36 months. Because the secondary ossification center forms around 4 months of age, it was felt that the center might have

249

a protective effect minimizing the likelihood of AVN. The rationale behind this is that, prior to formation of a secondary ossification center, the cartilage canals that provide vascularity to the epiphyseal cartilage of the femoral head are isolated within nonanastomosing vessel systems, whereas with the formation of the secondary ossification center anastomoses occur within the bony center. Therefore, damage to one set of vessels conceivably could be minimized by the shifting of flow through anastomotic channels. It is also possible that the bony center itself provides more stability to the head, although this appears unlikely. Yet another reason for the finding might be that the older the child, the more resistant the tissue to applied forces and the easier it is to control the amount of abduction. Regardless, the observation appears real and has been made subsequently by some but not all. Some have interpreted the finding to indicate that closed reduction should not be performed in the first 3 months of life. This appears far too radical a suggestion because it is abundantly clear that the best results are still obtained the earlier the reduction is performed. Suzuki and Yamamuro showed a relatively even rate of AVN in relation to the age treatment began in patients treated with the Pavlik harness. There actually were no cases of AVN in 18 patients treated in the first month of life, but after that results were similar: 15%, 2nd month; 17%, 3rd month; 21%, 4th month; 13%, 5th month; 20%, 6th month; 18%, 7th month. 4. PAVLIK: FUNCTIONAL METHOD OF TREATMENT

Pavlik had many interesting observations on the genesis of AVN in his report in which the "functional" method of treatment using the harness with stirrups (now known as the Pavlik harness) was first reported (222). Pavlik, working in the former Czechoslovakia, developed the functional harness treatment for all patients with joint dysplasias and dislocations beginning in 1951. The apparatus holds the hip and knee joints in flexion at about a right angle, and whereas it prevents the infant from extending the hip joint, all other motions of the hip are possible, including abduction, adduction, complete flexion, and internal and external rotation. The treatment is referred to as "functional" because active movement is the most important therapeutic factor in comparison to the previously used "passive-mechanical immobilization" of the affected hip after repositioning. Among the examples of the passive-mechanical method were the Lorenz forceful closed reduction hip spica and a method of Hanausek in which, with the benefit of a specially constructed apparatus, reduction was carried out slowly using the weight of the body. He reported a high incidence of femoral head necrosis in the passive-mechanical treatment, noting its occurrence as 2.94% of hypoplastic hips, 8.71% in subluxations, and 30% in dislocations. In dislocations treated by the method of Lorenz, the incidence reached as high as 60%. Pavlik reported on 1912 hip joints treated in 1424 patients with the functional method involving hypoplasia (640), subluxation (640), and luxation (632). He felt that the Frejka

250

CHAPTER 3 9 Developmental Dysplasia of the Hip

abduction splint offered insufficient support and a lack of functional mobility. He felt, in agreement with most others, that necrosis resulting from passive-mechanical methods occurred due to the persistent and rigid abduction and the elevated tone of the adductor muscles. He also noted that "I have never seen an aseptic necrosis of the femoral head in any untreated high dislocation," indicating the disorder to be iatrogenic. The common denominator in AVN was the "fixed pressure of the reduced femoral head against the floor of the acetabulum when there is no possibility of active movement in the hip joint. The increased tension in the adductors has an important influence that eventually affects the acetabulum." Pavlik's method promoted active motion from the beginning, which not only led to reduction but also promoted anatomical development. He did not specifically address the adductor tightness, feeling that the relaxation of tightness in the adductors occurred through active motion in the harness and was indeed one of the first signs of spontaneous reduction. When the adductors were relaxed, spontaneous reduction occurred during the active abduction movements of the hip. The reduction almost always was felt to maintain itself with very few instances of re-dislocation recurring. Spontaneous reduction could not be expected when the acetabulum was already filled with interposed tissue, and Pavlik advocated use of the functional treatment in children 8-9 weeks of age or younger. It could be used even as late as 6 months of age, though the child would have to be closely followed in relation to the result. He summarized by indicating that he tended to begin with functional treatment up to 1 year of age, although many in the later months subsequently would require open reduction. Of the 632 dislocated hips, 531 (84.1%) reduced with the closed abduction treatment and not a single case of femoral head necrosis occurred. In 101 hips (15.98%) reduction did not occur. These hips then were treated by the passive-mechanical method involving manipulation under anesthesia and immobilization and the rate of necrosis rose dramatically, occurring in 18 hips for an almost 18% rate. In a large number of patients, therefore, the AVN was noted only in cases treated by the passive-mechanical reduction and hip joint immobilization. 5. OTHER STUDIES OF AVN WITH TREATMENT OF C D H - D D H ; VALUE OF PREREDUCTION TRACTION

A large study on AVN of the femoral head in 254 hips treated for CDH conservatively was performed by Gregosiwicz and Wosko (99) from 5 orthopedic departments in Poland. The number of hips involved was 1211, of which 254 had some form of AVN (21%). The nature of the involvement was subdivided further using the Kalamchi and MacEwen categorization into 28% grade I, 6% Grade II, 41% Grade III, and 5% Grade IV (141). The highest risk findings for the development of AVN involved treatment in patients less than 6 months of age, those with severe dysplasia of the acetabulum, those having application of the Frejka

abduction pillow prior to appropriate repositioning of the head, immediate reduction without preparatory traction, and wide casting in abduction in the frog leg position. Conversely, the risk of AVN diminishes statistically in those who are aged 6-18 months, have mild dysplasia of the acetabulum, have prereduction traction, are repositioned in association with an adductor tenotomy, and are immobilized in a more physiologic position. Another major study showing a positive relationship of premanipulation traction to decreasing AVN at the femoral head in CDH was reported by Weiner et al. (344). They assessed 319 congenitally dislocated hips in which premanipulation skin traction was used with the hips slightly flexed and abducted not more than 45 ~. The force of traction varied depending on the size of the patient, but seldom exceeded 1 kg. All of the hips underwent premanipulation traction. When the duration of traction was from 0 to 6 days, the incidence of avascular necrosis was 43%. This diminished to 20% with traction from 7 to 13 days, 16% with traction from 14 to 20 days, and a low of 6% with traction from 21 to 27 days. Lengthier times showed a slightly increased rate. For the entire series, the incidence of avascular necrosis was 18% (59/319). There were large numbers of hips assessed in each group, ranging from 54 in the 0 - 6 day group to 93 in the 21-27 day group. The authors concluded that the duration of premanipulation traction was of great importance, with the lowest incidence occurring in those having 3 - 4 weeks of traction. They also assessed the incidence of avascular necrosis in relation to the age at which treatment was initiated. Here too their findings were consistent with those of others. There was a relatively high incidence of 14% AVN in those treated from 0 to 3 months of age when there was no secondary ossification center present. From 3 to 6 and from 6 to 12 months, the incidence of AVN decreased dramatically to 8% and then to 6%. Following 1 year of age, the incidence again increased in association with the more difficult hip displacement. In those treated from 12 to 18 months the incidence was 22%, from 18 to 24 months it was 44%, and from 24 to 36 months it was 53 %. The poorer prognosis for children more than 1 year of age when treatment was begun was demonstrated clearly and it continually worsened with time. The slightly higher incidence of AVN in the first 3 months was noted and related to the specific nature of the blood supply via the cartilage canals. During the time frame prior to the appearance of the femoral ossific center, avascular changes could not be identified even when they were severe. Tonnis actually has concluded that preoperative or prereduction adductor tenotomy has a detrimental effect because it is associated with an increased rate of AVN. Without tenotomy only a 2.7% rate was noted, with it 19.4%. On the basis of that subset study it was no longer performed with closed reduction (319). Quinn et al. specifically studied preliminary traction in the treatment of developmental dislocation of the hip (244). They retrospectively analyzed the

SECTION Xll ~ Avascular Necrosis as a Complication of Treatment results of traction treatment on 90 dislocated hips in 72 patients. Traction was continued for 3 weeks, after which closed reduction was performed. There was no significant difference in either the rate of successful closed reduction or the incidence of avascular necrosis compared to more recently published series in which preliminary traction was not used. They were unable to identify a subgroup of patients that clearly benefited from the use of traction in the treatment of DDH. Similar results were reported by Kahle et al. in relation to the value of preliminary traction (139). Thomas et al. noted no difference in the rate of AVN with open reduction in those who had preoperative traction compared with those who did not (309). A later study from the same institution assessing closed reductions found no beneficial effect from prereduction traction or adductor tenotomy (29). Roose et al. reviewed their experience with open reduction in 26 hips and showed no AVN in either the half who received preliminary traction or the other half who had no preliminary traction (255). Weinstein (346) and Weinstein and Ponseti (347) noted no difference in AVN in those who had undergone open or closed reduction with or without preliminary traction. Thomas et al. studied avascular necrosis after open reduction for CDH and compared it with the results from the same institution after closed reduction (309). AVN after open reduction of 87 dislocated hips occurred in 37% of cases. The open reduction itself was not considered to be the major causative factor because a similar percentage of patients developed AVN following closed reduction. Powell et al. studied 49 hips with CDH undergoing anterior open reduction after traction, some of whom had femoral osteotomy and some of whom did not (238). The incidence of AVN after open reduction alone was 25% (4/16). After open reduction and varus derotation osteotomy it was 28% (5/18); and after open reduction and innominate osteotomy it was 46.7% (7/15).

6. MILDER SEQUELAE OF AVN: COXA MAGNA, PREMATURE PHYSEAL CLOSURE, AND GREATER TROCHANTERIC OVERGROWTH One of the manifestations of an avascular event is the subsequent finding of a coxa magna deformity following surgical treatment of a CDH, although many instances of coxa magna appear to be due to hypervascularity alone with no evidence of prior AVN. Gamble et al. noted coxa magna, defined as a femoral head with a diameter 15% greater than the opposite side, in 33% of 64 dislocated hips at long-term follow-up (89). The factors correlating with coxa magna were femoral osteotomy (100%), open reduction (75%), and operation at a younger age (15.6 versus 35.8 months). Imatani et al. found coxa magna with femoral head size more than 20% greater than the opposite normal side in 35% of open reduction procedures. Excision of the limbus magnified both the degree and the frequency of coxa magna (130). Another of the structural sequelae of AVN is premature physeal closure, leading to a short neck. In its more severe

251

manifestations, this will be associated with relative overgrowth of the greater trochanter and a Trendelenburg gait. Stevens and Coleman assessed treatment for this particular complication (290). Most of their patients (30) developed a disorder following treatment of CDH, although 11 were the sequel to Legg-Perthes disease and 3 had other causes. For patients 8 years of age or less, epiphyseal arrest of the greater trochanter was effective, whereas those presenting with an established Trendelenburg limp and those 9 years of age or older were best treated by distal and lateral transfer of the greater trochanter. Relative trochanteric overgrowth in patients suffering from ischemic necrosis following treatment for CDH also was studied by Iwersen et al. (134). They documented the extent of the short neck or the relative trochanteric overgrowth by using the articulotrochanteric distance (ATD). The ATD in 29 patients with a Trendelenburg gait also was measured. The group was limited to 39 patients who developed ischemic necrosis of the femoral head in treatment for CDH but who did not have any femoral surgery. They also identified 29 patients with a positive Trendelenburg test. Children with an ATD equal to or less than 0 mm were most likely to have Trendelenburg gait. The mean ATD for the affected hip in their study was - 0 . 8 mm and for the normal hip was 21.7 mm. The worse the ischemic necrosis, the lower the ATD with the patients categorized by the Kalamchi group. Many studies in the literature also indicate that the Trendelenburg correlates well with the ATD. Edgren reported 25 Perthes patients with a positive Trendelenburg test, and of these the ATD was 0 to - 9 mm in 15, + 1 to +5 mm in 8, and greater than 8 mm in only 2 (134). Langenskiold and Salinius noted that the Trendelenburg test was positive in more than one-half of the cases in which the ATD was reduced to - 5 mm or less (134). Langenskiold reported the normal mean ATD for females aged 5-13 years at 16 + 3.6 mm and for males as 23 ___ 4 mm. Many authors recommend trochanteric surgery when the ATD approaches 0 mm.

E. Classification of Patterns of Avascular Necrosis Following Treatment of Developmental Dysplasia of the Hip 1. MASSIE Massie formulated one of the earliest categorizations of the sequelae of AVN, dividing the radiographic findings into three groups. Grade 1: This was the mildest of the groups (188). There was transient fragmentation of a portion of the epiphysis (secondary ossification center) followed by fragmentation and rapid repair (generally within 1 year) and resumption of a normally developing epiphysis after healing. At skeletal maturity, the head was of virtually normal shape and size. Grade 2 comprised 80% of those involved with AVN changes. There was initial fragmentation of the secondary ossification center but without the increased density seen in

252

CHAPTER 3 ~

DevelopmentalDysplasia of the Hip

Legg-Perthes disease. There was occasional irregularity of the epiphyseal line particularly centrally and flattening of the secondary ossification center, often associated with lateral extension of the secondary center bone. The metaphysis was broadened and the neck shortened, although characteristics seen in Legg-Perthes disease such as metaphyseal osteoporosis and cyst formation rarely were seen. There also was occasional premature physeal fusion. Grade 3 developed quickly and was markedly similar to the developmental findings in Legg-Perthes disease. Among the changes noted were early and severe convexity of the metaphyseal edge, rapid dissolution of the ossific center of the physis, slow regeneration of the secondary ossification center bone, progressive flattening of the metaphysis, subluxation of the proximal end of the femur, marked shortening of the femoral neck, and coxa magna deformity. Not all changes were present in each patient. Some with abnormal epiphyseal changes lacked the fragmentation of the secondary ossification center. There were two variants. Type 1 included hips with no fragmentation of the secondary center, convexity of the metaphyseal edge of the physis, and shortening of the neck. In type 2, the changes were similar to those in type 1 but radiographs revealed deformity resulting from molding of the femoral head against an oblique acetabular roof along with moderate to severe subluxation. Five major classifications have evolved from long-term studies of avascular necrosis, complicating the treatment of hip dislocation. 2. KALAMCHI AND MACEWEN

Kalamchi and MacEwen reviewed 119 patients with avascular necrosis and classified four groups (141). Group I: changes affecting the ossific nucleus. Changes are characterized by either delay in the appearance of the ossific nucleus or fragmentation of the secondary center. The overall longterm development, however, is reasonably good. Group II: lateral physeal damage. There is involvement of the secondary ossification center but also damage to the lateral part of the physis. There is delayed growth in the lateral segment of the femoral neck, premature lateral epiphyseal closure, and lateral tilt of the head into valgus. This worsens the uncovering of the femoral head in relation to the acetabulum. Group III: central physeal damage produces early changes in the ossific nucleus, but damage to the growth plate is more severe, is centrally located, and causes symmetrical retardation or cessation of growth of the entire femoral neck to produce a coxa vara deformity. Group IV: total damage to the head and physis. This is the most severe variant in which there is not only a delay and irregularity of the ossification of the secondary center but also a marked coxa vara and early femoral head irregularity, flattening, and eventual coxa magna. The femoral neck is short and widened. Age was related to the severity of avascular necrosis. The avascular necrosis was most severe in the group treated between birth and 6 months of age, with fewer and less severe problems after 6 months.

3. BUCHOLZ AND OGDEN

Bucholz and Ogden assessed the sequelae of avascular necrosis and defined four types (31). Type I ischemic necrosis was due to vascular occlusion at the circumferential ring of the base of the neck and occurred both medially and laterally, leading to a fragmentation of the ossific nucleus and a coxa vara appearance. Type II ischemic necrosis caused damage primarily laterally leading to irregularity of the secondary ossification center laterally and early lateral closure of the physis, which caused coxa vara but with a lateral tilt of the head. Type III damage was most severe with changes at lateral posterior, lateral anterior, and medial vessels. There was significant coxa vara with markedly delayed appearance of the secondary ossification center and a horizontal growth plate. Type IV ischemic necrosis primarily was medial leading to a medial femoral neck physeal arrest and tilting into varus of the head and neck, but relatively little shortening. 4. ROBERT AND SERINGE

Robert and Seringe reviewed growth problems of the proximal femur after treatment for congenital hip dislocation, selecting 100 hips that showed growth abnormalities due to AVN out of approximately 2500 treated cases (253). Their classification was based on 50 well-documented hips (Fig. 17). Group I involved irregular formation of the secondary ossification center alone and proceeded to good development with minimal sequelae. In group II the abnormalities were concentrated laterally involving the secondary ossification center, the lateral epiphyseal and physeal cartilage, and the lateral and proximal metaphyseal regions. This led to diminished growth laterally, a horizontal growth plate, a tendency to lateral subluxation of the femoral head, and relatively minimal shortening. Group III abnormalities involved global or total epiphyseal and proximal metaphyseal developmental changes, which were reflected in a marked coxa vara with an almost vertical growth plate. Group IV involved avascular necrosis medially of the secondary ossification center, the medial epiphyseal and growth plate cartilage, and the proximal medial metaphyseal regions. This also led to major deformity with a marked coxa vara, a vertical growth plate even more extensive than that noted in group III, and severe shortening. The final group V lesion was rare but involved isolated metaphyseal and adjacent physeal lesions, which also led to a mild coxa rata. 5. TONNIS AND KUHLMANN

Tonnis and Kuhlmann (315) reported four grades of changes as follows: Grade 1: the mildest grade of pathology in which the structure of the capital ossific nucleus is slightly granular and somewhat irregular, with margins somewhat indistinct. Generally, this condition is self-limiting and without sequelae. Grade 2: The margins of the ossific nucleus are more irregular, and its structure shows greater mottling and granularity than in grade 1 cases. Cystic changes may be

SECTION XII

Groupe

![

!

!1

~--.,, § L § ~"

( v/

f II

6. HIROHASHI E T A L . The classification of Hirohashi et al. is shown in Table IV (118). Figs. 18A and 18B show Siffert's outline of growth in the normal hip (Fig. 18A) and following specific localization of AVN changes (Fig. 18B). Figure 18C shows hip appearances following childhood AVN.

:

I

!I.

j/, ~

li

~"

- m~taohysaire EEpiphyso ~- externe ~"

F. Avascular Necrosis Associated with Immobilization Devices Other Than a Hip Spica

s §247247 I

i /-" ((

!

I I

I

!

/~ I

Epiphyso- m6taphysai

g,oba,

1/

\

Epiph~o - m6taphysair

interne

I

/

253

metaphyseal involvement is not apparent until valgus or varus type growth disturbances and shortening of the femoral neck have occurred.

Epipho~saire

..-...'Ji'~'K)

![

Avascular Necrosis as a Complication of Treatment

Classification

~=//~r

+#§

~

I

I

----

\

M.taphysai~e §

F I G U R E 17 Classification of Robert and Seringe of the negative sequelae of AVN of the hip. These include (from top) isolated secondary ossification center (epiphyseal) involvement, lateral epiphyseal-metaphyseal involvement, total epiphyseal-metaphyseal involvement, medial epiphysealmetaphyseal involvement, and metaphyseal (physeal-metaphyseal) involvement. The corresponding bony deformities are shown at right. [Reprinted from Robert and Seringe (1982), Rev. Chir. Orthop. 68:425-439, 9 Masson Editeur, with permission.]

present within the ossific nucleus. There may be punchedout defects that appear as a small lateral notch in the surface of the head. These changes tend to regress with time, sometimes leaving a mild flattening of the head. Grade 3: The ossific nucleus as a whole is fragmented or appears as a flat strip. Very small ossific nuclei may completely disintegrate. This grade may develop even before the ossific nucleus has appeared, in which case the necrosis will not become manifest for some months. Deformity of the femoral head and neck is apparent initially but may resolve if the physis is undamaged. Grade 4: There is involvement of the physis leading to serious growth impairment. Irregularities may be seen along both edges of the physis, though in some cases

Concern with the high incidence of avascular necrosis associated with the use of hip spicas and the earlier diagnosis of congenital hip dislocation led to the use of less rigid devices for immobilizing the hips in the reduced position. The Frejka abduction pillow, the Pavlik harness, and the Denis Browne abduction splint were introduced in Europe and subsequently were widely used (9, 64, 83, 129, 171,222, 235). These devices, however, are not necessarily fully protective against AVN. It is not solely the immobilization device used but associated factors such as patient age, degree of abduction, and relationship to adductor muscle tightness that predispose one to AVN. 1. FREJKA ABDUCTION PILLOW Long-term studies of the Frejka abduction pillow showed a fairly high incidence of avascular necrosis. Two large series in which congenital hip dislocation was treated by the Frejka abduction pillow show a 14% incidence of AVN in 113 out of 830 hips and a 31% incidence in 671 hips. In the former study the mild/moderate/severe distribution of AVN was 49%/14%/37%. In a smaller series of 84 dislocated hips, AVN occurred in 7% (9). Ilfeld and Makin reported on 7 cases of AVN with the Frejka pillow splint and recommended that it be altered to allow 90 ~ hip flexion but only 60 ~ abduction (129). 2. PAVLIK HARNESS

Reports on the Pavlik harness have shown a much wider range of problems, with some studies reporting a 0 incidence and others a generally lower but still significantly high incidence with values of 2.4%, 6% (9), 9%, and 11.2% (319). Grill et al. (100) in a series of 3611 developmental dysplasia hips noted a 2.4% incidence of AVN with Pavlik harness treatment, and in a literature review encompassing 3505 hips, the incidence was 2.5%. Touzet et al. reported a 5% incidence in 300 cases (322). Suzuki and Yamamuro reviewed results with the Pavlik harness used for treatment beginning from birth to 7 months of age in 220 reduced hips. The overall incidence of avascular

254

CHAPTER 3 " Developmental Dysplasia of the Hip

TABLE IV The Classification of Hirohashi el: a/. o f Avascular Necrosis o f Proximal Femur Following Treatment of Developmental Dysplasia o f the Hip a Findings at hip reduction Grade of necrosis Grade I (mild)

Partial damage

Grade IIa (anterolateral or lateral)

Grade lib (posteromedial or posterior)

Total damage

Grade III (severe)

Ossific nucleus present or ossific nucleus not yet visible 1. Appearance is on time or delayed, but no later than age 18 months 2. Appears subdivided into two or more parts 1. Appearance is on time or earlier than age 18 months in the posteromedial area 2. Growth disturbance of metaphysis laterally or anterolaterally 1. Appearance is on time or delayed but earlier than 18 months in anteromedial area 2. Growth disturbances of metaphysis posteromedially or only posteriorly Appearance delayed past 24 months of age

Ossific nucleus present 1. Temporary irregularity of border or fragmentation 2. Growth arrest Partial resorption with rest of nucleus showing evidence of regenerative capacity

Posteromedial or posterior resorption

Complete resorption of an existing nucleus

aModified from (118, 315).

necrosis was 16%. They concluded that the more severe the dislocation, the higher the rate of both failed reduction and avascular necrosis. Iwasaki reported an incidence of AVN of 7% in children treated as outpatients, but in a subset of more difficult dislocations treated as inpatients an AVN rate of 28% was noted in an early series.

3. DENIS BROWNE SPLINT Pool et al. documented a 2.5% incidence of AVN in 238 subluxed or dislocated hips treated with the Denis Browne splint alone following stable reductions and a 60.6% incidence in 33 hips treated with the splint after tenotomy and spica treatment initially (235). G. More Recent Reports of the Incidence of AVN in D e v e l o p m e n t a l Dysplasia of the Hip In spite of the use of prereduction traction, adductor tenotomy, gentle closed reduction, and positions of immobilization less extreme than those used previously, a disturbing continuation of the occurrence of avascular necrosis is evident from more recent publications. The large majority of these cover cases treated between 1975 and 1990, which represents the time frame during which the preceding clinical and experimental observations on the presumed causes of AVN had been published. Despite this, three recent studies report incidences of AVN in hip dislocations treated by closed reduction of 47% in 210 hips (29), 26% in 72 hips (80), and 45%

in 42 hips in a subset of male patients. It is becoming evident that, in patients who are not diagnosed in the first 6 - 8 weeks or who do not respond with an excellent result to the treatment in a Pavlik harness in the first 8-12 weeks of life, the subsequent more vigorous treatments continue to show a high level of AVN. What this would appear to represent is separation of a more benign form of DDH with a loose capsule, which responds to the relatively simple Pavlik harness treatment, and those having an abnormality that is greater structurally, shows more resistance to simple treatment, and has a high incidence of complication. AVN is most common in the more difficult cases that are detected relatively late, for example, after 3 months of age when secondary adaptive changes are beginning to occur and must also be treated, or in those that do not respond readily to the simpler abduction splinting methods of Pavlik or Frejka et al. It is those hips that require prereduction traction, adductor release, closed reduction under general anesthesia, immobilization in hip spica, and any of the various open reductive procedures that have a much higher incidence of AVN. The relatively uncomplicated DDH diagnosed in the perinatal period in which there is a dislocatable hip that repositions well with flexion and gentle abduction tends to have a low incidence of AVN. Several features pertaining to AVN complicating the treatment of congenital or developmental dysplasia of the hip can be recognized based on the extensive studies reported. As each of the five classifications points out, there is a considerable range of involvement and it really is only

SECTION Xll ~ A v a s c u l a r N e c r o s i s a s a C o m p l i c a t i o n o f T r e a t m e n t

F I G U R E 18 (A) Illustration by Siffert showing normal growth in the human hip. TRC; triradiate cartilage; LGP; longitudinal growth rate; TGP; trochanteric growth plate; and F-I, femoral neck isthmus. (B) Growth changes of the femoral head-neck region following physeal avascular necrosis. With global involvement at sites A, B, and C or with central involvement at B the femoral headneck junction has marked growth retardation but neither varus or valgus tilting of the head itself. When there is lateral growth cessation, there is diminution in height in the head-neck complex but relative valgus tilting of the head in relation to the acetabulum. When localization is at point C, there is head-neck growth diminution in length and there is tilting of the head. In each of these examples greater trochanteric growth continues. (C) Clinical examples of avascular necrosis following treatment of CDH-DDH are shown. [Parts A and B reprinted from Siffert, R. S. (1981). Clin. Orthop. Rel. Res. 160:14-29, 9 Lippincott Williams & Wilkins, with permission.]

255

256

CHAPTER 3 ~ Developmental Dysplasia of the Hip

those in the more serious or advanced categories that lead to clinical problems. In addition, because the major instances occur in the first few years of life, the remaining growth potential is extensive and many of the abnormalities are corrected with growth such as would be the situation in a LeggPerthes disorder, which occurs in a very young patient of 3-4 years of age. On the other hand, in cases that are severe, the fact that there are so many years of growth remaining can raise some problems to major clinical importance. The radiographic changes are quite consistent with those seen in Legg-Perthes disease. It is important to assess the nature of the avascular necrosis in each of the various reports rather than relying on only the number or percentage affected. For example, in the report by Pool et al. there was a 60.6% incidence of avascular necrosis in 20 of 33 patients treated with tenotomy and plaster spica followed by an abduction splint (235). Close assessment of the analysis, however, indicates that 17 of the 20 patients had a very mild grade I AVN using the Kalamchi and MacEwen classification. On the other hand, the dividing line between a mild and more severe AVN is uncertain, and thus any AVN bears with it the likelihood that with slightly greater pressure it would have been much more serious. The various papers also stress two other points. Most cases and indeed the more severe cases occur after relatively late diagnosis, which generally means 3 months of age or older because the secondary sequelae are greater and the treatment modalities correspondingly more harsh. When the overall spectrum is assessed, however, there is an increased percentage incidence of AVN when treatment is performed within the first 6 months of life. This almost certainly relates to the unique pattern of blood supply to the proximal femoral capital epiphysis prior to the formation of the secondary ossification center in which nutrition is provided by the cartilage canals, which are end vessels. Thus, if there is any AVN either throughout the entire epiphysis or in certain segments, there is little to no availability from the adjacent vasculature in the cartilage canals to become involved in the nutrition process. Kalamchi and MacEwen made several valuable clinical observations of the sequelae of AVN in relation to early treatment (141). They grouped their patients into four groups, the details of which were described earlier. In their assessment of limb length discrepancy secondary to the vascular insult, there was considerable variation. Limb lengths were equal in group I, whereas the difference averaged 2.5 cm in group II, 5.0 cm in group III, and greater than 7 cm in group IV. They reviewed 68 hips at skeletal maturity in relation to several clinical and radiographic features. Following the AVN they noted only 30 hips with a good grading, 20 were rated as fair, and 18 were poor. Some of the longterm problems were secondary to dysplasia due to imperfect reduction, but nevertheless the association of avascular necrosis with a tendency to meaningful downgrading of the end result is clear. Their classification, however, was based solely on the changes attributed to vascular insult and its

effect on the femoral head and physis and did not relate to the secondary mechanical problems resulting from residual acetabular dysplasia. Their study also pointed out that damage can remain dormant for many years such that the patients must be followed to the end of skeletal growth. They documented lateral physeal arrest leading to a femoral neck in valgus angulation in 35% of the patients. The most severe form of avascular necrosis came in the group having treatment from birth to 6 months of age. The surgical treatment of the established AVN deformity adhered to general treatments of abnormal hip pathology. The varus deformation was approached with valgus osteotomy, and relative overgrowth of the trochanter was treated either by physeal arrest if sufficient growth.was remaining or by distal transposition of the greater trochanter at skeletal maturity. Burgos-Flores studied 104 cases of unilateral congenital dislocation of the hip treated at his institution between 1977 and 1988. The patients had an average age of 12 months (range = 4-24) (33). The incidence of avascular necrosis was 37%. This relatively high number occurred even with the use of traction for an average of 5 weeks, an arthrogram posttraction under general anesthesia at which time stability and reduction were assessed, a high incidence of percutaneous adductor tenotomy, and, if closed reduction was unattainable, repositioning with open reduction with or without femoral and iliac osteotomies. Several variables were assessed in relation to the AVN. In those treated initially under 7 months of age, the incidence of AVN was 52%, and in those treated initially over 7 months of age, the incidence had decreased 30%. The Tonnis degree of AVN was 4 in 64% and 2 and 3 in 29%. When no adductor tenotomy was performed the AVN rate was 48%, which decreased dramatically to 16% when tenotomy was performed. The AVN rate in open reduction was 61% and in closed reduction 33%. These numbers denote the relatively higher occurrence at younger ages and the importance of adductor tenotomy. As treatment increases in degree of complexity, the secondary problems increase as well. Smith et al. noted an AVN rate of 20.6% (14 of 68 affected hips) treated by closed reduction and hip spica casting between 1984 and 1990 (283). Race and Herring compared CDH in those with and without AVN (245). The age of onset of treatment was younger in the 20 patients with AVN at 9.3 months than in the 39 patients without in which the average onset of treatment was 11.2 months. Prereduction traction was used in 40% of those developing AVN, whereas 72% having prereduction traction did not develop AVN. Similarly, adductor tenotomy was used in only 30% of those developing AVN, whereas 54% of those without AVN had the procedure. Extreme positions of abduction immobilization were used in 95% of those developing AVN, but in only 25% of those not developing AVN. The quality of the reduction also played a major role. The good/acceptable/poor reduction result in those without AVN was 18/11/5 hips, whereas in those with AVN the re-

SECTION XII ~ Avascular Necrosis as a Complication o f Treatment

sults were 6/1/11. Their recommended position of immobilization in cast after adductor tenotomy and reduction was one of greater than 90 ~ hip flexion, 30-40 ~ abduction, and neutral rotation. Gruel et al. noted a high rate of AVN of 48% (of 27 hips) in teratologic hip dislocations (101). Fogarty and Accardo showed the clear relation of increased abduction casting position to AVN (79). They assessed 222 congenitally dislocated hips that had been abducted in one of two fashions prior to closed reduction under general anesthesia. In both treatment groups, the hips were in 90 ~ flexion during the traction process. The patient was placed in Gallow's traction and the legs slowly separated over a period of 10 days until each hip had been abducted to approximately 90 ~ Since 1975, a similar method of traction was used but the hips had been abducted no more than 60 ~. In those in group I (abducted up to 90 ~ the incidence of total AVN was 17%, and in group II (abducted to not more than 60 ~ the incidence of total AVN was only 9%. Westin et al. reported an incidence of total AVN in 24% (48 of 209 patients) (349). In their opinion, prereduction traction and adductor tenotomy did not prevent avascular necrosis, and the major causative feature was abduction of the hip in the "frog" or markedly abducted posture. Tonnis reviewed two collective studies on CDH from a number of hospitals (317). One study compiled data on 730 dislocated hip joints treated by open reduction and the second study collected data on 4357 hip joints from Austria, Switzerland, and Germany treated with femoral and acetabular osteotomies. Preliminary traction was not found to influence significantly the rate of ischemic necrosis in the collective studies, but shortening osteotomies with the open reductions did reduce the AVN rate to 5.5%. Release of the iliopsoas and rectus femoris tendons also was felt to diminish pressure on the femoral head vessels. In the first study group, the rate of ischemic necrosis in open reductions was 8.2% for anterolateral approaches, 9.6% for inguinal, and 16.7% for Ludloff's medical approach. Open reduction with a simultaneous Salter osteotomy or acetabuloplasty increased the AVN rate to 10.3% and an associated varus osteotomy to 22.2%. The rate of AVN with open reduction alone was 8.4%. In an additional report on 3316 hip joints from a 20-hospital study, group al methods using the Lorenz position of hip spica immobilization had the highest incidence of necrosis, averaging 27% (319). The Lange position had 17% AVN, whereas the Fettweis ("human," Salter position) showed diminution to 2-5.5%. The feeling that the single most important factor leading to AVN was the extent of hip abduction was documented by Tonnis in remarkably clear fashion: 30-45 ~ abduction, AVN 2.5%, 46-50 ~ 4.9%; 51-60 ~ 8.7 %; 60 ~+, 16.7 %. The course of AVN complicating treatment for congenital dislocation of the hip was reviewed in 58 hips by Fisher and Cary (78). Common features in those developing AVN were inadequate prereduction traction, no adductor tenotomy, a

257

closed reduction, and immobilization of the hips either in very wide abduction or in extension and internal rotation. Very young children with no proximal femoral ossification center appeared to be particularly at risk. Brougham et al. assessed the incidence of AVN following closed reduction in 210 hips (29). Of these, 99 (47%) had some evidence of avascular necrosis, of which 81 were total and 18 partial. Contrary to previous reports, they felt that the incidence was not influenced by the age at reduction, the duration of traction, or the use of adductor tenotomy. They made the observation that the extreme ranges of AVN reported from 0% by Crego and Schwartzmann (49) to 67% by Esteve (65) was as indicative of the detail in assessment as it was of the occurrence of the disorder. No particular reasons were given to explain why adductor tenotomy and prereduction traction, felt to be so important by others, had little effect in the large series. Bensahel et al. reported 35 cases of vascular disorders of the proximal femur out of a total of 1500 cases of hip dislocation corresponding to an incidence of only 3%, although they eliminated from the study hips presenting a slight delay in epiphyseal maturation and cases in which radiographs showed discrete irregularities of the capital epiphysis that went on to excellent resolution (16). The extent of AVN in relation to specific treatments over several decades as reviewed primarily from European centers has been presented in articles by Sylkin (300) and Weber and Morgenthaler (342). Sylkin (300) showed the very high incidence of AVN reported in 19 papers reviewing the Lorenz technique; many studies ranged from incidences of 20 to 70%. Subsets of groups provided interesting information. In a study by Hohmann of 2358 cases, AVN following reduction in the first year of life was 30%, in the 2nd and 3rd years it was 24%, and after 3 years of age it was 10%. Krotschek showed (522 cases) how subluxation had slightly fewer incidences than dislocation (although all were very high) with AVN in subluxation-dislocation at 56.7% (67.7% in the 2nd and 3rd years of life, and 47.3%-61.5% after 3 years). Kaiser showed an overall incidence of 46.7% in 704 cases; 67.5% with treatment in the first 5 months of life, 57.3% from 6 to 10 months, 32.4% from 11 to 18 months, and 49.3% from 19 to 24 months. Hermann showed an overall rate of 76.9% (458 cases), with 27.2% in dysplastic hips, 31.1% in subluxation, and 92.4% in full dislocations. Other large series included Huber, 22.5% of 1073 cases, Becker, 68% of 340 cases, and Lange, 20-40% of 2200 cases. An additional 20 papers using variable techniques to specifically diminish AVN showed effectiveness. The functional harness treatments (including Pavlik, von Rosen, and other splints from Europe) greatly lowered the rate of AVN yet again from 1% to 22% in 17 papers. The long-term study by Fujioka et aL of Pavlik harness use even with persisting malreduction showed a 22% rate of AVN, although less rigid tightening subsequently was adopted (83). Weber and Morgenthaler (342) showed an AVN rate of 0-7.4% in 8 studies using the

258

CHAPTER 3 9 Developmental Dysplasia of the Hip

Fettweis cast ("human position"), with rates of 3.4% and 4.2% in the two largest studies of 388 and 174 patients, respectively. In a detailed review of multiple papers presenting results of differing immobilization splints, rates of AVN with the Pavlik harness ranged from 0.9% (Pavlik in 1912 cases) to 16% (Suzuki in 270 cases), with other large groups showing 1.7% (574 cases), 2% (520 cases), and 3% (1169 cases). Rates with the von Rosen-Barlow type splints in two of the largest studies were 0.2% (544 cases) and 0.4% (449 cases). The functional splinting series invariably include large numbers of patients with the relatively mild newborn dislocatability variants.

H. Long-Term Results Following AVN in Childhood CDH-DDH Virtually all long-term studies document considerable hip problems in adult life particularly following moderate and severe childhood AVN secondary to treatment of hip dysplasia. Cooperman et al. performed a long-term study of 30 congenitally dislocated hips in which AVN developed after closed reduction (47). The study assessed patients at an average of 39 years from the time of reduction; 24 of the 30 hips had moderate or severe osteoarthritis and 22 of the 25 patients had significant pain or loss of function or both by the time they were 42 years of age. The deformities caused by the AVN that led to subsequent osteoarthritis included loss of sphericity of the femoral head, lateral and proximal subluxation of the head, irregularity of the medial part of the femoral head, and acetabular dysplasia. All of the patients had been treated by closed reduction and hip spica cast immobilization. In the group of 109 patients studied originally, 54 (50%) developed AVN. These studies subsequently included only those who had not had major hip surgery in the interim and who could be located. At final follow-up, 29 of the 30 hips with AVN had radiographic evidence of osteoarthritis. Many patients with AVN had an associated subluxation of the hip during the childhood years, which itself led to the development of acetabular dysplasia and osteoarthritis. C. L. Thomas et al. studied 53 congenitally dislocated hips that developed AVN (308). They utilized the BucholzOgden classification, which defined four types of disorder. In the mildest type 1 variant, significant growth disturbances requiting surgical treatment were not observed. In the types 2, 3, and 4 disorders, however, progressively worse involvement led to a greater need for surgical intervention. Clinical concerns were limb length discrepancy, increased height of the greater trochanter relative to the femoral head and neck, lateral subluxation of the head, and acetabular dysplasia. Appropriate surgical treatments followed from the clinical findings and involved lower extremity length equalization, trochanter epiphysiodesis, and femoral or acetabular osteotomy as needed. In another study by I. H. Thomas et al. of 87 dislocated hips treated by open reduction, 32 developed AVN (37%) and of these only 45% had a good late result

(309). Robinson and Shannon studied 51 patients who developed AVN during treatment for CDH (254). A subset of 39 patients was followed for a mean of 23.5 years. The greater the involvement by AVN, the worse the eventual outcome. Particularly problematic was residual subluxation of the hips. Intervention of all types met with limited success such that the key to obtaining better long-term results was prevention of the AVN initially. All patients had undergone closed reduction and casting in the flexed and abducted position at a mean age initially of 18 months. Subsequent management was complicated due to the growth deformities caused by the AVN and the fact that most hips were not fully relocated, leading to the need to address both the lateral subluxation of the femur and the acetabular dysplasia. Keret and MacEwen noted a tendency for some deformities to worsen with growth and cautioned on the need to follow patients to skeletal maturity (146). Negative sequelae included a crescent-shaped epiphysis, medial bowing of the femoral neck (medial physeal growth decrease), lateral tilting of the capital epiphysis (lateral physeal growth decrease) with coxa valga positions and diminished acetabular coverage, premature physeal closure, relative trochanteric overgrowth, and a shortened femur.

I. MR Imaging to Detect Hip Ischemia Due to Extreme Immobilization Positioning AVN associated with treatment of DDH currently is diagnosed late, by plain radiographic criteria, months and even years after its onset when changes in the secondary center of ossification and growth sequelae have become evident. Detection of ischemia at the time of hip reduction, when it might be corrected, could help to prevent necrosis of the proximal femoral cartilage and bone. Gadolinium-enhanced MR imaging can demonstrate vascular perfusion to the physis, epiphyseal cartilage, and ossification center. We undertook an experimental study in piglets to determine whether gadolinium-enhanced MR imaging can detect physeal and epiphyseal ischemia in the proximal femur at an early stage when it is still reversible (137). The imaging technique now under investigation in children undergoing closed reduction and hip spica immobilization for DDH and early findings will be described in Section 2 that follows (136). 1. EXPERIMENTAL STUDIES We sought to determine whether gadolinium-enhanced MR imaging can detect early reversible ischemia of the capital femoral epiphysis and physis induced by hip hyperabduction in piglets. Thirteen 1- to 3-week-old piglets were placed in maximal hip abduction bilaterally and studied with dynamic gadolinium-enhanced MR imaging 1-6 hr later to assess ischemia of the 26 femoral heads. The abduction then was released, and they were allowed to walk freely for 1-7 days before they were reimaged with the hips in neutral po-

SECTION Xll ~ Avascular Necrosis as a Complication of Treatment sition to assess reperfusion. Enhancement was evaluated on MR images and compared with histologic findings. Ischemia after hyperabduction developed in all 26 cartilaginous epiphyses and in 85% of the femoral head physes. The secondary ossification center of the femoral head is present at birth in piglets. The most frequent abnormality was a sharply marginated nonenhancing area in the anterior part of the femoral head. A smaller area of ischemia developed in the posterior part of the femoral head, adjacent to the acetabular rim. The secondary center of ossification was ischemic in 56% of the hips after 1 hr of abduction and in all hips after 4 or 6 hr (p = 0.02). The overall severity of ischemia was greater with increasing abduction time (p < 0.01) and increasing degree of abduction (p < 0.01). There was partial reperfusion in 83% of the hips after 1 day of ambulation and complete reperfusion in all 26 hips (100%) after 1 week. Enhanced MRI was shown to detect early ischemia of the epiphyseal and physeal cartilage and the epiphyseal marrow. In piglets, the ischemia due to hyperabduction was reversible if corrected within 6 hr. Longer term studies with this model have not yet been performed.

a. Normal Epiphyseal and Physeal Enhancement (12 Hips) In six normal piglets, gadolinium administration resuited in rapid and intense enhancement of the physis, metaphyseal spongiosa, and epiphyseal vascular canals. The epiphyseal cartilage enhanced more gradually and to a lesser degree. By 10 min, the signal intensity of the epiphyseal cartilage was similar to that of the vascular canals. The physis had a higher signal intensity than the epiphyseal cartilage in both early and delayed images. The marrow was of low signal intensity in the epiphysis and metaphysis on unenhanced Tl-weighted images, suggesting that it was predominantly hematopoietic. It enhanced much less than the adjacent cartilage in both the epiphyseal ossification center and the metaphysis. b. Imaging in Hyperabduction (26 Hips) In all hips of 13 piglets, hyperabduction resulted in areas of decreased enhancement of the epiphyseal cartilage (Fig. 19). The typical abnormality involved predominantly the anterior epiphysis and had a sharp demarcation with the normal cartilage. These abnormalities were demonstrable only on post-gadolinium images, with no evidence of abnormality on the T2-weighted images or on any other preenhancement sequence. There was a second, less well-defined region of decreased enhancement in the posterolateral aspect of the femoral head. This region was small and was accompanied by a similar region of decreased enhancement in the adjacent posterior labrum. The degree of cartilaginous ischemia was greater in the anterior than in the posterior segments of the femoral head (p < 0.01) in the piglets abducted for 1 hr. This regional difference was not apparent in piglets abducted for more than 1 hr as ischemia involved most of the femoral head. The overall severity of ischemia was greater with increasing abduction time (p < 0.001) and increasing degree of abduction (p < 0.01). Ischemia was independent of the age of the

259

piglets. Physeal abnormalities were seen in 85% of femurs, and marrow abnormalities were seen in 69% of the capital femoral epiphyses. Of the latter, the secondary center of ossification was involved in 56% of the femurs subjected to 1 hr of abduction but in all of the femurs after 4-6 hr of abduction (p = 0.02). Figure 19A summarizes the typical pattern of ischemia after abduction. There was a marked difference between the enhancement curves of the perfused and nonperfused portions of the epiphyseal cartilage and between the perfused and nonperfused physeal cartilage. Differences between perfused and nonperfused enhancement ratios in epiphyseal cartilage and physis at the respective time points were found to be statistically significant for both the presence and absence of perfusion (p < 0.001) and time (p 0.0001). There were no differences in enhancement ratios between the left and fight hips. MR angiography was performed in two piglets, but there was insufficient resolution to show a site of vascular compression. c. Imaging after Ambulation (22 Hips) All hips of 11 piglets showed a decrease in the extent of ischemia following the release of abduction and resumption of normal ambulation. Reperfusion after 1 day of ambulation was partial in 83% of the epiphyses and complete in 17%. Reperfusion after 1 week of ambulation was complete in all hips. The decrease in degree of ischemia after more prolonged ambulation was statistically significant (p < 0.001). T2-weighted e images, however, were not different from those obtained during abduction. d. Histologic Sections (12 Hips) The secondary ossification centers of the femoral head and the greater trochanter were present in all specimens. The capital (head) secondary center was forming via the endochondral bone mechanism. Its marrow was in a developmental hematopoietic phase in each animal with no fat cells seen. Cartilage canals within the epiphyseal cartilage were abundant. They contained arteries (arterioles), capillaries, and veins (venules) in a connective tissue matrix. The cartilage canals were never present in the surface articular cartilage, but occasionally partially breached the physeal cartilage. There were no structural abnormalities of the cartilage of the epiphysis and physis, even in the piglets abducted for 6 hr. There also was no evidence of secondary center osteonecrosis, which at the light microscopic level would have been empty osteocyte lacunae and a fibrotic marrow. In one piglet, there was histologic evidence of marrow vessels structurally intact but devoid of red blood cells. This piglet initially had shown massive ischemia changes after 1 day of ambulation. In all other instances the normal and reperfused secondary center marrow had vessels filled with red blood cells. e. Summary Proximal femoral ischemia in the setting of developmental dysplasia of the hip occurs in epiphyses that are unossified or recently have developed secondary centers of ossification. Ischemia in these infants, therefore, is primarily of the cartilage component. Although avascular necrosis of the mature, ossified capital femoral epiphysis is detectable by MR imaging and scintigraphy, early diagnosis of

260

CHAPTER 3 ~ Developmental Dysplasia o]r the Hip

F I G U R E 19 (a) MR imaging of the immature pig hip held in full abduction shows a marked change in perfusion indicative of absent femoral head vascularity. Gadolinium-enhanced study shows a widely abducted femur with no signal in blackened appearing femoral head (arrow), whereas a normal signal persists with normal vascularity in distal femoral and proximal tibial epiphysis at far right. (B) Anteroposterior view of the widely abducted femoral head without vascular perfusion (arrow). In (C) the perfusion has returned to normal following release of the abduction at 6 hr with imaging done 1 week later. With release of the abduction full vascularity returned within 6 hr. The ability to image an avascular condition immediately upon its occurrence and to reverse that by changing the position of abduction is expected to have major positive consequences in minimizing posttreatment AVN. [Parts A - C reprinted from Jaramillo et al. (1996). Am. J. Roentgenol. 166:879-887, with permission.]

ischemia in the immature dysplastic epiphysis requires a technique that can detect perfusion of the cartilage. Gadoliniumenhanced MR imaging has been used for early detection of bone marrow ischemia in experimental animals and in patients with avascular necrosis of the femoral ossification center. We have evaluated the use of dynamic gadoliniumenhanced MR imaging for the detection of ischemia in the immature, mostly cartilaginous proximal femur. Our study demonstrates five points regarding ischemia of the proximal femur in piglets: (1) short-term hyperabduction consistently produces ischemia of the capital femoral epiphysis; (2) the ischemia is more severe with longer duration and greater degrees of abduction; (3) early ischemia of the cartilaginous epiphysis and physis is detectable by enhanced MR imaging even in the absence of bony ischemia; (4) ischemia of the

cartilage is not detectable by other MR sequences; and (5) this ischemia is reversible as abduction of 6 hr or less does not result in permanent abnormalities on MR images or histologic examination. Perhaps its ill effects also are reversible or preventable. In both piglets and infants, branches of the femoral circumflex vessels enter the capital femoral epiphyseal cartilage and course through it within vascular canals. These cartilaginous canals contain an arteriole, venule, and capillary plexus, all lying within connective tissue. The cartilage canals are present throughout the epiphyseal cartilage and occasionally traverse the physeal cartilage but are never present within the articular cartilage. The canals are distributed randomly and are entirely free of anastomoses. Nutrients (and gadolinium) presumably diffuse from the vascular

SECTION XII ~ Avascular Necrosis as a Complication of Treatment lumen into the intracanalicular perivascular tissue and finally into the cartilage. The enhancement curves of the perfused proximal femur obtained in our study support this sequence. Gadolinium is delivered shortly after injection to the vascular canals and physis and there is rapid enhancement of these structures. Enhancement in the epiphyseal cartilage is slower and less marked, which suggests that it occurs by diffusion. In the cartilaginous epiphysis and physis and in the vascular canals, the enhancement persists for several minutes after the injection. Therefore, there is little advantage in using very rapid sequences to image perfusion of the epiphyseal cartilage; enhanced Tl-weighted images produce higher resolution and less susceptibility than faster gradient-recalled echo sequences. The relationship between duration of abduction and severity of ischemia suggested by our data supports the need for diagnosing epiphyseal ischemia early. Our data also confirm the postulate that the severity of femoral ischemia is related to the degree of abduction. We found two patterns of early ischemia in the piglet capital femoral epiphysis: (1) anterior ischemia, which appears to be related to the obstruction of a larger vessel as it has very well defined borders, and (2) posterolateral ischemia, which is more likely to result from the direct compression of the cartilage of the femoral head against the acetabulum as it occurs on both sides of the joint. Anterior ischemia initially is cartilaginous, but with longer abduction the bony component of the femoral head becomes involved. Physeal ischemia occurred frequently, always with adjacent epiphyseal abnormality. In some hips, however, the physis was spared, despite severe epiphyseal ischemia. The discrepancy between epiphyseal ischemia and physeal involvement is puzzling as the physeal blood supply comes from the epiphyseal vessels. Discontinuation of the abduction appeared to result in complete reperfusion of the capital femoral epiphysis. Although no sequelae were demonstrated by MR imaging or histologic examination, it is uncertain whether mild injury, undetectable by these techniques, can occur after ischemia and result in permanent growth disturbance. Our study did not include any long-term follow-up of piglets with ischemia. Cartilage is relatively insensitive to hypoxia. Autoradiographic studies in animals show that the cartilage remains metabolically active for hours after the death of the animal. It is likely, therefore, that there is a period when hypoxia to the epiphyseal cartilage and physis can be corrected without permanent damage to these structures. In the newborn human, the anterior part of the femoral epiphysis and physis is perfused by the lateral circumflex femoral artery and the posterior part by the medial circumflex artery. During the first year of life, as the contribution from the lateral circumflex artery decreases, the anterior and lateral portions of the femoral head become more susceptible to vascular compromise. It is believed that compression of the epiphyseal vessels is a major cause of avascular necrosis. The compression may be extra-epiphyseal, between the psoas muscle and the acetabulum, or by the acetabular la-

261

brum pressing into the head-neck junction. Alternatively, direct pressure of the femoral head against the acetabulum may compress the cartilage canals within the epiphysis and cause ischemia by blocking the vessels within the epiphysis. In children, there are variable patterns of ischemia, suggesting that more than one vessel, and probably more than one mechanism, is responsible for the abnormalities. Avascular necrosis of the femoral head in patients with developmental dysplasia of the hip has not been documented in patients who have not undergone therapy. Ischemia appears to result primarily from immobilization in positions such as hyperabduction or extreme internal rotation, similar to the position we chose for our study. Previous experimental approaches in pigs using micro-opaque angiography and in dogs using a hydrogen washout technique following implantation of platinum electrodes into the secondary ossification centers, though valuable in demonstrating AVN, cannot be adapted readily to the clinical setting. Early detection of positional ischemia, however, could lead to a prompt change in the patient's degree of abduction and, hopefully, to the prevention of avascular necrosis. Gadolinium-enhanced MR imaging can detect ischemia and might become a useful clinical tool to determine whether a patient placed in a hip spica cast or other immobilization device is developing ischemia. It also provides excellent depiction of articular relationships and could be used instead of CT to verify proper reduction; MR imaging then would be partly a substitution rather than merely another addition to the routine of patient care. We are now beginning to apply MR imaging to children following open or closed reduction and hip spica casting. We have demonstrated that gadolinium-enhanced imaging detects ischemia in the cartilage and bone of the femoral head of piglets. MR imaging can be effective in detecting very early ischemia, and at this stage, within hours of casting, most or all of the changes should still be reversible. Efforts are underway to refine the technique to help prevent the significant hip pathology associated with avascular necrosis of the femoral head in infants with developmental dysplasia of the hip. 2. CLINICAL USE OF MR IMAGING TO DETECT VASCULARITY OF THE FEMORAL HEAD FOLLOWING CLOSED REDUCTION AND HIP SPICA IMMOBILIZATION FOR D D H

The findings referred to previously are now being applied to children undergoing treatment for DDH in our institution (136). As the previous clinical descriptions in this section indicated, the complication of AVN leads to considerable short- and long-term morbidity in DDH. Prevention or diminution of AVN would be of considerable benefit. We have assessed gadolinium-enhanced MR imaging of position and vascular enhancement of the femoral head in pediatric patients who have undergone reduction of hip dislocation. Within 24 hr of hip reduction and spica casting, we performed 25 gadolinium-enhanced MR studies in 18 infants and young

262

CHAPTER 3 9 Developmental Oysplasia of the Hip

FIGURE 20 MR imaging with gadoliniumenhancementof the human hip in spica cast following closed reduction of DDH shows markedly diminished perfusion of the left femoral head. Parts (a) and (b) are from the same hip at differinglevels of study. Note the visualizationof the secondary center of the right hip at top (a) and the epiphyseal vessels (linear streaks) at bottom (b). [Reprinted from Jaramillo et al. (1998). Am. J. Roentgenol. 170:1633-1637, with permission.]

children (15 girls and 3 boys) with 23 dysplastic hips. All but two patients underwent closed reductions. We also evaluated intraoperative arthrograms for obstacles to reduction and subsequent radiographs for avascular necrosis. In summary, MR images showed that all femoral heads were within the acetabulae, but several structures interfered with concentric eduction. Obstacles to reduction included a pulvinar (n = 16), infolding of the capsule (n = 9), interposition of the labrum (n = 21), and a hypertrophied ligamentum teres and transverse ligament (n = 21). All 50 femoral heads showed enhancement: 35 normally, 10 homogeneously but less than on the contralateral femoral head or the ipsilateral greater trochanter, and 5 with areas of focally decreased enhancement. Hips that showed decreased enhancement had undergone greater degrees of abduction (r = .38, p < 0.01). Gadolinium-enhanced MR imaging can reveal abnormalities of hip position and the proximal femoral epiphyseal and physeal vascularity that can occur after hip reduction (Fig. 20). Abnormalities of enhancement were more frequent in patients who had greater femoral abduction. The effect of decreased epiphyseal vascular enhancement is still uncertain. a. Femoral Enhancement Epiphyseal vascular canals and physeal enhancement were seen in all infants. Of 50 hips studied, 35 femoral heads showed normal gadolinium enhancement with contrast material showing numerous parallel vascular canals within the cartilaginous epiphysis and with prominent physeal enhancement. Ten femoral heads en-

hanced less than the contralateral heads without focal abnormality, and 5 femoral heads showed areas of focally decreased enhancement. No studies showed focal or global absence of vascular canals. In no case was the abnormality deemed severe enough at our current level of understanding to require a change in hip position and recasting. In the future, however, evidence of a certain degree of vascular occlusion would warrant immediate change of hip spica to improve vascular perfusion. Agreement between radiologists for the rating of epiphyseal enhancement was good with a K value of 0.47 (p < 0.0001). A significant correlation existed between greater abduction and more severe abnormality of enhancement (r = .38, p < 0.01) (Fig. 8). Only 2 of 14 femoral heads abducted less than 55 ~ showed abnormal enhancement. No enhancement defects were seen in hips abducted less than 50 ~. We have successfully obtained 25 studies without sedation in infants in hip spica casts. Success is attributed to the short imaging time ( < 2 min), the fact that the imaging can be repeated until satisfactory, and the substantial immobilization provided by the spica cast. The MR studies take longer than the CT examinations, but we have been able to image patients consistently in less than 15 min. Gadolinium is detectable in the vascular canals for approximately 10 min after the injection so the time window is larger. Ischemia of the femoral epiphysis during the treatment of developmental dysplasia of the hip can lead to proximal femoral growth disturbance, which in tum can result in coxa vara, femoral shortening, acetabular deformity, limping, pain, and osteoarthritis. Our data support the hypothesis that ischemia of the proximal femoral epiphysis is related directly to relatively excessive hip abduction after reduction. Gadolinium-enhanced MR imaging appears to demonstrate early detection of abnormal blood flow, leaving open the possibility of timely correction of the angle of abduction and prevention of avascular necrosis and its proximal femoral and acetabular abnormalities.

References 1. Akbas A, Unsaldi T, Korukliu O, Goze F (1995) The effect of physeal traction applied to the triradiate cartilage on acetabular growth. Internat Orthop 19:122-126. 2. Albinana J, Morcuende JA, Delgado E, Weinstein SL (1995) Radiologic pelvic asymmetry in unilateral late-diagnosed developmental dysplasia of the hip. J Pediatr Orthop 15: 753-762. 3. Allen RP (1962) Ischemic necrosis following treatment of hip "dysplasia." J Am Med Assoc 180:497-499. 4. Almby B, Grevsten S, Lonnerholm T (1978) Hip joint instability after the neonatal period. II. The acetabular growth potential. Acta Radiol Diag 20:213-222. 5. Almby B, Lonnerholm T (1978) Hip joint instability after the neonatal period. I. Value of measuring the acetabular angle. Acta Radiol Diag 20:200-212. 6. Andersen H (1962) Histochemical studies of the development of the hip joint. Acta Anat 48:258-292.

References 7. Andren L, Borglin NE (1961) Disturbed urinary excretion pattern of estrogen in newborns with congenital dislocation of the hip. 1. The excretion of estrogen during the first few days of life. Acta Endocrinol 37:423. 8. Artz TD, Levine DB, Lim WN, Salvati EA, Wilson PD, Jr (1975) Neonatal diagnosis, treatment and related factors of congenital dislocation of the hip. Clin Orthop Rel Res 110: 112-136. 9. Atar D, Lehman WB, Tenebaum Y, Grant AD (1993) Pavlik harness versus Frejka splint in treatment of developmental dysplasia of the hip: Bicenter study. J Pediatr Orthop 13: 311-313. 10. Azuma H, Taneda H, Igarashi H, Fujioka M (1990) Preoperative and postoperative assessment of rotational acetabular osteotomy for dysplastic hips in children by three-dimensional surface reconstruction computed tomography imaging. J Pediatr Orthop 10:33-38. 11. Badgley CE (1949) Etiology of congenital dislocation of the hip. J Bone Joint Surg 31A:341-356. 12. Baflow TG (1962) Early diagnosis and treatment of congenital dislocation of the hip. J Bone Joint Surg 44B:292-301. 13. Barlow TG (1966) Congenital dislocation of the hip in the newborn. Proc Roy Soc Med 5A: 1103-1108. 14. Barrett WP, Staheli LT, Chew DE (1986) The effectiveness of the Salter innominate osteotomy in the treatment of congenital dislocation of the hip. J Bone Joint Surg 68A:79-89. 15. Bennett WE (1908) Congenital Dislocation of the Hip. Birmingham: Percival Jones. 16. Bensahel H, Csukonyi Z, Huguenin P (1988) Vascular disorders of the proximal femur following treatment of congenital hip dislocation. Arch Orthop Traum Surg 107:372-376. 17. Bernays A (1878). Die entwicklungsgeschichte des kniegelenkes des menschen, mit bemerkungen uber die gelenke. Zeit f Anat Entwicklung 4:403-446. 18. Bernbeck R (1949) Die pathologische femurtorision und coxa volga (Ihre operative beseitigung durch intertrochantere femurosteotomie). Z Orthop 78:303-313. 19. Bertol P, Macnicol MF, Mitchell GP (1982) Radiographic features of neonatal congenital dislocation of the hip. J Bone Joint Surg 64B:176-179. 20. Bialik V, Benyamini O (1994) Hip osteotomies: Developmental dysplasia of the hip: Pathophysiology and surgical indications in the first two years of life. J Pediatr Orthop 3:1-4. 21. Blamoutier A, Carlioz H (1990) L'Osteotomie innominee de Salter dans le traitement de la luxation congenitale de hanche. Rev Chir Orthop 76:403-410. 22. Blockey NJ (1982) Congenital dislocation of the hip. J Bone Joint Surg 64B:152-155. 23. Bost FC, Hagey H, Schottstaedt ER, Larsen LJ (1948) The results of treatment of congenital dislocation of the hip in infancy. J Bone Surg 30A:454-468. 24. Bradford EH, Lovett RW, Brackett EG, Thorndike A (1904) Congenital dislocation of the hip. Boston Med Surg J 151: 85-96. 25. Brodhurst BE (1876) Congenital dislocations. In: Lectures on Orthopaedic Surgery, 2nd ed. pp. 158-169, London: J and A Churchill. 26. Brodhurst BE (1896) Congenital dislocation of the hip. In: Observation on Congenital Dislocations of the Hip, 3rd ed. pp. 5-33, London: J and A Churchill.

263

27. Bronner H (1927) Der diagnostische wert der luftfullung des gelenkraums bei der angeborenen huftverrenkung. Zentralbl f Chir 54:3237-3246. 28. Brougham DI, Broughton NS, Cole WG, Menelaus MB (1988) The predictability of acetabular development after closed reduction for congenital dislocation of the hip. J Bone Joint Surg 70B:733-736. 29. Brougham DI, Broughton NS, Cole WG, Menelaus MB (1990) Avascular necrosis following closed reduction of congenital dislocation of the hip. J Bone Joint Surg 72B:557-562. 30. Buchanan JR, Greer RB, Cotler JM (1981) Management strategy for prevention of avascular necrosis during treatment of congenital dislocation of the hip. J Bone Joint Surg 63A: 140-146. 31. Bucholz RW, Ogden JA (1978) Patterns of ischemic necrosis of the proximal femurmnonoperatively treated congenital hip disease. In: The Hip: Proceedings of the Sixth Open Scientific Meeting of the Hip Society. pp. 43-63, St. Louis: CV Mosby. 32. Burghard FF (1903) A discussion on congenital dislocation of the hip. Brit Med J ii:457-463. 33. Burgos-Flores J, Ocete-Guzman G, Gonzalez-Herranz P, Hevia-Sierra E, Amaya-Alarcon S (1993) Factors responsible for the development of avascular necrosis secondary to the treatment of congenital dislocation of the hip. Intern Orthop 17:305-307. 34. Carter CO, Wilkinson JA (1964) Genetic and environmental factors in the etiology of congenital dislocation of the hip. Clin Orthop Rel Res 33:119-128. 35. Castelein RM, Sauter AJM (1988) Ultrasound screening for congenital dysplasia of the hip in newborns" Its value. J Pediatr Orthop 8:666-670. 36. Castelein RM, Sauter AJM, de Vlieger M, van Linge B (1992) Natural history of ultrasound hip abnormalities in clinically normal newborns. J Pediatr Orthop 12:423-427. 37. Catterall A (1992) Assessment of adolescent acetabular dysplasia. In: Recent Advances in Orthopaedics. ed A Catterall, number 6. pp. 103-118, Edinburgh, Churchill Livingstone. 38. Cautru P (1892) Un cas de luxation coxo-femorale congenitale coincidant avec des malformations multiples. Rev Orthop 3:457-459. 39. Chen IH, Kuo KN, Lubicky JP (1994) Prognosticating factors in acetabular development following reduction of developmental dysplasia of the hip. J Pediatr Orthop 14:3-8. 40. Cherney DL, Westin GW (1989) Acetabular development in the infant's dislocated hips. Clin Orthop Rel Res 242:98-103. 41. Chuinard EG, Logan ND (1963) Varus-producing and derotational subtrochanteric osteotomy in the treatment of congenital dislocation of the hip. J Bone Joint Surg 45A: 1397-1408. 42. Chung SMK (1976) The arterial supply of the developing proximal end of the human femur. J Bone Joint Surg 58A: 961-970. 43. Clarke J (1896) Trans Brit Orthop Soc 1"36-39. 44. Clarke NMP (1986) Sonographic clarification of the problems of neonatal hip instability. J Pediatr Orthop 6:527-532. 45. Coleman SS (1968) Congenital dysplasia of the hip in the Navajo infant. Clin Orthop Rel Res 56:179-193. 46. Coleman SS (1974) The incomplete pericapsular (Pemberton) and innominate (Salter) osteotomies. Clin Orthop Rel Res 98: 116-123.

264

CHAPTER 3 9

Developmental Dysplasia of the Hip

47. Cooperman DR, Wallensten R, Stulberg SD (1980) Postreduction avascular necrosis in congenital dislocation of the hip. J Bone Joint Surg 62A:247-258. 48. Crawford AH, Carothers TA (1982) Hip arthrography in the skeletally immature. Clin Orthop Rel Res 162:54-60. 49. Crego CH, Jr, Schwartzmann JR (1948) Follow-up study of the early treatment of congenital dislocation of the hip. J Bone Joint Surg 30A:428-442. 50. Crellin RQ (1974) Innominate osteotomy for congenital subluxation and subtrochanteric osteotomy in the treatment of congenital dislocation of the hip. J Bone Surg 45A:1397-1408. 51. Crock HV (1980) An atlas of the arterial supply of the head and neck of the femur in man. Clin Orthop Rel Res 152:17-27. 52. Cruveilhier J (1849) Luxation congenitale de l'articulation coxo-femorale. In: Traite D'Anatomie Pathologique, volume 1. pp. 482-512, Paris: J-B Bailliere. 53. Dega W (1964) Schwierigkeiten in der chirurgischen reposition der veralteten kongenitalen subluxation des huftgelenkes bei kindern. Beitr Orthop Traumatol 11:642-647. 54. Delgado-Baeza E, Albinana-Cilveti J, Miralles-Flores C (1992) Why does pelvic deformity occur in experimental dislocation of the growing hip? J Pediatr Orthop 12:376-383. 55. Delpech J (1828) De L'Orthomorphie. Par Rapport a L'espece humaine. 56. Deutschlander C (1904) Zur beurteilung-der unblutigen reposition der angeborenen huftverrenkungen. Deut Z Chir 73: 1-48.

57. Dimeglio A, Kaelin A, Bonnel F, De Rosa V, Couture A (1994) The growing hip: Specifications and requirements. J Pediatr Orthop 3B:135-147. 58. Dimeglio A, Pous JG (1983) Le traitement chirurgical de la hanche luxee apres l'age de 5 ans. Rev Chir Orthop 69: 613-622. 59. Dunn PM (1969) Congenital dislocation of the hip (CDH): Necropsy studies at birth. Proc Roy Soc Med 62:1035-1037. 60. Dunn PM (1976) Perinatal observations on the etiology of congenital dislocation of the hip. Clin Orthop Rel Res 119: 11-22.

61. Dunn PM (1976) The anatomy and pathology of congenital dislocation of the hip. Clin Orthop Rel Res 119:23-27. 62. Dupuytren G (1826) Un deplacement originel ou congenital de la tete des femurs. Rep Gen Anat Physiol Path et Clin Chin 2:82-93 [reprinted and translated (partial) in Clin Orthop Rel Res 33:308]. 63. Dyson PHP, Lynskey TG, Catterall A (1987) Congenital hip dysplasia: Problems in the diagnosis and management in the first year of life. J Pediatr Orthop 7:568-574. 64. Elsworth C, Walker G (1986) The safety of the Denis Browne abduction harness in congenital dislocation of the hip. J Bone Joint Surg 68B :275-277. 65. Esteve R (1960) Congenital dislocation of the hip. A review and assessment of results of treatment with special reference to frame reduction as compared with manipulative reduction. J Bone Joint Surg 42B:253-263. 66. Exner GU (1988) Ultrasound screening for hip dysplasia in neonates. J Pediatr Orthop 8:656-660. 67. Eyre-Brook AL, Jones DA, Harris FC (1978) Pemberton's acetabuloplasty for congenital dislocation or subluxation of the hip. J Bone Joint Surg 60B: 18-24.

68. Fabry G, MacEwen GD, Shands AR (1973) Torsion of the femur: A follow-up study in normal and abnormal conditions. J Bone Joint Surg 55A: 1726-1738. 69. Faciszewski T, Kiefer GN, Coleman SS (1993) Pemberton osteotomy for residual acetabular dysplasia in children who have congenital dislocation of the hip. J Bone Joint Surg 75A: 643-649. 70. Fairbank HAT (1930) Congenital dislocation of the hip: With special reference to the anatomy. Brit J Surg V17:380-416. 71. Falliner A, Hasssenpflug J (1994) Der einfluss der sonographie auf diagnose und behandlung der sogangeborenen huftgelenksluxation. Z Orthop 132:505-511. 72. Farrell BP, Howorth MB (1935) Open reduction in congenital dislocation of the hip. J Bone Joint Surg 17:35-42. 73. Farrell BP, von Lackum HL (1926) Congenital dislocation of the hip. A report of 310 cases treated at the New York Orthopaedic Dispensary and Hospital. J Bone Joint Surg 8:551-556. 74. Felts WJL (1954) The prenatal development of the human femur. Am J Anat 94:1-44. 75. Ferran JL, Dimeglio A, Lebouco N, Couture A, DeRosa V (1994) Three-dimensional computed tomography of the infantile hip. J Pediatr Orthop, Part B, 3:131-134. 76. Fettweis E (1968) Sitzhochstellungsgips bei huftgelenksdysplasien. Arch Ortho Unfoll-Chir 63:38-51. 77. Fettweis E (1977) Die behandlung der kongenitalen huftgelenksdysplasie mit dem Sitz-Hock-Gips. Z Orthop 115: 648-649. 78. Fisher RL, Cary JM (1978) Avascular necrosis complicating congenital dislocation of the hip. Intemat Orthop 2:229-240. 79. Fogarty EE, Accardo NJ (1981) Incidence of avascular necrosis of the femoral head in congenital hip dislocation related to the degree of abduction during preliminary traction. J Pediatr Orthop 1:307-311. 80. Forlin E, Choi IH, Guille JT, Bowen JR, Glutting J (1992) Prognostic factors in congenital dislocation of the hip treated with closed reduction. J Bone Joint Surg 74A:1140-1152. 81. Foumet-Fayard J, Kohler R, Michel CR (1988) Resultats de l'osteotomie innominee de Salter dans la dysplasie residuelle de la hanche chez l'enfant. Rev Chir Orthop 74:243-251. 82. Fritsch EW, Schmitt E, Mittelmeier H (1996) Radiographic course after acetabuloplasty and femoral osteotomy in hip dysplasia. Clin Orthop Rel Res 323:215-225. 83. Fujioka F, Terayama K, Sugimoto N, Tanikawa H (1995) Long-term results of congenital dislocation of the hip treated with the Pavlik harness. J Pediatr Orthop 15:747-752. 84. Gabuzda GM, Renshaw TS (1992) Reduction of congenital dislocation of the hip. J Bone Joint Surg 74A:624-631. 85. Gade HG (1974) The surgical treatment of osteoarthritis of the hip-joint. Acta Chir Scand 95, Supp 120:32-36. 86. Gage JR, Winter RB (1972) Avascular necrosis of the capital femoral epiphysis as a complication of closed reduction of congenital dislocation of the hip: A critical review of 20 years experience at Gillette Children's Hospital. J Bone Joint Surg 54A:373-388. 87. Gallien R, Bertin D, Lirette R (1984) Salter procedure in congenital dislocation of the hip. J Pediatr Orthop 4:427-430. 88. Galloway HPH (1926) The open operation for congenital dislocation of the hip: Special reference to results. J Bone Joint Surg 8:539-550.

References

89. Gamble JG, Mochizuki C, Bleck EE, Rinsky LA (1985) Coxa magna following surgical treatment of congenital hip dislocation. J Pediatr Orthop 5:528-533. 90. Ganz R, Klaue K, Vinh TS, Mast JW (1988) A new periacetabular osteotomy for the treatment of hip dysplasias. Technique and preliminary results. Clin Orthop Rel Res 232: 26-36. 91. Gardner E, Gray DJ (1950) Prenatal development of the human hip joint. Am J Anat 87:163-211. 92. Gill AB (1948) The end results of early treatment of congenital dislocation of the hip. J Bone Joint Surg 30A:442-453. 93. Graf R (1983) New possibilities for the diagnosis of congenital hip joint dislocation by ultrasonography. J Pediatr Orthop 3:354-359. 94. Graf R (1984) Classification of hip joint dysplasia by means of sonography. Acta Orthop Traum Surg 102:248-255. 95. Graf R (1984) Fundamentals of sonographic diagnosis of infant hip dysplasia. J Pediatr Orthop 4:735-740. 96. Grant JCB (1972) An Atlas of Anatomy, 6th edition. Baltimore: Williams and Wilkins. 97. Grawitz P (1878) Ueber die ursachen der angebornen huftgelenk-verrenkungen. Arch f Path Anat Physiol 74:1-14. 98. Greenhill BJ, Hugosson C, Jacobsson B, Ellis RD (1993) Magnetic resonance imaging study of acetabular morphology in developmental dysplasia of the hip. J Pediatr Orthop 13: 314-317. 99. Gregosiewicz A, Wosko I (1988) Risk factors of avascular necrosis in the treatment of congenital dislocation of the hip. J Pediatr Orthop 8:17-19. 100. Grill F, Bensahel H, Canadell J, Dungl P, Matasovic T, Vizkelety T (1988) The Pavlik harness in the treatment of congenital dislocating hip: Report on a multicenter study of the European pediatric orthopedic society. J Pediatr Orthop 8:1-8. 101. Gruel CR, Birch JG, Roach JW, Herring JA (1986) Teratologic dislocation of the hip. J Pediatr Orthop 6:693-702. 102. Guyot J (1990) Atlas of Human Limb Joints, 2nd edition. Berlin, Baltimore: Williams and Wilkins. 103. Haines RW (1933) Cartilage canals. J Anat 68:45-64. 104. Haines RW (1942) The evolution of epiphyses and of endochondral bone. Biol Rev 17:267-292. 105. Hall JE (1986) Pelvic osteotomy in the early treatment of CDH. Adv Orthop Surg 3-11. 106. Hallel T, Salvati EA (1977) Premature closure of the triradiate cartilage: A case report and animal experiment. Clin Orthop Rel Res 124:278-281. 107. Hangen DH, Kasser JR, Emans JB, Millis MB (1995) The Pavlik harness and developmental dysplasia of the hip: Has ultrasound changed treatment patterns? J Pediatr Orthop 15: 729-735. 108. Hansson G, Althoff B, Bylund P, Jacobsson B, Lofberg AM, Lonnerholm T (1990) The Swedish experience with Salter's innominate osteotomy in the treatment of congenital subluxation and dislocation of the hip. J Pediatr Orthop 10:159-162. 109. Harcke HT, Kumar SJ (1991) The role of ultrasound in the diagnosis and management of congenital dislocation and dysplasia of the hip. J Bone Joint Surg 73A:622-628. 110. Harris NH (1976) Acetabular growth potential in congenital dislocation of the hip and some factors upon which it may depend. Clin Orthop Rel Res 119:99-106.

265

111. Harris NH, Lloyd-Roberts GC, Gallien R (1975) Acetabular development in congenital dislocation of the hip. J Bone Joint Surg 57B:46-52. 112. Harrison TJ (1961) The influence of the femoral head on pelvic growth and acetabular form in the rat. J Anat 95:12-24. 113. Hefti F (1995) Spherical assessment of the hip on standard ap radiographs: A simple method for the measurement of the contact area between acetabulum and femoral head and of acetabular orientation. J Pediatr Orthop 15:797-805. 114. Heine J, Felske-Adler C (1985) Ergebnisse der behandlung der kongenitalen huftluxation durch offene reposition unds beckenosteotomie nach Salter. Z Orthop 123:273-277. 115. Heinrich SD, Missinne LH, MacEwen GD (1992) The conservative management of congenital dislocation of the hip after walking age. Clin Orthop Rel Res 281:34-40. 116. Hilgenreiner H (1925) Zur fruhdiagnose und fruhbehandlung der angeborenen huftgelenkverrenkung. Med Klin 21:1385-1388 [Reprinted in translation in J Pediatr Orthop 6:202-214, 1986]. 117. Hintsche E (1928) Untersuchungen an stutzgeweben. I. Ueber die bedentung der gefasskanale im knorpelnach nach befunden an distalen enda des menschlichen schenkalbeines. Z Mikroskop Anat Forsch 12:62-126. 118. Hirohashi K, Kanbara T, Narushima M, Lee Y (1982) A radiographic study of ischemic necrosis following the treatment of congenital dislocation of the hip. J Jap Orthop Assoc 56: 927-928. 119. Hjelmstedt A, Asplund S (1983) Congenital dislocation of the hip: A biomechanical study in autopsy specimens. J Pediatr Orthop 3:491-497. 120. Hoaglund FT, Low WD (1980) Anatomy of the femoral neck and head, with comparative data from caucasians and Hong Kong Chinese. Clin Orthop Rel Res 152:10-16. 121. Hoffa A (1892) Zur operativen behandlung der angeborenenhuftgelenksverrenkung. Zentralbl Chir 19:921. 122. Hoffa A (1898) Lehrbuch der orthopaedischen chirurgie. Stuttgart. 123. Howorth MB (1952) Congenital dislocation of the hip-technic of open reduction. Ann Surg 135:508-519. 124. Howorth MB (1963) The etiology of congenital dislocation of the hip. Clin Orthop Rel Res 29:164-179. 125. Howorth MB (1977) Development of present knowledge of congenital displacement of the hip. Clin Orthop Rel Res 125: 68-87. 126. Howorth MB, Smith HW (1932) Congenital dislocation of the hip treated by open operation. J Bone Joint Surg 14: 299-313. 127. Humphrey (1888-1889) The angle of the neck with the shaft of the femur at different periods of life and under different circumstances. J Physiol 23:273-282. 128. Hurrell DJ (1934) The vascularization of cartilage. J Anat 69: 47-61. 129. Ilfeld FW, Makin M (1977) Damage to the capital femoral epiphysis due to Frejka pillow treatment. J Bone Joint Surg 59A:654-658. 130. Imatani J, Miyake Y, Nakatsuka Y, Akazawa H, Mitani S (1995) Coxa magna after open reduction for developmental dislocation of the hip. J Pediatr Orthop 15:337-341. 131. Ippolito E, Ishii Y, Ponseti IV (1980) Histologic, histochemical, and ultrastructural studies of the hip joint capsule and

266

CHAPTER 3 ~

DevelopmentalDysplasia of the Hip

ligamentum teres in congenital dislocation of the hip. Clin Orthop Rel Res 146:246-258. 132. Ishii Y, Weinstein SL, Ponseti IV (1980) Correlation between arthrograms and operative findings in congenital dislocation of the hip. Clin Orthop Rel Res 153:138-145. 133. Iwasaki K, Management after application of the Pavlik harness in congenital dislocation of the hip. Arch Orthop Trauma Surg 106:276-280. 134. Iwerson LJ, Kalen V, Eberle C (1989) Relative trochanteric overgrowth after ischemic necrosis in congenital dislocation of the hip. J Pediatr Orthop 9:381-385. 135. Jacquemier M, Jouve JL, Bollini G, Panuel M, Migliani R (1992) Acetabular anteversion in children. J Pediatr Orthop 12:373-375. 136. Jaramillo D, Villegas-Medina O, Laor T, Shapiro F, Millis MB (1998) Gadolinium-enhanced MR imaging of pediatric patients after reduction of dysplastic hips: Assessment of femoral head position, factors impeding reduction and femoral head ischemia. AJR 170:1633-1637. 137. Jaramillo D, Villegas-Medina OL, Doty DK, Dwek JR, Ransil BJ, Mulkern RV, Shapiro F (1996) Gadolinium-enhanced MR imaging demonstrates abduction-caused hip ischemia and its reversal in piglets. AJR 166:879-887. 138. Judet J (1958) L'avenir des luxations congenitales de la hanche reduites. Rev Chir Orthop 44:152-175. 139. Kahle WK, Anderson MB, Alpert J, Stevens PM, Coleman S (1990) The value of preliminary traction in the treatment of congenital dislocation of the hip. J Bone Joint Surg 72A: 1043-1047. 140. Kahle WK, Coleman SS (1992) The value of the acetabular teardrop figure in assessing pediatric hip disorders. J Pediatr Orthop 12:586-591. 141. Kalamchi A, MacEwen GD (1980) Avascular necrosis following treatment of congenital dislocation of the hip. J Bone Joint Surg 62A:876-888. 142. Karadimas JE, Holloway GMN, Waugh W (1982) Growth of the proximal femur after varus-derotation osteotomy in the treatment of congenital dislocation of the hip. Clin Orthop Rel Res 162:61-68. 143. Kashiwagi N, Suzuki S, Kasahara Y, Seto Y (1996) Prediction of reduction in developmental dysplasia of the hip by magnetic resonance imaging. J Pediatr Orthop 16:254-258. 144. Kasser JR, Bowen JR, MacEwen GD (1985) Varus derotation osteotomy in the treatment of persistent dysplasia in congenital dislocation of the hip. J Bone Joint Surg 67A:195-202. 145. Keetley CB (1900) Orthopaedic Surgery, Chapter 14, pp. 273-302, London: Smith, Elder. 146. Keret D, MacEwen GD (1991) Growth disturbance of the proximal part of the femur after treatment for congenital dislocation of the hip. J Bone Joint Surg 73A:410-423. 147. Kirmisson E (1894) Contribution a l'etude de la pathogenie et du traitement des luxations congenitales de la hanche. Rev Orthop 5:186-215. 148. Kirmisson E (1905) Nouvel exemple de luxation congenitale de la hanche recueilli sur un foetus mort-ne. Rev Orthop 6: 319-324. 149. Kleinberg S, Lieberman HS (1936) The acetabular index in infants in relation to congenital dislocation of the hip. Arch Surg 32:1049-1054.

150. Koenig F (1891) Bildung einer knochernen hemmung fur den gelenkkopf bei der kongenitalen luxation. Zentralbl Chir 17: 146-147. 151. Lack W, Windhager R, Kutschera HP, Engel A (1991) Chiari pelvic osteotomy for osteoarthritis secondary to hip dysplasia: Indications and long-term results. J Bone Joint Surg 73B: 229-234. 152. Lagrange J, Dunoyer J (1962) La vascularisation de la tete femorale de l'enfant. Rev Chir Orthop 48:123-137. 153. Lance M (1925) Constitution d'une butee osteoplastique dans les luxations et subluxations congenitales de la hanche. Le Presse Med 33:945-948. 154. Lance M (1937) Malformation Luxante. Luxation congenitale de le hanche. In: Traite de Chirurgie Orthopedique, Ed. L Ombredanne, P Mathieu, volume 4. pp. 3047-3179. 155. Lang P, Genant HK, Steiger P, Lindquist T, Moore S, Skinner SR (1989) Three-dimensional digital displays in congenital dislocation of the hip: Preliminary experience. J Pediatr Orthop 9:532-537. 156. Lange M (1930) Die endresulte der unblutigen behandlung der angeborenen huftverrenkung. Z Orthop 90:270. 157. Langenskiold A, Laurent LE (1966) Development of the concepts of pathogenesis and treatment of congenital dislocation of the hip. Clin Orthop Rel Res 44:41-49. 158. Langenskiold A, Sarpio O, Michelsson J-E (1962) Experimental dislocation of the hip in the rabbit. J Bone Joint Surg 44B:209-215. 159. Laurenson RD (1959) The acetabular index. J Bone Joint Surg 41B:702-710. 160. Laurenson RD (1965) Development of the acetabular roof in the fetal hip. J Bone Joint Surg 47A:975-983. 161. Lauritzen J (1974) The arterial supply to the femoral head in children. Acta Orthop Scand 45:724-736. 162. Law EG, Heistad DD, Marcus ML, Mickelson MR (1982) Effect of hip position on blood flow to the femur in puppies. J Pediatr Orthop 2:133-137. 163. Le Damany P (1904) Un defant de la hanche humaine. J Anat Physio140:1-21. 164. Le Damany P (1905) La luxation congenitale de la hanche. Anatomie pathologique, lesions pneumonitoires. Rev Chir 32:27-36. 165. Le Damany P (1908) Die angeborene huftgelenksverrenkung. Z Orthop Chir 21:129-169. 166. Le Damany P (1914) Congenital luxation of the hip. Am J Orthop Surg 11:541-567. 167. Le Demany P (1904) Une nouvelle theorie pathogenique de la luxation congenitale de la hanche. Rev Chir 29:175-199, 370-397. 168. Lee DY, Choi IH, Lee CK, Cho TJ (1991) Assessment of complex hip deformity using three-dimensional CT image. J Pediatr Orthop 11:13-19. 169. Lee J, Jarvis J, Uhthoff HK, Avruch L (1992) The fetal acetabulum: A histomorphometric study of acetabular anteversion and femoral head coverage. Clin Orthop Rel Res 281: 48-55. 170. Lempicki A, Wierusz-Kozlowska M, Kruczynski J (1990) Abduction treatment in late diagnosed congenital dislocation of the hip. Follow-up of 1010 hips treated with Frejka pillow 1967-1976. Acta Orthop Scand Supp 236, 61:1-30.

References 171. Lepage G, Grosse (1901) Luxation congenitale de la hanche droite chez un nouveau-ne atteint de malformations multiples. Rev Orthop 2:257-270. 172. Leveuf J (1941) Etude des resultats eloignes de la reduction non sanglante des luxations congenitales de la hanche. J Chir 57:11-39. 173. Leveuf J (1941) Etude des resultats eloignes de la reduction sanglante des luxations congenitales de la hanche. J Chir 57: 117-133. 174. Leveuf J (1947) Primary congenital subluxation of the hip. J Bone Joint Surg 29:149-162. 175. Leveuf J, Bertrand P (1941) Notre orientation actuelle dans le traitement des luxations congenitales de la hanche. J Chir 57: 29-64. 176. Lima C, Esteve R, Trueta J (1960) Osteochondritis in congenital dislocation of the hip: A clinical and radiographic study. Acta Orthop Scand 29:218-236. 177. Liu JSC, Kuo KN, Lubicky JP (1996) Arthrographic evaluation of developmental dysplasia of the hip. Clin Orthop Rel Res 326:229-237. 178. Loeffler F (1923) Operative behandlung veralteter kongenitaler hufluxation. Ergeb Chir Orthop 16:484-515. 179. Loew F, Niethard FV, Schneider E (1990) Die prognostische bedutung klinischer und radiologischer parameter fur die revalgisierung des schkelhalses nach intertrochanterer derotations varisierungsosteotomie. Z Orthop 128:592-597. 180. Lorenz A (1895) Uber die unblutige behandlung der angeborenen huftverrenkung mittels der funktionellen belastungesmethode. Zentralbl Chir 22:761. 181. Lorenz A (1920) Die sogenannte angeborene huftgelenks verrenkung, ihre pathologic und therapi. Enke, Stuttgart. 182. Ludloff K (1911) Die angeborene huftluxation mit besonderer berucksichtigung der luxationspfanne. Ergeb Chir Orthop 3: 529-570. 183. Ludloff K (1913) The open reduction of the congenital hip dislocation by an anterior incision. Am J Orthop Surg 10:438. 184. MacKenzie IG (1972) Congenital dislocation of the hip. The development of a regional service. J Bone Joint Surg 54B: 18-39. 185. Mader G, Brunner CH, Ganz R (1979) 10 Jahres resultaten der beckenosteotomie nach Salter. Orthopade 8:30-35. 186. Mardam-Bey TH, MacEwen GD (1982) Congenital hip dislocation after walking age. J Pediatr Orthop 2:478-486. 187. Massie WK (1951) Vascular epiphyseal changes in congenital dislocation of the hip. J Bone Joint Surg 33A:284-306. 188. Massie WK (1956) Congenital dislocation of the hipmlts causes and effects. Clin Orthop Rel Res 8:103-121. 189. Massie WK, Howorth MB (1950) Congenital dislocation of the hip. Part I. Method of grading results. J Bone Joint Surg 32A:519-531. 190. Massie WK, Howorth MB (1951) Congenital dislocation of the hip. Part II. Results of open reduction as seen in early adult period. J Bone Joint Surg 33A:171-190. 191. Mau H (1961) Spontaneous straightening of the femoral neck following adduction osteotomy in congenital dysplasia of the hip. J Bone Joint Surg. 43A:285. 192. Mau H, Dorr WM, Henkel L, Lutsche J (1971) Open reduction of congenital dislocation of the hip by Ludloff's method. J Bone Joint Surg 53A:1281-1288.

267

193. McKay DW (1974) A comparison of the innominate and the pericapsular osteotomy in the treatment of congenital dislocation of the hip. Clin Orthop Rel Res 98:124-132. 194. McKibbin B (1970) Anatomical factors in the stability of the hip joint in the newborn. J Bone Joint Surg 52B:148-159. 195. Melzer C, Wulker N (1990) Potential mistakes in hip-joint sonography. Arch Orthop Trauma Surg 109:126-130. 196. Michaels G, Matles AL (1970) The role of the ligamentum teres in congenital dislocation of the hip. Clin Orthop Rel Res 71:199-201. 197. Michelsson J-E, Langenskiold A (1972) Dislocation or subluxation of the hip. Regular sequels of immobilization of the knee in extension in young rabbits. J Bone Joint Surg 54A: 1177-1186.

198. Milgram JW (1976) Morphology of untreated bilateral congenital dislocation of the hips in a 74-year-old man. Clin Orthop Rel Res 119:112-115. 199. Milgram JW, Tachdjian MO (1976) Pathology of the limbus in untreated teratologic congenital dislocation of the hip: A case report of a 10-month-old infant. Clin Orthop Rel Res 119:107-111. 200. Millis MB, Murphy SB (1992) Use of computed tomographic reconstruction in planning osteotomies of the hip. Clin Orthop Rel Res 274:154-159. 201. Millis MB, Kaelin AJ Schluntz K, Curtis B, Hey L, Hall JE (1994) Spherical acetabular osteotomy for treatment of acetabular dysplasia in adolescents and young adults. J Pediatr Orthop Part B 3:47-53. 202. Mitchell GP (1963) Arthrography in congenital displacement of the hip. J Bone Joint Surg 45B:88-95. 203. Mitchell GP (1972) Problems in the early diagnosis and management of congenital dislocation of the hip. J Bone Joint Surg 54B:4-12. 204. Mittelmeier H (1964) Zur kombination von intertrochanterer femurosteotomie und pfannendachplastik. Beitr Orthop 11: 624. 205. Mittelmeier H (1990) Huftchirurgie im kleinkindesalter. Z Orthop 128:365-368. 206. Morel G (1983) L'osteotomie de Salter dans le traitement de le malformation luxante de la hanche apres l'age de le marche. Rev Chir Orthop 69:157. 207. Morgan JD, Somerville EW (1960) Normal and abnormal growth at the upper end of the femur. J Bone Joint Surg 42B: 264-272. 208. Morin C, Harcke HT, MacEwen GD (1985) The infant hip: Real time US assessment of acetabular development. Radiology 157:673-677. 209. Morville P (1936) On the anatomy and pathology of the hip joint. Acta Orthop Scand 7:107-142. 210. Murphy SB, Kijewski PK, Millis MB, Harless A (1990) Acetabular dysplasia in the adolescent and young adult. Clin Orthop Rel Res 261:214-223. 211. Murray RO (1965) The etiology of primary osteoarthritis of the hip. Brit J Radiol 38:810-824. 212. Nichols GW, Schwentker EP, Boal DKB (1986) Correlation of anatomy and ultrasonographic images in the infant hip: An experimental cadaver study. J Pediatr Orthop 6:410-415. 213. O'Brien T, (1985) Growth-disturbance lines in congenital dislocation of the hip. J Bone Joint Surg 67A:626-632.

268

CHAPTER 3 ~

Developmental Dysplasia of the Hip

214. O'Brien T, Millis MB, Griffin PP (1986) The early identification and classification of growth disturbances of the proximal end of the femur. J Bone Joint Surg 68A:970-980. 215. O'Hara JN (1989) Congenital dislocation of the hip: Acetabular deficiency in adolescence (absence of the lateral acetabular epiphysis) after limbectomy in infancy. J Pediatr Orthop 9:640-648. 216. O'Hara JP, Dommisse GF (1983) Extraosseous blood supply to the neonatal femoral head. Clin Orthop Rel Res 174: 293-297. 217. Ogden JA (1982) Dynamic pathobiology of congenital hip dysplasia. In: Congenital Dislocation of the Hip, Ed MO Tachdjian, Chapter 4. pp. 93-144, New York: Churchill Livingstone. 218. Ogden JA (1974) Changing patterns of proximal femoral vascularity. J Bone Joint Surg 56A:941-950. 219. Ortolani M (1976) The Classic: Congenital hip dysplasia in the light of early and very early diagnosis. Clin Orthop Rel Res 119:6-10. 220. Paci A (1887) Nuovo contributo alla patologia della lussacione iliaca del femore. Arch Att: Soc Ital di Chir 3:444. 221. Paletta JB (1820) Exercitations Pathologicae. Mediolani. 222. Pavlik A (1957) Die funktionelle behandlungsmethod mittels riemenbugel alsprinzip der konservativen therapie bei angeborenen huftgelenksverrenkungen der sauglinge. Z Orthop 89:341 [Reprinted in translation in Clin Orthop Rel Res 281: 4-10, 1992]. 223. Pemberton PA (1965) Pericapsular osteotomy of the ilium for the treatment of congenital subluxation and dislocation of the hip. J Bone Joint Surg 47A:65-86. 224. Pemberton PA (1974) Pericapsular osteotomy of the ilium for the treatment of congenitally dislocated hips. Clin Orthop Rel Res 98:41-54. 225. Perkins G (1928) Signs by which to diagnose congenital dislocation of the hip. Lancet i:648-650. 226. Peterson HA, Klassen RA, McLeod RA, Hoffman AD (1981) The use of computerized tomography in dislocation of the hip and femoral neck anteversion in children. J Bone Joint Surg 63B: 198-208. 227. Pettersson H, Theander G (1979) Ossification of femoral head in infancy. I. Normal standards Acta Radiol Diag 20: 170-179. 228. Pettersson H, Theander G (1979) Ossification of femoral head in infancy. II. Ossification in infants treated for congenital dislocation. Acta Radiol Diag 20:180-192. 229. Pitzen P (1923). Das menschliche femur wahrend seiner entwicklung. Arch Anthrop 19:57-80. 230. Plaster RL, Schoenecker PL, Capelli AM (1991) Case report: Premature closure of the triradiate cartilage: A potential complication of pericapsular acetabuloplasty. J Pediatr Orthop 11: 676-678. 231. Poggi A (1890) Contributo alla cura cruenta della lussazione congenita coxo-femorale unilaterale. Arch Ortop 7:105 [Reprinted in translation in Clin Orthop Rel Res 98:5-7, 1974]. 232. Ponseti I (1944) Causes of failure in the treatment of congenital dislocation of the hip. J Bone Joint Surg 26:775-792. 233. Ponseti IV (1978) Growth and development of the acetabulum in the normal child. Anatomical, histological and roentgenographic studies. J Bone Joint Surg 60A:575-585.

234. Ponseti IV (1978) Morphology of the acetabulum in congenital dislocation of the hip. Gross, histological and roentgenographic studies. J Bone Joint Surg 60A:586-599. 235. Pool RD, Foster BK, Paterson DC (1986) Avascular necrosis in congenital hip dislocation: The significance of splintage. J Bone Joint Surg 68B:427-430. 236. Potocki (1905) Luxation congenitale de la hanche gauche chez un nouveau-ne. Rev Orthop 6:325-338. 237. Pous JG, Camous J-Y, E1Blidi S (1992) Cause and prevention of osteochondritis in congenital dislocation of the hip. Clin Orthop Rel Res 281:56-62. 238. Powell EN, Gerratana FJ, Gage JR (1986) Open reduction for congenital hip dislocation: The risk of avascular necrosis with three different approaches. J Pediatr Orthop 6:127-132. 239. Pravaz CG (1837) Du traitement de la luxation congenitale du femur. Bull Acad Med 2:579. 240. Putti V (1926) Congenital dislocation of the hip. Surg Gyn Obstet 42:449-452. 241. Putti V (1929) Early treatment of congenital dislocation of the hip. J Bone Joint Surg 11:798-809. 242. Putti V (1933) Early treatment of congenital dislocation of the hip. J Bone Joint Surg 15:16-21. 243. Putti V (1934) Resultati della cura incruenta della lussazione congenita dell'anca. Chir Deg Org di Movimento 20:93-112. 244. Quinn RH, Renshaw TS, DeLuca PA (1994) Preliminary traction in the treatment of developmental dislocation of the hip. J Pediatr Orthop 14:636-642. 245. Race C, Herring JA (1983) Congenital dislocation of the hip: An evaluation of closed reduction. J Pediatr Orthop 3:166-172. 246. Ralis Z, McKibbin B (1973) Changes in shape of the human hip joint during its development and their relation to its stability. J Bone Joint Surg 55B:780-785. 247. Ramsey PL, Lasser S, MacEwen GD (1976) Congenital dislocation of the hip. Use of the Pavlik harness in the child in the first six months of life. J Bone Joint Surg 58A: 1000-1004. 248. Reeves HA (1885) Bodily Deformities and Their Treatment. A Handbook of Practical Orthopedics, Chapter 18. pp. 292305, Philadelphia: P Blakiston. 249. Reichelt A, Hansen L (1978) The development of the dysplastic hip joint after an intertrochanteric varus rotation osteotomy. Internat Orthop 2:241-243. 250. Reikeras D, Bjerkreim I, Kolbenstvedt A (1983) Anteversion of the acetabulum and femoral neck in normals and in patients with osteoarthritis of the hip. Acta Orthop Scand 54:18-23. 251. Rejholec M, Stryhal F (1991) Behavior of the proximal femur during the treatment of congenital dysplasia of the hip: A clinical long-term study. J Pediatr Orthop 11:506-513. 252. Renshaw TS (1981) Inadequate reduction of congenital dislocation of the hip. J Bone Joint Surg 63A:1114-1121. 253. Robert H, Seringe R, Jr (1982) Les troubles de development de l'extremitie superieure du femur apres traitement de la luxation congenitale de hanche. Rev Chir Orthop 68:425439. 254. Robinson HJ, Jr, Shannon MA (1989) Avascular necrosis in congenital hip dysplasia: The effect of treatment. J Pediatr Orthop 9:293-303. 255. Roose PE, Chingren GI, Klaaren HE, Broock G (1979) Open reduction for congenital dislocation of the hip using the Ferguson procedure. J Bone Joint Surg 61A:915-921.

References

256. Roser W (1864) Die lehre von den spontanluxationen. Arch Heilkunde 5:542-545. 257. Roser W (1879) Ueber Angeborene Huftverrenkung. Arch Klin Chir 24:309-313. 258. Roth A, Gibson DA, Hall JE (1974) The experience of five orthopedic surgeons with innominate osteotomy in the treatment of congenital dislocation and subluxation of the hip. Clin Orthop Rel Res 98:178-182. 259. Rungee JL, Reinker KA (1992) Ossific nucleus eccentricity in congenital dislocation of the hip. J Pediatr Orthop Res 12: 61-67. 260. Sainton R (1893) Etude sur l'anatomie de la hanche chez l'enfant et sur la pathogenie de la luxation congenitale du femur, I. Rev Orthop 4:262-284. 261. Sainton R (1893) Etude sur l'anatomie de le hanche chez l'enfant et sur la pathogenie de la luxation congenitale du femur, II. Rev Orthop 4:352-382. 262. Salter RB (1961) Innominate osteotomy in the treatment of congenital dislocation and subluxation of the hip. J Bone Joint Surg 43B:518-539. 263. Salter RB (1968) Etiology, pathogenesis and possible prevention of congenital dislocation of the hip. Can Med Assoc J 98: 933-945. 264. Salter RB (1974) Osteotomy of the pelvis. Editorial comment. Clin Orthop Rel Res 98:2-4. 265. Salter RB, Dubos J-P (1974) The first 15 years personal experience with innominate osteotomy in the treatment of congenital dislocation and subluxation of the hip. Clin Orthop Rel Res 98:72-103. 266. Salter RB, Kostuik J, Dallas S (1969) Avascular necrosis of the femoral head as a complication of treatment for congenital dislocation of the hip in young children: A clinical and experimental investigation. Can J Surg 12:44-61. 267. Scaglietti O, Calandriello B (1962) Open reduction of congenital dislocation of the hip. J Bone Joint Surg 44B:257-283. 268. Schoenecker PL, Bitz DM, Whiteside LA (1978) The acute effect of position of immobilization on capital femoral epiphyseal blood flow. J Bone Joint Surg 60A:899-904. 269. Schoenecker PL, Lesker PA, Ogata K (1984) A dynamic canine model of experimental hip dysplasia. J Bone Joint Surg 66A:1281-1288. 270. Schuler P, Feltes E, Kienapfel H, Griss P (1990) Ultrasound examination for the early determination of dysplasia and congenital dislocation of neonatal hips. Clin Orthop Rel Res 258: 18-26. 271. Schulze K-J, Schmidt H, Hillig M (1980) Nachuntersuchungsergebnisse der azetabuloplastik nach Dega. Z Orthop 118: 905-914. 272. Schwartz DR (1965) Acetabular development after reduction of congenital dislocation of the hip. J Bone Joint Surg 47A: 705-714. 273. Scoles PV, Boyd A, Jones PK (1987) Roentgenographic parameters of the normal infant hip. J Pediatr Orthop 7:656-663. 274. Sedillot (1836) Luxations du femur en haut et en dehors, ou dans la fosse iliaque exteme. J des Connais Med-Chir 3:307-318. 275. Seringe R, Bonnet C, Katti E (1990) The pathogenesis and natural history of congenital dislocation of the hip. Fr J Orthop Surg 4:492-502 (Rev Chir Orthop 1990;76:391-402, 1990).

269

276. Seringe R, Cressaty J, Girard B, Francoual C (1981) L'examen orthopedique de 1500 nouveau-ne en maternite. Chir Pediatr 22:365-387. 277. Severin E (1939) Arthrography in congenital dislocation of the hip. Acta Orthop Scand 10:372-375. 278. Severin E (1941) Congenital dislocation of the hip joint. Late results of closed reduction and arthrographic studies of recent cases. Acta Chir Scand 84, Supp 63:1-142. 279. Severin E (1950) Congenital dislocation of the hip. Development of the joint after closed reduction. J Bone Joint Surg 32A:507-518. 280. Shapiro F (1998) Epiphyseal and physeal cartilage vascularization: A light microscopic and tritiated thymidine autoradiographic study of cartilage canals in newborn and young postnatal rabbit bone. Anat Rec 252:140-148. 281. Sievers (1927) Fortschr Geb Rongen 35:16-28. 282. Siffert RS (1981) Patterns of deformity of the developing hip. Clin Orthop Rel Res 160:14-29. 283. Smith BG, Millis MB, Hey LA, Jaramillo D, Kasser JR (1997). Post reduction computed tomography in developmental dislocation of the hip. Part II: predictive value for outcome. J Pediatr Orthop 17:631-636. 284. Smith WS, Ireton RJ, Coleman CR (1958) Sequelae of experimental dislocation of weight-beating ball-and-socket joint in a young growing animal. J Bone Joint Surg 40A:1121-1127. 285. Somerville EW (1982) Displacement of the Hip in Childhood. pp. 1-12, Berlin: Springer-Verlag. 286. Stalder G, Jani L (1978) Erbberatung bei orthopadischen erkrankungen. Z Orthop 116:555-557. 287. Stanisavljevic S (1964) Diagnosis and Treatment of Congenital Hip Pathology in the Newborn. Baltimore: Williams and Wilkins. 288. Steel HH (1973) Triple osteotomy of the innominate bone. J Bone Joint Surg 55A:343-350. 289. Steindler A (1950) On congenital dislocation of the hip. In: Post Graduate Lectures on Orthopedic Diagnosis and Indications, volume 1. pp. 216-242, Springfield, Illinois: CC Thomas. 290. Stevens PM, Coleman SS (1985) Coxa breva: Its pathogenesis and rationale for its management. J Pediatr Orthop 5:515-521. 291. Stewart RJ, Patterson CC, Mollan RAB (1986) Joint ossification of the normal capital femoral epiphysis. J Bone Surg 68B:653. 292. Strayer LM (1943) The embryology of the human hip joint. Yale J Biol Med 16:13-26. 293. Sullivan JA, Vanhoutte JJ, Remondino B (1982) Femoral neck anteversion in perspective. Clin Orthop Rel Res 183: 185-191. 294. Sundberg SB, Clark B, Foster BK (1986) Three-dimensional reformation of skeletal abnormalities using computed tomography. J Pediatr Orthop 6:416-420. 295. Sutherland DH, Greenfield R (1977) Double innominate osteotomy. J Bone Joint Surg 59A: 1082-1091. 296. Suzuki S (1995) Deformity of the pelvis in developmental dysplasia of the hip: Three-dimensional evaluation by means of magnetic resonance imaging. J Pediatr Orthop 15:812-816. 297. Suzuki S, Yamamuro T (1986)Correlation of fetal posture and congenital dislocation of the hip. Acta Orthop Scand 57: 81-84.

270

CHAPTER 3 9

Developmental Dysplasia of the Hip

298. Suzuki S, Yamamuro T (1990) Avascular necrosis in patients treated with the Pavlik harness for congenital dislocation of the hip. J Bone Joint Surg 72A:1048-1055. 299. Suzuki S, Yamamuro T (1993) The mechanical cause of congenital dislocation of the hip joint. Dynamic ultrasound study of 5 cases. Acta Orthop Scand 64:303-304. 300. Sylkin NN (1990) Die femurkopfnekrose bei luxationshuften und ihre prognose (eine literaturubersicht). Z Orthop 128: 584-591. 301. Szoke N, Kuhl L, Heinrichs J (1988) Ultrasound examination in the diagnosis of congenital hip dysplasia of newborns. J Pediatr Orthop 8:12-16. 302. Tanaka T, Yoshihashi Y, Miura T (1994) Changes in soft tissue interposition after reduction of developmental dislocation of the hip. J Pediatr Orthop 14:16-23. 303. Taussig G, Delor MH, Masse P (1976) Les alterations de croissance de l'extremite superieure du femur: leur apport a la connaissance de la croissance normale essai de deductions therapeutiques. Rev Chir Orthop 62:191-210. 304. Teot L, Dussault RG, Bosse JP, Pous JG (1983) Acetabuloplastie par greffon pedicule d'epiphyse de crete iliaque chez le chien en croissance. Internat Orthop 7:1-9. 305. Terjesen T, Runden TO, Johnsen HM (1991) Ultrasound in the diagnosis of congenital dysplasia and dislocation of the hip joints in children older than two years. Clin Orthop Rel Res 262:159-169. 306. Terjesen T, Runden TO, Tangerud A (1989) Ultrasonography and radiography of the hip in infants. Acta Orthop Scand 6: 651-660. 307. Terver S, Dillingham M, Parker B, Bjorke A, Bleck E, Levai JP, Teinturier P, Viallet JF (1982) Etude de 1'orientation reelle du cotyle grace a la tomodensiometrie. J Radiol 63:167-173. 308. Thomas CL, Gage JR, Ogden JA (1982) Treatment concepts for proximal femoral ischemic necrosis complicating congenital hip disease. J Bone Joint Surg 64A:817-828. 309. Thomas IH, Dunin AJ, Cole WG, Menelaus MB (1989) Avascular necrosis after open reduction for congenital dislocation of the hip: Analysis of causative factors and natural history. J Pediatr Orthop 9:525-531. 310. Toby EB, Koman LA, Bechtold RE (1985) Magnetic resonance imaging of pediatric hip disease. J Pediatr Orthop 5: 665-671. 311. Toby EB, Koman LA, Bechtold RE, Nicastro JN (1987) Postoperative computed tomographic evaluation of congenital hip dislocation. J Pediatr Orthop 7:667-670. 312. Tonnis D (1976) An evaluation of conservative and operative methods in the treatment of congenital hip dislocation. Clin Orthop Rel Res 119:76-88. 313. Tonnis D (1976) Normal values of the hip joint for the evaluation of X rays in children and adults. Clin Orthop Rel Res 119:39-47. 314. Tonnis D (1977) Indikation und wirksamkeit verschiedener operationsmethoden bei angeborener huftdysplasie. Z Orthop 115:726-737. 315. Tonnis D (1987) Congenital Dysplasia and Dislocation of the Hip. Berlin: Springer-Verlag. 316. Tonnis D, Bruning K, Heinecke A (1994). Lateral acetabular osteotomy. J Pediatr Orthop Part B 3:40-46. 317. Tonnis D (1990) Ischemic necrosis as a complication of treatment of CDH. Acta Orthop Belg 56:195-206.

318. Tonnis D (1990) Results of newborn screening for CDH with and without sonography and correlation of risk factors. J Pediatr Orthop 10:145-152. 319. Tonnis D (1990) Surgical treatment of congenital dislocation of the hip. Clin Orthop Rel Res 258:33-40. 320. Tonnis D, Sprafke K (1977) Neuew ergebnisse der weiterewntwickelten azetabuloplastik nach Lance in verbindung mit der detorsionsvarusierungs osteotomie. Z Orthop 115:743-752. 321. Tonnis D, Storch K, Ulbrich H (1990) Results of newborn screening for CDH with and without sonography and correlation of risk factors. J Pediatr Orthop 10:145-152. 322. Touzet, Chaumien JP, Rigault P, Padovani JP (1990) Le harnais de Pavlik dans la luxation congenitale de hanche avant un an. Acta Orthop Belg 56:141-147. 323. Trueta J (1957) The normal vascular anatomy of the human femoral head during growth. J Bone Joint Surg 39B:358-394. 324. Trueta J, Harrison MHM (1953) The normal vascular anatomy of the femoral head in adult man. J Bone Joint Surg 35B: 442-461. 325. Tschauner C, Klapsch W, Baumgartner A, Graf R (1994) "Reifungskurve" des sonographischen alpha-winkels nach graf unbehandelter huftgelenke im ersten lebensjahr. Z Orthop 132:502-504. 326. Tucker FR (1949) Arterial supply to the femoral head and its clinical importance. J Bone Joint Surg 31B:82-93. 327. Vartan CK (1945) The behaviour of the fetus in utero with special reference to the incidence of breech presentation at term. J Obstet Gyn Brit Emp 52:417-434. 328. Verneuil (1852) Luxation congeniale du femur par maladie articulaire chez un foetus a terme. Gas des Hop, Nov 11: 530-531. 329. Verneuil A (1890) Sur les luxations pretendues congenitales de la hanche. Rev Orthop 1:23-35. 330. Visser JD, Jonkes A, Hillen B (1982) Hip joint measurements with computerized tomography. J Pediatr Orthop 2:143-146. 331. von Lanz T (1951) Uber umwegige entwicklungen am menschlichen huftgelenk. Schweiz Med Wochen 81:1053-1056. 332. von Rosen S (1962) Diagnosis and treatment of congenital dislocation of the hip joint in the newborn. J Bone Joint Surg 44B :284-291. 333. Wagner VA, Gembruch V, Schmitt O, v Deimling V, Hansmann M (1994) Sonographische untersuchungen des fetalen huftgelenks. Z Orthop 132:497-501. 334. Walker JM (1980) Morphological variants in the human fetal hip joint. J Bone Joint Surg 62A: 1073-1082. 335. Walker JM (1981) Histological study of the fetal development of the human acetabulum and labrum: Significance in congenital hip disease. Yale J Biol Med 54:255-263. 336. Walker JM (1983) Comparison of normal and abnormal human fetal hip joints: A quantitative study with significance to congenital hip disease. J Pediatr Orthop 3:173-183. 337. Walker JM, Goldsmith CH (1981) Morphometric study of the fetal development of the human hip joint: Significance for congenital hip disease. Yale J Biol Med 54:411-437. 338. Walmsley T (1915) The epiphysis of the head of the femur. J Anat Physio149:434-440. 339. Walmsley T (1916) A note on the ligament of Weitbrecht. J Anat Physiol 51:61-64. 340. Watanabe RS (1974) Embryology of the human hip. Clin Orthop Rel Res 98:8-26.

References 341. Waters P, Kurica K, Hall J, Micheli LJ (1988) Salter innominate osteotomies in congenital dislocation of the hip. J Pediatr Orthop 8:650-655. 342. Weber M, Morgenthaler M (1994) Die behandlung der dysplasie-bzw, luxationshufte mit dem hochgips nach Fettweis. Z Orthop 132:260-271. 343. Wedge JH, Wasylenko MJ (1979) The natural history of congenital disease of the hip. J Bone Joint Surg 61B:334-338. 344. Weiner DS, Hoyt WA, O'Dell HW (1977) Congenital dislocation of the hip: The relationship of premanipulation traction and age to avascular necrosis of the femoral head. J Bone Joint Surg 59A:306-311. 345. Weiner LS, Kelley MA, Ulin RI, Wallach D (1993) Development of the acetabulum and hip: Computed tomography analysis of the axial plane. J Pediatr Orthop 13:421-425. 346. Weinstein SL (1987) Natural history of congenital hip dislocation (CDH) and hip dysplasia. Clin Orthop Rel Res 225: 62-76. 347. Weinstein SL, Ponseti IV (1979) Congenital dislocation of the hip. Open reduction through a medial approach. J Bone Joint Surg 61 A:221-227. 348. Werndorf R (1912) The pathology and therapy of congenital dislocations of the hip. Am J Orthop Surg 1-11. 349. Westin GW, Ilfeld FW, Provost J (1976) Total avascular necrosis of the capital femoral epiphysis in congenital dislocated hips. Clin Orthop Rel Res 119:93-98. 350. Wiberg G (1939) Relation between congenital subluxation of the hip and arthritis deformans. Acta Orthop Scand 10: 351-371. 351. Wiberg G (1941) Studien uber das normale arthrographiebild des huftgelenks bei kleinkindern. Z Orthop 72:35-47. 352. Wientroub S, Green I, Terdiman R, Weissman SL (1979) Growth and development of congenitally dislocated hips reduced in early infancy. J Bone Joint Surg 61A:125-130.

271

353. Wientroub S, Tardiman R, Green R, Salama R, Weissman SL (1981) The development of the normal infantile hip as expressed by radiological measurements. Internat Orthop 4: 239-241. 354. Wilkinson JA (1963) Prime factors in the etiology of congenital dislocation of the hip. J Bone Joint Surg 45B:268-283. 355. Wilkinson JA (1966) Breech malposition and intrauterine dislocations. Proc Roy Soc Med 59:1106-1108. 356. Wilkinson JA (1972) A postnatal survey for congenital displacement of the hip. J Bone Joint Surg 54B:40-49. 357. Wilkinson JA (1992) Etiologic factors in congenital displacement of the hip and myelodysplasia. Clin Orthop Rel Res 281: 75-83. 358. Williams GD, Martin CH, Mclntire LR (1934) Origin of the deep and circumflex femoral group of arteries. Anat Rec 60: 189-196. 359. Windhager R, Lack W, Schiller Ch, Kotz P (1990) Die beckenosteotomie nach Salter in der behandlung der kongenitalen huftludysplasie unter besonderer berucksichtigung der beckenkippung. Z Orthop 128:575-583. 360. Wolcott WE (1943) The evolution of the circulation in the developing femoral head and neck. Surg Gynec Obstet 77: 61-68. 361. Yamamuro T, Chene S-H (1975) A radiological study on the development of the hip joint in normal infants. J Jap Orthop Ass 49:421-439. 362. Zahradnicek J (1934) Beitrag zur reposition der hohen angeborenen buftverrenkung. Arch Klin Chir 180:353-368. 363. Zanoli R (1966) Early diagnosis and late results of the early treatment of congenital dysplasia of the hip. In: Excerpta Medica, International Congress Series, number 116, pp. 33-35, Amsterdam, Excerpta Medica.

CHAPTER

4

Le99-Calve-Perthes Disease I.

Definition

VII.

II. Original Recognition of Disorder III. Clinical Profile IV. Early Pathologic Reports of Cell and Tissue Changes in LeggCalve-Perthes Disease V. Subsequent Pathologic Reports with Better Defined Clinical and Radiographic Correlations VI. Early Correlation of Radiographic with Histopathologic and Clinical Features of Legg-Perthes from the Incipient Stage to the Residual Stage

VIII. IX. X. XI.

I. DEFINITION

Pathoanatomic Changes and Their Relation to Clinical, Radiologic, and Other Imaging Findings Lower Extremity Length Discrepancies with Legg-Perthes Disease Prognostic Indicators during the Active Disease Process

Classifications Defining Results Based on Appearances at Skeletal Maturity at the End of Repair Treatment Approaches to Legg-Perthes Disease

Waldenstrom from Stockholm, Sweden (275), Jacques Calve from Berck, France (33), Georg Perthes from Leipzig, Germany (211), and Arthur Legg from Boston, USA (168). Paul Sourdat, working in the same hospital as Calve, described and illustrated the radiographic changes of the hip indicative of a previously uncategorized disorder of childhood in his thesis (Paris) and in an accompanying article in 1910 (256). Considerable controversy followed over the next 25 years concerning which individual had priority in making the initial finding of what within the span of a few years became recognized as a distinct entity. Sundt outlined the chronology in full detail, including the dates on which the works initially were presented at scientific meetings as well as their publication dates (260). Waldenstrom first presented his work in March 1909 at a surgical conference in Stockholm. He accurately described the radiologic and clinical picture, but his priority of recognition was minimized somewhat because he attributed the disorder to a mild form of tuberculosis. In June 1909, Legg presented his work at a meeting in Hartford, Connecticut, and in his thesis in July 1909, Sourdat reported on eight abnormal cases of childhood hip disease as seen radiographically as part of a larger study on childhood hip disease, with some cases illustrated in an article "La coxalgie en radiographie" published in 1910 in Archives Provinciales de Chirurgie (256). The clinical and radiologic findings of his subset of disorders were described as coxa vara but came to be recognized as coxa plana or Legg-Perthes. Waldenstrom published his article, "Der Obere Tuberkulose Collumherd," in the Zeitschrift fur Orthopadische Chirurgie in 1909 (275). Legg published his article "An obscure affection of the hip-joint" in the Boston Medical and Surgical Journal on February 17, 1910 (168). Calve published his

Legg-Calve-Perthes disease is a disorder of the immature hip in which there is necrosis and subsequent repair of the bone and marrow of the femoral capital epiphyseal ossification center, accompanied by a variable amount of damage to the epiphyseal and physeal cartilage and by a range of secondary changes in the shape and size of the articular cartilage, the epiphyseal model of the femoral head, the bone of the femoral head, the physeal cartilage, and the acetabulum. The structural changes of the femoral head and acetabulum at skeletal maturity can vary from near normal to highly irregular, the more severe of which predispose the hip to degenerative arthritis in later years.

II. O R I G I N A L R E C O G N I T I O N OF DISORDER

A. General Review The disorder was recognized shortly after the development of radiography when structural abnormalities of the proximal femur were noted in children who presented with relatively minor limping, discomfort, and a diminished hip range of motion. Clinical and radiographic correlation enabled the new disorder to be distinguished from previous known causes of hip pathology in children such as fracture, rickets, septic arthritis, and tuberculous arthritis. Papers describing the clinical and radiographic findings along with comments on the possible etiology were presented and published within a very short period of time in 1909 and 1910 by Henning 272

SECTION II ~ Original Recognition o f Disorder

article "Sur une form particuliere pseudo-coxalgie. Greffee sur des deformations characteristiques de l'extremite superieure du femur" in the Revue Chirurgie in July 1910 (33). Finally, Perthes published his article "Uber arthritis deformans juvenilis" in the Deutsche Zeitschrififur Chirurgie in October, 1910 (211). The disorder shortly became known in most countries as Legg-Calve-Perthes disease, but LeggPerthes disease, Calve disease, or simply Perthes disease also commonly were used. Waldenstrom used the term coxa plana to describe the disorder, based on one of the characteristic early radiographic features of the secondary ossification center of the proximal femur (276). Calve, in 1921, said he preferred the term "coxa plana" and suggested that the disorder should be called "coxa plana" (34). Legg also commented on "coxa plana" as a desirable term (172). In addition, he had coined the term "osteochondral trophopathy," but fortunately it never caught on (169).

B. Legg Legg's brief article was accompanied by four radiographs showing various stages of the deformity (168). He presented five cases and characterized the following general facts, which indicated the age of occurrence from 5 to 8 years, a history of injury and limping, radiographic thickening about the neck of the femur, absence of pain and constitutional symptoms, little or no spasms, and no limb shortening. He discussed the possible causes of the disorder, ruling out congenital dislocation of the hip, rickets, and syphilis, but considering an indirect cause by "injury or displacement at the epiphyseal line which could then affect nutrition of the head," which he noted came mostly from the neck region. The poorly nourished epiphysis in relating to the acetabulum then would become flattened. He commented on the hyperemic condition in the neck as a result of "a deranged circulation" and discussed briefly that an underlying infection might be the root cause. Legg concluded, however, that "it does not seem probable to me that the change in the head in this case is secondary to the infection in the neck for we see many cases of infection in the neck, and in none of these have I seen the condition described present in the head." In his final paragraph, he indicated that he could "make no claim to any definite conclusion" as to causation and hoped that "by further study their true etiology may be determined." Over succeeding years, Legg came to support more strongly the traumatic conception of etiology with damage to the epiphyseal blood supply occurring due to occult or subclinical physeal movement and vessel irritation as they passed upward from the surface of the neck, over the physis, and into the head. Legg returned to the question of etiology when he reviewed 75 cases showing the disorder (170). Of these 25 gave a history of an insidious onset, 26 gave a history of distinct trauma preceding the limp, and 24 showed the disorder after reduction of a congenital dislocation of the hip.

273

In his review of possible disorders, including those mentioned by other authors, he convincingly ruled out the possibilities that the disorder was secondary to preexisting rickets or that it was of a congenital origin or due to infection. He noted that Allison and Moody also felt that the disorder "is a disturbance of the line of epiphyseal growth and that it depends for its typical development upon changes in circulation which destroy the nice balance which exists between the metaphysis and epiphysis in growing bones" (2). Legg continued to support the theory he advanced in his initial work and indicated again that "a trauma about the hip may cause an injury at the epiphyseal line whereby the nutrient vessels to the epiphysis are blocked and atrophy and flattening follow." The impression, therefore, that the disorder was "due to a circulatory disturbance" was present from the earliest descriptions, although it was unclear whether this was due to trauma or some other cause in the works of various writers. Legg himself felt that the disorder generally was due to trauma. Legg later described 40 cases in "The End Results of Coxa Plana" that he had followed to skeletal maturity since his original description (171). He clearly noted that "there are different types of end results" and cautioned that it often was not until skeletal maturity that the ultimate change in structure was finalized. The results appeared better in the "mushroom"-shaped head, examples of which show coxa magna with a generally spherical or ovoid head, than in the "cap" type in which the secondary ossification center initially was flattened and narrower than the adjacent neck and in which radiographs often showed "fragmentation of the epiphyseal bone center." Legg's observations on treatment remain interesting. He indicated that "it has been my experience however that relief from weight beating has in no way affected the end result, that is in those cases in which I have not allowed weight bearing, the end result has not varied in any way from those in which weight bearing on the affected leg throughout the disease has been allowed." He stated in 1916 that "there is never a necessity for operation in the disease" (169). C. Calve Calve presented 10 cases, which he was able to subdivide from a group of 566 hip disorders (33). He was able to make his distinction from other disorders based on differences in the clinical and radiographic appearance. The articular response to the disorder was chronic or subacute but of short duration and improved with diminution of activity. Radiographic deformities that characterized the disorder involved coxa vara, enlargement of the femoral head, atrophy and deformation of the secondary ossification center of the head, and a total absence of bony destruction. The age of occurrence was between 3.5 and 10 years. The general state of health in all was excellent. Many had the initial clinical appearance of rickets, but one never found any signs of

274

CHAPTER 4 9 Le~ly-Calve-Perrhes Disease

hereditary syphilis. All presented with mild signs of hip arthritis with some limitation of movement, spontaneous pain, and muscular atrophy. The two constant radiographic signs were a certain degree of coxa vara and bony hypertrophy of the femoral head. The coxa vara ranged between 90 and 120 ~ and hypertrophy characterized both the head and neck of the proximal femur. The neck was shorter as well as wider than normal and the head was larger. There were atrophy and deformation of the secondary ossification center bone. The deformation always was seen and in his opinion constituted the most important characteristic of the disorder. In each case, the normal appearance of the bony epiphysis of the femoral head was modified greatly. The secondary ossification center was flattened, lamellated, and smaller than normal initially and did not cover the entire upper end of the femur as in the normal. Calve also noted that the physeal growth cartilage line was winding, irregular, and strongly convex within. In some instances the secondary ossification center scarcely was visible and almost linear in appearance. It was not a uniform bone mass as in the normal but was subdivided into two or more parts disseminated within the central cartilage mass. (This appearance now is referred to as fragmentation.) Gradually the small bony island increased in size and eventually reconstituted a single mass. There was a total absence of wear in either the bone or the cartilage, and the articular space in particular was maintained. Recuperation following diagnosis of the disorder in a child was relatively quick with movements returning shortly, although abduction particularly with large degrees of coxa vara appeared to be most limited. The spontaneous pain also tended to disappear relatively shortly. In this and a subsequent article, Calve made reference to the work of his colleague Sourdat, who published findings on hip radiography in his Paris thesis in July 1909 and in a related article. Calve indicated that in no instance was there ever any trace of an abscess. He felt that the disorder represented a thoroughly transient arthritis of short duration, which developed at the superior end of the femur and was followed by coxa vara, enlargement of the femoral head, and atrophy of the secondary ossification center of the proximal femur but with conservation of the normal relationships of the articular surfaces of the joint. The deformation in the bone actually occurred prior to the symptoms. One would have to consider tuberculosis of the hip as a cause, and in fact that was the initial clinical diagnosis prior to radiography in each case. Calve outlined in great detail the major differences in this disorder from tuberculosis, feeling that a tuberculosis diagnosis was erroneous and had to be rejected as a cause for the disorder. Associated findings of a tuberculous lesion such as ulcer, dystrophic changes, and soft tissue invasion were never seen, and the course appeared to be much shorter than that seen in tuberculosis. There was never any trace of an abscess nor did the disorder lead to ankylosis. He even discarded the possibility that the disorder represented an abnormal form of tuberculosis. He rejected the possibility of hereditary syphilis, arthritis of infancy, and other deformities

such as enchondromas, exostoses, or primary coxa vara. The possibility of a rickets disorder was raised; in rickets, there was curvature of the neck but the epiphysis itself was normal and the head retained its shape and size. A purely mechanical cause did not appear to explain the total phenomenon. Calve felt that the development of this particular anomaly perhaps could be due to the presence of abnormal and retarded osteogenesis, and that was about the closest he felt he could get to an explanation. He summarized his discussion of etiology by stating, in a straightforward fashion, that it was impossible to reach a conclusion as to the cause but that the entity described was a type of hip arthritis that did not correspond to any previous description. Calve then presented 10 case reports with some radiographs but with each case primarily illustrated by an artist's drawing of the radiographic features of the pelvis and proximal femur. Each of the illustrations is readily recognizable as a Legg-Calve-Perthes disorder. The ages involved ranged from 3.5 to 10 years and averaged 7 years. Calve summarized two illustrative cases in 1921 (34-36). In both, there were radiographic studies over a several-year period including one in which a normal X ray was taken 1 year prior to the onset of symptoms and another in which there was an entire absence of bone in the initial radiograph in the proximal femoral capital epiphysis and then reconstitution of bone over a several-year period. He indicated that the development of the disorder was silent clinically and preceded symptoms by some period of time. In addition, the destruction of the ossification center could be total but regeneration took place regardless, and eventually it resulted in regeneration to a voluminous epiphyseal nucleus. The disorder involved partial or total destruction of the epiphyseal bone nucleus "without injury to the neighboring articulation." He summarized with five clear conclusions that remain accurate today (34, 35). 1) Osteochondritis is not a congenital affection; 2) The phase of the invasion of the epiphyseal nucleus is latent from the clinical viewpoint; 3) The clinical phase corresponds to the phase when the child first begins to complain of pain and is considerably later than the real beginning of the trouble; 4) At the beginning of the clinical phase, there is a corresponding radiographic picture showing an established and characteristic lesion--a laminated and fragmented epiphysis; 5) That the regeneration of the osseous epiphyseal nucleus occurs progressively as the osseous fragments augment in volume, approach each other, reunite one by one, and finally form a single mass. This regeneration continues through the following years and tends to a return to the normal form. He continued to stress that the term coxa plana was descriptive only, and the etiology remained controversial and the pathogenesis obscure. The disorder, however, clearly was acquired. Calve was never an advocate of specific or dramatic therapy. He indicated the need only for a simple treatment, namely, to rest the child during a several-week period early on when discomfort was greatest and then rest again if discomfort recurred.

SECTION II ~ Original Recognition o f Disorder

Calve summarized the first 12 years of work on the disorder in a detailed presentation (36). The age of occurrence was between 5 and 10 years, and the disorder presented with some discomfort and some slightly decreased ranges of hip movement. Abduction, in particular, was affected. The diagnosis was made definitively by radiographs. Characteristic of the disorder was the dissociation between the often marked and abnormal radiographic findings and the relatively benign clinical appearance of the affected child. The articular space tended to be enlarged radiographically due to relatively more cartilage and relatively less bone. He stressed that the overall volume of the head remained relatively the same, although the bony mass was smaller than on the normal side initially. The femoral neck almost always was wider than normal in radiographs as well. As the disorder evolved, the radiographic appearance of the secondary ossification center was flattened and then fragmented. Regeneration of bone played a major role in the eventual repair phenomenon. Calve recognized that in many instances, although repair occurred, the head itself was not fully reconstituted to a normal form. Often the femoral head repaired itself such that it was enlarged and flattened and thus poorly adapted to the adjacent acetabulum. It was that degree of nonadaptation of the large head and relatively small acetabulum that led to subsequent discomfort. Calve thus clearly pointed out that this subsequent problem in coxa plana was as much due to imperfect repair as it was due to the fact that repair did not occur.

D. Perthes Perthes reported initially on six of his own cases, one of which was bilateral (211, 212). Perthes noted his first case in an 11-year-old boy early in 1909 and within 1 year reported that he had seen six similar cases. He reported on the growing awareness of the fact that this represented a newly appreciated disorder with characteristic clinical and radiographic findings. Many childhood hip ailments previously diagnosed as coxa vara or tuberculosis now could be distinguished as being due to this new disorder. The first and most important changes were seen initially in the head of the femur, with five of the seven initially affected hips showing deformity of the proximal epiphysis, whereas in two, the epiphysis (by which he meant the secondary ossification center) had disappeared except for a small residue. Perthes noted the first change to be flattening of the superior portion of the head of the femur with a decrease in thickness of the bony epiphysis. The superior and medial portions of the epiphyseal line formed a marked angle with each other, which was close to a fight angle. He noted clinically that flexion virtually was full, whereas limitation of motion was most marked in terms of abduction, adduction, and rotation. The round shape of the femoral head had changed to that of a truncated cone. In the early stages, the only changes were in the proximal femoral epiphysis with the neck radiologically normal. With time, the proximal epiphysis of the femur disappeared al-

275

most completely, and with further time, the head of the femur assumed a mushroom shape. The neck eventually was involved, appearing shortened and thickened, whereas the femoral head eventually lost its normal shape and the proximal femur showed a varus position (211-214). Perthes also noted the participation of the acetabulum with enlargement in most of the severe cases. The head of the femur was noted to be positioned somewhat laterally and there was relative overgrowth of the greater trochanter. Perthes distinguished the symptoms from those of the more commonly appreciated tuberculosis. Among the key differentiating features were the facts that motion in the early stages was not limited in all directions but often only in one, pain was not a prominent early symptom, there was shortening of the limb caused by the deformity, and a positive Trendelenburg gait occurred. He offered the opinion that the frequency and importance of the disorder were much greater than noted previously and that trauma did not play a role in causation. The first pathological changes involved flattening of the proximal femoral epiphysis with progressive loss of height of the epiphysis with time. Due to the fact that the disorder had just been recognized and described there were few guidelines to therapy. Perthes stressed that immobilization was to be avoided, with passive range of motion recommended. In a second more detailed paper in 1913, Perthes further clarified the clinical and radiographic criteria involving 15 new cases (213). He underscored the importance of the initial site of pathology in the subchondral bone in differentiating the lesion from osteoarthritis, which primarily affected the articular cartilage. In the second work, he named the disorder osteochondritis deformans juvenilis, replacing the term arthritis deformans juvenilis from his initial work (213).

E. Waldenstrom Waldenstrom presented his initial work in 1909 (275). Although he defined the disorder as being tuberculous in nature, the drawings accompanying each of his cases are readily recognizable as a Legg-Calve-Perthes disorder or, as he subsequently referred to it, coxa plana. He presented seven of his own cases and three of his colleague SindigLarsen. He noted that the head maintained its articular cartilage coveting at all times and that the acetabulum generally changed its shape secondarily, adapting to the altered oval shape of the head. As an addendum, Waldenstrom referred to a radiographic study of the hip by Sourdat in which this new variant of childhood hip disease was illustrated. The latter was unable to give the etiology, but in no case did he list tuberculosis as a cause. Waldenstrom continued to publish his updated observations on the disorder for several years (276-280). He subsequently divided changes of coxa plana into four stages (278). Stage I, which he referred to as the evolutionary period, could be as long as 3-4 years and was itself divided into two parts: (a) the initial stage, which lasted for 6 months to 1 year. The epiphysis, by which was meant the secondary ossification center, became dense,

276

CHAPTER 4 ~ Leyy--Calve--Perrhes Disease

flattened, and uneven at its margin. The neck also frequently had areas of decalcification especially just adjacent to the epiphyseal growth plate. The articular cartilage was felt to be of normal height. (b) The fragmentation stage, which lasted for an additional 2-3 years. During this stage, the secondary ossification center became extremely flattened and divided and often was in three or more separate segments. It generally was smaller than that on the opposite side. Stage II, the healing period, continued from 1 to 2 years as bone of the secondary ossification center became reconstituted and uniform in appearance and density. Stage III, the growing period to the conclusion of normal growth, was the period in which the coxa plana reached final form. Stage IV, the definite stage, was that in which the appearance of the acetabulum and head at skeletal maturity represented the final result. Waldenstrom indicated the extreme variability of the end results. Eventually some flattening of the head and acetabulum occurred in all. It was mainly the anterior and superior portion of the head that was enlarged, frequently to the extent that it lay outside the normal confines of the joint. In terms of the degree of eventual deformity, he divided coxa plana into three groups. In the first group, the head preserved a rounded form and structural delineation of the head, neck, and greater trochanter was seen clearly. In the second group, the upper and frontal part of the enlarged head approached the greater trochanter and was seen in a lateral view to be greatly enlarged and lying outside the joint. In the third group, the joint surface of the head was uneven and frequently excavated. The upper pole of the head frequently was triangular in shape and often lower than the tip of the greater trochanter. Waldenstrom believed that the flattening of the secondary ossification center was the principal sign of the pathological process and was present from the earliest stages. He reviewed his concepts for development after a period of 30 years of involvement with treating the disorder (279, 280). In some instances, when the child presented with only a few weeks of limping, the initial radiograph was normal, at least on the anteroposterior films. He stressed, however, that it was the lateral view that was of most importance in the earliest time periods. On one lateral radiograph, he demonstrated nicely the crescent sign with continuity of shape of the articular surface laterally along with a small rim of subchondral bone. He clearly commented on the narrow strip of subchondral bone in its normal position adjacent to the articular cartilage. Underneath this strip, he noted the radiolucent zone caused by a "subchondral resorption of necrotic bone." In studying cases that eventually became bilateral, Waldenstrom was able to note that previously normal hip joints were present that subsequently developed the disorder, thus concluding that the condition arose in a joint that primarily was normal. He again stressed his feeling that the patients should be treated conservatively, by which he meant 1-2 months of rest in bed and then periodic use of crutches.

Waldenstrom presented his most detailed long-term considerations in 1922 (277) and 1923 (278). "On Coxa Plana" (278) is particularly valuable in terms of the clinical as well as radiographic progression. Waldenstrom commented that "the epiphyseal line becomes more crooked" and that "not infrequently the acetabulum undergoes similar changes." His comments regarding therapy remain of interest. His early impression was that "no intervention should be undertaken. No improvement is to be expected from it." This was because the prognosis was so good that the possibility existed that "operation of the joint may have recourse to great harm." He mentioned the possibility of permanent extension of the hip with rest in bed or use of an ambulant brace. These treatments were problematic because it could take 5 or 6 years for the entire disorder to evolve. The question then was raised as to whether the treatment had to be continued for the whole time, and if not which time frame was most important within which treatment should occur. Waldenstrom basically felt that decreased activity would be a sufficient approach so as not to further stress the hip. He commented that "the development in treated and non-treated cases gives us no clue." Other than bed rest during the few weeks of the acute phase, he primarily recommended observation.

E Sourdat Sourdat wrote in detail on the use of radiography to define hip disease, particularly in the childhood years (256). Among the examples shown, he pointed out that "there exists as well disturbances of ossification of the head, resulting in atrophy of the epiphyseal head, thickening of the neck which deviated into varus, broadening of the articular space, which simulated tuberculosis of the hip and were taken for that prior to the radiographic examination." Sourdat felt that the cases probably were not that rare. Two unilateral cases were illustrated in his article: In one case, he pointed out the abnormal shape of the femoral head epiphysis (broad and flattened bone), the largeness of the space between the secondary ossification center and the medial pelvic sourcil, and the thickness of the neck and its varus position. The other case illustrated a fragmented head with a thickened neck. Several of the early articles were illustrated by line drawings of original radiographs; a few of these from the works of Sourdat (256), Perthes (210), and Waldenstrom (275) are shown in Figs. 1A-1E.

III. C L I N I C A L P R O F I L E A. General Features The primary etiology remains unknown, but the clinical and radiographic changes are secondary (1) to an avascular occurrence or series of occurrences that cause variable necrosis of the secondary ossification center of the proximal femur

S E C T I O N ill ~ Clinical Profile

277

FIGURE 1 Linedrawings of original radiographs were used in early articles to outline the developing hip deformity. Parts (A) and (B) from Sourdat (256) show two separate unilateral cases clearly recognizable as the Legg-Calve-Perthes disorder. Parts (C) and (D) from Perthes (212) show an irregular shape of the secondary ossification center of the femoral head, lateral displacement of the femoral head and neck, early lateral lysis of the secondary ossification center, and a metaphyseal cyst (C) and an enlarged ovoid-shaped head and an acetabulum that has reconstituted to a congruent position in relation to the misshapen head (D). In part (E) three illustrations from different patients from the work of Waldenstrom (275) show characteristic appearances through varying stages of the disorder.

and of the adjacent epiphyseal and physeal cartilage and (2) to biomechanical and biological responses during the long repair phase. It has an approximately 4:1 male predilection and is bilateral in 10-18% (68). The bilaterality incidence in seven series between 1949 and 1978 ranged from 9 to 17% (43). It can present between the ages of 2.5 and 13 years, but the large majority of patients present between 5 and 10 years of age. If the initial radiographs reveal bilateral involvement, with both sides showing the same extent of involvement and with clinical symptoms mild to absent, there should be a high suspicion that the child has a skeletal dysplasia, commonly multiple epiphyseal dysplasia. Bilateral involvement in Legg-Perthes disease does not occur simultaneously but at differing times, with the second hip involved several months to a few years after the first. The child usually will present with a several-week to several-month history of limping and only minimal to moderate discomfort, which can be in either the hip or the groin region or sometimes in the distal medial thigh adjacent to the knee with the hip area experiencing no discomfort. A definable traumatic episode rarely if ever is associated with the presentation of this disease. In a very

small proportion of patients an initial episode of synovitis will have occurred, with the Legg-Perthes condition not revealing itself for several weeks or months. The bone always will repair itself, but the shape of the femoral head varies at skeletal maturity from normally spherical to markedly irregular. The disease cycle from initiation to full reconstitution is very slow. Edgren, in an assessment of 326 affected hips in 276 patients, documented the mean duration of the disease to be 4 years 4 months (68). The mean duration of the initial stage was 5.6 months, the fragmentation stage was 10.8 months, and the reparative stage was 32 months.

B. Epidemiologic Features of Legg-CalvePerthes Disease Goff produced a detailed review of the Legg-Perthes literature from 1883 to 1960, particularly in relation to theories of causation (91). As is frequently the case, there were some papers in the decades prior to specific definition of the disorder in 1909-1910 in which the observers began to develop an awareness that some seemingly severe childhood

278

CHAPTER 4 ~

Le99--Calve-PerthesDisease

hip disorders recovered spontaneously, thus differentiating themselves from the infectious and dislocation etiologies appreciated at the time. Goff felt the disorder to be complex, with multiple causes acting on a genetically conditioned individual. He noted the deceleration of growth prior to onset of the disorder. Others have linked Legg-Perthes to the spectrum of developmental bone disorders previously referred to as the osteochondroses--a term not favored currently--and sought to describe systemic disorders possibly underlying each of the disorders (64, 126, 249). Several features of Legg-Perthes disease have been recognized over several decades of study. 1. ROLE OF DIMINISHED STATURE AND DELAYED BONE MATURATION Retrospective studies have confirmed the clinical impression that many patients with this condition are small in stature. A review from our hospital demonstrated all patients to be within the normal height range, but height at disease presentation was below the mean in 53% of the children (246). If those at the mean and below were considered, 68% were in this range with only 32% above the mean. Assessment of standing height at skeletal maturation demonstrated that the same relative distribution as in the initial assessment persisted. Fifty-nine percent of the patients had a final standing height below the mean, although the vast majority still remained within the normal range. Although skeletal maturation occurred late, a compensatory "catch-up" phenomenon in regard to total height was not seen. At the time of initial assessment shortly after disease presentation, there was a marked delay in skeletal age as related to chronologic age in the large majority of patients, with 83% demonstrating a skeletal age less than the chronologic age by 3 months or more, 11% demonstrating chronologic and skeletal ages within 3 months of each other, and only 6% demonstrating a skeletal age greater than the chronologic age. Sixty-eight percent of the patients with skeletal age retardation at the time of presentation or shortly after were from 1 to 3 or more years delayed (246). Thus, it is surmised that Legg-Calve-Perthes disease is caused by a local phenomenon at the hip superimposed on a generalized skeletal condition, which in itself appears to have no clinical significance aside from a tendency to stature within the low-normal range. A very large number of studies from several countries has confirmed the delay of skeletal maturation in patients with Legg-Perthes disease. Carpal maturation generally is markedly delayed, proceeds very slowly, and precedes the clinical and radiologic onset of disease. As early as 1940, Gill found many with retardation of bone maturation in Perthes (88). The percentage of patients with bone retardation increased with the increasing precision of observation. Bloch-Michel et al. found a delay in maturation of bone age in 75% in a series of 54 cases (20), with similar findings by Fisher (77), Reichelt et al. (227), Lauritzen (166), Harrison et al. (105),

and Burwell et al. (29), as well as our own series mentioned earlier (246). DeGuembecker and Duriez studied 321 cases, assessing maturation by the Greulich and Pyle atlas of carpal development (58). They noted 78% of patients with bone age less than chronologic age, bone age less than or equal to chronologic age in a further 19%, and only 3% in whom the bone age was greater than the chronologic age. Bohr performed a lengthy epidemiologic assessment of delay in skeletal age in relation to Perthes, studying 223 children with the disorder (21). In the 154 patients from Denmark, the delay in skeletal age at diagnosis was a mean of 19.5 months in boys and 15.7 months in gifts. The total (male and female) mean delay in skeletal age in the patients from Denmark was 18.8 months. In 25 patients from the Faroe Islands it was 16.2 months, and in 44 patients from Greenland it was 12.6 months. Katz and Siffert noted almost 30% of 116 children with a significant delay in skeletal maturation of greater than 2 standard deviations (SD) (146). Girandy and Osman found 119 of 184 (65%) Perthes patients with markedly retarded skeletal maturation (greater than 2 SD below the mean), 42 (23%) 1-2 SD below the mean, and 23 (12%) slightly below the mean, with no patients showing maturation greater than the mean, for age (89). Although a delay in skeletal maturation was present during the initial stages in most children with Legg-Perthes, this did not seem to influence e the duration of the disease or the ultimate result. Kristmundsdottir et al. carefully studied carpal bone development in 27 girls with Perthes disease by consecutive radiographs and noted delayed skeletal maturation to be both frequent and considerable at the time of initial occurrence (160). The findings then were extended to a longitudinal study of carpal bone development in 125 children with Perthes (98 boys, and 27 girls). The bone age delay was severely abnormal at 3-5 years chronologic age. The mean age at diagnosis for 34 boys with skeletal standstill was 4.49 _+ 1.07 years compared to 7.28 _+ 2.40 years for 51 boys without such standstill (161). The conclusion reached by each of the investigators in the field is that patients with Legg-Perthes showed elements of a generalized endochondral disorder manifesting itself as a delay in the transformation of cartilaginous epiphyses to bony epiphyses and that Legg-Perthes itself represented a local hip-related complication of a mild generalized chondrodystrophy. An anthropometric study of massive detail in 232 children with Perthes disease was performed by Burwell et al., in which mildly but statistically significantly impaired and disproportionate growth in patients with the disorder was noted (29). There was disproportionate skeletal growth although it was the forearm, hand, and foot that were more impaired than more proximal segments. Cannon et al. documented markedly altered growth patterns in children with Perthes (39). Though noting height and weight well below the mean at the time of onset of the disorder, subsequent measurements showed that, from that point forward, growth velocities continued to be elevated significantly throughout

S E C T I O N I!1 9 Clinical Profile

development compared to control groups. For much of childhood, their growth progressed at a greater velocity than that in the mean population. They postulated that this growth rate in association with a delay in epiphyseal bone formation or skeletal maturation might well produce susceptibility of the epiphyseal regions to trauma and minor ischemia. Most convincing is a study by Eckerwall et al. based on the fact that it represents a longitudinal study of the same children followed throughout the growth period. They assessed 110 children with Legg-Calve-Perthes disease and were able to utilize data from birth to skeletal maturity. Most other studies cited have been cross-sectional, such that the same individual was not followed from beginning to end, but rather the data available for the Legg-Perthes children were compared with other standards from normal children studied separately. They concluded that the children were slightly shorter at birth and that they remained short throughout the entire growth period to maturity. At skeletal maturity, the boys were 4.4 cm and the girls 2.5 cm below the reference mean. Their growth velocity, however, was normal at the time of diagnosis, at the prepubertal time, and during puberty (67). Eckerwall et al. concluded that there was no abnormal growth pattern from the onset of symptoms to the time of diagnosis. They noted normal growth and growth velocity at diagnosis and a continuance of the slightly reduced height at maturity in children with Legg-Perthes. 2. NONHEREDITARY DISORDER

Genetic causation has not been implicated in Perthes disease. Wynne-Davies and Gormley studied 217 patients along with their parents and unaffected siblings and showed an extremely low frequency of Perthes disease among relatives with no obvious pattern of inheritance (287). Harper et al. also showed a relatively small likelihood of any genetic predisposition to the disorder (102). 3. SUBTLE ABNORMALITIESOF CONTRALATERALHIP Harrison and Blakemore studied radiographs of the apparently normal hip on the contralateral side of 153 children with unilateral Perthes disease (106). In 48.4% of patients, they felt that some irregularity of the surface outline of the bone of the femoral capital epiphysis was noted, with the vast majority of those present at the time of initial radiograph. These changes were present in only 10.4% of a control series of 153 children. Similar findings of subtle abnormalities in the opposite hip were reported by Katz (141) in 33 of 190 children with unilateral Perthes disease (17%), and Chivabongs (47) indicated that changes in the "normal" hip of patients with unilateral Perthes disease were found in many instances. The authors concluded that the changes in the otherwise normal epiphysis were consistent with the theory that the femoral capital epiphysis in a young child may be vulnerable to stress, with contour irregularities at one end of the spectrum and frank Perthes disease at the other.

279

4. ABNORMALITIES OF THROMBOLYSIS

It has been noted by Glueck et al. that 33 (75%) of 44 children with Legg-Perthes disease had coagulation abnormalities (90). Twenty-three had thrombophilia (a deficiency of antithrombotic factors C or S with an increased tendency toward thrombosis), 19 of these 23 had protein C deficiency, and 4 had protein S deficiency. Seven children had a high level of lipoprotein A, a thrombogenic atherogenic lipoprotein associated with osteonecrosis in adults, and 3 children had hypofibrinolysis (a reduced ability to lyse clots). Protein C or S deficiency, hypofibrinolysis, or a high level of lipoprotein A may result in thrombotic venous occlusion of the femur, which leads to venous hypertension and osteonecrosis of the femoral head. The authors concluded that early diagnosis of these clotting abnormalities might well open the door for pharmacological preventative therapy and potentially minimize the Legg-Perthes process. Conditions associated with an increased tendency for thrombosis are considered as thrombophilic or hypercoagulable states, whereas impairment of intravascular lysis of clots is called hyperfibrinolysis. Normally there is a balance between thrombosis and fibrinolysis. The hematologic screen for this particular observation involved measurement of fibrinolytic activity and levels of protein C, protein S, C4b binding protein, antithrombin III, lipids, and lipoprotein A, as well as a prothrombin time. In a separate study, Glueck et al. also demonstrated that heritable, high plasminogen activator-inhibitor activity with consequent hypofibrinolysis in adults is associated with idiopathic osteonecrosis. They have gone on to speculate that hyperfibrinolysis could cause inadequate lysis of venous thrombi in the head of the femur, obstruction of osseous venous drainage, and venous hypertension of bone leading to osteonecrosis. These observations currently await independent confirmation. 5. TRANSIENT SYNOVITIS IN PERTHES DISEASE

For many years, the feeling has persisted that occasional episodes of transient synovitis of the hip would be followed several months later by Legg-Perthes disease. There have been many reports on this possible correlation following initial mention of the matter by Ferguson and Howorth in 1934 (75). The cause of the Perthes disease was felt to be the effect of a synovial effusion and elevated intra-articular pressure on the vessels supplying the femoral head, which are intraarticular in the subsynovial region and thus vulnerable to compression by tamponade. Epidemiologic studies have not been definitive in this regard. As Kallio et al. point out, most patients with Perthes disease have never had a previous episode of hip pain attributable to transient synovitis, and most cases of transient synovitis make a permanent recovery even in the absence of any specific treatment (137). In addition, Legg-Perthes disease has not been noted after other conditions that cause major intra-articular hip joint synovitis, such as infective or juvenile rheumatoid disorders. In a detailed review of the literature involving 17 studies, they noted 6 reports

280

CHAPTER 4 ~

Legg--Calve--Perrhes Disease

that failed to document any incidence of Legg-Perthes disease after transient synovitis, 1 report in which there was a 1% incidence (only 1 of 101 patients), and other reports showing a variable incidence of 2.8% (1/32), 4.8% (1/21), 5.1% (6/117), 5.9% (6/102), 7.7% (1/13), 8.6% (9/105), 8.7% (2/23), and even as high as 13.2% (5/38) and 17.7% (11/62). In an effort to resolve this issue, Kallio et al. performed a prospective study in 119 children with transient synovitis or other causes of synovial effusion and elevated intra-articular pressure. During a followup of 1 year, no cases of LeggPerthes disease were diagnosed, and they concluded that the concept that Legg-Perthes disease can develop as a result of the period of elevated intra-articular pressure found in transient synovitis was not validated. The diagnosis of synovitis in their study was made by clinical history and examination, ultrasonography, plain radiographs with anteroposterior and frog lateral projections, and, in many instances, intra-articular pressure determinations. The patients were admitted to the hospital for treatment, which involved bed rest only. The group particularly stayed away from skin traction with the hip in extension because of their earlier demonstration of extremely high intra-articular pressure in this position. They felt that some instances of Perthes disease in previous studies might have been caused by treatment of transient synovitis using traction with the hip in an extended position. Previous studies had indicated that intra-articular pressure was highest in that position, with a mean of 130 mm Hg being measured, whereas intra-articular pressure with the hip in flexion had a mean of only 17 mm Hg in transient synovitis, which rarely exceeded the estimated arteriolar pressure (136). Aspiration was not felt to influence the result because none of the patients without synovial aspiration developed Perthes disease either. They also commented that from their previous study, when effusion had been aspirated, it returned to a preaspiration level within 1 day and then subsided over 5-10 days, much as occurred without aspiration. 6. VENOUS HYPERTENSION IN THE PATHOGENESIS OF L E G G - P E R T H E S DISEASE

An interesting phenomenon in relation to the etiology of Legg-Perthes has been the demonstration that arterial ischemia alone is not involved exclusively in causation, but rather the problem may be associated with a venous origin. Suramo et al. had observed that arterial occlusion as a primary event was not accepted by all as the cause of Legg-Perthes and that venous stasis could be the initiating feature leading to arterial occlusion (262). Such vessel hemodynamics had been implicated in adult osteoarthritis. They performed intraosseous venograms in 28 hips with Legg-Perthes and 20 normal hips, showing that in the normal hip contrast medium drained rapidly into local veins with none passing into the diaphysis, whereas in Legg-Perthes hips in the initial and early repair stages some contrast medium always flowed into the diaphysis, with flow into local veins reduced markedly. When the venous and intraosseous pressures rise, the capillary nutrition of the femoral head decreases.

Green and Griffin assessed intraosseous pressure in 23 Legg-Perthes and 23 normal femoral necks (95). When saline was injected intraosseously, the pressure was significantly higher in the diseased hips. Intraosseous venograms all were normal in control hips but mildly to severely abnormal in Legg-Perthes. The observation of venous abnormality of the proximal femur was substantiated. Liu and Ho studied 32 patients with Legg-Perthes disease using venography, measurement of intraosseous and intra-articular pressures, arthrography, and bone scan with methylene diphosphonate (175). The arterial flow to the affected head was decreased only minimally, but there was marked disturbance of the venous drainage in the diseased hip, elevated intraosseous pressure in the affected femoral neck, and increased intraarticular pressure in the involved compared with the normal side. An animal model then was created in which venous drainage was obstructed and intraosseous pressure of the femoral head and neck was elevated, following which avascular necrosis developed in 11 of 20 dogs. It has been shown by Arnoldi, Linderholm, and Mussichler that venous engorgement and intraosseous hypertension are present in osteoarthritis of the hip and by Arnoldi and Linderholm that obstruction of venous drainage can lead to increasing intraosseous pressure and consequent bone necrosis (4). Work by Kemp has shown that transient venous occlusion can produce coxa plana (149). In multiple experiments, he injected fluid into the hip joints of puppies, producing intra-articular tamponade sufficiently great to hinder venous return with subsequent demonstration of diminished vascular perfusion of the head by autoradiography (149, 150). The phenomenon appears well-illustrated, but the causal relationship between venous drainage and intraosseous pressure is not clear. 7. MULTIPLE EPISODES OF INFARCTION IN THE CAUSATION OF L E G G - P E R T H E S DISEASE

Although the blood supply of the femoral head now is well-studied (266, 267), no specific anatomic vascular anomaly underlying Legg-Perthes has been noted. The observation was made by Sanchis et al. that Perthes-like lesions of the proximal capital femoral epiphysis in young puppies did not occur after single episodes of avascularity, but were reproduced in two puppies whose femoral heads were devascularized on two occasions separated by several weeks (241). The initial devascularization process obtained by dislocating the femoral head and cauterizing the arterial supply led to some necrosis of the bone and cartilage, but this invariably was followed by repair. Once the second devascularization occurred, changes similar to those found in human Perthes with lytic lesions of the head were produced. Chondrocytes in the depth of the cartilage were dead and the area beneath what had once been hyaline articular cartilage showed histologically an area of poorly vascularized connective tissue containing occasional dead bone trabeculae. Histologic sections showed the original necrosis from the first devascularization, followed by repair, followed by damage subsequent to the second devascularization process.

SECTION IV ~ Early Pathologic Reports of Cell and Tissue Changes

When the concept of two or more episodes of infarction causing Perthes disease was utilized in examining histologic specimens, several observers then noted findings consistent with the theory. Their findings were not particularly those reported in the dog, but rather the finding of necrotic woven bone in the femoral head secondary ossification center, which by definition led to an interpretation of at least two episodes of devascularization. Because woven bone is indicative of a repair state, it would only be seen in children of the age at which Perthes occurs following damage to the existing lamellar bone and marrow. If the woven bone then appears necrotic, as evidenced by empty osteocyte lacunae, one would have to propose that a second and clearly separate episode of avascularity had occurred.

IV. E A R L Y P A T H O L O G I C R E P O R T S OF CELL AND TISSUE CHANGES IN LEGG-CALVE-PERTHES DISEASE Perthes made an initial description of the histopathology in 1913 from a small piece of a femoral head removed at the time of surgical intervention from a 9-year-old boy with a 2year involvement (213). The articular joint cartilage was normal. The cartilage within the epiphysis structurally was irregular. Cartilage islands within the bone were prominent, leading him to call the disorder an osteochondritis. The bone was described only as hard. Histopathologic studies in the first decade or so after the initial clinical and radiographic description of the disease were somewhat confusing; some represented Perthes disease usually in its late stages, whereas other of the cases today would not be recognized as the disorder. Understanding of the histologic findings evolved shortly, however, and it widely became recognized that the disorder was characterized by damage to the secondary ossification center of the femoral head, which led to massive subchondral bone and marrow necrosis and initially left the articular cartilage intact. It thus was recognized that normal articular cartilage combined with necrosis of the subchondral bone and marrow were the two histopathologic hallmarks of Perthes disease. These criteria enabled histologic differentiation between Perthes disease and a primary arthritis deformans occurring in the childhood years. Perthes changed his initial term for the disorder, arthritis deformans juvenilis, to osteochondritis deformans juvenilis based on the early recognition of the site of primary pathology.

A. Zemansky, 1928 Zemansky reviewed the extensive literature, most of which was published in German, in his 1928 paper and studied a single case in great detail (291). He reviewed the histopathology from what he considered to be the best documented cases from the literature, which were those described by Phemister (216), Axhausen (9-11), Heitzmann (113), Riedel

281

(231), Walter (281), Konjetzny (159), and Rockemer (235). The case he described represented a late-onset Perthes with the patient 16 years of age at presentation. The initial X ray demonstrated flattening of the head of the femur with extension of the capital femoral epiphysis posteriorly and laterally onto the neck. The medial part of the head was not involved. The neck was wider than normal and somewhat shortened. The treatment involved manipulation of the hip under anesthesia followed by application of a plaster of Paris hip spica. Symptoms persisted after treatment. The primary aim of therapy was described by the orthopedic surgeon, Dr. Royal Whitman, one of the major hip surgeons of his era, as follows" "the original intent in this case was as the head of the femur was flattening under pressure, to put it in an extreme abduction, that the point of pressure might be removed to the outer part of the head." Considerable deformity persisted, and eventually the head and neck of the femur were removed surgically and available for assessment. The changes described represented a well-advanced stage of the disorder. The macroscopic appearance showed the head of the femur to be markedly flattened and thrown into 3 main folds with deep crypts in between. Surface cartilage was clear, white and glistening and followed the irregularities of contour without apparent loss of substance.., in only one small area, the edge of the specimen laterally, was a thoroughly regular spherical contour observed. On the outermost circumference of the head the cartilage is somewhat lifted up from the underlying bone and hung in several shreds with pieces of the capsular ligament attached to it. The appearance of the freshly cut hemisection then was described. The articular cartilage varied in thickness from 0.5 to 5 mm. Changes were noted in the immediate subchondral area. Throughout the entire extent of the surface of the specimen the cartilage was separated from the apparently normal bone by reddish tissue of putty-like consistency. This band varied in depth and its deepest part was no more than one centimeter from the surface cartilage. In one area it occupied a cleft between bone and cartilage 5 millimeters deep and 2 millimeters long and communicated with the surface through the tear in the cartilage. Embedded in the reddish material were several small islands of whitish tissue having the consistency of the cartilage. These were loose.., a distinct epiphyseal line was nowhere to be seen. However, on tracing the surface cartilage laterally it was found to be continuous with one of the islands just mentioned, which suggested that the island represented the remains of the epiphyseal line. The bony tissue further down appeared grossly normal, its trabeculae being apparently of normal thickness. This description presented a clear-cut recognition of the histologic appearance of the subchondral radiolucency in Perthes. Microscopically the articular cartilage was normal in appearance in the individual areas, but it varied in thickness. Variations of cartilage thickness were due to encroachment

282

CHAPTER 4 ~ Le99-Calve-Perthes Disease

of the process from below. Cartilage islands were seen within the bone substance. Even at this late stage "the process is not the same in all areas." The overlying normal cartilage covered a "completely necrotic spongiosa." This was described as a striking change. In the deeper levels one encounters necrotic and fragmented bone lamellae lying in the matrix of degenerated marrow and blood. Still deeper one sees the fragment of necrotic bone surrounded by a loose granulation tissue containing many blood vessels and multinucleated giant cells (osteoclast) . . . . The granulation tissue forms a more or less definite wall around the necrotic areas and becomes continuous with a zone of dense fibrosis of the marrow at their junction with the normal bone lamellae below. This fibrous marrow gradually passes over into a poorly cellular fatty marrow. The process was not uniform, with adjacent areas showing replacement by a more or less dense fibrous tissue rich in capillaries but poor in cellular elements. "Besides the cartilage formation directly from connective tissue, there is abundant osteoid tissue laid down on the basis of pre-existing bone lamellae." This osteoid tissue was absent where the necrosis was most marked and was present only where the necrotic tissue had been resorbed and replaced by fibrous tissue. By far the greatest amount of newly formed bone was of this "appositional" variety. Throughout the involved area the picture of empty lacunae and resorption was seen frequently, and it was not uncommon to see resorption of bone by osteoclasts on one side of the bone lamella and osteoblastic activity on the other. No obliterative thickening of any blood vessels was seen. The summary presented six primary pathological findings: (1) extensive subchondral necrosis of bone and marrow; (2) complete destruction of the epiphyseal line; (3) fragments of dead bone, surrounded by richly vascular granulation tissue containing many multinucleated giant cells; (4) fibrous tissue replacement of necrotic areas; (5) osteoid tissue formation from fibrous tissue and preexisting bone lamellae; and (6) dilated blood vessels in the undersurface of the cartilage. Previous descriptions of the histopathology noted by Zemansky as well as some other descriptions from the same era are reviewed next.

B. Schwarz, 1914 Schwarz provided an early detailed review of 22 patients with Legg-Perthes disease from Perthes' surgical clinic (244). Each case description was accompanied by a camera lucida drawing of the radiograph. In these cases we can recognize the entire spectrum of the disorder from the earliest changes to the late long-term deformities. Many case presentations featured two or three radiographs from different times showing the evolution of the disorder. The appearance of lateral subluxation of the femur, flattening of the secondary ossification center, widening of the secondary ossification

center, fragmentation of the secondary ossification center, and the longer term appearance of the entire flattened femoral head are shown clearly. One histologic specimen was presented. This allowed for early correlation between radiographic and histologic findings. Areas of rarefaction on the X ray corresponded to cartilage tissue that was interspersed between the bone trabeculae of the secondary center. There was death of some lamellae and in particular an absence of normal cellular red marrow with its replacement by a fatty marrow. Schwarz felt the process was an active one in the sense that it occurred in a previously normally developing epiphysis. The occurrences were evidence that the bony epiphysis, "because of impaired nutrition, dies focally and with time to a progressively larger extent. The destroyed bone is then replaced by cartilaginous tissue which is much less susceptible to nutritional disturbances." With time the cartilage again was transformed into bone, and eventually the whole epiphysis consisted of healthy bone. The flattening of the head was considered to be secondary to mechanical causes "brought about by pressure of body weight; a type of marginal bulge formation is produced by the overhang of the lateral portions of the spread out epiphysis." The nutritional disturbance was felt to be secondary to a disturbance of the normal blood supply. The ultimate basis of the disorder was felt to be trauma in which there was a "loosening" of the epiphysis without displacement and subsequent vascular irregularities due to that. Operative inspection of the hip of a 7-year-old boy showed the surface of the head of the femur to be flattened and indented and to extend widely over the neck. Histologically the joint cartilage was normal. In the subchondral region there were islands of bone and cartilage surrounded by fibrous tissue and fatty marrow.

C. Phemister, 1920 Gross inspection at the time of surgery in a 10-year-old boy showed the surface of the head of the femur to be markedly distorted and flattened but with the articular cartilage unimpaired (216). The subchondral region was curetted and was noted to contain necrotic debris and several sequestra. There was subchondral necrosis of bone and marrow in a richly cellular granulation tissue, with giant cells (osteoclasts) surrounding dead bone fragments. Phemister is widely considered to be the first observer to define clearly bone necrosis as the primary histopathologic finding. All subsequent studies noted the prominent bone and marrow necrosis.

D. Axhausen, 1923 Autopsy study of the head of the femur of a 9-year old child showed that the spherical shape of the head was preserved, with the only surface irregularity seen being a small depression at its highest point (9-11). The articular cartilage was intact. The diaphysis was normal as was the epiphyseal line.

SECTION IV ~ Early Pathologic Reports of Cell and Tissue Changes

283

became necrotic. He felt that necrotic bone in and of itself was an important causative factor for callus formation.

E. Heitzmann, 1923

F I G U R E 2 Photomicrograph, reproduced from the work of Axhausen (10), shows a specimen from the femoral head of a child with Perthes disease. Note the intact although ruffled articular cartilage, the deeper staining region of necrotic subchondral bone, the subchondral fracture with a space left by the collapse of adjacent trabeculi, and normal bone deeper in the section.

A 9-year-old boy had a resection of the head of the femur. The surface of the head was practically flat, having a slight convexity only at its outer part (113). The epiphyseal growth plate cartilage lay only 6 mm from the surface. Between the two cartilage surfaces was a broad area of necrotic bone and marrow enclosed in granulation tissue. Many osteoclasts were seen resorbing the necrotic bone. One section of the physis was interrupted by granulation tissue passing into the adjacent metaphyseal bone. Resection also was performed in a 13-year-old girl. The head was remarkably flattened with wrinkling and furrowing of the otherwise normal cartilage surface. The epiphyseal line had disappeared almost entirely. The joint cartilage was well-preserved, but in a cleft underlying that cartilage were sections of necrotic bone and marrow. Fibrous tissue penetrated this area just below the cartilage and contained a fine network of newly synthesized bone.

F. Riedel, 1923 The subchondral bone and marrow were necrotic and surrounded by an outgrowth of young connective tissue interposed between the surface and epiphyseal cartilage. Early new bone formation was seen with synthesis on the old lamellae. Axhausen was the first to clarify in great structural detail the concept of aseptic bone necrosis, which is now referred to as either avascular necrosis or osteonecrosis of unknown cause. In landmark studies, he produced necrosis by electrocautery and followed the subsequent repair phenomena. A color photomicrograph drawing from one of his earlier works clearly shows the necrotic bone region in the subchondral area and the overlying normal cartilage. The bone lacunae in the dead segments are empty of cells, whereas the new repair bone synthesized on the necrotic cores clearly is living with both osteocytes in the newly formed lacunae and osteoblasts lying on the bone surface. Many of his clinical studies involved arthritis secondary to subchondral bone necrosis, such as was seen in osteochondritis dissecans of the knee and other sites and Perthes disease as well. In a paper published in 1924, he clearly illustrated a histopathologic section from a patient with Perthes showing viable overlying articular cartilage and the necrotic bone underneath. The cartilage surface is ruffled but intact. Underneath, one can see the necrotic subchondral bone involving both dead trabeculae and debris in the marrow. There is a fracture through the subchondral bone. The bone deeper in the secondary center is viable (Fig. 2). In his work, he points out that aseptic bone necrosis frequently was seen in many types of bone disorder and, thus, was central to an understanding of histopathology. He also pointed out that, even in fractures, the ends of the fracture fragments

The head of the femur was resected in a 10-year-old boy (231). The surface of the head was pressed flat and strongly indented. In the region of attachment of the ligamentum teres was a concave depression. On histologic section there was a small free bone fragment beneath the depression, which itself was surrounded with extensive necrotic bone and marrow. The epiphyseal line was nearly completely destroyed. Osteoblastic activity and lacunar resorption were occurring simultaneously, often on the same bony lamellae. Around the necrotic bone fragments was a highly cellular granulation tissue with many giant cells. Farther away the marrow was fibrotic, and in some areas cellular osteoid was seen. In a 9year-old boy the surface cartilage of the head of the femur showed a crescent-shaped depression, beneath which the bone felt very soft. The subchondral region was softened.

G. Walter, 1925 Resection of the femoral head in a 17-year-old showed the head to be flattened and mushroom-shaped, with its surface markedly distorted and denuded of much of its cartilage (281). This patient also had polyarticular rheumatism.

H. Konjetzny, 1926, 1934 The head of the femur was resected in a 17-year-old boy (159). The surface still was distinctly spherical with the cartilage intact throughout. There was an extensive infarct of subchondral necrosis to a depth of 1.5 cm. Microscopically, complete bone and marrow necrosis was observed, on the border of which was granulation tissue with many osteoclasts.

284

CHAPTER 4 ~ Legg-Calve-Perthes Disease

A second case was obtained at autopsy of a child 8 years after onset of Perthes. The head of the femur was markedly deformed and along with the neck and greater trochanter formed one large cylindrical mass. The surface cartilage was, for the greater part, normal. The secondary center was diminished markedly in size and the epiphyseal line was nearly completely destroyed. The articular cartilage adjacent to the ligamentum teres was 2 - 4 times the normal thickness. The subchondral bone trabeculae were markedly thickened and contained necrotic fragments composed of dead bone, but at the periphery the bone and marrow were normal. The diameter of the acetabulum was one-third larger than normal.

I. Delchef, 1926 Delchef reported on an autopsy study of a hip in a 9-yearold girl with Legg-Perthes noted 1 year previously (60). The secondary ossification center was flattened and irregular, but the overlying cartilage model of the head remained intact, although it followed the undulations of the flattened bone center. At one point, the cartilage remained stretched over an underlying cavity in the epiphysis. Transverse sectioning of the head showed a small linear cystic cavity visible between the articular cartilage and the bone of the head. The growth plate was horizontal. The acetabulum retained its normal shape.

J. Rockemer, 1927 Resection of the head of the femur in a 13-year-old boy showed a normal spherical shape of the head with intact cartilage (235). The cartilage was of uniform thickness but separated from the underlying bone for about two-thirds of its extent. In the subchondral region was soft reddish tissue, and in one area was a small bone sequestrum. The epiphyseal line was intact. Surface cartilage histologically was normal, but the subchondral bone was completely necrotic with spicules surrounded by blood and granulation tissue. The wellvascularized granulation tissue had many giant cells. The bone lamellae were undergoing resorption with some osteoblastic activity seen.

K. Lippmann, 1929 Lippmann studied a femoral head obtained from a resection for operative correction for advanced Perthes in a 12-yearold girl (174). At operation the head was flattened and the articular cartilage extended outward to the neighborhood of the trochanter. The acetabulum was normal. The surface cartilage was described macroscopically as being clear and white but marked with superficial indentation, "as if in accommodation to a shrinking of the underlying substance." On hemisection the surface cartilage was, with minor variations, of uniform thickness and normal appearance, and the epiphyseal line appeared normal and unbroken except in one small region. Microscopically the articular and epiphyseal

cartilages also were structurally normal. The greatest part of the bone of the secondary ossification center, however, was a necrotic mass. The bone lacunae were empty and the marrow was homogeneous in nature, with destruction of all cellular appearances. In the region of the secondary ossification center just above the growth plate, some small areas of normal bone and marrow persisted. The small regions of physeal cartilage that were interrupted on macroscopic exam showed mesenchymal tissue interposition on histologic exam. Four sections of tissue were cut from different regions. In one both bone and cartilage were normal, but in the other three the findings described earlier were noted. One of the line drawing illustrations showed a fracture through the region of necrotic subchondral bone lying just below and conforming to the shape of the articular cartilage surface. The massive subchondral bone and marrow necrosis involved approximately half of the femoral head. At its periphery, avascular and mesenchymal tissue reaction was seen. Lippmann clearly indicated that "the gross femoral head deformity is secondary to collapse of the underlying necrotic bone." He also concluded, similar to Zemansky, that vascular occlusion was the most likely cause of the disorder. The pathologic hallmark of the disorder was the subchondral bone and marrow necrosis along with granulation tissue and an initially intact articular cartilage. The physeal cartilage was intact in some but interrupted in others.

L. Nagassaka, 1930 Nagassaka reported a histological study of eight cases of Legg-Perthes disease (201,202). He confirmed the extensive necrosis of bone trabeculae and marrow in the secondary ossification center of the femoral head and also recognized the ingrowth of vascular granulation tissue. He interpreted this "intermingled state" as representing evidence of both necrosis and beginning repair. He felt that, in some instances, the articular cartilage was not normal and on occasion appeared to be degenerated and also necrotic. In relation to the necrotic subchondral bone and marrow, he felt that the area invariably was surrounded, absorbed, and even replaced by newly formed young connective tissue. There often was a crushing of the trabeculae of the dead bone, which he attributed to compressive forces. Newly formed osteoid was seen as part of the repair response, as was new cartilage on occasion. The epiphyseal line generally was normal but on occasion also was destroyed. Cartilage islands characteristically were seen within the region of the secondary ossification center, due primarily, Nagassaka felt, to callus from the ingrowth of the new tissue. He noted that it was incorrect to interpret the characteristic changes on the radiograph as showing progression of the disease, but rather was impressed by the fact that histologic correlation showed evidence of "a steady recovering process of the necrotic femoral head." He was one of the first who recognized that, even at the early stage of the disease when extensive necrosis of bone and marrow was present, there was also onset of the recovery

SECTION V ~ Subsequent Pathologic Reports with Better Defined Correlations

process (202). Although the radiographic changes appeared to continue in spite of treatment, they themselves "do not necessarily mean the progress of the disease, but we rather have to recognize it as the steady recovering one in the necrotic area."

M. Summary of Histopathologic Changes after Two Decades of Study (Zemansky) A review of Zemansky's own case plus those of others began to show a recognizable spectrum of change in the histopathology (291). All cases showed some deformity of the surface contour of the head of the femur, which varied from a slight central depression to marked flattening and furrowing. In all cases, however, the surface articular cartilage was intact. The condition of the epiphyseal line varied considerably from being well-preserved in those cases of relatively short duration, to interrupted in those of longer duration, to showing nearly complete destruction in some others. The fact, however, that it initially was intact along with the articular cartilage indicated that the disorder was not primary in either articular or epiphyseal cartilages. Zemansky clearly noted that "the most outstanding finding in all of the cases was the extensive subchondral necrosis of bone and marrow and the presence of a richly vascular granulation tissue. That the granulation tissue contained many multinucleated giant cells in a large majority of instances was not to be wondered at as they are to be expected where dead bone is present." He concluded that inflammatory cells did not represent a pathologic feature of the disorder because many of those cells normally were present in the marrow anyway and identification of them need not indicate a pathologic response. Fibrosis of the bone marrow and lacunar resorption generally were present but also were expected in any process in which there was extensive bone destruction. In cases of moderate duration, osteoid tissue was noted growing in relation to the granulation tissue, and in the case noted after 8 years of the disorder, solid new bone was observed. These findings represented spontaneous healing. Zemansky recognized, however, that the healing was slow and generally indicated "how little can be expected from this method of healing." This statement may reflect the fact that the specimens available for assessment were in the most severe group and often in those appearing at relatively late ages in which the collapse was greatest and the repair least. Zemansky reviewed the possible etiologic factors listed involving tuberculosis, syphilis, tickets, constitutional factors, arthritis deformans, infection, trauma, and vascular obliteration. Discussion of a possible trauma etiology is particularly interesting, and he noted correctly that the fragmented bone really could not be accounted for as an immediate sequel of crushing injury because the fragmentation actually develops slowly over several months' time. In assessing embolic vascular occlusion, Zemansky indicated the very early impression that vascular changes were the cause of the problem. He indicated that "all agree that the whole pathologic picture of

285

Perthes disease closely resembles an infarct . . . . the differences of opinion arise in the question of how the blood supply to the region is cut off." Although he could not ascribe the disorder to an embolization that simply picked out the small vessels of the femoral head, he remained convinced that "a disturbance of nutrition through some sort of occlusion of the vascular supply other than embolic would best account for the pathological picture seen in Perthes disease." Trauma to the vascular supply of the proximal end of the femur represents one of the first detailed theories of etiology and was offered by Schwarz, who worked in Perthes' clinic. Zemansky's observation has a familiar ring to it in which he indicated that "no one seems to have paid much attention to Schwarz's statements probably because Perthes himself did not believe in them." Zemansky himself, however, also felt that these viewpoints represented the closest approach obtained thus far in the solution of the problem. He concluded that "the theory of vascular disturbance of the femoral epiphysis from trauma is considered the most likely one to explain the pathogenesis of the disease." Zemansky summarized his findings on pathogenesis by indicating that "during childhood the blood supply of the epiphysis of the head of the femur by way of the periosteum and ligamentum teres is just adequate to maintain its nutrition and in certain children perhaps differs in its anatomic arrangement. Following trauma, even though unnoticed, one or all of these blood vessels may be sufficiently damaged to interfere with the nutrition of the epiphysis and necrosis of the subchondral bone and marrow takes place." Zemansky then clearly articulated the impression that the necrotic bone is readily fractured by weight bearing, with resultant deformation of the surface contour. Repair is slow by means of fibrous tissue replacing the bony fragments. A small amount of osteoid tissue is formed from the fibrous tissue and by the surrounding live bone lamellae but the blood supply of the region is too inadequate to bring about solid bony replacement. Eventually an equilibrium is reached where the blood supply from the diaphysis is able to take care of whatever is left of the epiphysis. He felt in summary that this often took several years. "Likewise as the joint accommodates itself to the presence of the deformed femur head the limp and limitation of motion disappears. It is at this point that secondary arthritis deformans may well appear. The mechanical factors are certainly conducive to its development."

V. S U B S E Q U E N T P A T H O L O G I C R E P O R T S WITH BETTER DEFINED CLINICAL AND RADIOGRAPHIC CORRELATIONS

A. Ferguson and Howorth, 1934 Ferguson and Howorth surgically explored 21 hips with Legg-Perthes disease in the course of managing 83 hips with the disorder (75). Many were felt to have the early

286

CHAPTER 4 ~

Legg--Calve--Perthes Disease

stages of a subacute arthritis. In patients with more advanced disease, some had drilling of the femoral head bone in an effort to hasten repair. Six of the hips had arthrotomy early in the active stage of the disorder, at which time the synovial membrane was always thickened, soft, fragile, very vascular, and often irregular with villus formation. The periosteum was usually thickened and edematous. The capsule was usually thickened, slightly edematous, and more vascular than normal. The contour and appearance of the visible portion of the cartilage of the femoral head were normal. The synovial fluid was not abnormal. On microscopic examination, the synovial membrane was usually edematous, contained clusters of lymphocytes, and was often villous; the capsule and periosteum were chronically inflamed in most cases. Ferguson and Howorth clearly pointed out that the synovitis was present in the absence of any deformity to gross examination of the femoral head, which led them to suggest that the synovitis was a primary or at least early finding in the disorder and not one totally secondary to deformity of the femoral head. Two hips were explored early in the stage of repair and the other 13 hips were explored later in the repair stage. The hips exposed in the reparative stage showed a synovial membrane that was no longer inflamed but rather smooth, inelastic, tough, thin, and avascular. The periosteum and capsule also were scarred and inelastic. In some instances, the cartilage of the femoral head was flattened, and in others, it had proliferated at the margin with the development of a pannus. The cartilage otherwise was normal in appearance in all hips. On microscopic examination, the periarticular soft tissues were scarred extensively and contained thickwalled vessels of small lumen. When the secondary ossification center of the femoral head was drilled, the authors reported that in relatively advanced repair "in several cases very dense bone was encountered and in other cases certain portions were soft." The bone of the femoral neck usually was normal in consistency. In cases that appeared radiographically to have a cap or mushroom deformity, such distinctions were not possible on inspection of the femoral head at surgery. This would support the concept that the shape of the cartilage surface was relatively normal even though the bone of the secondary center was abnormal in shape. Ferguson and Howorth then correlated the radiographic features with some of the observations made at surgical arthrotomy. They again noted that "prior to the development of any other roentgenograph changes, distention of the capsule of the hip may be demonstrable." In the active stage of the disorder, they commented on the differing areas of density within the head, which they interpreted as evidence of disease activity (necrosis) going on simultaneous with evidence of repair. They also noted that "new dense areas may appear after the epiphysis is well into the stage of repair," which they interpreted as a relapse. This finding also is consistent with the more recently documented evidence of mul-

tiple episodes of infarction rather than a single causative episode. Ferguson and Howorth clearly noted that the repair phase with revascularization was underway when areas of radiolucency appeared. They note that "irregular ossification, which appears the more alarming, marks the presence of repair; also that as 2 or more dense areas may not lose their density at the same time, activity (by which is meant the occurrence of necrosis) and repair may be present concurrently in different parts of the same femoral head." They again pointed out that often there was no flattening of the cartilage of the head to correspond to the flattened, ossified portion shown in the radiograph. "At operation, it was constantly found that the limitation of motion was due to inelasticity of the synovial membrane and capsule rather than to bony or cartilaginous block." In the residual stage in cases with marked flattening and widening of the head, abduction and rotation then were limited on a mechanical basis. Even in patients who healed with good shaping of the femoral head, sometimes there was a distinct limitation of motion due to soft tissue scarfing. They felt, therefore, that the impairment of function was "usually and largely due to the effects of inflammation of the soft parts." Thus, they felt that the treatment of the synovitis by rest early in the phase of the disease was of great importance in maintaining long-term function.

B. Gall and Bennett, 1942 Gall and Bennett studied a patient with Perthes disease who died at 13 years of age due to severe renal disease (renal osteitis fibrosa cystica), which led to severe chronic renal insufficiency (85). Because the latter was associated with hyper-parathyroidism and marked osteoporosis, the case does not represent a study of Perthes alone. The fact that it is bilateral with more or less the same degree of involvement also raises questions as to the diagnosis. The long bones showed increased radiolucency and there was delayed epiphyseal development at the lower end of the radius and ulna. Medications were not listed. The findings resembled previous descriptions of Perthes, but bone repair was not a prominent feature undoubtedly due to the other systemic features of the patient. Macroscopic exam showed the femoral head to be flattened markedly along its superior surface but covered by smooth, normal appearing articular cartilage. The acetabulum was normal. The neck was markedly shortened and widened. The epiphyseal growth plate cartilage was narrow, irregular, and destroyed in places. The most prominent changes in the secondary ossification center were in the lateral two-thirds. The articular cartilage of the femoral head was structurally normal. On the undersurface of the articular cartilage there were occasional fragments of necrotic subchondral bone, whereas in other areas the cartilage tissue immediately was adjacent to underlying granulation tissue. A wide band of highly vascularized connective tissue penetrated deeply into the necrotic bone mass through disrupted

SECTION V ~ Subsequent Pathologic Reports with Better Defined Correlations segments of the epiphyseal cartilage. Marrow underneath the subchondral bone segment was necrotic and contained a flattened cystlike crevice consistent with a subchondral fracture.

C. Haythorn, 1949 Haythorn reported a unique experience assessing specimens of the proximal femoral capital epiphysis obtained by curettage in 33 cases considered clinically, pathologically, and radiographically to be Legg-Perthes disease (111). The tissue was obtained as part of a treatment approach originally performed by P. B. Steele of Pittsburgh, in which the necrotic bone and marrow tissue of the femoral head epiphysis was removed and replaced with bone graft from the femoral neck in the hope that the repair response would be hastened and thus allow for better long-term results. Haythorn indicated that the pieces removed were large enough to determine the character of the pathological lesion and the relative degree of change, but tissue orientation could not be determined because the bone was curetted. The study allowed for correlation of the radiographic appearance and the gross macroscopic appearance of the femoral head, as well as the histologic appearance. It was specifically noted that the cartilage model of the head was more regular and intact than the radiographic appearance of the bone of the secondary ossification center, which often was fragmented, flattened, and crushed. The report indicates that "the head was not flattened as it appeared in the roentgenogram but remained spheroidal; and the cartilage appeared normal except for thickening. The gross appearance of the acetabulum was normal." When the joint was opened, the cartilage distortion and flattening were much less evident than had been expected from the plain radiographs. The majority of the heads in this instance were described as spheroidal, smooth, and glistening, although some reports indicated the cartilage to be thin. Some showed slight flattening with cartilage of uneven thickness, and on occasion focal degeneration of the cartilage was described. The bone tissue obtained by curettage was described as "soft, necrotic, containing bony spicules and bits of cartilage." One additional value of this study was that tissue was obtained at an earlier period in disease evolution than in most studies reported up to that time. The microscopic changes in the tissue removed by curettage followed a similar pattern. The histopathologic appearance was described in relation to seven changes. (1) Degenerative changes. Degenerative or necrotic changes were present in each instance. At the mildest end of the spectrum there was necrosis of the marrow with normal marrow cells and fatty stroma absent. On occasion there was evidence of fibrous tissue repair of the marrow. In more advanced lesions, necrosis was extensive and complete involving the marrow and cartilage and bony spicules from broken trabeculae. (2) Crushing in necrotic areas. On occasion there was crushing or jamming together internally of the degenerated elements, but the entire femoral head did not appear crushed.

287

The surface and periphery of the femoral head were more resilient in situ than expected and permitted the transmission of pressure to its contents. Bits of cartilage, bone, and granulation tissue were compressed together and the intertrabecular spaces were filled with homogeneous debris. (3) Reparative processes. Healing commonly was found, which varied from the simple replacement of marrow to more elaborate repair. Islets of healing were found immediately adjacent to degenerated areas. Areas of osteogenesis were seen with cellular woven bone rimmed by surface osteoblasts. (4) Giant cell reactions and cystic reactions. Multinuclear foreign body giant cells were seen, as were multinucleated cells more typical of osteoclasts. (5) Changes in cartilage. There was loss of polarity in the cartilage cells. Many pieces of cartilage were penetrated by arteries and fibrous granulation tissue. (6) Vascular changes. As a rule there were no vessels in the degenerated areas. Capillaries and arteries were present, however, in marrow spaces particularly in relation to newly synthesized fibrous tissue. (7) Resemblance to scurvy and tickets. This section is quite imprecise but did indicate that both osteoblasts and osteoclasts were seen in ossifying areas. The author concluded that the changes were constant within fairly narrow limits and included aseptic (avascular) necrosis of marrow, cartilage, and bone, crushing of trabecular fragments, congruent degeneration and repair, partial ossification of displaced cartilage tissue, and various regions of cystic degeneration. It was felt that the removal of the necrotic contents in the head should be beneficial because "since it eliminates debris, it may lead to the formation of cysts and giant cell reactions." No long-term results of these interventions were published.

D. Jonsater, 1953 Jonsater performed a major study of Legg-Perthes disease, concentrating on core biopsies of involved femoral heads and correlations with the plain radiographic appearance of the bone and with hip arthrograms, which outlined the shape of the cartilage model of the femoral head (133). The study involved 44 biopsies in 34 patients of whom 26 were boys and 8 girls. The ages at the time of the examination ranged from 3 years 2 months to 11 years 3 months. The changes were unilateral in 27 and bilateral in 7. All other types of hip disorder were excluded rigorously. The staging system of Waldenstrom was used for correlation. 1. INITIAL STAGE a. Bony Changes The histologic picture in this stage was dominated by marked necrosis of bone. In some instances the trabecular structure was well-maintained, but in others the trabeculae were broken and split into small fragments. The marrow spaces, when preserved, were filled with a mass of necrotic tissue. The trabeculae frequently gave the appearance of being pressed into one another and into smaller fragments under the influence of loading. There were no

288

CHAPTER 4 ~ Legy-Calve-Perrhes Disease

signs of bone regeneration during this stage. No giant cells were seen, and there were no signs of inflammation. b. Cartilage Changes Cartilage from the periphery of the secondary ossification center generally was included in the biopsy samples. As a rule, the bone necrosis did not reach farther than to the inner margin of the zone of ossification, but in some cases it was noted that the endochondrally formed bone had also become necrotic. Jonsater interpreted this to indicate that necrosis began in the inner part of the secondary center and gradually involved the periphery. In some of the cases there was evidence of change in the cartilage involving the innermost layers, which appeared somewhat devitalized. These cartilage changes were immediately adjacent to areas in which bone necrosis reached so far out that it also embraced the most newly forming endochondral bone. He concludes, correctly in our viewpoint, that the necrosis occurring in the bone and the changes in the cartilage have the same cause. Jonsater concludes that "necrosis of the bone is the dominating histopathological change in the initial stage." Although he had no definitive indication of the cause of the disorder, he felt that the bone necrosis was ischemic in origin. The early increased bone density was due to the trabeculae being compressed into a smaller space. Waldenstrom had noted (1934) the subchondral thinning early on. This was in the anterior and superior part of the epiphysis and was described as a thinner streak shape formation separated from the joint cartilage by only a thin layer of bone. He felt that the separated space was caused by resorption. Jonsater's studies showed no signs of resorption at this stage. He felt that the bone necrosis left the secondary ossification center intact for a long time, whereas new endochondral bone formation occurred on the undersurface of the cartilage. When subjected "to loading the necrotic bone is compressed with the cartilage which in this age is thick and elastic. The attached ossification zone springs back into position when the pressure is removed whereby a gap arises which in all probability is filled up with tissue fluid." This represents an early description of what is referred to currently as a subchondral fracture represented by the crescent sign on plain radiography. 2. FRAGMENTATION STAGE

a. Bony Changes Bone necrosis continued to appear prominently, but the trabeculae did not show the uniformly dead appearance as in the early stages. Reconstitution of the normal bone marrow was more prominent at this stage. Characteristic of the microscopic picture in the fragmentation stage is the profuse occurrence of the morphological elements which indicated a regeneration of the decomposed bone. One can see.., a tissue that is profuse in cells and resembling connective tissue which in more or less broad formation grows into the necrotic bone. This tissue which is profusely vascularized in most cases is probably not connective tissue in a real meaning of the word but rather a tissue the cells . . . are determined to become bone . . . . Early areas of osteoid formation are seen.

Giant cells also were present and often arranged like osteoclasts in depressions into necrotic trabeculae. New bone formation often progressed to a stage in which osteoblasts lined up on the surface of the bone. b. Cartilage Changes Cartilage changes from the undersurface of the articular cartilage were localized to the basal regions and only occurred where the bone necrosis reached right out to the endochondral ossification zone. The fragmentation stage was characterized by bone necrosis and also by changes of a reparative nature rich in new blood vessels, giant cells, and osteoclasts. The fragmentation stage thus represents one in which active repair of previously necrotic bone is occurring. Jonsater found relatively few cartilage islands, which had been a characteristic finding in early reports. He felt that the radiolucent regions in the femoral head in the fragmentation stage represented areas in which connective tissue was prominent and in which differentiation was only in the early phase of osteoid formation.

3. REPARATION STAGE

a. Bony Changes Bone necrosis was still seen but was markedly diminished from the previous stages, whereas normal cellular bone had increased in relation to that seen in the fragmentation stage. The bone marrow also reconstituted to a great extent and much of the new bone was lamellar. b. Cartilage Changes The changes seen here also were limited to the basal layer of the articular cartilage. There often was evidence for lowered vitality of the cartilage, and within some regions the degenerative changes gave the appearance of an embryonal cartilage. In a histologic sense, the reparative phase showed a greater tendency toward repair than the fragmentation phase. No tissue was available in Jonsater's study from the definitive or final repair stage, but all evidence pointed to the complete reconstitution of the cancellous bone with the major problem relating to the shape of the femoral head and in particular its articular surface.

4. CONCLUSION Jonsater was able to make a series of relative correlations. The radiographic initial stage of Waldenstrom corresponded histologically to a high degree of necrosis in the secondary ossification center of both bone and bone marrow. The subchondral thinning, by which Jonsater appears to refer to the crescent sign, arises because the necrotic bone is compressed by loading or muscular tension, after which the elastic joint cartilage with a thin layer of attached endochondral bone springs back and leaves a gap in the subchondral region. This offers a partial explanation of both the crescent sign and the increased density of the secondary ossification center. All bone of the secondary ossification center is formed by the endochondral mechanism, not just that on the undersurface of the articular cartilage. Cartilage changes of the articular region of a degenerative nature occur but are localized to the basal layer of the joint cartilage and only show themselves in places in which the bone necrosis reaches as far as the

SECTION V ~ Subsequent Pathologic Reports with Better Defined Correlations cartilage-bone interface, indicating that the cause of the bone necrosis and of the cartilage change is the same. There is necrosis both of the secondary ossification center and of the epiphyseal cartilage surrounding it, including the basal layer of the articular cartilage, often referred to as the miniplate, and also part of the epiphyseal growth plate. The fragmentation stage is a repair stage in which histologic study shows both necrotic persisting bone and newly synthesized repair bone.

E. Ponseti, 1956 Ponseti obtained biopsy specimens from two children with clinical and radiographic findings typical of the early stage of Legg-Perthes disease (219). The punch biopsies were 1 cm in diameter through the central portion of the neck and head. In case 1 (7-year-old girl), the bone and marrow were necrotic. The line of demarcation between the femoral head and the epiphyseal plate was uneven. Some cartilage cells were seen between the bone trabeculae of the femoral head. The joint cartilage appeared normal except for the deeper layers adjacent to the subchondral bone. The epiphyseal growth plate itself was abnormal structurally, and it was observed that "endochondral ossification did not take place; this left tongues of cartilage in the metaphysis." No necrotic bone was seen, however, in the metaphysis or in the femoral neck. In case 2 (11-year-old boy), bone and marrow from the secondary ossification center were necrotic except for a small area adjacent to the growth plate. In some areas bone repair was seen. The line of demarcation between the femoral head secondary ossification center and the epiphyseal growth plate was very uneven. The joint cartilage appeared to be normal. The epiphyseal growth plate showed areas of disorganization, with the cartilage cells clustered and the cartilage matrix unevenly stained, fibrillated, and crossed by extensive clefts filled with blood such that the endochondral ossification was deranged. Ponseti focused upon the necrosis of the femoral head bone and the associated disruption of the epiphyseal plate. He indicated, however, that faulty epiphyseal plate cartilage was a primary lesion. Most now feel that because the blood supply of the secondary ossification center and of the epiphyseal growth plate come from the same source both can be affected in Legg-Perthes. Changes to the bone of the secondary ossification center more readily are seen radiographically, however, and most descriptions of the disorder concentrate on them. Irregularities of proximal femoral epiphyseal growth plate, however, have been commonly described by Haythorn, Perthes, Riedel, and Zemansky, among others.

F. Mizuno, Hirrayama, Kotani, and Simazu, 1966 These authors reported on histologic studies of cases of Legg-Perthes disease obtained by needle biopsy or thin seg-

289

mental resections (194). The age distribution was between 3 and 12 years with a mean of 6 years. They pointed out that the fragmentation stage terminology of Waldenstrom is based on a radiographic appearance, which is associated histologically with areas of necrotic persisting bone interspersed with areas in which necrotic bone has been resorbed and replaced by newly synthesized osteoid. They used the terminology of initial stage, intermediate stage, and late stage to describe the changes with time. 1. INITIAL STAGE Tissue was obtained by a needle biopsy. Findings are similar to those of Jonsater with a characteristic feature being extensive avascular necrosis of the bony trabeculae and the bone marrow. There is no trace of tissue reaction at this phase. The sphericity of the articular surface is maintained as is its gross appearance. There are some extensive areas of degeneration in the articular cartilage. Changes are noted in the chondrocytes, in the matrix with fibrillar formation, and overall, mechanically, involving bending and folding of the articular surface and occasional transverse slits or fissures. 2. INTERMEDIATE STAGE In many instances the articular surface loses its glossy whitish appearance. The sphericity is reported as lost in most cases; on occasion craterlike depressions are seen and in rare instances even ulceration, allowing the underlying bone to become visible. There is substitution of necrotic bone and cartilage by newly formed tissues, vascularized fibrous tissue ingrowth, atypical endochondral ossification, and an impression that the changes are complicated by repeated microtrauma. The histologic sections thus showed a mixture of modes of new bone formation. This included deposition of new bone on preexisting cores of necrotic trabecular bone; bone formation by the endochondral mechanism, sometimes with repair cartilage callus transformation and sometimes from continuation of normal endochondral bone formation; and bone formation from fibro-osseous or fibrocartilaginous loci. The source of the repair tissue was felt to be 4-fold: from the ligamentum teres, retinacular tissue of the neck, transphyseal vascular perforation, and some repair from the remaining unaffected bone marrow. Reference also is made to the metaphyseal cysts. These are primarily due to a tonguelike process of epiphyseal plate cartilage left behind by the growing epiphysis, which looks like a cystic formation on X ray. True cysts rarely are seen. 3. LATE STAGE Bone repairs continue in the secondary ossification center until it becomes a uniform radiodense mass. There is tendency to a rounded appearance of the head, but the overall cartilage model may have become sufficiently deformed that sphericity is never completely recovered.

4. MECHANISMOF HEAD DEFORMITY Several drawings were presented indicating possible modes of the entire femoral head deformity. One difference

290

CHAPTER 4 ~ Legg--Calve--Perthes Disease

from previous studies is the description of vertical cracks from the articular surface down to the secondary ossification center, described as being "not infrequently observed." Other reports have not confirmed this, but many of the observations by Mizuno et al. appear to have been made at open arthrotomy during relatively early, active phases of the disorder. These authors propose primarily a mechanical deformation of the cartilage model of the head once necrosis has occurred due to compressive or tangential shearing stresses. The compression effect may alter mechanically the normal spherical shape of the articular surface into an ovoid shape; it may cause collapse of the articular surface into the softened necrotic subchondral bone, also leading to an ovoid shape; it may lead to a more focal collapse of cartilage into the necrotic bone, forming a central depression; or it may crack the articular surface into the softened subchondral regions, leading to flattening and lateral extrusion of the model of the head. An added tangential or shearing stress would cause a lateral shift of the head eventually leading t o an articular surface crack or a subchondral dead space due to compression of the necrotic bone, leading to the shaping changes listed for the purely compressive mode.

G. Dolman and Bell, 1973 These authors reported a patient who died at 6 years of age with unilateral Legg-Perthes disease complicated by severe glomerulonephritis (62). Macroscopically, the femoral head was flattened but the articular cartilage was smooth although depressed at the insertion of the ligamentum teres. The central region of the secondary ossification center consisted of dead bone with empty lacunae and amorphous debris in the marrow spaces. The bone trabeculae and marrow were normal at the lateral margin of the epiphysis as well as medial to the insertion of the ligamentum teres. The necrotic center was bordered by a zone of fibrous tissue, which had areas of high cellularity with cells clearly identifiable as osteoblasts and osteoclasts. Osteoblasts were seen to synthesize new bone on dead trabeculae. The articular cartilage was viable and its surface smooth, but endochondral ossification on its undersurface was not seen. The epiphyseal plate chondrocytes normally were aligned, but areas of irregularity were seen with cartilage fragments projecting into both the secondary ossification center and the metaphysis. The authors felt that the increased bone density radiographically was the result of compression, with more trabeculae being in a unit area than normal, and of the presence in the marrow spaces of amorphous radiodense material referred to as bone dust. The radiolucent zone of the "head within the head" appearance was due to the concentration of granulation and fibrous tissue devoid of bone around the central necrotic bone. They also noted that bone repair largely occurred by intramembranous ossification. The epiphyseal plate also was affected, being cracked and buckled focally. The cartilage cells failed to align and calcify at both the undersurface of the articular cartilage and the epiphyseal plate.

H. Larsen and Reiman, 1973 Histologic studies were obtained from 13 patients with Legg-Perthes disease undergoing intertrochanteric osteotomy (164). At surgery the joints were opened and a thin wedge of tissue was removed from the anterior surface of the femoral head, consisting of the articular cartilage and a small part of the periphery of the ossification center. The terminology used for assessment involved stages 1, 2, and 3, which were initial, intermediate, and delayed based on preoperative radiographs. In the early stages the cartilage of the femoral head either was normal in shape or on occasion it was flattened slightly. This study is of particular value because the cartilage surface is assessed from the joint surface through the entire articular surface down to the epiphyseal cartilage and just barely into the bone of the secondary ossification center. In each of the five stage 1 specimens there was slight to pronounced proliferation in the basal part of the cartilage. In stages 2 and 3 there also was evidence of islands of cartilage present within the secondary ossification center. This indicates that the cartilage continues to grow and proliferate but does not effectively undergo ossification due to the lack of appropriate vascularity within the secondary ossification center. The arthrogram showed the cartilage contour of the femoral head to be almost normal despite pronounced flattening of the ossification center. In addition, the acetabulum in these cases was always normal. An interesting observation was that "in no case was an increase in the synovial fluid observed when the hip joint was opened." "In four cases there was some thickening of the joint capsule but the synovial membrane was always considered to be normal." Biopsy from the synovium showed only slight hyperemia in some cases but generally no abnormality. The authors had the impression that the cartilage thickness was greater than normal, but this would be difficult to confirm without a very detailed comparative study. There was newly developed bone of osteoblastic apposition. The articular cartilage, though thickened, was always normal in the superficial layer. "Basally there was sign of slight irregular proliferation." In the intermediate stages the articular cartilage also was thick with "pronounced proliferation basally." It is not clear whether the description of this island of cartilage is truly pathologic because it may represent perhaps normal endochondral bone formation on the undersurface of the articular cartilage whose lower margins are always more active.

I. McKibbin and Ralis, 1974; McKibbin, 1975 These authors pointed out the possibility, based on previous experimental work, that Legg-Perthes was the end result of more than one episode of infarction of the femoral head (186, 187). They assessed the femoral head of a boy aged 9 years who had been treated for 2 years prior to his accidental death. This important study demonstrated the articular cartilage to be healthy with the surface shiny, although the head was deformed "by a large central indentation." On hemisec-

SECTION V ~ Subsequent Pathologic Reports with Better Defined Correlations tion the thickness of the articular cartilage was very irregular, being twice as great in some places compared with others. The ossification center was flattened and irregular with the basal portion adjacent to the growth plate containing blood vessels, whereas the more superficial parts were avascular. The trochanteric epiphysis and metaphysis of the neck were normal. The growth plate was irregular, appearing to be breached by transphyseal vessels centrally. The secondary ossification center demonstrated normal bone at its lower part adjacent to the growth plate. The trabeculae were thicker with osteoblastic new bone appositional formation. In the superficial zone there was total avascularity and bone death. The trabeculae had empty lacunae and were themselves broken and disordered. In some instances "the trabeculae lay so close together as to suggest that they had suffered the effects of mechanical compression." Other trabeculae were thicker than normal and, on closer study, appeared to represent new bone synthesized on older bone except that lacunae in both layers were empty. McKibbin and Ralis interpreted the finding of dead repair bone on persisting original bone to indicate that there had been more than one episode of necrosis. Repair could have occurred only on the basis of revascularization, and necrosis of repair bone therefore must indicate another episode of major infarction. There was a surrounding zone of fibrous granulation tissue, which had the appearance of an advancing front approaching the dead bone. There was a concentration of osteoclasts adjacent to this area of invasion. The irregular thickness of the articular cartilage was confirmed, although for the most part this was healthy. There were areas in which endochondral ossification was not occurring. The growth plate also was irregular in thickness and at its center there was a deficiency through which granulation tissue connected epiphysis and metaphysis. The authors noted that cartilage overlying the dead bone not only was thickened but showed necrosis of the basal chondrocytes and no endochondral sequence. This would indicate that the inner vascularity was affected even though synovial diffusion continued. They concluded that the head had been the site of a previous infarct, following which almost complete vascularization had taken place, following which another avascular episode occurred. McKibbin reviewed the changes defined by the assessment of gross and histologic examinations from human specimens. In relation to the articular cartilage, he pointed out that the cartilage overlying the secondary ossification center was composed of the surface articular cartilage and also the underlying epiphyseal cartilage, which was effectively a second type of tissue in that it was responsible for growth of the femoral head in the interstitial fashion as well as serving as the endochondral growth plate region for expansion of the periphery of the secondary ossification center. The articular cartilage continued to survive because it derived its nutrition from the synovial fluid by diffusion. He pointed out the importance of recognizing that the cartilage at the periphery of the secondary ossification center received its nutrition primarily from the same vascular supply that nourished the

291

bone and marrow. Because the latter clearly were affected by infarction in the Perthes process, it was not surprising that the adjacent cartilage of the endochondral sequence, sometimes referred to the cartilage on the undersurface of the articular cartilage, also suffered invariable necrosis in many instances. This clearly was evident histologically by the absence of chondrocytes and empty chondrocyte lacunae adjacent to the secondary center. McKibbin then reviewed the histologic changes in the secondary ossification center in association with the necrosis. One of the main differentiating features of the pathoanatomy in relation to the end result was whether the area of infarction involved the entire head or whether it was more localized, at which time it was concentrated in the superior and anterolateral aspects of the subchondral region of the secondary center. The findings within the bone region were variable but within a specific range nevertheless. In those instances in which bone repair occurred, the stability of the secondary center was maintained and collapse was minimal. The areas that became most problematic were where osteoclastic absorption predominated over bone repair, leaving areas of decreased structural integrity. These often were filled initially with fibrous or fibrocartilaginous material, which imparted the radiolucent appearance radiographically. The long delay in the transformation of these regions to bone appeared to enhance the likelihood of bony collapse. McKibbin also pointed out that the growth plate was involved in many instances of Perthes disease. This too was not overly surprising once one understood the nutrition patterns, because the blood supply of the epiphysis was responsible not only for the secondary ossification center but, via cartilage canals, for nutrition of the physis itself. He noted that previous studies had shown the physis to be completely absent in some long-standing cases, whereas in others during the active phase of the disorder it had been breached at its center by a plug of tissue linking secondary ossification center and metaphyseal bone. Lesser irregularities in the plate characterized by a wandering or irregular path of physeal tissue had long been described. 1. MECHANISM OF HEAD DEFORMITY

McKibbin reviewed the mechanical theories for head deformation and indicated that, although some of the deformation was secondary to mechanical weakness of the bone, not all deformation could be attributed to it. Trabecular fragmentation and subsequent osteoclastic resorption clearly weakened the internal structure and would allow some degree of collapse. He indicated, however, that there was not a clear mechanical stress-collapse relationship. In discussing the central depression of the head characteristic of the relatively early stages of Legg-Perthes disease, McKibbin noted that the overall area of deformity in the normal position of weight beating would "be well within the confines of the acetabulum." He felt, therefore, that the central depression, which has been commonly described in Perthes, "suggests it is the result of growth failure and absorption from within

292

CHAPTER 4 ~

Legg-Calve-Perrhes Disease

rather than from pressures from without." He strongly supported, therefore, a more biological cause of the head deformation with mechanical factors playing a relatively secondary role. One aspect of the misshapen features was that the joint cartilage and adjacent epiphyseal cartilage would grow in an asymmetric fashion with those regions laterally more favored because of less stress on the one hand and better nutrition on the other. McKibbin also noted that the revascularization process often was interrupted such that the pathologic process operated intermittently. The concept of multiple infarctions appeared valid, and this concept of a series of repeated major infarcts made more understandable the chronicity and slow and halting repair in what was otherwise a relatively small region of abnormality.

J. Jensen and Lauritzen, 1976 Morphological studies were performed in two cases examined at necropsy (132). Case 1: A boy 4 years 10 months of age died following appendicitis, having been treated for Perthes for 1.5 years. At necropsy, there was no increased hip joint fluid on the involved side. The left femoral head was flattened and its neck was broadened with coxa vara. The surface cartilage was intact. The involved femoral head showed strands of cartilage and connective tissue extending from the surface to the epiphyseal plate. The epiphyseal growth plate was irregular. The bony trabeculae in the secondary ossification center were arranged in a haphazard way. The epiphyseal area was divided into several parts by strands of fibers and hyaline cartilage, and fibrous connective tissue had replaced the bone marrow. The bone trabeculae were thick and irregular. New bone formation was seen. A narrow zone within the deeper layers of the surface cartilage was necrotic and replaced by granulation tissue. Growth plate function appeared normal where present. Several trabeculae contained a nucleus of lamellar bone with a layer of immature woven bone covering the surface. "In most of these trabeculae both the central lamellar bone and the peripheral woven bone were necrotic." The fact that woven bone was necrotic also would support the occurrence of infarcts at differing time periods. The second case was a 6-year-old boy treated for LeggPerthes for 1.5 years who died from a brain tumor. At necropsy the hip joints were normal in terms of their capsule and amounts of fluid. The joint cartilage was smooth and intact although the femoral head was flattened and irregular. The upper femoral epiphysis was divided into several parts by septae of fibrous tissue and islands of cartilage. The femoral neck was widened slightly and there was a moderate coxa vara. The bone structure of the epiphysis was irregular with bone and bone marrow being replaced to a large extent by cartilage, fibrous tissue, and granulation tissue. Cells of the deeper layers of the surface cartilage were necrotic. The bone trabeculae were composed of both lamellar and woven bone with no remnants of necrotic bone detected. The authors

stressed, however, the degenerative alterations of the cells in the basal areas of the surface cartilage. They also interpreted changes in the bone in the first case as representing at least two episodes of ischemia. They felt that the vascular fault in Legg-Perthes was within the retinacular vessels.

K. Inoue, Freeman, Vernon-Roberts, and Mizuno, 1976 The histologic appearance of 57 femoral head biopsy specimens in Perthes disease was studied (128). Studies were performed from biopsy specimens taken from the femoral head with a special needle after opening the joint capsule. The cylinder of bone removed was 0.4 cm wide and up to 1.5 cm long. At least one infarct was identified in each hip. Any example of woven bone that normally is not seen in the postnatal femoral epiphysis is considered to have formed during repair following the Perthes insult. Definite pathologic evidence of more than one infarct was seen in 51% of the hips. The characteristic histologic change that enabled that diagnosis to be made was the presence of dead repair tissue. Because the tissue was obtained only by small core biopsies, it was felt that the technique alone was highly likely to underestimate the true incidence of double infarction. The findings of double infarction support the concept that the deformation of the femoral head and the chronicity of Perthes may be due at least as much to repeated episodes of infarction as to purely mechanical factors. These observations followed upon experimental studies in the dog in which repeated episodes of infarction were produced and assessed histologically. The histopathologic hallmark of repeat infarction is the occurrence of necrosis in mature lamellar bone, which is followed by the synthesis of woven bone on its surface and evidence that the woven bone itself undergoes ischemia as its lacunae become empty as well. No cause for the infarcts was identifiable. The authors suggested that the chronicity of the disorder was due to the repeated episodes of infarction. These events also led to abnormalities of growth, which were considered to cause the irregular shape of the head rather than the previously held theory of mechanical deformation secondary to weight bearing.

L. Inoue, Ono, Takaoka, Yoshioka, and Hosoya, 1980 A Japanese study assessed 57 biopsy specimens from the capital nucleus of the femoral head with Perthes disease (129). The authors also concluded that the necrosis was due to repeated episodes of infarction. The Perthes was classified as initial, intermediate, or late stage. The Perthes biopsies in the initial stage show complete necrosis following one episode of infarction. Some repair tissue had already extended into the capital epiphyses. In the intermediate stage most specimens showed a zone of mature granulation tissue, ini-

SECTION V ~ Subsequent Pathologic Reports with Better Defined Correlations tially without bone formation and formally advancing as a front of repair.

M. Catterall et al., 1982a,b (a) Catterall and colleagues studied the hips from two children with Perthes disease who died from unrelated causes (45). A 12-year-old boy, 2 years into his Perthes disease, died due to chronic nephritis. Radiographs showed a Catterall group 3 Perthes disorder. At the time of necropsy the disease process was well into the healing phase but without serious deformity of the head. Macroscopic exam indicated the superior and lateral margins of the femoral head to be flattened slightly and the cartilage somewhat fibrillated, but overall sphericity was well-maintained. The medial, posterior, and lateral segments of the epiphysis were composed of normal trabecular bone and marrow. The articular cartilage over these regions was normal with normal endochondral ossification on its deep surface. Pathological features were concentrated in the central and anterior segments. The articular cartilage was thick but varied in thickness in some areas compared with others. Subchondral clefts were still seen. Centrally the cartilage was continuous with fibrocartilage in the previous region of the secondary ossification center, and this itself was being invaded by vascular tissue in which some woven bone was being formed. Appositional bone formation was occurring along with remodeling and irregular thickening of the trabeculae. Bone necrosis was seen in only a few instances. The epiphyseal plate was quite irregular in shape rather than being curvilinear. Case 2: An 8-year-old boy presented with Perthes disease but died shortly afterward from lymphocytic lymphoma. The Perthes disease was in a healing phase. Gross and specimen radiograph illustrations are shown in Figs. 3A-3C. The involved femoral head was enlarged and flattened on its superior surface. The femoral head had grown too large for the acetabulum and hinged on its lateral rim, causing a depression of the head surface. The acetabular roof also had become flattened. There were adaptive secondary changes in the acetabulum. Macroscopically the head was broad, flat, and indented at its outer rim by the acetabulum. Certain trabeculae were showing remodeling. There was a large segment of necrotic bone in the central region of the femoral head surrounded above by the articular cartilage, below by epiphyseal cartilage, and to each side by reactive tissue both fibrous and fibrocartilaginous in nature. The peripheral articular cartilage continued to grow and ossify on its inner layers. The growth plate was abnormal in all areas with distortion of the cartilage columns. Often there was cartilage within the metaphysis. These cases served to illustrate the variability of the extent of involvement, the infarction leading to necrotic bone, trabecular fractures, and the disorganized repair process. Bone formation in the damaged femoral head occurred by two mechanisms: (1) replacement of necrotic trabeculae by

293

the intramembranous mechanisms of new bone laid down on the trabeculae and (2) much more irregular mechanisms involving endochondral bone formation and frequently apparent metaplasia from fibrous and fibrocartilaginous tissue. In cases in which there was extreme distortion of the femoral head, the bone necrosis and repair appeared to be complicated by subchondral fracture and epiphyseal collapse. Appositional bone predominated when the marrow viability returned, a repair characterized by invasion of granulation tissue. A callus response may occur in relation to fracture and predispose one to fibrocartilage formation. The authors felt that fibrocartilage was observed "in areas where trabecular bone has previously been crushed and removed." Comments also were made on the metaphyseal lesions. In these cases extensive areas of cartilage tissue positioned in the metaphysis were noted particularly at the anterior and lateral aspects. The growth plate had lost its normal structure in these regions. The more deformed second case demonstrated how, when the femoral head was enlarged and deformed, significant degrees of ossification within the cartilage further complicated the problem by imparting rigidity to the lateral aspect of the femoral head. (b) Catterall and several colleagues, all of whom had personal experience assessing femoral head tissue from patients with Legg-Perthes disease in terms of the underlying histopathology, reviewed six whole femoral heads and core biopsies in five other cases (44). Cases were categorized by the Catterall group I-IV approach. Secondary Ossification Center: There is only one case in group I in which the overall shape of the femoral head was maintained. There was a central area of bone resorption in the subchondral area. Necrosis was not seen histologically. The defect at the anterosuperior margin seen radiographically contained fibrocartilage continuous with the overlying articular cartilage. The material was cellular, the bone deep to it had persisting cores of cartilage, and at the stage of assessment it had no active ossification. Vascular granulation tissue was invading from below. In groups II and III, in the superior part of the secondary ossification center, trabeculae were necrotic and many were fragmented; in the subchondral region, the trabeculae were thickened with many cement lines and the marrow was necrotic. In the region above the growth plate, the trabeculae were thickened with central necrosis and viable appositional new bone formation on their surface; the marrow was replaced by granulation tissue, and lying between this and the necrotic area was avascular tissue. In the medial and lateral segments, trabecular bone had a normal appearance. Two types of bone repair were seen. In relation to the necrotic trabeculae, vascular connective tissue invaded the region and synthesized new bone on the dead trabeculae. Active resorption by osteoclasts also was observed. Where the trabeculae were fragmented, the marrow was necrotic and the residual trabeculae were thickened and necrotic. In these areas only

294

CHAPTER 4 ~ Legg--Calve--Perrhes Disease

F I G U R E 3 Examples of deformed femoral heads in Perthes disease compared with the normal contralateral head and acetabulum. (A) Gross specimen shows the central depression of the femoral head (left) characteristic of early deformation in many cases of Perthes disease. The head also is larger than that on the normal contralateral side (right) and the acetabulum is misshapen slightly as well. (B) Specimen photographs following hemisection of the hip are shown involving the affected (right) and normal (left) hips. Note the significant flattening of the cartilage model of the femoral head, the associated acetabular shape change, the small and irregular appearance of the secondary ossification center, and the irregular line of physeal cartilage. This photo also serves as an example of the clinical phenomenon of hinge abduction. (C) Specimen radiographs of Perthes hip (right) and normal side (left) show fragmentation of the necrotic and repair bone in the affected secondary ossification center along with the waviness of the physeal line, early change in shape of the adjacent acetabulum, lateral subluxation of the head and neck region, and widening of the neck. [Reprinted from Catterall, A., et al. (1982). J. Bone Joint Surg. [Br] 64B: 276-281 (45), with permission.]

SECTION V 9 Subsequent Pathologic Reports with Better Defined Correlations fibrous tissue and fibrocartilage were present, with all the bone having been resorbed. A chondroid tissue formed from the fibrous tissue similar to what is seen in an immature fracture callus. In group IV, fibrocartilaginous material was present at the site of the original secondary center. Necrotic bone trabeculae were seen, which appeared to indicate successive episodes of ischemia and remodeling. At the periphery there were bony trabeculae showing one episode of infarction only. Articular Cartilage: The articular cartilage was thicker than that of normal controls. Where it overlaid necrotic bone there were areas of necrosis in its deep part and normal endochondral ossification was absent. In some sites, the deep surface of the cartilage was undergoing resorption and replacement by fibrocartilage. Epiphyseal Growth Plate: This was abnormal in every case. Interference with ossification was greater than normal, with collections of cartilage stretching unossified into the metaphyseal region. Metaphyseal Changes: Four types of change were noted in the metaphyses: (1) adipose tissue was present on occasion in increased focal collections; (2) radiographs sometimes showed osteolytic lesions with well-defined sclerotic margins, which histologically were composed of fibrocartilage that often was immediately adjacent to the growth plate; (3) where the growth plate was widened along with the head and neck often there were disorganized columns of unossifled cartilage streaming down into the metaphysis but no necrosis was observed; and (4) extension of the growth plate down the side of the femoral neck in cases in which there was deformity of the femoral head. The authors also observed that, when compared with normal controls, the contralateral radiologically nonaffected side articular cartilage and growth plate showed abnormalities suggesting a preexisting condition. Assessments in group IV femoral heads suggest that the whole epiphysis had been involved in the ischemic process and that there may have been repeated episodes of infarction. There are widespread attempts at repair by remodeling of the thickened trabeculae and by ossification of thickened fibrocartilage. The study was felt to confirm the fact that the extent of any infarction is variable and consistent with the degree of radiological involvement. This would be absent in group I, localized in groups II and III, and extensive in group IV. The prognosis for any case of Perthes disease was felt to be "proportional to the degree of infarction present within the epiphysis." The metaphyseal lesions were felt, for the most part, to be accumulations of unossified cartilage derived from the growth plate.

N. Ponseti et aL, 1983 Histologic studies were done from the lateral protruding aspect of the femoral head and the small portion of the adjacent

295

5

r

k

d.

8

13

e.

f.

20

26

g.

h.

l

39

F I G U R E 4 A series of drawings based on lateral radiographs of an affected hip in a 7-year-old girl shows the characteristic findings as the disorder progresses to full healing with a coxa magna, slightly flattened head, and corresponding shape changes in the lateral acetabulum.

physis and femoral neck obtained at surgery from five boys 8-12 years of age, who had severe limitation of hip abduction as well as discomfort in association with their Perthes disorder. The authors felt that the articular cartilage was normal but that the epiphyseal cartilage immediately underlying it was structurally abnormal, containing sharply demarcated areas of hypercellular and fibrillated cartilage with prominent blood vessels. The lateral physeal margin also was irregular. Ultrastructural examination revealed many irregularly oriented large collagen fibrils and variable amounts of proteoglycan granules. It was not possible to be certain whether the epiphyseal cartilage abnormalities were primary or secondary, but Ponseti et al. postulated that abnormalities of the epiphyseal cartilage matrix could lead to a collapse and necrosis of the femoral head followed by abnormal ossification (220). Figure 4 illustrates the sequence of changes from onset to healing as seen on lateral radiographs. Figure 5 summarizes the histopathologic changes and the dual mechanisms of head deformation. Varying mechanisms underlying the pathogenesis of deformity have been proposed (107, 144, 187, 194, 239, 240, 250).

296

CHAPTER 4 9 Le~7~7-Calve-Perthes Disease

PATHOGENESIS OF FEMORAL HEAD - ACETABULAR DEFORMATION IN LEGG-CALVE-PERTHES DISEASE

A mixture of biologic and mechanical-physical phenomena The initiating event in LCP is considered to be avascularity of the proximal femoral capital secondary ossification center. The cause of the avascularity is unknown. This leads to necrosis of bone and marrow of the secondary ossification center, which can be partial or complete. The area always involved with necrosis is the subchondral bone of the secondary ossification center immediately adjacent to the articular and epiphyseal cartilage. With more extensive to complete involvement of bone the necrosis progresses towards the bone adjacent to the epiphyseal cartilage overlying the physeal cartilage. The bone and cartilage of the greater trochanter always remain normal throughout the entire LCP process. The plain x-ray at this time is normal since there has been no collapse of the secondary ossification center and its radiodensity remains normal due to structural persistence of bone matrix with its mineralization undisturbed. The radiodensity is the same whether or not the osteocytes and marrow are alive or necrotic. With time (weeks) the cartilage model of the head continues to increase in size and remains spherical since cartilage growth is unaffected receiving nutrition by diffusion from the synovial fluid of the hip joint. Both bone scan and MR imaging will show abnormalities however since they reflect the vascularization of tissues. The bone scan in particular is "cold". The secondary ossification center will not increase in size compared to the nonaffected side as the avascular state stops the endochondral bone mechanism. Events to this stage in the pathogenesis of the disorder have been biologic in nature. As the child continues to walk there is a tendency to subchondral bone fracturing within the necrotic outer portion of the secondary ossification center. The articular cartilage remains viable and initially remains normally shaped even with the subchondral. fracture. The articular cartilage surface with an underlying strut of (necrotic) subchondral bone "springs back" to retain the normal surface curvature. Radiographs show the crescent shaped radiolucency of the subchondral bone in particular on lateral projections. The radiolucent space is due to absent bone trabeculae at the fracture site, early fibrous tissue ingrowth and beginning resorption of necrotic bone. The bone of the secondary ossification center now generally appears flatter on radiographic projections at its most superior aspect. This occurs due to the collapse of the subchondral trabeculae at and immediately inferior to the fracture site. The bone also appears denser radiographically owing to the fact that the necrotic trabeculae are crushed and displaced into the more central parts of the secondary ossification center and interdigitate with one another in a closer fashion. The combination of necrotic interdigitating trabeculae and new bone formation with early repair leads to the increased bone density. With continued motion and weight beating of the hip there can be depression of part of the articular cartilage leading to a central and lateral depression of the articular cartilage surface and even to surface furrowing in more severe states. The combination of compression and associated shear stresses which are clearly mechanical in nature lead to the articular surface collapse, furrowing and occasional oblique rupture, The term coxa plana can involve two phenomena, which must be carefully distinguished. That seen on a plain radiograph involves the malformation of the superior surface of the secondary ossification center owing to necrosis, physical fracture and collapse of the subchondral bone. It requires arthrography or MR imaging of the articular cartilage surface to determine whether or nor it also is flattened. If collapse of the articular cartilage surface occurs then at full healing the coxa plana can represent both secondary ossification center bone and articular cartilage surface flattening. If the articular surface remains spherical then full repair to a normal state can occur. Coxa magna refers to an enlarged head, which is one of the frequent sequelae of the disorder. Some of the widening of the head may be due to physical collapse of the subchondral bone and spreading of the adjacent articular and epiphyseal cartilages. Much of the coxa magna however is due to increased cartilage model growth at the periphery in association with the synovitis associated with the LCP. The fact that the head in LCP eventually becomes larger than that on the noninvolved side implies increased growth rather than simple mechanical malpositioning. Since growth is mediated by expansion of the cartilage model and since the cartilage receives its nutrition from the synovial fluid which is increased in the disorder a true biologic increase in size has occurred. The neck is often wider reflecting the excess of repair. As the cartilage model of the head widens the neck itself will be wider since it derives much of its structure from the endochondral physeal sequence but neck widening is also due to periosteal new bone apposition. The repair phenomenon involves both head and neck of the femur and is associated with increased blood supply of the periosteum working its way from the circumferential ring upwards. The neck is frequently shorter than normal (coxa breva) since the LCP often involves diminished physeal growth in whole head involvement. The physis receives its nutrition from the blood of the secondary ossification center. Coxa vara occurs owing to diminished growth of the head-neck complex in association with the continuing growth of the nonaffected greater trochanter. The acetabular contour frequently adheres to the shape of the femoral head in particular in those developing LCP under 9 years of age. The acetabulum grows by modeling of the articular and epiphyseal cartilage which is very sensitive to adjacent pressures. In relation to the radiographic and MR imaging appearance of secondary ossification center bone the following must be borne in mind. Where fragmentation is seen (alternating areas of radiodensity and radiolucency) the radiodensity is owing to compressed necrotic trabeculae and new bone formation on the old necrotic trabeculae (more bone per unit area). Areas of radiolucency occur because of bone resorption by osteoclasts leading to less bone per unit area and the presence either of fibrous tissue, fibrous tissue with vascularization, early osteoid with incomplete mineralization, and islands of cartilage tissue from the epiphyseal region not appropriately incorporated into bone. FIGURE 5 Textsummarizesthe gross and histopathologic changes in Legg-Perthes disease and the apparent mechanismsof head, neck, and acetabular deformation.

SECTION VII ~ Pathoanatomic Changes

VI. E A R L Y C O R R E L A T I O N O F RADIOGRAPHIC WITH HISTOPATHOLOGIC AND CLINICAL FEATURES OF LEGG-PERTHES FROM THE INCIPIENT STAGE TO THE RESIDUAL STAGE Ferguson and Howorth described in great detail the radiographic features of Legg-Perthes disease (coxa plana) as early as 1934 (75). A brief summary of their findings remains informative today. Incipient Stage: The capsule is distended, the cartilaginous space is widened (especially inferiorly), and there is slight flattening of the crest of the head and slight widening of the epiphyseal growth plate. Active Stage--Early: There is a denser radiographic line at the crest or base of the head (by which is meant the secondary ossification center), and the secondary center is shallower and broader than on the normal side. Active Stage--Late: The radiographic appearance of the neck shows irregularity (by which is meant the lack of uniform radiodensity), there are dense radiographic areas within the head, and there is a greater degree of widening of the joint space, shallowness of the head, and broadening of both the head and neck. Occasionally there is some decalcification of the innominate bone and upper half of the femur (which would be due to decreased weight bearing and perhaps increased generalized vascularity) and protrusion of the head laterally beyond the ilium due to broadening of the head and widening of the cartilaginous joint space inferiorly. Transition from the Active to the Reparative Stage: There is a decrease of density in previously dense areas of the femoral head bony center, with the appearance of irregular ossification as the density decreases. New areas of density may have appeared (in the secondary ossification center) whereas the first areas were becoming ossified irregularly. Occasionally there are irregular areas of ossification in cartilaginous areas at the margins of the epiphyseal line. Obliquity of the roof of the acetabulum and lengthening of the acetabulum are seen. Occasionally there is a semblance of separation of a fragment of bone at the surface of the head due to the extensive loss of texture in developing irregular ossification, which they defined as "osteochondritis-dissecans type of coxa plana." Reparative Stage: The dense areas of bone in the secondary ossification center are replaced almost completely by irregular ossification, which itself is becoming more uniform in density. There is a coarse texture of the trabeculation of the femoral neck. Occasionally there is a semblance of a cavity in the head or in the neck adjacent to the head due to extensive loss of texture in association with irregular ossification. Reparative StagemLate: The texture (by which is meant bone density radiographically) of the affected areas tends toward the normal. The secondary ossification center of the

297

head becomes less flat, with the rounded contour being derived from the highest remaining point of the head. The broadened or widened head may have spread down over the neck in the regions in which the edge of the head had been irregularly ossified. The acetabulum tends to conform to the contour of the head. There is shortening of the neck due to deficient growth in length, and the joint space becomes less wide (with increased new bone formation of the secondary ossification center). Residual Stage: The lateral margin of the head where it has grown down over the neck develops a dense line crossing the neck often quite near the trochanter. When this bony density is present, the process of the entire disorder may be regarded as complete, although occasionally there is a little further repair to be accomplished. The head has become rounded unless pressure of the upper lip of the acetabulum on a subluxated or broad femur has interfered with this rounding. This detailed list refers to several of the sequelae of the Legg-Perthes disorder, which came to be quantified and better understood only over several succeeding decades. Among the concepts described that are now increasingly recognized are the widened cartilaginous joint space, which is a feature of persisting cartilage growth in the absence of adjacent secondary ossification center bone formation, and the protrusion of the head laterally beyond the ilium, which is due to broadening of the secondary ossification center of the head and widening of the cartilaginous tissues, which continue to grow. The fragmentation stage is well-described with alternating areas of bone density and radiolucency signifying the nonuniform rate at which bone repair occurs. The response of the acetabulum to the disorder is evident in descriptions of the obliquity of the roof of the acetabulum, the lengthening of the acetabulum, and the eventual conformity of the acetabulum to the contour of the misshapen femoral head. The delayed repair of a central fragment of bone at the surface of the head, the so-called osteochondritis dissecans lesion, is described as the late reparative stage when bone density becomes uniform. The rounded contour of the head is reestablished in some instances, although in others the shape of the head is imperfect due to pressure of the outer rim of the acetabulum on the subluxated or broadened head.

VII. P A T H O A N A T O M I C C H A N G E S AND THEIR RELATION TO THE CLINICAL, R A D I O L O G I C , AND O T H E R IMAGING FINDINGS

A. Overview of Plain Radiographic Changes in Legg-Perthes Disease Figure 6 summarizes the many early and intermediate stage plain radiographic changes that have been recognized. Additional imaging techniques using scintigraphy, arthrography,

298

CHAPTER 4 9

Leyg--Calve--Perthes Disease

computerized tomography, and magnetic resonance imaging provide increasingly more valuable information. In Legg-Calve-Perthes disease there is always eventual full repair of the necrotic bone of the secondary ossification center with normal bone. Long-term problems occur in those situations in which the articular cartilage surface of the femoral head relates imperfectly to the acetabulum due to changes in its size and shape during the long disease process. The extensive radiologic and anatomic changes that accompany Legg-Calve-Perthes disease are associated with the repair process as well as with the initial necrotic insult. Much of the confusion surrounding the radiographic changes in the condition rests on the absence of histologic specimens for correlative study in individual cases, such that our perceptions of the disease have come almost exclusively from twodimensional radiographic projections of the three-dimensional femoral head, neck, and acetabulum. The radiograph does not show the cartilage structure, whereas the bone detail represents a summation of the cell and tissue processes throughout the head, neck, and acetabulum. Histologic studies, both in human tissue and in animal models, demonstrate considerable nonuniformity of necrosis and subsequent repair. The additive picture given by the radiograph can be quite confusing; certain parts of the head may be noninvolved, certain parts may be necrotic with repair not yet having begun, other parts may have suffered necrosis with the repair process well underway, and certain areas of repair bone may suffer additional episodes of necrosis. The following radiographic studies presented major early interpretations for the disorder. 1. FREUNO Freund presented one of the earliest and most detailed studies of the developmental changes in Perthes disease (81). He illustrated the lateral epiphyseal notching as part of the early repair response and also the fracture cleft in the subchondral spongiosa early in the disease process. His line drawings reveal the entire spectrum of the disorder. 2. BRAILSFORD

Brailsford defined 12 radiographic stages of Perthes from the beginning of the disorder to complete healing (26). He noted that changes occurred often for a period of 4 years or more from the time the disease was recognized, during which time he referred to the bone as being in a "plastic state." He also pointed out that the secondary changes "began to develop in the roof of the acetabulum of the affected hip joint" in association with the femoral head changes. He defined the term "plastic" to indicate a bone or epiphysis that is not capable of withstanding normal pressure without becoming deformed. He considered the plastic state to last until the epiphysis was reorganized completely and a normal density of epiphyseal bone structure had been reestablished. In this early use of the term, therefore, there was no implication as to whether the plasticity was biological or mechan-

12

1

5

( F I G U R E 6 The early and intermediate stage plain radiographic changes in Legg-Perthes disease are illustrated in this figure. (1) Bulging of the joint capsule; (2) smaller secondary ossification center than opposite normal side; (3) increased medial joint space; (4) subchondral crescent sign; (5) rarefaction (lysis) in lateral outline of epiphysis (Catterall sign); (6) lysis of superolateral margin of neck (metaphysis), Gage sign; (7) band-shaped strip of metaphyseal radiolucency; (8) translucent area of medial metaphysis; (9) thickening of physis; (10) widened neck; (11) tear-shaped widening of medial pelvis; (12) acetabular roof changes, osteoporosis, increased upward slope. Modified from Edgren (68), with permission.

ical. Brailsford's 12 stages of radiographic appearances in Perthes disease follow: (1) an increase in the density of the femoral capital epiphysis; (2) a relative increase in the joint space; (3) osteoporosis of the adjacent extremity of the diaphysis [three types of osteoporosis could be found, including (a) linear zones of translucency following the lines of the principal bone trabeculae, which thus ran obliquely downward and outward from the epiphyseal line; (b) a zone of translucency of almost uniform depth across the proximal diaphysis or neck; and (c) one or more circumscribed areas of translucency near or close to the epiphyseal growth plate that had the appearance of cysts]; (4) the dense epiphysis begins to show the signs of compression and impression fractures; (5) the appearance of fragmentation, in which the epiphysis appears to be broken into a number of dense fragments with a loss of homogenous appearance; (6) the epiphysis is compressed and flattened further, with some of the lateral fragment of the secondary center appearing to be displaced beyond the lateral margin of the roof of the acetabulum; (7) gradual absorption of the dense islands of bone; (8) compression and expansion of the proximal end of the diaphysis, by which is meant widening of the neck; (9) a faint outline of a regenerated epiphysis in which the dense fragments appear to be undergoing absorption; (10) the last dense epiphyseal bone is absorbed, being replaced by an area of relative osteoporosis; (11) the neck shows increased deposition of new bone with the obliteration of previous osteoporosis; and (12) the cancellous structure of the epiphysis assumes the radiographic appearance of a normal bone.

SECTION VII 9 Pathoanatomic Changes

Brailsford indicated that stages 1-3 were present from 1 to 3 months after the onset of symptoms, stages 4 - 8 tended to be present from 3 to 18 months, stages 9-11 from 1.5 to 4 years, and stage 12 was not reached until the fourth year or later. He considered the state of plasticity to be present up until the completion of the 1 lth stage. He also felt strongly that immobilization treatment was mandatory throughout the entire plastic stage. Changes occurred in the acetabulum either at the same time or subsequent to the healing of those seen in the epiphysis. The acetabulum tended to be compressed, which was most evident at the upper and lateral borders where the bone projected beyond the normal boundary. Brailsford also noted the common patterns of resulting deformity at the end of growth. In the first, in which pressure had been distributed evenly over the plastic femoral head and acetabulum, the head was uniformly expanded over the adjacent neck, whereas the acetabulum was flattened out and enlarged to accommodate the large femoral head. In the second group, in which the lateral fragments of the epiphysis were "squeezed beyond the lateral border of the roof of the acetabulum during regeneration," the displaced head fragment developed and eventually came to lie with its more medial border against the lateral border of the roof of the acetabulum serving to limit abduction. This indeed would represent an early description of what is currently known as "hinged abduction." 3. KEMP AND BOLDERO

Kemp and Boldero felt that lateral displacement of the femoral head was the earliest radiographic sign, preceding the apparent increase in the overall cartilage space by about 2 weeks ( 151). They felt that lateral displacement of the head of the femur was always seen provided adequate radiographs were taken in the early stages. The finding often was subtle such that specific measurements were required because the increase in displacement often was only 1-2 mm. The apparent increase in the cartilage space was due to the arrest of growth of the epiphyseal secondary center. The lateral displacement was real and not dependent on positioning. The lateral displacement, again defined as the earliest radiological sign of Perthes disease, occasionally disappeared after a period of rest, and they felt it was due to hyperemia of periarticular and intra-articular tissues. The third stage of this triad involved a dense appearance of the bone of the secondary center, by which time the disorder was well-established. The lateral displacement was basically defined as an increase in the medial joint space. The anterolateral displacement was confirmed by arthrography to be due to soft tissue enlargement in the acetabular fossa. The radiological increase in joint space also was accounted for by the temporary arrest of growth of the secondary ossification center. The next stage of the disorder, the radiological increase in density of the femoral capital epiphysis, appeared to be due to initiation of the repair phase with new lamellar bone laid down on per-

299

sisting dead trabeculae. In later stages, collapse of part of the head of the femur led to physical compression of the trabeculae within a smaller unit area and sequestration of tissues in the intertrabecular spaces, which acted as nucleation centers for new mineralization. 4. CAFFEY Caffey also reviewed several basic findings particularly concerning early radiologic changes in Legg-Perthes disease (31). He too felt that slight lateral displacement of the femoral head either with or without slight diminution of the size of the femoral ossification center was among the earliest changes noted radiologically. The frog lateral position or Lauenstein position was best to note the early subchondral changes in the femoral head. The earliest structural change in the osseous nucleus, other than relative diminution in size, was the "early segmental fracture," which appeared in the subchondral region and was noted particularly on the lateral radiograph. The segmental fracture was noted most clearly in the frog lateral position, which involved the femur being abducted and rotated externally. There was a sharply defined "submarginal strip of diminished density at the anterolateral and superior segment of the ossification center." It was usual that the fracture line was visible in the frog lateral position, but not seen in the anteroposterior projection. The subchondral fracture often was associated with a linear area of blackness due to the absence of tissue in that region. Caffey showed excellent examples of the subchondral fracture line, including its variable extent, its position in the subchondral spongiosa, and its tendency to heal or disappear as other changes in the secondary center bone density occurred. He also alluded to the characteristic changes in density, including fragmentation of the secondary ossification center as well as adjacent metaphyseal changes. He documented the fact that the affected ossification center was slightly smaller than its normal counterpart in 16 of 30 cases, whereas it was equal in size in 13 cases. Lateral displacement of the femoral head in the acetabulum also was commonly noted, as was the fact that it often was quite small in extent exceeding 2 mm in 26 of 30 cases but being more than 5 mm in only 2. The subchondral fracture of the ossification center was clearly present in 24 of 30 cases, whereas in 5 patients none was seen. Of these 24 cases with fractures, 19 were submarginal or subchondral in the anterolateral superior quadrant of the affected ossification center. When present, the fracture invariably was seen in the frog lateral position but relatively infrequently in the anteroposterior position. Caffey felt that, by increasing the abduction and external rotation by only a few degrees, the fracture line became more apparent and more extensive. We will now discuss the progression of pathoanatomic changes, as we can surmise them from the extensive histopathologic studies reported previously, from the initial insult onward and relate it to plain radiographic findings.

300

CHAPTER 4 ~ Legg--Calve--Perthes Disease

B. Pathologic Changes and Their Demonstration by Varying Imaging Modalities Including Ultrasonography, Scintigraphy, Magnetic Resonance Imaging, and Computerized Axial Tomography At a certain time necrosis of the secondary ossification center of the developing femoral head occurs. The condition invariably is referred to as an avascular necrosis, but no convincing cause of the avascularity has been demonstrated. Studies on the extraosseous blood supply of the head and neck region have always been normal. Necrosis of the bone and marrow cells in the secondary ossification center of the femoral head has been demonstrated convincingly. The developing bone of the greater and lesser trochanteric epiphyses remains normal. At the time immediately prior to the etiologic insult, the developing femoral head is in a normal relationship to the developing acetabulum. The labrum is appropriately positioned and the capsule is intact. The secondary ossification center radiologically is the same size and density as that on the opposite side. We can surmise that histologic studies would demonstrate normal epiphyseal cartilage development, central bone synthesis of the secondary ossification center by the endochondral mechanism with bone surrounding the calcified cartilage cores, healthy appearing osteocytes within the bone tissue and an active marrow tissue producing hematopoietic precursor cells, osteoblasts that lay down additional bone, and osteoclasts that resorb the newly formed bone and cartilage. The epiphyseal cartilage, articular cartilage, and physeal cartilage also would appear structurally normal. At the time that necrosis occurs, a plain radiograph would demonstrate no change on the involved side in comparison to the living and perfectly normal side. The initial pathologic insult affects the bone and marrow cell population of the secondary ossification center of the proximal femoral capital epiphysis, rendering the osteocyte lacunae empty and the marrow cavity filled with necrotic tissue but leaving the already present radiodense bone and calcified cartilage intact at that particular point in time. The overall size and shape of the femoral head are still normal compared to the opposite side, as is the femoral head-acetabular relationship and the size and radiodensity of the secondary ossification center. As patients virtually never present at this time, it can be assumed that the condition is either asymptomatic or so minimally symptomatic that virtually full physical activity persists. The evolving clinical and radiologic picture, therefore, is seen to be secondary to the necrotic state and reflects the sequelae of that necrosis as well as the repair processes. If the end result is less than perfect, it is as much a consequence of disordered or poorly regulated repair as it is of the primary necrosis itself. Diagnosis at this stage occasionally is made on the basis of imaging studies with much higher resolution than plain radiographs and with the ability to assess the soft tissue components. The vast majority of patients present with some

discomfort and limping, and on physical examination have a diminution of hip motion. A synovitis is present, as shown by early reports of arthrotomy and microscopic assessment. Some maintain that the synovitis is primary, leading to either vessel sclerosis or extrinsic tamponade on the basis of the synovial effusion. Others feel that the synovitis, though early, is secondary to early structural changes in the secondary ossification center--such as microtrabecular subchondral fracturemnot yet demonstrable at least by plain radiographs. Those who feel that the synovitis may be causative point to the occurrence of Perthes several weeks to months after clinical episodes of transient synovitis of the hip. Imaging techniques with much higher resolution than plain radiographs are used increasingly in the early stages of Legg-Perthes. A bone scan will be "cold" in the earliest stages of Legg-Perthes disease due to the avascularity and the fact that the repair response, which is accompanied by revascularization, has not yet begun (Fig. 7A). It is most valuable at this stage; with revascularization, the scan becomes positive (Fig. 7B). There is, however, insufficient resolution with the scan to quantitate or accurately localize the repair response. Studies of relatively large numbers of patients were reported by Sutherland et al. (263) and Fisher et al. (78). Both groups confirmed the value of the bone scan in the early diagnosis of Perthes, often with plain radiographs still normal, where the scan characteristically shows a marked absence of activity in the femoral capital femoral epiphysis. Revascularization also could be defined with increased and patchy activity within the head. The most detailed report on scintigraphy and LeggPerthes disease was based on a study done in Chicago by Tsao et al. They reported on 44 consecutive patients who underwent serial technetium 99 and m-diphosphonate bone scintigraphy using scintigraphic magnification (pinhole imaging) in both anterior and frog leg lateral projections (268). The use of pinhole magnification plus serial studies greatly enhanced the resolution of the technique. They analyzed 47 consecutive cases in 44 children in a retrospective fashion. All of the patients were followed with serial studies at an average 3- to 4-month interval during the initial year of follow-up or until a definitive pathway of revascularization was recognized. They developed a scintigraphic classification involving an A pathway and a B pathway (Figs. 8A and 8B). In the A pathway of repair, there are four stages. In stage 1A, using what is referred to as the Conway-Dias scintigraphic staging system, the femoral head epiphysis is void of scintigraphic activity in both the anterior and frog lateral projections. In stage 2A, a distinct column of scintigraphic activity is seen in the posterolateral portion of the femoral epiphysis in the anterior projection, whereas in the frog lateral position this column of activity rotates medially and thus is partially obscured by the overlying acetabular activity and is not well-defined. In stage 3A, the lateral column of scintigraphic activity is present and has extended medially in the anterior projection and anteriorly in the frog

SECTION VII ~ Pathoanatomic Changes

301

FIGURE 7 The bone scan is particularly useful for diagnosis in the earliest stages of Legg-Perthes disease prior to plain radiographic changes. (A) "Cold" bone scan of the right hip with no uptake in the femoral head is shown early in Legg-Perthes disease. The scan was performed because the patient had a known Perthes disorder of the opposite left hip and had developed discomfort on the right. (B) Positive bone scan from the opposite hip is shown during the repair phase of the Perthes disorder. (C) The anteroposterior radiograph with hips in abduction shows the normal appearing femoral head on the right and the repair phase of the Perthes disorder on the left. Part (A) shows the right hip bone scan and part (B) shows the left hip bone scan at the same time.

lateral projection. In stage 4A, complete revascularization has occurred and scintigraphic activity is demonstrated throughout the entire epiphysis in both projections. A different pattern of revascularization is documented in the B pathway. In stage 1B, the entire capital femoral epiphysis is devoid of scintigraphic activity in both anterior and frog leg positions. In stage 2B, scintigraphic activity extends centrally from

from the base of the capital femoral epiphysis and is observed in both anterior and frog leg lateral projections. The lateral column, however, is absent. In stage 3B, scintigraphic activity further extends into the base of the capital femoral epiphysis and involves approximately half of the epiphysis in both anterior and frog lateral projections. In stage 4B, there is complete revascularization with scintigraphic activity

302

CHAPTER 4 ~ L e y g - - C a l v e - - P e r t h e s Disease

FIGURE 8 The Conway-Dias scintigraphic classification of repair patterns in Legg-Perthes disease is shown. (A) The A pathway demonstrates repair beginning at the superolateral aspect of the femoral head and sweeping medially. The white dots represent vascularization and bone synthesis activities. (B) The B pathwaydemonstratesrepairbeginningjust abovethe physeal region and then passing progressively in a hemispheric fashion to involve the entire head. [Reprinted from Tsao, A. K., et al. (268) J. Pediatr. Orthop. 17: 230-239, 9 1997 by Lippincott Williams & Wilkins, with permission.]

demonstrated throughout the entire epiphysis. In their review, which was correlated with plain radiographs, 20 hips followed the A pathway, 20 followed the B pathway, and a third group of 7 converted from the A pathway to the B pathway during the course of revascularization. Tsao et al. felt their classification was valuable in the sense that no patients in pathway A went on to develop head at-risk signs, whereas 18 of 20 patients in pathway B and 7 of 7 patients in the regression pathway exhibited head at-risk signs. In addition, the bone scintigraphy classification preceded the radiographic head at-risk signs by an average of 3 months, allowing for earlier assessments as to the direction of repair. They concluded that, in pathway A, the early appearance of the lateral column on bone scintigraphy was an indication of uncomplicated revascularization of the femoral head, whereas the B pathway represented a slower rate of revascularization and healing and subsequently was associated with greater deformity. Ultrasonography has been used in the assessment of the hip with Legg-Perthes disease. Ultrasonography is particularly valuable in the earliest stages as it is able to demonstrate

the synovitis component of the disorder. Synovitis can occur in association with the earliest changes in the secondary ossification center, as revealed by plain radiographs, or on occasion even prior to these radiographic changes in the earliest stages of Legg-Perthes or perhaps in the evolution of the disorder from a transient synovitis (83). Wirth et al. (268) described their ultrasonography findings in detail in 25 hips over the course of the evolution of the complete disease. They found it most valuable early on in particular in relation to the detection of synovitis with the early disease and also assessing its diminution in association with the early relief of weight bearing. They were able to grade the synovitis and the degree of capsular distention from 0 where it was absent to 3+ where there was a large effusion. They felt that ultrasonography would be valuable clinically in determining when immobilization had been sufficient that one could proceed to bracing or surgical therapies. The ultrasonographic findings also were valuable at more advanced stages in assessing the size of the femoral head and any extrusion particularly of the cartilage model, which was occurring in relation to the more lateral aspect of the acetabulum.

SECTION VII 9 Pathoanatomic Changes

FIGURE 9 Exampleof MR imaging studyin Legg-Perthes disease is shown. The head is misshapen, there are differing signal intensities within the femoralhead ossificationcenter (white arrowpointsto area of subchondral necrosis), and there is an increase of synovialfluid within the hip joint (black arrow).

MR imaging has the capability of assessing on one examination each of the parameters revealed by the plain radiograph, ultrasonography, and arthrography. It can reveal areas of marrow avascularity, epiphyseal cartilage avascularity with diminished to absent cartilage canal signals, differing signal intensities from areas of dead and viable bone, the cartilage shape of the articular surface of the femoral head .and acetabulum, the physis with normal uniform or abnormal irregular signal, and the metaphyseal bone and marrow structure. An example of MR imaging is shown in Fig. 9. An early description of MR imaging for Legg-Perthes disease was presented by Scoles et al. (245), and several studies have been presented since then in an effort to outline the clinical applicability of the method. Pinto et al. used MR imaging in two patients suspected of having Perthes clinically but whose radiographs and bone scans were normal (218). The MR imaging showed definite avascular necrosis in the affected hip, and subsequently the radiographic changes developed. The use of MR in establishing early diagnosis thus is clear. At present, there is little clinical reason for early diagnosis because treatment modalities are not specific, but hopefully in the future early diagnosis will play a meaningful role in management. Henderson et al. assessed 24 hips with Legg-Perthes using 49 MRI studies (144). They confirmed that, early in the disease process, MRI more clearly delineated the extent and location of areas of involvement than did the plain radiographs. Hoffinger studied 20 hips with MR images and compared them with plain radiographs (124). These authors commented on the difficulty, even with MR imaging, of determining whether the so-called metaphyseal cyst rested in the metaphysis itself or was due to projectional changes and in essence was an epiphyseal lesion. They indicated that 48% of hips with Legg-Perthes and metaphyseal changes on plain radiographs had no such corresponding lesions on MRI. It is unclear why this would be the case. The other 52% of hips had physeal widening and epiphyseal-physeal irregularity on MRI. An additional 44% of hips with no plain radiographic metaphyseal changes did show metaphyseal changes on MRI. This study raises important points but

303

clearly is a long way from a definitive determination. Uno et al. studied a correlation between MR imaging and pinhole scintigraphy (bone scan) (269). They felt that MR imaging depicted the extent of involvement of the femoral epiphysis more clearly than did the bone scans. The uptake of bone scan corresponded with high or normal intensity of T2weighted images on MRI. Uno et al. noted the evolution of bone scan from a cold period with no uptake initially, corresponding to the avascular necrosis, to the progressive increase in uptake with repair, to subsequent return to full uptake at the termination of the repair phase. The MR is helpful in relation to early diagnosis, extent of bone involvement of the secondary ossification center, involvement of the physis and adjacent metaphysis, and in particular the shape of the cartilage model of the head. Bos et al. also used MR imaging to aid in the assessment of hips with Perthes disease particularly in the early phases (23). MRI exactly depicted the infarct zone in the femoral head before typical radiographic changes were seen. In plain radiographs in their patients, the subchondral fracture line was visible in 10 of 15 hips, but it was seen on all MRIs. With time, the MRI was able to delineate the extent of the infarct zone in relation to the surrounding living zone well before clear radiographic definition. There was particularly good delineation when the area of bone infarction did not actually involve the growth plate itself. More recent studies have begun to utilize MR imaging to assess more specific factors during the repair phase. Eckerwall et al. followed the remodeling of the femoral head during the repair phase particularly after proximal femoral varus osteotomy (66). They found the method excellent to assess the sphericity of the femoral head and felt that the assessment of deformation of the femoral head was much less reliable on conventional radiography than on MR imaging. The MR was felt to be superior because of its ability to image the cartilage model of the head. Sales de Gauzy et al. found their ability to assess lateral subluxation to be improved using the MR imaging compared with the plain radiograph (236). In some instances there was no change, but in others the femoral head was well-contained on the plain radiograph but subluxated on MRI, because the MR was highly effective in assessing the thickening of the cartilage portion of the femoral head. Computerized axial tomography also has b ~ quite valuable, particularly in the late stages of the disorder when most of the cartilage model of the femoral head and acetabulum has been transformed to bone, in assessing the structural relationship of the head and neck region to the acetabulum. Kim and Wenger have demonstrated in great detail the value of three-dimensional computed tomography in outlining and assessing the shape of the femoral head and its relationship to the acetabulum. This technique is being used increasingly in some centers in preoperative planning for surgical hip reconstructions. It not only defines the shape of the femoral head and acetabulum but also gives an appreciation of rotational abnormalities, which can then be corrected surgically (156, 157).

304

CHAPTER 4 ~

Legg--Calve--Perrhes Disease

F I G U R E 10 Radioactive labeling studies in the rabbit demonstrate the role of nutrition by diffusion from the synovial fluid supplying the articular cartilage, the epiphyseal cartilage, and the peripheral regions of the secondary ossification center in the immature animal. This study from an immature rabbit humeral head demonstrates uptake of tritiated proline in each of the three regions mentioned previously. The proline is incorporated into collagen and also serves as a marker of new bone deposition in the secondary center (arrow). The label was applied directly to the articular surface at open operation by dropping the radioactive solution onto the articular cartilage.

C. Subsequent Pathologic Changes Presenting as a Relative Decrease in Size of the Involved Secondary Ossification Center The first, subtle plain radiographic change of bony tissue in Legg-Perthes disease reflects the cessation of new bone deposition in the secondary ossification center on the involved side, whereas normal bone formation persists on the uninvolved side. The secondary ossification center on the involved side appears smaller by failing to grow relative to the continuing growth of the secondary center on the opposite side. The first plain radiographic change, therefore, in the secondary ossification center is its size rather than changes in either the density or shape. This finding is noted relatively infrequently because most patients do not present until later in the disease process. According to Edgren (68), it was first noted by Bergmann (15) and has been commented on by several other observers, including Blanchard (19), Bergstrand and Norman (16), Edgren (68), Katz (141, 145), Kemp and Boldero (149), and Axer and Schiller (6).

D. Nutrition of the Proximal Femoral Epiphysis and Its Bearing on Legg-CalvePerthes Diseases The epiphyses have a dual source of nutrition via diffusion from the synovial fluid and via the cartilage canals and intraosseous vessels. When avascularity of the secondary ossification center occurs, it can be expected that synovial diffusion persists to supply the articular and adjacent epiphyseal cartilage portions of the femoral head such that the cartilage mass of the head itself continues to grow (Fig. 10). At this stage the cartilage model of the femoral head remains normal in size, even though the secondary ossification center

ceases to increase in size because the intraosseous vascular supply is not present to allow the endochondral sequence to convert the central epiphyseal cartilage areas to bone. Because children rarely present for diagnosis at this stage, remaining either asymptomatic or minimally so, they continue to walk and run on the involved hip.

E. Gage Sign-Catterall Sign: Lateral-Proximal Neck Convexity-Lateral Epiphyseal Lysis Catterall pointed out that an early radiologic sign of repair in Perthes was a small osteoporotic segment forming a radiolucent "V" on the lateral side of the epiphysis (40-43). He considered this to be an "at risk" sign for relatively poor prognosis and attributed the original description to Gage, thus referring to the finding as Gage's sign. A review of Gage's original article (84), as clarified by Schlessinger and Crider (242), shows that Gage described a "convexity of the upper border of the fight (affected) femur in comparison with the left (normal) which is practically straight." He noted the same convexity to be present in films and illustrations of over 30 cases of Perthes disease that had been examined, whereas in over 250 films of other disorders the finding was not seen. He felt that the curved upper border was related to the well-recognized "thickening" of the neck in the well-established case and that it represented the first and perhaps earliest sign of a developing Perthes disorder. Gage's sign is shown in Figs. 11A and 11 C, whereas the lateral epiphyseal indentation, referred to by some as Catterall's sign, is shown in Figs. 11B and 11C.

F. Subchondral Fracture and Crescent Sign Bone tissue normally responds to stress by repairing any microdamage and strengthening itself as osteoblasts synthe-

SECTION VII ~ Pathoanatomic Changes

305

FIGURE 11 Radiologicsigns relating to lysis at the lateral proximal part of the neck (Gage's sign) and lateral epiphyseal lysis (Catterall's sign) are shown. (A) Gage's sign is illustrated in this figure taken with permission from his paper (84). The rounded area of the superolateral aspect of the neck (arrow) on the right is the abnormal finding. The left proximal femur is normal. (B) Illustration showing the lateral epiphysealradiolucency (arrow)interpretedby Catterall as an early "at risk" sign. [Modifiedfrom Schlesingerand Crider (242), J. Pediatr. Orthop. 8:201-202, 9 1988 LippincottWilliams & Wilkins, with permission.] (C) Radiograph shows a Gage sign (solid arrow) and a Catterall sign (open arrow).

size osteoid, which quickly calcifies to become bone. As the secondary ossification center is necrotic at this stage of disease evolution, with its osteocyte lacunae empty and the marrow cavity filled with necrotic debris rather than hematopoietic and osteoprogenitor cells, the bone is unable to respond to any stresses or microfractures it undergoes in normal fashion. Microfractures thus occur within the necrotic secondary ossification center and collapse of the subchondral bone follows. The cumulative minute injuries lead to macroscopic collapse in the absence of the usually present and immediately instituted repair process. A space thus is formed in the subchondral bone region that is due to the subchondral fracture through stressed necrotic bone and the localized collapse of the necrotic trabeculae. The overlying articular cartilage remains alive due to its separate source of nutrition from the unaffected synovial fluid. Initially it springs back to maintain the normal shape of the articular surface with a thin rim of subchondral bone remaining with the articular surface. A definitive relatively early plain radiographic sign of Legg-Calve-Perthes disease is this subchondral lucency, which is referred to as the crescent sign (Figs. 12i-12iv) (204). This is best demonstrated at the anterosuperior region of the femoral head on the frog lateral radiograph. The plane and extent of the subchondral fracture in relation to the radiographic projection partially determine whether the crescent sign will be demonstrated. Approximately 50% of initial biplanar radiographs in patients with Legg-Calve-Perthes disease demonstrate this sign. The extent of the radiolucent region is felt by some to indicate the amount of femoral head necrosis because the radiolucency is related to the adjacent necrotic bone. One would suspect that, if greater efforts were made to demonstrate the phenomenon using more radiographic projections at slightly differing angles of obliquity, it would be documented more frequently. Arthrographic and magnetic resonance imaging studies of

the femoral head at this stage usually show the articular and epiphyseal cartilage to be intact and of normal shape with the pathoanatomic changes still limited to the secondary ossification center. As noted previously, the size of the center will have remained the same over a several-week to -month period in the absence of bone growth and, thus, is smaller relative to the opposite side, which continues its normal growth. The histopathogenesis of the subchondral radiolucency or crescent sign involves: (1) necrosis of subchondral bone, (2) fracture through the necrotic bone, (3) collapse of the necrotic microtrabeculae with continued stress on the hip, (4) springing back to normal position of the articular cartilage with its thin attached rim of subchondral bone, and (5) increase in the extent of the subchondral radiolucency by early invasion of vascularized repair tissue, which allows osteoclastic resorption of the necrotic trabeculae and proliferation of radiolucent tissues of a fibrous, fibrocartilaginous, and early osteogenic nature. Ferguson felt that an important element in lateral extrusion of the femoral head was the occurrence of a lateral epiphyseal fracture of a viable segment of the head. This particular lesion had not been described and was different from the widely recognized subchondral fracture within necrotic bone (76).

G. Increased Radiodensity of the Secondary Ossification Center As the method used to follow the progression of LeggCalve-Perthes disease historically has involved primarily two-dimensional plain radiographs of the hip, interpretation of the variable and subsequently changing radiodensity of the femoral head is of great importance. Increased radiodensity of secondary ossification center bone is shown in Figs. 13i-13iv. There are three ways in which the radiodensity

306

CHAPTER 4 ~ Leg~t--Calve--Perthes Disease

FIGURE 12 The subchondral fracture often can be identified on plain radiographs. It is seen best in the frog lateral view, although on occasion it also is evident on the anteroposterior projection. The radiologic finding is referred to as the crescent sign because of the shape of the radiolucent defect (solid arrow) that is present within the subchondral bone on the inner surface of the articular cartilage. The radiolucency is seen because the subchondral bone (open arrow) immediately adjacent to the periphery of the epiphyseal cartilage remains attached to the undersurface of the articular-epiphyseal cartilage and at least in the early stages springs back to its normal position after the fracture, allowing the radiolucent area to appear. Four examples are shown (A-D).

of the secondary ossification center of the femoral head involved in L e g g - C a l v e - P e r t h e s disease can increase. (1) Microfractures of the femoral head compress necrotic bony trabeculae into a smaller area, thus increasing their relative radiodensity. The compression is concentrated initially just below the lucent crescent or subchondral fracture. (2) When repair in the femoral head commences it is associated with a revascularization process, which causes new bone to be synthesized on persisting spicules of dead trabecular bone. The combination of new and old bone, both of which are calcified, leads to an increase in calcified tissue per unit area and the radiodensity increases. (3) The absence of a blood supply to the femoral head while the blood supply to the femoral neck remains normal allows normal resorption to occur in the neck but prevents any resorption from occurring in the head. Thus, the normal radiodensity in the neck persists, whereas the cycle is changed for the head. This relative change is not as important as (1) and (2).

H. Alternating Areas of Radiodensity and Radiolucency On occasion, one sees an almost uniformly radiodense ossification center in Legg-Perthes disease, but shortly thereafter alternating and irregularly positioned areas of radiodensity and radiolucency appear. This is referred to as the fragmentation stage (Figs. 14i-14iv). The presence of fragmentation radiographically is a sign that repair of the necrotic bone is occurring. Diminution of radiodensity is a clear sign that

revascularization of the secondary ossification center is occurring because it is only by the mediation of a renewed blood supply that osteoclastic resorptive cells remove the necrotic bone. The revascularization also provides a source of new bone, forming osteoblasts as well. The occurrence of new bone synthesis and resorption of necrotic bone are not uniform either throughout the femoral head or even within microregions of the secondary center, such that the plain radiographic appearance of the secondary ossification center is changing and variable for several months. The matter is complicated further by the extent of necrosis, partial or complete head involvement, and by the occurrence of single or multiple episodes of infarction. The mechanism of repair of necrotic cancellous bone generally is favorable to partial and occasionally complete maintenance of its structure. If all necrotic bone was to be resorbed prior to new bone formation, collapse of the surrounding articular and epiphyseal cartilage would be inevitable and extreme. New woven bone is deposited on necrotic lamellar bone trabeculae initially, however, and the latter then are resorbed gradually to allow for normal bone reconstitution. Lamellar bone then is synthesized on cores of woven bone, which itself is resorbed, in the final stages of repair. Kenzora et al. have demonstrated well the synthesis of new bone on necrotic cores of dead lamellar bone in an adult rabbit model of femoral head necrosis (153). In certain patients, resorption appears to predominate such that the secondary ossification center virtually disappears prior to the time when bone forming cells renew their synthetic activity. In other patients, the fragmentation appears to

SECTION VII ~ Pathoanatomic Changes

F I G U R E 13

307

Examples of increased radiodensity of secondary ossification center bone are shown in parts (A) through (D).

be much better controlled with small areas of radiodensity and radiolucency occurring in a speckled pattern throughout the secondary ossification center. Ultimately, all necrotic bone is resorbed and the density of the secondary ossification center becomes the same as that of the neck and the opposite normal side. The long-term prognosis is dependent primarily on whether the cartilage model of the head, involving the articular and epiphyseal cartilage regions, has maintained its sphericity during the stages of secondary ossification center necrosis, resorption, and reconstitution. The plain radiograph shows the density changes and the relative location of synthesis and resorption in the bone of the secondary ossification center. The ultimate determinant of the final result, however, is the size and shape of the femoral head and its structural relationship to the acetabulum. The pathologic cause in Legg-Perthes leads to a dissociation in the normal cartilage-bone growth of the femoral head, with the cartilage growth initially unaffected. Because the bone always reconstitutes in this disorder, it is actually the response of the cartilage model that is the prime deter-

minant of the final result. Bone eventually is synthesized within the cartilage model via the endochondral ossification sequence, but it is the shape of the overall cartilage model that dictates the bone deposition pattern.

I. Responses of the Cartilage Model of the Femoral Head 1. COXA MAGNA

Coxa magna refers to the situation in which the head is bigger than that on the normal unaffected side (Figs. 15i and 15ii). It is a common sequela in Legg-Calve-Perthes and is evidence of an imperfectly controlled, but somewhat hyperactive repair process. Coxa magna can only occur from an excess of repair because the cartilage model grows to more than its expected size. The articular and adjacent epiphyseal cartilage at the superior weight beating surface of the head continue to grow via diffusion of nutrients from the synovial fluid. The synovial fluid does not appear to be diminished by the pathoanatomic process; indeed, it would appear

308

CHAPTER 4 ~

Legg-Calve-Pe~hes Disease

F I G U R E 14 Altemating areas of radiodensity and radiolucency within the secondary ossification center represent the fragmentation stage. Radiodense areas represent regions of persisting old necrotic lamellar bone on which new repair woven bone has been deposited. Radiolucent areas imply vascular invasion, which allows for osteoclastic resorption of necrotic bone and its replacement by granulation tissue that has not yet synthesized mineralized bone. Radiographs from different patients range from early fragmentation (A) to progressively later repair stages (F).

F I G U R E 15 Coxa magna refers to the situation in which the affected head is bigger than that of the normal unaffected side. (A) The example shown is indicative of the fact that repair not only has occurred on the right but that it has been controlled imperfectly because the repair side is now larger than the normal side. (B) An additional example of coxa magna is shown.

SECTION VII 9 Pathoanatomic Changes

that the continuation of the normal synovial diffusion nutritional source is supplemented by the stimulation from the excessive blood supply brought into the hip region during the repair process to yield a greater mass of tissue than otherwise would have been present. There can be two possible types of coxa magna: one with no long-term negative consequences and the other with more serious long-term consequences. In the first situation, the growth of the acetabulum may correspond with that of the femoral head such that the femoral head-acetabular relationship remains congruent and does not lead to long-term problems. The coxa magna is associated with a still spherical head. Long-term problems arise when the growth of the femoral head outpaces that of the acetabulum and the head not only is large in relation to the acetabulum but also is misshapen such that it relates imperfectly to the acetabulum. The size and shape of the cartilage model of the femoral head can be assessed by arthrography and magnetic resonance imaging. 2. COXA PLANA Coxa plana refers to a flat femoral head. The term was used initially to describe the radiographic appearance of the bone of the secondary ossification center (Fig. 16). Many still use this term to refer to the Legg-Calve-Perthes entity. The dissociation between the continuing growth of the cartilage model of the head and the disordered bone formation of the secondary ossification center must be borne in mind in interpreting radiographic images. In the early stages of LeggPerthes disease, this dissociation frequently leads to a flattened secondary center (Fig. 16A) but a persisting spherical cartilage model of the head. It is essential to define what is indicated by use of the term. In a prognostic sense it is the shape of the cartilage model that is most important; a coxa plana with a flattened articular surface is far more serious than a coxa plana of the secondary center with the cartilage model remaining spherical. It is the shape of the cartilage model and, thus, of the articular surface that is the crucial determinant of the quality of the end result. The flatness of the superior regions of the secondary ossification center can be due to one or both of subchondral collapse of necrotic bone and deposition of new bone as part of the repair process in the lateral parts of the head in which vascularity is first restored or in which diffusion from the articular cartilage is achieved more easily. The more worrisome type of coxa plana is that that persists at the termination of the repair process in which the overlying articular cartilage has collapsed to conform to the shape of the deformed secondary ossification center (Fig. 16b). When an arthrographic study demonstrates lack of sphericity of the articular cartilage, then a true coxa plana with probable long-term deleterious consequences concerning osteoarthritis has occurred. 3. LATERAL SUBLUXATION OF THE FEMORAL HEAD Hip radiographs frequently demonstrate more lateral positioning of the femoral head (Figs. 17A-17B) than on the

309

normal side as early as the repair phase and often persisting in the residual phase of Legg-Calve-Perthes disease. In the repair phase, there are three possibilities for this radiographic appearance. (1) The presence of fluid or hypertrophy of the synovium in the deep medial part of the hip joint might push the head laterally and thus into what appears to be a subluxed position. (2) The cartilage model of the femoral head may remain well-seated in the acetabulum in its normal position, but ossification might occur preferentially in a more lateral position within the epiphyseal cartilage of the femoral head. Ossification medially and inferiorly might be inhibited, whereas renewed ossification with repair and perhaps in association with nutrition by synovial fluid diffusion is most likely to be present at the superolateral regions. One then would postulate that the cartilage model of the head continues to grow in size and maintain an appropriate relationship to the acetabulum, but the secondary ossification center itself would appear to shift laterally. Instead of ossification occurring uniformly in a hemispheric conformation, the vascularization and pressure phenomena would tend to favor repair bone formation at the superolateral aspect of the head and, thus, give the appearance of a lateral subluxation. (3) The entire cartilage head may be enlarging as part of the repair response, i.e., becoming a coxa magna, such that the secondary center appears to be positioned more laterally. The importance of arthrographic or magnetic resonance imaging to study the shape and size of the cartilage model of the femoral head in this situation is evident, as is the need for extremely cautious interpretation of such studies. For example, epiphyseal ossification or calcification occurring lateral to the outer bony margin of the acetabulum frequently is construed as occurring in association with so-called subluxation, but as noted earlier may simply indicate that the lateral portions of the continually enlarging femoral head epiphyseal cartilage are more susceptible to revascularization and new bone formation than the deeper and inferior regions. It is not certain, therefore, that lateral calcification as seen during the middle to late stages of the repair process is a poor prognostic sign. If a coxa magna with a spherical head is evolving, this lateral calcification will not be significant. If, however, a coxa plana is evolving, it will. 4. COXA VARA Because the site of pathologic change is in the femoral head while the developing greater trochanter is unaffected, it is not surprising that abnormal sequelae might ultimately diminish growth of the head-neck region, leading to a coxa vara deformity. The avascular phenomenon does not affect just the bone of the secondary ossification center but also the epiphyseal cartilage and the physeal cartilage because the blood supply to the physeal cartilage in terms of its growth potential comes from the epiphyseal side. In many cases of Legg-Perthes disease, radiographic studies show not only abnormalities of the secondary center bone but also irregularities in the growth plate cartilage. In some instances these

310

CHAPTER 4 ~

Legg--Calve--Perthes Disease

F I G U R E 16 Coxa plana refers to a flat femoral head. In part (A) a plain radiograph shows the flattened orientation of the bone at the secondary ossification center. In early descriptions of the disorder, the term coxa plana referred only to the secondary ossification center shape, because that was the only structure that could be visualized on the plain radiograph. If the cartilage model of the femoral head retains its sphericity, then the flattening of the secondary ossification center bone may represent only a transient phase with subsequent bone repair following the still spherical articular and epiphyseal cartilage model and leading to normal or close to normal reconstitution of the bone. The more serious variants of the coxa plana situation occur when the articular cartilage collapses such that the model is imperfect when repair occurs and at the termination of the repair phase at skeletal maturity a flattened femoral head and articular cartilage persist. This is shown in part (B).

F I G U R E 17 Lateral subluxation of the femoral head is a relatively worrisome sign in Legg-Perthes disorder. On occasion an early lateral subluxation will recede particularly in those instances in which it is due to hypertrophy of the synovium in the deepest medial part of the acetabulum. In general, however, later subluxation of the femoral head occurs in association with repair in which a larger cartilage model of the femoral head and the medial acetabulum is present compared with the normal side and in particular one in which the lateral acetabular hypertrophy does not occur to as great an extent. The area of main concern is medial (arrows); note the increased space between the medial bony acetabulum and medial neck.

SECTION VII ~ Pathoanatomic Changes

311

chanter, the femoral head and neck often are placed in a relative coxa valga position. This occurs because slowing and/or premature closure of the head-neck physis often are more marked laterally than centrally or medially. When this occurs continuing central and medial growth positions the head into a valgus relationship to both the acetabulum above and the neck and shaft below. In this sense, even though there is overgrowth of the trochanter, it is not strictly accurate to refer to the relationship as one of coxa vara. Great care must be taken, therefore, to indicate whether it is the trochanter-head relationship that is being described, which often in Legg-Perthes disease turns out to be a coxa vara deformation, or whether it is the acetabulum-head-neckshaft relationship that is being described, which often leads to a relative coxa valga positioning of the head-neck axis.

FIGURE 18 An exampleof coxa vara in Perthes disease due to relative overgrowthof the normalgreater trochanteris shown.The tip of the greater trochanter (arrow)is at the same level as the superiorsurface of the femoral head.

appear to correct, but in others permanent damage to the growth of the femoral neck occurs, leading to the coxa vara deformity (Fig. 18). As is the case with avascular necrosis complicating treatment of congenital hip dislocation, the avascular sequelae in relation to the physis can cross the entire width of the growth plate, leading to symmetrical shortening. On occasion, however, it is concentrated either medially, leading to coxa vara associated with a more vertical growth plate, or laterally, leading to an overall coxa vara with a tendency to lateral obliquity of the growth plate and a tilting of the head into valgus and out of the bony acetabulum. In addition, approximately 25% of patients who suffered from Legg-Perthes disease have premature cessation of longitudinal growth of the head-neck region toward the end of skeletal maturity, often several years after repair of the avascular necrosis apparently has occurred. This worsens any coxa vara and increases shortening. Coxa vara is a frequent residual finding in cases of LeggPerthes disease. There are two different ways, however, in which the position of the proximal femur has been described in Legg-Perthes disease. One way refers to the position of the femoral head solely in relation to the greater trochanter. Diminished growth and frequent premature fusion of the physis of the capital femoral epiphysis in association with continued normal growth of the greater trochanteric epiphysis will change the relationship of the tip of the trochanter to the most superior portion of the femoral head, such that the trochanter is at the level of or higher than the articular cartilage of the head. In this sense, therefore, a coxa vara deformity can be defined. When the head and neck alone, however, are related to the more distant femoral shaft, without attention being paid to the position of the greater tro-

J. Assessment of Cartilage Model of Proximal Femur Using Arthrography Plain radiographs clearly outline the shape of the bone of the epiphysis (the secondary ossification center) and adjacent neck as well as that of the acetabulum. The even more crucial cartilage surfaces, however, of the femoral head and acetabulum plus their relationship to each other cannot be assessed by plain radiographs, particularly when there are a few years of growth remaining. Arthrograms have been used for several decades to outline the cartilage contours of joints and can be highly informative in Legg-Perthes disease (Figs. 19A-19C). It is well-recognized that the articular surface of the proximal femur maintains its sphericity and appropriate relationship to the acetabulum in the early phases of the disorder, even though the secondary ossification center is abnormal. Jonsater (133), Goff (91), and Katz (143) all published arthrographic studies of hips in the early stages of Perthes and indicated that there was little to no collapse of the articular surface and that there was general preservation of its sphericity for a considerable period of time after bone changes of the secondary center were noted. Herzog performed arthrotomies on three patients with early LeggPerthes and reported that in all there was a well-rounded cartilaginous head despite radiographic findings of a flattened bony epiphysis (121). The arthrogram outlines the shape of the femoral head and can also enable determination of its size and relationship to the adjacent cartilage of the acetabulum. Coxa magna thus can be diagnosed as can flattening of the superior and lateral articular surface as determined by the outline of the dye, and in addition by pooling of the dye, which is not seen when normal sphericity has been maintained. An increase in the medial joint space also can be detected and some information can be gained as to whether there is incongruity of relationship to the acetabulum, at which time there is pooling of the dye medially, or whether the increase is due to coxa magna and increased thickness of the medial acetabular and/or medial femoral head cartilage, at which time sphericity would be maintained. By utilizing

312

CHAPTER 4 ~ Le~tg--Calve--Perthes Disease

F I G U R E 19 Arthrograms play a major role in defining the shape and size of the cartilage model of the femoral head in LeggPerthes disease and in showing its relationship to the acetabulum. It is these relationships of the cartilage that are far more important in assessing the long-term prognosis for the hip than the shape of the bony model of the secondary ossification center alone. Note the normal spherical outline of the cartilaginous femoral head even though the secondary ossification center is misshapen (A). In part (B) the cartilage model also is normal, whereas the secondary center is small and irregular. The normal triangular thistle sign is seen laterally with dye on either side of the fibrocartilaginous labrum. In part (C) the labrum is positioned normally. The cartilage model of the head remains normal even over the metaphyseal cyst.

multiple projections, the earliest changes in the shape of the cartilage model can be detected. Axer and Schiller showed that 7 of 8 hips in the initial stage of the disease had definite, even though mild, anterolateral flattening, which became even more noticeable in later stages of the disease (6). Katz presented his studies of arthrograms in 25 children with Legg-Perthes disease and found good correlation with data presented earlier by Jonsater (143). Jonsater, who also correlated his studies with plain radiographs and histologic biopsy specimens, found that in the initial stage of LeggPerthes the femoral head customarily kept its spherical shape, whereas slight alterations in shape occurred and then increasingly worsened during the repair phases. Katz also confirmed that, in the early initial stages of Legg-Perthes disease, the arthrogram revealed regularity of the outline of the femoral head with minimal or no alteration in spite of major structural changes in the bony epiphysis. The arthrograms were most valuable during the fragmentation and reparative stages because they showed whether the articular surface was remaining spherical or flattening to conform to the irregularity of the bony epiphysis. Once the final or definitive stages had been reached, the arthrogram was of less value because the enlarged bone mass of the secondary center and the close to mature articular cartilage had a basically parallel outline. The arthrogram also was shown to be valuable in assessing early stages of coxa magna. The acetabular articular cartilage also could be determined in relation to its shape and thickness, and early medial widening was noted. The arthrogram is most helpful if there is pooling of dye, indicating a concavity or depression particularly of the superolateral surface. It also is most helpful in assessing the relationship of the acetabulum and particularly its articular

and fibrocartilaginous labrum to the femoral head. Because enlargement and lateral extrusion of the femoral head occur and bone formation also occurs laterally, this important relationship is best revealed by the arthrogram. The lateral acetabular margin often is shown to have depressions and indentations into the femoral head, especially at the site of the labrum. Gallagher et al. also defined structural abnormalities that were clarified by the arthrogram in relation to the plain radiograph (86). Arthrography also has been found to be a more sensitive indicator of early and relatively mild lateral subluxation. Moberg et al. compared plain radiographs and arthrograms in 76 hips with Legg-Perthes disease (195). Lateral subluxation was documented in 43 of 76 hips both by plain radiographs and by arthrography. In the remaining 33 hips, however, plain radiographs could not verify any subluxation, but arthrography showed subluxation, in 32 of the 33. Calculation of the acetabulum head index on plain radiographs involved assessment of the horizontal distance from the innermost surface of the ossified epiphysis or secondary center medially to the outermost margin of the bony acetabulum. The arthrogram, however, allowed the index to be determined from the innermost surface of the cartilaginous head to the outermost margin of the cartilaginous acetabulum. Because the femoral head-acetabular head relationship in early Perthes is determined by the cartilage models, the arthrogram is slightly more sensitive than a plain radiograph.

K. Responses of the Physis The physis almost always is affected in Legg-Perthes disease because its blood supply on the epiphyseal side is

SECTION VII 9 Pathoanatomic Changes

derived from the same vessels as that of the secondary ossification center. In cases that heal uneventfully, it continues to function or resumes its normal function relatively early on. In many instances, however, it is damaged focally or completely, and disordered function follows. The earliest change is a widening of the physis. Obliquity then can be noted in some, a finding associated with asymmetric early premature closure. If the physis tilts medially, then medial closure predominates and a coxa vara is occurring; if it tilts laterally, then lateral closure leads to a coxa valga. If closure occurs centrally, then there is no obliquity of the physis, the head-neck-shaft angle is unchanged, but flattening of the physis is minimized. The physeal changes often occur late, well into the repair phase, and often they are not manifested until the final year or two of skeletal growth even if full repair appears to have been accomplished in the secondary ossification center years previously. Edgren points out that fusion tends to occur first centrally. In those cases in which premature growth plate closure occurred (1 year or more prior to the opposite normal side), the femoral head was spherical in only 5, elliptical (oval) in 13, and irregular in 24, whereas with closure at the normal time 18 were spherical, none elliptical, and only 2 irregular. Severe metaphyseal changes almost always were associated with premature physeal closure, whereas slight metaphyseal changes correlated with normal closure.

L. Sagging Rope Sign Many patients with a relatively severe Legg-Perthes disorder demonstrate a characteristic radiographic finding well into the repair phase in which there is an opaque line in the femoral neck. This was described in detail by Apley and Wientroub and referred to as the "sagging rope sign" (3). The sign consists of a thin opaque line in the upper femoral metaphysis, which is curvilinear with the concavity toward the head and upper neck region (Fig. 20). In the anteroposterior view it extends laterally from the inferior border of the neck in a curved fashion to the medial region of the neck, generally at the head-neck junction. The line always appears to be contained within the neck regardless of the radiologic view. Apley and Wientroub were able to relate it to a relatively severe disorder in which negative growth sequelae had occurred, such that it was almost always seen with a short and thick femoral neck, a flattened secondary ossification center, premature closure of the growth plate, and often overgrowth of the greater trochanter. All of these findings were consistent with relatively severe damage in a Legg-Perthes disorder. They also noted its presence in other hip disorders with comparable pathology, including multiple epiphyseal dysplasia and avascular necrosis following treatment of a congenital hip dislocation. The sign, when present, was indicative of a Perthes disorder in which significant negative growth sequelae had occurred, and if there was still growth

313

F I G U R E 20 The sagging rope sign (arrows) is indicative of considerable growth damage.

remaining, these clearly could worsen with time if not appreciated. Apley and Wientroub felt that the radiographic density was due to increased bone tissue within the femoral neck and that it represented damage to the growth plate with a marked metaphyseal reaction reminiscent of a Harris growth arrest line, to which it was felt to be analogous. Catterall also commented on the sagging rope sign but felt that the line represented the anterior margin of the femoral head as it angulated acutely to join the metaphyseal area, such that the cortical bone of the periphery of the head and the adjacent metaphyseal bone were represented as a dense line on the anteroposterior view (43). Clarke et al. confirmed the sagging rope sign in severe cases of Perthes, noting it as a late finding in 13 of 78 patients, 7 of whom had fair results and 6 poor. They demonstrated that the radiodense line represented the anterolateral edge of a very deformed femoral head. A three-dimensional computerized tomographic reconstruction of the hip region in a case of Legg-Perthes disease by Kim and Wenger also showed the inferolateral margin of the flattened femoral head to coincide exactly with the shape and site of the "sagging rope sign" as noted on the plain radiographs (155).

M. Responses of the Femoral Neck (Metaphysis) 1. WIDENEDFEMORALNECK A widened femoral neck, often referred to simply as the metaphysis in the Perthes literature, frequently is seen in the

314

CHAPTER 4 ~ Leyg--Calve--Perthes Disease

F I G U R E 21 A widened femoral neck (arrows) almost always accompanies the repair response in Perthes disorder.

repair phase of Legg-Perthes disease (Fig. 21). Studies by Robichon et al. (233) have shown that this is an absolute increase in width and not an appearance just relative to a smaller femoral head. The femoral head normally is much wider than the adjacent femoral neck. The femoral neck periosteum thus serves as an osteoclastic cellular layer as well as providing support via its outer fibrous layer. In LeggPerthes disease, however, the repair phase is characterized by a generalized increase in vascularity to the hip region. The growth in width of the cartilage model of the femoral head actually is increased (coxa magna) in the repair phase, and the blood supply to the neck is increased because the vessels supplying the head pass along the surface of the neck. This phenomenon appears to switch the inner cell layer of the femoral neck periosteum to an osteoblastic phase. New bone formation is enhanced in association with the increased blood supply passing up the femoral neck and eventually entering the head as part of the revascularization repair process. During this repair phase, therefore, the widened femoral neck develops due to increased appositional bone formation from the inner layer of the periosteum and due to endochondral growth in relation to the widened cartilage model of the femoral head and physis as it develops a coxa magna. Although the widened femoral neck is an associated radiologic sign and provides interesting biological considerations, it is of no clinical concern either as a prognostic indicator or as an eventual functional problem. 2. METAPHYsEAL BONE IRREGULARITY AT F E M O R A L NECK-PHYsEAL JUNCTION The usual linear appearance of the femoral neck metaphyseal bone adjacent to the physeal cartilage often is irregular during the course of the disorder. This actually would appear to represent a physeal growth irregularity, leading to a less

F I G U R E 22 Metaphyseal radiolucencies referred to as cysts are a common radiographic finding in Legg-Perthes disease. They are present immediately adjacent to the physis. On occasion they transverse the width of the neck (arrows).

than uniform scaffold on which metaphyseal bone could be deposited. 3. FEMORAL N E C K (METAPHYSEAL) e

RADIOLUCENCIES, SOMETIMES REFERRED TO AS CYSTS

The presence of radiolucent regions in the femoral neck has long interested those studying Legg-Perthes disease. These are seen on plain radiographs and often are considerable in size (Figs. 22 and 23A-23C). They almost invariably are adjacent to the physis and have two general shapes. On occasion, linear radiolucencies are seen that may or may not reach the lateral cortical border of the neck. More often, the radiolucencies are circular to oval in outline. Smith et al. defined the radiolucent changes as being either transverse zones or cystic and further subdivided the cystic changes as being rounded radiotranslucent areas with a well-defined edge most often, associated with marked rarefaction, or surrounded by markedly sclerotic changes less frequently (252). They noted radiographic abnormalities in the metaphysis in 120 of 134 affected hips (90%). The metaphyseal changes generally were seen simultaneous with epiphyseal changes or after the epiphyseal changes, but in no instances did the metaphyseal changes precede those in the epiphysis. The metaphyseal cystic changes always were immediately underneath changes in the epiphysis. The site of metaphyseal change reflected the fact that physeal involvement tended to predominate laterally in those instances in which it was not complete. They defined the area of greatest metaphyseal change as being central in 20 of 72 instances, central lateral in 20, and lateral in 10 with 18 of these 30 involving extension to the lateral cortex; involvement of the entire transverse diameter of the metaphysis was

SECTION VII ~ Pathoanatomic Changes

315

FIGURE 23 Metaphysealcysts usually are localized (A, B). Sclerotic margins simplyindicate a lengthierpresence and no recent size changes.

seen in 11, 7 central medially, and 4 medially only. Efforts were made to correlate the metaphyseal changes with the end result. As the authors noted, this was difficult because categorization of the end result in Perthes overall is difficult. Certain observations seem valid, however. The absence of cystic change within the metaphysis indicated a favorable prognosis. The site of the cystic change could not be correlated directly with the result except in those cases in which extension of the cystic process to involve the lateral metaphyseal or neck cortex was observed because 13 of 17 such cases were judged to be in the poorest category. They commented that cystic changes within the metaphysis appeared to reflect the general severity of the disease process and had some prognostic significance as indicated earlier, even though it was a phenomenon secondary to the primary epiphyseal insult. Ellis pointed out that metaphyseal lesions had been appreciated from the earliest descriptions of Perthes himself (69). Ellis also noted the metaphyseal radiolucency as either an ill-defined band stretching across the metaphysis or a more focal single or multilocular cystic area. He summarized his work in 1984, pointing out that where metaphyseal radiolucency involved the lateral border of the neck or metaphysis the percentage of collapse of the lateral pillar was higher and the result poorer. Ellis used the metaphyseal lesion as a guide to containment therapy. If there was no metaphyseal lesion, he felt it was safe to bear weight with rest recommended only for irritable hip symptoms. He was not concerned with neck or metaphyseal expansion in and of itself, but rather with expansion associated with lateral cortex radiolucency. The absence of expansion simply indicated that the disease was at an early stage or was mild in degree, whereas expansion indicated that healing was well underway. If the lateral neck-metaphyseal cortex was radiolucent or decalcified in his term by the metaphyseal lesion, then it was essential to protect the hip by containment, whereas if the radiolucency did not reach the lateral cortex, it was safe for the patient to

bear weight without containment as long as the hip was not irritable. He indicated that "lateral cortical decalcification of the metaphysis indicates a vulnerable hip which must be contained until the lesion has recalcified; otherwise, weight bearing will not affect the result." In those hips in which there was no evidence of decalcification of the lateral metaphyseal cortex regardless of the appearance of changes more medially, all developed congruous femoral heads irrespective of treatment. He thus pointed out the importance of the maintenance of the lateral pillar, with good prognosis occurring if it had never been radiolucent or decalcified or if it had regained radiodensity as a consequence of repair. Mindell and Sherman commented on metaphyseal cystic changes and noted that they generally appeared late after epiphyseal involvement, that when present they were situated opposite the region of epiphyseal involvement, and that they tended to heal earlier than the epiphyseal lesions (192). They indicated that, where the lesions were large, the incidence of ultimately poor results was greater. Catterall described metaphyseal lesions of increased dimension in those children with larger areas of epiphyseal involvement (43). Katz and Siffert in a large series of 337 patients noted metaphyseal lesions only in 31% (145). They felt there was twice the likelihood of a poor result when metaphyseal cysts were seen. Although most observers placed them in the metaphysis of the neck where they appear to be on plain X rays, some concern was expressed that they might indeed represent radiolucent regions in the epiphysis itself and merely appeared on plain radiographs in the metaphysis due to the geometry of the head-neck, in which the peripheral portion of the head on an anteroposterior radiographic projection essentially overlays the physis and the neck due to a mushroom-shaped conformation of the head. The use of MR imaging, however, demonstrated the head, physis, and neck regions by images only a few millimeters thick and showed the cystic regions to be clearly within the neck.

316

CHAPTER 4 ~

Le~tg-Calve-Perthes Disease

Most histopathologic studies show the cysts to be persisting, nonossifying physeal cartilage. They appear to be related to abnormalities in bone formation presumably in relation to the abnormal cartilage synthesized by the physeal regions. They also may relate to abnormal stresses in a child who has both Perthes disease on the one hand and is undergoing weight relieving treatment on the other.

N. Femoral Neck Anteversion Femoral neck anteversion has been studied in Legg-Perthes disease, much as it has in other hip disorders. No major changes in anteversion have been noted. Fabry performed a detailed study in which he noted no change in anteversion compared with patients without Perthes disease or with the contralateral hip. In 864 anteversion studies on normal children from 1 to 16 years of age, he noted the average anteversion at 24.14 ~ whereas in 241 anteversion studies in 160 patients with Legg-Perthes disease the average anteversion was 24.96 ~ A total of 118 studies carried out on the unaffected hips of 118 unilateral cases showed an average anteversion of 25.12 ~ (73). Other authors have reported usually slight increases in anteversion in Perthes disease (155). Kim and Wenger also documented increased anatomic anteversion using classic criteria for measurement in their studies of Perthes disease with three-dimensional computerized tomography (155). They subsequently went on to develop the concept of "functional retroversion" of the femoral head in LeggPerthes disease based on their interpretation of weight bearing as determined from the three-dimensional CT reconstructions.

O. Responses of the Greater Trochanter The greater trochanter is not affected by the necrosis occurring within the femoral capital epiphysis and continues its growth in a normal fashion. This growth occurs by elongation at the greater trochanter epiphysis and by appositional growth at the tip and lateral side walls. When femoral head and neck growth lags, there is a relative overgrowth of the trochanter, the tip of which may come to lie above, and often well above, the superior surface of the femoral head (Fig. 24). This occurrence leads to a diminution of hip abduction and a Trendelenburg gait due to the relative laxity of the attached gluteus medius and minimus muscles.

P. Responses of the Acetabulum In many patients with Perthes disease acetabular development is affected but most have considered the changes to be secondary to the abnormal shape of the femoral head (Figs. 25A25C). Perthes occurs during the active growing years but changes in shape of the femoral head occur slowly over a several-month to few-year period, such that the acetabulum can respond by modeling itself to the shape of the head. If the patient has a fairly rapid response to the disorder and heals

FIGURE 24 The greater trochanter grows normally in Legg-Perthes disease. It is important to assess the position of the tip of the greater trochanter in relation to the articular surface of the femur as this relationship serves as an index of femoral head-neck physeal growth. In this late-stage radiograph the tip of the trochanter lies proximalto the femoral head articular surface.

the condition with a round femoral head and very little coxa magna deformity, the acetabulum maintains its normal shape. In no instance does the acetabular change appear worse than that of the corresponding femoral head, a finding that has been taken to indicate the entirely secondary and reactive nature of the acetabular changes. In those patients with a few years of growth remaining who develop a large femoral head and who also lose sphericity of the head, the tendency is for an acetabular dysplasia to develop. In the most favorable instances, however, the shape of the acetabulum corresponds to that of the femoral head allowing for a continuing good relationship, which is referred to as a congruent fit. In many instances, however, the acetabulum, though dysplastic, does not expand laterally and the coxa magna deformity leaves the head relatively uncovered in relation to the normal side. Those cases of Perthes disease that have the poorest longterm prognosis tend to occur in patients 11 years of age and older when the disorder is first seen. One of the reasons for the poor long-term prognosis is the fact that by this age the acetabulum has reached a stage of development in which there is insufficient growth remaining to effectively mold it into a shape similar to that of the femoral head. In the older patients, the acetabulum remains at its normal size, depth, and shape, and the changes in shape of the femoral head develop without any meaningful change in acetabular structure. These responses of the acetabulum are quite important because many patients with large and misshapen femoral heads do well in long-term clinical studies because the acetabulum also has changed its shape to a relatively congruent stress-free position.

SECTION VII ~ Pathoanatomic Changes

317

FIGURE 25 The shape of the acetabulum must be reviewed carefully to assess the overall response to Perthes disorder. The younger the patient, the more likely the acetabulumwill conformto the shape of the femoral head. The older the patient at the time of disease occurrence, the less likely the acetabulum will respond in a congruent fashion. A normal acetabulum at skeletal maturity following treatment of the right hip with an abductionbrace (A), a normalacetabulumfollowingproximalfemoralvarus osteotomy(B), and a misshapen acetabulumfollowingabductionbrace treatment (C) are shown.

Information has accumulated slowly, however, to indicate that acetabular changes begin very early in Perthes disease and that the specific growth profile of the acetabulum is such that it is not totally passive in relation to the femoral head in Perthes disease. Halkier noted the altered teardrop shape on the involved side in Perthes disease (he referred to this as the "tear-shaped phenomenon") (98). The teardrop is that part of the pelvis that corresponds medially to the inner surface of the pelvis, laterally to the medial floor of the acetabulum, above to the triradiate cartilage, and below to the margin of the limbus. He noted that in Legg-Perthes disease there was almost constantly a relative increase in width of the so-called tear shape on the affected side and that it was seen both early and throughout the disorder. He went into great detail to dispel the fact that it was projectional in nature and thus felt that it was a genuine bony change. The very early appearance of the increase in width also has been noted by others, but so far no specific prognostic features have been associated with it. Joseph has directed attention to acetabular changes in Legg-Perthes disease and also has noted subtle abnormalities in growth beginning from the earliest phases of the Perthes disorder (134). The characteristic acetabular changes that he identified and quantified involved osteoporosis of the roof, irregularity of acetabular contours, premature fusion of the triradiate cartilage, changes in thickness of the articular cartilage, and changes in overall dimensions of the acetabulum. Bone scans were performed on 27 children with 31 hips involved with Perthes disease. In every unilateral case the radioisotope uptake was higher on the affected side in at least two of the three regions of interest, which were the acetabular roof, the area adjacent to the triradiate cartilage, and the medial acetabular wall. This was postulated to be due to the

synovitis, which affects the entire hip region in Perthes disease, but nevertheless it showed the altered growth phenomena to be associated not just with the femoral head itself. The acetabulum was shown to be affected early in the course of the disease and it was shown that there was a significant increase in its metabolic activity, which, due to the age of the patients, clearly resulted in altered growth. Arthrograms had long demonstrated that both the acetabular and femoral articular cartilages were thickened in Perthes disease. Even aside from abnormal shaping relationships to the femoral head, acetabular growth changes were due to (1) thickening of the acetabular cartilage throughout, (2) increased appositional growth at the triradiate cartilage leading to bicompartmentalization and widening of the acetabulum, and (3) premature fusion of the triradiate cartilage, leading to subsequent diminished growth of the acetabulum particularly in its lateral spread. Joseph felt that premature triradiate fusion occurred in 39 of 135 hips (29%). There has been relatively late recognition of the facts that acetabular remodeling can continue to skeletal maturity and that a considerable amount of improvement can follow after primary healing of the femoral head. A study by Kamegaya et al. found that acetabular remodeling did not correlate either with the age at primary healing or with the sphericity of the femoral head at primary healing (138). They documented considerable acetabula-head relationship improvement between the time of primary healing and skeletal maturity. As a consequence of this observation, they often deferred decisions on acetabuloplasty until skeletal maturity. Kamegaya et al. felt that changes in subluxation of the hip during the active phase of the disease were the most important factors in predicting final acetabular coverage at maturity and that it was the position of the femoral head rather

318

CHAPTER 4 9 Ley~t--Caive--Perthes Disease

than its shape that was the most important factor in acetabular growth and remodeling. They classified the amount of subluxation by the teardrop distance. Yngve and Roberts documented acetabular hypertrophy in Perthes disease in a quantitative fashion (288). Most of their study was performed during the active phase of the disease, and thus they concluded that acetabular and femoral head overgrowth occurred in a parallel fashion. They pointed out well the concept of acetabular bicompartmentalization, which led to a larger acetabulum that, however, was not smooth in contours. The medial part of the acetabular enlargement was due to cartilage hypertrophy, and the lateral shaping was due to the position of the femoral head in relation to the more lateral part of the acetabulum. They felt that bicompartmentalization was a poor prognostic sign because the spherical or ovoid femoral head did not have a congruent fit to the acetabulum throughout the entire range of motion. Medial acetabular cartilage thickening previously had been diagnosed primarily on arthrographic studies by Katz in which the majority of cases (33 of 37) had arthrographically increased femoral head sizes, with 29 of 36 also showing medial acetabular cartilage thickening. The arthrographic studies of Axer and Schiller (6) found that lateral displacement of the femoral head in relation to the acetabulum was not due to swelling of medial acetabular soft tissues causing a subluxation, but rather to asymmetric cartilage growth with lateral placement of the head due to thickening of both femoral head and medial acetabular cartilage. Acetabular changes during growth in Perthes disease were studied radiographically in 62 cases by Cahuzach et al. (32). They felt that the length of the acetabular roof did not change with time, but the opening angle and the diameter of the acetabulum increased beginning in the earlier stage of the disease in half of the cases. The length of the acetabular roof was defined as the distance between the external osseus edge of the roof and the superior and external edges of the Y growth plate. The diameter of the acetabulum was measured using the Mose circles and the opening angle of the acetabulum, which was a combination of the acetabular index at the superior aspects and inferiorly with what they referred to as the angle of depth, which is an angle between Hilgenreiner's line and the line joining the superior and external edges of the Y growth plate to the most inferior point of the pelvic teardrop. Cahuzach et al. felt that the end results were better when the opening angle did not change.

Q. Remodeling in the Residual Phase of the Disease between the Termination of Healing and Skeletal Maturity Some form of remodeling can continue at least until skeletal maturity and to a much less extent even afterward (118, 228, 255). Efforts have been made to assess the remodeling potential in the residual or final stage between the time that virtually full repair of the femoral head ossification center

has occurred and skeletal maturity. In those patients having the disorder before the age of 6 or 7, it is possible to determine when repair is complete for the most part, but in those having the disorder at 9-10 years of age or later, one of the reasons for the generally poor result is that repair is relatively slow and may continue until skeletal maturity. Katz attempted to determine the late modeling changes in LeggPerthes disease by superimposing sequential radiographs of the proximal femur on each other in the time prior to skeletal maturity (147). For the most part, there was no major change in categorization, which in his paper involved assessment using the Catterall groupings. Occasionally, there was evidence of improvement by one grade, although this tended to be counterbalanced by worsening by one grade in others. The one particular area, however, structural proximal femur changes occurred related to premature closure of the proximal femoral capital epiphysis with the continuing growth of the greater trochanteric epiphysis. This premature closure served to increase the lower extremity length discrepancy when present. In addition, the location of the closure in the transverse axis of the physis defined whether angular deformity of the head and neck would occur. If the closure was uniform or concentrated centrally, there was no change in the femoral head and neck axis in relation to the long axis of the femur. When closure predominantly was medial, the coxa vara would increase, and when the closure was lateral, there would be a relative valgus position of the head and neck in relation to the shaft although still in association with the greater trochanteric overgrowth. Allen performed a similar analysis of shaping changes in the proximal femur with Legg-Perthes disease utilizing computer graphic analysis (1). His study demonstrates the value of tracking changes in shape with time in a far more systematic manner than can be obtained either from measuring indices, which are relatively crude, or from merely observing shape changes qualitatively. The method also shows how it is the cartilage model changes that represent the significant ultimate determinants of the final result in Legg-Perthes disease, whereas the bony changes in the secondary ossification center serve as relatively crude and only secondary or derivative changes. Harry and Gross also have shown the value of computerized analysis of digitized tracings of proximal femoral structure in patients of varying time periods in an effort to improve quantitative data (110). Herring et al. noted that the shape of the femoral head continued to change for 34 years after apparent reossification (118). The heads became progressively rounder in 36% (49 of 136 hips) and progressively flatter in 11% (15 of 136 hips). The final outcome is not determined, therefore, until skeletal maturity. Growth in the residual phases of the disorder right up to skeletal maturity remains important clinically. The changes that occur can change the result for the better due to femoral head-acetabular remodeling or for the worse due to premature physeal closure, but are amenable to management if caught early.

SECTION VII ~ Pathoanatomic Changes

R. Imperfect Healing of Legg-Calve-Perthes with Persistence of an Osteochondritis Dissecans Lesion at Skeletal Maturity On occasion, there is incomplete bone repair at the superior surface of the femoral head, leaving an osteochondritis dissecans-like lesion persisting at skeletal maturity (Fig. 48B). Generally there is a localized small area of increased density visible in the superior weight bearing part of the epiphysis, usually in the subchondral region. There is a surrounding area of radiolucency, but in most reported cases there is little to no discomfort. Although the occurrence is infrequent, many reports have appeared including those by Ratliff (226), Freund (81), Brailsford (26), Evans (70), Freehafer (80), Hallel and Salvati (99), Osterman and Lindholm (207), Morris and McKibbin (196), and Bowen et al. (25). Bowen et al. found an osteochondritis dissecans lesion in 3% of a large number of patients (14 of 465) with Perthes disease. For those who were symptomatic, arthroscopic surgery was used to remove the loose fragment.

S. Hinge Abduction: Imperfect Healing with a Flattened Femoral Head and a Superolateral Prominence Impeding Smooth Abduction This particular variant of imperfect healing in Legg-Perthes disease was recognized more recently and has been referred to as hinge abduction. In this situation there is flattening of the cartilage model of the femoral head and a superolateral prominence underneath and lateral to the edge of the acetabulum, which impedes free abduction. This finding initially was described by Grossbard (96) and Catterall (42) (Fig. 3B). Hinge abduction is present when the outer part of the misshapen femoral head hinges on the lateral lip of the acetabulum as the femur is abducted. Kruse et al. considered widening of the medial cartilage space of more than 2 mm on abduction with associated narrowing superolaterally as evidence of hinge abduction (162). In those who develop this disorder and continue to walk on the affected hip, degeneration is much quicker than would otherwise be the case. Negative effects of hinge abduction, as defined by Reinker, include compression and flattening of the femoral head at its superolateral aspect, flattening of the lateral edge of the adjacent acetabulum, and imperfect position of the labrum. The disorder is felt to occur during the middle to late fragmentation phase when a central depression in the head areais often seen. In turn this leads to continuation of lateral subluxation with stretching of the medial and inferior joint capsule and further compression of the femoral head laterally. Much of this is associated with collapse of the lateral pillar. There is increasing recognition that this is a fairy early development in many patients with Legg-Perthes disease. In a study of 106 cases, the disorder was suspected in 26 and with arthrographic and specific radiographic studies hinge abduction was detected in 19 of these (229). The finding is a poor

319

Causes of Femoral Shortening in Legg-Perthes Disease 1. The Primary Disorder Necrosis and collapse of 2 ~ ossification center bone/marrow Growth damage to epiphyseal and physeal cartilage

\ 2. Varus Osteotomy Especially with varus correction > 15 ~ and surgery after 8 years of age

3. Premature Physeal Closure Occurs in 25% +

FIGURE 26 The three possible causes of femoral shorteningin LeggPerthes disease.

prognostic sign and treatment is mandated, with the nature of the treatment being dependent on the stage of healing phase at which the hinging is detected. Reinker stresses that hinge abduction is an important indicator of a poor outcome if left untreated and that attempts to recognize and treat it are important even while the disorder is evolving and the head is in the healing phase but has not completed it (229).

T. Femoral Shortening as a Sequel to Legg-Perthes Disease Shortening of the involved femur is a frequent occurrence in Legg-Calve-Perthes disease. There are three possible causes of any shortening (Fig. 26). 1. PHYSEAL AVASCULARITY Shortening at disease termination, though not always a result, is not unexpected because the cause of the condition is an avascular event or series of events of the proximal capital femoral epiphysis. The epiphyseal blood supply is responsible not only for supplying the bone of the developing secondary ossification center but also for providing the stimulus to growth of the head-neck growth plate. Variable

320

CHAPTER 4 ~ Legg--Calve--Perthes Disease

patterns of length deformity are seen (246). In the most benign situation, complete reconstitution of the femoral head and neck region occurs with no long-term sequelae. Most patients develop a shortening in association with the Perthes disorder, although in many this will correct itself either partially or fully as part of the repair phase. 2. PREMATURE PHYSEAL CLOSURE Edgren documented a shortened femur in Legg-Perthes disease in almost all patients at the time of skeletal maturity with a mean discrepancy of 1.6 cm and a range of 0.2-3.5 cm. He also noted tibial shortening in some averaging 0.4 cm when present and ranging from 0.2 to 1.2 cm. Edgren reported an incidence of 12.8% premature physeal closure in his large series of 326 hips (68). Some interference with physeal growth was noted by Keret et al. in 90% of patients; in 25% of patients there also was premature closure of the proximal femoral growth plate as a delayed sequel, which further increased any discrepancy during the final year or two of skeletal growth (154). We also documented premature physeal closure as a delayed sequel, leading to the type IV lower extremity length discrepancy pattern (see Section VIII below) (246). Thompson and Westin noted 49 of 192 hips (25.5%) with premature physeal closure (264). The closure in their patients began laterally but proceeded to involve the entire physis, predisposing one to a short valgus femoral neck and a prominent greater trochanter. Barnes reported on 22 cases of premature physeal closure in Perthes disease (13). The shortening in 19 unilateral cases, most of whom had previous varus osteotomies, was a mean of 1.92 cm (range = 0.5-3.0 cm) and was present in all. Six had shortening of 2.5 cm or more and 13 of 2.0 cm or more. Premature physeal closure of the capital femoral physis was commented on by Bowen et al., who noted a 23% premature closure rate in 430 hips with Perthes disease (24). There were two patterns of femoral closure, the most common being central closure beginning at the center of the physis and progressing to the periphery usually in symmetric fashion. The neck shaft angle remained normal but overgrowth of the greater trochanter allowed for an overall coxa vara appearance. The second common pattern found closure concentrated laterally, which led to asymmetry as the head was tilted laterally in relation to the acetabulum as central and medial physeal growth continued. In these patients, the femoral head tilted laterally and tended to develop an ovoid shape with the adjacent lateral margin of the acetabulum deficient. In the study by Keret et al., 80 patients with unilateral coxa plana who had been treated conservatively were followed to skeletal maturity (154). Physeal involvement was inferred by the radiographic presence of premature physeal closure, overgrowth of the greater trochanter, change in physeal shape, lateral protrusion of the capital nucleus, and medial bowing of the femoral neck. They noted premature physeal closure in 25% of the affected femoral heads and noted a direct correlation between the severity of physeal involvement and the ultimate deformity of the head. They

commented on the obvious importance of following all patients with Perthes disease to skeletal maturity. Keret et al. felt, however, that much of the physeal injury was related to epiphyseal stress fractures. If major epiphyseal compression and distortion occurred, then damaging loads would fall on the germinal cells of the physis, which normally are shielded by the intact articular cartilage and bone of the capital epiphysis. The fractures would mechanically flatten the germinal layer of the physis such that full healing does not occur. The possibility of micro-transphyseal bone bridges also was raised. All patients with Legg-Perthes disease should be followed to skeletal maturity so as not to overlook the premature closure of the physis, which can occur several years after the onset of the disease and after repair appears to have been complete with growth reestablished. It is due to excess pressure on the lateral physis inhibiting growth, physeal damage due to vascular insufficiency of the primary disorder, and transphyseal bone bridge formation. The physis may continue to grow, with apparent healing of the secondary ossification center necrosis, but be unable to sustain full growth to the normal time of skeletal maturation. 3. TREATMENT-RELATED SHORTENING The treatment method chosen in Perthes can have an additional effect on length discrepancy at skeletal maturity. During the era in which unilateral abduction non-weightbearing bracing was prominent, decreased function of the affected limb increased the discrepancy because the normal growth stimuli to the distal femur and the tibia also were absent. Once brace use was discontinued, however, resumed function of the limbs allowed for an increase in the growth rate of the other normal physes and a diminution of the discrepancy (246). At present unilateral bracing is rarely used as it was found to be ineffective because patients tilted into pelvic obliquity, thus limiting the effectiveness of the planned abduction of the involved hip and increased coverage gained in that fashion. Operative intervention is used commonly at present in many centers. In those having femoral head coverage improved by innominate osteotomy, a secondary benefit of the procedure is a gain in lower extremity length of 1.0-1.5 cm as the pelvic osteotomy site is opened and stabilized by the inserted graft. Proximal femoral varus and derotation osteotomy also provides for better coverage of the femoral head in the standing position, but varus osteotomy almost invariably shortens the femur and further increases any length discrepancy. Mirovsky et al. provided a detailed comparative study of residual shortening after varus osteotomy for Perthes disease compared with a group of patients treated in weight relieving braces (193). They documented the residual shortening in 43 patients having subtrochanteric varus derotation osteotomy and 47 treated with the weight relieving brace, each of whom had reached complete or near-complete skeletal maturity. In both groups the average shortening was 0.9 cm. In a group of patients treated conservatively by Edgren, 50 patients had an average

SECTION VIII ~ Lower Extremity Length Discrepancies with Legg--Perthes Disease residual shortening of 1.5 cm, whereas a group of 30 patients treated conservatively by Gower and Johnston had a residual shortening of 1.6 cm. Due to concerns about shortening that arose during the development of the operative technique of varus derotation osteotomy, changes were instituted. In those having a closing wedge resection, 15 patients had an average shortening of 1.1 cm. In those who had a reversed half-wedge resection (in which a closing wedge was removed medially and then replaced into the lateral opening), the average shortening in 5 patients was 0.8 cm. Finally, when an open wedge was used allowing the gap to heal with new bone, the average shortening was diminished in 4 patients to 0.4 cm. Another way of looking at the data indicated that, after osteotomy, 58% of 55 patients had shortening of the affected limb, 27% had limb length equal to that of the opposite side, and 8 patients (15%) actually had some lengthening (although these patients mostly had a derotation without varus osteotomy). As in other series, those who had osteotomy up to 8 years of age had less residual shortening due to the remodeling potential than those who had osteotomy after 8 years of age. Poorer results in terms of more shortening also were seen in those with fair and poor results and in those who had less remodeling of the varus deformity. In 34 osteotomized patients under 7 years of age, at the onset of symptoms the residual shortening was 0.6 cm, and in 21 patients more than 7 years old when symptoms started, the average residual shortening was 1.1 cm. The data still indicate, however, that in some patients shortening can be considerable. In patients operated at a mean of 8.5 years (6 patients), the average residual shortening was 2.3 cm with a range from 1 to 3.75 cm. Some of this was felt to be due premature physeal closure. Leitch et al. noted a 6% incidence of leg length discrepancy greater than 2 cm after both nonoperative and operative treatment. They also quantified the articulotrochanteric distance (ATD), which was less than 5 mm in 23% of patients, 43% of whom had a positive Trendelenburg sign. They also noted a significantly lower mean ATD in patients treated by femoral varus osteotomy, which they felt should be avoided in patients over 8 years of age (173). No discrepancies were noted after innominate osteotomy. Thus, it is extremely important to follow patients with Legg-Perthes to skeletal maturity to assess shortening caused by three possible mechanisms: (1) the Perthes disorder, (2) premature cessation of growth in adolescence, and (3) femoral shortening by varus osteotomy. A detailed review of our experience with lower extremity length discrepancies in Legg-Perthes disease follows.

VIII. L O W E R E X T R E M I T Y LENGTH DISCREPANCIES WITH LEGG-PERTHES D I S E A S E A study of lower extremity length discrepancies in LeggPerthes disease from Children's Hospital, Boston, will be

321

reviewed (246). This series did not include patients having innominate or proximal femoral osteotomies.

A. Maximum Total Femoral and Tibial Discrepancy during Growth Years During the course of the condition, the average maximum total femoral and tibial discrepancy in all patients was 2.14 cm. In those requiring epiphyseal arrest (31 patients) it was 2.99 cm, and in those not requiting epiphyseal arrest (116 patients) it was 1.91 cm. The maximum average discrepancy reached during the course of assessment in 147 patients was 1.5 cm or more in 113 patients (77%) and 1.0 cm or more in 139 patients (95%).

B. Femoral and Tibial Discrepancy at Skeletal Maturity The average final total femoral and tibial discrepancy in the entire group was 1.21 cm. The final average discrepancy in the entire series with or without epiphyseal arrest was 1.5 cm or greater in 50 patients (34%). In the group that had epiphyseal arrest the final average discrepancy was 1.27 cm, and in the group that did not it was 1.21 cm.

C. Maximum Femoral Discrepancy Maximum femoral discrepancy in all cases averaged 1.38 cm. In those not having arrest the average maximum shortness was 1.18 cm, and in those having arrest it was 2.09 cm. The final femoral difference in those not having arrest, which is indicative of the extent of spontaneous correction, averaged 0.92 cm. As the maximum femoral difference was 1.18 cm and the final difference 0.92 cm, the average spontaneous correction was 0.26 cm.

D. Maximum Tibial Discrepancy The average maximum tibial difference in all cases was 0.93 cm. In those not having arrest the average was 0.84 cm, and in those having arrest it was 1.28 cm. The final tibial difference in the nonarrest cases averaged 0.30 cm, indicating average spontaneous correction of 0.54 cm. The time of maximum femoral discrepancy rarely coincided with the time of maximum tibial discrepancy.

E. Developmental Patterns of Discrepancies in Legg-Perthes Disease When the extent of the discrepancy in centimeters was charted in relation to age, it became evident that not all discrepancies had increased continually with time. A series of patterns of the developing discrepancies was identified and a classification made. These were presented initially as types A, B, C, and D, but are referred to here using Developmental Pattern Classification types 1-5 from our review of several

322

CHAPTER 4 ~ Legg--Calve--Perthes Disease

disorders causing lower extremity length discrepancies. The type 1 developmental pattern group (discrepancy increasing continually with time) occurred in 21, the type 3 pattern in 60, the type 4 pattern in 10, and the type 5 pattern in 49, in which the discrepancy occurred, reached a plateau, and then partially or completely corrected without surgery. The marked slowdown of growth in the femoral head in Legg-Calve-Perthes disease associated with necrosis, subchondral fracture, and collapse leads to shortening, the occurrence of which has been recognized for decades (172). The average maximum femoral and tibial shortening in all patients was 2.14 cm. In the 21% of patients who eventually had epiphyseal arrest to correct discrepancies, the average maximum femoral and tibial discrepancy was 2.99 cm. A significant contribution to the lower extremity shortening in this series was from the ipsilateral tibia, with the average maximum tibial discrepancy being 0.93 cm due to disuse related to unilateral immobilization during treatment. The large majority of patients in this series were treated with the abduction patten bottom splint for 1.5-4 years, with good correlation noted between the tibial discrepancy and the time of immobilization. This study demonstrated that the repair process alone can lead to meaningful correction of discrepancies. When the discrepancies were plotted against age, differing developmental patterns were outlined, referred to as types 1, 3, 4, and 5, and illustrated by four case studies. The pattern types relate directly to the disease and repair phenomena rather than to skeletal age. The four patterns seen indicate that careful continuing assessment of lower extremity length discrepancies should improve long-term results. A good correlation was found between the average age at presentation and the final developmental pattern of the discrepancy that developed. In the type 1 pattern the discrepancy continues to increase with time. The average age at presentation in this type was the oldest of the four groups at 8.7 years. The older patients have a relatively larger amount of necrosis and, more importantly, less time for repair. The type 5 pattern, in which the discrepancy was corrected partially or entirely by the repair process, was associated with the youngest average age at presentation at 5.3 years. In these patients there was adequate time to allow for generally excellent repair to occur. In type 3, in which the discrepancy reached a plateau, the average patient age at presentation was intermediate between that of types 1 and 5 at 6.5 years. These numbers do not represent absolute correlations; they do, however, provide an indication of how awareness of the developmental pattern classification can help in determining the outcome of discrepancies with time. The type 4 pattern, if not appreciated, can worsen a discrepancy just before skeletal maturation due to premature closure of the proximal femoral capital epiphysis. There were 10 instances of this phenomenon documented when both femoral and tibial measurements were plotted. The femoral discrepancy alone showed this conformation in 14 (10%). Much of the type 5 pattern in which correction

occurs, or at least spontaneous correction that removes the discrepancy from the clinically significant group, appears to be due to the release of unilateral immobilization at the termination of therapy. The discrepancies in this group of patients are partially due to growth slowdown relating to the proximal femoral capital epiphysis, but some of the discrepancy also is due to growth slowdown in the distal femoral and both tibial physes in relation to prolonged unilateral immobilization. This phenomenon also was well-documented by Willner (285). He reported on 55 patients with unilateral LeggPerthes disease who were treated by crutches and slings, such that there was unilateral unloading while the opposite limb continued with full function. Subsequent growth studies from the beginning of treatment, throughout treatment, and continuing for as long as 5 years following treatment showed that shortening caused by the deformation of the femoral head remained unchanged, but shortening caused by immobilization of the femur distal to the greater trochanter and the tibia corrected with time. When treatment began, no observable difference in leg length could be found in most of the cases. As expected, however, the affected legs of all cases were shorter during the treatment and this shortening became more marked as treatment continued. The study then concentrated on growth changes in lower extremity lengths of the femur distal to the trochanter and the entire tibia with time. The shortening in the limb distal to the trochanter of the femur was at a mean of 8.0 mm and that in the tibia was 7.9 mm at the completion of treatment. At the beginning of treatment the mean discrepancy was only 1.7 mm short, but at 6-month intervals until 36 months of age the shortening continually increased from 6.3 to 10.3 to 17.8 to 18.8 to 22.0 mm at 36 months. Once the immobilization was released, there was an equally impressive diminution in leg length discrepancy with time. Six months after the end of treatment the discrepancy had decreased to a mean of 16.7 mm, and at 12-month intervals the value continued to diminish to 12.9, 9.9, 8.6, 4.6, and 4.2 mm at 60 months. Their illustrations fully confirm the existence of the type 5 pattern also noted in our study. Those 31 patients having distal femoral epiphyseal arrest had an average maximum total femoral and tibial discrepancy of 2.99 cm, with a final total discrepancy averaging 1.27 cm. The existence of a large head with lateral extrusion or subluxation was such that the physician planning arrest frequently wished the involved extremity to remain slightly shorter such that, in gait, head coverage would be greater. The sluggish skeletal age maturation appeared to have made timing prediction more difficult than in other diseases, and failure to appreciate the differing developmental patterns outlined here on occasion led to inaccuracies in timing. This was especially true with type 4 patients in which arrest occasionally was done late or not at all, with the end stage increase in discrepancy missed due to failure to appreciate the early closure of the proximal femoral plate. Change in

SECTION IX ~ Prognostic Indicators During the Active Disease Process

the femoral head-greater trochanter relationship is an early radiologic indicator of this occurrence.

IX. P R O G N O S T I C I N D I C A T O R S D U R I N G THE ACTIVE DISEASE PROCESS

A. General Considerations The prognosis in Legg-Calve-Perthes disease is extremely variable, ranging from reestablishment of a completely normal hip to development of a fiat misshapen head with a complete lack of congruency with the adjacent acetabulum. The treatments used also are quite variable, ranging from observation only in very young patients of 3 or 4 years of age, to long-term bracing, to femoral or pelvic osteotomy. Great efforts have been expended in attempting to define, during the active phase of the disease, what the ultimate prognosis will be so that observation alone will not be prolonged in those who could benefit from more active intervention, and, equally importantly, long-term bracing or surgery will not be done in those who might be expected to heal relatively uneventfully with minimal to no intervention. In spite of extensive efforts, however, to correlate various radiographic parameters with the eventual outcome, one of the most helpful predictive factors remains the age of occurrence of the disorder. There is not an absolute correlation between the age of occurrence and the end result, but this clinical feature still remains the simplest and best indicator. At the time of initial occurrence of the disorder, by which is meant the time of the vascular insult, plain radiographs would be perfectly normal. Currently it is possible to diagnose the disorder at the earliest stage on the basis of a negative bone scan in which the involved side shows no uptake in the secondary ossification center of the proximal femur. The condition only rarely is diagnosed at this stage currently because it usually is brought to medical attention by clinical hip or thigh discomfort with a several-week to -month history, by which time early plain radiographic changes have occurred. At the time of disease occurrence, however, we have no ability to project the future outcome other than the age of the patient at the time. The various radiographic classifications that have evolved have as a major weakness the fact that the disease process is well underway and usually well into the repair stage at the time that the radiographs are taken. Classification by radiographic appearance alone does not distinguish the age of the patient or the stage of the repair process. Two-dimensional radiographs are imperfect in estimating the three-dimensional geometric shapes of the femoral head-acetabular relationships. Another problem is the clinical demonstration that one does not need perfectly normal hip structure at skeletal maturity to assure several decades of normal pain-free function. Several categorizations have been developed over the past three decades, which are designed to provide prognostic

323

and treatment guidelines. The simplest approach involves the age at which the initial disease occurs. Other plain radiographic categorizations have provided much useful information in assessing the evolution of a Perthes condition but still lack total specificity in terms of directing therapy as they tend to provide information in the revascularization and residual phases of the disorder. Thus, they have a tendency to define what has already occurred rather than defining, early on, what will occur.

B. Age of Occurrence of the Disease The age at which the Legg-Calve-Perthes disorder occurs remains the single best prognostic indicator of the outcome. Although there is not an absolute correlation, the best results occur in those developing the disorder at the younger end of the spectrum at ages 3, 4, and 5 years, and the worst results occur in those developing the disorder at the older end of the spectrum from ages 10 to 13 years. Ippolito et al. reported on an adolescent group developing Legg-Perthes from 13 to 15 years of age. All developed pain and restriction of motion between skeletal maturity and 39 years of age (130). Many other reports have documented that, when the disorder develops in late childhood or adolescence, the long-term resuits usually are poor. It is those who develop the condition between the ages of 6 and 9 who have the widest variation of end results and who represent the group in which more precisely defined treatment could allow for the most optimal results. Even within this relatively narrow age group, it is those acquiring the disorder at 8 or 9 years that have the poorer prognosis. Vila Verde et al. noted that, in children older than 9 years of age, the results invariably were poor irrespective of the head at risk criteria of Catterall or treatment measures (274). The observation that the younger ages of occurrence predisposed one to a better result is one of the most uniform in Legg-Perthes studies, a field in which great diversity of opinion is common. The work of Snyder points this out but also reminds us that deformed femoral heads are seen after repair even in those 5 years of age and younger (253). The explanation for the finding appears to rest on the fact that the younger the patient the greater the relative amount of cartilage epiphyseal tissue present compared to bone in the femoral head. The amount of bone to be repaired thus is smaller, and because it is the cartilage model that is responsible both for shaping the femoral head and for the subsequent growth, the fact that there is a relatively large amount of cartilage that is less affected by the disorder leads to the improved repair capability. In addition, the greater period of time from healing of the secondary ossification center to skeletal maturity allows the remodeling capability of femoral head and acetabular cartilage to assert itself. In contrast, the poorest results occur in the older age groups, at which time the proportion of secondary ossification center bone in relation to epiphyseal cartilage is much greater and the converse

324

CHAPTER 4 ~

Legg--Calve--Perrhes Disease

situation applies, with far more bone to be repaired and far less cartilage available to serve either as a preventative to collapse or as a source for head shape remodeling. In those with excellent or good long-term results, the acetabulum has molded itself to the slightly irregular shape of the head allowing for a situation referred to clinically as congruency, which appears to have a major saving effect in terms of eventual osteoarthritis. The acetabulum in particular loses its capacity for shaping responses toward the end of the first decade.

C. Plain Radiographic Classifications Three major categorizations have been developed over the past two decades primarily based on the appearance of the femoral head during the revascularization phase (40, 117, 237). These are referred to widely, provide much useful information about the response of a particular femoral head to the Legg-Perthes disorder, and have been the focus of numerous studies designed to correlate the eventual outcome with the findings of the classification. The three most commonly used categorizations are those of Catterall (40) and Salter-Thompson (237) and the most recently described lateral column classification of Herring et al. (117). There are major problems, however, with these categorizations both in theory and, as clinical studies continue to show, in correlation with eventual outcomes. Each of the three categorizations uses a plain radiographic approach. There are high interobserver differences in gradings particularly with the Catterall classification. In each of the three there is clear potential for the change of classification with temporal progression of the disease. The major shortcoming of these approaches is that they focus on the bone pattern of the secondary ossification center of the femoral head, whereas the final outcome of any Legg-Perthes disorder is dependent primarily on the shape of the cartilage model of the head. During the stage of the evolution of the disorder, the cartilage model is much larger than the secondary ossification center, and frequently abnormal shaping of the secondary center is not reflected in the shape of the cartilage model, which may, and frequently does, remain spherical. Two additional plain radiographic classifications, those of O'Garra (205) and Hirohashi et al. (123), show yet other ways to assess the variable findings. 1. O'GARRA O'Garra commented on the prognostic value of lateral radiographs of the hip in particular (205). The lateral view demonstrated the disorder earlier and more accurately than the anteroposterior view. Two basic patterns of Perthes were defined: (1) an "anterior" group, in which the anterior onehalf or two-thirds of the head of the femur was affected and any metaphyseal changes if present were in the anterior onethird of the neck, and (2) involvement of the whole epiphysis, in which the whole femoral head was affected. O' Garra

indicated that Perthes disease in which only the anterior part of the head was involved had a good prognosis, although those with full head involvement had a poorer prognosis. This concept was later expanded in particular by Catterall to a more inclusive consideration. 2. CATTERALL CLASSIFICATION The Catterall classification is based on an interpretation of the extent of femoral head involvement on both anteroposterior and lateral plain radiographs (40-43). A precursor to this descriptive approach was the definition of two types of Legg-Perthes disease by O'Garra, who noted an "anterior" variant and a "whole head" variant. In the former, the anterior one-half or two-thirds of the femoral head was affected as seen best on the lateral radiograph. The healing was far better in those with less than full involvement (Fig. 27). The Catterall classification involves groups I-IV progressing from the most limited involvement to the most extensive involvement. The initial impression was that the prognosis worsened with an increasing grade of classification. There is a considerable element of accuracy to this, but the correlation is not complete and its use as a clinical tool other than for groups I and IV is problematic. Group I: Only the anterior part of the epiphysis is involved. The abnormality is seen most clearly on the frog lateral view, which highlights the anterior from the posterior regions of the head. There is no collapse of the bone or cartilage model of the proximal femur and the height of the epiphysis is maintained. There is complete absorption of the involved segment with no sequestrum formed. Group II: More of the anterior part of the epiphysis is involved than in type I. The involved segment, after a phase of absorption, undergoes collapse although reference is to the bone of the secondary ossification center rather than to any specific assessment of the epiphyseal and articular cartilage. The area of increased bone density, which Catterall refers to as a sequestrum, is central on the anteroposterior projection with medial and lateral bone structure maintained. Group III: Only a small part of the epiphysis is not sequestrated. This invariably refers to the portion that is medial and somewhat posterior. The anteroposterior radiograph demonstrates the "head within a head" phenomenon. It also shows the collapsed central sequestrum with some medial and lateral head continuity seen. Group IV." The whole epiphysis is involved and appears on the anteroposterior view with complete collapse of the epiphysis, producing a dense line. There is flattening of the head. The term epiphysis refers to the secondary ossification center. On the lateral radiograph there is no viable posterior segment. Metaphyseal changes may be extensive. A theoretical critique of the Catterall classification is easy to make, although it in no way diminishes the value of the work in providing a previously nonexistent framework for study of the disorder. If one considers the evolution of the disorder, then at the initial moment of osteonecrosis the plain

SECTION IX ~ Prognostic Indicators During the Active Disease Process

oup I ,

325

G r o u p II

9

"

%

No metaphyseal reaction No sequestrum No subchondral fracture line

Sequestrum present--junction clear Metaphyseal reaction~ antero/lateral Subchondral fracture line--anterior half

F.

!.,',

~ I

Group

~up

t

Sequestrum m large m junction sclerotic Metaphyseal reaction m diffuse antero/lateral area Subchondral fracture line ~ posterior half

Whole head involvement Metaphyseal reaction m central or diffuse Posterior remodelling

FIGURE 27 The Catterall classification involves groups I-IV. [Reprinted from Catterall, A. (1981). Clin. Orthop. Rel. Res. 158: 41-52, 9 LippincottWilliams & Wilkins, with permission.]

radiograph will show a normal appearing secondary ossification center. Thus, at the initial time of occurrence no prognostic feature is truly available, assuming an X ray is done at this stage, because all will be graded as class I or even appear to scarcely allow a diagnosis to be made. As noted in the earlier sections of this chapter, the radiographic changes of the secondary ossification center that ensue are a result not only of the osteonecrosis but also primarily of the repair response in which new bone is laid down on old bone, resorption of necrotic areas is occurring more or less simultaneously, and shaping changes of the cartilage model of the femoral head occur. If one looks at virtually any sequential series of X rays from a patient with Legg-Calve-Perthes disease, it is not infrequent to note that the classification would change depending on the time the radiograph was taken in relation to the disease stage. If one considers the initial appearance in which the secondary centers essentially are normal and then the appearance of a group IV lesion, it is evident that the head does not collapse overnight to a flattened radiodense image but rather goes through a series

of changes that essentially would move the patient through groups II and III. The Catterall classification rating given thus may well be dependent on the time the radiograph was taken during the disease process rather than the extent of involvement. The classification is helpful for those individuals who do not progress beyond group I or II. It can be difficult to determine, however, during the course of the disorder whether an individual is in a type II categorization as the final extent of involvement or simply is passing through this stage to a worsening type. Van Dam et al. used the Catterall classification to assess 50 hips with Legg-Perthes (272). The rating changed in 40% of the hips when they were classified before they had reached the fragmentation stage compared with only 6% changing when classified after fragmentation. In conjunction with this categorization, Catterall developed "head at risk" signs, which were considered indicative of a worsening prognosis. Although some of these signs may indicate a poor prognosis, their consideration in relation to the underlying pathoanatomy indicates that they may not

326

CHAPTER 4 ~ Le~tf--Calve--Perthes Disease

necessarily have such a poor prognosis. He defined four "head at risk" signs. (1) The Gage sign: This was defined by Catterall as a small osteoporotic segment, which forms a radiolucent " V " notch on the lateral side of the epiphysis. As noted earlier, Gage originally described lysis leading to convexity of the upper-outer border of the femoral neck as being an early radiologic sign of Perthes disease. This may not always be a head at risk sign, because revascularization occurs from the lateral and posterior aspect of the head and neck region from the retinacular vessels concentrated there, and because the initial response is resorption of the necrotic bone, it is not surprising that early lysis tends to predominate in this region. (2) Lateral subluxation: As noted earlier it frequently is unclear whether the radiologic appearance described as lateral subluxation represents a true subluxation in terms of displacement of the head of the femur in relation to a normal acetabulum, which presumably would occur due to the head being pushed out of the acetabular space by increased fluid or synovium, or whether it represents a lateralization appearance of the bone of the neck and secondary ossification center, with the enlarging radiolucent cartilage model of the head giving the appearance of a subluxed position. If the latter is the case, then the so-called lateral subluxation is not a true displacement but rather indicative of the development of a coxa magna. As Stulberg has shown, an isolated coxa magna is not a bad prognostic sign particularly in those situations in which the acetabulum is changing its shape to relate to the enlarged femoral head. (3) Calcification lateral to epiphysis: Catterall uses the term epiphysis to refer to the secondary ossification center. Lateral calcification is not surprising because in the vast majority of patients with Legg-Calve-Perthes disease the cartilage model of the head is larger than the normal side and early new bone repair formation tends to occur laterally because that is the earliest point of ingress of repair vessels. (4) Angle of the epiphyseal line: By this term, Catterall refers to the orientation of the epiphyseal growth plate, which normally has a slight obliquity to it. In many patients with Perthes, the line casts a horizontal radiographic shadow because of the fact that the medial one-third of the physis maintains its growth, whereas the lateral two-thirds tends to show growth diminution. An additional or fifth "head at risk" sign has come to be attributed to Catterall; this is defined as metaphyseal cysts or diffuse metaphyseal reaction by others. Studies have been performed in relation to the value of the head at risk concept in assessing prognosis. Vila Verde et al. assessed 75 hips with the Perthes disorder followed to skeletal maturity and concluded that patients with "head not at risk" had better results regardless of treatment (274). In children greater than 9 years of age, the results almost invariably were poor irrespective of head at risk designation. Within the younger age group, however, the presence of head at risk findings tended to worsen prognosis particularly with an increased number of findings. The group assessed head at risk signs involving lateral subluxation, lateral ossi-

fication, the Gage sign, and metaphyseal cysts, although the horizontal plate was not assessed in detail. They felt that lateral subluxation, present in 41 hips, and lateral ossification, present in 34 hips, were the most relevant signs of head at risk, particularly in the 31 hips in which they occurred together. The Gage sign was relatively rare, and metaphyseal cysts, though present, were not felt to relate to ultimate prognosis. They felt that within the head at risk group, in particular those under 9 years of age, results were improved significantly with varus derotation proximal femoral osteotomy. The head not at risk group was concentrated in the younger patients. Murphy and Marsh also assessed head at risk factors in relation to prognosis using each of the five criteria (200). Twenty-eight patients were assessed. The most common risk factor was lateral subluxation, with diffuse metaphyseal reaction being the least common. The most common risk factor with a poor result was lateral subluxation. No absolute correlations were noted, but in general those with 3-5 risk factors had a poorer prognosis than those with 0 - 2 risk factors. In the latter group, there were 11 good results, 3 fair results, and 1 poor result. They found classification by the Catterall degree of epiphyseal involvement approach to be difficult in particular at the time of initial diagnosis because the classification tended to change as the disease progressed. On the other hand, they found the five head at risk factors to be more accurately predictive of the course of Perthes disease. Dickens and Menelaus found the Catterall groupings to be applicable in the senses that a spectrum of group I-IV abnormalities could be defined reliably with good agreement between different observers and that groups I and II had a much better ultimate prognosis than those defined as groups III and IV (61). They did note changes with time in the group, however, and felt that the final gradation could not be determined definitely often for a period of up to 8 months from presentation. They also assessed their patients by the head at risk criteria and found, similar to others, that the factors most frequently indicating that the result would be poor were lateral calcification and lateral subluxation or displacement of the femoral head. The horizontal epiphysis was not felt to be of correlative value.

3. SALTER-THOMPSON SUBCHONDRAL FRACTURE CLASSIFICATION This classification is based on the impression that only that part of the epiphysis (secondary ossification center) underlying the subchondral fracture eventually is resorbed (237). The extent of the subchondral fracture thus is felt to represent an important early indicator of the eventual amount of femoral head bone involvement. Only two groups are defined in this categorization, with group A showing involvement of less than half of the head and group B showing involvement of more than half of the head (Fig. 28). The subchondral crescent when present can be seen on both anteroposterior and frog lateral X rays, although the frog lateral

SECTION IX ~ P r o g n o s t i c Indicators D u r i n g

ANTERIOR-POSTERIOR

LATERAL

ANTERIOR-POSTERIOR

LATERAL

MAXIMUM RESORPTION

MAXIMUM RESORPTION

~

ANTERIOR

ANTERIOR-POSTERIOR

327

SIJBCHONDRAL FRACTURE

B

SUBCHONDRAL FRACTURE

the Active Disease Process

SUPERIOR

LATERAL

ANTERIOR-POSTERIOR

R

)

SUPERIOR

LATERAL

SUBCHONDRAL FRACTURE

C

SUBCHONDRAL FRACTURE

ANTERIOR-POSTERIOR

LATERAL

ANTERIOR-POSTERIOR

LATERAL

MAXIMUM RESORPTION

\

MAXIMUM RESORPTION

~~

~iOR

ANTERIOR ANTERIOR-POSTERIOR

LATERAL

SUPERIOR

ANTERIOR-POSTERIOR

LATERAL

SUPERIOR

FIGURE 28 The Salter-Thompsonsubchondralfracture classificationis shownin parts (A-D). In each illustrationthe subchondral fracture line is shown along with the adjacent necrotic area, which is determinedon the basis of the extent and position of subchondral radiolucency. [Reprintedfrom (237), with permission.]

is somewhat more helpful in providing documentation using this approach. Limitations of this categorization also have become apparent with time. In large clinical studies, including that of the original description, only about 50% of patients have radiographs during the course of their assessments that demonstrate the subchondral crescent. This radiographic finding is not present for very long and in many patients has disappeared by the time of initial X ray. In some it may never occur. The sign also is dependent to a certain extent on the plane of the radiographic projection. In routine clinical instances only two radiographs are taken, and the subchondral sign can be missed unless the defect is relatively large. For those wishing to use this indicator it would appear mandatory to take multiple radiographic projections, perhaps even under fluoroscopic assessment, to detect any lucent crescent that is present. MR imaging has a greater likelihood of detecting subchondral fracture. Studies from other centers

show poor correlation of the sign with ultimate outcome, even though the initial paper indicated that "in all hips the extensive subchondral fracture correlated precisely with the subsequent extent of maximum resorption." 4. LATERAL PILLAR CLASSIFICATION: HERRING E T AL. This plain radiographic classification is based on the observation that fragmentation of the secondary ossification center in Legg-Calve-Perthes occurs in distinct anatomic sectors of the femoral head (117). The head is divided into three sectors or pillars: the lateral, middle, and medial regions. By fragmentation the authors are referring to what we have defined as interspersed areas of lysis. This classification divides the patients into A, B, and C categories during the fragmentation stage of the disease (Fig. 29). In group A there is no involvement of the lateral pillar, which is

CHAPTER 4 9 Le99--Calve--Perthes Disease

328

Lateral Pillar Classification of Legg-Perthes Disease

Normal Assessed

Group B

Group A

Group C

in fragmentation stage from antero-posterior hip radiograph

Group A Group B Group C

F I G U R E 29

Height of lateral pillar normal Height of lateral pillar > 50% Height of lateral pillar < 50%

The lateral pillar classification of Herring et

al.

radiographically normal. In group B more than 50% of the lateral pillar height is maintained, and in group C less than 50% of the lateral pillar height is maintained. Many of the criticisms directed against the previous two categorizations also apply here. The phenomenon being assessed is the bone of the secondary ossification center, whereas the shape of the cartilage model, which cannot be determined in plain radiographs, is not addressed. It is well-known that cartilage shape does not correspond during the developing stages of a Legg-Perthes disorder with the shape of the secondary bone center. Because the categorization is made during the fragmentation stage, which is well into the repair phase, it essentially documents what has already occurred rather than acting as an early prognostic determinant. If we again consider the example of a patient diagnosed very early just after the osteonecrotic insult, the lateral pillar classification would be grade A. As resorption occurs progressively over time,

none

epiphvsis

!

5. HIROHASHI ET AL. Hirohashi et al. developed a classification for Perthes dis-

ease based on the extent of both epiphyseal and metaphyseal involvement as indicated in lateral hip radiographs (Fig. 30) (123). They felt their results correlated well with the classification, with the greater the extent of involvement of both epiphysis and metaphysis, the more severe the end result.

2

I

taphysis

the patient would slip into the grade B and then grade C classifications. The paper indicates specifically that the classification does not change during the course of the disorder. If one accepts the accuracy of the example we gave earlier, it is hard to see how this can be true. Classification is not possible at the time of an initial film. Nevertheless, if the patient does not advance beyond stage A, then follow-up would indicate that involvement was less severe, and certainly if the patient is in group C then by definition involvement is severe. However, if a patient is diagnosed in category C at the time of the initial radiograph, then the categorization is not indicating what will happen but rather what has already happened. The Herring et al. group felt that interobserver reliability was highest for their classification system, lower for the Catterall classification, and lowest for the head at risk determinations. Farsetti et al. have assessed retrospectively radiographs of 49 patients at the stage of fragmentation and again at skeletal maturity (74). They found the classification to be relatively easy to apply and reliable if the radiograph for assessment is at the fragmentation stage of the disease. A total of 10 of 11 group A hips showed good reconstruction of the femoral head. Good results were noted in group B hips when the patients were less than 9 years of age at diagnosis, and all 11 group C patients showed hip deformity at follow-up.

or

slight

( ~ > in width)

( > ]/z in width)

mild

(< '2) !-I

!!

3

extensive

moderate

!-3

!-2

/y

moderate

(, ._,) !1-1

I! - 2

Iii - 1

ill - 2

~ 1 1 _ 3

severe

!!!

total

III - 3

F I G U R E 30 The extent of epiphyseal and metaphyseal involvement as indicated in lateral hip radiographs is the basis of the classification of Hirohasai et al. [Reprinted from Hirohasai et al. (1980). Internat. Orthop. 4:47-55, copyright notice of Springer Verlag, with permission.]

SECTION X 9 Results Based on Appearances at Skeletal Maturity

They also used their categorization to influence treatment approaches. Categorization follows the "the worse it is, the worse it will be" approach. The epiphyseal abnormality was defined as mild (I), moderate (II), and severe/total (III) types, whereas metaphyseal involvement was defined as none/ slight (1), moderate/less than one-half the width of the metaphysis involved (2), and extensive/greater than one-half the width of the metaphysis involved (3). A grouping of nine possible gradations was then formed with, for example, moderate epiphyseal and moderate metaphyseal involvement being graded a 11-2 type.

6. MINIMALPERTHESDISEASE Herring et al. noted a small subset of patients, 12% (24 of 193), with focal Perthes disease and a benign natural history (116). A total of 10 of the 24 had only anterior head involvement as described by O'Garra and by Catterall as group I. Other areas noted, however, were 7 posteromedial, 3 lateral, and 4 central. There was no sequestrum formation or collapse, and the localized density changes resolved as in more extensive patterns of involvement. Patients had either no treatment or relatively brief brace treatment.

D. Comparison of Classification Schemes Efforts have been made to compare the validity of the various plain radiographic classification schemes. It is important to recognize that, whereas the classification schemes themselves are subjective, comparisons between them are even more so. Nevertheless, because the ultimate value of any classification scheme is the ability of a wide group of individuals to use it in a clinically valuable fashion, such reports are of interest. Mukherjee and Fabry compared the assessment of 116 hips with Legg-Perthes disease using both the Salter-Thompson and the Catterall classifications in relation to their prognostic value at the end of the disease process (199). They felt that both of the classifications had a high degree of prognostic significance. They also assessed the Catterall head at risk factors, which involved the Gage sign, metaphyseal reaction, calcification lateral to the epiphysis, horizontal line of the growth plate, and lateral subluxation. Other major problems of interpretation occur because the patients were treated by three different modalities, including no treatment, plaster abduction casts, various types of braces, and proximal femoral varus osteotomy. Mukherjee and Fabry found that the Salter-Thompson classification was easier to make and valued its use to form the basis of decisions on management particularly in the early stages of the disease. In terms of prognostic importance, they felt that the Salter-Thompson group A could be considered almost equal to Catterall groups I and II and Salter-Thompson group B to Catterall groups III and IV. A high correlation coefficient thus was established between the two classifications. In spite of all findings, however, they reached the conclusion that it was the age at presentation and the lateral subluxation of the femoral head that were the most important adjuncts in decid-

329

ing between conservative or surgical containment. Mukherjee and Fabry felt that the Salter-Thompson classification could not be done on 22 hips (of 116) because early radiographs were unavailable. This would indicate that the SalterThompson classification was possible in approximately 80% of cases, which was much higher than that reported by Catterall at 25% and others at 59%. Ritterbusch et al. compared the predictive value of the lateral pillar classification with that of Catterall (232). They felt that the lateral pillar classification was a significantly better predictor of Stulberg outcome than the Catterall classification. Christensen et al. (48) and Hardcastle et al. (101) both reported low degrees of interobserver agreement when using the Catterall approach and questioned whether the classification should be used to form the basis of treatment decisions.

E. More Recent Clarifications of Poor Prognostic Signs With increasing study, early radiographic signs of poor prognosis in Perthes disease are becoming better defined. The use of the term early is relative; in a sense these findings are indicative of a fairly advanced pathological process and they reflect as much a delay in appreciation of the disorder or rapid advancement of the disorder as they do of the fact that the condition is being analyzed shortly after its development. The first finding of negative prognostic value is that of hinge abduction. Definition of this occurrence is aided greatly by the use of arthrography, which allows for examination of the range of motion of the hip under fluoroscopic control and thus demonstrates the hinge effect with the extremes of abduction. The second group of negative findings defined by Yrjonen et al. includes (1) lateral calcification extending far laterally outside the epiphysis toward the greater trochanter; (2) deformation and widening of the femoral head before the fragmentation stage; (3) deformation and widening of the femoral neck in the initial phases of the disease; (4) early sclerotic changes in the metaphysis; and (5) a sclerotic secondary center surrounded by a ring of less dense repair tissue (290). Most observers would agree that points 1-3 are indicative of a marked disorder, although the changes described in points 4 and 5 can be seen in cases that subsequently heal with minimal negative sequelae.

X. CLASSIFICATIONS DEFINING RESULTS BASED ON APPEARANCES AT SKELETAL MATURITY AT THE END OF REPAIR A. General Considerations The long-term effects of childhood Legg-Calve-Perthes disease have been determined by several excellent clinical studies. It is evident that imperfect looking hip radiographs can still be compatible with a full or virtually full normal functioning of the hip throughout the patient's early and mid-adult

330

CHAPTER 4 ~ Le~tg--Calve--Perthes Disease

life. It is essential to take these studies into consideration and expand them by additional studies. Of the residual radiographic findings in Legg-Perthes disease at skeletal maturity, there is good agreement that a widened femoral neck is of no prognostic or functional significance concerning ultimate arthritis. Coxa magna often has been presumed to be a bad prognostic sign, but long-term studies indicate that in certain conditions coxa magna itself has no long-term deleterious effects. If the coxa magna is associated with a spherical head, which relates in a congruous fashion to an also somewhat enlarged acetabulum, minimal to no long-term problems occur. The early presence, therefore, of an apparently laterally subluxed head and calcification lateral to the outer border of the acetabulum may not necessarily be longterm poor prognostic signs if the cartilage model of the head remains in an appropriate shape and relationship to the associated cartilage model of the acetabulum. If the coxa magna is associated with a coxa plana, then a poorer longterm result can be expected. If a coxa plana persists at skeletal maturity, the overlying articular cartilage will not conform to the acetabulum and future problems can be expected. A minimal coxa vara, if associated with a round femoral head in relation to an appropriately shaped acetabulum, should not produce any problem. It may produce a mild Trendelenburg gait, but this should be readily controllable with either abductor muscle strengthening, epiphysiodesis of the greater trochanter, or distal transfer of the greater trochanter at skeletal maturity.

B. Sundt Classification One of the earliest classifications of the end result of a LeggPerthes disorder is that of Sundt based on 172 hips (261). He defined the shape of the femoral head into four categories: (I) spherical, (II) oval, (III) cylindrical, with or without hypertrophy of the greater trochanter, and (IV) quadrangular. Sundt considered categories I and II to represent favorable results, III less favorable, and IV unfavorable. He reviewed the long-term results from 172 hips, providing one of the earliest and most detailed assessments of the evolution of the disorder. He recognized that it was malformation of the head and neck of the femur that was the primary determinant of a poor result, with acetabular deformity occurring secondarily. In these cases, the end result would be an osteoarthrosis "evoked by the incongruence of the articular surfaces" of the femur and acetabulum. The whole purpose of treatment was to prevent deformation of the upper end of the femur and thereby "avoid an incongruity between the articular surfaces."

C. Quantitative Indices of Femoral H e a d - Acetabular Repair Extensive efforts have been made to quantitate the outcome of a Legg-Perthes disorder.

1. MOSE CONCENTRIC CIRCLE TEMPLATE METHOD A well-accepted index of sphericity of the head is that developed by Mose (198). He used a concentric circle template, with each circle separated by 2 mm, which was placed over a radiographic image of the femoral head from anteroposterior and lateral projections. The Mose template method is designed to assess the shape of the head in its entirety to determine whether it was spherical, and if not then the relative degree of deformity. To be classified as spherical with a good result, the surface of the head must follow the same circle on the template with no variation in both frontal and lateral views. Variation up to 2 mm is considered a fair longterm result, and a head with variation greater than 2 mm is considered poor (Figs. 31A-31C). Additional indices were used particularly when nonsphericity was present. A number of ratios were established. 2. EPIPHYSEALINDEX (EYRE-BROOK) The epiphyseal index describes the proportion between the height and width of the epiphysis (secondary ossification center) x 100 (Figs. 32 and 33) (72). For children under the age of 7 years, the normal is 45-55; for those over 7, it is 35-45. The epiphyseal index of Eyre-Brook expresses the flattening of the epiphysis and registers the height of the epiphysis from the growth plate to the highest point of the epiphyseal surface contour divided by the width of the epiphysis. 3. EPIPHYSEAL QUOTIENT (SJOVALL) This value is derived by comparing the height and breadth of the involved epiphysis (epiphyseal index) with that of the uninvolved contralateral side (Fig. 33) (251). Sjovall converted the epiphyseal index to an epiphyseal quotient by dividing the epiphyseal index of the affected head by that of the uninvolved side. A good result is between 75 and 100%, fair is 50-75%, and poor is less than 50%. Mose and others interpreted an epiphyseal quotient of 60% to be the dividing point between normal sphericity (greater than 60%) and an abnormal flattened (less than 60%) state. 4. ACETABULAR-HEAD INDEX (HERNDON AND HEYMAN) The acetabula-head index is that part of the femoral head ossific nucleus covered by the bony acetabulum divided by the entire ossific nucleus • 100 (Figs. 34A and 35) (115, 122). MR imaging has been used to assess the cartilage model of the femoral head covered by the cartilaginous lateral acetabular margin divided by the entire cartilaginous width of the head x 100 by Sales de Gauzy et al., and arthrography has assessed the same parameters (Moberg et al.) (Fig. 35). 5. COMPREHENSIVE QUOTIENT (HERNDON AND HAYMAN) Herndon and Hayman developed a comprehensive quotient composed of values from the epiphyseal quotient (the height and length of the epiphysis), the head-neck quotient

SECTION X

A

~

9

Results Based on Appearances at Skeletal Maturity

331

SUBJECT-R.K. <

"1

~ SPHERICAL HEA

FIGURE 32 The epiphyseal index establishedby Eyre-Brookregisters the flatteningof the epiphysis in Legg-Perthes comparedto the normal hip. The index is determined for the abnormal hip by the measurementof secondary ossification center height, divided by the width, and multiplied by 100. [Reprintedfrom (190), with permission.]

B

\

SUBJECT-R.C.

SPHERICAL HEAD

E.Q.:49 %

C

~

tients were calculated as a percentage of the normal hip, and each of these was then added and the result averaged to give a percentage of normal. These various measurements are illustrated in Fig. 34B. Axer used this measure in his initial paper on subtrochanteric osteotomy in Perthes; 9 0 - 1 0 0 % was very good, 80-90% good, 7 0 - 8 0 % fair, 6 0 - 7 0 % poor, and less than 60% bad (5).

'

~ SUBJECT"O.M.

6. CENTER-EDGE ANGLE (WIBERG) This value measures the relationship of the femoral head to the acetabulum. The CE angle is formed by a vertical line through the center of the femoral head and a second line beginning at the center and extending to the outer edge of the acetabulum. The greater the angle, the better the femoral head coverage. Normal is 20 ~ or greater, 15-19 ~ is fair, and less than 15 ~ is poor. 7. JOINT SURFACE QUOTIENT (MEYER) This measure was considered to best represent the shape of the head (190). The normal range was greater than 85.

IRREGULAR HEAD/ .

.=

FIGURE 31 The Mose concentriccircle templatemethodis an index of sphericity used to determinehead shape in Legg-Perthesdisease.Parts (A-C) from the work of Katz outline good, fair, and poor results, respectively,as determined by the Mose criteria. [Reprinted from (142), with permission.]

(the distance from the intertrochanteric line through the center of the neck to the top of the other secondary center divided by the nalTowest width of the neck), the acetabular quotient (the depth of the socket divided by the width of the socket), and the acetabulum-head quotient (the width of that part of the bony epiphysis covered by the socket divided by the entire width of the bony epiphysis) (115, 122). All quo-

The Mose rating and the epiphyseal quotient are two determinations that can be made at skeletal maturity with a quite reasonably high degree of interobserver agreement. Examples of this are shown from the work of Katz (142). In

/

NORMAL HIP A'= Z l m r n B': 54 m m

"~ B"

j

S

LI[OG'PERTNE5 HIP A= 17rnm

B:61 rnrn ,.o[x_~.xmo:39 INOEx~xl o0=Za QUOTIENT:~-gB-7ZZ

FIGURE 33 The epiphysealquotientwas derivedby Sjovallby comparing the epiphyseal index of the abnormal hip to that of the normal hip. The quotient is obtained by dividing the epiphysealindex of the affectedhead by that of the normal uninvolvedhead. [Reprintedfrom (122), with permission.]

332

CHAPTER 4

9

Leyg--Calve--Pe~hes

A

.

Disease

9

I

B

~

B

AC['I"ABULUM- HEAD Q U O T I E N T NORMAL HIP LrGG-PERTH[$ HIP K=54mrn A:53mm 8"= 5 8 r a m B:61 m m

,.,.'x~x,oo:93 ,.o.'x~x,oo:07 QUOTIENT:07-94-~.- 7.

EHEN5 QUOTI E NT NORMAL HIP

LKGG-IPIEICTHr..~NIP EPI PHYSEAL 72 HEAD-NECK IOO ACETABULAR 100 ACETADULUM" HEAD AVERAGE or COMPRIrH[N$1Vs 9Z7..o( NORMAl..

FIGURE 34 (A) The acetabular-head index is derivedby measuring that part of the femoral head ossific nucleus coveredby the bony acetabulum, divided by the entire width of the ossific nucleus, multiplied by 100. (B) The acetabular-head quotient (left) is determined by dividing the acetabular-head index on the Perthes side by the acetabular-head index on the normal side. The comprehensivequotient of Heyman and Herndon is shown at right. [Part A reprinted from (86), with permission; part B reprinted from (122), with permission.]

Fig. 31, both AP and lateral tracings show a spherical head in which the circular outline of the bony margin does not deviate from a marking line on the template. Good results have been seen with the epiphyseal quotient above 60%. In those graded fair, the shape of the surface of the femoral head was spherical, but the epiphysis was crescent-shaped and the deviation on the template was at 2 mm. The epiphyseal quotient was less than 60%. Cases that were poor scarcely required use of the template, but the femoral heads were nonspherical and the radius of the circle making up the outline of the head differed in the lateral and frontal views. Katz subsequently demonstrated that the epiphyseal quotient alone was the most valuable and that the composite quotient of Herndon and Heyman, even though it represented four measurements, added little. He also showed that a linear relation existed between the comprehensive and the epiphyseal quotients.

~= I I

I

8. EPIPHYSEAL EXTRUSION INDEX (GREEN,

BEAUCHAMP, AND GRIFFIN) These authors developed a new measurement of subluxation of the femoral head referred to as epiphyseal extrusion and found it to be an important predictor of prognosis in L e g g - P e r t h e s disease (94). The measurement was made on an anteroposterior hip and pelvic radiograph with the hips in neutral rotation and neutral abduction-adduction. They felt that this index could be used during the course of the disease to assess prognosis. When the epiphyseal extrusion was more than 20%, the prognosis was poor, and conversely, when it was less than 20%, the prognosis was good. Epiphyseal extrusion is a measure of the diseased ossific nucleus lateral to Perkins' line measured in millimeters along a line perpendicular to Perkins' line ( A - B ) divided by the width of the opposite normal femoral head measured in millimeters along the epiphyseal plate (Fig. 36). The quotient obtained is multiplied by 100 to obtain a percentage of the femoral head extruded from the acetabulum.

~ I .

,

FIGURE 35 The acetabular head index determinedfrom a plain radiograph measures bony parameters (left). The acetabular-head index, however, also can be measured from magnetic resonance images or from hip arthrograms in which the cartilage model of the femoral head covered by the entire cartilage model of the acetabulum is related to the entire cartilaginous diameter of the femoral head (right). [Reprintedfrom Sales de Gauzy, J., et al. (1997). J. Pediatr. Orthop. 6B:235-238, 9 Lippincott Williams & Wilkins, with permission.]

~

~{[i

\

FIGURE 36 The epiphyseal extrusion index is a measurement of subluxation of the femoral head referred to as epiphysealextrusion. It measures the involved ossific nucleus lateral to Perkins' line, divided by the width of the opposite normal femoral head measured along the epiphyseal plate, multiplied by 100. [Reprinted from (86), with permission.]

SECTION X ~ Results Based on Appearances at Skeletal Maturity

333

/

/

/ i[ I

I

FIGURE 37 The relative position of the tip of the greater trochanter and the superior articular surface of the femoral head is determined by the articulotrochanteric distance (ATD). [Reprinted from Leitch, J. M., et al. (1991). Clin. Orthop. Rel. Res. 262:178-184, 9 Lippincott Williams & Wilkins, with permission.]

9. TROCHANTERIC HEIGHT This represents the position of the tip of the greater trochanter with respect to the superior portion of the femoral head. The distance is referred to as the articulotrochanteric distance (ATD) (Fig. 37). The measurement was refined and used by Stulberg et al. (259), Sponseller et al. (257), and Leitch et al. (173) (Fig. 38).

FIGURE 38 An index of the articulotrochanteric distance (ATD) can be determined by relating the tip of the greater trochanter to the quadrant of the femoral head, divided into quarters, opposite which it lies. [Reprinted from (257), with permission.]

head in the same way as Jonsater but assesses the head deformity by measuring from the midpoint of the transverse diameter to the nearest point on the joint surface (Fig. 39) (247). The index is the value S(/ 89 Measurements can be done in both anteroposterior and lateral projections. The normal index is 1.0. In a second publication, Shigeno and Evans documented that patients with loss of sphericity had significantly lower arthrographic indices (248).

D. S t u l b e r g Classification

10. CAPUT INDEX (JONSATER) The caput index is the ratio of the height to half the greatest width of the femoral head but as determined from an arthrogram rather than from a plain radiograph (133). The greatest width of the head is measured, and from the center point of this line that connects the two points the height is measured at right angles to this line. The arthrogram thus includes the cartilage of the femoral head and is felt to be a more accurate predictor of end results because it is the cartilage model that ultimately is important. This often is shaped differently from the secondary center and initially is closer to the normal sphericity than the bone of the secondary center. The maximum diameter of the femoral head is measured (D) along with the perpendicular distance from the midpoint of this line to the joint surface (H). The index is the value H / 8 9 (Fig. 39). The normal value in a spherical head is 1.0. 11. ARTHRoGRAPHIC INDWX (SHIGENO AND EVANS) Often the subchondral bone and overlying cartilage collapse not along the perpendicular line drawn in the caput index measurement but rather anterior and superior to it such that the caput index can underrepresent the severity of focal deformity. Shigeno and Evans developed an arthrographic index that measures the transverse widest diameter of the

The long-term studies by Stulberg et al. defined five patterns of hip structure based on plain radiographic appearances at skeletal maturity following childhood Legg-Perthes disorder (259). The classification indicates the extent of hip malformation and remodeling reached at the termination of the repair phase at skeletal maturity. This classification has good

I FIGURE 39 The caput index of Jonsater, shown at left from a hip arthrogram, measures the cartilage model of the head. At right, the arthrographic index of Shigeno and Evans also uses the arthrogrambut quantifies the greatest extent of femoral head collapse, assessing head deformity by measuring from the midpoint of the transverse diameter to the nearest point on the joint surface. [Reprinted from Shigeno, Y., and Evans, G. A. (1996). J. Pediatr. Orthop. 5B:44-47, 9 Lippincott Williams & Wilkins, with permission.]

334

CHAPTER 4 ~ Le~ty--Calve--Perthes Disease

FIGURE 40 Radiographicexamples of Stulberg types I-V are shown in parts (A-E): (A) type I, post proximal femoral varus osteotomy;(B) type II, post abductionbracing with mildright coxa magna; (C) type III; (D) type IV; and (E) type V.

prognostic valueconcerning arthritic development over the subsequent decades. The types of residual appearances are graded as follows: class I, completely normal hip joint; class II, spherical femoral head associated with one or more of coxa magna (larger than normal although spherical femoral head), short neck, or dysplastic, abnormally steep acetabulum; class III, nonspherical, mushroom-shaped, ovoid head plus coxa magna, short neck, and steepened acetabulum; class IV, flat femoral head with coxa magna, short neck, and steepened acetabulum; and class V, flat femoral head of normal size with a normally shaped neck and acetabulum. Radiographic examples are shown in Fig. 40. This study remains of extreme value in assessing the long-term sequelae of Legg-Perthes disorder. The authors point out that hip deformity produced by Legg-Perthes disease clinically is problematic only when it leads to painful and disabling osteoarthritis in particular in the early to mid-adult years. They

noted that most previous studies had equated good and excellent results with the shape of the femoral head, which was graded using the criteria of Mose. Those methods equated the quality of the result with the degree of sphericity of the femoral head, with the implication that a lack of sphericity would quickly lead to early deterioration. In fact, long-term follow-up studies showed that patients even with moderately deformed hips radiographically often had relatively mild symptoms well into their adult years. 9The Stulberg classification thus concentrated on features other than sphericity alone. In this classification system, class 3, 4, and 5 included hips with poor Mose ratings; however, only the hips in class 5 were associated with the development of painful osteoarthritis in early adulthood. The hips in classes 3 and 4 were associated with a relatively benign clinical and radiographic course in spite of significant residual deformity. This study was significant not only for

SECTION X ~ Results Based on Appearances at Skeletal Maturity the relatively large number of long-term patients assessed (88:28 from the University of Iowa, 27 from the Hospital for Sick Children in Toronto, and 33 from the Karolinska Institute in Sweden) but also for the average age at followup which was 47.3 years and virtually the same in each of the three institutions. Patients thus were followed on average for three decades following skeletal maturity. The class 5 deformity usually occurred in children whose onset of symptoms was after the age of 9 years. The deformity was characterized by a flattened femoral head but a normal acetabulum and a femoral neck of normal length. Those disorders occurring relatively late lack the ability for remodeling to occur, an observation also reported by Mose et al. This was the only pattern associated with significant osteoarthritis before the age of 60 years. This hip is referred to as aspherically incongruent, which is the characteristic pattern leading to the rapid development of discomfort and arthritis. The head invariably was only partially involved and usually in the anterosuperior quadrant. Important findings of the study were that many patients with deformed hips (classes 3 and 4) do quite well and that radiographic signs of osteoarthritis begin to develop only when the patients are in their 40s and 50s. The symptoms develop more slowly in class 3 than in class 4. The age at occurrence again was noted to be a prognostic finding, with deformities in classes 3 and 4 occurring in children whose average age at the onset of symptoms was 2-3 years younger than those in class 5. The acetabulum thus had sufficient growth remaining to remodel during the active stage of the disease to conform to the deformed femoral head and produce a hip that was described as having aspherical congruity. This was determined to protect the hip from arthritis until late adulthood, allowing for essentially full function through the early and middle adult years. Most of the hips in these groups had extensive involvement of the femoral head (Catterall types III and IV) and severe metaphyseal changes. Patients in classes 1 and 2 function essentially normally throughout adulthood. These patients had the earliest age at onset of the disease, averaging 1-2 years younger than patients in classes 3 and 4. All of the hips in classes 1 and 2 had a spherical femoral head and a correspondingly shaped acetabulum, allowing them to be considered as spherically congruent. This study raised the extremely important question as to whether containment treatment was likely to alter the natural course of Perthes disease. Patients with coxa magna, which by definition represents heads not fully contained, often developed congruency with the acetabulum, which itself was dysplastic and did not go on to arthritic changes.

E. Butel, Borgi, and Oberlin Grading System Butel et al. proposed a grading system to assess the result of a Perthes disorder incorporating both clinical and radiographic parameters (30). Stage I: excellent, there were no clinical problems, the head was spherical radiologically, the quotients were normal, and the angle of Wiberg was greater than 25 ~ Stage II: good, the lower extremity length discrep-

335

ancy was 15 mm or less, there were slight alterations of hip rotation, the head was flattened between 0 and 2 mm (Mose criteria), the quotients were normal, and the angle of Wiberg was 25 ~ or more. Stage III: average, there was occasional pain, the length discrepancy was 20 mm or less, there was a slight limitation in mobility, the head was flattened between 0 and 2 mm radiographically (Mose criteria), and the quotients were slightly abnormal with the angle of Wiberg between 20 ~ and 25 ~. Stage IV: mediocre, there was some limping or occasional discomfort, the length discrepancy was greater than 20 mm, there was a distinct limitation of mobility, head flattening was between 2 and 4 mm (Mose criteria), and the angle of Wiberg was less than 20 ~ Stage V: poor, limping and discomfort were frequent, there was significant hip stiffness, the length discrepancy was greater than 20 mm, head flattening and irregularity by the Mose criteria were greater than 4 mm, and the angle of Wiberg was less than 15 ~

F. Additional Long-Term Studies of Adult Responses to a Childhood Perthes Disorder Long-term effects were assessed by McAndrew and Weinstein, who were able to follow 37 affected hips for an average period of 47.7 years (183). They began to note meaningful degenerative changes judged by the increasing number of hip arthroplasties done only in the fifth or sixth decade of life. Their 1984 report was based on a retrospective review of patients diagnosed between 1920 and 1940. Of 112 patients seen initially, data were available on 35 with 37 affected hips. In previous long-term studies of the same group of patients, Gower and Johnston concluded that 86% of hips were still functioning well in patients at an average age of 45 years, but in the same institution, with the patients at an average age of 56 years, many more had come to arthroplasty and only 40% had maintained a good level of function (93). McAndrew and Weinstein also noted a strong statistical correlation between a decrease in ratings over the follow-up period and the age at onset of the disease. The age at onset had a significant beating on the long-term outcome, with younger patients having the better outcome. In their study, the dividing line between better and poorer prognoses appeared to be at the age of 8 years. Eaton reviewed 100 hips 10-45 years after therapy (65). All were treated by non-weight-bearing in recumbency followed in some by the ischial weight beating brace. Hip containment as a principle of treatment was not used. He documented 64% good-excellent results, 17% fair, and 19% poor. Prolonged bed rest did not lead to improved results. The earlier the age of onset, the better the result. Symptomatically, the hips did not appear to worsen in the 20- and 30-year follow-ups. Ratliff reviewed 34 hips observed for an average of 30 years with a range from 25 to 40 years (225). Twenty-five of the hips had been treated by immobilization on frames and 1 in a hip spica for an average period of 18 months (6-24 months), whereas 8 patients received no treatment during the

336

CHAPTER 4 ~ Le~t~t--Calve--Perthes Disease

active phase of the disease. The average age of the patients was 38 years and the average follow-up was 30 years. Ratliff determined that the hips remained good in 15 patients, fair in 11, and poor to very poor in 8. He felt that "with rare exceptions, the clinical and radiographic condition of these hips in 1965 was not changed from that of 1953." In a functional sense, the patients were doing even better, with 30 of the 34 reporting either no pain or only slight aching and activity being normal in 29 of 34. He indicated in his general assessment that treatment was worthwhile. Deterioration in the condition of the hip was rare up to 40 years of age. He felt his findings to be similar to those of Danielsson and Hernborg, who examined patients 33 years after onset and concluded that 28 of the 33 had painless hips. Perpich et al. also observed that patients followed into their fifth and sixth decades of age generally did well but that worsening of osteoarthritic status continued in a gradual fashion from that time onward (209). In referring to their own work plus that of others, the average length of followup was correlated with severe degenerative joint disease (DJD) as follows: average follow-up 27 years (32 patients), 6% DJD; 29 years (40 patients), 12%; 33 years (35 patients), 20%; 40 years (88 patients), 25%; and 57 years (16 patients), 85%. Catterall reviewed a subset of adolescent patients with Perthes who were developing hip pain. Most of these displayed flattening of the superolateral femoral head with hinge abduction (46). He felt that approximately 10% of patients with Perthes developed deterioration, needing surgery for pain and limp before 35 years of age. Studies from Birmingham, England, by Clarke and Harrison in which a large group of patients had been followed for several decades did, however, show a significant subset of patients with Perthes having the painful sequelae of coxa plana in mid-adult life (51). This study assessed 31 patients, although they were assessing those who were presenting because of the discomfort. Patients were seen who were under 20 years of age, between 20 and 30 years old, and greater than 30 years old, with each group having roughly the same number of patients. Characteristic of the radiographic changes were the mushroom-shaped deformity of the femoral head and premature epiphyseal fusion, with frank osteoarthritis noted particularly in those older than 30 years of age. Another study showing an increase of symptomatic hips in mid-adult life was reported by Yrjonen (289). He reported on 106 hips who had conservative noncontainment treatment and were studied after an average of 35 years, ranging from 28 to 47 years. The radiographic result by fairly strict criteria was poor at skeletal maturity in 65 hips. At an average age of 43 years, the radiographic signs of osteoarthritis were found in 51 of 106 (48%). By that age 5 patients had already had hip arthroplasty, with 13 more having symptoms that would justify recommendation of the procedure. As in virtually all other series it was the patient's age at diagnosis and the shape of the femoral head at skeletal maturity that were the best prognostic factors.

In the following section, treatment approaches and results will be presented in great detail. This section indicates, however, that the results noted clinically for a period of 3-5 decades appeared to be reasonably good regardless of type of treatment and even in the presence of less than anatomically or radiographically normal hips. These findings made treatment programs even more difficult to outline because a "perfect" anatomic result did not seem mandatory or routinely attainable. Precautionary warnings thus have accompanied many reviews of the Legg-Perthes entity that radical treatments could be considered as "exposing the patient to an unjustifiable risk." McKibbin noted that many of the then current reports showed "the small differences which could be discerned between the treated and untreated groups" (187). He also cautioned in 1975 that "when we are assessing the methods of treatment, it must be constantly born in mind that a high proportion of satisfactory results can be anticipated with no treatment at all and the differences to be expected from treatment are probably small and must be considered against the background of the natural history." As the following section will indicate, even today extensive efforts persist to define which group would benefit most from intervention and which type of intervention, surgical or nonoperative, is most effective.

XI. T R E A T M E N T A P P R O A C H E S T O

LEGG-PERTHES DISEASE Treatment approaches have evolved in a gradual and rather uncertain fashion. Specific therapy has been impeded by a lack of knowledge as to the specific etiology, the prolonged several-year period during which healing occurs, and the variable but often similar results that occur with markedly different therapies. We will review the earliest detailed studies of Flemming Moiler (79) and Sundt (261) initially and then proceed to subsequent approaches. For several decades, therapy has been based on three generally intuitive approaches: (1) non-weight-bearing; (2) hastening bone repair of the secondary ossification center; and (3) femoral head containment with or without weight bearing. Specific reviews commenting on various aspects of treatment are most helpful in developing an overview of this complicated area (43, 119, 120, 176, 243,283).

A. Early Major Reviews of Treatment Approaches Sundt reviewed his own and other studies presented up to 1949 and concluded with many that therapy was contributing very little to any improvement beyond that of natural history (261). The investigators most responsible for defining the entity, Calve, Legg, Perthes, and Waldenstrom, each felt that the therapies used initially had little to no influence on the course of the disorder. Early treatment in the first half of the twentieth century was based on the assumption that

SECTION Xl ~ Treatment Approaches to Legg--Perthes Disease

the femoral head was soft and thus predisposed to deformity with weight bearing. The treatment, therefore, involved varying forms and varying time periods of immobilization with avoidance of weight bearing. The early studies impressed Sundt that treatment by immobilization appeared to have no influence as to whether the final result was good. He had concluded, on the basis of his early studies, that "the pathological changes in the diseased hip passed through the same phases of development and with the same rapidity in the treated as in the untreated cases." He commented that "a completely nihilistic attitude towards a routine systematic treatment of the disease" had developed. 1, FLEMMING MOLLER Flemming Moller presented one of the earliest and most detailed studies of the results of a Legg-Perthes disorder in 1926 discussing 74 cases, 35 of these being his own (79). He confirmed the findings of Schwarz, who with Perthes performed the earliest studies on the disorder. The disease could develop such that there was almost normal structure of the hip with healing, although some patients develop considerable clinical problems characterized by limping and decreased range of motion of the joint. Flemming Moiler also noted that the "results are by no means always as ideal as might be desired." The detailed clinical assessment involved the study of flexion, extension, abduction, adduction, medial and lateral rotation, gait, the presence or absence of pain, shortening, and the Trendelenburg sign. In his 35 patients, there was no shortening documented in 18, whereas in the others the shortening ranged from 0.5 to 2.0 cm. When the result was clinically imperfect at skeletal maturity, it was restricted mobility or decreased range of motion of the joint that was most prominent. The most prominent limitation was found with abduction in virtually all cases, although most also had some diminution of flexion. Internal rotation was greatly restricted. Flemming Moiler felt that shortening was almost an invariable occurrence and the gait almost always was abnormal. He presented accurate drawings from the radiographs that clearly depict the shape of the femoral head, the trochanters, and the adjacent acetabulum in each instance. A series of drawings for each patient shows the evolution of the shaping changes with time. In most instances, there are four drawings provided per patient. The drawings depict the final result in a much clearer fashion than verbal descriptions or various measured indices. Virtually the entire spectrum of structural abnormalities that came to be better defined over the subsequent decades was illustrated here. Even at this early time period, it was stated that "the earlier in life a child gets the Calve-Perthes disease, the more favorably and with less resulting deformity will the process heal." The converse also was stated that, when Perthes developed around the age of puberty, there would not be sufficient time for the process to heal completely. For the perfectly healed cases, the average age at the beginning of the disease was 6 years, in the cases healing with slight restriction it was 9.2 years, and in the poor cases the average age at the beginning of the disease

337

was 12 years. "It is thus in reality the age of the child that is determining the extent of the final deformity." Studies of patients in their late 20s and early 30s led him to state that "the deformities resulting from Calve-Perthes disease afford a strong disposition to the development of a subsequent arthritis deformans." He summarized his work by indicating that approximately three-fourths of the patients healed with perfectly good restoration of hip function. The results described were clinical in nature, with 78.4% with no noticeable clinical changes beyond a very slight limp and 21.6% in a second group with movement at the hip considerably restricted as well as a permanent limp. Of these, 12% of the total already had a limp, continual hip pain, and decreased work capacity due to restricted mobility at the joint. The potential for long-term arthritis clearly was established in the markedly deformed heads. He concluded that "we cannot by any kind of physical therapy of the hip correct either the severe or the less severe deformities which occur in the different cases." He therefore recommended as little active therapy as possible, confining oneself to symptomatic treatment. 2. SuNI)X Sundt felt on reviewing the work of Danforth and Gill (88), the strongest advocates of complete non-weight-bearing, that their results did not appear different from those commonly found either in cases treated symptomatically only or in wholly untreated cases (261). He reviewed his 172 hips using his four-fold classification listed earlier in terms of final result categorization. He observed that the younger the patient, the better the prognosis for Legg-Perthes disease as the deformity of the head would be less pronounced. Of 93 hips with onset of the disorder before the 8th year of life, 50 were treated and 43 were untreated. In 60 cases, the disease appeared after the 8th year of life with 38 treated and 22 untreated. Some of his patients had been treated by immobilization with partial confinement to bed during an average period of 2 years. In other cases treatment had been discontinued when the diagnosis was confirmed not to be tuberculosis; and in others symptomatic treatment only had been instituted. All of these were considered in the treatment group, and the average duration of patient immobilization was 5.1 months. Eighty-eight patients had been treated and 65 untreated, but no case had been treated systematically the whole time from the beginning of the disease to its termination, by which he meant that he did not approach the 4 years of treatment done in other programs. He agreed in a great majority of cases that "the disease leads to deformation of the head." There were only 11 cases with no deformation, with the head being of the normal spherical shape (category I). In 56 cases, the head was oval or ovoid (category II), in 78 cases it was cylindrical (category III), and in 8 cases it was square or cuboid (category IV). The younger the patient, the more favorable the prognosis whether the patient had been treated or not. His end results actually showed better results with nontreated rather than with treated cases. The

CHAPTER 4 9 Legy-Calve-Perthes Disease

338

TABLE I Results in Perthes Disease in Treated a n d U n t r e a t e d Patients (Sundt)

Grade of result Group I (before 8th Year) Treated types Untreated

I and II

Percentage 42%

III and IV

58%

I and II

48%

III and IV

52%

Group II (after 8th Year) Treated types Untreated

I and II

42.1%

III and IV

57.9%

I and II

45.5%

III and IV

54.5%

"result for the untreated joints in both age groups is more favorable than for the treated cases," with similar favorable results in categories I and II occurring in either approach as well as similar results in categories III and IV. He indicated that the results "made distinctly evident that the treatment has nothing to say in this connection." The results from his work are listed in Table I. Categories I and II represent favorable results and III and IV unfavorable results. Sundt indicated that any treatment that involved avoidance of weight bearing or complete immobilization would have to start as early as possible to be of any value. If there was an altered shape of the femoral head radiographically, from that point on the treatment would be fruitless. His results showed this to his satisfaction, indicating that the most favorable results were seen in the group of patients in which treatment was begun in the first 6 months of the onset of the disorder. He concluded that "the deformation of head and neck proceeds on the same lines in all cases, treated as well as entirely untreated, and whether the treatment has consisted in complete avoidance of weight bearing or (and) in immobilization." Sundt also considered that observation to be correct in his assessment of the serial radiologic investigations reported by active therapy proponents such as Danforth, Gill, and others. Published radiographs showed that changes in shape continued to progress during rest, treatment, or immobilization and often reached extreme degrees of flattening and fragmentation, an observation also made in his own material. Sundt made the interesting observation that, although one of the end results frequently is reported as a coxa vara, in many instances that appearance is due to the overgrowth of the greater trochanter, whereas if one simply looks at the angle of the head-neck in relation to the shaft, generally there is a valgus deformation due to earlier closure of the physis laterally with continued growth medially. He noted no difference with regard to the size of this angle between

treated and untreated cases and pointed out that the disease did not lead to true coxa vara but rather to coxa valga whether treated or untreated. Coxa valga of the head in relation to the neck (and not considering the greater trochanter) tended to be produced in all cases except 9 in which there was a slight coxa vara. Subluxation of the upper end of the femur was a radiologic finding seen very frequently. At the end stage, there was subluxation in 80 hips and no subluxation in 66 with treatment showing no positive influence; the treated patients had more subluxation than the untreated. Sundt performed a detailed assessment of end stage and nontreated patients in relation to several criteria involving discomfort, limp, and a positive Trendelenburg test. In those patients who had pain, there were more treated than untreated patients. Treatment did not prevent the occurrence of limping, treatment had no positive influence on diminishing the Trendelenburg gait, and the range of motion of the hip also was not markedly different in the two groups. The major long-term criterion of effective treatment, however, would involve the spherical shape of the head following treatment, but even here he concluded that "the treatment of L e g g Perthes has had no influence as regards the final form of the head." Severe deformation of the head occurred in 12.4% of his cases, quite similar to Flemming Moller's findings. Sundt concluded "accordingly, it appears the treatment had no influence as regards the shape of the head at the termination of the primary phase of the disease." The secondary osteoarthritis (arthrosis deformans coxae) was "first and foremost due to incongruence of the articular surfaces in the acetabular cavity on account of the deformation of the head and the danger of such arthrosis is therefore greatest where the deformation is most pronounced." He concluded his extensive investigation with the following words: "My investigations, therefore, fully support the conclusion reached by Flemming Moller, namely that Legg-Perthes is not the 'innocent disease' it was first supposed to be, but they also show that measures of treatment fail to prevent the occurrence of the secondary arthrosis which is the cause of the troubles that may follow in the wake of the disease." 3. DIVERSITY OF OPINION AFTER SEVERAL DECADES CONCERNING VALUE OF TREATMENT

Unlike most disorders, the effect of treatment on the end result often is unclear with Legg-Perthes disease. Many patients have done extremely well in terms of hip development with minimal or no treatment, and other patients treated over a several-month to few-year period have had poor results. An enormous amount of effort thus has been expended by the orthopedic profession to define which patients do well with minimal to no treatment, and, when treatment appears needed, studies have attempted to define not only the best treatment but also the appropriate time of intervention and length of treatment as well. As noted earlier, in the first couple of decades after delineation of the entity, each of the four major physicians initially involved with the disorder felt that

SECTION Xi ~ Treatment Approaches to Leoo--Perthes Disease treatment interventions had been ineffective in influencing the final outcome. The clear abnormality of the hip both in a clinical and certainly in a radiologic sense has always influenced management because very few physicians and very few families were comfortable allowing the changes to run their course without some attempts at improving the child's fate. Thus, it is very difficult to determine a true natural history result with the disorder. A detailed review of the literature, particularly those studies in the early and mid decades of the century, does turn up an appreciable number of instances in which no treatment was performed. Although the bias of the reporting physician must be taken into consideration, it truly does appear that simply advocating no treatment at all to any patient with Legg-Perthes disease would tend to place the intermediate and long-term results into the poorest range. The paper by Sundt referred to in the preceding section actually concluded that, by a small margin, those patients who had no treatment were slightly better in terms of intermediate result than those having treatment. It is important to recognize, however, that in the "treatment" group he included all forms of management, even those consisting of only a few weeks of non-weight-bearing, although he indicated that the average period of immobilization was approximately 5 months. Even so, 42-45% of patients had a favorable radiographic result in his grades I and II, with again a few percentage point advantage to those untreated. These numbers highlight the problem of interpretation of results because almost one-half of the patients did quite reasonably well in the absence of intervention. The percentages of good results in nontreated patients are reasonably high in other series, with Ratliff documenting 33% (225), Danielson 40%, and in a group reviewed by Catterall as high as 51% (41, 43). The work of Meyer perhaps best defined the relatively poor results with no or only symptomatic treatment (190, 191). He concluded that the radiological results improved as the efficiency of relief from weight bearing increased. This was determined by the degree of sphericity of the head at and beyond skeletal maturity. Only 19% of patients with no or only symptomatic treatment had spherical heads, and this increased to 61% with ambulatory relief from weight bearing, 73% with bed rest without traction, and 87% with bed rest with traction. Mose et al. reported on a series of patients studied at an average age of 65 years who had not had any primary treatment and noted a high frequency of irregularly shaped heads radiographically and an incidence of osteoarthritis of greater than 85% (197). This represents a greater than 50-year follow-up and certainly is greater than most other groups having had a coherent treatment program, but nevertheless these numbers still appear higher than those expected from treatment groups. Thus, there is some reasonable, although far from definitive, investigational evidence that treatment is beneficial in general over the broad spectrum of the Perthes disorder. It thus appears reasonable to continue to assess the specific groups benefiting from treat-

339

ment, the specific time to institute the treatment, the treatment itself, and the length of time to continue the treatment. It thus is not surprising that, even into more recent times, the concept of no or minimal treatment has been considered and practiced by some. A fairly recent study by Norlin et al., referred to in more detail later, concluded that a passive form of management probably was less favorable than treatment approaches but nevertheless led to a relatively large number of reasonably good results (203).

B. Range of Approaches to the Disorder The early but highly detailed studies showed the relative ineffectiveness of treatment regimens adopted (almost always immobilization regimens) in the sense that there did not appear to be a direct definable correlation between a specific therapy and a good result. Since that time, six basic approaches, for the most part intuitive rather than scientific, have directed management for Legg-Perthes disease. The approaches are categorized as (1) nontreatment, (2) symptomatic treatment only, (3) non-weight bearing, (4) containment of the femoral head, (5) weight beating, and (6) hastening of bone repair by surgical intervention on the secondary ossification center. These are outlined in Table II. Two of these treatment principles may of necessity be combined, for example, containment therapy is done either with weight bearing or with a non-weight-beating approach. A brief summary of each follows. 1. No TREATMENT There are some types of patients who can be managed with no treatment or observation only. Those fitting into this category are very young children 3, 4, and occasionally 5 years of age, especially because long-term results are most favorable in terms of remodeling potential in those youngest at the time of disease occurrence. The other group of patients seemingly requiring little to no treatment are those with the "anterior" head only involvement as defined by O'Garra (205) or by the Catterall group I designation. Often the age and partial involvement criteria are present simultaneously. Strict adherence to this approach entails no treatment. If a child with one or both of these criteria is asymptomatic, having no pain and a persisting full range of movement of the hip in spite of continued activity, many physicians recommend no intervention but only continuing careful followup. Others extend the no treatment regimen to older and more radiologically involved patients as long as a full and painless range of motion of the hip persists, although that situation is seen infrequently. Although most patients in the youngest age group do well, it is essential to note that fair to poor results are seen on occasion in this age group even with treatment. As best we can determine, the development of symptoms or the demonstration of shaping abnormalities of the articular cartilage surface by arthrography or magnetic resonance imaging should shift a patient to a treatment group.

340

CHAPTER 4 ~ Legg-Calve-Perthes Disease TABLE II

Spectrum o f A p p r o a c h e s to L e g g - P e r t h e s T h e r a p y

Approach

Rationale

1. Nontreatment 2. Symptomatic treatment only 3. Non-weight bearing-noncontainment A. Recumbency Bed rest with or without traction, abduction splint, or hip spica cast B. Ambulatory Crutch, sling, ischial weight bearing brace 4. Containment A. Non-weight beating Recumbency-abduction casts, braces, or frame Ambulatory, unilateral abduction-internal rotation-flexion brace. B. Weight beating Abduction plaster Abduction brace Proximal femoral osteotomy Innominate osteotomy 5. Weight bearing With containment, see 4B With no containment, see 1 or 2 6. Hastening of bone repair of secondary ossification center Drilling Drilling and bone grafting Resection of necrotic bone Resection of necrotic bone and bone grafting

Treatment of no demonstrated value. Treatment not needed or not effective. Malformation of head is primarily on a mechanical basis; head soft during disease process and weight beating causes deformity.

Malformation of head occurs primarily on a biological basis due to asymmetric growth and repair. Head not physically soft; containment within acetabulum allows latter to mold repair process leaving head round; weight bearing with head contained does not cause deformity but rather enhances sphericity. Incorporates both mechanical and biological principles of deformation.

Removal of necrotic bone with or without addition of graft quickens repair response and minimizes excesses of repair on cartilage model or head.

,

The nontreatment approach, regardless of the various categorizations, occasionally is adopted by physicians unimpressed by results from the various treatment regimens used. 2. SYMPTOMATIC TREATMENT ONLY FOR PAIN AND DECREASED MOTION

Rest remains the primary approach to a symptomatic hip. This can be accomplished by a series of methods involving bed rest with or without traction or, if the child is reasonably cooperative, the use of crutches with the involved side either non-weight bearing or partially weight bearing. Such measures generally result in a decrease in the associated synovitis. The early pain also is felt by many to occur in relation to the subchondral fracture through necrotic bone. Decreased use of the hip during this phase of the disorder would diminish pain. It also conceivably diminishes collapse of the femoral head, but no controlled studies of this specific point have been made. Some practitioners treat only the symptomatic hip, often on two or more occasions, to minimize or elimi-

nate discomfort and rehabilitate one to an improved range of motion. Once comfort and motion have been regained, unprotected weight bearing resumes, again with the feeling that treatment regimens have not proven themselves of definitive merit in improving the long-term structure of the hip. Strict adherence to this approach entails symptomatic treatment only regardless of the required frequency of rest periods. Those physicians proceeding with treatment also will use rest and non-weight bearing as an initial management both for comfort of the child and for awareness that the brace or surgical therapies cannot be done as effectively unless a pain-free hip with a normal or at least virtually normal range of motion has been achieved. A comfortable range of motion is better for the patient in general and also allows for appropriate seating of the head in relation to the acetabulum. If subsequent treatments, whether they be operative or nonoperative, are planned, the results appear much better if they are performed in the presence of a nonsymptomatic or mini-

SECTION Xl ~ Treatment Approaches to Legg-Perthes Disease

mally symptomatic hip with as free a range of motion as possible. Mose reviewed several studies in 1964; 14 papers reported that in cases with no treatment less than 20% had good results, with more than 60% poor results (196). More recently, Norlin et al. responded to the observation that their previous treatment protocols seemed to lead to extremely variable results by following 20 patients whose only active treatment involved management of severe pain by bed rest and traction for a few days, following which they were "told to live as normally as possible" (203). The average age at onset was 8 years (range = 3.9-15.4 years), and the age at follow-up was 22.4 years on average (range = 17.328.1 years). Subjective symptoms were assessed by the Iowa hip score system and radiographs were assessed using the Mose criteria. On the Iowa hip scale, 2 of the 20 scored a total of 100 points with the average score being 90 (range 74-100). Thirteen of the 20 patients had full function, but only 6 of the 20 had no pain. Norlin et al. concluded that their passive form of management was "probably less favorable compared to others" in the sense that they felt they had used the terms good, fair, and poor as in other papers. They noted the total absence of a spherical head at follow-up in any of the patients. They also noted the absence of correlation between the clinical outcome and the primary Catterall classification. Although this study was far from definitive, and although the authors suggested that active treatment was preferable, there still were a relatively large number of untreated patients with surprisingly good Iowa hip scores and radiographic assessments. Even those who take an active interventional approach to the treatment of Legg-Perthes disease often will opt for this category or the no treatment category of management if the patient is 5 years of age or under or if he or she appears to be maintaining the disorder in the Catterall group I. In virtually all studies, the Catterall group I patients do well and often with minimal to no treatment. On the other hand, even though most series have commented that the younger patient does better, it is important to recognize that one cannot automatically assume that those 3, 4, or 5 years of age at disease onset will proceed to a good or excellent result with minimal to no intervention. It is essential to individualize treatment for particular patients. Snyder studied 31 patients with 40 involved hips in whom Legg-Perthes developed when they were less than 5 years of age (253). He commented that Ralston (224) had noted no correlation of results with age at onset and that other studies, though showing better results in the younger group, still noted that a significant subset had fair and on occasion poor results. Snyder also, in a relatively small group of patients, noted better resuits in the younger group but commented that "there were significant numbers who did not do well." He went on to stress that it was important not to incorrectly interpret the fact that the younger patient does better to mean that all

341

young patients with the disease will do well. He also made the point that it is virtually impossible to compare differing series because a wide variety of radiographic measurements tend to be used. An individual grading of a certain deformity as fair or poor often does not translate either in the young or in the middle-aged into a clinically problematic result. Snyder listed nine reports documenting better results in the young. These include Broder with fair and poor results in 45% overall but only 19% in those under age 5 years at disease onset (27), Eaton with 36% and 0% (65), Evans with 31% and 13% (70), and Katz with 13% and 6.5% (142). Gossling documented that 96% of patients with poor results were 6 years old or more at onset of the disease, Fleming (92) and Herndon and Heyman (115) indicated that the only factor of real prognostic significance was age at onset, Moiler reported poor results with average age of onset of 12 years and good results at 6 years (79). Salter indicated that patients with disease onset at age 4 years or less required no treatment because the results were quite good (238, 240). Finally, Sundt stated that the younger the child, the better the prognosis (261). 3. NON-WEIGHT BEARING APPROACH-NO PARTICULAR ATTENTION TO FEMORAL HEAD POSITION a. General Review The non-weight bearing approach was based on the intuitive feeling that the femoral head healed in a misshapen fashion because the patients had been allowed to bear weight on the hip when it was in a softened state due to the disease process; collapse of the spherical head and its repair in a deformed shape occurred due to mechanical reasons. It was felt, therefore, that the prevention of weight bearing during the prolonged repair phase would maintain femoral head sphericity and lead to a good longterm result. The non-weight bearing treatment could be either (1) nonambulatory by prolonged recumbency with or without traction and occasionally using hip spica or (2) ambulatory using a sling, crutches, or an ischial weight bearing brace on the affected side but with no attempt to specifically position the femoral head in relation to the acetabulum. Immobilization was a commonly used approach in the first several decades following recognition of the disorder. Immobilization was attempted with bed rest and, on occasion, with casting for what we would now consider heroic periods of time varying from 1 to 4 years. Brailsford stated that serial X rays showed definite proof of the benefit of prolonged immobilization and indicated that the bone of the affected joint was plastic and could be deformed by pressure. If the joint was immobilized during the plastic stage, although the series of changes was uninterrupted, organization of the affected bone proceeded and no deformity occurred (26). He defined radiographic stages in the full evolution of the disorder. The plastic stage during which modeling changes continued to occur could last as long as 4 years. Brailsford concluded from his extensive studies that "the best results

342

CHAPTER 4 ~ Le99--Calve--Perrhes Disease

will be obtained in those cases which are recognized early, and kept immobilized for as long a period as the radiographic appearance suggests plasticity of bonemeven to the length of 4 years and that there is little justification for immobilization during only a part of the time that the bone shows plasticity." He subsequently confirmed this view in 50 of 71 cases. Eyre-Brook (72), Danforth (57), Gill (88), and others supported the systematic avoidance of weight bearing until the bone had regained its normal structure, even if this took 18-24 months or longer. Danforth reported on non-weightbearing treatment extending for 3 - 4 years (57), Pike (217) for almost 5 years, Edgren (68) a mean of 28 months, Goff (91) a mean period of 27 months, Evans (70) a mean of 25 months, and Pedersen and McCarroll (209) 18 months. Although Danforth had stimulated interest in the United States in treating Legg-Perthes disease by bed rest or without weight beating, his 1934 report assessed only 3 patients. A much greater stimulus to utilization of the non-weight bearing treatment came from Gill, whose detailed report presented in 1940 assessed 20 patients (88). Gill utilized prolonged rest in bed using Buck's extension on both legs until regeneration was well advanced as shown radiologically. Following that, the child was allowed to walk using a walking brace and crutches. He indicated that "when we are able to carry out this method without interruption, the end results are practically perfect hips." He then presented sequential X rays from several cases in which the results indeed were excellent. It is interesting to note, however, that 10 of his illustrated cases had an average age of occurrence of 5.5 years with a range from 3.5 to 9 years. Most now recognize that with such a young group of patients results would tend to a more favorable range regardless of treatment. A poor result was shown bilaterally in a case that occurred in a male at 10 years of age who had no treatment. In his 20 charted cases of Legg-Perthes disease, the average age of occurrence was 6.3 years with a range from 3.5 to 12 years. Upon review of the X rays it is not surprising that Gill concluded that "treatment by rest in bed with hip extension leads to such complete restoration that in the final result, it is difficult to detect any deformity of head, neck, and acetabulum." On the other hand, the fact that patients in the entire series averaged only 6.3 years of age and those presented in detail only 5.5 years subsequently led others including Sundt to conclude that the results really were not fully attributable to the non-weight-bearing therapy itself. The study, however, carded great weight in the orthopedic community certainly in the United States over the next several years. Well into the 1960s, the major treatment approach was non-weight beating based on the premise that all or part of the head was structurally softened in Legg-Perthes disease such that weight bearing would lead to a misshapen head at skeletal maturity. In the most extreme forms, the patients were kept in bed away from weight bearing for prolonged periods as long as 4 years, although in most instances, after still lengthy periods of time between 1 and 2 years, the pa-

tients were switched to braces. These unilateral braces still attempted to maintain a non-weight bearing status as they were abduction braces with weight borne on the ischium. Some of the more detailed studies of non-weight bearing management follow.

b. Long-Term Recumbency with Traction (Eyre-Brook) One of the earliest detailed studies of Legg-Perthes disease and the use of non-weight bearing therapy by traction in recumbency was by Eyre-Brook (72). He clearly stated his reason for preferring the non-weight bearing approach, indicating "I picture each head as being in a softened condition and suffering progressive deformity from stresses which are greater with increasing years, and I look upon the best treatment as that which will completely rid the femoral head of all stresses and allow uninterrupted consolidation and natural growth. The treatment must relieve the head of all crushing forces, not only those of weight bearing, but also those of muscular contraction." In terms of the recumbent or bed therapy, one could use a plaster hip spica, special types of frames, or simply sliding traction in bed. Eyre-Brook favored the latter. In terms of ambulatory treatments, he referred to the weight relieving caliper, the patten-ended caliper, and the use of either with crutches. Although these reduced weight bearing, they did not protect the femoral head from all crushing forces in that muscle force across the joint still was quite strong. He thus treated patients on bed rest using sliding traction with a 6- to 10-1b weight and pulley and the foot of the bed raised. No splinting was used and the hip was treated in extension. The average duration of treatment was from 18 to 24 months with longer periods sometimes needed. Eyre-Brook categorized his cases into those under and over 7 years of age and also into three stages depending on the original condition of the head radiographically at the inception of treatment. In stage 1, there was a femoral epiphysis with little or no flattening, in stage 2, typical flattening, fragmentation, and cervical thickening, and stage 3 showed more advanced changes with recalcification. He studied the end results using a quotient he developed and referred to as the epiphyseal index. The height of the epiphysis over the breadth of the epiphysis • 100 on an anteroposterior radiograph gives the epiphyseal index. The normal index under 7 years of age ranged between 45 and 55, whereas over 7 years it ranged between 35 and 45. He developed principles that appear valid today, namely, that prognosis was dependent upon (1) the age of the child, which controls both the severity of the natural course of the disease and the stresses to which the diseased femoral head, by which is meant the ossification center, is subjected, (2) the stage to which the disease had progressed when treatment began; and ultimately (3) the effectiveness of treatment. He felt that traction and recumbency provided the most satisfactory treatment in particular for those under 7 years of age, or older than this when the disorder was diagnosed early.

c. Long-Term Recumbency with Abduction Splint (Pike and Gossling) The conservative treatment utilized by Pike

SECTION Xl 9 Treatment Approaches to Legg--Perthes Disease

at Newington, CT involved recumbency carded out over a prolonged period of time until the entire stage of regeneration of the head was complete as shown radiographically (217). It was felt that the complete avoidance of weight bearing was essential to allow the head of the femur to assume a natural and normal state. The child initially was placed on bed rest to allow pain and spasm to subside. The recumbency was accompanied by use of an abduction splint worn night and day, although the abduction was not carried to the extreme and appeared to be in the range of 30-40 ~ bilaterally. Mobility was allowed on wheeled sleds. The average length of recumbency was 27 months in 29 patients assessed in detail. The study indicated 10 excellent and 14 good results (83% excellent and good), with 3 fair, 1 poor, and 1 unclassified, leaving 17% fair and poor results. This regimen was accompanied by full-time schooling in the facility and daily physical therapy to provide range of motion and muscle toning exercises. Gossling reviewed additional patients from the Newington facility treated in the same way at a later date (92). He assessed 109 hips from 96 patients. The good/fair/poor percentages in 52 hips followed for 4.5 years were 73%/13.5%/ 13.5% and in 57 hips followed for an average of 24 years 57.8%/19.3%/22.9%. d. Non-Weight Bearing Treatment with Hip Spica Cast (Perpich et aL) Perpich et al. reviewed 40 patients (41 hips) treated between 1937 and 1958 in recumbency and nonweight bearing, with hip spica casts (210). No specific attempt to position the hip into a fully contained position was made. The time in plaster averaged 11.8 months, with 1 week out of the cast for physical therapy every 3 months. The evaluations occurred at an average of 30 years posttreatment (range = 14-40 years). The clinical gradings were 34 good (83%), 2 fair (5%), and 5 poor (12%). The radiologic gradings were 13 good (33%), 9 fair (23%), and 17 poor (43%). Perpich et al. concluded that reasonable results were obtained with non-weight bearing. e. Bed Rest and Skin Traction followed by Ambulatory Treatment in an Ischial Weight Bearing Brace (O'Hara et al.) There were 52 hips available for assessment in 46 patients (206). Forty of the 46 were treated with bed rest followed by ischial weight bearing, with the others having variable methods. The average time in brace was extensive, with those in the Catterall group I averaging 22 months, group II 24 months, and groups III and IV 27 months. The anatomic results throughout the series showed 33% (17) good, 37% (19) fair, and 31% (16) poor. When assessments ranged from Catterall I through IV, the best results showed 71% good in Catterall group I and the worst results were 58% poor in groups III and IV. Much as in other series, good anatomical results correlated with earlier age of onset, less quantity of head involvement, and lack of lateral subluxation. The authors compared their results to historical series and came to the conclusion that treatment with the ischial weight bearing brace and patten extension appeared to pro-

343

duce results worse than no treatment at all. Using the good/ fair/poor triad, the numbers in this series were 33%/37%/ 31%. Similar percentages of good results in nontreated patients were noted by Danielson at 40% and Ratliff at 33%. In Catterall's review of 46 hips without treatment, the percentages were 51%/24%/17%. f. Bed Rest Followed by Weight Relieving Caliper (Evans) Evans studied 52 hips with Legg-Perthes disease in which the initial treatment was bed rest, which varied between 2 and 24 months and averaged 10 months, followed by use of a weight relieving caliper for an average period of 15 months (70). Overall treatment therefore was approximately 25 months. Follow-up varied from 10 to 36 years. Much of the emphasis in determining the result was on the radiographic appearance of the head. By using these criteria, the results were good in 15 cases, fair in 21 cases, and poor in 16 cases. Evans focused on the age at onset of the disease as being important, with good results generally not occurring with an onset over 6 years of age and the result always being poor if the age at onset was over 8 years. Girls tended to a slightly lower percentage of good results than boys. Evans did report, however, that except with the worst shaped heads function was excellent. g. Treatment by Harness or Sling and Crutches (Kelly et al.) This study of 80 hips, evaluated at an average followup of 22.4 years, was designed to determine the long-term results in patients treated with a weight relieving sling or harness and crutches, but not according to the concepts of containment (148). Their results showed good responses in 64 patients, fair in 9, and poor in 7. Of 58 patients with the Catterall grouping III or IV, 42 had good results. The Catterall classification could not be applied accurately for an average of 8 months after onset of disease. Kelly et al. concluded that the vast majority of patients can be treated successfully by enforced relief from weight beating on the involved extremity without any attempt at containment of the femoral head within the acetabulum. 4. FEMORAL HEAD CONTAINMENT TREATMENT WITH NON-WEIGHT BEARING

a. General Review The next approach to Legg-Perthes therapy, referred to as the containment approach, is based on the premise that remodeling of the head to maintain sphericity would be favored by placing the head most deeply and concentrically into the acetabulum at all times to allow the acetabulum to serve as a template controlling tissue deposition and shaping. Because the serious long-term sequelae of Legg-Calve-Perthes disease involve either a flattened head (coxa plana) or an enlarged head (coxa magna) with deformity, physicians have attempted to position the femoral head deeply into the acetabulum at the beginning of treatment and keep it there during the active phase of the disease process. The acetabulum is considered to serve as a template, retaining and molding the shape of the repairing femoral head. The femoral head initially is of normal size and the opposing

344

CHAPTER 4 ~ Leg~t--Calve-Perthes Disease

femoral head and acetabular articular surfaces are of normal shape. The femoral head is seated most deeply in the acetabulum by abducting, internally rotating, and flexing the femur. In that position, forces on the femoral head by the acetabulum will be relatively uniform and the acetabulum should continue to shape the repair of the femoral head such that it heals with its spherical shape intact. It is important to recognize, however, that the femoral head is simply larger than the acetabulum and, thus, never fits into it fully. The containment theory thus positions the vulnerable weight bearing part of the femoral head (anterior-superior-lateral) and that part most susceptible to necrosis and subchondral fracture into the relatively better covered or protected position. The principle of treatment adhered to by most in the orthopedic profession over the past few decades is one defined as containment therapy in which the femoral head is positioned as fully into the acetabulum as possible, allowing the latter to serve as a template molding the shape of the repairing head. Containment therapy can be either non-weight bearing or weight bearing. It is the non-weight bearing approach that actually incorporates both the mechanical and biological concerns about deformation of the femoral head and aims to address both simultaneously. The abduction is obtained by casts, braces, or specially constructed frames and generally was carried out using bed rest or a recumbent position. A unilateral brace, however, can allow the patient to remain ambulatory and still be treated using the non-weight bearing and containment approaches. The second major subdivision of containment therapy stresses the fact that malformation of the head occurs primarily on a biological basis due to asymmetric growth and repair and thus allows weight bearing because the head is not considered to be physically soft. Weight bearing containment therapy can be performed either by using cast and braces or by repositioning the hip components surgically. b. Treatment by Prolonged Recumbency and Femoral

Head Containment by Abduction-Internal Rotation Cast Containment Therapy This approach was well-articulated by Harrison and Menon in 1966 using "containment of the femoral head" by abduction and internal rotation positioning held by abduction or broomstick plasters but maintaining strict non-weight-bearing (103). The method was credited to A. O. Parker of Cardiff, Wales, who instituted it in 1929. They treated 37 patients with this technique and compared them with 37 treated by a variety of other methods. They considered the broomstick abduction containment method to be superior using several criteria. The lower limbs were placed in plaster casts from the groin to the malleoli with the knees slightly flexed. Abduction to place the entire femoral head epiphysis within the acetabulum along with some degree of internal rotation was performed, checked by radiography, and held in position by a wooden bar between the legs. There was unrestricted use of the flexion-extension range of hip motion, but lateral and rotatory motion was prevented by the crossbar and weight bearing was not al-

lowed. In detailed comparative studies of appropriate pairs, Harrison and Menon described a decisive superiority in favor of the broomstick abduction treatment in 12 pairs, as compared with 6 in which both control and abduction treatments were equal and 6 in which the control treatment was superior. The theory underlying the treatment was articulated clearly. It was recognized that the weight of the erect body was a powerful deforming force on a weakened femoral head, but as Pauwels had indicated the hip also was exposed to major tension-compression stresses as well when the lower limb was moved by a recumbent patient due to the associated muscle forces and limb weight. It was theorized that deforming tendencies could be reduced to a minimum in Legg-Perthes disease by "containment of the femoral head." They noted that, with the lower limbs in the anatomical extended position, a portion of the femoral head still projected laterally beyond the lateral edge of the acetabulum such that any compression forces still would tend to flatten part of the head, leaving the uncovered rim untouched. These compressive forces will be marked in the ambulatory patient but still significant with the patient supine. The only way to spare the femoral head would be to abduct and internally rotate it enough to bring the epiphysis inside the acetabulum. Failure to introduce the femoral head epiphysis completely into the acetabulum will result in unequal compression of various areas of the epiphysis with resulting deformation. "If the head is contained within the acetabular cup, then like jelly poured into a mold, the head should be the same shape as the cup when it is allowed to come out after reconstitution." Harrison and Menon felt that the alternative was to arrange for abduction and internal rotation adequate to introduce the epiphysis completely into the acetabulum. Reports of this approach had advocates. c. Containment with Abduction Splint: Recumbency with and without Biologic Nondisplacement Osteotomy (Kendig and Evans and Bohr) Kendig and Evans assessed 52 hips in 49 patients, all of which were treated in recumbency in abduction splints for containment (152). Twentysix of the 52 hips had proximal femoral osteotomy relatively early in the disease course to determine whether healing was faster. Although these were planned as non-displacement osteotomies solely for their biological effects on the rate of healing, 38% healed anatomically, 31% with < 2 0 ~ varus, and 31% with >20% varus, The overall results in Catterall III and IV cases were 44% good, 29% fair, and 29% poor. The results in both groups, however, were the same as assessed by the Mose criteria, healing rate, and epiphyseal quotient. The authors concluded that proximal femoral osteotomy did not hasten repair and that any benefits of the procedure were mechanical due to increased containment rather than biological due to more rapid repair. The average time in the abduction frame was 20.3 months in the osteotomy group and 18.8 months in the nonsurgical group. Bohr also concluded that osteotomy alone did not have much influence on the result of the disease particularly in

SECTION Xl ~ Treatment Approaches to Legg--Perthes Disease

terms of hastening repair (22). He performed partial intertrochanteric osteotomy along with application of a metal plate in an effort to promote healing and shorten the time of hospitalization. The procedure was performed in 42 patients who were being treated with bed rest and traction, but results compared with 133 cases treated with bed rest and traction alone showed no difference either in result or in the time required for healing.

d. Containment and Non-Weight Bearing: Long-Term Recumbency with Abduction Traction Followed by Abduction Plaster Casts (Brotherton and McKibbin) Brotherton and McKibbin reported on a long-term study of 102 hips in 82 patients treated by a rigorous conservative regimen devised in their region (Cardiff, Wales) approximately 50 years earlier by A. O. Parker, who thus was credited as the developer of the containment principle (28). The patients were recumbent throughout treatment. Treatment involved initial management with a period of traction with wide abduction of both legs followed by abduction plasters with the limbs rotated internally once the fragmentation stage was identified radiologically. The average time spent in traction was 11 months and in rotation plasters 15 months. All treatment was in the hospital and ranged from 6 to 44 months. Followup averaged 17 years (range = 10-35 years). The results were among the best reported with the 102 hips distributed as 90 good, 10 fair, and only 2 poor. A comparative study by Ratliff using a variety of methods but over the same period of time showed a distribution good/fair/poor/very poor of 19/19/7/5. When compared with untreated controls, the percentage distributions of recumbent abduction/nontreated were good 60%/57%, fair 31%/19%, and poor 9%/24%. The greatest contribution of this approach thus was in the most severe groups. In Catterall III patients treated there were 18% poor but untreated there were 44% poor; in Catterall IV patients treated there were no poor results compared with 53% poor untreated. The method also was slightly better than femoral osteotomy in a head at risk subset of 70 recumbency abduction/34 osteotomy showing good 49%/53 %, fair 38%/23.5%, and poor 13%/23.5%. The report by Brotherton and McKibbin remains highly informative for several reasons. It presents a long-term review of 102 hips with a mean follow-up interval of 17 years. The patients were treated in one center with an extremely rigorous conservative regimen in which the patients were kept in the hospital for an average period of 26 months, during which time they were confined to bed with the legs in wide abduction and later in abduction plasters. The entire treatment was carried out with the patient in the hospital, thus allowing for careful assessment and also physical therapy to help retain range of motion. As the authors indicate, the results are among the best, if not the best, for any longterm extensive series. It clearly appears that the regimen favorably modified the course of the disorder. The treatment regimen addressed both the mechanical and biological theories regarding the causation of femoral head deformity in

345

Legg-Perthes disease. Most treatment regimens address only one. By keeping the patients recumbent, weight bearing clearly was reduced and, thus, the mechanical effect of deformation was minimized. By utilizing the abduction principle, the head remained as well-contained as possible within the depths of the acetabulum, thus encouraging the biological aspects of the repair phenomenon to lead to a rounded head or at least one congruent with the shape of the acetabulum, which served as the template for remodeling. Of particular note was that the benefits of this treatment were most marked in the more severely affected cases, with a clear shift to fair and excellent results from poor results with comparable methods particularly in the Catterall group III and IV patients. The prolonged therapy was interpreted to indicate that, even in the residual stages of the disorder, remodeling of the head can occur. The mobility of the hip also was maintained because the abduction treatment still allowed for hip flexion and extension. It was felt that containment alone was not the only active and beneficial treatment, especially because it has been shown that the femoral head is never contained completely within the acetabulum regardless of orientation. In breaking down their patients by various groupings, the authors concluded that the regimen described offered no benefit compared with the natural history in Catterall groups I and II, whereas in group III the results were only marginally better than in associated osteotomy groups. In group IV, however, the femoral head was totally involved, but benefit was clear because these patients in almost all series had the worst results.

e. Abduction Splinting, Ambulation, but Non.Weight Bearing: Birmingham Splint A unilateral abduction splint designed by Harrison and colleagues in Birmingham, England, has as its basis a non-weight bearing approach along with use of the containment principle. Both the mechanical and biological concepts of deformity thus are treated with this approach. The unilateral abduction brace builds the position of flexion, abduction, and internal rotation for appropriate positioning of the femoral head in the acetabulum (104, 108). The child walks with crutches. The brace was worn for almost 24 hr per day except for during physical therapy and bathing. Extensive experience by this group has accumulated such that in 1982 a report of 213 patients treated from 1960 to 1977 was presented (108). Two hundred completed the course and were available for assessment. During the first part of their series, there was prolonged splint use with the affected foot off the ground for an average of 23 months (range = 4-38 months). Subsequently, efforts were made to decrease use of the splint, and in a smaller subset of patients the average period of time spent in splint was diminished to 16.3 months with no deterioration of resuits shown. In their detailed review, it was difficult to provide a simple good/fair/poor result, but Harrison et al. did consider overall that 65% of the results could be considered as examples for effective treatment in relation to the shape of the femoral head, 19% were acceptable with the femoral

346

CHAPTER 4 9 Le~tg--Calve--Perthes Disease

head shape unaltered during therapy, and 16% therapeutic failures with the hip joints more abnormal in shape than when the treatment began. Their multiple assessments also were made in terms of the various quotients and assessments of sphericity with the Mose criteria as well as the CE angle of Wiberg. Harrison et al. concluded that the patients with the Catterall type I grouping did not need therapy. Other patients, however, were felt to benefit from the intervention. 5. FEMORAL HEAD CONTAINMENT TREATMENT WITH WEIGHT BEARING

a. General Review Treatment regimens in the mid decades of the twentieth century postulated development of a misshapen femoral head on the basis of its softness during the repair process and attempted to prevent weight bearing. As the repair process takes 1-2 years or more, such prolonged treatment presented a potentially intolerable situation for the child and family. In many instances, the treatments were carried out in institutions to allow for nursing supervision and schooling. Subsequent feeling has been that the entire head is not biologically soft but that its deformity relates to (1) articular surface collapse in relation to adjacent regions of subchondral necrosis, fracture, and bony collapse and (2) poorly constrained repair. With nonconcentric positioning of the head in relation to the acetabulum, the most affected superolateral regions of the head are flattened by pressure from the outer margins of the acetabulum, including the soft tissue components of labrum and cartilage as well as bone. The tendency, therefore, more recently has been to encourage range of movement and weight beating of the hip but in the abducted, internally rotated, and slightly flexed position. Patients were allowed to be ambulatory at all stages of the repair process with the weight bearing status not limited because "softness" of the femoral head no longer was considered an accurate interpretation of the pathoanatomic findings. Containment therapy with ambulation and weight beating can be done either using abduction with casts or braces or surgically using either proximal femoral varusderotation or innominate osteotomies. A strong stimulus to the value of containment therapy came from a study by Kahmi and MacEwen (135). Treatment without containment in 117 hips implied bed rest with extension traction, Snyder sling, or an ischial weight bearing brace, whereas containment therapy in 55 was nonsurgical involving an abductioninternal rotation non-weight beating brace or weight bearing abduction Petrie casts. Patients treated with containment showed significantly more good and fair results. The more severe the involvement, Catterall III and IV groups, the larger the number of poor results with noncontainment management. For patients in group I or II but under 6 years of age, management made little difference because almost all results were good. Beyond that age and grade containment clearly seemed beneficial. b. Weight Bearing and Containment Therapy by Bilat. eral Abduction-Internal Rotation Casts The bilateral ab-

duction cast technique with ambulation and weight bearing was described by Petrie and Bitenc in 1971 (215). This technique addresses each of the desired hip positions of abduction, flexion and internal rotation. Abduction of 45 ~ on each side and 10-15 ~ of internal rotation are obtained by use of an abduction bar attached to each long leg cylinder cast just above the ankle. Flexion of the knees in cast induces a compensatory hip flexion when the patient is walking. The patients are able to walk with the casts on, although in a very awkward fashion, so that limited motion of the hip is possible. A main problem with this technique involves the stiffness of the knees, which occurs with prolonged cast immobilization. Generally it is recommended that the casts be removed at 6-week intervals to allow a day or two of range of motion of the knees, although Petrie and Bitenc went 3-4 months between changes. Excellent results with this method, referred to as a "functional" method, were reported by Kiepurska from Poland (155). Perthes disease in 334 affected hips was treated by the principle of containment with weight beating using the "broomstick" cast, placing the lower extremity and hip in abduction and internal rotation with walking and weight bearing encouraged. The end results were evaluated radiographically using the Heyman and Herndon quotients. In the 334 hips, the results were very good in 216 and good in 92 with a 92% very good or good result noted. In 23 cases results were fair and in 3 they were poor, yielding 8% fair and poor combined. As in other series the fair and poor resuits were found in children older than 9 years of age, with total head involvement and the presence of risk factors. The average time of treatment was 13 months. Percutaneous adductor myotomy was needed in some patients to obtain abduction prior to beginning casting. Traction also was used to quieten the hip initially. The casts were removed every 8 weeks and the child was placed back in bed until the knees regained at least 90 ~ flexion. The principle of containment by abduction casting with ambulation was supported by Richards and Coleman in a subset of patients with severe coxa plana who had developed lateral subluxation (230). They had painful limitation of hip motion that precluded abduction for brace use, but previous experience by this group using osteotomy with such patients had been unsatisfactory. The 22 patients underwent closed reduction under general anesthesia often supplemented by percutaneous adductor tenotomy and application of Petrie casts. At follow-up, they assessed 9 spherically congruent, 12 aspherically congruent, and 1 incongruent hip, factors they felt supported the principle of containment. A good result with treatment using abduction-internal rotation casts is shown in Figs. 41A-4 IN.

c. Weight Bearing and Containment Therapy by Bilateral Abduction Brace The Newington ambulation-abduction brace is a bilateral long-leg brace holding the hips in 45 ~ abduction and 10~ internal rotation (56, 97). It uses the same principle as the abduction casts discussed earlier but was felt

SECTION XI 9 Treatment Approaches to Legg--Perthes Disease

F I G U R E 41 A good result with treatment using abduction-internal rotation casting and full weight bearing is shown. (A) The initial anteroposterior radiograph already indicates slight lateral subluxation and slight flattening of the superolateral bone of the secondary center. (B) The initial frog lateral radiograph shows a remnant of the subchondral crescent and slight irregular mottling of the bone particularly of the superolateral and anterior aspect of the femoral head. The teardrop is widened. (C) Anteroposterior radiograph several weeks later shows almost uniformly increased density of the secondary center bone and persistence of the subchondral crescent. (D) Similar findings are seen on the lateral view, although there already is some increase in radiolucency at the anterior portion of the head. (E) Anteroposterior view of the pelvis with the hips in the abduction-internal rotation casts shows full containment of the involved left femoral head. The line of the femoral head-neck physis is in line with the outer bony margin of the acetabulum. Note whole head involvement with secondary center denseness, relative osteoporosis of the acetabulum, femoral neck, and trochanteric regions, and widened teardrop on the involved side. (F) The uniform radiodensity of the secondary ossification center is broken by a V-shaped notch of radiolucency at the lateral epiphyseal margin. This is defined as a Catterall sign and considered to be a head at risk sign. (G) Subsequent anteroposterior radiograph shows increased lysis at the lateral margin of the secondary center, a sign of revascularization with resorption of necrotic bone. The rounding of the upper outer border of the femoral neck is the Gage sign. ( H - J ) Progressive

347

348

CHAPTER 4 ~ Legy--Calve-Perrhes Disease

FIGURE 41 (continued) fragmentation with increasing resorption of the previously uniformly dense bone of the femoral head. (K-N) Increasing new bone formation is shown in anteroposterior and lateral radiographs. The final two radiographs show excellent reconstitution of bone in both anteroposterior and lateral projections, with the slowest area of repair being central. The sphericity of the femoral head relates well to that of the adjacent acetabulum. There is a mild coxa magna. Physeal growth has continued well with only a slight change from normal in the articulotrochanteric distance.

to be easier and more physiologic for the patient. Curtis et al. reviewed 19 patients with results of 12 (63%) good, 4 (21%) fair, and 3 (16%) poor (56). When adjustments were made eliminating 3- and 4-year-old patients from assessment, the ambulation-abduction treatment still compared favorably over previous methods. In addition, it was felt that the final results for the Newington ambulation-abduction brace were similar to those for the long-term recumbency brace technique from the same institution. Current bracing techniques have favored the use of bilateral abduction apparatuses, which stabilize the pelvis and position both femoral heads deep into the acetabulum. A commonly used bilateral abduction brace in North America has been the Scottish Rite or Atlanta Brace. This is a lightweight brace that is tolerated reasonably well by patients. Although abduction can be obtained, the brace has two major shortcomings in terms of its effectiveness. (1) Flexion at the hip frequently is not built into the brace or readily is

overcome by the activity of the patients so that anterior head coverage is compromised. (2) There is no control for internal rotation and the patients actually walk with the lower extremities and thus the hips in external rotation, which clearly worsens the coverage particularly in the crucial anterior segments that are more involved than the posterior segments with the necrotic process. Another major problem with any brace therapy is the issue of compliance. Relatively few children can be kept in the braces continually for a severalmonth period. It remains unclear biologically whether brace management can affect the outcome of the disorder and, if it does, what period of use is optimal. Studies of the effectiveness of the Scottish Rite abduction brace have not been particularly favorable. The initial report on the Scottish Rite abduction brace was by Purvis et al. (222). In 41 patients they reported a good result in 15 (37%), fair in 17 (42%), and poor in 9 (22%). The patients at the start of therapy had been in Catterall groups II, III, and IV.

SECTION XI ~ Treatment Approaches to Legg--Perthes Disease

The brace was widely adopted, however, perhaps because of the relative ease of use and patient acceptance. Two detailed studies, however, cast major doubts on the effectiveness of the therapy in relation to other modalities. In a review of 31 patients (34 hips) with severe Legg-Perthes disease of Catterall groups III and IV, Martinez et al. showed no hips with a good result, 12 (35%) with a fair result, and 22 (65%) with a poor result (180). The hips were assessed using the Mose criteria. When the Stulberg end stage classification was used, there were no class I results, 41% class II, 18% classes III and IV, and 2 (6%) class V. A second series of 34 patients (38 hips) also involving the more severe Catterall grades III and IV was assessed (188). By the end of treatment, there were no Stulberg class I results, 3 class II, 24 class III, 6 class IV, and 1 class V. When the Catterall end stage results system was used, the results were equally unimpressive, with 4 (11%) having a good result, 11 (31%) fair, 21 (45%) poor. These results really were comparable with earlier studies involving no specific therapy. It should be noted, however, that the relatively more severe Catterall III and IV patients were treated. The assessment of this group indicated that external rotation in the brace often was as much as 55 ~ Those advocating use of the varus-derotation osteotomy actually indicate that it is the internal rotation component that is the more effective repositioning tool and that it frequently is in the 20-25 ~ range. Examples of excellent, good, and fair results with the Scottish Rite brace are shown in Figs. 42A-42H, 43A-43F, and 44A-44I, respectively.

d. Weight Bearing and Containment Therapy by Surgical Intervention Surgical repositioning of the head and neck of the femur in relation to the acetabulum or of the acetabulum in relation to the head and neck of the femur had been used beginning in the 1950s. Such surgical interventions are done on the basis of the same weight bearing containment principle used in bracing and casting, namely, that positioning of the femoral head deeply into the acetabulum is advisable, allowing weight beating during the healing phase such that coverage is congruous and the forces on the repairing head are uniform. Both proximal femoral varus-derotation osteotomy and innominate osteotomy have been used. The head is positioned appropriately even in the uptight standing position postsurgery so that walking can continue in normal fashion once the osteotomy has healed. Proximal Femoral Varus-Derotation Osteotomy: This operation is designed to provide a varus position plus some increased flexion and internal rotation of the femoral head and neck in relation to the acetabulum to improve coverage of the femoral head even when the patient is walking. Once the osteotomy has healed and range of motion rehabilitation has been completed, the patient can be fully ambulatory with the head in the most favorable position at all times. Many excellent results have been achieved with the use of this technique. Once the osteotomy has healed and the patient has regained ambulation, concern about patient compliance

349

is eliminated. The amount of time under active treatment is significantly less than with nonoperative recumbent or ambulatory bracing management. One concern, however, with varus osteotomy is that it further shortens the involved femur and thus increases the likelihood of the need for contralateral epiphyseal arrest toward the end of skeletal growth. There was an early expectation that proximal femoral osteotomy might speed up the rate of healing. A study by Marklund and Tillberg, however, comparing the various stages of Perthes in 25 osteotomy and 33 nonoperated patients showed no evidence that osteotomy hastened the repair phase (179). Serial quantitative scintigraphic studies by Lee et al. in 25 hips with Legg-Perthes treated by proximal femoral osteotomy showed that local blood flow to the femoral head did not increase significantly postsurgery (167). Several years previously, Kendig and Evans performed a control study of proximal femoral non-displacement osteotomy in Perthes disease and found that it did not hasten repair (152). Initial use of the proximal femoral varus osteotomy to reposition the femoral head in Legg-Perthes disease was reported by Soeur and De Racker in 1952 (254). In a long article detailing the Legg-Perthes entity, the static theory of causation of Legg-Perthes disease was outlined and based on mechanical considerations. The reasoning, though interesting, was completely speculative. They proposed that the necrosis of the femoral head in coxa plana resulted from static forces pressing the femoral head against the acetabulum. In coxa plana, it was proposed that there was a disequilibrium of the normal static forces and that these abnormal pressures led to the degenerative state and disappeared when the condition was corrected. The normal head-neck-shaft angle progressively diminished with growth, and it was this angle that established the equilibrium between the forces around the hip. The ultimate position served to balance the varus and valgus forces established by the various muscle groups. It was felt that tension on the hip was diminished with coxa vara positioning and increased with valgus positioning. It was proposed that, with coxa plana, there was a preponderance of valgus forces, which led to an increase in the angle of inclination and to subluxation. In coxa plana, the angle of inclination always was greater than that of the normal femur of the same age. Sundt also made this observation, noting that in 172 hips with coxa plana all with the exception of 9 presented in valgus (261). By valgus, the authors are referring to the head-neck shaft angle without reference to the greater trochanter. This position primarily is due to decreased growth of the physis laterally and continuing growth medially. The coxa plana developed with a predilection for those hips in valgus position. In association with the valgus deformation, subluxation of the head in relation to the acetabulum occurred. Coxa plana thus was due to a disequilibrium between two antagonistic muscle groups, which led to a valgus position and a characteristic subluxation of the femoral head. Soeur and De Racker then went on to derive therapeutic implications from this observation.

350

CHAPTER 4 ~

Le~ty--Calve--Perthes Disease

F I G U R E 42 An excellent result is shown in a series of radiographs documenting the progression of Legg-Perthes disorder using the Scottish Rite abduction brace from inception to skeletal maturity. (A) Anteroposterior radiograph shows hips in abduction in the Scottish Rite brace. A period of traction was used to quieten the hip prior to utilization of the brace. Treatment was started with the hip in the early fragmentation state. Note the increased teardrop width on the involved right side. (B) Anteroposterior radiograph shows fragmentation of the femoral head ossification center, lateral subluxation, and a lateral metaphyseal cyst surrounded by sclerosis. The lateral radiograph (C) shows excellent containment of the femoral head, an increased joint space indicative of thickening of the acetabular cartilage and the femoral articular-epiphyseal cartilage with slight flattening of the secondary ossification center, an anterolateral metaphyseal cyst with sclerosis, and a widened physis. (D) Anteroposterior radiograph shows slight lateral subluxation of the femoral head, a wavy but intact physis, increased new bone formation of the secondary ossification center, and early evidence of a coxa magna. The lateral radiograph (E) still shows fragmentation with varying areas of radiodensity of the secondary ossification center, although a spherical subchondral margin can now be seen throughout the entire secondary center. There has been considerable healing of the metaphyseal cyst. The sphericity of the acetabulum has been well-maintained. (F) In the lateral projection, the sphericity of the femoral head and adjacent acetabulum has been well-maintained with more new bone formation seen. (G, H) Anteroposterior and lateral projections show excellent healing at skeletal maturity with essentially normal relationships in both projections. There is a mild coxa magna on the involved fight side.

SECTION Xl ~ T r e a t m e n t A p p r o a c h e s t o Legg--Perthes Disease

351

F I G U R E 43 A good result using the Scottish Rite abduction brace is shown in this series of radiographs. In this boy, usage was spotty and there were recurrent episodes of tightness to abduction, which could be only partially controlled with bed rest prior to having the boy resume walking again. The final series of radiographs, however, show good sphericity of the femoral head, good congruency of the acetabulum although there is a coxa magna deformity, and slight persisting lateral subluxation. (A) Radiograph shows the abducted position of the femoral head shortly after commencement of brace usage. Note the smaller secondary ossification center on the right. (B) Anteroposterior radiograph shows progression of changes to increased radiodensity and a slight bony flattening of the secondary ossification center. (C) Fragmentation phase of femoral head involvement is well underway with lysis of the previously seen almost uniformly dense secondary center. Note decreased density of the subchondral bone of the acetabulum plus loss of normal curvilinear shape. (D) Increased new bone is seen with both the superior surface of the acetabulum and the bony outline of the secondary center showing flatness. The medial joint space is widened indicating lateral subluxation, and there is increased lateral bone deposition of the secondary center also indicative of a developing coxa magna. Anteroposterior (E) and lateral views (F) show further new bone formation. The sphericity of the subchondral bone is reestablished in both anteroposterior and lateral projections and there is good congruity with the acetabulum, although lateral subluxation and coxa magna persist. The physis is irregular but still present and there is no coxa vara or change in the articulotrochanteric distance.

Previous therapies were r e v i e w e d and found wanting, with the observation again made that deformations of the femoral head led to a position of coxa valga and subluxation whether or not the patient had been treated and that later deforming arthritis was inevitable. The extremely long time that patients were treated with bed rest immobilization again was stressed. If Soeur and De Racker's conception of the pathogenesis was correct, modification of the static conditions of the hip subsequently should influence the ossification of the secondary center. Evidence was presented that a distal transplantation of the greater trochanter had served to m o v e the proximal femur into a more valgus position with rapid worsening of the necrotic sequelae. The more benefi-

cial biomechanical approach was a varus osteotomy. The osteotomy in their patient was intertrochanteric and repositioned the femoral head more deeply into the acetabulum by abducting the proximal fragment and adducting the distal fragment. The angle of inclination decreased from 138 to 125 ~ No internal fixation was used, and a bone graft was placed in the interfragmentary region with the patient immobilized in a hip spica. Three months later, the authors felt that excellent healing of the secondary ossification center had occurred and no additional treatment was performed. The adduction or varus osteotomy, closing the angle of inclination, allowed for a rapid reconstitution of the femoral head. The procedure thus was done not to specifically improve

352

CHAPTER 4 ~

Legg--Calve--Perrhes Disease

F I G U R E 44 The unpredictability of the response to treatment in Legg-Perthes disorder is shown in this series of radiographs. The patient was extraordinarily cooperative throughout the entire period of treatment in a Scottish Rite abduction brace. Treatment began at 6 years of age, which generally is a favorable time. Considerable lateral subluxation soon occurred and persisted even though the range

SECTION Xl ~ Treatment Approaches to Legg--Perthes Disease

coverage of the femoral head or to improve containment, but rather to reestablish the normal static equilibrium of forces around the hip. The procedure also served to correct the subluxation of the head in relation to the acetabulum. The proximal femoral osteotomy transformed the coxa valga into a varus position, and at the same time, caused the subluxed head to reposition itself in the hip joint cavity. The static equilibrium was reestablished and the forces around the hip resumed their normal position and normal values. Pressure against the epiphysis stopped, vascular invasion of the necrotic regions that brought along repair occurred, and, in the authors' opinion, the procedure served to hasten the processes of spontaneous repair in coxa plana. In summary, increased compression at the proximal femur resulted in a disequilibrium between the muscular actions and led to the avascular necrosis and destruction of adjacent vessels. This then "causes subchondral fractures or those in the epiphysis." The practical deductions presented seemed to open the possibility of a new treatment, namely, adduction or varus osteotomy. Other than this report on the use of one operation and the theory behind it, no subsequent studies were done. Subsequently, Axer (5), beginning in 1958, and Somerville (255) (also in 1958) used femoral osteotomy combining both varus and internal rotation to better position the femoral head in Perthes disease. The rationale for use of the procedure, however, was to enhance containment. The procedure was designed to keep the most lateral and anterior segments of the femoral head covered by the acetabulum, centering the whole "plastic" epiphysis inside the joint cavity and keeping it well-covered by the roof of the acetabulum. With walking, forces would be redistributed and contribute to the molding of a more normal joint. It also was felt that the motion and normal weight bearing would enhance not only the mechanical but also the physiological conditions of the hip and in particular enhance resorption of the necrotic epiphyseal bone and its replacement by new living bone. Axer reported on 12 procedures, analyzing the results with clinical and radiographic data including measurements for the comprehensive quotient of Herndon and Heyman (5).

353

He defined results as 5 very good, 4 good, 1 fair, and 1 poor. These results compared favorably with those from nonoperative regimens and greatly simplified treatment for the child. Axer et al. later reported results in 70 hips in 66 children (7, 8). By this time the subtrochanteric varus, extension, and derotation osteotomy (SVEDO) had become better defined. Arthrograms aided in the decision to operate. Surgery was done when superolateral flattening had already been identified. The overall results were 60%/23%/17% good/fair/poor. Once the osteochondral femoral head was "contained within the acetabulum, reciprocal modeling of the biologically plastic hip joint during a full range of motion and normal weight bearing occurs." If the procedure was done later in the repair phase, the deformed head was unable to remodel. The authors stressed close assessment of each patient. Spherical heads well-covered were not treated with surgery, and osteotomy (SVEDO) was done "when the femoral head abnormally protrudes anterolaterally beyond the confines of the acetabulum and the epiphysis becomes locally flattened." The operative approach was quickly adopted by others and currently remains the more frequently used operation in an effort to improve containment in the early and mid-stages of Perthes disease. The recommended degree of varus angulation has ranged between 10 and 29 ~ Klisic et al. recommended 15~ of varusization and about 20 ~ of derotation (158). Trias reported 20-30 ~ of varus at surgery (265). Somerville indicated that internal rotation should not exceed 2025 ~ and varus 10-15 ~ (255). Indeed, he felt that, if possible and as determined by preoperative radiographs, internal rotation alone should be the corrective element. Once healing occurred and range of motion was regained, the child resumed full activities. Somerville felt that surgical containment and early activity gave results as good as and generally better than results in the far more prolonged bracing methods. There is a concern that varusization can be excessive, leading to marked worsening of the Trendelenburg component of the gait. In general, one seeks full coverage of the secondary ossification center by the bony acetabulum on the one hand and maintenance of the proximal tip of the greater

F I G U R E 44 (continued) of motion in abduction was close to full throughout the entire treatment program. At skeletal maturity there is a coxa magna, coxa plana, and considerable incongruence between the femoral head and the acetabulum. In some views the femoral head is flattened considerably, particularly in its lateral two-thirds, and there is lateral subluxation as evidenced by the widened medial joint space and the relation of the femoral head to the lateral margin of the acetabulum. (A) Anteroposterior radiograph shows increased density of the left femoral head along with early metaphyseal rarefaction and lateral head-neck displacement. (B) Frog lateral view demonstrates the subchondral radiolucent crescent and increased density of the secondary center. (C) Anteroposterior radiograph shows increased density of the secondary ossification center with early radiolucency indicative of revascularization at its most lateral aspect. There is prominent metaphyseal rarefaction extending across the width of the neck with adjacent sclerosis. (D) Increased flattening of the secondary ossification center and lateral extrusion of its margin. (E) Anteroposterior view shows relatively rapid resorption of the dense bone of the secondary center, increased lateralization of the head-neck complex, and increasingly poor differentiation of the physis. (F) The femoral head continues to repair with bone, but now there is not only lateral subluxation but also proximal migration of the head-neck complex. The boy remained asymptomatic throughout this entire period with a full range of abduction. (G) The frog lateral shows flattening of the bone of the secondary center at the anterolateral aspect. (H) Anteroposterior view at skeletal maturity shows a markedly deformed head with coxa plana and marked lateral displacement. The acetabulum is slightly dysplastic but clearly not fully congruent with the shape of the femoral head. The sagging rope sign is seen. (I) The frog lateral view shows flattening of the lateral two-thirds of the femoral head and a dysplastic acetabulum.

354

CHAPTER 4 ~ L e 9 9 - C a l v e - P e r t h e s

Disease

Proximal Femoral Osteotomy for Legg-Perthes Disease Varus 10o-25 ~ 9 Head above tip of greater trochanter 9 Lateral physis to outer bony lip of acetabulum 9 Minimize varus if patient > 8 years of age.

Post.

Internal Rotation 10o-25 ~

Extension Also increases coverage of anterior head. Important not to flex proximal fragment.

After Axer- SVEDO = s_ubtrochanteric y_arus extension

d_erotation __osteotomy

F I G U R E 45 Surgical considerations for proximal femoral varus osteotomies are shown.

trochanter below that of the top of the femoral head on the other. The amount of derotation (or increased internal rotation) generally is in the range of 15 ~, which is obtained by extemally rotating the distal fragment prior to fixation. Puranen and Heikkinen felt optimal internal derotation to be 1525 ~ (222).

The final component of correction, stressed by some but not all, is the addition of a few degrees of extension to the osteotomy site to increase coverage of the anterior aspect of the femoral head in the upright position. Surgical considerations in the use of the proximal femoral varus-derotation osteotomy for Legg-Perthes disease are shown in Fig. 45. Considerable experience has been derived from studies with proximal femoral varus osteotomy over the years. A detailed study by Weiner et al. involving 79 proximal femoral varus osteotomies for Perthes identified possible problems as excessive varus angulation after osteotomy minimizing remodeling of the proximal femur with subsequent growth, greater trochanteric overgrowth creating an abductor lurch, and excessive clinical shortening (282). There is good indication in the literature that surgically induced varus in most instances will correct with time. Mirovsky et al. measured the postosteotomy varus angle in 44 patients and noted progressive improvement over the initial angle in each case. In

20 patients in whom the varus angle initially was 18~ or more, rapid improvement took place within 10-16 months with slow progress continuing over a number of years. When the correction was least, the shortening was greatest. In a group of 23 children whose femoral shafts were remodeled completely, the mean shortening was only 0.8 cm after an average follow-up of 9 years. In 9 patients who had a residual varus between 5 and 14~ the average residual shortening was 1.6 cm after a mean follow-up of 8.3 years (193). Menelaus reported that varus will remodel as the proximal femur grows, with the potential for remodeling being greater in children less than 8 years of age at surgery (189). Evans et al. noted an average correction of varus deformity of 13~ (71), Clothier 21 ~ (52), Heikkinen and Puranen 10~ (112), Weiner et al. 11 o (282), and Trias 10-20 ~ (265). Evans et al. noted a substantial decrease in the potential for correction of varus deformity after 9 years of age (71). They felt that, for patients older than 7 years of age, varus should not be less than 110 ~ and for those older than 9 years, varus should not be less than 120 ~ The recommendation of Weiner et al. was that varus correction should not diminish the head-neckfemoral axis by less than 105 ~ because angulation in this range did not remodel as well and appropriate correction could be achieved with an angulation greater than 105 ~ (282). Greater trochanteric epiphysiodesis at the time of varus osteotomy was beneficial and should be considered strongly at the time of surgery. One of the largest reports on varus osteotomy in Perthes disease involved 112 procedures reported by Hoikka et al. (125). The results worsened with increasing age of the patient at surgery, and results were particularly poor in those operated at 9 years of age or older. They felt that the critical age period at operation for optimal results seemed to be up to 8 or 9 years, after which good results seldom were obtained. Long latency between onset of symptoms and surgical intervention also correlated with a worse prognosis, as demonstrated by many others. Hoikka et al. observed no correlation between the overall results and Catterall grouping or signs of head at risk. They felt that the age of the patient at operation was the most important prognostic factor. They were able to note a subchondral fracture in 59% of the hips, but "the overall result could not be correlated with the extent of the subchondral fracture." They also felt that the quality of the end result--good, fair, or poor--could not be quantified in relation to the degree of varusization or neck shaft angle. They felt that femoral osteotomy must not be performed routinely in patients older than 9-10 years of age, but other interventions should be taken into consideration. Cordeiro spoke favorably of intertrochanteric varus osteotomy in 60 cases of Legg-Perthes with 56.6% good and fair results (55). Patients operated at the age of 9 years or older tended to show poor clinical and radiologic results. The epiphyseal quotient was used for radiologic correlation and was considered to be an effective value. Good results were seen with values greater than 60 and fair and poor less

SECTION XI ~ Treatment Approaches to Legg--Perthes Disease than 60. The varus osteotomy was not accompanied by derotation, but medial displacement of the femoral shaft was felt to be important. A detailed study of the results and morbidity for varus osteotomy following Perthes disease was performed by Karpinski et al. (140). They detailed instances of the failure of surgical varus to resolve, limb shortening, and the need for additional surgery to remove the implant. In Catterall groups II-IV, 75.6% of hips either had improved or were unchanged with 24.4% failures. They compared their results in 55 operated hips with a series of 200 hips treated by the Birmingham containment splint as reported by Harrison et al. The same grading systems were used for comparison. The results generally were similar with an overall success rate of 75% in the operated series and 84% in the orthotic series. Haraldsson (100), Laurent and Poussa (165), and McElwain et al. (184) felt that varus-derotation osteotomy improved results they had obtained previously by nonoperative methods. Three detailed studies from the Hospital for Sick Children, Great Ormond Street, and the Royal National Orthopedic Hospital, London, have assessed the effects of proximal femoral osteotomy. In an initial study, it was concluded that containment by femoral osteotomy is the treatment of choice in patients with at risk signs, provided that severe deformity had not already occurred (177). This was based on a comparative assessment with one untreated group and another treated by methods other than containment. Use of the varus osteotomy began in 1963. Overall results in the osteotomy group were good 28 (58%), fair 11 (23%), and poor9 (19%); this compared with percentages in the untreated controls of 43% good, 28% fair, and 29% poor, whereas in the uncontained treated controls the results were similar with 46% good, 26% fair, and 28% poor. The method not only was better than others to a slight extent but, more importantly, also was simple, precise, time saving, and effective. It was important, however, to do the procedure when the at risk signs had appeared but not to delay as further irreversible damage would occur. Group I patients did not warrant surgery and neither did groups II-IV with hips not at risk. A subsequent study in 1980 compared 63 hips in the more severe group III and IV categories contained by femoral osteotomy with 85 untreated hips and found a dramatic benefit, with 50.7% of treated patients developing congruous spherical heads in contrast to only 14.1% of those untreated (38). They felt that, of the dual principles of Perthes care, relief from weight beating did not appear to improve the benefits of containment; therefore, containment was felt to be vastly more important than non-weight beating. They also presented data demonstrating the intuitive feeling that the earlier the surgery was done, the better the result. Good results averaged only 7 months delay following clinical onset of the disease, fair results 8.5 months, and poor results over 14 months. In a 1990 study, the results of proximal femoral osteotomy were assessed at maturity by Coates et al. (53). Indications for intervention were Catterall groups II, III, and

355

IV with head at risk signs. They again concluded that the results of femoral osteotomy were considerably better than those conservatively treated hips in all age groups, except for those under 5 years. Twenty-eight hips (58%) were in either Stulberg class I or II with a good prognosis, 16 were in class III, 3 in class IV, and only 1 in class V. They also compared their results with the work of Ippolito et al., who had reviewed the long-term results of 61 patients with Perthes treated with bed rest, traction, and prolonged weight relief (130). In terms of the Stulberg classification, Ippolito et al. showed 38% I and II and 62% III-V, whereas Coates et al. showed 58% I and II and 42% III-V. There were some additional effects of the surgery that require a note; shortening was present in 14 hips (29%) and a positive Trendelenburg sign was seen in 12 (25%). Menelaus has indicated that stiffness of the hip preoperatively with proximal femoral osteotomy does not have as marked serious consequences as it does with innominate osteotomy (189). He cautions, however, that whereas correction of the varus created by the surgery occurs with growth in those less than 8 years of age, it corrects poorly after 8 years such that proximal femoral osteotomy in the latter group should aim for containment correction by internal rotation and extension of the proximal fragment, with the varus component not done. The value of proximal femoral osteotomy appears to lie in better positioning of the femoral head rather than in speeding the rate of repair. Early advocates of osteotomy had spoken of enhanced healing. Marklund and Tillberg showed no difference in the rate of healing in 25 patients treated by osteotomy and 33 treated by splinting (179). Kendig and Evans showed no difference in healing rate in 26 patients treated in recumbency in an abduction frame and 26 who also had a non-displaced proximal femoral osteotomy (152). Somerville noted no evidence of an increased rate of repair (255). Clancy and Steel (49) found no improvement in healing with incomplete intertrochanteric osteotomy performed specifically to hasten repair, and Bohr (22) found no improvement with partial intertrochanteric osteotomy. Jani and Dick strongly supported the value of immediate varus osteotomy once diagnosis was made and the hip quietened down with bed rest and traction (131). Their results were far better than when varus was delayed until head at risk signs had appeared. This is instructive. Although the better results with immediate surgery might be attributed to the inclusion of patients destined to do well anyway, it also is possible that early repositioning allowed the controlling and molding features of the acetabulum to provide the very benefits the theory of containment postulates. Characteristic responses to the proximal femoral varusderotation osteotomy are shown in Figs. 46A-46E and 47A-47G. Innominate Osteotomy: Innominate osteotomy for a carefully chosen subset of patients with Legg-Perthes disease was proposed by Salter, who described the procedure initially

356

CHAPTER 4 ~

Le99--Calve--Perrhes Disease

F I G U R E 46 (A) Anteroposterior left hip radiograph shows whole head involvement with uniform radiodensity of the secondary ossification center but good position of the head in relation to the acetabulum. (B) Frog lateral radiograph shows whole head involvement with clear evidence of the crescent sign indicative of the subchondral fracture. (C) Proximal femoral varus osteotomy was chosen to allow the patient, who was an extremely active boy, to become weight bearing in the shortest period of time. Intraoperative film shows the AO blade plate, varus osteotomy, which was combined with derotation, and the slight medial displacement of the distal fragment. (D) The osteotomy healed uneventfully. The femoral head now is into the fragmentation stage. It still maintained its sphericity and appropriate position within the acetabulum. The tip of the greater trochanter remains below the most superior point of the articular surface of the femoral head. (E) Good reconstitution of the bone of the head is shown. It has maintained its sphericity and position within the acetabulum.

F I G U R E 47 (A) Anteroposterior radiograph shows Legg-Perthes disease of the right hip in an 11-year-old boy. The femoral head is already somewhat misshapen and a subchondral crescent sign can be seen. (B) The frog lateral projection shows the extensive crescent sign indicative of the area of necrosis plus the subchondral fracture. Note the retained sphericity of the femoral head and adjacent acetabulum. (C) Anteroposterior radiograph following proximal femoral varus-derotation osteotomy is shown. Stabilization was with the AO blade plate. There was slight medial displacement of the distal fragment. Varusization was relatively slight as shown by the fact that the tip of the greater trochanter remains beneath the most superior point of the articular surface of the femoral head. The subchondral crescent sign can still be identified. (D) The lateral radiograph shows the blade plate and the femoral head, which is more radiodense than the adjacent neck and in which the subchondral crescent sign still can be clearly identified. (E) In the anteroposterior radiograph the osteotomy has healed and there is early fragmentation of the femoral head. Containment is still excellent by plain radiographic criteria. (F) Bone repair virtually is complete in the femoral head with sphericity well-maintained. Premature physeal closure has occurred, however, and the tip of the greater trochanter is now above the most superior edge of the articular surface of the femoral head. There is excellent head-acetabular congruity and the acetabulum has expanded laterally to enhance coverage. (G) In the frog lateral view, good congruity persists between head and acetabulum, although the acetabulum is somewhat dysplastic and the femoral head has a coxa magna structure with the mushroom shape deformity.

SECTION XI ~ T r e a t m e n t A p p r o a c h e s to Lecjg-Perthes Disease

357

358

CHAPTER 4 ~ Le~lg--Calve--Pe~hes Disease

for the dysplastic acetabulurn in association with developmental dysplasia of the hip (238-240). Salter felt that containment was essential to obtaining a spherical head and a consequent good-to-excellent long-term prognosis. He felt that subluxation would lead to deformation because loss of containment pressure by the edge of the acetabulum would cause stress concentrations on the femoral head that lead to progressive deformity. He felt that, in the pathogenesis of long-term arthritis secondary to Perthes disease, "residual femoral head deformity is the only predisposing abnormality that is significant and also the only one that is preventable." The only two major valid prognostic factors were the age at onset, with the younger patient having the better prognosis, and the loss of containment, which referred to subluxation, extrusion, and limitation of motion of the femoral head. The most important treatment principle would incorporate weight beating methods with containment of the head. The advantage of operative treatment in the older children was that the period of restriction of the child was less than 2 months, after which containment was permanent and no specific additional treatment was needed. Innominate osteotomy was indicated for many over 6 years of age with moderate-to-severe involvement and loss of containment. Contraindications for innominate osteotomy included those at either end of the disease spectrum, namely, those with a very good prognosis with minimal involvement at any age and all children under 6 years of age and those with a very bad prognosis which included persistent restriction of hip motion and established deformity as shown by arthrogram. Salter commented on 110 innominate osteotomy procedures, comparing them to a control group of 38 hips treated several years previously in a weight relieving brace without a containment principle. In the nonoperated group, 37% had good results, 66% had either good or fair results, and 34% had poor results. In the innominate osteotomy group, 77% had good results, 94% had either good or fair results, and only 6% had poor results. Salter frequently stressed the fact that, whereas treatment could prevent a deformity of the femoral head, it probably could not correct or reverse a preexisting deformity. Arthrography thus was important to determine the shape of the cartilage model of the femoral head. One of the reasons that Salter's results were better than others was his stated reluctance to operate on many of the more severe variants. Although he and others have reported excellent results with the innominate osteotomy, some centers have experienced problems using the procedure for Legg-Perthes disease and it has not been widely adopted. Advantages of the procedure include increased coverage of the head in its anterior and lateral aspects along with the fact that the procedure lengthens the involved limb to compensate for any shortening in relation to the osteonecrosis of the capital femoral epiphysis. Occasional major problems have occurred, however, with increased postoperative stiffness of the hip. Adherence to the strict criteria for use of the procedure is essential: (1) the operation should not be performed until a virtual full range of motion of the affected hip has

been regained by rest and nonoperative means; (2) there is almost perfect congruity of the hip joint; (3) age of onset of disease of older than 6 years of age; (4) total head involvement; and (5) extrusion-subluxation of the femoral head. If the patient is still in an active synovitis phase with decreased range of motion and discomfort at the time of performance of the innominate osteotomy, the likelihood of prolonged postoperative stiffness appears to be greatly increased. Menelaus (189) has stressed good preoperative abduction to within 20 ~ of the range of abduction of the opposite hip, and Klisic et al. (158) suggest that the operation is not advisable in the presence of subluxation. Even if the hip has good preoperative movement, the operation tends to increase pressure on the femoral head, which itself can lead to further collapse of the subchondral region. Salter also stresses release of the tendinous portion of the iliopsoas muscle at the musculotendinous junction, and any residual contracture of the adductor muscles should be released by a subcutaneous tenotomy. Salter reported that, in 110 hips treated by innominate osteotomy, the results were good in 77%, fair in 17%, and poor in 6% (with the average age at operation of the 6% with a poor result relatively older at 10 years of age). These results were much better than in his control series of 38 nonoperative hips. A subsequent report by Robinson et al. in 27 patients treated with this procedure also noted clinically good or fair results in 88% with poor results in 12% (234). All but 2 were Catterall III or IV at the time of surgery. Overall favorable results also were reported by Barer (12) with 14 good, 6 fair, and 3 poor and by Dekker et al. (59) with 5 excellent, 10 good, 9 fair, and 1 poor. Maxted and Jackson were pleased with the results of innominate osteotomy in those with whole head involvement and noted circular-spherical heads in 78% (28/36) radiologically at completion of healing (182). In the study, 15 of the 36 were less than 6 years of age at osteotomy. Other studies have been somewhat more equivocal. Ingman et al. reviewed 38 cases of Perthes treated by innominate osteotomy and felt that there was the same high incidence of poor results as in a control group matched by age and Catterall grouping and treated with hip spica alone (127). Overall, the clinical results were similar in the two groups, although somewhat better in those not operated. One of the poor results postosteotomy eventually had an arthrodesis. Figure 48 shows an example of a hip that underwent an innominate osteotomy not in accordance with criteria now accepted and subsequently required hip arthrodesis. The good/fair/poor results were 14/15/9 for osteotomy and 17/10/6 for spica. The single most significant factor determining the end result in both groups was age, with those 8 years and under doing much better than those over 8 years of age. The Adelaide, Australia, group continued to use the procedure, however, but sought to improve results by using stricter criteria more in line with Salter's early suggestions. When 27 patients (11 from the original group) were reported 9 years later, clinical and radiologic good/fair/poor results in percentages were 66/26/8 and 56/40/4, respectively (208).

SECTION XI ~ Treatment Approaches to Legg--Perthes Disease

359

FIGURE 48 Radiographsshow a poor result following an innominateosteotomydone in a young teenagerwell beyond the currently accepted criteria for patient selection. When seen subsequently the patient had a fixed flexion, adduction, and external rotation contracture of 35~ 15~ and 30~ respectively, which was nonresponsive to conservative managementand physical therapy and which subsequently underwent a salvage procedure with hip arthrodesis. (A) A standing anteroposteriorfilm of the pelvis shows the normal right hip and the adducted and markedly deformed left hip. This representedthe maximumamount of abduction obtainable. Pelvic obliquity also was present. (B) The misshapen left femoral head is shown. There also is an osteochondritis dissecans type lesion. (C) Hip arthrodesis allowed for a firm union and markedly improved position. Note the disappearance of the pelvic obliquity and improved alignment of the femoral shaft in relation to the left hemipelvis. (D) Anteroposteriorradiograph of the pelvis following removal of the nail. The patient remainedrelatively symptom-freeover the next 15 years.

Although these results clearly were better, two factors must be recognized. Indeed, these factors are important in assessing all reports on Perthes management. First, in the later series, 15 of the 27 patients were younger than 6 years of age, a group in which a high percentage of excellent and good results is found regardless of therapy. Only 1 of the 27 patients was over 8 years of age. Second, each of the 11 patients from the first series included in the second series had an improving result due to continuing remodeling up to the time of skeletal maturity. This shows the need to wait until skeletal maturity before definitely grading the result of any therapy. What is clear, and thus worrisome, is that many of the papers discussed in this chapter did not exclusively review patients only at skeletal maturity or beyond. Canale et al. reviewed 15 patients having innominate osteotomy with whole head involvement and lateral subluxa-

tion of the femoral head (37). They had 40% good, 33% fair, and 27% poor results, but stressed that the procedure was performed specifically in a relatively severe group of patients with poor prognosis but good potential for repair. Stevens et al. reported on 70 patients at an average follow-up of 4.3 years with 54% good, 19% fair, and 27% poor results (258). Seven patients (10%) developed postoperative hip stiffness marked enough to require additional hospitalization. The stiffness developed in patients who had stiffness at the time of surgery and were 8 months or more into the symptomatic phase of their condition. A poor result following innominate osteotomy performed in disregard of Salter's criteria is shown in Figs. 48A-48D. The teenage patient with severe head deformity developed stiffness and an adduction contracture and eventually underwent hip arthrodesis.

360

CHAPTER 4 ~ Legg-Caive-Perrhes Disease

e. Weight Bearing Therapy without Containment Weight bearing during the repair phase was adopted when the concept of the "softness" of the entire femoral head was abandoned. It invariably is combined with containment and has been reviewed in preceding Sections a-d. Weight bearing without containment attempts is included for the sake of completeness of the spectrum but really indicates either no treatment or symptomatic treatment only. 6. HASTENING OF BONE REPAIR BY OPERATIVE PROCEDURES ON THE SECONDARY OSSIFICATION CENTER

Efforts have been made to quicken bone repair in the secondary ossification center by using techniques helpful in enhancing such repair in other areas of the body and in other disorders. a. Drilling of the Secondary Center of Ossification This procedure was performed occasionally in an effort to hasten the revascularization process. Drilling was done through either the greater trochanter or femoral neck, but no detailed long-term results were reported (75). The negative features of this approach are clear. Because the femoral head articular and physeal cartilages effectively merge there is no safe entry point into the secondary ossification center without at least slightly damaging physeal or articular cartilage. Any vascularity the procedure would bring to the secondary center would have other negative effects on the two cartilage systems. b. Drilling and Bone Grafting Secondary center drilling plus the addition of a bone graft also was attempted as reviewed by Edgren (68). Utilization of a graft of nonviable bone is not attractive when placed within a bed of necrotic tissue. c. Resection of Necrotic Bone Repair was felt to be hastened by resection of the necrotic secondary center bone in separate reports by Haythorn (111) and Bernbeck (17, 18). No positive results or widespread adoption of the technique occurred. The theory behind this approach is interesting, but the current findings of multiple episodes of infarction could affect any repair tissue. d. Resection of Necrotic Bone with Addition of Cancellous Bone Graft Cathro and Kirkaldy-Willis added a cancellous bone graft following resection of necrotic bone (68). One might well expect a quickening of the repair response, much of which is concerned with the need to remove necrotic dead bone before repair bone can predominate. The removal process is very slow and is accompanied by collapse of the head and coxa magna, probably associated with the increased vascularity to the entire region.

C. Other Factors Concerning Results 1. COMPARISON OF DIFFERING TREATMENT TECHNIQUES PERFORMED BY THE SAME INSTITUTIONS AND REPORTED IN THE SAME PAPERS There are some instances in which papers report comparison of treatment techniques within the same institution.

These have value in the sense that the same internal controls for assessment are used. Some of these comparisons have been reported earlier. Fulford et al. reported two groups of children with Perthes disease randomized at the beginning of treatment, with one group treated by bed rest and skin traction followed by either a weight relieving caliper or proximal femoral varus osteotomy (82). They noted the outcome to be similar in both groups, with the end result predicted more effectively by the arthrographic shape of the femoral head at presentation than by the Catterall classification. The series of Karpinski et al. reported earlier showed similar results with the varus proximal femoral osteotomy compared to those treated with the Birmingham splint (140). The report used the same criteria of assessment, Harrison's radiologic autoassessment method, although different institutions were involved. Sponseller et al. compared the results of 42 femoral osteotomies with 49 innominate osteotomies done in Catterall group III and IV hips and at an average follow-up of 9 years using the Mose criteria, the Iowa hip rating system, and the classification of Stulberg et al. (257). There were no statistically significant differences between the results after the two operations. Cooperman and Stulberg assessed ambulatory containment treatment in Perthes disease in four large series involving 72 patients treated with crutches alone, 58 treated with a Scottish Rite orthosis, 48 with a Newington abduction orthosis, and 70 with proximal femoral osteotomy (54). Patients with partial head involvement under 8 years of age at disease onset with minimal lateral subluxation and no proximal subluxation fared equally well with crutch walking treatment alone compared with any of the three ambulatory containment treatments studied. With greater degrees of subluxation, all three containment groups were superior to crutch walking and all were equally effectix/e. In patients with partial head involvement between 8 and 12 years of age with or without lateral or proximal subluxation, all three ambulatory containment methods were superior to crutch walking. Femoral osteotomy and Newington abduction orthoses were equally effective. Patients in general with partial head involvement over 12 years of age at disease onset had poor results. For patients with total head involvement under 8 years of age at disease onset, femoral osteotomy was correlated with a lesser degree of femoral head flattening than either brace or crutch treatment. This represented the only section of the study in which femoral osteotomy clearly was superior. In whole head involvement with patients over 8 years of age at disease onset, all three ambulatory containment methods were equally effective again with all better than crutch walking. The end results were evaluated according to the Mose criteria and the Stulberg classification scheme. Cooperman and Stulberg concluded that crutch walking essentially was equivalent to total nontreatment. Curtis et al. compared the Newington ambulationabduction brace program with an age-matched series treated

SECTION Xl 9 Treatment Approaches to Le00--Perthes Disease in recumbency with abduction (56). The end results were similar at (good/fair/poor) 63/21/16 in the ambulatory brace group and 66/17/17 in the recumbency group (percentage distributions). A subsequent study from the Newington group by Evans et al. using different patients compared 17 ambulation-abduction brace patients with a matched group having proximal femoral varus-derotation osteotomies (71). Treatment results were equal in the osteotomy group with good/fair/poor in 63%/11%/26%. Uyttendaele et al. studied conservative nonoperative treatment in Perthes disease and concluded that noncontainment methods were as effective or even slightly more effective than containment approaches (271). A total of 59 hips were treated by conservative noncontainment and 31 by conservative containment, although the specific methods were not mentioned. In categorizing overall results as good, fair, or poor, the percentages with noncontainment were 44%/19%/37% and with containment were 35%/26%/39%. Mindell and Sherman assessed two comparative groups involving ambulatory treatment without weight beating using either crutches or an ischial weight bearing brace and nonambulatory recumbent treatment with prolonged periods of bed rest with or without traction (192). They felt that the results basically were the same in the ambulatory compared with the nonambulatory group. Patients under 5 years of age had better results than those over 8 years, with the 5-8 group intermediate. Their overall results showed 41 satisfactory, 11 fair, and 21 poor. Karadimas performed a comparative study of four different conservative treatment methods in 96 patients with unilateral Perthes disease treated between 1939 and 1968. Twenty patients were treated in broomstick abductioninternal rotation plaster casts and bed rest, 30 in an abduction frame and bed rest, 25 with skin traction and bed rest with no restriction of hip movement, and 21 in a weight relieving ambulatory caliper brace. All patients were assessed the same way with documentation of the radiographic results including use of the Comprehensive quotients of Herndon and Heyman. Each of the groups was comparable in terms of patient profile. The best results occurred with the broomstick plaster casts, which utilized the principle of containment of the femoral head with full protection from weight bearing. Progressively less effective results were achieved with the abduction frame, skin traction, and the weight relieving caliper (which gave the poorest results). The final radiographic shape of the heads was spherical in 65% with the broomstick plasters, 43% using the abduction frame, 36% with skin traction, and only 24% with caliper treatment. In terms of the comprehensive quotient, good or excellent results were achieved in 17 of 20 (85%) with the broomstick plaster and 22 of 30 (73%) with the abduction frame. No patient treated with the abduction plaster casts had a poor result and in only 3 was the result graded as fair. Lahdes-Vasama et al. compared the outcome in Perthes in unselected patients after femoral varus osteotomy and

361

conservative treatment with a Thomas splint. In patients with less than 50% femoral head involvement (Salter-Thompson group A), no advantage was derived from the operation (163). In hips with more than 50% head involvement (Salter-Thompson group B), the operative method yielded slightly better coverage and sphericity of the femoral head. 2. LENGTH OF TREATMENT TIME IN NONOPERATIVE CONTAINMENT AND NONCONTAINMENT APPROACHES

Even when treatment is performed for this disorder, considerable uncertainty has existed as to how long treatment should continue. Variability was great because many physicians felt there had to be complete reossification to uniform bone density of the epiphyseal bone of the femoral head, such that patients often were continued in recumbency or immobilized in braces for as long as 3 years and occasionally longer. Many physicians progressively began to shorten the time of immobilization both due to their feeling that it was not necessary and also due to familial pressure because of the adverse social effects on the child of such prolonged treatment. Ferguson and Howorth as early as 1934 found that treatment to full bony reestablishment was not needed and that full weight bearing could be resumed once the reparative phase was underway (75). This involved early reossification in which new bone began to form in the secondary ossification center. Thompson and Westin performed a detailed study assessing the results of discontinuing treatment in the early reossification phase (264). In a large group comprising almost 200 hips, variable treatment programs had been used such that comparative analysis could be made. Some patients were treated with noncontainment methods, including ischial weight beating braces, slings, recumbency, crutches, or a spica cast. In another large group of patients, treatment by conservative containment was utilized involving either abduction casts or abduction braces. Results were uniformly satisfactory with none poor in Catterall grade I and II patients, regardless of whether containment or noncontainment was used in those who were treated nonoperatively. When groups III and IV were added to the assessment, results in containment and noncontainment groups again remained essentially the same with the overall series producing 76% satisfactory and 24% poor results. It is evident that most of the patients had the grade III or IV categorization and also that the poor results were seen in the latter two groups only. The study concentrated on 84 hips treated until there was complete restoration of the capital femoral epiphysis and 81 hips whose treatment was discontinued in the reossification phase. All hips treated by containment were in the latter group. The difference for the patient was considerable because those treated to complete healing averaged 30 months of therapy and the other group showed treatment time markedly diminished to between 6 and 12 months. Thompson and Westin noted that "when comparing the results of these 2 groups at the time of discontinuing treatment and the last clinic visit, there were no significant differences." Indeed

362

CHAPTER 4 ~ Legy--Caive--Perthes Disease TABLE llI Results of Legg--Perthes Disease Based on Age of Disease O n s e t Irrespective of Treatment Type a

Age at onset of disease

No. of hips

<6 years

220

6-8 years

234

> 8 years

114

Good

Fair

Poor

127 (58%) 92 (39%) 19 (17%)

70 (32%) 84 (36%) 28 (25%)

23 (10%) 58 (25%) 67 (59%)

aBased on Herring (120).

there was a tendency toward further improvement the longer the patient was assessed to skeletal maturity. The immobilization was discontinued when the subchondral reossification was present in both anteroposterior and lateral radiographs. Thus, they noted that 81 conservatively treated hips, which were mobilized early and followed for an average of 5 years, presented results identical to those treated until the time of complete restoration of the capital femoral epiphysis. 3. RESULTS FROM MODERN SERIES IN RELATION TO AGE AT ONSET OF DISEASE

Herring has summarized in tabular form the results reported in several papers based on the age at disease onset (119). In Table III, we have categorized the results, which again show the most favorable responses, regardless of the type of therapy, in those less than 6 years of age at disease onset and the poorest results in those 9 years of age or older.

growth plate, which occurs in roughly 25% of patients with Legg-Perthes disease (Fig. 26). If the relative overgrowth is early in its phase of development, epiphysiodesis of the greater trochanter can stabilize the relationship of the greater trochanter to the head and neck regions. Matan et al. studied two groups of their patients, one group of 8 who had varus osteotomy alone and a group of 20 who had simultaneous varus osteotomy and greater trochanteric epiphysiodesis (arrest) (181). They concluded that the latter approach was preferable to stabilize the proximal femoral relationships. The latter group had a better range of motion, less abductor weakness, and less pain. Another approach is to wait until the disorder is well into the repair phase and there is relatively little or no growth remaining, at which time distal transfer of the greater trochanter minimizes the coxa vara position and allows for relative strengthening of the abductor muscles. The advantages of trochanteric advancement in LeggPerthes disease were reported by Lloyd-Roberts et al., who noted that all 9 hips with Perthes disease treated in this fashion had improvement (178). Doudoulakis reported on 35 hips with premature closure of the subcapital growth plate who were treated with trochanteric advancement (63). All of the trochanters fused without complication, and the Trendelenburg sign, which was positive in 24 hips preoperatively, became negative postoperatively. In addition, hip abduction was increased in 28 of the 30 hips reviewed. In this series, five of the procedures were done for Perthes disease with the large majority for sequelae of DDH. 2. DISTAL FEMORAL EPIPHYSEAL ARREST ON THE CONTRALATERAL SIDE TO TREAT LOWER EXTREMITY LENGTH DISCREPANCY

D. Late-Stage Surgical Intervention to Treat the Sequelae of L e g g - P e r t h e s Disease 1. INTERVENTIONS ON THE GREATER TROCHANTER

Greater trochanteric overgrowth is a common feature of Legg-Perthes disease. The avascular necrosis that affects the secondary ossification center and the epiphyseal and physeal cartilage does not affect the vascularity of the greater trochanter, which continues its growth. The asymmetry of growth between the greater trochanter and the head-neck regions leads to a coxa vara deformity (head-neck-trochantershaft) and almost always to an awkwardness of gait, which has both a short limb and an abductor muscle weakness component. The latter, which can be assessed on the basis of a positive Trendelenburg sign, can be minimized or corrected fully by surgical intervention on the greater trochanter. The relative overgrowth of the greater trochanter and its position in relation to the shortened femoral head and neck can be worsened further by proximal femoral varus osteotomy done to improve the relationship of the femoral head to the acetabulum and by premature closure of the proximal femoral

Many patients with Legg-Perthes disease have a clinically significant lower extremity length discrepancy. This can occur from the avascular necrosis of the femoral head and neck region itself and can be worsened by proximal femoral varus osteotomy and by premature physeal closure. In a certain percentage of patients, the disorder is sufficiently great to mandate distal femoral epiphyseal arrest on the contralateral side to allow for diminution of the discrepancy. Some physicians feel that slight shortening of 1 cm or less of the involved side is preferable as it enhances coverage of the femoral head with walking. If shortening, however, is greater than 2 cm, concern remains that the relative adduction positioning of the contralateral normal hip can lead to osteoarthritic changes in the later decades of life. Management of lower extremity length discrepancies in LeggPerthes disease also will be referred to in Chapter 8. 3. ACETABuLAR RECONSTRUCTION Acetabular reconstruction can be attempted in those patients usually in the late teenage or early adult years who have sufficient incongruence that they become symptomatic.

SECTION Xl ~ Treatment Approaches to Legg--Perthes Disease On occasion, children and adolescents with severe LeggPerthes disease involving symptomatic major deformity might require surgical intervention in an effort to remove pain, increase motion, or better support the hip through the late adolescent and early adult years. It is estimated that roughly 10% of patients with Legg-Perthes disease develop deteriorating symptoms that require surgery before the age of 35 years. Pain, instability, and decreased hip motion represent the major problems. Generally this is attributed to an incongruent joint representing early osteoarthritis. Acetabular remodeling continues until skeletal maturity such that acetabular procedures should be deferred until that time. There is, however, increasing recognition of two entities that, if treated appropriately, can add many years of function to the hip. These include loose body formation from foci of osteochondritis dissecans and the prominence of the superolateral margin of the femoral head beneath and lateral to the rim of the acetabulum, producing hinge abduction. a. Chiari Osteotomy Bennett et al. reported on Chiari osteotomies done for the treatment of painful subluxation of hips following Perthes disease (14). They documented improved radiographic appearance, particularly involving the center-edge angle and the percent coverage of the femoral head. The average age at surgery was relatively young at 10 years, which, although "old" for Perthes disease, is young for the Chiari osteotomy. There is not wide experience with this procedure in the growing age group. b. Acetabular S h e l f Arthroplasty Kruse et aL performed shelf arthroplasty for patients with severe Legg-Perthes disease and compared them with a group not having the procedure (162). The shelf arthroplasty was performed in 20 hips following a period of traction and adductor tenotomy to improve subluxation position. The average duration of followup was 19 years. The second group consisted of 18 hips managed nonoperatively reviewed at an average duration of follow-up of 28 years. In the arthroplasty group, 17 of the patients were graded as Catterall group IV and 3 as Catterall group III. The mean age at onset of disease in the arthroplastic group was 7.9 years and 6.9 years in the nonoperated group. The Catterall class was somewhat less severe in the nonoperated group with 12 class III and 6 class IV. Patients in the operative group did better than those in the nonoperated group. Postsurgery, the operative group had very significant improvement in the Mose sphericity measurement and in the center-edge angle. Acetabular coverage was much better and the average Iowa hip rating score in the operative group was 91 points, compared with 81 points in the nonoperated group. When function was analyzed, the operative group had less pain and a positive Trendelenburg sign was seen less frequently postsurgery. Hinge abduction, which was producing the incongruity, was eliminated in 11 of 14 hips after surgery. They felt that the shelf operation had proven itself useful in their patients in whom lateral displacement, flattening, and enlargement of the femoral head had occurred. Shelf arthroplasty also was favored by

363

Reinker in late cases of hinge abduction not responsive to traction, abductor release, and abduction casting (229). Generally it is warranted for instability due to imperfect lateral coverage of an enlarged, misshapen head. Willett et al. (284) reported on 20 patients and van der Heyden and Tongerloo (273) on 25 patients with generally satisfactory results. 4. CHEILECTOMY Garceau (87) and later McKay (185) removed the superolateral prominence of the deformed femoral head that was positioned beneath the lateral to the outer acetabular rim preventing free hip abduction. Although a mechanical block to free movement was removed, the bone was left denuded of cartilage within the joint, a nonphysiologic response. The procedure is done infrequently at present. 5. PROXIMAL FEMORAL VALGUS OSTEOTOMY If the superior articular cartilage surface of the femoral head already is flattened at the time of surgery, varus osteotomy will change the position of the head but still not allow for any meaningful congruency between the misshapen articular surface of the head and the relatively spherical acetabulum. Some now wait until much bone repair and remodeling has occurred, at which time a proximal femoral valgusextension osteotomy positions the smoothest area of the articular surface, which is invariably the medial one-third, against the most prominent weight bearing dome of the acetabulum. Good results were reported by Catterall (46) and Harrison (109) and in subsequent larger reports by Quain and Catterall (26 cases) (223) and Urlus et al. (17 cases) (270). The primary indication for valgus and extension osteotomy is what is referred to as "hinge abduction," which occurs with prominence at the superolateral margin of the head in association with a flattened region medial to it. The area of prominence lies underneath and lateral to the lateral margin of the acetabulum and produces what is referred to as hinge abduction. When the hip is abducted, the femoral head does not glide smoothly within the acetabulum but moves eccentrically, hinging on the superior lip of the acetabulum and increasing the width of the medial joint space as defined radiographically or in particular by arthrography. Once the bony prominence has formed, the lateral subluxation becomes fixed and any attempt at abduction results in a hinge process that worsens the degeneration of the articular surfaces (46, 109). In many, the hip does not abduct beyond neutral and may even be fixed in an adducted position. The hip thus becomes stiff and painful and intervention generally is needed. Quain and Catterall reported satisfactory results with abduction-extension osteotomy in 26 hips. The hinging is best documented under general anesthesia with a hip arthrogram and fluoroscopy. In addition, the most congruent position of the head in the acetabulum can be determined, which is usually between 15 and 30 ~ of adduction with 1030 ~ of flexion. The abduction-extension osteotomy then is performed to place the most spherical part of the femoral

364

CHAPTER 4 ~ Leg~t--Calve--Perthes Disease

head in the weight bearing position in relation to the acetabulum. The procedure serves to minimize or eliminate pain, eliminate the fixed adduction and flexion deformity, improve movement, and also somewhat correct any shortening. Those developing hinge abduction tend to have developed the Perthes disorder at a relatively late age. Urlus et al. described 18 hips treated by the abduction-extension osteotomy with satisfactory to good results. The mean age at surgery was 9 years 6 months (range = 6-14 years), whereas in the Quain and Catterall series patients were somewhat older at an average age of 13.4 years (range = 8-23 years). An arthrogram was used both for diagnosis and for deciding on the optimal position for surgery. The best congruence for the head portion was between 10 and 30 ~ of femoral adduction in all hips. Extension also was built into the osteotomy to correct any flexion contracture. The procedure, when performed toward the end of the repair phase or in early adulthood, is designed as somewhat of a salvage procedure rather than a procedure to enhance the normal sphericity of the head (Figs. 49A-49F). Kim and Wenger used three-dimensional CT scan reconstructions to study femoral head-acetabular relations prior to surgery for major end-stage Perthes deformity. In many of their patients, such studies led to proximal femoral valgus osteotomies, with internal rotation and flexion, rather than extension, added to better position the femoral head (157). The importance of preoperative imaging studies to assess optimal head position is made clear.

6. APPROACHESTO HINGE ABDUCTION BASED ON DISEASE STAGE The concept of hinge abduction for many has become a key indicator of prognosis in the sense that, if present and untreated, the likelihood of degenerative disorders of the hip is raised greatly. Treatment guidelines are becoming clearer. When the disorder is appreciated in the active or fragmentation stage of the disorder at a time when improvement of the cartilage model is feasible at least theoretically, steps are taken to improve containment. These involve the use of traction to rest the hip, percutaneous adductor tenotomy to increase the range of abduction, and then definitive use of a containment approach once the range of motion has been improved and the position of the head within the acetabulum obtained. For some this involves use of the abduction and internal rotation plaster or "broomstick" technique with weight bearing allowed; for others the use of proximal femoral varus and derotation osteotomy is preferred. In general, however, the hinge abduction phenomenon is not detected until relatively late in the disorder. At these stages, generally irreversible flattening and a misshapen femoral head have formed in terms of both the cartilage model and the underlying bone. Three basic approaches then are taken to this disorder. As noted earlier, one can either remove the superolateral prominence by cheilectomy or attempt to cover it by shelf arthroplasty. Increasingly, however, the treatment recommended by Catterall and others is used, which involves a

proximal femoral abduction or valgus osteotomy along with extension to position the most spherical part of the femoral head into the deepest or weight bearing portion of the acetabulum.

E. S u m m a r y of Treatment A p p r o a c h e s Legg-Perthes is one of the more difficult disorders for the pediatric orthopedist to treat. Although our knowledge of the disorder has increased over the decades, major deficiencies in our understanding still exist at each level of assessment. The etiology remains unknown, which clearly limits our ability to prevent the disorder and even to apply specific therapies. The clinical picture also is extremely variable. Many patients present well into the disease progression, such that deformity of a meaningful extent already has occurred at the time that therapy is beginning. The results are extremely variable and not strictly in correlation with any of the parameters used for assessment, whether it be the age at presentation or each of several classification schemes. Because not an inconsiderable number of patients do well either with no therapy or with only minimal therapy, and because many patients who undergo extensive types of therapy involving prolonged brace use or extensive surgeries do poorly, a certain degree of uncertainty pervades the issue of management. Categorization of the end result of the disorder is imprecise and makes comparative interpretation difficult. The following three factors must be considered in reviewing papers on results published in the literature. (1) Studies done on hips in which repair appears to be complete but skeletal maturation has not been reached do not truly represent end results because femoral head and acetabular remodeling in the final years of growth can improve sphericity and congruity or, on occasion, worsen appearance with or without premature physeal closure. (2) Both radiologic and clinical parameters must be assessed because often there is a great discrepancy between the two particularly in adolescence and the early adult years. Clinical function almost always is better, often markedly better, than radiographic appearance. (3) There still is imperfect understanding of which radiographic parameters predispose one to clinically meaningful worsening with time, such that a difference in interpretation can lead one group of observers to downgrade a hip based on one set of findings that others rate more favorably. It also is evident that long-term assessments several years to decades after the cessation of growth continue to show variable results and are subject to differing interpretations. Although long-term studies show that many patients with mildly to moderately deformed hips do well for several decades, the tendency is increasing to use the criterion of whether total hip replacement surgery was required as an index of success of treatment. Although this has apparent statistical validity, another series of problems arises because the use of the total hip procedure itself is quite variable in the adult orthopedic community and particularly in differing regions of the world.

SECTION Xl ~ Treatment Approaches to Legg--Perthes Disease

F I G U R E 49 Proximal femoral valgus osteotomy can be used to place the medial portion of the femoral head, which generally remains spherical, into a more effective superior weight bearing position. The valgus osteotomy better positions the more spherical medial portion of the head as well as transfers the trochanteric tip more distally, favoring diminution of the Trendelenburg gait. (A) Anteroposterior radiograph of the right hip shows the Legg-Perthes disorder at diagnosis. A subchondral fracture can be seen. (B) The frog lateral view shows increased radiodensity of the secondary ossification center and irregular shaping of the subchondral bone. (C) A proximal femoral varus osteotomy healed uneventfully, but greater trochanteric growth continued to be more marked than femoral head and neck growth, leading to worsening of the varus deformity. (D) A severe coxa vara deformity followed with the femoral head pointing medially and inferiorly. Abduction was limited markedly and the Trendelenburg gait was marked. (E) Proximal femoral valgus osteotomy stabilized with an AO blade plate placed the femoral head into a much better anatomic relationship to the weight bearing portion of the acetabulum. (F) At skeletal maturity the sphericity of the femoral head was quite good in relation to the acetabulum, although the head clearly was larger and the lateral half was uncovered.

365

366

CHAPTER 4 9 Legg--Calve--Pe~hes Disease

There seems to be little objective support for any uniform or dogmatic approach to the Legg-Perthes disorder. What follows is an overview of approaches to treatment based on our personal experience and particularly on the extensive review conducted in this chapter. (1) Comparative studies from several centers as well as intuitive feelings about the femoral head and acetabular responses lead us to continue to support the value of treating many, although not all, patients with this disorder. The data supporting the value of treatment are by no means rigorous and definitive, and even today one can make a reasonably strong case that nontreated patients considered in large numbers do as well as the broad spectrum of treated patients. The crucial issues in terms of treatment therefore continue to be those patients needing treatment, the appropriate time for initiating treatment, the specific type of treatment, and the length of time the treatment is continued. (2) The "no treatment" approach in a strict sense simply allows the child to remain active except when, particularly in the early phases of the disorder, discomfort is such that some form of rest, crutch walking, and analgesia is performed. Closely allied with this approach is that referred to as "symptomatic treatment only" in which the child is followed quite carefully but symptoms of diminished range of motion and discomfort only are treated again specifically with bed rest, crutches, and analgesia when needed. Most centers will now accept that epidemiologic studies have defined two groups of patients who can be managed successfully in this way. These are considered to be those at the extreme youngest end of the disease spectrum at 3-4 years of age and those with partial or anterior head involvement as defined by O'Garra and categorized by Catterall as group I. Virtually all series report excellent results with this group with or without treatment. In a strict sense, therefore, "symptomatic treatment only" implies a higher degree of involvement than the "no treatment" approach, with frequent clinical assessments and radiographs performed. As long as the patient has a full or virtually full range of motion, no limp, and no discomfort, ambulation is allowed. Free activity is interrupted only by the presence of symptoms at which time conservative measures are instituted for the few days to few weeks needed for the symptoms to disappear. Those particularly unimpressed by the more rigorous brace and surgical therapies will extend symptomatic treatment to older patients and to those with more advanced radiographic gradings employing rest and decreased activity only based on discomfort and limping. (3) Even allowing for marked differences in clinical and radiographic interpretations of the results, certain observations appear accurate in relation to the varying modes of treatment performed during this century. When there is strict adherence to the principles of therapy outlined in the various sectors, comparable results can be found relating to extremely prolonged recumbency, often from 1-1.5 to 3 or more years, containment therapy with recumbency again for

prolonged periods of time, containment therapy that is nonweight-bearing with some type of brace, and containment therapy with weight bearing as is done in early surgical intervention with either the proximal femoral varus or innominate osteotomy. A review of the papers reported in this chapter in each of several treatment groups frequently will show good to excellent results in 60%, fair results in 25%, and poor results in 15%. There are two possible interpretations of similar values with varying techniques: each of the techniques has a degree of similar effectiveness or, on a negative note, each of the treatments is of equally limited value. We are willing to interpret the data, although not with complete 100% confidence, to indicate that conscientious application of several therapies will lead to reasonably good results. It is our feeling that no particular regimen has been shown definitively to be superior to any other in terms of the end result. Where differences lie, however, is in the time required for treatment and the degree to which the child has to be immobilized because of the therapy. The value, therefore, of surgical intervention involving proximal femoral varus osteotomy or innominate osteotomy is not necessarily in any documented series of improved results but rather in the fact that results similar to markedly prolonged recumbency and bracing can be achieved in a relatively short period of time. (4) The mode of deformation of the femoral head has been defined by some as mechanical in nature and by others as biological in nature, reflecting altered growth patterns due to the disease state rather than to specific pressures to which the head is subjected. In our opinion, both mechanisms are involved in causing deformation in differing stages of the disorder (Fig. 5). There is a certain degree of mechanical collapse of the subchondral bone and shortly thereafter the articular surface relatively early in the disease. The concept of the subchondral fracture is real based on both radiographic and histopathologic studies and would appear to correlate reasonably well with the symptomatic period of discomfort relatively early in the disease phase. Although the entire femoral head is not softened, it is evident from early pathologic descriptions that the subchondral region particularly in the area of the subchondral fracture is filled with necrotic material and truly is softened in a mechanical sense. The localized collapse of the femoral head thus appears mechanical in nature. Most of the deformation subsequent to that, however, appears biological in the sense that it is due to altered growth patterns as the necrotic focus, which is often the entire secondary ossification center plus the adjacent cartilage, is repaired. Histopathologic studies dating back over several decades have shown necrosis of epiphyseal and physeal cartilage as well as of the secondary ossification center bone. The value of containment therapy lies in helping to control new tissue deposition into the spherical pattern during the long repair phase and perhaps, if done early enough, in minimizing collapse due to subchondral necrosis. It is not definitive alone, however.

SECTION Xl ~ Treatment Approaches to Lecjcj--Perthes Disease (5) Many studies document well the fact that remodeling of both the femoral head and the acetabulum, responding to the abnormal shape of the femoral head, continues until skeletal maturity. Analyses that cease with growth remaining, even though there has been complete reossification of the secondary center of the femoral head, often overstate the deformation because in the remaining few years of growth a form of acceptable biological congruency can be reestablished. At a minimum, therefore, follow-up of results should assess the femoral head and acetabular shape at skeletal maturity. (6) A large number of papers define better results as occurring in the younger age groups. This remains one of the best documented findings throughout the Perthes literature. Also present in some papers is the fact that treatment results are better the earlier into the disease state intervention is begun (68). Indeed, in the vast majority of pediatric orthopedic disorders, early diagnosis and early treatment lead to improved results. In the Perthes condition, however, often there has been a reluctance to intervene early because many patients seemingly do well with minimal to no treatment. In our opinion this delay has allowed deformation to occur in several patients such that treatment is begun when a deformity that cannot be fully repaired already is present. It is our opinion, based on the extensive review of the literature in this chapter plus reasoning in terms of the appropriate time for intervention, that the best time to begin definitive treatment is when the sphericity of the femoral head is still intact, which means at the time that diagnosis has been made. It is well-demonstrated, for example, in tibia vara (Blount disease) that surgery between 2 and 4 years of age leads to a far higher number of good and excellent results and far fewer recurrences of deformity than those in whom surgery is delayed beyond 4 years of age. Regardless of the mode of treatment decided upon, therefore, it appears to us that results would be markedly improved if treatment was begun and rigidly adhered to while femoral head sphericity was still present. If a patient is observed for 6, 8, or 12 months or longer, in particular to determine the category of classification into which he or she will evolve, a golden time frame often is lost such that treatment is begun once deformity already has occurred. Should the patient remain in one of the better categories, then the period of observation or use of minimally effective treatment is not a problem. Should the patient, however, progress from a spherical femoral head to a flattened cartilage model of the femoral head under observation, much would have been lost during the period of observation or utilization of ineffective therapy. The key to a more aggressive management program, therefore, is careful assessment of clinical symptoms, involving limp, pain, and decreased range of motion, and increased use of imaging modalities such as arthrography, CT and MR imaging, and three-dimensional reconstructions to define the shape of the cartilage model of the femoral head and its relation to the acetabulum.

367

If symptoms persist or recur quickly with reambulation after an early period of observation, there should be little reluctance to proceed to the definitive therapy if one indeed is seeking to maximize results. If recumbency is chosen, the data from decades ago appear convincing that, if performed with containment, which involves some form of traction or cast device, results ultimately will be good although the time of treatment will be relatively prolonged. Indeed, the best approach to both mechanical and biological theories of causation is the use of recumbency to minimize mechanical stress along with containment to enhance biological modeling. If containment therapy alone is chosen, then there must be accurate assessment to determine that optimal containment indeed is being achieved. Such positioning occurs not just with proximal femoral abduction but also with internal rotation and slight flexion. If surgical treatment is chosen, it would appear that results would be improved with earlier intervention. At the time that the disorder occurs, even with whole head involvement, the plain radiograph is normal or close to normal and the cartilage model of the femoral head is intact and spherical. The Catterall and lateral pillar classifications, though helpful in terms of prognosis, indicate what has already happened to the femoral head. Although some patients who might do well without surgery would undergo intervention, overall the results should be markedly improved if one assumes that the theory underlying the treatment is valid, namely, the effect of template molding of the femoral head shape by the acetabulum based on the containment principle. This approach, therefore, stresses early diagnosis and, if commitment is made to a surgical program, early performance of surgery once the immediate symptoms have quietened. Observation of a patient over a severalmonth period, much of which is symptomatic, allows the golden opportunity to pass when intervention on a spherical head could have been achieved. We thus feel that observation, investigation, and aggressive management should be telescoped into the earliest framework of the disorder, quietening the symptoms with immobilization and then repositioning the head in its still spherical state. The possibility exists that the outcome of Legg-Perthes disorder is determined very early after the subchondral necrosis is established, perhaps over a few-day to few-week period, even though ultimate repair occurs over several months to several years and is not really complete until skeletal maturity. It is well-documented that there is no need to continue therapy until there has been full reossification of the secondary center. Treatment regimens even with bracing and recumbency, therefore, have been markedly shortened with no apparent deleterious effects. This process of gradually shortening treatment regimens has gone on for the past few decades. It is at the opposite end of the spectrum, the early beginnings of the disorder, where little attention has been paid. This is understandable because initial occurrence of the disorder universally is believed to be asymptomatic, and even early symptoms are relatively mild such that

368

CHAPTER 4 ~ Legff-Calve-Perthes Disease

patients present not infrequently with the fragmentation phase already underway. Greater attention to earlier symptoms, in particular those that appear to be associated with the subchondral fracture, might pay benefits in terms of earlier treatment. Indeed, there is a meaningful subset of patients who have considerable discomfort and limping, much of which may be associated with the subchondral fracture. It is important to recognize that plain radiographs do not necessarily show the infraction, and MR imaging appears to increase the recognition of this component of the disorder. In most theories of pathogenesis of deformity, it is the subchondral fracture and subsequent radiolucency associated with it that play a major role in articular surface deformity. Therefore, diagnosis of the fracture or of the subchondral crescent sign should mandate immediate non-weight-bearing, protection of the hip by recumbency and gentle abduction, and an early decision regarding definitive therapy. In most orthopedic departments, the time after diagnosis often involves a severalweek period in which either a decision is made about brace usage or consideration is given to discussions of surgery, but the hip itself is not particularly protected. Although we currently have no control over necrosis of epiphyseal and physeal cartilage, we might be better able to control the repair of the necrotic bone, which still serves as a major support structure for the spherical femoral head. It is clear that multiple insults of ischemia occur in Legg-Perthes because histopathologic sections show woven bone, which at this age clearly is a sign of repair bone, that subsequently becomes necrotic itself. What is not clear is the reason for either the initial or the subsequent ischemic event. It is conceivable that repair bone necrosis occurs due to instability and weakness of the bone matrix of the secondary center, such that with weight bearing the fragile revascularization process is damaged mechanically, leading to additional episodes of ischemia. One positive feature of Legg-Perthes disease is the fact that the bone of the femoral head eventually will repair itself fully. This is not true in the adult counterparts of avascular necrosis. The problem during the childhood years is that repair is exceedingly slow, and many of the structural features of Legg-Perthes disorder result from poorly controlled, excessive, and asymmetric repair such that the cartilage models and ultimately the bone models of femoral head and acetabulum are imperfect. Whereas the ultimate improvement in Legg-Perthes management would be detection and early treatment of the primary cause, it appears conceivable that, in the absence of such events, hastening and better control of the repair process would improve results. The previous discussion reviewed our recommended approach. Much of the difficulty with repair stems from the relatively large amount of bone of the secondary ossification center that becomes necrotic and the relatively slow rate of dead bone resorption and new bone deposition due to the fact that the blood supply to the region is tenuous. Thus, it is attractive theoretically to remove the necrotic bone surgi-

cally and insert materials to enhance repair quickly. This also would require maintenance and perhaps augmentation of the blood supply. Hastening of bone repair was attempted by Haywood and others several decades ago, as reviewed by Edgren (68), but no definitive value was found. Any surgical enucleation clearly would have to be utilized in conjunction with hip spica cast immobilization as the femoral head would become markedly more fragile in terms of shape maintenance than currently is the case. If repair could be enhanced such that healing occurred over a several-week to few-month framework, as occurs in fracture or osteotomy, then such hip spica immobilization would be acceptable. Insertion of structural composites and bone growth factors would have to be shown to be extremely effective in experimental animals before such intervention is warranted because the capacity for damaging the femoral head to an extent even greater than occurs by the natural history of the disease is present. The finding that multiple episodes of infarction characterize some cases of Legg-Perthes disease further complicates primary intervention efforts. If the repeating episodes are due to trauma of fragile repair tissue, then an approach of quicker, more biological repair would not be a problem. If it was due to other more intrinsic problems, then repair tissue still would be subject to repeat damage. Nevertheless, a biological solution seems feasible based on the following criteria. (1) the disorder eventually heals itself and enhancement of bone healing is feasible seeking to quicken the rate of repair; (2) the arterial blood supply is intact; (3) surgical drilling of the metaphysis can diminish venous hypertension; (4) removal of the necrotic secondary ossification center bone in those with whole head involvement with minimal damage to articular, epiphyseal, and physeal cartilages technically is feasible and should shorten the fragmentation phase of repair wherein all necrotic bone of the head eventually must be removed; (5) protection of the sphericity of the articular cartilage in hip spica is feasible, but only if relatively rapid repair of the secondary center can be induced; (6) bone growth factors such as bone morphogenetic and angiogenic proteins can induce more rapid repair; and (7) structural composites can provide a scaffolding both to support the femoral head outer cartilage model and to provide a framework for osteoid deposition.

References 1. Allen BL, Jr (1997) Graphic analysis of femoral growth in young children with Perthes disease. J Pediatr Orthop 17: 255-263. 2. Allison N, Moody EF (1915) Osteochondritis deformans juvenilis (Perthes disease). Am J Orthop Surg 13:No. 2. 3. Apley AG, Wientraub S (1981) The sagging rope sign in Perthes disease and allied disorders. J Bone Joint Surg 63B: 43-47. 4. Arnoldi CC, Linderholm H (1971) Intercalcanean pressure in various forms of dysfunction of the venous pump of the calf. Acta Chir Scand 137:21-27.

References

5. Axer A (1965) Subtrochanteric osteotomy in the treatment of Perthes' disease: A preliminary report. J Bone Joint Surg 47B:489-499. 6. Axer A, Schiller MG (1972) The pathogenesis of the early deformity of the capital femoral epiphysis in Legg-CalvePerthes syndrome (LCPS). Clin Orthop Rel Res 84:106-115. 7. Axer A, Schiller MG, Segal D, Rzetelny V, Gershuni-Gordon DH (1973) Subtrochanteric osteotomy in the treatment of Legg-Calve-Perthes' syndrome. Acta Orthop Scand 44: 31-54. 8. Axer A, Gershuni DH, Hendel D, Mirovski Y (1980) Indications for femoral osteotomy in Legg-Calve-Perthes disease. Clin Orthop Rel Res 150:78-87. 9. Axhausen G (1923) Der anatomische krankheitsablauf bei der koehlerschen krankheit der metatarsalkopfchen und der perthesschen krankheit des huftkopfes. Archiv f Klin Chir 124:530-542. 10. Axhausen G (1924) Uber den abgrenzungsvorgang am epiphysaren knochen (Osteochondritis dissecans Konig). Virch Arch Path Anat 252:458-518. 11. Axhausen G (1975) The Classic: Clinical and experimental observations about the etiology of arthritis deformans and especially the significance of the aseptic bone and cartilage necrosis. Clin Orthop Rel Res 106:2-6. 12. Barer M (1978) Role of innominate osteotomy in the treatment of children with Legg-Perthes disease. Clin Orthop Rel Res 135:82-89. 13. Barnes JM (1980) Premature epiphyseal closure in Perthes' disease. J Bone Joint Surg 62B:432-437. 14. Bennett JT, Mazurek RT, Cash JD (1991) Chiari's osteotomy in the treatment of Perthes' disease. J Bone Joint Surg 73B: 225-228. 15. Bergmann E (1927) Theoretisches, klinisches und experimentelles zur Frage der aseptischen Knochennekrosen. Deutsche Ztschr f Chir 206:12-87. 16. Bergstrand I, Norman O (1961) Die Krankheiten des Huftgelenks im Kindesalter. Der Radiologe 1:76. 17. Bernbeck R (1951) Zur Pathogenese der jugendlichen huftkopfnekrose. Archiv fur Orthop Unfallchir 44:164-200. 18. Bernbeck R (1951) Kritisches zum Perthes-Problem der Hufte. Archiv fur Orthop Unfallchir 44:445-472. 19. Blanchard W (1917) Treated and untreated osteochondritis juvenilis of the hip. J Am Med Assoc 69:1060-1063. 20. Bloch-Michel H, Benoist M, Salomon A, Duriez J, Morel G, Heripret G (1960) Consequences chez 1-adulte de l'osteochondrite de l'hanche. Rhumatologie 4:170-195. 21. Bohr HH (1979) Skeletal maturation in Legg-Calve-Perthes' disease. Internat Orthop 2:277-281. 22. Bohr HH (1980) On the development and course of LeggCalve-Perthes disease (LCPD). Clin Orthop Rel Res 150: 30-35. 23. Bos CFA, Bloem JL, Bloem RM (1991) Sequential magnetic resonance imaging in Perthes' disease. J Bone Joint Surg 73B:219-224. 24. Bowen JR, Schreiber FC, Foster BK, Wein BK (1982) Premature femoral neck physeal closure in Perthes' disease. Clin Orthop Rel Res 171:24-29. 25. Bowen JR, Kumar VP, Joyce JJ, III, Bowen JC (1986) Osteochondritis dissecans following Perthes' disease. Arthroscopic operative treatment. Clin Orthop Rel Res 209:49-56.

369

26. Brailsford JF (1935) Osteochondritis. Br J Radiol 8:87-134. 27. Broder H (1953) The late results in Legg-Perthes' disease and factors influencing them. A study of 102 cases. Bull Hosp Joint Dis 14:194-216. 28. Brotherton BJ, McKibbin B (1977) Perthes disease treated by prolonged recumbency and femoral head containment: A long-term appraisal. J Bone Joint Surg 59B:8-14. 29. Burwell RG, Dangerfield PH, Hall DJ, Vernon CL, Harrison MHM (1978) Perthes' disease: An anthropometric study revealing impaired and disproportionate growth. J Bone Joint Surg 60B:461-477. 30. Butel PJ, Borgi R, Oberlin P (1980) L'association du traitement chirurgical a la decharge ambulatoire dans la maladie de Legg-Calve-Perthes: Experience de quinze ans portant sur 112 cas. Acta Orthop Belg 46:423-427. 31. Caffey J (1968) The early roentgenographic changes in essential coxa plana: Their significance in pathogenesis. Am J Roent 103:620-634. 32. Cahuzac JP, deGauzy JS, Vidal H, Gaubert J (1992) The acetabular opening angle in Perthes' disease: Radiographic study of 62 unilateral cases. Acta Orthop Scand 63:278-281. 33. Calve J (1910) Sur une forme particuliere de pseudo-coxalgie. Rev Chir 42:54-84. 34. Calve J (1921) Osteochondritis of the upper extremity of the femur. J Orthop Surg 3:489-492. 35. Calve J (1921) Coxa plana. La Press Med 29:383-385. 36. Calve J (1921) Quelques considerations sur la coxa plana denommee a tort osteochondrite de l'epiphyse superieure du femur. La Press Med 29:828-829. 37. Canale ST, D'Anca AF, Cotler JM, Snedden HE (1972) Innominate osteotomy in Legg-Calve-Perthes disease. J Bone Joint Surg 54A:25-40. 38. Canario AT, Williams L, Wientroub S, Catterall A, LloydRoberts GC (1980) A controlled study of the results of femoral osteotomy in severe Perthes' disease. J Bone Joint Surg 62B:438-440. 39. Cannon SR, Pozo JL, Catterall A (1989) Elevated growth velocity in children with Perthes' disease. J Pediatr Orthop 9:285-292. 40. Catterall A (1971) The natural history of Perthes' disease. J Bone Joint Surg 53B:37-53. 41. Catterall A (1980) Natural history, classification and X-ray signs in Legg-Calve-Perthes' disease. Acta Orthop Belg 46:346-351. 42. Catterall A (1981) Legg-Calve-Perthes syndrome. Clin Orthop Rel Res 158:41-52. 43. Catterall A (1982) Legg-Calve-Perthes' Disease. Edinburgh: Churchill Livingstone. 44. Catterall A, Pringle J, Byers PD, Fulford GE, Kemp HBS, Dolman CL, Bell HM, McKibbin B, Ralis Z, Jensen OM, Lauritzen J, Ponseti IV, Ogden PJ (1982) A review of the morphology of Perthes' disease. J Bone Joint Surg 64B:269-275. 45. Catterall A, Pringle J, Byers PD, Fulford GE, Kemp HBS (1982) Perthes' disease: Is the epiphysial infarction complete? A study of the morphology in two cases. J Bone Joint Surg 64B:276-281. 46. Catterall A (1986) Adolescent hip pain after Perthes' disease. Clin Orthop Rel Res 209:65-69. 47. Chivabongs A (1972) Verlauft die Perthes'sche erkrankung vorwiegend einseitig? Z Orthop 11:418-425.

370

CHAPTER 4 ~ Le~ty--Caive--Pe~hes Disease

48. Christensen F, Soballe K, Ejsted R, Luxhoj T (1986) The Catterall classification of Perthes' disease: An assessment of reliability. J Bone Joint Surg 68B:614-615. 49. Clancy M, Steel HH (1985) The effect of an incomplete intertrochanteric osteotomy on Legg-Calve-Perthes' disease. J Bone Joint Surg 67A:213-216. 50. Clarke NMP, Harrison MHM, Keret D (1983) The sagging rope sign: A critical appraisal. J Bone Joint Surg 65B: 285-287. 51. Clarke NMP, Hal~son MHM (1983) Painful sequelae of coxa plana. J Bone Joint Surg 65A:13-18. 52. Clothier JC (1979) The behavior of upper femoral osteotomies performed for Perthes' disease. J Bone Joint Surg 61B: 517-518. 53. Coates CJ, Paterson JMH, Woods KR, Catterall A, Fixsen JA (1990) Femoral osteotomy in Perthes' disease: Results at maturity. J Bone Joint Surg 72B:581-585. 54. Cooperman DR, Stulberg SD (1986) Ambulatory containment treatment in Perthes' disease. Clin Orthop Rel Res 203: 289-300. 55. Cordeiro EN (1980) Femoral osteotomy in Legg-CalvePerthes disease. Clin Orthop Rel Res 150:69-72. 56. Curtis BH, Gunther SF, Gossling HR, Paul SW (1974) Treatment for Legg-Perthes disease with the Newington ambulationabduction brace. J Bone Joint Surg 56A:1135-1146. 57. Danforth MS (1934) The treatment of Legg-Calve-Perthes disease without weight beating. J Bone Joint Surg 16:516-534. 58. DeGuembecker W, Duriez J (1981) Osteochondrite de la hanche et maturation osseuse. Rev Chir Orthop 67:473-478. 59. Dekker M, van Rens TJG, Slooff TJJH (1981) Salter's pelvic osteotomy in the treatment of Perthes' disease. J Bone Joint Surg 63B:282. 60. Delchef J (1926) Apropos de l'etiologie et de la pathogenie de l'osteochondrite de la hanche. Relation d'un cas. Rev Orthop 13:5-13. 61. Dickens DRV, Menelaus MB (1978) The assessment of prognosis in Perthes' disease. J Bone Joint Surg 60B:189-194. 62. Dolman CL, Bell HM (1973) The pathology of LeggCalve-Perthes disease: A case report. J Bone Joint Surg 55A:184-188. 63. Doudoulakis JK (1991) Trochanteric advancement for premature arrest of the femoral head growth plate: 6-year review of 30 hips. Acta Orthop Scand 62:92-94. 64. Duthie RB, Houghton GR (1981) Constitutional aspects of the osteochondroses. Clin Orthop Rel Res 158:19-27. 65. Eaton GO (1967) Long term results of treatment in coxa plana. J Bone Joint Surg 48A:1031-1042. 66. Eckerwall G, Hochbergs P, Wingstrand H, Egund N (1997) Magnetic resonance imaging and early remodeling of the femoral head after femoral varus osteotomy in Legg-CalvePerthes disease. J Pediatr Orthop 6B:239-244. 67. Eckerwall G, Wingstrand H, Haaglund G, Karlberg J (1996) Growth in 110 children with Legg-Calve-Perthes' disease: A longitudinal infancy childhood puberty growth model study. J Pediatr Orthop 5B: 181-184. 68. Edgren W (1965) Coxa plana: A clinical and radiological investigation with particular reference to the importance of the metaphyseal changes for the final shape of the proximal part of the femur. Acta Orthop Scand Supp 84:1-129.

69. Ellis W (1984) Metaphysis in Perthes disease: A method of assessment and selection for treatment. J Pediatr Orthop 4: 731-734. 70. Evans DL (1958) Legg-Calve-Perthes' disease: A study of late results. J Bone Joint Surg 40B:168-181. 71. Evans IK, Deluca PA, Gage JR (1988) A comparative study of ambulation-abduction bracing and varus derotation osteotomy in the treatment of severe Legg-Calve-Perthes disease in children over 6 years of age. J Pediatr Orthop 8: 676-682. 72. Eyre-Brook AL (1936) Osteochondritis deformans coxae juvenilis or Perthes's disease: The results of treatment by traction in recumbency. Br J Surg 24:166-182. 73. Fabry G (1980) Anteversion in Legg-Calve-Perthes' disease. Acta Orthop Belg 46:352-354. 74. Farsetti P, Tudisco C, Caterini R, Potenza V, Ippolito E (1995) The Herrring lateral pillar classification for prognosis in Perthes disease: Late results in 49 patients treated conservatively. J Bone Joint Surg 77B:739-742. 75. Ferguson AB, Howorth MB (1934) Coxa plana and related conditions at the hip. J Bone Joint Surg 16:781-803. 76. Ferguson AB, Jr (1975) The pathology of Legg-Perthes disease and its comparison with aseptic necrosis. Clin Orthop Rel Res 106:7-18. 77. Fisher RL (1972) An epidemiological study of Legg-Perthes disease. J Bone Joint Surg 54A:769-778. 78. Fisher RL, Roderique JW, Brown DC, Danigelis JA, Ozonoff MB, Sziklas JJ (1980) The relationship of isotopic bone imaging findings to prognosis in Legg-Perthes disease. Clin Orthop Rel Res 150:23-29. 79. Flemming Moiler P (1926) The clinical observations after healing of Calve-Perthes' disease compared with the final deformities left by that disease, and the bearing of those find deformities on the ultimate prognosis. Acta Radiol 5:1-36. 80. Freehafer AA (1960) Osteochondritis dissecans following Legg-Calve-Perthes disease. A report of one case. J Bone Joint Surg 42A:777-782. 81. Freund E (1930) Zur deutung des rontgenbildes der Perthesschen krankheit. Fortschr Roent 42:435-464. 82. Fulford GE, Lunn PG, Macnicol MF (1993) A prospective study of nonoperative and operative management for Perthes' disease. J Pediatr Orthop 13:281-285. 83. Futami T, Kasahara Y, Suzuki S, Ushikubo S, Tsuchiya T (1991) Ultrasonography in transient synovitis and early Perthes' disease. J Bone Joint Surg 73B:635-639. 84. Gage HC (1933) A possible early sign of Perthes' disease. Br J Radiol 6:295-297. 85. Gall EA, Bennett GA (1942) Osteochondritis deformans of the hip (Legg-Perthes disease) and renal osteitis fibrosa cystica. Arch Path 33:866-878. 86. Gallagher JM, Weiner DS, Cook AJ (1983) When is arthrography indicated in Legg-Calve-Perthes disease? J Bone Joint Surg 65A:900-905. 87. Garceau GJ (1964) Surgical treatment of coxa plana. J Bone Joint Surg 46B:779-780. 88. Gill AB (1940) Legg-Perthes disease of the hip: Its early roentgenographic manifestations and its cyclical course. J Bone Joint Surg 23:1013-1047.

References

89. Girandy BR, Osman MZ (1968) Longitudinal growth and skeletal maturation in Perthes disease. Radiol Clin N Am 6: 245-251. 90. Glueck CJ, Crawford A, Roy D, Freiberg R, Glueck H, Stroop D (1996) Association of antithrombotic factor deficiencies and hypofibrinolysis with Legg-Perthes disease. J Bone Joint Surg 78A:3-13. 91. Goff CW (1962) Legg-Calve-Perthes syndrome (LCPS): An up-to-date critical review. Clin Orthop Rel Res 22: 93-107. 92. Gossling HR (1972) Legg-Perthes disease. Part II. Treatment by recumbency. In: Instructional Course Lectures, American Academy of Orthopedic Surgeons, vol. 22. pp. 296-301, St. Louis: CV Mosby. 93. Gower WE, Johnston RC (1971) Legg-Perthes disease. Long-term follow-up of 36 patients. J Bone Joint Surg 53A: 759-768. 94. Green NE, Beauchamp RD, Griffin PP (1981) Epiphyseal extrusion as a prognostic index in Legg-Calve-Perthes disease. J Bone Joint Surg 63A:900-905. 95. Green NE, Griffin PP (1982) Intraosseous venous pressure in Legg-Perthes disease. J Bone Joint Surg 64A:666-671. 96. Grossbard GD (1981) Hip pain during adolescence after Perthes' disease. J Bone Joint Surg 64B:572-574. 97. Gunther SF, Gossling HR (1972) Legg-Perthes disease. Part IV. Treatment by abduction-ambulation. In: Instructional Course Lectures, American Academy of Orthopedic Surgeons, vol. 22. pp. 305-309, St. Louis: CV Mosby. 98. Halkier E (1956) The "tear-shaped phenomenon" in CalvePerthes' disease. Acta Orthop Scand 25:287-295. 99. Hallel T, Salvati EA (1976) Osteochondritis dissecans following Legg-Calve-Perthes disease. J Bone Joint Surg 58A: 708-711. 100. Haraldsson S (1973) Derotation varisation osteotomy in the treatment of Perthes' disease. Acta Orthop Scand 44: 105-108. 101. Hardcastle PH, Ross R, Hamalainen M, Mata A (1980) Catterall grouping of Perthes' disease: An assessment of oberserver error and prognosis using the Catterall classification. J Bone Joint Surg 62B:428-431. 102. Harper PS, Brotherton BJ, Cochlin D (1976) Genetic risks in Perthes' disease. Clin Genet 10:178-182. 103. Harrison MHM, Menon MPA (1966) Legg-Calve-Perthes disease: The value of roentgenographic measurement in clinical practice with special reference to the broomstick plaster method. J Bone Joint Surg 48A:1301-1318. 104. Harrison MHM, Turner MH, Nicholson FJ (1969) Coxa plana: Results of a new form of splinting. J Bone Joint Surg 51A: 1057-1069. 105. Harrison MHM, Turner MH, Jacobs P (1976) Skeletal immaturity in Perthes' disease. J Bone Joint Surg 58B:37-40. 106. Harrison MHM, Blakemore ME (1980) A study of the "normal" hip in children with unilateral Perthes' disease. J Bone Joint Surg 62B:31-36. 107. Harrison MHM, Burwell RG (1981) Perthes' disease: A concept of pathogenesis. Clin Orthop Rel Res 156:115-127. 108. Harrison MHM, Turner MH, Smith DN (1982) Perthes' disease: Treatment with the Birmingham splint. J Bone Joint Surg 64B:3-11.

371

109. Harrison MHM (1986) A preliminary account of the management of the painful hip originating from Perthes' disease. Clin Orthop Rel Res 209:57-64. 110. Harry JD, Gross RH (1987) A quantitative method for evaluating results of treating Legg-Perthes sydrome. J Pediatr Orthop 7:671-676. 111. Haythorn SR (1949) Pathological changes found in material removed at operation in Legg-Calve-Perthes disease. J Bone Joint Surg 31A:599-611. 112. Heikkinen E, Puranen J (1980) Evaluation of femoral osteotomy in the treatment of Legg-Calve-Perthes disease. Clin Orthop Rel Res 150:60-68. 113. Heitzmann O (1923) Epiphysenerkrankungen im wachstumsalter. Klin Wochenschr 2:397-399. 114. Henderson RC, Renner JB, Sturdivant MC, Greene WB (1990) Evaluation of magnetic resonance imaging in LeggPerthes disease: A prospective, blinded study. J Pediatr Orthop 10:289-297. 115. Herndon CH, Heyman CH (1952) Legg-Perthes disease: An evaluation of treatment by traction and ischial weight-bearing brace. J Bone Joint Surg 34A:25-46. 116. Herring JA, Lundeen MA, Wenger DR (1980) Minimal Perthes' disease. J Bone Joint Surg 62B:25-30. 117. Herring JA, Neustadt JB, Williams JJ, Early JS, Browne RH (1992) The lateral pillar classification of Legg-CalvePerthes disease. J Pediatr Orthop 12:143-150. 118. Herring JA, Williams JJ, Neustadt JN, Early JS (1993) Evolution of femoral head deformity during the healing phase of Legg-Calve-Perthes disease. J Pediatr Orthop 13:41-45. 119. Herring JA (1994) The treatment of Legg-Calve-Perthes disease: A critical review of the literature. J Bone Joint Surg 76A:448-458. 120. Herring JA (1996) Management of Perthes' disease. J Pediatr Orthop 16:1-2. 121. Herzog FG (1950) A study of Perthes' disease. Proc Roy Soc Med 54:1102. 122. Heyman CH, Herndon CH (1950) Legg-Perthes disease. A method for measurement of the roentgenographic result. J Bone Joint Surg 32A:767-778. 123. Hirohashi K, Kanbara T, Kuroda K, Okajima M, Hyashi M, Shimazu A (1980) Perthes' diseasemA classification based on the extent of epiphyseal and metaphyseal involvement. Internat Orthop 4:47-55. 124. Hoffinger SA, Henderson RC, Renner JB, D'Ales MC, Rab GT (1993) Magnetic resonance evaluation of "metaphyseal" changes in Legg-Calve-Perthes' disease. J Pediatr Orthop 13:602-606. 125. Hoikka V, Lindholm TS, Poussa M (1986) Intertrochanteric varus osteotomy in Legg-Calve-Perthes disease: A report on 112 hips. J Pediatr Orthop 6:600-604. 126. Holdsworth FW (1966) Epiphyseal growth: Speculations on the nature of Perthes's disease. Ann Roy Coil Surg Engl 39:1-16. 127. Ingman AM, Paterson DC, Sutherland AD (1982) A comparison between innominate osteotomy and hip spica in the treatment of Legg-Perthes' disease. Clin Orthop Rel Res 163:141-147. 128. Inoue A, Freeman MAR, Vernon-Roberts B, Mizuno S (1976) The pathogenesis of Perthes' disease. J Bone Joint Surg 58B: 453-461.

372

CHAPTER 4 ~

Le~lg--Calve--Perrhes Disease

129. Inoue A, Ono K, Takaoka K, Yoshioka T, Hosoya T (1980) A comparative study of histology in Perthes' disease and idiopathic avascular necrosis of the femoral head in adults (IANF). Internat Orthop 4:39-46. 130. Ippolito E, Tudisco C, Farsetti P (1985) Long-term prognosis of Legg-Calve-Perthes disease developing during adolescence. J Pediatr Orthop 5:652-656. 131. Jani LFH, Dick W (1980) Results of three different therapeutic groups in Perthes' disease. Clin Orthop Rel Res 150:88-94. 132. Jensen OM, Lauritzen J (1976) Legg-Calve-Perthes' disease: Morphological studies in two cases examined at necropsy. J Bone Joint Surg 58B:332-338. 133. Jonsater S (1953) Coxa plana: A histo-pathologic and arthrographic study. Acta Orthop Scand Supp 12:1-98. 134. Joseph B (1989) Morphological changes in the acetabulum in Perthes' disease. J Bone Joint Surg 71B:756-763. 135. Kahmi E, MacEwen GD (1975) Treatment of Legg-CalvePerthes disease. Prognostic value of Catterall's classification. J Bone Joint Surg 57A:651-654. 136. Kallio P, Ryoppy S (1985) Hyperpressure in juvenile hip disease. Acta Orthop Scand 56:211-214. 137. Kallio E Ryoppy S, Kunnamo I (1986) Transient synovitis and Perthes' disease: Is there an etiological connection? J Bone Joint Surg 68B:808-811. 138. Kamegaya M, Shinada Y, Moriya H, Tsuchiya K, Akita T, Someya M (1992) Acetabular remodelling in Perthes' disease after primary healing. J Pediatr Orthop 12:308-314. 139. Karadimas JE (1971) Conservative treatment of coxa plana. A comparison of the results of different methods. J Bone Joint Surg 53A:315-325. 140. Karpinski MRK, Newton G, Henry APJ (1986) The results and morbidity of varus osteotomy for Perthes' disease. Clin Orthop Rel Res 209:30-40. 141. Katz JF (1965) "Abortive" Legg-Calve-Perthes disease or developmental variation in epiphyseogenesis of the upper femur. J Mt Sinai Hosp NY 32:651-659. 142. Katz JF (1967) Conservative treatment of Legg-CalvePerthes disease. J Bone Joint Surg 49A: 1043-1051. 143. Katz JF (1968) Arthrography in Legg-Calve-Perthes disease. J Bone Joint Surg 50A:467-472. 144. Katz JF (1970) Legg-Calve-Perthes disease. The role of distortion of normal growth mechanisms in the production of deformity. Clin Orthop Rel Res 71:193-198. 145. Katz JF, Siffert RS (1975) Capital necrosis, metaphyseal cyst, and subluxation in coxa plana. Clin Orthop Rel Res 106:75-85. 146. Katz JF, Siffert RS (1977) Skeletal maturity in Legg-CalvePerthes disease. Determination based on bone age of carpal centres. Internat Orthop 1:227-230. 147. Katz JF (1980) Late modeling changes in Legg-CalvePerthes disease (LCPD) with continuing growth to maturity. Clin Orthop Rel Res 150:115-124. 148. Kelly FB, Jr, Canale ST, Jones RR (1980) Legg-CalvePerthes' disease: Long-term evaluation of noncontainment treatment. J Bone Joint Surg 62A:400-407. 149. Kemp HBS (1973) Perthes' disease: An experimental and clinical study. Ann Roy Coll Surg Engl 52:18-35. 150. Kemp HBS (1981) Perthes' disease: The influence of intracapsular tamponade on the circulation in the hip joint of the dog. Clin Orthop Rel Res 156:105-114.

151. Kemp HS, Boldero JL (1966) Radiological changes in Perthes disease. Br J Radiol 39:744-760. 152. Kendig RJ, Evans GA (1986) Biologic osteotomy in Perthes disease. J Pediatr Orthop 6:278-284. 153. Kenzora JE, Steele RE, Yosipovitch ZH, Glimcher MJ (1978) Experimental osteonecrosis of the femoral head in adult rabbits. Clin Orthop Rel Res 130:8-46. 154. Keret D, Harrison MHM, Clarke NMP, Hall DJ (1984) Coxa planamThe fate of the physis. J Bone Joint Surg 66A: 870-877. 155. Kiepurska A (1991) Late results of treatment in Perthes' disease by a functional method. Clin Orthop Rel Res 272: 76-81. 156. Kim HT, Wenger DR (1997) "Functional retroversion" of the femoral head in Legg-Calve-Perthes disease and epiphyseal dysplasia: Analysis of head-neck deformity and its effect on limb position using three-dimensional computed tomography. J Pediatr Orthop 17:240-246. 157. Kim HT, Wenger DR (1997) Surgical correction of "functional retroversion" and "functional coxa vara" in late LeggCalve-Perthes disease and epiphyseal dysplasia: Correction of deformity defined by new imaging modalities. J Pediatr Orthop 17:247-254. 158. Klisic P, Blazevic U, Seferovic O (1980) Approach to treatment of Legg-Calve-Perthes disease. Clin Orthop Rel Res 150:54-59. 159. Konjetzny GE (1934) Zur pathologie und pathologischen anatomie der Perthes-Calve'schen krankheit (osteochondritis coxae deformans juvenilis). Acta Chir Scand 74:361-378. 160. Kristmundsdottir F, Burwell RG, Hall DJ, Marshall WA (1986) A longitudinal study of carpal bone development in Perthes' disease: Its significance for both radiologic standstill and bilateral disease. Clin Orthop Rel Res 209: 115-123. 161. Kristmundsdottir F, Burwell RG, Harrison MHM (1987) Delayed skeletal maturation in Perthes' disease. Acta Orthop Scand 58:277-279. 162. Kruse RW, Guille JT, Bowen JR (1991) Shelf arthroplasty in patients who have Legg-Calve-Perthes disease: A study of long-term results. J Bone Joint Surg 73A:1338-1347. 163. Lahdes-Vasama TT, Marttinen EJ, Merikanto JEO (1997) Outcome of Perthes' disease in unselected patients after femoral varus osteotomy and splintage. J Pediatr Orthop 6B: 229-234. 164. Larsen EH, Reimann I (1973) Calve-Perthes disease, with special reference to histological observations. Acta Orthop Scand 44:426-438. 165. Laurent LE, Poussa M (1980) Intertrochanteric varus osteotomy in the treatment of Perthes' disease. Clin Orthop Rel Res 150:73-77. 166. Lauritzen J (1975) Legg-Calve-Perthes' disease. A comparative study. Acta Orthop Scand Supp 159:1-137. 167. Lee DY, Seong SC, Choi IH, Chung CY, Chang BS (1992) Changes of blood flow of the femoral head after subtrochanteric osteotomy in Legg-Perthes' disease: A serial scintigraphic study. J Pediatr Orthop 12:731-734. 168. Legg AT (1910) An obscure affection of the hip joint. Boston Med Surg J 162:202-204. 169. Legg AT (1916) Osteochondral trophopathy of the hip joint. Surg Gynec Obstet 22:307-323.

References

170. Legg AT (1918) Remarks on the etiology of the flattening of the upper femoral epiphysis: A study of seventy-five personally observed cases. Am J Orthop Surg 16:448-452. 171. Legg AT (1927) The end results of coxa plana. J Bone Joint Surg 9:26-34. 172. Legg AT (1927) In "Discussion." J Bone Joint Surg 9:36. 173. Leitch JM, Paterson DC, Foster BK (1991) Growth disturbance in Legg-Calve-Perthes disease and the consequences of surgical treatment. Clin Orthop Rel Res 262:178-184. 174. Lippmann RK (1929) The pathogenesis of Legg-CalvePerthes' disease based upon the pathologic findings in a case. Am J Surg 6:785-791. 175. Liu S-L, Ho T-C (1991) The role of venous hypertension in the pathogenesis of Legg-Perthes disease: A clinical and experimental study. J Bone Joint Surg 73A:194-200. 176. Lloyd-Roberts GC (1982) The management of Perthes' disease. J Bone Joint Surg 64B: 1-2. 177. Lloyd-Roberts GC, Catterall A, Salamon PB (1976) A controlled study of the indications for and the results of femoral osteotomy in Perthes' disease. J Bone Joint Surg 58B:31-36. 178. Lloyd-Roberts GC, Wetherill MH, Fraser M (1985) Trochanteric advancement for premature arrest of the femoral capital growth plate. J Bone Joint Surg 67B:21-24. 179. Marklund T, Tillberg B (1976) Coxa plana: A radiological comparison of the rate of healing with conservative measures and after osteotomy. J Bone Joint Surg 58B:25-30. 180. Martinez AG, Weinstein SL, Dietz FR (1992) The weightbeating abduction brace for the treatment of Legg-Perthes disease. J Bone Joint Surg 74A:12-21. 181. Matan AJ, Stevens PM, Smith JT, Santore SD (1996) Combination trochanteric arrest and intertrochanteric osteotomy for Perthes' disease. J Pediatr Orthop 16:10-14. 182. Maxted MJ, Jackson RK (1985) Innominate osteotomy in Perthes' disease: A radiological survey of results. J Bone Joint Surg 67B:399-401. 183. McAndrew MP, Weinstein SL (1984) A long-term follow-up of Legg-Calve-Perthes disease. J Bone Joint Surg 66A: 860-869. 184. McElwain JP, Regan BF, Dowling F, Fogarty E (1985) Derotation varus osteotomy in Perthes disease. J Pediatr Orthop 5:195-198. 185. McKay DW (1980) Cheilectomy of the hip. Orthop Clin N Am 11:141-160. 186. McKibbin B, Ralis Z (1974) Pathological changes in a case of Perthes' disease. J Bone Joint Surg 56B:438-447. 187. McKibbin B (1975) Recent developments in Perthes disease. In: Recent Advances in Orthopaedics, ed B McKibbin, No. 2. Edinburgh: Churchill Livingstone. 188. Meehan PL, Angel D, Nelson JM (1992) The Scottish Rite abduction orthosis for the treatment of Legg-Perthes disease: A radiographic analysis. J Bone Joint Surg 74A:2-12. 189. Menelaus MB (1986) Lessons learned in the management of Legg-Calve-Perthes disease. Clin Orthop Rel Res 209:41-48. 190. Meyer J (1966) Treatment of Legg-Calve-Perthes disease. Assessment of therapeutic results with particular reference to the value of traction in bed. Acta Orthop Scand Supp 86: 1-111. 191. Meyer J (1977) Legg-Calve-Perthes disease. Radiological results of treatment and their late clinical consequences. Acta Orthop Scand Supp 167:1-131.

373

192. Mindell ER, Sherman MS (1951) Late results in LeggPerthes disease. J Bone Joint Surg 33A: 1-23. 193. Mirovsky Y, Axer A, Hendel D (1984) Residual shortening after osteotomy for Perthes disease. J Bone Joint Surg 66B: 184-188. 194. Mizuno S, Hirayama M, Kotani PT, Simazu A (1966) Pathological histology of the Legg-Calve-Perthes disease with a special reference to its experimental production. Med J Osaka Univ 17:177-209. 195. Moberg A, Kaniklides C, Rehnberg L (1993) Arthrography in Legg-Calve-Perthes disease: Evaluation of lateral subluxations. J Pediatr Orthop 2B: 156-158. 196. Morris ML, McGibbon KC (1962)Osteochondritis dissecans following Legg-Calve-Perthes disease. J Bone Joint Surg 44B:562-564. 197. Mose K, Hjorth L, Ulfeldt M, Christensen ER, Jensen A (1977) Legg-Calve-Perthes' disease: The late occurrence of coxathrosis. Acta Orthop Scand Supp 169:1-39. 198. Mose K (1980) Methods of measuring in Legg-CalvePerthes disease with special regards to the prognosis. Clin Orthop Rel Res 150:103-109. 199. Mukherjee A, Fabry G (1990) Evaluation of the prognostic indices in Legg-Calve-Perthes disease: Statistical analysis of 116 hips. J Pediatr Orthop 10:153-158. 200. Murphy RP, Marsh HO (1978) Incidence and natural history of "head at risk" factors in Perthes' disease. Clin Orthop Rel Res 132:102-107. 201. Nagasaka K (1930) Histological study of 8 cases of osteochondritis deformans coxae juvenilis (Calve-Legg-Perthes) [English abstract] Fukuoka Acta Med 23:4. 202. Nagasaka K (1930) On the pathogenesis and etiology of the osteochondritis deformans coxae juvenilis (Calve-LeggPerthes) [English abstract] Fukuoka Acta Med 23:4. 203. Norlin R, Hammerby S, Tkaczuk H (1991) The natural history of Perthes disease. Internat Orthop 15:13-16. 204. Norman A, Bullough P (1963) The radiolucent crescent linemAn early diagnostic sign of avascular necrosis of the femoral head. Bull Hosp Jt Dis 24:99-104. 205. O'Garra JA (1959) The radiographic changes in Perthes' disease. J Bone Joint Surg 41B:465-476. 206. O'Hara JP, Davis ND, Gage JR, Sundberg AB, Winter RB (1977) Long-term follow-up of Perthes' disease treated nonoperatively. Clin Orthop Rel Res 125:49-56. 207. Osterman K, Lindholm TS (1980) Osteochondritis dissecans following Perthes' disease. Clin Orthop Rel Res 152: 247-254. 208. Paterson DC, Leitch JM, Foster B K (1991) Results of innominate osteotomy in the treatment of Legg-Calve-Perthes disease. Clin Orthop Rel Res 266:96-103. 209. Pedersen HE, McCarroll HR (1951) Treatment in LeggPerthes disease. J Bone Joint Surg 33A:591-600. 210. Perpich M, McBeath A, Kruse D (1983) Long-term followup of Perthes disease treated with spica casts. J Pediatr Orthop 3:160-165. 211. Perthes G (1910) Uber arthritis deformans juvenilis. Deut Zeit f Chir 107:111-159. 212. Perthes G (1981) The Classic. Concerning arthritis deformans juvenilis. Clin Orthop Rel Res 158:5-9. 213. Perthes G (1913) Ueber osteochondritis deformans juvenilis. Archiv Klin Chir 101:779-807.

374

CHAPTER 4 ~ Le~tg--Calve--Perrhes Disease

214. Perthes G (1924) Uber osteochondritis deformans coxae. Klin Wochenschr 3:513-516. 215. Petrie JG, Bitenc I (1971) The abduction weight-beating treatment in Legg-Perthes disease. J Bone Joint Surg 53B: 54-62. 216. Phemister DB (1921) Operation for epiphysitis of the head of the femur (Perthes disease). Arch Surg 2:221-230. 217. Pike MM (1950) Legg-Perthes disease: A method of conservative treatment. J Bone Joint Surg 32A:663-670. 218. Pinto MR, Peterson HA, Berquist TH (1989) Magnetic resonance imaging in early diagnosis of Legg-Calve-Perthes disease. J Pediatr Orthop 9:19-22. 219. Ponseti IV (1956) Legg-Perthes disease: Observations on pathological changes in two cases. J Bone Joint Surg 38A: 739-750. 220. Ponseti IV, Maynard JA, Weinstein SL, Ippolito EG, Pous JG (1983) Legg-Calve-Perthes disease: Histochemical and ultrastructural observations of the epiphyseal cartilage and physis. J Bone Joint Surg 65A:797-807. 221. Puranen J, Heikkinen E (1976) Intertrochanteric osteotomy in the treatment of Perthes' disease. Acta Orthop Scand 47: 79-88. 222. Purvis JM, Dimon JH, III, Meehan PL, Lovell WW (1980) Preliminary experience with the Scottish Rite Hospital abduction orthosis for Legg-Perthes disease. Clin Orthop Rel Res 150:49-53. 223. Quain S, Catterall A (1986) Hinge abduction of the hip. Diagnosis and treatment. J Bone Joint Surg 68B:61-64. 224. Ralston EL (1961) Legg-Calve-Perthes disease: Factors in healing. J Bone Joint Dis 43A:249-260. 225. Ratliff AHC (1967) Perthes' disease: A study of thirty-four hips observed for thirty years. J Bone Joint Surg 49B: 102-107. 226. Ratliff AHC (1967) Osteochondritis dissecans following LeggCalve-Perthes' disease. J Bone Joint Surg 49B:108-111. 227. Reichelt A, Huke B, Port J (1973) Das verhalten der handstiefskelton wicklung beijuvelinen osteochondrosen und der epiphyseolysis capitis femoris. Z Orthop 11:763-767. 228. Reinker KA, Larsen IJ (1983) Patterns of progression in Legg-Perthes disease. J Pediatr Orthop 3:455-460. 229. Reinker KA (1996) Early diagnosis and treatment of hinge abduction in Legg-Perthes disease. J Pediatr Orthop 16:3-9. 230. Richards BS, Coleman SS (1987) Subluxation of the femoral head in coxa plana. J Bone Joint Surg 69A:1312-1318. 231. Riedel G (1922) Beitrag zur pathologischen anatomie der osteochondritis deformans coxae juvenilis. Zentralbl f Chir 49:1447-1450. 232. Ritterbusch JE Shantharam SS, Gelinas C (1993) Comparison of lateral pillar classification and Catterall classification of Legg-Calve-Perthes' disease. J Pediatr Orthop 13:200-202. 233. Robichon J, Desjardins JP, Koch M, Hooper CE (1974) The femoral neck in Legg-Perthes' disease. Its relationship to epiphyseal change and its importance in early prognosis. J Bone Joint Surg 56B:62-68. 234. Robinson HJ, Jr, Putter H, Sigmond MB, O'Connor S, Murray KR (1988) Innominate osteotomy in Perthes disease. J Pediatr Orthop 8:426-435. 235. Rockemer K (1927) Zur histopathogenese der Perthesschen krankheit, an der hand eines fruhfalles. Frank Zeit f Path 35: 1-21.

236. Sales de Gauzy J, Kerdiles N, Baunin C, Kany J, Darodes P, Cahuzac JP (1997) Imaging evaluation of subluxation in Legg-Calve-Perthes disease: Magnetic resonance imaging compared with the plain radiograph. J Pediatr Orthop 6B: 235-238. 237. Salter RB, Thompson GH (1984) Legg-Calve-Perthes disease: The prognostic significance of the subchondral fracture and a two-group classification of the femoral head involvement. J Bone Joint Surg 66A:479-489. 238. Salter RB (1973) Legg-Perthes disease. Part V. Treatment by innominate osteotomy. In: Instructional Course Lectures, American Academy of Orthopedic Surgeons, volume 22. pp. 309-316, St. Louis: CV Mosby. 239. Salter RB (1984) The present status of surgical treatment for Legg-Perthes disease. J Bone Joint Surg 66A:961-966. 240. Salter RB (1980) Legg-Perthes disease: The scientific basis for the methods of treatment and their indications. Clin Orthop Rel Res 150:8-11. 241. Sanchis M, Zahir A, Freeman MAR (1973) The experimental simulation of Perthes disease by consecutive interruptions of the blood supply to the capital femoral epiphysis in the puppy. J Bone Joint Surg 55A:335-342. 242. Schlesinger I, Crider RJ (1988) Gage's signmRevisited! J Pediatr Orthop 8:201-202. 243. Schoenecker PL (1986) Legg-Calve-Perthes disease. Orthop Rev 15:561-574. 244. Schwarz E (1914) Eine typische erkrankung der oberen femurepiphyse. Beitr Klin Chir 93:1-61. 245. Scoles PV, Yoon YS, Makley JT, Kalamchi A (1984) Nuclear magnetic resonance imaging in Legg-Calve-Perthes disease. J Bone Joint Surg 66A:1357-1363. 246. Shapiro F (1982) Legg-Perthes disease: A study of lower extremity length discrepancies and skeletal maturation. Acta Orthop Scand 53:437-444. 247. Shigeno Y, Evans GA (1996) Revised arthrographic index of deformity for Perthes' disease. J Pediatr Orthop 5:44-47. 248. Shigeno Y, Evans GA (1996) Quantitative correlation between the initial and final femoral head deformity in Perthes' disease. J Pediatr Orthop 5:48-54. 249. Siffert RS (1972) Osteochondrosis of the proximal femoral epiphysis. In: Instructional Course Lectures, American Academy of Orthopaedic Surgeons, volume 22. pp. 270-285, St. Louis: CV Mosby. 250. Siffert RS, Feldman DJ (1980) The growing hip: The dynamic development of the normal adult hip and the deformed hip of Legg-Calve-Perthes' disease. Acta Orthop Belg 46:443-476. 251. Sjovall H (1942) Zur frage der behandlung der coxa plana. Acta Orthop Scand 13:324-353. 252. Smith SR, Ions GK, Gregg PJ (1982) The radiological features of the metaphysis in Perthes disease. J Pediatr Orthop 2:401-404. 253. Snyder CR (1975) Legg-Perthes disease in the young hip-Does it necessarily do well? J Bone Joint Surg 57A:751-759. 254. Soeur R, DeRacker CH (1952) L'aspect anatomopathologique de l'osteochondrite et les theories pathogeniques qui s'y rapportent. Acta Orthop Belg 18:57-102. 255. Somerville EW (1971) Perthes' disease of the hip. J Bone Joint Surg 53B:639-649. 256. Sourdat P (1910) La coxalgie en radiographie. Arch Prov Chir 19:330-350.

References 257. Sponseller PD, Desai SS, Millis MB (1988) Comparison of femoral and innominate osteotomies for the treatment of Legg-Calve-Perthes disease. J Bone Joint Surg 70A: 1131-1139. 258. Stevens PM, Williams P, Menelaus M (1981) Innominate osteotomy for Perthes' disease. J Pediatr Orthop 1:47-54. 259. Stulberg SD, Cooperman DR, Wallensten R (1981) The natural history of Legg-Calve-Perthes disease. J Bone Joint Surg 63A: 1095-1108. 260. Sundt H (1920) Malum Coxae Calve-Legg-Perthes. Zentralblatt f Chir 22:537-538. 261. Sundt H (1949) Further investigations respecting malum coxae Calve-Legg-Perthes with special regard to the prognosis and treatment. Acta Chir Scand Supp 148:1-101. 262. Suramo I, Puranen J, Heikkinen E, Vuorinen P (1974) Disturbed patterns of venous drainage of the femoral neck in Perthes' disease. J Bone Joint Surg 56B:448-453. 263. Sutherland AD, Savage JP, Paterson DC, Foster BK (1980) The nuclide bone-scan in the diagnosis and management of Perthes' disease. J Bone Joint Surg 62B:300-306. 264. Thompson GH, Westin GW (1979) Legg-Calve-Perthes disease: Results of discontinuing treatment in the early reossification phase. Clin Orthop Rel Res 139:70-80. 265. Trias A (1978) Femoral osteotomy in Perthes disease. Clin Orthop Rel Res 137:195-207. 266. Trueta J, Morgan JD (1960) The vascular contribution to osteogenesis I. Studies by the injection method. J Bone Joint Surg 42B:97-109. 267. Trueta J, Amato VP (1960) The vascular contribution to osteogenesis. III. Changes in the growth cartilage caused by experimentally induced ischaemia. J Bone Joint Surg 42B: 571-587. 268. Tsao AK, Dias LS, Conway JJ, Straka P (1997) The prognostic value and significance of serial bone scintigraphy in Legg-Calve-Perthes disease. J Pediatr Orthop 17:230-239. 269. Uno A, Hattori T, Noritake K, Suda H (1995) Legg-CalvePerthes disease in the evolutionary period: Comparison of magnetic resonance imaging with bone scintigraphy. J Pediatr Orthop 15:362-367. 270. Urlus M, Stoffelen D, Fabry G (1992) Hinge abduction in avascular necrosis of the hip: Diagnosis and treatment. J Pediatr Orthop 1B:67-71. 271. Uyttendaele D, De Kelver L, Croene P, Fabry G (1980) Conservative treatment in Perthes' disease: A comparison between containment and noncontainment methods of treatment. Acta Orthop Belg 46:414-421. 272. VanDam BE, Crider RJ, Noyes JD, Larsen LJ (1981) Determination of the Catterall classification in Legg-CalvePerthes disease. J Bone Joint Surg 63A:906-914.

375

273. Van der Heyden AM, Tongerloo RB (1981) Shelf operation in Perthes' disease. J Bone Joint Surg 63B:282. 274. Vila Verde VMR, Gomes Peres JFS, Costa B AA (1985) Value of the head at risk concept in assessing the prognosis in LeggCalve-Perthes disease. J Pediatr Orthop 5:422-427. 275. Waldenstrom H (1909) Der obere tuberkulose collumherd. Zeitschr f Orthop Chir 24:487-512. 276. Waldenstrom H (1920) Coxa plana, osteochondritis deformans coxae, Calve-Perthes'sche Krankheit, Legg's disease. Zentralbl f Chir 47:539-542. 277. Waldenstrom H (1922) The definite form of the coxa plana. Acta Radiol 1:384-394. 278. Waldenstrom H (1923) On coxa plana. Acta Chir Scand 55: 577-590. 279. Waldenstrom H (1934) The first stages of coxa plana. Acta Orthop Scand 5:1-34. 280. Waldenstrom H (1938) The first stages of coxa plana. J Bone Joint Surg 20:559-566. 281. Walter H (1925) Zur histologic und pathogenese der Perthesschen krankheit und der aseptischen knorpel-knochennekrosen uberhaupt. Arch f Orthop Unfall-Chir 23:672-696. 282. Weiner SD, Weiner DS, Riley PM (1991) Pitfalls in treatment of Legg-Calve-Perthes disease using proximal varus osteotomy. J Pediatr Orthop 11:20-24. 283. Wenger DR, Ward WT, Herring JA (1991) Legg-CalvePerthes disease. J Bone Joint Surg 73A:778-788. 284. Willett K, Hudson I, Catterall A (1992) Lateral shelf acetabuloplasty: An operation for older children with Perthes' disease. J Pediatr Orthop 12:563-568. 285. Willner S (1975) Difference in leg length in children with coxa plana during and after treatment using unilateral unloading. Acta Orthop Scand 46:102-107. 286. Wirth T, LeQuesne GW, Paterson DC (1993) Ultrasonography in Perthes' disease: Clinical relevance and influence on treatment. Internat Orthop 17:300-304. 287. Wynne-Davies R, Gormley J (1978) The aetiology of Perthes' disease. J Bone Joint Surg 60B:6-14 288. Yngve DA, Roberts JM (1985) Acetabular hypertrophy in Legg-Calve-Perthes disease. J Pediatr Orthop 5:416-421. 289. Yrjonen T (1992) Prognosis in Perthes' disease after noncontainment treatment: 106 hips followed for 28-47 years. Acta Orthop Scand 63:523-526. 290. Yrjonen T, Poussa M, Hoikka V, Osterman K (1992) Poor prognosis in atypical Perthes disease: Radiographic analysis in 19 hips after 35 years. Acta Orthop Scand 63:399-402. 291. Zemansky AP, Jr (1928) The pathology and pathogenesis of Legg-Calve-Perthes' disease (Osteochondritis juvenilis deformans coxae). Am J Surg 4:169-184.

CHAPTER 5

Coxa Vara in Developmental and Acquired Abnormalities of the Femur Slipped Capital Femoral Epiphysis, Proximal Femoral Focal Deficiency, Infantile Coxa Vara, Coxa Vara with Conyenital Short Femur, Coxa Vara with Skeletal Dysplasias I. II.

Coxa Vara: General Overview

III.

Developmental Abnormalities of the Femur

Slipped Capital Femoral Epiphysis

IV.

Infantile Coxa Vara

I. C O X A V A R A : G E N E R A L O V E R V I E W

defined this way, the term coxa vara includes the broad range of developmental and acquired disorders that predispose one to that shape of the proximal femur.

A. Terminology Coxa vara is not a disease; it is a descriptive term that refers to the deformed position of the head and neck of the femur in relation to the long axis of the shaft. In a strict structural sense, coxa vara refers to deformity of the proximal femur in which the head-neck axis on an anteroposterior radiograph forms an angle with the long axis of the femoral shaft that is 110 ~ or less (Fig. 1). In a more dynamic sense, coxa vara exists when any developing deformity of the proximal femur repositions the femoral head-neck axis into a more varus conformation than previously existed in a relatively rapid and pathological fashion not compatible with the normal developmental diminution of the head-neck-shaft angle. In the childhood years that angle passes from a value as high as 145 ~ to the normal adult range of 120-130 ~. When

B. Causes of Coxa Vara 1. H I S T O R I C A L R E C O G N I T I O N OF T H E D E F O R M I T Y

There are many causes of coxa vara and the terminology used to describe the various types has been imprecise, due in part to a lack of certainty concerning causation in many cases. At the end of the ninteenth century and the beginning of the twentieth, clinical, radiographic, and pathoanatomic studies both defined the coxa vara deformities of childhood and adolescence and began to differentiate them into specific entities (58, 61, 106, 115). The most common variant of coxa vara today is slipped capital femoral epiphysis, which was referred to commonly in earlier times as epiphyseal coxa vara or adolescent coxa vara. Careful consideration of the deformity leads to differentiation of coxa vara into four broad categories: acquired coxa vara, congenital coxa vara with a spectrum of femoral abnormalities, infantile coxa vara, and coxa vara associated with generalized skeletal dysplasias. The initial clinical description of the varus nontraumatic hip deformity is credited to Fiorani, who in 1881 reported 15 patients in which limping was considered to be due to bending of the femoral neck caused by rachitic softening (75). Both Elmslie (66) and Key (144) pointed out that his descriptions of the cause almost certainly were wrong, but the pathologic varus deformity of the proximal femur formally was introduced. In 1888, Mueller reported on a bending of the neck of the femur in what now appears to have been an adolescent coxa vara of slipped capital femoral epiphysis (194, 195). Kocher (153) and Hofmeister (115) were the first to use the term coxa vara in 1894 in separate articles. Hofmeister reported on 45 cases, showing that some had their origin in early childhood (115). The earliest report of congenital coxa vara was by Kredel in

I F I G U R E 1 Illustration of normal proximal femur (left) and coxa vara (right). The abnormal side has a markedly decreased head-neck-shaft angle of only 75 ~ One definition of coxa vara defines it as present when the angle is less than 110~

376

SECTION I ~ Coxa Vara: General Overview 1896 (157). In 1905 Hoffa described two cases of coxa vara, now recognizable as infantile coxa vara, including clinical, radiographic, and histologic assessments to distinguish the disorder as a primary one and not secondary to trauma or tickets (114). Large numbers of papers were presented in the first decade of the twentieth century as radiographs began to reveal the coxa vara deformity much more frequently. The primary site of deformity was recognized to be variable, occurring in either the head, physis, neck, or trochanteric region. The paper by Elmslie, "Injury and Deformity of the Epiphysis of the Head of the Femur: Coxa Vara," remains of great value more than 90 years after presentation (66). He indicated that coxa vara "has not received the careful attention to its pathological anatomy which is essential if the cause of the condition is to be satisfactorily explained," and "more than one anatomical class of deformity is included under coxa vara and that in a majority of cases the origin of the deformity is connected with the mode of growth of the neck of the femur." In regard to congenital coxa vara, Elmslie noted that it was very difficult to prove the congenital origin of this disorder, although he did note that "occasional congenital imperfections of the upper epiphysis of the femur do occur and a few cases have been recorded." These have, without exception, shown other associated defects of the limb such as " . . . talipes equinovarus.., or defective growth of the femur." Elmslie listed a classification of coxa vara to include (1) congenital, (2) infantile, (3) adolescent (a) acute traumatic, (b) traumatic but following the injury after an interval, and (c) without trauma, (4) secondary to diseases causing softened bone, including rickets, osteitis deformans, and osteomalacia, (5) following acute or chronic inflammatory processes, tuberculous disease, and epiphysitis, (6) osteoarthritis, and (7) traumatic nonunion of fractures of the femoral neck. Another of the older but still informative classifications of coxa vara is that of Key (144). This represented primarily an anatomic location classification with four groupings. In group A, capital coxa vara, the underlying disorders were Legg-Perthes disease, arthritis, and destructive diseases of the femoral head such as tuberculosis and sepsis. Group B included epiphyseal coxa vara with an idiopathic type, currently called slipped capital femoral epiphysis, and a traumatic type, referring to epiphyseal growth plate fracture separations. In group C, cervical coxa vara, the causes were congenital deformity including unreduced congenital dislocation of the hip, femoral neck fractures with malunion, and many other disorders such as rickets, osteogenesis imperfecta, tuberculosis, infection and neoplastic disorders, and some skeletal dysplasias. In group D, trochanteric coxa vara, the causes were those disorders listed for group C as well as intertrochanteric fracture with malunion. A brief overview of the four broad categories of coxa vara consistent with current concepts follows. Individual papers through the decades have employed terms in variable fash-

377

ion, and it is important to focus on the cases being described rather than the terminology used in any particular article. 2. ACQUIRED COXA VARA Acquired forms of coxa vara occur as a sequel to: (1) avascular necrosis of the femoral head following treatment of developmental dysplasia of the hip, (2) Legg-Perthes disease, (3) septic or tuberculous arthritis of the hip, (4) fractureseparation of the proximal capital femoral epiphysis with avascular necrosis, (5) slipped capital femoral epiphysis, and (6) bone diseases that predispose one to either physeal or structural bone weakness, such as nutritional rickets, vitamin D resistant rickets, renal osteodystrophy, fibrous dysplasia, osteopetrosis, and osteogenesis imperfecta in which the coxa vara is characterized by a lateral bowing of the head, neck, and proximal metaphysis such that the tip of the greater trochanter comes to lie close to or even above the superior edge of the head. The triangular neck fragment is not seen in any of these acquired variants of coxa vara. 3. CONGENITAL COXA VARA In congenital coxa vara there are clear developmental abnormalities of the proximal end of the femur, whose onset dates from the embryonic period. This variant of coxa vara is associated with a spectrum of deformity patterns, including those described as proximal femoral focal deficiency. Moderate involvement encompasses coxa vara with a congenital short and bowed femur. 4. INFANTILE COXA VARA Infantile coxa vara is a condition in which deformity is isolated to the proximal femur without abnormalities of the middle or distal femur or the rest of the skeleton. This term was used originally by Elmslie in 1907 and has been adopted, albeit slowly, by many others especially over the past 25 years because the disorder appears to develop after birth rather than being truly congenital. Other terms for this disorder are developmental or cervical coxa vara. This entity was described often as congenital coxa vara in the first half of the century and still is described this way by some. It may be unilateral or bilateral. It is characterized by a pathognomonic bony discontinuity of the inner surface of the femoral neck appearing as a triangular-shaped bone fragment bordered by the physis medially and superiorly and a vertical radiolucent fissure laterally with its base along the inferior surface of the neck. The disorder is not congenital in that many patients have been described in whom the hip radiographs from the first year of life were normal but who subsequently developed the coxa vara deformity shortly afterward. 5. COXA VARA ASSOCIATED WITH GENERALIZED SKELETAL DYSPLASIAS Coxa vara can occur in association with many of the generalized skeletal dysplasias. It is particularly common with

378

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

spondyloepiphyseal dysplasia congenita and cleidocranial dysostosis and also has been described in variants of multiple epiphyseal dysplasia and metaphyseal dysostosis. In the skeletal dysplasias it is common for the acetabulum to be abnormal as well. Some cases of coxa vara in the skeletal dysplasia group have the isolated neck fragment and some do not. The variable primary sites of deformity are illustrated in Figs. 2 A - 2 E

C. Clinical Presentation of Coxa Vara Due to the wide variety of causes of coxa vara, the clinical presentation is not uniform. Some clinical features of any coxa vara deformity are universal regardless of etiology; foremost among these are a waddling gait, a positive Trendelenburg sign, and limitation of hip abduction and usually internal rotation. Unilateral cases of coxa vara are characterized by a lower extremity length discrepancy that is short on the involved side.

D. Imaging Assessments in Coxa Vara 1. PLAIN RADIOGRAPHY Plain radiography clearly illustrates the coxa vara deformity. The differences in bone and cartilage development are striking particularly in a unilateral situation in which the opposite hip serves as a normal control. In most cases of coxa vara, except those occunfng as part of a generalized skeletal dysplasia, the femoral head is round and the acetabulum is developed appropriately at least in the early stages. The greater trochanter, however, rides higher in relation to the femoral neck and head than normal, and in severe cases the tip of the greater trochanter is higher than the most proximal part of the femoral head. One index of assessment is the height or level of the superior surface of the femoral head secondary ossification center related to the superior bony tip of the greater trochanter. In many instances there is a large, nonunited, triangular, inverted V fragment at the medial proximal region of the neck adjacent to the physis, a radiologic finding that is diagnostic of coxa vara. The growth plate of the greater trochanter appears to be normal, although because of its relative prominence it is somewhat more horizontal than that on the opposite side. In patients with a skeletal dysplasia there usually is a marked delay in the appearance of the proximal femoral capital secondary ossification center, and in many instances it does not appear even as late as adolescence. Quantitation of the deformity can be imprecise. A common measurement is the angle made by the head-neck axis with the femoral shaft on an anteroposterior plain radiograph. This can be difficult to measure as in most instances of coxa vara the neck is short and determination of the central long axis through the head and neck is somewhat subjective. The proximal femoral capital growth plate is

more vertical than normal and measurement of its obliquity is valuable. It may be wider or narrower than normal as well. The shape of the cartilaginous femoral head can be determined by other modalities, the two most common of which are arthrography and MR imaging. 2. ARTHROGRAPHY An arthrogram can be extremely useful in outlining both the shape of the femoral head and its relationship to the adjacent acetabulum. This study is useful particularly in the first decade of life and particularly in those with a coxa vara present in association with a skeletal dysplasia in which there frequently is a delay in appearance of the secondary ossification center as well as acetabular maldevelopment.

3. M R IMAGING MR imaging is particularly valuable in assessments of coxa vara. The images outline the shape of the femoral head as well as that of the adjacent acetabulum and are able to distinguish both cartilage and bone components. A further value of MR imaging is the ability to assess marrow signals, the structure, and, to a certain extent, the function of the growth plate of the femoral head-neck complex and in particular to contrast it with that of the greater trochanter, which is either normal or certainly much less involved in any dysplastic process. 4. CT SCANNING CT scanning provides a three-dimensional indication of the relationship of the femoral head to the acetabulum and to the femoral neck and shaft. It is of great value in assessing the degree of posterior slippage of the head relative to the neck in slipped capital femoral epiphysis. 5. ULTRASONOGRAPHY Kallio et aL (136) and Terjesen (252) have used ultrasonography to assess femoral head and neck position and shape and to diagnose intra-articular effusion in slipped capital femoral epiphysis.

H. S L I P P E D C A P I T A L FEMORAL EPIPHYSIS A. Terminology Slipped capital femoral epiphysis is a hip disorder of adolescence characterized by a change in position of the femoral head in relation to the neck at the level of the physeal cartilage. The head is positioned more posteriorly and medially in relation to the femoral neck. Although the head is described as moving into the abnormal position by the term slipped capital femoral epiphysis, in reality it is the neck and

SECTION II ~ Slipped Capital Femoral Epiphysis

femoral shaft that displace primarily by externally rotating, leaving the head positioned posteriorly and medially.

B. Evolving Clinical Awareness and Description of the Disorder Recognition of the occurrence of displacement of the capital femoral epiphysis, but without clarity as to cause, evolved slowly over several centuries. Howorth has written a superb detailed review of the early history of studies on slipped capital femoral epiphysis (SCFE) (117). Par6 recognized the displacement of the proximal capital femoral epiphysis in 1572 and indicated that "the epiphysis of the head of the femur sometimes becomes disjointed and separates in such a way that the surgeon is muddled, thinking it may be luxation and not separation of the epiphysis of this bone (206)." Petit in Maladies des Os (1709) commented on "the slipping or the separation of the neck from the epiphysis which forms the head" (211). The epiphysis that formed the head was attached to the underlying bone by a cartilage present only during the growing years, such that "the slipping can occur only in young subjects where the cartilage which joins the epiphysis has not yet ossified." Another French surgeon, Duverney, in his Treatise des Maladies des Os in 1751, also pointed out that slipping of the epiphysis could occur only prior to adulthood (65). Most separations of the head from the neck of the femur were true fractures and "not a slipping of the epiphysis which experts have claimed." Separation "with very little effort" could occur in inflammatory disorders, however, in which the periosteum was separated from the entire circumference of the neck of the bone. In the early and middle 1800s, observers continued to comment on displacement of the upper end of the femur in situations in which trauma alone appeared insufficient, such that a transphyseal separation denoted some form of pathology. Many of the patients described were in their mid to late teens and in many instances had trauma of varying intensity. In the pre-radiographic era, understanding of SCFE was obscured by its inclusion with the multietiologic coxa vara entity and by failure to distinguish it from acute traumatic epiphyseal fracture-separations. The condition came to be recognized widely as causing a coxa vara deformity in the late 1800s. With increasing numbers of assessments and particularly with the onset of radiography in the first decade of this century, the disorder came to be defined more clearly. The earliest and most accurate description of what we now refer to as slipped capital femoral epiphysis generally is attributed to Mueller based on his 1888 article "Deflection of the Femoral Neck in Childhood. A New Syndrome" in which he reported on four patients (194, 195). Among the characteristic clinical features of SCFE that Mueller described are the following: the disorder occurs in young individuals in the mid-teens without apparent special cause or as the result of any preceding painful trauma. The presentation involves

379

weariness and a gradual shortening of the affected limb. The symptoms develop gradually and do not confine the patient to bed, and patients in general are otherwise quite healthy. The shortening is apparent, although the distance from the trochanter to the lateral malleolus is equal with the trochanter 2-3 cm higher on the affected side. The limb is externally rotated. Hip flexion rarely is diminished with rotation and abduction being somewhat more restricted. The mobility of the hip is excellent such that the expectation is that the head remains positioned in the acetabulum and that no serious changes are to be found within the joint itself. The shortening is due to the fact that the diaphysis had shifted on the epiphysis, so that the angle between the shaft and the neck of the femur is lessened, thus leading to the coxa vara terminology. Whitman also described the slipped capital femoral epiphysis entity in 1894 (268). He was aware of and referred to the work of Mueller, titling his article "Observations on Bending of the Neck of the Femur in Adolescents." He described four patients, 11, 15, 16, and 17 years of age. The symptoms in each instance indicated a gradual development over several months to years; the individuals presented with a limp and considerable shortening of the involved extremity. The main limitation of movement was that of abduction, but smoothness of the general hip motion indicated that the femoral head had remained well-positioned in the acetabulum. The symptoms worsened with time with the shortening becoming more marked, the greater trochanter becoming more prominent, and an adduction deformity of the hip occurring with compensatory lateral curvature of the spine. Often there were complaints of fatigue and discomfort referred to the trochanteric region and the front of the thigh, and a waddling or rolling gait often was described. Whitman commented on the marked limitation of abduction to onethird of normal frequently and in some cases restriction of internal rotation of the hip. He pointed out that Roser in 1843 had described a hip specimen from a subject who died at 24 years of age and in whom a marked deformity of the hip had been present. Inspection showed the joint surfaces to be normal and the head of the bone to remain in the acetabulum, with the neck being greatly displaced downward and forward. By the end of the nineteenth century there was widespread recognition of the possibility of displacement of the proximal femoral epiphysis and the fact that it often occurred with minimal injury (103). One of the earliest surgical treatments was a valgus osteotomy in the subtrochanteric region, performed initially by Keetley (139, 140). With increased clinical awareness of coxa vara and wider use of radiographs, the relatively frequent occurrence of displacement of the proximal femoral capital epiphysis was recognized. Because some form of injury frequently accompanied the displacement, the traumatic separation of the epiphysis of the head of the femur as an entity became recognized. Those disorders in which slipping was accompanied by little or no injury or discomfort were

380

CHAPTER 5 9

Coxa Vara in Developmental and Acquired Abnormalities of the Femur

A

Various Sites of Deformation in Coxa Vara

Septic Arthritis (hip)

Slipped Femoral Epiphysis

SECTION I! ~ Slipped Capital Femoral Epiphysis

referred to as adolescent coxa vara, a precursor term for slipped capital femoral epiphysis. Elmslie, in his classic article on coxa vara, clearly recognized that in the traumatic category "the violence has been slight" in many cases and that "in several there has been more than one accident" (66). He continued that "it is a very remarkable fact that this severe injury in a majority of cases gives rise to comparatively trivial symptoms. Many of the patients are able to get up and walk away after the mishap and some are even unaware that they have injured the hip at all until several days or even weeks later. With time and with continuation of walking the adduction deformity became more apparent." When seen some months after the accident, in every case there is the deformity of the ordinary variety of coxa vara. Elmslie summarized that traumatic separation of this epiphysis (which we now recognize in most instances to be slipped capital femoral epiphysis) (1) was much more common than formerly believed, especially in adolescence; (2) was produced often by very slight violence and sometimes by several separate accidents; (3) had immediate symptoms that often were of no great severity; and (4) had resulting deformities clinically or pathologically indistinguishable from ordinary adolescent coxa vara. Kocher used the term coxa vara to describe a triple deformity of the neck of the femur (153), which was bowed upward and forward and was twisted such that it was hyperextended in relation to the head. The limb, therefore was held in adduction, external rotation, and hyperextension. Hofmeister used the term to refer to the diminished angle of the neck with the shaft with the more distal extremity maintained in adduction (115). Key used the term epiphyseal coxa vara to denote slippage of the epiphysis, encompassing the femoral head, on the neck (144). He reported an early series in 1926; in 24 patients, the male:female ratio was 2.4:1 and he recognized that involvement was earlier in the female, with the average age of onset in girls at 11.3 years and that in boys at 14.3 years. In 250 cases from the earliest literature on slipped epiphysis the male:female ratio was 3.3:1 with the average age of recognition being 13.5 years in girls and 15.4 years in boys. The condition often was bilateral. The clinical description involving both history and physical examination provided by Key in 1926 scarcely has been improved upon today. The

381

patients present with hip or thigh discomfort and a limp. The initial symptoms frequently are quite mild and often of several months duration. Trauma is not a specific etiologic factor, although a relatively mild episode of trauma often brings the condition to clinical awareness. The characteristic physical finding involves a tendency of the involved lower extremity to move into external rotation with flexion of the hip. Gait also is characterized by external rotation or out-toeing on the involved side. The clinical exam reveals a diminution of internal hip rotation and in some instances an increase in external hip rotation. There is no tendency toward a fixed flexion contracture, and if the patient is sufficiently comfortable there actually is an increase in extension of the hip on examination in the prone position. Some shortening can be detected, and this is proportionally greater with increased amounts of slippage. There may be some thigh atrophy if the symptoms have been present for longer than 2 or 3 months. As in most hip disorders of childhood, the discomfort may be centered either primarily or exclusively about the medial distal thigh with no complaint of hip area discomfort.

C. Etiology of Slipped Capital Femoral Epiphysis 1. INITIAL T H E O R I E S : T R A U M A T I C , B O N E DISEASE, AND STATIC

The three common etiologic theories early in the twentieth century involved those of a traumatic, bone disease, or static origin. The trauma theory came to be discounted relatively early because there often was little history of trauma, although initially it was proposed by many. In addition, the slipped capital femoral epiphysis or epiphyseal coxa vara frequently was bilateral, being as high as one-third of cases in one series and at least 15% in Key's summation of 250 cases (144). Many early writers considered the disorder to be secondary to tickets, but increasing examination of patients and pathoanatomy studies led most to believe that slipped capital femoral epiphysis or epiphyseal coxa vara was a separate entity. Frangenheim felt that rickets was not the cause in any of his six specimens and that in all specimens there was some abnormality in the epiphyseal cartilage (81). He classified the disorder as a local growth disturbance or chondrodystrophy and reported good awareness with the

F I G U R E 2 (A) Variable primary sites of coxa vara deformity are illustrated. There is value in interpreting a coxa vara deformity based on the primary site of pathology. The deformity can occur due to a primary pathological process in the head, physis, neck, or upper part of the shaft. In some disorders two or more contiguous sites are affected. (B) In the acquired forms of avascular necrosis of the secondary ossification center of the head, the effects lead to diminished head height with both collapse and decreased growth and sometimes shortening of the neck due to diminished physeal vascularity with continued normal growth of the greater trochanter. Common examples are (Bi) AVN secondary to treatment of DDH and (Bii)) with Legg-Perthes. (C) With slipped capital femoral epiphysis, it is the slippage of the normal head mediated through the physis that leads to the deformity. (Di, Dii) In skeletal dysplasias and infantile coxa vara the physis and adjacent neck are the site of abnormality. (E) In tickets the widened physis and "softened" bone of the neck and upper shaft predispose to bowing. (F) In disorders such as osteogenesis imperfecta it is primarily the structural weakness of the bone of the neck and upper shaft that leads to the bowing, with the head and neck regions becoming depressed relative to the tip of the greater trochanter

382

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

overweight phenomenon. Hofmeister first proposed the static theory for the adolescent variant, which referred to the mechanical effects of increased patient weight in relation to a hip or proximal femur that otherwise was normal (115). Kirmisson reported on the frequent association of epiphyseal coxa vara with obesity (148). The static theory soon had many adherents, emphasizing the bending of the neck of the femur by increased weight beating requirements. It gradually gained favor as it was felt that the epiphyseal line was the point of least resistance in slippage in the adolescent hip. Currently, it is felt that obesity, certain anatomic and physiological characteristics of the proximal femur, and a series of specific medical disorders strongly predispose one to development of the deformity.

2. OBESITY The only uniform etiologic finding in all studies over several decades is the fact that the large majority of patients are obese. The obesity may be absolute with patient weight "off the charts," by which is meant greater than the 97th percentile, or it may be relative with the weight within the normal range but at a markedly higher percentile than height. 3. ANATOMIC-PHYSIOLOGICAL FEATURES Four prominent anatomic-physiological features appear to serve as predisposing causes. a. Diminution o f Periosteal Thickness as Skeletal Maturity Is Reached Toward the end of skeletal growth, the periosteal thickness lessens relative to the size of the underlying bone and cartilage and its supportive periphyseal role is diminished. Key was a prominent advocate of this predisposing cause (144). b. Slippage Related to Bone Age (Growth Spurt) Exner has shown that the time of occurrence of the slipping is related most closely to bone age and to the peak of the growth spurt (68). A narrow bone age range at the time of slippage indicates that skeletal maturation is a key contributing feature. The chronological ages in his series of 23 (8 girls, 15 boys) were a mean of 12.9 years (range = 10.716.3 years) in the girls and 14.5 years (range = 12.2-17.4 years) in the boys, but the skeletal ages were 13.1 years (range = 12-13.9 years) in girls and 14.9 years (range = 14-15.6 years) in boys. Loder et al. also pointed out the narrow range of bone age within which SCFE occurred (170). c. Retroversion o f the H e a d - N e c k Area Patients with SCFE have been shown to have relative retroversion of the head-neck axis, which appears to predispose one to slippage. Gelberman et al. documented a decreased angle of femoral anteversion in slipped capital femoral epiphysis as shown by computerized axial tomography (CT) measurements (85). The mechanical forces acting across the proximal femoral physis would be altered by this rotational abnormality, leading to an increased shear stress that predisposed to failure of the physis. Twenty-five patients were assessed with the degree of anteversion determined in 39 involved

hips by computerized axial tomography. The mean amount of anteversion for all of the involved hips was 1.0 +_ 0.2 ~ In 18 hips seen at the time of original diagnosis, the amount of anteversion was - 0 . 7 _ 7.4 ~ whereas in 21 hips seen 7 years after operative treatment, the mean anteversion measurement was 2.5 _+ 8.7 ~ The amount of anteversion for hips in both groups was markedly less than the predicted mean for individuals of the same age. The mean amount of anteversion of the unaffected hips of patients with a unilateral slip was +6.3 _ 8.2 ~ At the time of this study, there were no standards of anteversion by CT, but many other studies had been done using biplanar radiography. Fabry et al. had documented the amount of anteversion at 20 _ 7 ~ for normal children between the ages of 10 and 13 years and 15 + 8~ for those between the ages of 14 and 16 years (69). Values for anteversion from adults by CT study were available from Reikeras et al., who in a study of 94 normal hips found the mean anteversion to be 13 _ 7 ~ (223). It was evident that, in patients with SCFE, the values for anteversion were substantially lower than normal, indicating a relative retroversion. This feature of relative retroversion appeared to be a causative effect in the sense that the magnitude of the retroversion was just as great in those seen early in the course of the disease as it was in those seen later. In addition, the magnitude of the retroversion was as great in patients with a mild slip in which relatively little remodeling would have occurred as it was in those with a moderate or severe slip. Independent confirmation of the finding of a low extent of femoral anteversion in association with SCFE was presented by Jacquemier et al. in a review of 25 cases (126). Their study showed that the risk of slipped upper femoral epiphysis was distinctly greater even for the opposite hip if femoral anteversion had a low value. Their assessment of femoral anteversion in SCFE was by CT studies, with comparative CT studies from a second group of 127 children seen for other pediatric orthopedic disorders. The mean value for femoral anteversion in those patients with SCFE was 9.8 ~ with a range from - 2 3 ~ to + 45 ~ whereas the value was the same from the contralateral side, which had not slipped, ranging from 0 to 28 ~ In the 127 children studied for other disorders without slip, with ages ranging from 5 to 16 years, the mean femoral anteversion was 31 ~ whereas in those from 8 to 16 years of age, the mean value for femoral anteversion was 25 ~ The mean value for femoral anteversion in SCFE of 9.8 ~ was significantly less than that of the non-SCFE group at 25 ~ The statistical study of Jacquemeir et al. showed that the risk of developing slipped upper femoral epiphysis is 15 times higher for a femur with a low femoral anteversion than for one with normal femoral anteversion. The retroversion aligned the hip in a functional retroverted position during weight beating and thus increased the sheafing stresses of the junction of the head and neck. Braenkel et al. in 15 cases found a femoral anteversion of 3.06 ~ and 8.5 ~ in SCFE, which also indicates considerable

SECTION I! ~ Slipped Capital Femoral Epiphysis

relative retroversion (35). In normal studies, conventional radiographic assessments of anteversion by Teinturier and Deschambre were 15-25 ~ at 12 years of age (251). Pritchard and Perdue performed three-dimensional force analyses on the hips of 50 normal patients and 50 with SCFE to assess the effects of mechanical factors (218). The SCFE patients had reduced resistance to shear because of increased body weight and a decreased neck shaft-plate shaft angle. In addition, slipped epiphysis patients with relative retroversion were found, by modeling, to generate increased sagittal plane shear stress of the proximal-femoral growth plate. During running, these mechanical factors were increased such that shear failure of the growth plate could be documented mechanically in obese patients. They calculated that, when forces were applied to a hip retroverted 10~ more than normal, the growth plate would experience 20% more shear stress. Subtle anatomical variations in SCFE patients reduced resistance of the growth plates to shear forces particularly during running. d. Increased Obliquity o f Proximal Femoral Growth Plate An additional mechanical feature predisposing one to the likelihood of a slip is the increased obliquity of the proximal femoral growth plate with growth. Mirkopulous et al. measured the physeal angle or slope of the proximal femoral capital epiphysis on standard anteroposterior radiographs in 307 hips in children aged 1-18 years to establish normal values (188). Similar measurements then were made on the affected slip and nonaffected nonslip sides of 107 patients with unilateral SCFE. In the normal patients, there was an average increase of 14~ in the slope of the proximal femoral physis between the ages of 1 and 18 years. The maximal increase occurred between 9 and 12 years. Within the unilateral slip group, the age groups 10-12 years and 13-15 years specifically were of interest. The affected and unaffected sides within these groups were compared with those of age-matched controls and each other. In those children aged 10-12 years with unilateral SCFE, there was an average 11.2~ increase in slope on the slipped side compared with age-matched controls. On the nonslip side, there was an average 4.8 ~ increase in slope compared with age-matched controls. Comparison of the slip side to the nonslip side in unilateral patients showed an average 6.5 ~ in the slope on the slipped side. Similar findings were found in the age group 13-15 years, with the slipped side showing an average 8.0 ~ higher slope than the age-matched controls. The nonslipped side showed an average 4.0 ~ higher slope than age-matched controls. Within the patients with SCFE, the slipped side revealed an average 5.2 ~ higher slope than the nonslipped side. Increased obliquity of the physis was yet another structural feature associated with SCFE, with increasing obliquity predisposing one to instability. Mirkopulous et al. documented a progression in the slope of the proximal femoral physis from birth through adolescence and a significant increase in the measured slope of the growth plate in children with unilateral SCFE on both the affected and unaffected

383

sides when compared with age-matched controls. Speer also determined that there was a 13.7 ~ progression in the slope of the proximal femoral growth plate between the ages of 1 and 18 years, that the maximal increase occurred between the ages 9 and 11 years, and that there were 15 and 5~ higher slopes on the affected and unaffected sides, respectively, in patients with unilateral SCFE compared with controls (188, 243). Speer reviewed the vast number of suggested pathogenetic mechanisms for slipped epiphysis, incorporating studies of clinical, biological, biomechanical, experimental, and theoretical factors suggested over several decades (243). He then outlined four basic categories contributing to the tendency for slipping, which included factors leading to (1) increased physeal height, (2) altered geometry of the physis and adjacent bone in relation to the linear or planar orientation of the physis and the physeal inclination angles, (3) abnormal excessive loading of the physis, which would alter the shear and compression stresses, and (4) deficient physeal and periphyseal components, which also served to alter the tensile and hydrostatic forces. Alexander suggested that increased shearing stresses acted on the femoral head and neck when susceptible patients were in the sitting position (6). Rennie theorized that because physeal tilt preceded slipping the obliquity was due to diminished growth from compression of the back (posterior region) of the femoral neck (224). The relative increase of the length of the front of the femoral neck was considered to reflect primary failure of growth of the posterior neck. Most still feel that the slippage itself is primary with altered orientation of the head-neck structure (retroversion, physeal obliquity) predisposing to its occurrence. 4. MEDICAL DISORDERS PREDISPOSING TO SLIPPED CAPITAL FEMORAL EPIPHYSIS There is a distinct subgroup of patients who have primary nonskeletal disorders eventually predisposing them to slipped capital femoral epiphysis. They account for approximately 5% of cases of SCFE in series from large pediatric institutions. The underlying conditions involve (1) endocrine disorders such as hypothyroidism, craniopharyngioma with its associated panhypopituitarism, hypopituitarism due to other causes, growth hormone therapy, and primary hyperparathyroidism (due to parathyroid adenoma); (2) radiotherapy done previously and involving the pelvic region for the treatment of Wilm's tumor, rhabdomyosarcoma, neuroblastoma, or other soft tissue sarcomas of the abdominal or pelvic regions; and (3) renal osteodystrophy. These will be reviewed in detail in Section II.F. D. P a t h o a n a t o m y 1. OBSERVATIONS AND DESCRIPTIONS Detailed pathoanatomic studies have helped to solidify understanding of the entity. In this section we review

384

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

descriptions that seem consistent with a slipped capital femoral epiphysis. Mueller described a specimen from a severe chronic slip in which the upper border of the neck was about twice the normal length, whereas the lower border was shorter and sharply curved (194, 195). The articular cartilage on the femoral head was normal. The epiphyseal line was visible as a narrow strip, and all cartilage present appeared to be normal. The femoral neck gave the impression of being molded by downward pressure from above. Due to the change in position of the head and neck in a slipped epiphysis, the superior surface of the neck follows a horizontal, slightly convex curve, whereas in the normal the superior neck outlines a concavity to the cartilage border. On the superior surface, the distance from the tip of the greater trochanter to the line of the cartilage articular surface in the slipped hip was 7 cm, whereas that of the normal was only 3.5 cm. Along the inferior border of the neck from the lesser trochanter to the cartilage border of the head the deformity was opposite. The distance in the pathologic specimen amounted to 1.2 cm compared to nearly 4 cm in the normal. A clear shift in the trabecular pattern of the epiphyseal and diaphyseal bone was noted in attempts to conform to the deformed anatomy. Where the epiphysis joined the diaphysis there was a much sharper convex curve on the superolateral aspect than normal, giving the epiphysis a distinctly crescent shape, which lay like a cap over the diaphysis such that the diaphysis simulated an impacted fracture driven into the epiphysis. The femoral head thus was placed in an extreme position of abduction. Because the patients continued to walk, the inner architecture of the bones changed corresponding to this altered form. The thickness of the compact or cortical bone was greatest along the inferior border of the specimen. On the surface the compact substance reached to the inferior end of the epiphyseal line: "this thickening of the arch is as an expression of the changing form of the bone and essential part of the greater load on the inner circumference of the bone." Mueller indicated that "from this compacta (cortex) a system of bone trabeculae radiates upward and inward separately thereby differentiating it from the normal in that the individual trabeculae are of considerable thickness and remain distinct in the large marrow spaces." It follows, therefore, that because of the considerable enlargement of the proximal circumference of the femoral neck, the trabeculae must be distributed over a larger surface. A further difference from the normal was that "the most strongly developed trabeculae lie in the neck and the head while in the normal the densest network of trabeculae are present from the inferior medial part of the femoral neck towards the superior periphery of the joint surface of the femoral head." In the slipped epiphysis femur, the strongest and thickest set of trabeculae was found at the upper and medial portions of the neck. There was no change in the consistency of the bone, and nothing abnormal was seen on the surface of bone or in the joint cartilage. Similarly, upon microscopic study, there was no evidence of specific changes with the bone and

marrow having a normal appearance. By the time of the study, the epiphyseal growth plate cartilage was absent and the trabecular bone was characterized by large amounts of osteoblasts and osteoclasts, indicating the synthesis and remodeling functions in relation to the deformed architecture. Mueller discussed the possible causes of the deformation. The disorder was "a completely isolated disease of the femoral neck; anatomically except for the change in shape, no deviation from the normal can be identified in the specimen macroscopically or microscopically." Each of many disorders was discarded as the cause, and he indicated that some atypical form of rickets was possible but that there was no direct proof of this. Koeher, describing three cases, noted from one resected specimen that the head and neck were bent obliquely downward and backward ( 153). Torsion of the neck especially was marked in one case; the head rotated upon the neck in such a manner that, with the head in its normal relationship, the femur was hyperextended. The upper border of the neck appeared lengthened. Where the head was not in contact with the acetabulum, the articular cartilage was thin. The epiphyseal line was irregular and the physeal cartilage was interrupted by ingrowing metaphyseal bone. Haedke described one specimen in which the articular cartilage was markedly thinned in places (95). The epiphyseal line was very irregular in its upper portion but well-formed below, although in the central region it was traversed by areas of young bone. The reactive bone of the posterior-inferior neck was shown well. It was bridging the gap between the neck and the posterior medial margin of the displaced head. Sehlessinger also described the marked torsion of the neck with the displacement of the head (230). He attributed the deformity to the gradual and continuous slipping of the epiphysis. The epiphyseal line was narrowed and the structure of the physeal cartilage was irregular, with small islets of cartilage and necrotic cells and nuclei seen. He illustrated the lengthened and markedly convex superior surfaces of the head and neck and the markedly shortened, concave, and thickened inferior neck surface. Frangenheim described six resection preparations in two papers with varying degrees of displacement (81, 82). The pathogenesis was unclear. The physeal cartilage generally was normal with the cells, however, being irregular in distribution. The cartilage was sometimes vascular, and endochondral ossification in various stages was present in the cartilage matrix. Marked posterior displacement was shown in excellent gross preparations cut in half, which also illustrated physeal and articular cartilage. In one specimen the uncovering of the superolateral portion of the neck was well shown along with the irregular thickness of the displaced physis and its replacement by bone in focal areas (82). In another specimen the full posterior displacement of the head left slightly more than one-half of the (lateral) neck uncovered. The fibrocartilaginous tissue of the posterior physis was thickened, and the medial edge of the head and physis abutted the inferior surface of the neck. Hofmeister de-

SECTION II ~ Slipped Capital Femoral Epiphysis

scribed several cases of coxa vara due to SCFE as well (115). He illustrated well the reactive bone formation of the posterior surface of the neck adjacent to the medial edge of the markedly displaced epiphysis. Sprengel described the pathology of SCFE in two patients based on a study of hip resections. Both patients, 17 and 18 years of age, had severe slips. The cause was considered to be consistent with a low-grade trauma rather than with any pathologic disease process. The epiphysis was structurally normal, although malpositioned, with a smooth articular cartilage surface. The periosteum was intact inferiorly, and callus was present between the fragments and was contracted at the inferior and medial concavity between epiphysis and neck. Grashey illustrated chronic posterior slippage with a clear pathoanatomic specimen (91). The physis appeared fully obliterated or fused. The posterior-medial edge of the epiphysis was displaced against the inferior-posterior neck in the intertrochanteric notch region, and the anterior convex curbed surface of the neck was markedly longer than the posterior concave shortened surface. Elmslie summarized well the first few decades of pathoanatomic study in which several resected specimens were available for gross and microscopic analysis (66). These almost exclusively came from the German literature because concern about more serious disorders at that time led to surgical resection until the pathoanatomy was clarified. He summarized the microscopic changes by reducing them to two considerations, both of which appear accurate today. (1) There was irregularity in growth of the epiphyseal cartilage with irregular ossification leading to the production of columns of cartilage cells, which were irregular in that some contained several cells in a capsule (recognized today as a chondrocyte clone), and occasionally to the inclusion of cartilage islands in the bone (now interpreted as evidence of irregularity in the process of conversion of physeal cartilage to metaphyseal bone). (2) Absorption of bone was noted in some areas with the formation of new bone in others. Elmslie interpreted both of these findings as being secondary changes adapting the bone shape mechanically to its new position. He felt that all of the alterations noted could be explained as resulting from injury due to "the altered pressure conditions to which the epiphyseal region is being subjected." The observed histologic and gross morphologic changes, therefore, were considered to be the result of the deformity rather than being the primary cause of it. The clinical, radiographic, and pathoanatomic studies appeared consistent with relatively minimal trauma, often in separate incidents, leading to the increasing deformity. Elmslie stressed the importance of appreciating the fact that relatively slight accidents could produce the coxa vara, such that it was important to examine patients carefully and assess even minimal displacements radiographicaUy in order to make an early diagnosis and minimize subsequent severe deformity. Kleinberg and B u c h m a n described the gross appearance of a severe slip of the proximal capital epiphysis as defined at arthrotomy (152). The slipping of the head occurred

385

downward and backward with the epiphyseal plate wedge shaped with its base anteriorly and superiorly. The femoral head was fixed firmly to the neck such that it could be displaced only by surgical levering. No fracture line was seen and the head and physis were continuous with the neck. Kleinberg and Buchman commented on "the impression of a gradual wandering of the head downward and backward resulting from a plastic change in the epiphyseal plate." The femoral neck bulged anteriorly and appeared longer at its anterior and superior aspects. The periosteum adjacent to this region was atrophied markedly, and the border of the epiphyseal plate was anteriorly and superiorly articulated with the acetabulum. The articular cartilage of the femoral head, however, was normal in appearance as was that of the acetabulum. The joint fluid was normal in appearance and quantity. Part of the epiphyseal plate was resected during surgical treatment. Pathologic assessment showed the line of demarcation between physeal cartilage and metaphyseal bone to be more irregular than normal. The arrangement of the proliferating cartilage cells was irregular and the number of cells per column was diminished. The resting cartilage above was fairly well preserved, but some splits were seen often with invasion of blood vessels from the metaphysis whereas others were filled with granulation tissue. In other areas, regions of fibrous bone and osteoid were surrounded by resting cartilage. This was interpreted as a reparative response. Newly formed woven bone at the metaphyseal junction was placed irregularly. The adjacent marrow in the femoral neck showed a moderate amount of fibrosis. With moderate stages of deformity the junction between the physeal cartilage and metaphyseal bone still was evident but was now highly irregular. The columnar arrangement of the proliferating cartilage was lost almost completely. Newly formed bone continued to be deposited but was positioned highly irregularly in relation to the normal trabecular pattem. Endochondral bone formation appeared to be at a minimum and fibrosis of the marrow was increased. In the severely advanced stages of the disorder, the architecture of the growth zone was extremely irregular. The growth plate in particular had very little structure and increased areas of woven bone. From the physis, irregularly shaped tongues of more or less degenerated cartilage extended into the bony zone. The arrangement of the newly formed trabeculae also was confused and irregular with little normal patterning seen. Endochondral bone formation was replaced entirely by the formation of fibrous or woven bone, which was undergoing active transformation. There was active osteoclastic resorption and osteoblastic deposition with the formation of large sheaths of lamellar bone interspersed with new areas of osteoid. There also were areas of localized bone necrosis, whereas other areas had marked vessel dilation. The marrow was completely fibrous. R a m m s t e d t described seven cases, including four resections, and defined the level of pathology at the physis where the slip occurred due to junctional weakness (221).

386

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

The importance of the periosteum in anchoring the head (epiphysis) to the neck was stressed. It was the intact periosteum that allowed walking to continue and often stability to be maintained, even in the face of physeal loosening. If the periosteum stretched or tore, the head slipped downward and backward, symptoms worsened, and external rotation deformity of the lower extremity worsened. The process was recognized as gradual. Key: Growing bone was less able to withstand strain than adult bone, and in many adolescents a disproportion was present between the patient weight and the strength of the femoral neck (144). Careful pathological studies emphasized "that the neck is not actually bent, but the epiphysis is displaced and the neck remodeled by synthesis of bone on the inferior concavity to meet the altered mechanics of the hip." Key felt that "in our series, the unaffected hips were normal except that in some instances there appeared to be a broadening of the epiphyseal line." Hip pathology had not been studied early in the disorder, and in all specimens reported, the original pathology has been obscured by secondary changes owing to the slip, changes in the blood supply, and attempts at repair. Key noted that the difficulty might not lie either in the bone or in the epiphyseal cartilage but rather in the periosteum of the femoral neck. He recognized that a strong periosteum spanned the epiphyseal line in the growing child and was a chief factor in holding the head in place. In adolescence, the periosteum begins to atrophy, producing a point of weakness at the epiphyseal line. A rapid growth spurt also is associated. Key indicated that "it is not difficult to imagine that during this period of rapid growth, the periosteum spanning the epiphyseal line is stretched, thinned, and consequently weakened thus permitting the head to be easily loosened." Balensweig described 20 cases of SCFE in 18 patients with varying degrees of epiphyseal separation of the proximal femur (18). He reported an early case of avascular necrosis of the femoral head in association with a slip of the capital femoral epiphysis. The average age at disease presentation was 13.3 years, although he felt that the actual disturbance occurred around 11 years of age, long before the patients presented. In many instances there was mild trauma and also evidence of endocrine dysfunction. Often there was shortening in these cases, which ranged from 1/8to 1 in. with the average being approximately 89 in. Many of the changes seen radiographically anticipated the observations noted later. Abnormalities were found, including widening of the epiphyseal line, haziness of the epiphyseal line, fragmentation of the epiphyseal line, a short broad femoral neck, outward rotation of the shaft at the femur, and downward displacement of the femoral head, although Balensweig recognized that "what actually takes place [is] an upward displacement of the shaft." Later changes involve premature ossification of the growth plate, occasional diminution of the joint space, irregularity of the head, and flattening of the head resembling Perthes disease. There was lessening of the

angle between the neck and shaft of the femur accompanying the coxa vara position. Even at this early date Balensweig recognized that those cases complicated by moderate or marked epiphyseal separation that were subjected to forcible manipulations resulted in varying degrees of osteoarthritis. Phemister in the United States and Axhausen in Germany had described the entity of avascular necrosis of the head of the femur a few years previously but not at that time in relationship to slipped capital femoral epiphysis. Sutro studied three cases of slipped capital femoral epiphysis of the femur, finding no histologic evidence of rickets, osteomalacia, or osteitis fibrosa (248). He interpreted the abnormality as "a fracture through the upper epiphyseal plate of the femur and through some of the contiguous osseous trabeculae." The epiphyseal plate histologically showed only scattered foci of degeneration, usually close to tears or fractures of the physis. Many of the other changes were secondary to the fracture. Sutro discussed the structure of the developing hip in detail and then referred to weight bearing forces acting on the hip, which would be maximal in relation to developmental changes of the neck between the ages of 10 and 14 years. He concluded that "variations in the weight bearing forces at the hip joint may be a predisposing factor in the causation of the fracture. The normal tilting of the capital epiphysis which is the result of normal developmental and mechanical forces is the basis for the condition." Howorth described the microscopic pathology of the proximal femur and hip joint in SCFE as determined at open operation in 17 pre- or early slips, 23 moderate slips, and 28 marked slips (116, 118). Those with moderate slips were examined in the process of undergoing operation for stabilization, whereas those with severe slips were assessed at the time of open reduction or wedge osteotomy. Howorth later increased his case material to 232 hips with slipped epiphysis seen at open operation (120). He had a wealth of material with various degrees of slippage because of his tendency to perform the open transphyseal pegging operation, thus allowing for gross examination and microscopic sections from the core biopsy. He also performed biopsies of the synovial membrane and periosteum. In the earliest phases of the disorder, the synovial membrane was swollen, edematous, and hyperemic (118). The periosteum and to a lesser extent the hip joint capsule also were slightly swollen. No gross changes were visible in the head or the acetabulum. Microscopic sections showed hypervascularity of the synovial membrane with edema and occasional infiltrates of plasma cells and small lymphocytes. He commented on "softening and decalcification at the junction of the epiphyseal plate and the neck." The zone of slipping was between the distal surface of the physis and the metaphysis of the proximal neck. With a sudden severe slip there could be complete discontinuity in this zone. Usually the periosteum remained attached to the margin of the head, but on occasion it was pulled loose superiorly and anteriorly. The displacement of the head on the neck nearly always was

SECTION !1 ~ Slipped Capital Femoral Epiphysis

downward and posteriorly and the head was tilted into varus. As the displacement tended to occur gradually, or in small amounts with repeated episodes, continuing stimulation of the intact periosteum posteriorly and inferiorly occurred, which was considered a mechanistic response to enhance stabilization of the head. The zone of the slip was visible grossly anteriorly and superiorly as a bluish zone sometimes with tiny islands of bare bone seen. No free blood was found in the joint unless there had been severe trauma and a sudden recent slip. The periosteum tended to be stripped from the neck posteriorly and inferiorly, and this angle gradually filled with callus. With successive slips, there were several zones of callus and as this callus ossified it became more dense and appeared to be part of the neck. With displacement of the head posteriorly and downward there was a projecting hump of neck proximally and anteriorly, although gradual absorption occurred and this became smooth and rounded. This was the region, however, that could impinge against the anterior and superior margins of the acetabulum with abduction. With time the proximal surface of the neck tended to reunite to the adjacent physis, but cell activity understandably often was disorderly, and areas of nonunion, fibrosis, and cartilage penetration into the metaphysis were seen grossly and microscopically. Changes in the synovial membrane and periosteum tended to persist in the early slipping phase, but in time the swelling and edema subsided and fibrosis set in. With further slipping there was additional disorganization, degenerative changes of the physis, and continuing reactive new bone formation in the adjacent periosteum. Callus formed particularly in the inferior and posterior concavity and gradually matured and participated in the union of head and neck. With time the physis was totally obliterated and stability in the displaced position was regained. Howorth noted that "necrosis does not occur unless there has been severe trauma with marked compromise of the circulation or circulatory damage from manipulation, surgical dissection with open reduction or osteotomy of the neck." In the residual stage, the lesion healed and the epiphysis was solidly united to the neck. The synovium, periosteum, and capsule were fibrotic and callus had become mature bone almost indistinguishable from the neck itself. The contours of the neck were smoother. Lacroix and Verbrugge studied an entire femoral head and neck from a 16-year-old boy with a severe slip, who had developed the disorder 3 years previously (161). Histologic sections were taken throughout the entire extent of the growth plate from the superior surface of the neck, from three intermediate quadrants of the growth plate, and from the inferior head-neck surface. In the most superior onethird of the growth plate, the epiphyseal cartilage was replaced by fibrocartilaginous tissue. The fibrous tissue in particular was vascularized. In the central one-third, there were small remnants of fibrous and cartilaginous tissue but the growth cartilage tended to be replaced by fibrillar tissue. In the lowest portion of the epiphyseal cartilage, three nod-

387

ules of endochondral ossification were enclosed between a thin layer of cartilage tissue and a thick layer of fibrous tissue. The cartilage in the physis, instead of undergoing an endochondral sequence, when present tended to be hypercellular and to evolve into fibrous tissue, which then was invaded by capillaries. The superior neck at the head-neck juncture was thinned and angulated downward. The trabeculae were being absorbed by osteoclasts when in contact with the periosteum. In the inferior concavity between the head and neck, the process was reversed with a thickened periosteum producing new trabeculae of bone. Osteoblasts were abundant. As the head slipped posteriorly, the periosteum was tightened on the superior neck and loosened on the inferior surface. Bone was absorbed on the upper surface and synthesized on the lower surface. The primary cause of the disorder was considered to lie in the epiphyseal cartilage, which, instead of producing endochondral bone, was transformed into fibrous tissue. There would seem, in our opinion, to be ample reason to interpret the physeal changes, though real, as secondary to the slip rather than causative of it, an interpretation also reported by Elmslie several decades earlier. Intrinsic abnormalities of the growth plate, however, also were postulated by Ponseti and co-authors (124, 214). Three core biopsies of the femoral head and neck were done with a 1-cm-diameter needle in an 11-year-old female, an 11.5year-old male, and a 14-year-old male (214). In the first case the bone and bone marrow of the femoral head and neck were normal and the joint cartilage was unremarkable. The epiphyseal plate was very wide, and the chondrocytes appeared normal but were grouped in clusters rather than being in orderly rows. These clusters were separated by fibrillated septae, some of which were wide. There were extensive clefts extending through the middle of the growth plate. Endochondral ossification appeared to proceed normally in some areas, whereas in others it had ceased completely. Areas of endochondral ossification appeared deep in the neck. In the second case, the findings were similar. There was a large cleft in the epiphyseal plate filled with amorphous debris. Vascularization was occurring and endochondral ossification persisted, but areas of cartilage existed within the metaphysis in which bone conversion normally would have occurred. In the third case, the plate was wide in some regions, but in others it looked structurally normal. Where the plate was wide, the cartilage cells were clumped in clusters separated by small clefts, some of which were penetrated by blood vessels. This case represented a preslip. The authors concluded from their cases that the epiphyseal plate in SCFE initially was wide and greatly disrupted with rows of proliferating cartilage cells disorganized and the cells grouped in clusters. There was fibrillation of the cartilage matrix in large clefts with necrotic debris and occasional blood vessels. Endochondral ossification was interrupted in many areas but progressing in others. Islands of cartilage cells were seen in the metaphysis. There were no signs of tickets, osteomalacia, osteoporosis, or infection. Ponseti et al. felt that the

388

CHAPTER 5 " Coxa Vara in Developmental and Acquired Abnormalities off the Femur

primary lesion was a loss of cohesion of the cartilage matrix, presumably due to an alteration in the chemical composition of the mucopolysaccharides. It appears just as reasonable to interpret the changes seen as secondary to slippage rather than causative of it. In a 1981 study, three additional core biopsies from two boys, 12 and 13 years of age, and one 10-year-old girl with SCFE were studied (124). The growth plate in general was thicker than normal, with chondrocytes arranged in large clusters rather than in columns and large accumulations of matrix separating the chondrocyte clusters. All slips were mild. Control studies from autopsy femurs were done at corresponding ages. In SCFE, the thickness of the growth plate was irregular with many regions increased in thickness from normal. The resting zone of cartilage appeared normal. The proliferating zone contained small nests of short columns of chondrocytes enveloped in a compact territorial matrix. Longitudinally oriented loose bundles of matrix fibers were strongly alcian-positive. Occasional clefts were seen along the long axis. The hypertrophic zone was thickened markedly with clefts in the increased fibrillar matrix. Capillary invasion of the lower part of the growth plate was uneven, and in many areas the hypertrophic cells appeared resistant to capillary penetration and were found in the metaphysis. The orderly progression of chondrocyte maturation in the growth plate was disrupted and the usually finely tuned process of chondrocyte proliferation, hypertrophy, and degeneration was interfered with. Chondrocyte ultrastructure appeared normal. In the proliferative zone, the two major histochemical abnormalities were increased glycoprotein staining in the matrix and increased proteoglycan staining in the extraterritorial matrix. The most striking abnormalities were found in the markedly thickened hypertrophic zone, with the chondrocyte clusters surrounded in a periodic acid-Schiff-positive, glycoprotein-rich, territorial matrix. The cartilage clusters appeared to calcify their matrix irregularly, and many large cartilage columns were found in the metaphysis surrounded by trabecular bone. Ippolito et al. concluded that "the maturation of the hypertrophic cells is altered for some reason and few cells progress to degeneration." There was "an abnormal accumulation and distribution of glycoprotein and proteoglycan in the matrix." They felt that the altered and disorganized matrix of the hypertrophic zone may inhibit chondrocyte degeneration, matrix calcification, capillary invasion, and endochondral ossification, with subsequent thickening and slippage of the growth plate. The dissolution or weakening of the matrix reflected in the finding of clefts, fraying, and granular metamorphosis may either predispose to or result from the slippage of the epiphysis. A major problem with these studies is not the descriptions of the pathoanatomy, but rather the interpretation of their possible causes. The extrinsic causations of extra weight and femoral retroversion seem compelling. Intrinsic causation, by which is meant primary physeal abnormality, appears highly unlikely because all other physes are and remain normal.

Core biopsies of the physis in 21 cases of SCFE as reported by Agamanolis et al. showed findings similar to earlier reports (3). The resting zone was normal. Proliferative and hypertrophic zones were the sites of the main abnormality but were not felt to be as wide as previously described. There was a decrease in chondrocyte cellularity with a relative excess of matrix, marked structural disarray, frequent lobular arrangement of chondrocytes, slits, displacement of cartilage islands into the diaphysis, and reactive callus formation. Histochemical observations did not show a consistent change in the amount of proteoglycans or structural glycoproteins. The authors felt that the changes seen were irregular and secondary. Photomicrographs showed severe misalignment and distortion of the growth plate architecture. Matrix staining was irregular and there were foci of chondrocyte clusters. Ultrastructural studies were not remarkable (4). Cell abnormalities involving chondrocyte degeneration and death were seen not only in the lower hypertrophic zone but at all levels of the proliferative and hypertrophic zones. There was a deficiency of the supporting collagenous framework of the growth plate, which implied a slipping mechanism due to a possible defect in collagen production by chondrocytes. The morphology or distribution of matrix granules of proteoglycan did not appear abnormal. Agamanolis et al. felt the life cycle and turnover of chondrocytes were disturbed. These findings also can be considered as secondary to slipping. Figures 3A-3C illustrate characteristic pathoanatomic specimens. 2. CARTILAGE NECROSIS (CHONDROLYSIS) IN SLIPPED CAPITAL FEMORAL EPIPHYSIS Waidenstrom was the frst to describe cartilage necrosis (chondrolysis) of the femoral head with slipped capital femoral epiphysis, recognizing it as a specific complication (256). Both acetabular and femoral head cartilages were affected, and failure of nutrition of the cartilage by the synovial fluid was at fault. It generally followed varying types of therapy for moderately or severely affected hips. Grossly there was marked thinning to complete disappearance of the joint cartilage. Radiographs showed osteopenia of both the femoral head and acetabulum and progressive joint space narrowing. The pathoanatomy of cartilage necrosis in SCFE became defined more clearly over the next few decades. Ponseti and Barta described a case that came to arthroplasty in midadolescence (215). Photomicrografts showed that the joint cartilage had been replaced extensively by fibrocartilage with minimal cellularity. Studies by Cruess also defined the pathology of acute necrosis of cartilage (56). A synovial biopsy showed chronic inflammation and fibrosis in the early months of the disorder, but eventually 2 years later joint worsening led to cup arthroplasty and availability of the femoral head specimen for assessment. The capsule was thickened greatly and adherent to the neck. The articular cartilage

SECTION II ~ Slipped Capital Femoral Epiphysis

was destroyed almost completely. There was only a thin rim of cartilage over the femoral head bone and this was exclusively fibrocartilaginous. Bone was exposed both on the femoral head and in the acetabulum. The cartilaginous regions were markedly destroyed. The bone was covered with a thin rim of tissue, which in some parts was exclusively fibrous and in other areas was fibrocartilage. The basilar layers of cartilage did retain some cartilaginous appearance. The bone itself was normal in appearance as was the marrow. Similar changes often were found on the acetabular cartilage as well. Areas of hypocellularity were seen in the cartilage, and there were other regions in which the chondrocytes had been destroyed completely over large focal regions. Lance et al. reported on several cases of chondrolysis examined histologically (162). Fissures of the joint cartilage were common in the midst of the cartilage. These generally were covered with fibrinous material and they often were numerous and deep, fragmenting the cartilage into small segments and often passing to the underlying subchondral bone. The characteristic hyaline appearance of the cartilage was lost, exposing a fibrillar collagen network. The chondrocytes were increased for a while in size and number and often were packed together in cartilage clones. On occasion the cartilage would differentiate into a fibrous or fibrocartilaginous tissue, and sometimes spicules of bone were seen within. The subchondral bony lamella was generally thin and the adjacent marrow of the bone showed increased vascular tissue active with new bone formation. The synovium early on was thrown into thickened fronds, the centers of which were fibrous and hypervascularized but without acute inflammatory cells. When studying the synovial-cartilage junction, Lance et al. noted that the cartilaginous fibrils were somewhat less numerous than normal but appeared structurally normal at both light and electron microscopic levels. Opposite the junction of the lower regions of the cartilage and the subchondral bone the collagen fibers were more numerous but otherwise structurally unremarkable. It was opposite the superficial part of the cartilage surface that the most abnormal alterations were seen in which the collagenous fibrils were thickened slightly, often being 3 - 4 times their normal width as measured ultrastructurally. They were fragmented and disoriented in that area and the chondrocytes themselves appeared to be in various stages of degeneration. Jerre found cartilage necrosis in two patients who had no treatment other than crutches (131). Ingram et al. reviewed their own and previous cases in a detailed review of chondrolysis (123). The synovium appeared reddish and edematous in the early stages and thick, fibrosed, and scarred later on. The capsule tended to be thickened also. The articular cartilage was uniformly thin and its surface often was irregular and characterized by pitting, erosion, and fraying. Microscopic changes showed synovium with edema, congestion with infiltration, with round cells involving lymphocytes, plasma cells, and histiocytes in the early phases. Chronic hemorrhage sometimes was seen.

389

There tended to be capillary proliferation, vascular congestion, and edema. Villus formation was common with subsynovial fibrosis in more prolonged instances. The articular cartilage showed marked degeneration similar to what would be seen in osteoarthritis. The cartilage was thinned with fibrillation and fissuring working from the surface downward. The proteoglycans were diminished. The cell density was decreased and there was a tendency for cartilage transformation to a fibrocartilaginous appearance. The subchondral bone plate was active with a tendency to thickening. The appearance, therefore, was quite nonspecific and indicative of early joint inflammation and subsequent cartilage degeneration proceeding to an osteoarthritic state. Although there is little doubt that some cases evolve following pinning of a slip in which the tip of the pin penetrates the articular cartilage, review of the literature from the 1930s on shows an extremely wide variety of treatment programs leading to cartilage necrosis, including reports of some instances in which no treatment at all was performed. The disorder appears to occur, however, in moderate to severe deformities in which the corresponding treatment is more radical. Among the treatments that have been associated with chondrolysis are closed reduction, nailing or pinning of the slipped epiphysis, open reduction, wedge osteotomy of the neck in association with open reduction, prolonged immobilization either prior to or after surgery, strong traction, manipulation, and plaster spica immobilization. The most common treatments leading to the disorder involve closed reduction, open reduction, and wedge osteotomies, with hip spica immobilization after each of these three approaches worsening the stiffness. Unlike avascular necrosis, which essentially is always a complication of treatment, acute necrosis of cartilage can occur in the absence of any treatment.

E. Interpretation of the Studies on Pathogenesis and Pathoanatomy Kleinberg and Buchman noted the temporal aspects of slipping, indicating that displacement and slipping of the femoral head could occur in one of three ways: "it may be very gradual over a considerable period, it may be very sudden, or may be gradual and then be completed by rapid displacement" (152). These words clearly fit with the concept of displacement being chronic, acute, or acute on chronic. Kleinberg and Buchman then related the histologic appearances to considerations of the pathogenesis of deformity. The picture was reminiscent of a repair bone situation in an environment with some, but not complete, stability. Their interpretation incorporated an understanding of the development of the hip toward the end of the growth period and the various mechanical features contributing to the slip. With the approach of puberty, several anatomic and physiological features became evident. The epiphyseal plate between the head and the neck of the femur had changed its position with growth from the horizontal position common in childhood

390

CHAPTER 5 9

Coxa Vara in Developmental and Acquired Abnormalities of t h e Femur

F I G U R E 3 Pathoanatomic specimens from slipped capital femoral epiphysis are shown. (Ai) A severe slip of the head into a medial and posterior direction is seen. The head remains round and the articular cartilage is intact. There has been a relative external rotation positioning of the femoral neck and shaft with a prominent lesser trochanter. There is a rounded but angular prominence of the superior surface of the neck, which plays a major role in limiting abduction. [Derived from (244).] (Aii) The gross specimen shown

SECTION II ~ Slipped Capital Femoral Epiphysis to an oblique plane. There also was a decrease in the thickness of the periosteum and the retinacula of Weitbrecht, which supports the epiphyseal plate and helps to hold the head firmly attached to the neck. With growth the femoral neck had become longer and its angle with the shaft lessened while the density of bone structure had decreased. In addition, there was a relative gain in body weight and an increase in activity. The first features produced physiological weakness at the epiphyseal area, and the latter two conditions produced an increase in the stress and sheafing strain on the epiphyseal area. When any of these changes was slightly beyond the normal limit, the balance became disturbed and pathological changes could occur. On occasion the obliquity of the femoral head was increased, and in other instances there might be more rapid thinning of the periosteum than normal. The situation also was worsened by relatively rapid growth rates in which the growth plate was wider than usual and consequently weakened by increases in body weight or excessive physical activity, which would impose a greater strain than the part could withstand. The primary intrinsic features of the proximal femur leading to a slipped epiphysis are the oblique position of the proximal femoral epiphysis in relation to the neck, the gently curvilinear shape of the proximal femoral capital physis in relation to the neck, the relative thinning of the periosteum during the phase of rapid growth in early adolescence, and retroversion of the proximal femur. Each of these features tends to predispose one to slipping of the head in relation to the neck, but obesity remains the primary extrinsic feature associated with this disorder because the extreme weight of the patient places excessive stress on the physis in relation to the intrinsic factors mentioned earlier. Due to the fact that the slippage is slow, there is not a complete loss of continuity of growth plate cartilage. Excellent studies show oblique positioning of the cartilage cell columns through the physis particularly involving the zone of proliferating cells and occasional slits in the cartilage matrix. There is a persisting continuity of cartilage matrix, however, from the epiphyseal secondary ossification center to the adjacent metaphyseal bone. This disorder in architecture has

391

been considered by some to be a primary causative occurrence, although to most it appears to be the secondary result of the slippage, which must of necessity lead to disarray in the normally well-structured cartilage columns. In association with the chronic slippage, the periosteum reacts in an attempt to stabilize the head on the neck. There is a wellstructured periosteum of the femoral neck that inserts by its outer fibrous layer into the femoral head epiphyseal cartilage just above the physis. This periosteum normally is associated with the remodeling process of the femoral neck, which is exclusively resorptive in nature. Under changed stimuli, however, the inner cambial layer of the periosteum can become osteoblastic; the reactive process serves to widen the neck by bone deposition. With slippage of the head medially and posteriorly relative to the neck, the periosteal regions on the concave side frequently demonstrate new bone formation similar to what one observes after a healing fracture on the concave area of any bone. This can be considered as a mechanism to increase the stability of the head and minimize further slippage. Of clinical importance in treatment, however, is the fact that this new bone formation minimizes the ease with which any reduction can be performed. As the blood supply also is extensive on the medial and posterior surfaces, efforts to reduce the displaced femoral head can lead to further damage to the blood supply by tearing the reactive, shortened, and thickened periosteum, its associated vessels, and the newly synthesized bone. An additional finding is resorption and smoothing of the superior-lateral region of the femoral neck as part of the remodeling process. In the vast majority of instances, positioning of the head is medial and posterior in relation to the femoral neck. Occasionally, the head slips posteriorly only with lateral slips of the head into valgus described, but extremely rare. In those situations in which the head slips completely, the head has nestled into the intertrochanteric region from which position it cannot displace further. The distal fragment (the neck and shaft) moves into the externally rotated position and also tends to migrate proximally as indicated by the position of the greater trochanter in relation to the uppermost level of the femoral head. In cases of moderate and severe slip, therefore, the

FIGURE 3 (continued) in part (i) has been sectioned, allowing for a view of the interiorof the bone. The physis is thickened. There is a shortening of the capsular tissues inferiorlybetween the posterior aspect of the head where it nestles againstthe medial and inferior surface of the neck. This will be the site of periosteal reaction and new bone formation. This is a virtually completeslip with no further displacement of the head able to occur. [Derived from (244).] (Aiii) The neck is at right and the head at left. This is a major and almost complete slip with the head passing posteriorto the neck region. The physis is widened. There has been rounding of the neck superiorly and this appears somewhatelongated. The area of new bone formation at the inferior surface of the neck at its posterioraspect is shown. [Derived from (61).] (Aiv) The head is markedlydisplaced medially and posteriorly in relation to the neck. The articular surface of the head remains intact, but that of the physis cartilage is irregular. [Derived from (82).] (Av) The neck has been approximatelyone-third uncovered at the upper left with slippage occurring posteriorly. This is a lateral projection of the head-neck region and normally would point straight upward along the neck. Note the physeal irregularitywith thinning and replacementby bone at other sites. [Derivedfrom (82).] (B) Illustration from Sprengel depicts the posterior and medial slippage of the head. Approximatelyone-half of the superior surface of the neck is uncovered(d'). Reactivebone on the inferiorsurface of the neck has formed (X). Thickenedphyseal tissue is seen (e) as well as reactive fibro-osseustissue (C). [Derived from (244).] (C) The formation of reactive bone tissue at the inferior and posterior surface of the neck adjacent to the physis was recognizedearly on. Illustration at fight shows the reactive bone spur (a~).This radiographic sign is a feature of chronicity. The adjacent periosteum with its enclosed vasculature is shortened and thickened. [Derived from (61).]

392

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormalities of the Femur

individual will have an abnormal gait characterized by shortening, a Trendelenburg limp due to relative abductor muscle weakness, an external rotation deformity, and discomfort. Because the slippage occurs gradually there is controlled stretching and compensation of the vessels supplying the proximal femoral capital epiphysis, and avascular necrosis of the secondary ossification center does not occur even with complete slips. There also is virtually no tendency to chondrolysis, which is characterized by a marked diminution in the thickness of the articular cartilage surface leading to a decreased joint space. Both of these complications of slipped capital femoral epiphysis are caused in the vast majority of instances by treatment rather than being part of the natural history of the disorder. It is evident that some patients will have a slipped capital femoral epiphysis that spontaneously stabilizes with varying degrees of displacement in the absence of treatment. Many patients present with osteoarthritis in mid to late adult life whose radiographs give an appearance consistent with a mild or moderate slipped capital femoral epiphysis, which had never been realized by the patient or come to specific treatment. Slipped capital femoral epiphysis is a chronic process that occurs gradually over a several-week to several-month period. The displacement frequently is worsened by an episode of relatively mild trauma, although the trauma itself would not generally be considered sufficient to cause a fracture or to account for slippage in an otherwise normal hip. This factor has lead some to use the term "acute on chronic" to describe a variant of the disorder. Although there may well be some accuracy in this description, there are possible problems in using the concept as a diagnostic category from which certain treatment options will evolve. Milch was one of the earliest to recognize that "it was the neck of the femur which slipped, not the head" (186). He felt that it was not the capital epiphysis that slipped, at least initially, because the head remained constantly in its normal relation to the acetabulum. It was the external rotation or anteversion of the femoral neck and thus of the entire femur that really shifted such that the primary and characteristic change in the position of the neck of the femur was not one of varus but unequivocally one of anteversion. The slipping thus was a rotation not so much of the head posteriorly but of an external movement of the upper end of the femoral neck and shaft beneath the epiphyseal plate of the head. When a series of X rays in varying positions were taken, the neck was noted to point forward and the head was posterior to the plane of the femoral neck. This finding coincided with the fact that external rotation of the lower extremity was among the earliest of the clinical symptoms and stressed that it was the external rotation of the neck (tending to an anteverted position) that really was the precipitating event in the slipping. It currently is recognized that the femoral head actually is in some relative retroversion to the normal plane prior to slippage, but it is still felt by most that the external

rotation of the neck and shaft is the primary event with the head reacting and passing to a posterior position secondarily. Griffith also assessed the position of the displaced epiphysis in relation to the neck in laboratory studies and in clinical cases (93). In the morphological studies the proximal ends of the femurs of six children aged 10-15 years dying of unrelated disorders were removed and the periosteum was divided circumferentially. Displacement of the epiphysis was readily performed and radiographs were taken following displacement. Griffith felt that displacement was almost exclusively posterior in the disorder. When such displacement was performed on the human specimens and the specimen was placed with the neck of the femur at right angles to the X-ray beam, the epiphysis appeared on the radiograph to lie directly behind or posterior to the neck. If the specimen was placed in relative external rotation with relation to the X-ray beam, the epiphysis appeared to lie on the medial aspect of the neck. Therefore, he felt that the appearance of medial slipping was due to the effect of parallax. Similar X rays were taken with clinical patients, and he again concluded that the epiphysis almost invariably was displaced in the plane that lies posteriorly at right angles to the anteversion plane of the neck of the femur. There was no true medial displacement of the epiphysis with respect to the neck, and appropriate studies served to define the plane of displacement as being strictly posterior. With displacement the anterior aspect of the neck subsequently was resorbed and new bone was synthesized posteriorly. As the epiphysis slipped downward and backward the periosteum and epiphyseal vessels were stripped off the posterior aspect of the neck and new bone was synthesized beneath this periosteum. Ireland and Newman also demonstrated clinical specimens from postmortem analyses of slipped epiphysis (125). In a severe untreated slip the epiphysis rotated backward around the axis of the cylindrical upper growth plate of the femur. The degree of slip that can occur is limited by abutment of the back of the epiphysis against the back of the femoral neck. In such situations the epiphysis has rotated approximately 90 ~ posteriorly. If the extent of slip was expressed in terms of the diameter of the femoral growth plate as seen on the lateral radiograph, the maximum slip really corresponded to one of approximately three-fourths of the diameter. Ireland and Newman thus reasoned that slip of one-third of the diameter corresponded to one of 45 ~ and half the diameter to 60 ~. The plane of slippage defined earlier is illustrated in Figs. 4A-E.

F. Medical Disorders Predisposing to Slipped Capital Femoral Epiphysis 1. ENDOCRINEDISORDERS Delayed skeletal maturation is a common finding in endocrinopathies of any type. The likelihood of the presence of an underlying endocrine disorder is raised markedly in

SECTION II ~ Slipped Capital Femoral Epiphysis

FIGURE 4 (A) Anteroposterior hip radiograph shows minimal slippage on right and a normal hip on left. The dotted line drawn along the superior surface of the neck passes within the secondary ossification center on the normal side but only touches it when minimal slipping has occurred. (B) Radiograph shows early bilateral displacement. The growth plate is widened with linear opacities and increased epiphyseal side sclerosis. (C) The initial plane of slippage of the head primarily in a posterior direction is illustrated. Radiographicfrog lateral projection showsa slightly widened physis.

those who have a slip before 10 years of age or after 15 years of age. The pinning in situ treatment is effective, however, in these disorders. a. Hypothyroidism Hypothyroidism represents the most common endocrine disorder associated with SCFE. The first report of hypothyroidism and SCFE was by Lewin in 1928, who documented a bilateral slip in a 9-year-old hypothyroid girl (167). Heyerman and Weiner reported 7 cases of SCFE appearing at the same time as a juvenile hypothyroid state (110). Studies in hypothyroidism have documented changed growth plate structure. The physis generally is thickened, the appearance of secondary ossification centers is delayed and irregular, and the skeletal age is retarded. Irregularity and alteration of the metaphysis adjacent to epiphyses other than

393

the hip and similar to some of the changes in rickets tend to be seen. Delayed skeletal age and a higher than normal incidence of bilaterality characterized hypothyroid slips. Puri et al. presented 9 patients whose slips were associated with primary juvenile hypothyroidism; in all patients the slip occurred or the symptoms developed in the affected hip before the hypothyroidism was diagnosed (219). Bilateral slippage was seen in 6 of 9, and all patients were characterized by obesity, shortened stature, and delayed skeletal maturation. When the weight percentile was related to skeletal age, 5 of 8 patients were well beyond the 97th percentile, 2 were at the 95th percentile, and only 1 was within the normal range at the 50th percentile. That patient, however, was only at the 15th percentile in terms of height so relative obesity was noted even there. The delay in skeletal maturation ranged between 2 and 6 years. Puri et al. reviewed the previous literature on juvenile hypothyroidism and SCFE, documenting 18 cases. A large majority of those also had a marked delay in skeletal maturation and obesity tended to be present as well. Radiographically hypothyroidism was characterized by generalized growth retardation as indicated by delays in skeletal maturation, for example, delay in the normal time of appearance of the secondary ossification centers, and at the other end of the growth period, delayed closure of the growth plate. Descriptions in the older literature of fragmentation of various secondary centers refer to congenital hypothyroidism; the juvenile onset hypothyroidism is not characterized by that finding. b. Broad Range o f Other Endocrine Disorders Wells et al. cited the incidence of SCFE in patients with associated endocrine disease (267). In a review of 131 patients with SCFE, 9 (6.9%) had an associated endocrine problem. Of these, one was hypogonadal with Turner syndrome, 4 had panhypopituitary endocrinopathy with 3 of these due to craniopharyngioma, and 4 were hypothyroid of various etiologies. Although relatively few of these patients were obese in the sense of having weight beyond the normal range, 8 of the 9 patients displayed a body weight percentile much greater than the height percentile, with the mean percentile of weight at 63.6 and the mean percentile of height at 24.7. As in most endocrinopathies, the average chronological age of slipping was somewhat later than in idiopathic SCFE with the mean at 14.3 years. There was a mean bone age delay in relation to chronological age of 4.6 years as patients with endocrinopathies frequently close their physes at a later time period. Almost all patients had the acute on chronic designation. When the patients were followed to skeletal maturity, all of the 9 had bilateral slips. Wells et al. recommended that T4, TSH, and bone age determinations be performed on all patients with slipped capital femoral epiphysis initially. Mann et al., however, in a prospective study felt that routine hormonal studies in all patients with slipped capital femoral epiphysis produced a very low yield and they did not recommend such testing (177). Puri et al. also reviewed published

394

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

reports of SCFE associated with other endocrine disorders, noting a high incidence of craniopharyngioma, nonspecific hypopituitarism, hypogonadal states, and growth hormone deficiency states (219). Most patients with craniopharyngioma and hypopituitarism have an associated hypothyroidism. A more recent detailed review of SCFE associated with endocrine disorders was by Loder et al., assessing 85 patients described in several papers, including 6 of their own, and subdividing the group by endocrinopathy (171). Many interesting features of the cumulative study were presented, which clarified issues raised in previous studies in which there were insufficient patients to offer definitive corroboration. In patients from their own institution, 115 had SCFE with 6 having endocrinopathy, an incidence of 5.2%. In the cumulative series, the most common primary endocrine diagnosis was hypothyroidism in 34 patients (40%), growth hormone deficiency in 21 (25%), with others accounting for 35% of disorders. Panhypopituitarism, often caused by craniopharyngioma, also was a common cause. Other less frequently seen disorders were hypogonadism, hyperparathyroidism, growth hormone excess, and cases of multiple endocrine neoplasia. The entities of hypothyroidism, decreased growth hormone, and panhypopituitarism including craniopharyngioma accounted for 82% of the endocrinebased SCFE cases. A craniopharyngioma leading to variable endocrinopathies was present in 20 (24%). In those patients adequately documented 61% had bilateral involvement. The age of diagnosis was 13.2 + 6.2 years for the endocrine disorder, but the age at diagnosis at the first SCFE was later at 15.3 + 5.3 years. Bone age was significantly retarded with a mean of 11.6 years, whereas the chronological age was a mean of 16.5 years. Idiopathic SCFE usually is diagnosed between 10 and 16 years of age, but in the endocrine-associated group 29 of 81 (36%) were seen with their first slip when less than 10 or greater than 16 years (21 patients) of age. Only those with hypothyroidism or growth hormone deficiency were less than 10 years of age, whereas all other endocrinopathy patients when first seen at an atypical age were greater than 16 years of age. The distribution of time prior to diagnosis and severity were not markedly different from the uncomplicated SCFE. There were 74 chronic and 13 acute slips. Regarding severity, there were 48 mild, 20 moderate, 12 severe, and 2 preslips. The prevalence of bilaterality was 61%, whereas in a review of non-endocrineassociated SCFEs it was 37%. Most of the hypothyroid children, 23 of 33 (70%), had the endocrine diagnosis made at presentation of the first slip, whereas all of the growthhormone-deficient children had the endocrine diagnosis made before diagnosis of the slip. The majority of the children were of short stature with 58% (26 of 45) in the lower 10th percentile. Of these however, 32% (16 of 50) were obese at greater than the 95th percentile. This relative obesity further supported the role of excess weight in relation to SCFE. c. Hyperparathyroidism Primary hyperparathyroidism due to underlying parathyroid adenoma has been described

as an associated cause of slipped capital femoral epiphysis (30, 147). The primary hyperparathyroidism was due to an adenoma in three reported cases. Two of the three cases were bilateral. The time of occurrence was 13.5 years of age in one patient, who was also obese, and 16 years of age in another. The hyperparathyroid disorder was characterized by marked widening of the femoral growth plate and considerable proximal metaphyseal bone resorption. Additional X rays showed widening and irregularity of the epiphyseal plates of the proximal humerus, distal femur, and tibia as well as sub-periosteal resorption of bone in the finger phalanges. Laboratory tests were characteristic with elevated calcium, decreased phosphorus, increased alkaline phosphatase, and increased levels of PTH-C and PTH-MID. The parathyroid adenoma was diagnosed by a combination of CT scanning, ultrasonography, and MR imaging. In one instance, the bone resorption quickly corrected and no further slippage occurred in association with removal of the parathyroid adenoma alone, whereas in the other patients pinning in situ was performed after removal of the adenoma. McAfee and Cady presented four cases of atypical SCFE associated with hypopituitarism, renal osteodystrophy, and two instances of radiation therapy for rhabdomyosarcoma of the ischium and a malignant sarcoma of the right sacroiliac joint (183). Their extensive literature review indicated the endocrinologic and metabolic disorders associated with SCFE. They detailed the diagnostic steps in the investigation of atypical SCFE. Of particular note is the need to perform more detailed assessments in patients younger than 10 or older than 16 years of age. Other considerations are extremes of height or weight, although this is present in any patient with SCFE, sudden or unexpected changes in growth, precocious puberty or menarche, and actual cases of known endocrinopathy or post-radiation treatment for tumors. In many instances, with endocrinopathies it is the slip that occurs first with the underlying diagnosis presenting only afterward. In those with known endocrine, metabolic, or neoplastic disorders, symptoms of SCFE often are missed because they are confused with the underlying disorder and no particular attention is paid to the hips until slippage is extreme. 2. POST-RADIOTHERAPY Another subset of patients with SCFE has had previous episodes of radiation therapy for control of tumors. The radiation, which often precedes the slip by several years, is centered in the lower abdominal or pelvic regions. The earliest awareness of the association of radiotherapy with slipping of the upper femoral epiphysis was described by Wolf et al. in 1977 (273). Many reports have confirmed the reality of this association. Chapman et al. reported SCFE in five patients following radiotherapy given for management of Wilm's tumor, abdominal neuroblastoma, pelvic yolk sac tumor, scrotal rhabdomyosarcoma, and retroperitoneal neuroblastoma (47). The increased incidence of SCFE in these patients following radiotherapy is highly statistically signif-

SECTION II ~ Slipped Capital Femoral Epiphysis

icant. The instance of bilaterality in this group also is quite high, with 60% showing bilateral slips. Although there has been much speculation about the cause of the slip, the effect of the radiation on the proximal femoral capital growth plate is not clear. The slip generally occurs several years after the treatment with the times of occurrence in the preceding series ranging from 7 to 9 years. Barrett described two cases following rhabdomyosarcoma radiation therapy with patients presenting 3 and 9 years following treatment (21). A review of the literature presented 18 additional cases. The incidence of bilaterality was 50%. Although the cases occurred several years following therapy, the overall age in the post-radiotherapy cases still was significantly lower than that in uncomplicated SCFE, with the mean occurring at 11 years. Associated chemotherapy also may contribute to the slipping. Radiation, however, appears to be the primary agent because many postradiation patients develop the disorder in the absence of chemotherapy. Walker et al. reported two cases of SCFE following courses of radiation therapy and chemotherapy (257). Radiation therapy in which the portal included both femoral heads was for malignant lymphoma and prostatic rhabdomyosarcoma. The critical level of rads needed to contribute to the SCFE phenomenon has not been determined, but levels in excess of 2400 rad have been reported with a range between 2400 and 6000 rad thought to be the causative level in the various reviews. 3. SLIPPED EPIPHYSES IN JUVENILE RENAL OSTEODYSTROPHY The third major association of SCFE with other medical disorders is in patients with childhood renal osteodystrophy. Brailsford described the influence of renal tickets on the slipping of the proximal epiphysis of the femur in 1933 (34). He described in detail 2 cases showing that the deformity was due to slipping of the epiphysis of the head of the femur rather than due to bowing of the proximal end of the femur. This correlation was substantiated by Mehls et al., who noted slipped epiphysis in 11 out of 112 children with renal osteodystrophy (184). Their histologic studies, along with those of Krempien et al. (158), showed that epiphyseal displacement was not due to traumatic physeal or metaphyseal fractures. The disorder was associated with hyperparathyroidism with elevated serum parathyroid hormone levels, with a good correlation noted between the serum PTH levels and osteoclastic resorption and periosteal fibrosis. Osteoclasts and chondroclastic resorption greatly diminished the continuity between physeal and metaphyseal tissues. The hypertrophic zone was thinned, and the well-ordered trabeculae of calcified cartilage surrounded by new bone were absent or severely distorted. Instead, fibrous tissue and poorly organized woven bone were present, and it was through this weakened area that slippage occurred. Sub-periosteal resorption also was intense, weakening the supportive periphyseal bone of the groove of Ranvier region. The radiolucent zone

395

between the secondary ossification center and the metaphysis is not similar to that in rickets in which there is poorly or nonmineralized cartilage but rather is due to the accumulation of woven bone and fibrous tissue. In severe examples, no continuity of physeal cartilage and metaphyseal bone trabeculae was seen. Histological studies showed that epiphyseal slippage in uremic children was the ultimate consequence of osteitis fibrosa in which the physeal-metaphyseal junction was filled by dense fibrous tissue interspersed with poorly organized woven bone. In the growth zone of the uremic children, the cartilaginous growth plate is separated from the metaphyseal bone by chondroclastic removal of hypertrophic cartilage and its replacement primarily by dense fibrous tissue but occasionally by structurally imperfect woven bone. The response to vitamin D therapy was excellent in most cases. Many primary renal disorders contributed to the end-stage osteodystrophy phenomenon. The characteristic X-ray features of epiphyseal slipping in uremia showed an irregular growth plate, a radiolucent and stippled adjacent metaphysis, and sub-periosteal metaphyseal and diaphyseal resorption. The metaphyseal trabeculae were coarse, medical treatment involved the use of vitamin D3, and in many instances surgical pinning was also performed. In those patients younger than 8 or 9 years of age, treatment with a central nonthreaded pin to stabilize the physis during healing along with medical therapy often is recommended as premature fusion would lead to further problems in the structure of the proximal femur. Hartjen and Koman described treatment in very young patients using a specially constructed screw in which the distal threads were removed but those in the neck adjacent to the head were left intact to allow for some stabilization (104). This disorder represents essentially the only one in which slipped epiphysis occurs at areas other than the proximal femur. The study of Mehls et al. documented occurrences of slipped epiphyses in virtually all patients at the distal radius and distal ulna, with fairly frequent occurrences also at the distal femur, distal tibia, and proximal humerus (184). Nixon and Douglas commented on the presence of severe secondary hyperparathyroidism in two of their cases (198). Internal fixation was performed as well as medical therapy with a good response. In 1883 Lucas described renal tickets of adolescence characterized by bending of the bones seemingly due to epiphyseal displacement, although he did not specifically allude to true epiphyseal slipping (174). None of the deformities described were in the hips, but knock knees (genu valgum) and flat feet (valgus ankles) were reported. Shea and Mankin reviewed eight cases from the literature and described three of their own (237). Floman et al. reported bilateral slipped upper femoral epiphysis in an 18-year-old male with renal osteodystrophy (80). Successful bone fusion without additional slippage occurred without orthopedic management following treatment involving subtotal parathyroidectomy, chronic hemodialysis, and renal transplantation.

396

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormaflties o f the Femur

A 33% slip

Post.

50% slip

nt.

F I G U R E 5 Two common ways of measuring the degree of slippage are used. (A) Lateral radiographs illustrate the angular tilt of the head as well as the extent of neck uncovering. (B) A quantitative measurement of the extent of slippage is the deformation angle defined by Southwick and modified by others. [9 1996 American Academy of Orthopaedic Surgeons. Reprinted from the Journal of the American Academy of Orthopaedic Surgeons, Volume 4(4), pp. 173-181 with permission.]

G. Types of Classification for the SCFE Entity 1. DURATION OF SYMPTOMS

Classification based on the duration of symptoms leads to description of the entity as being either acute, chronic, or acute on chronic. a. Acute Slip An acute slip is one in which there is a relatively sudden onset of symptoms, which have been present for less than 2 weeks in some categorizations or less than 3 weeks in others. b. Chronic Slip A chronic slip is one in which symptoms have been present for more than 3 weeks. c. Acute on Chronic Slip An acute on chronic slip implies the existence of a chronic slip in which symptoms have been present for more than 3 weeks onto which is superimposed a recent and sudden worsening of symptoms and deformity caused by further acute displacement, which is generally associated with an appreciable degree of trauma. 2. SEVERITY OF DISPLACEMENT BASED ON RADIOGRAPHIC APPEARANCE

Radiographic classifications are based on the extent of displacement of the femoral head with respect to the femoral neck. Several stages of slippage based on the amount of displacement can be defined ranging from a preslip, to a mild, moderate, or severe slip; and finally to a complete slip. In the preslip stage the patient has some hip and thigh area discomfort, although both anteroposterior and lateral radiographs show no loss of position of the femoral head in relation to the neck. The radiographic finding that leads to a diagnosis of preslip is evidence of slight widening and irregularity of the physis associated with bone resorption on the metaphy-

seal side. There may well be some loss of internal rotation on clinical examination due to an associated synovitis. Mild, moderate, and severe slips can be quantified by the amount of the epiphyseal displacement relative to the neck. If the translocation is less than one-third of the diameter of the superior surface of the neck, the disorder is referred to as mild, if it is one-third to one-half of the diameter it is considered to be moderate, and if greater than one-half it is severe (Fig. 5A). In a complete slip the femoral head is nestled into the intertrochanteric notch on the posterior surface of the neck. Other variants of this categorization technique were proposed by Bianco, with 0-33% being mild, 33-67% moderate, and 67100% severe (25), and by Kulick and Denton, who further subdivided into four grades beginning with from 0-25% as grade I and concluding with 75-100% as grade IV (160). A widely used quantitative radiographic classification of the severity of displacement measured on a frog lateral view is based on the angular deformity between the long axis of the shaft and neck and the long axis of a line drawn through the center of the femoral head perpendicular to a line linking the two edges of the physis of the epiphysis (Fig. 5B). The angle of head tilt is determined by subtracting the normal side angle from that of the affected side. Where bilateral slips are present, the angle measured minus 12 ~ is considered to be the deformation angle. This index, described by Southwick (241) and modified by Ingram et al. (123), considers an angle less than 29 ~ to represent a mild slip, 30-50 ~ to represent a moderate slip, and greater than 60 ~ to represent a severe slip. Cohen et al. derived a slip index from computerized tomographic scans, which determined the head-neck angle (51). This increased with the further degree of posterior slipping of the head relative to the neck.

SECTION II ~ Slipped Capital Femoral Epiphysis

397

3. STABLE-UNSTABLE CLASSIFICATION (LODER E T AL. )

Loder et al. pointed out the importance of the stability of the head on the neck in relation to prognosis (169). They divided patients into those whose slips were stable and unstable. Patients with stable hips were able to walk, whereas those with unstable slips could not bear weight, even with the use of crutches. They felt that this classification was useful in predicting the likelihood of AVN because that complication developed in approximately 50% of patients with unstable slips but in less than 10% of those with stable slips. The stability also correlated with prognosis, with a satisfactory result in 96% of stable hips but a diminution to only 47% satisfactory in unstable hips. The stability classification focuses on events at the physeal junction rather than on the time of their occurrence. In the original study by Loder et al. of 55 slips that would have been classified as acute with symptoms less than 3 weeks in duration, 25 were stable and 30 were unstable (169). The prognosis correlated with the stability rather than the temporal aspect. Slips designated as acute on chronic by the temporal approach were not found to be within one group but could be either stable or unstable. It was felt that the end result was more dependent on the latter (stable-unstable) categorization and that the acute on chronic term was of little value in directing management expectations. AVN developed in 14 of 55 children, all of whom had unstable hips. 4. ULTRASONOGRAPHIC ASSESSMENT OF STABILITY AND REMODELING (KALLIO E T AL.)

Kallio et al. studied 26 cases of SCFE using ultrasonographic appearances, which enabled them to assess whether the hip joint had an effusion, whether remodeling had occurred, and the degree of displacement (136). They proposed a new classification based on a combination of physeal stability and chronicity. An acute SCFE was characterized by an effusion, which itself indicated physeal instability or recent progression. A chronic SCFE was characterized by an absence of effusion but evidence of remodeling based on rounding of the anterior aspect of the metaphysis, which also indicated a duration of symptoms of at least 3 weeks. In an acute on chronic situation, there was both effusion, indicating instability, and remodeling, indicating a previous stable or chronic state (Fig. 6). In a subsequent study, Kallio et al. felt that the true prevalence of instability may be higher than diagnosed by the clinical criteria of Loder because they noted an effusion by ultrasound in 60% (33 of 55 cases) in an unselected series (135). They felt that both clinical weight beating status and sonographic study would lead to a more accurate prediction of physeal stability or instability. A slip was considered very unlikely to be unstable in a child able to bear weight and with no sonographic effusion. Serial radiographs showed partial or complete spontaneous reduction in 22% (12 of 55 cases) with an average change of 15 ~ in the

F I G U R E 6 The ultrasonographic categorization of Kallio et al. assesses the degree of femoral head displacement relative to the neck, the presence or absence of a joint effusion, and the presence or absence of remodeling of the neck. [Reprinted from Kallio et al. (1993), Clin. Orthop. Rel. Res. 294:196-203, 9 Lippincott Williams & Wilkins, with permission.]

head-neck angle. When serial sonographic studies were performed, reduction was noted in 7 out of 20 cases (35%) with an average change of 3.7 mm in displacement. Bone scans were done. They felt that all stable hips had normal epiphyseal vascularity. Of the hips that showed subsequent reduction, 12 had had a bone scan on admission with 3 showing initial epiphyseal avascularity and only 1 progressing to symptomatic AVN. They thus felt that AVN primarily was due to the events surrounding the initial injury rather than to those relating to any spontaneous reduction. 5. VALGUS SLIPPING

The vast majority of slips result in varus positioning of the head in relation to the proximal femur. There have been rare reports of valgus slips, however. Howorth documented 3 valgus slips out of 234 (119). Imhauser recognized the valgus slip (121). Segal et al. reported 2 patients with valgus slips (233). It has been suggested that a preexisting coxa valga and increased femoral anteversion might be predisposing (239). Treatment by pinning in situ has been successful. Valgus position of the femoral head can occur as a late sequel

398

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormalities of the Femur

to developmental dysplasia of the hip, but is due to a partial lateral growth plate arrest secondary to avascular necrosis of the head rather than to mechanical slippage (87).

H. Epidemiologic Characteristics of Slipped Capital Femoral Epiphysis: Age, Sex, Weight, Symptom Time, Bilaterality, and Associated Disorders 1. GENERALOVERVIEW Slipped capital femoral epiphysis is the most common hip condition of adolescence. Early treatment when there is minimal slipping can lead to excellent results. Delay in diagnosis allows further slippage and deformity to occur, necessitating more intricate or extensive surgery with increased complications, most of which then predispose one to degenerative hip disease. It was emphasized as long ago as the 1930s that early treatment leads togood results, yet delay between the onset of symptoms and the time of diagnosis continues (182, 271). Reports persist today on large numbers of osteotomies for moderate affd:~severe slips with controversy still surrounding methods of treatment and the potential hazards of the various surgical procedures. Data compiled over the past 4 decades, however, seemingly repeat the same message concerning age and weight distribution at the time of diagnosis. The problem of delayed diagnosis is due to a considerable extent to the fact that early symptoms are relatively mild. The discomfort is frequently present in the thigh or knee region only. The characteristic clinical finding on hip examination is decreased internal rotation. The anteroposterior and particularly the lateral hip radiographs are diagnostic with even minimal slippage and at times with no slippage at all. Bilateral involvement in slipped capital femoral epiphysis has often been reported to be approximately 20% but now is recognized to be much higher. The frequency is of great clinical importance as it relates to the need for (1) careful initial assessment of the often asymptomatic contralateral hip, (2) consideration of prophylactic treatment of the opposite hip, and (3) continuing assessment of opposite hip involvement until skeletal maturity. As second side slips can occur several months to years after the first side, the true incidence of bilaterality can be ascertained most accurately only by following a large number of patients to skeletal maturity. 2. CHILDREN'S HOSPITAL, BOSTON, CLINICAL EPIDEMIOLOGIC DATA Several parameters~were studied in 236 patients with slipped capital femoral epiphysis seen at the Boston Children's Hospital from 1954 to 1981 (236). All diagnoses were based on plain anteroposterior and frog lateral radiographs as well as clinical:rriteria. Assessments included (1) sex distribution, (2) ag~'at time of initial diagnosis, (3) symptom time prior to diagmisis, (4) unilateral or bilateral involvement, (5) weight at time of initial diagnosis, and (6) associated conditions with SCFE. In those with bilateral involvement,

it was determined whether both hips were involved at initial presentation or whether the second hip became involved later. In order to provide an accurate indication of whether the disorder was truly unilateral it was necessary to follow a patient until skeletal maturity. Cases were considered to be definitively unilateral when the opposite hip was shown radiographically to be fused or fusing without slippage. The weight at initial diagnosis was charted on the National Center for Health Statistics percentiles as adapted for the Ross Laboratory charts. The sex distribution, age at diagnosis, and weight at diagnosis also were related to bilaterality. Eleven patients (4.7%) had slipped epiphyses in association with recognized predisposing causes, including postirradiation of sarcomas, renal disorders, and endocrinopathies due to craniopharyngioma, hypothyroidism, and panhypopituitarism. a. Sex Distribution There were 236 patients, 159 male and 77 female (2.1:1). b. Age at Time of Initial Diagnosis For males the average age at diagnosis was 13 years 2 months, with a range from 5 years 7 months to 16 years 11 months. Ninety percent of cases presented between the ages of 11 and 15 years 11 months. For females the average age at diagnosis was 11 years 5 months, with a range between 8 years 6 months and 15 years 3 months. Ninety percent of cases presented between the ages of 9 and 12 years 11 months. The determinations were made on 217 patients, 150 males and 67 females (2.2:1). The data are illustrated in Fig. 7A. There was no difference in the average age at initial involvement for both males and females in unilateral and bilateral groups (Table I). c. Symptom Time prior to Diagnosis The symptom time prior to diagnosis in the entire series was 0 - 3 weeks, 14%; 4-11 weeks, 30%; 3-5 months, 30%; 6-12 months, 12%; and > 1 year, 14%. In an effort to determine whether the time between the onset of symptoms and the time of diagnosis had changed during the period of study, the number of cases per decade was assessed. The bulk of the cases were in the 1960s and 1970s. The relatively few cases from the 1950s were combined with those from the 1960s, and cases from the 1970s were combined with those from the 1980s. There was no change from the onset of symptoms to time of diagnosis between the two periods. These determinations were made

TABLE I Average Age a t Initial Involvement in Unilateral and Bilateral Groups a

Unilateral Male 13 years + 3 months

Female 11 years + 8 months

Bilateral Male 13 years + 3 months

Female 11 years + 1 month

aBased on 183 patients with accurate data.

S E C T I O N !1 ~ S l i p p e d C a p i t a l F e m o r a l E p i p h y s i s

399

A 40

FEMA L E 90%

35

i

3O

I

:;'52015105 lll~

0 25

90~ MALE

20 15 10 5 0

.....

5%" +%" P-7" +o_+,, 9':'-9"1o010"110--1''1 12~ ' i+~ " ~4~ '' +5o_~+,morn,' AGE A T DIAGNOSIS

FIGURE 7 Datafrom the Children's Hospital, Boston, study concerning age at the time of initial diagnosis. (A) Age at the time of initial diagnosis. (B) Symptomtime prior to diagnosis.

on 175 patients, 116 males and 59 females (2:1). (Fig. 7B and Table II). d. Unilateral and Bilateral Involvement These determinations were made on the full series of 236 patients. The male and female findings were quite similar. There was bilateral involvement in 92. There was unilateral involvement in 122, who were followed to skeletal maturity with radiographs demonstrating opposite side fusion without slippage. When the percentage of those with bilateral involvement is calculated in relation to this group, a 43% incidence of bilaterality is noted. There were 22 patients who still had an open proximal femoral physis on the opposite hip. These

TABLE II

patients either had been lost to follow-up or did not return for assessment because they seemingly were fine. It is important to consider them in the statistical assessment in efforts to reach the most accurate determination of the incidence of bilaterality as some of them almost certainly developed a second side slip late. It has been noted both in this series and elsewhere that second side involvement may be minimally symptomatic or sometimes asymptomatic. If none of these 22 developed a second side slip the incidence of bilaterality would still have been a minimum of 39%, and if all slipped the incidence would rise to 48%. As neither extreme would have been expected to happen, an intermediate

S y m p t o m Time Prior t o D i a g n o s i s from t h e Children's Hospital, Boston, Study a 1950s

1960s

1970s

1980s

Entire series

4 8

15 45

37 53

3 10

59 116

Female Male

72

103

Male

Female

Symptom time

Female

Male

Total

11 (15%)

7

4

0-3 weeks

3

10

13 (13%)

24 (14%)

20 (28%)

18

2

4-11 weeks

15

17

32 (31%)

52 (30%)

24 (33%)

15

9

3-5 months

11

17

28 (27%)

52 (30%)

4 (6%)

3

1

6-12 months

7

11

18 (17%)

22 (12%)

13 (18%)

10

3

1 year +

3

9

12 (12%)

25 (14%)

aBased on 175 patients with accurate data.

400

CHAPTER 5 9 Coxa Vara in Developmental and Acquired AbnormaBties o f the Femur TABLE III W e i g h t o f Patients w i t h Slipped Capital Femoral Epiphysis a t t h e Time o f Diagnosis a Percentile Number of patients

Males 114 Females 53 Total 167

>95th

>90th

>75th

>50th

48

65

87

95

45

62

79

89

47

64

84

93

aWeight charted on the Ross Laboratory charts derived from the National Center for Health Statistics percentiles, Hamill PVV et al. (1979) Am J. Clin Nutr 32:607-629. Based on 167 patients with accurate data. FIGURE 8 Unilateral and bilateral involvement from the Children's Hospital, Boston, study is illustrated. value appears likely. The 5 female patients were 10 (2) and 12 years (3) of age at latest assessment, and the 17 male patients were 9, 11, 12 (5), 13 (7), and 14 (3) years of age. As these ages fall within the peak 90% range of age at occurrence, the likelihood of some additional but unrecorded slips is high. If we project that as few as 6 of these 22 (27%) had second side slips, then the overall incidence would persist at 42%. A determination of 4 0 - 4 2 % bilaterality, therefore, is supported by the data in this study. The data are illustrated in Fig. 8. In those with unilateral involvement, a left-sided predominance of 82 to 62 (1.3:1) was noted with the slight left predominance seen in both males and females. In those with bilateral involvement, 46 were bilateral at the time of initial assessment and 46 developed an opposite side slip from 1 week to 2.5 years after the initial slip. e. Weight at Initial Diagnosis The weight of 47% of patients at the time of diagnosis was above the 95th percentile; 64% were above the 90th percentile; 84% above the 75th percentile; and 93% above the 50th percentile. The percentages were virtually the same in males and females. The determinations were made on 167 patients, 114 males and 53 females. (2.2:1). (Table III). There was no difference in the weight percentile distribution at initial involvement for males and females in the unilateral and bilateral groups (Table IV). f. Associated Conditions Only 5% of cases were associated with other conditions, which included slips postirradiation of pelvic, retroperitoneal, and abdominal sarcomas (5), renal disorders (2), and endocrinopathies such as craniopharyngioma, hypothyroidism, and panhypopituitarism (4). 3. EPIDEMIOLOGIC DATA IN RELATION TO OTHER STUDIES a. Gender Predominance In all large studies males are affected more commonly than females. The male:female ra-

tio of 2.1:1 in our series is the same as that found in two large American series (141, 272) and differs only slightly from previous European reports [with the exception of Jerre's report (131) from Sweden, which showed a 4.9:1 male predominance, although a subsequent Swedish report (107) noted a 1.9:1 ratio]. Sorenson reported a 3:1 male: female ratio in 101 Danish patients seen between 1957 and 1964 (240). Burrows noted a 1.5:1 male:female ratio in 100 British patients (41). Durbin and Hansson et al. documented a 2.3:1 male:female ratio in 532 patients in southern Sweden studied from 1910 to 1982, although the ratio decreased to 1.5:1 in more recent decades (102). A huge multinational study by Loder et al. of 1993 slipped epiphyses documented a 1.4:1 male:female ratio (168). b. Age at Diagnosis The age at diagnosis is strikingly similar in reports dating as far back as 1931. In our series the average age of female involvement was 11 years 5 months.

TABLE IV W e i g h t Percentile Distribution o f Patients with Unilateral a n d Bilateral Involvement a Weight percentile at initial involvement >95th

Unilateral Male Female

22 12 34 (41%)

Bilateral Male Female

20 8 28 (39%)

90-95th

60-89th

9 7

13 3

16 (19%) 10 6 16 (22%)

aBased on 155 patients with accurate data.

16 (19%) 13 8 21 (29%)

<60th

10 7 17 = 83 (21%) 4 3 7 = 72 (10%)

SECTION II ~ Slipped Capital Femoral Epiphysis TABLE V

401

S u m m a r y o f Age, W e i g h t , a n d Bilaterality D a t a C o n c e r n i n g Slipped Capital Femoral Epiphysis from Previous Major Studies

Author(s) (country)

Number of patients

Male: female

Ferguson and Howorth (USA) (73)

70

Wilson, Jacobs, and Schechter (USA)

Age at occurrence Male

Female

1.3:1

75% between 12 and 14 years

89% between 10 and 13 years

21%

65% overweight

40

2.4:1

14 years (peak)

11 years (peak)

28% (12% bilateral initially; rest later up to 2 years)

75% with obesity slight slipping; 34% obese; severe slipping 50% obese; bilateral 72% obese

66

2:1

13 years

11 years

32% (67% bilateral initially; rest noted on average 1 year later)

13 years (median)

11 years (median)

A considerable proportion of patients with slipped epiphysis are very much overweight.

90% between 11 and 16 years

94% between 9 and 14 years

Weight of 49% of males and females for age was greater than the 95th percentile. 90% boys and 86% girls greater than the 50th percentile.

(272)

Cowell (USA) (53)

Kelsey, Keggi, and Southwick (USA) (141)

431

2.2:1

Kelsi, Acheson, and Keggi (USA) (142)

Bilaterality

Jerre (Sweden) (131)

166

4.9:1

Henrikson (Sweden)

81

1.9:1

13.5 years mean

11.8 years mean

22%

Burrows (England)

100

1.5:1

15 years

11 years

23%

Sorenson (Denmark) (240)

101

3:1

532

2.3:1

11 years (peak), 76% between 10 and 14 years 11.8 years mean (1950-1969)

25% (one-half bilateral initially; one-half later)

Hansson et al. (Sweden) (102)

15 years (peak), 80% between 12 and 16 years 12.7 years mean (1960-1969)

Inconsistent with other data; not recorded at time of origin.

Weight

23% (one-half bilateral initially and one-half later; incidence raised to 41.8% at final study)

(107)

Each major series referred to in Table V noted the peak year of female occurrence to be at age 11 years. The average age of male involvement in our series was 13 years 2 months with other series showing male peaks at 13, 14, and 15 years. The large multinational series of Loder et al. showed the

Data presented as weight related to height. "The excess of abnormally heavy is much more pronounced than the corresponding excesses of tall people."

25% males and 18% females

average age of female involvement to be 12 years and that of males to be 13.5 years (168). This narrow age range can be used effectively to target groups at risk to enhance earlier diagnosis. The youngest patient described with idiopathic SCFE in the literature was a 5 year 9 month old obese female

402

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

with bilateral involvement (138), and one of our male patients was 5 years 7 months old. Milgram and Lyne reviewed epiphysiolysis of the proximal femur in very young children, describing 14 cases in patients 3 years of age or younger (187). There were no instances of slipped capital femoral epiphysis. Ten of the 14 hips were associated with myelomeningocele (4) or the battered child syndrome (6) and the others were associated with major trauma. A pure slip is virtually impossible before 4 years of age because the cartilaginous end of the proximal femur is composed of the femoral head-neck-greater trochanter as a single mass. The average age at initial presentation was the same in the unilateral and bilateral groups. It might have been expected, perhaps, that the bilateral group would have presented at earlier time periods, but this was not borne out by the data. Of further interest is the observation by Hansson et al. that the age of occurrence has decreased gradually throughout the century from a mean of 16 years in males and 12.6 years in girls to 12.7 years in boys and 11.8 years in gifts. Even though the chronological age at diagnosis is relatively specific and narrow for both males and females, Exner has shown that the time of occurrence of a slip is even more closely related to the bone age and the peak of the growth spurt than it is to chronological age (68). The chronological ages of girls at the time of the diagnosis ranged from 10.7 to 16.3 years with a mean of 12.9 years, whereas the bone ages ranged only from 12 to 13.9 years with a mean of 13.1 years. In boys the chronological ages ranged from 12.2 to 17.4 years, whereas the bone age ranges were narrower from 14 to 15.6 years with a mean of 14.5 years. He concluded that it was the skeletal maturation rate rather than chronological age itself that determined the time of occurrence of a slip. As early as the 1930s it was recognized that the onset of symptoms and displacement usually occur during the rapid growth period of adolescence. Loder et al. also studied the narrow window of bone age in children with slipped epiphyses (170). They utilized assessments from pelvic radiographs in 30 children with slips. The chronological age range was 98 months but the skeletal age range was only 50 months. c. Bilateral I n v o l v e m e n t Bilaterality in our series was much higher than indicated from most previous reports. An incidence of 21-32% has been reported in several large reviews from 1957 to 1986, whereas reviews of 10 large series from 1926 to 1948 reported bilaterality from 4.4% to 28%. Loder et al. noted 22% bilaterality (168). The necessity of assessing patients to skeletal maturity is clear as many patients who present unilaterally have been shown to develop opposite side involvement as long as 4 years later. Failure to follow patients to maturity would decrease the amount of bilaterality noted. Our series documents a 40-42% incidence of bilaterality. Two previous groups have adjusted their figures upward to report an approximate 40% incidence of bilaterality on the basis of later and continuing studies. Jerre noted a final incidence of 41.8% after an initial assessment of 22.2% (128, 131), and Klein and colleagues referred to

bilaterality eventually as being 41% following an initial impression of 22% (149, 150). Jensen et al., in a long-term study of 62 patients, eventually found bilateral involvement in 30 of 62 (48%) (127). Presently, many studies have documented the increasing incidence of bilaterality from time of initial presentation to skeletal maturity. Hagglund et al., using the same technique as Billing and Severin on a much larger number of patients, 260, documented bilateral SCFE in 61% at skeletal maturity (99). Schreiber documented a 65% incidence of bilaterality in 100 cases (232). The highest documented incidence was by Billing and Severin, who performed radiographic assessments of the configuration of the proximal femur at maturity on 81 patients who had suffered slipped capital femoral epiphysis and concluded that 80% were bilateral (26). Due to the high resolution and standardization of their technique, they referred to slippage that others looking at regular radiographs either would not note or would consider to be a variation of normal. It is possible that some of these slips would not become of clinical importance, but the findings are of great interest in confirming that the incidence of bilaterality is much higher than realized. Half of the bilateral cases in our series were such at time of the initial diagnosis, whereas the other side became involved later at periods varying from 1 week to 2.5 years. This finding is similar to others. Loder et al. noted 60% simultaneous presentation and 40% sequential (168). In 82% the second slip was diagnosed within 18 months. Wilson et al. noted that 48% presented simultaneously and 52% with the second slip from 1.5 to 48 months after the first (272). The average time between slips in several series is about 1 year, with the ranges similar to that noted previously. The extent of bilateral involvement points to the importance of careful assessment of the contralateral hip at initial presentation and the need for continuing assessment to skeletal maturity. If we set aside the rigorous study of Billing and Severin, slipped capital femoral epiphysis still has a 4 0 60% incidence of bilaterality. In those who have bilateral involvement approximately one-half are bilateral at initial assessment, whereas the other half can develop opposite side involvement anywhere from 1 week to 3.5 years later. The average age at initial presentation does not differ in the unilateral and bilateral groups. The weight percentile distribution at initial presentation does not differ in the unilateral and bilateral groups. A treatment matter relating to bilaterality is the question of whether, in a unilateral slip, the contralateral hip should be pinned prophylactically. Both support for and disagreement with the principle of prophylactic pinning continue to be raised. The advantage of pinning is that a second side slip, which conceivably could be detected late in a moderate or severe stage, is prevented. The disadvantage is that many hips that would have remained normal would undergo operation with a procedure that still has an appreciable complication rate. Jerre et al. estimated that, if bilateral pinning had been performed prophylactically in 61 patients

SECTION II ~ Slipped Capital Femoral Epiphysis

initially treated with unilateral slips, 36 of the operations (59%) would have been unnecessary (128). This also is supported by our study in which 25% of involved hips are bilateral at time of initial presentation with another 25% slipping over the remaining time period to skeletal maturity. With careful assessment periodically during the remaining years of growth and description of concern to the family, virtually all second side slips will be diagnosed in the early or mild phase. We also feel that the dangers of a complication in a hip that is and would remain normal suggest close continuing assessment rather than prophylactic pinning. d. Side Predominance The left side is invariably affected more commonly than the fight. Hansson et al. noted a 2:1 left:right ratio in their large study of 532 Swedish patients (102). Alexander summarized the side incidence from 480 cases in the literature, noting a 1.6:1 left:fight ratio (6). Loder et al. found 1.5:1 left:fight in 1266 unilateral cases (168). Our study documented 1.3:1 left:fight in 144 unilateral cases. e. S y m p t o m Time to Diagnosis The symptom time to diagnosis remains relatively long. Although there is not absolute correlation between the length of time that a person is symptomatic prior to diagnosis and the degree of slippage at diagnosis, nevertheless there is a positive correlation. Cowell noted 67% excellent results with symptoms less than 3 months; 36% excellent results with symptoms more than 6 months, and only 28% excellent results with symptoms greater than 9 months (53). Kulick and Denton noticed increasingly poor results with a greater degree of slippage. With the mildest grade I slip good results were seen in 93% of 74 patients, in grade II good results diminished to 74% in 23 patients, in grade III it was 64% good in 11 patients, and in grade IV only 40% good in 10 patients (159). Aronson et al. also showed progressive worsening of results the greater the degree of slippage (15). In a study of 80 patients, excellent and good results were seen in 86% with mild slippage, 55% with moderate slippage, and only 27% with severe slippage. Conversely, poor results increased with the severity of slip with only 15% in the mild group, 45% in the moderate group, and 46% in the severe group. Hansson et al. also noted a persisting delay in time to diagnosis following the onset of symptoms in more recent decades, with 35-40% with symptoms exceeding 3 months (102). The longer a person remains active on an involved hip, the greater the likelihood of advanced slipping at time of diagnosis. f. Obesity Obesity remains the overwhelming and prime etiologic factor (143, 148, 168, 199). At the time of diagnosis, 47 % of patients in the Children's Hospital, Boston, study were above the 95th percentile in weight. This almost exactly corresponds with the study involving school children in Connecticut and several southwestern states, which noted 49% with weights above the 95th percentile (141, 142). There was no difference in the weight percentile distributions in the unilateral and bilateral groups. It might have been expected that the bilateral group would have had a

403

larger concentration of heavier, obese patients, but this was not borne out by the data. Noble and Hauser commented in 1926 that in adolescent tibia vara (SCFE) "a feature of considerable significance is the almost invariable occurrence of obesity" (199). Loder et al. noted slipped epiphysis to be "overwhelmingly a disease of obesity; 63.2% of the children were in the upper tenth percentile for weight" (168). Epidemiologic features from many studies are summarized in Table V. g. Associated Medical Conditions There are many reports of associations of SCFE (1) with endocrine disease or treatment states including hypothyroidism, craniopharyngioma, panhypopituitarism, and those undergoing growth hormone therapy; (2) a few years after pelvic or abdominal radiation for soft tissue sarcomas; and (3) in those with severe chronic renal disease. Our series documents a 5% incidence of associated diseases in a large group, and other studies also point out a similar percentage in large series from specialized pediatric institutions. (See also Section II.E 1-3.) 4. SCHOOL SCREENING RECOMMENDATIONS BASED ON EPIDEMIOLOGIC DATA

The data presented previously both from Children's Hospital, Boston (236), and from several other series suggest procedures readily adaptable for earlier diagnosis. If screening procedures are developed to include only those greater than the 50th percentile in weight, then one-half of the age group at risk can be eliminated from screening with virtually no risk of missing cases. If the screening cutoff point is the 75th percentile, then assessment of only one-quarter of the age at risk population would still yield early diagnosis of approximately 85% of cases. If those at and above the 90th percentile are assessed, 10% of the age at risk population, the yield would still would be 65%, and if the path of least resistance is taken and only those above the 95th percentile are assessed, that being only 5% of the population at risk, almost 50% of cases would have been provided with early warnings. Calculations of the incidence of SCFE in relation to the population between 7 and 16 years of age have been made in American and Swedish studies with very good correlation. The Swedish study noted that in the age group at risk the annual occurrence was from 2 to 13 cases in 100,000, whereas the Connecticut study noted the incidence to be 8.3 in 100,000 (141). Choice of an age range in males of 11 to 15 years inclusive at the 50th, 75th, 90th, or 95th percentiles as previously described, plus ages 8, 9, 10, and 16 years who are above the 95th percentile coverage, would be excellent. The same would apply for females aged 9-12 years inclusive and above the 95th percentile for ages 8, 13, and 14 years. In reference to the Connecticut data, if one screened all age at risk children above the 50th percentile, then one would expect to pick up 8 cases per 50,000 children. Screening only for those weighing at the 95th percentile and up would yield approximately 4 cases per 5,000 children.

404

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormalities of the Femur

Screening need not be complicated as mechanisms for awareness of the at risk patient already are in place. Public high schools record student weights annually as part of school health examinations. Pediatricians in offices, hospital clinics, and health maintenance organizations record weight as part of any general assessment. By charting the weights on percentile charts those at risk become readily apparent. Depending on the age and weight screening ranges chosen, any child at risk for slipped capital femoral epiphysis could have (1) a letter or discussion with patients and parents concerning symptoms of hip, thigh, or knee discomfort, limp, and external rotation of the lower extremity; (2) immediate hip examination in those who are symptomatic; (3) clinical examinations every 3-4 months in those above the 95th weight percentile; and (4) anteroposterior and lateral hip radiographs as indicated by early symptoms and clinical findings. By concentrating on those who are extremely obese, patient and family warnings supplemented by physical examinations and lateral hip radiographs would provide a level of awareness sufficient to allow for much earlier diagnosis. Philip D. Wilson wrote in 1938 (271) that "the diagnosis of slipping of the upper femoral epiphysis at an early stage, when the displacement is still minimal and not sufficient to cause any permanent functional impairment, permits an entirely different type of treatment from that which is necessary when serious deformity is present. Since early diagnosis depends chiefly upon the general practitioner, there is need of better instruction in the pathological and clinical appearance of this disease." The age and weight ranges of this condition are narrow and consistent, and the registering of a child's weight is a virtually universal occurrence in school nurse and pediatrician offices already. Slipped capital femoral epiphysis represents a condition in which screening techniques are easily applicable and could lead to beneficial results in limiting degenerative hip disease.

I. Diagnostic Imaging Studies Diagnostic imaging studies involve anteroposterior, frog lateral, and true lateral radiographs of the proximal femur. The true or "shoot-through" lateral view at times can show a greater degree of displacement than frog lateral views, which often are oblique radiographs of the head and neck region. CT scanning defines both the slippage of the head and the retroversion of the neck. Ultrasound is most helpful in defining the presence or absence of any effusion, the amount of head-neck displacement, especially if mild, any reduction (usually inadvertent) occurring during the process of treatment, and the remodeling of the anterior femoral neck metaphysis, which is a sign of chronicity greater than 3 weeks. Magnetic resonance imaging is used by some to assess marrow reactivity on either side of the physis as an index of physeal stability. It also assesses fluid in the joint and retroversion, although these two features are shown more easily or more clearly by ultrasonography and CT scanning, re-

spectively. Bone scintigraphy can detect early femoral head avascularity. Arthrography rarely if ever is used at present to assess slips. Guzzanti and Falciglia showed that careful application of plain radiographic techniques produced measurements of slip severity with a high level of concordance with CT measurements (94).

J. Therapy 1. GOALS OF TREATMENT The purpose of treatment is to produce a hip that at skeletal maturity has a femoral head-neck-shaft relationship as close to normal as possible but in association with normal head sphericity, normal articular cartilage, well-vascularized bone in the femoral head, and no clinically significant shortness compared to the opposite side. Treatment in slipped capital femoral epiphysis is designed primarily to stabilize the head on the neck, preventing further slippage by inducing premature growth plate fusion. A second goal for some but not all practitioners is to reposition the head to restore partially or fully the anatomy of the proximal femur in those instances in which displacement has been severe or complete and, in some instances, moderate. In patients with a preslip or mild to moderate slipping, prevention of further slippage appears sufficient. Stabilization of the femoral head on the neck is brought about by causing premature fusion of the proximal femoral capital epiphysis using some form of transphyseal fixation. In severe and complete slips, the stabilization procedure is sometimes performed (1) in association with open reduction and cervical resection or shortening osteotomy designed to restore the position of the femoral head in relation to the acetabulum and the femoral neck and shaft or (2) with compensatory basicervical or intertrochanteric osteotomy, another approach to repositioning that can be done at the same time as, or generally several months after, in situ stabilization. Controversy continues as to what degree of slippage mandates repositioning of the head in relation to the neck, what technique should be used to reposition the head, and when, if ever, repositioning should be done. 2. EVOLUTION OF TREATMENT MODALITIES UP TO THE 1960S: LESSONS LEARNED

a. Surgical Intervention, Closed Manipulation, and Hip Spica Immobilization Virtually all of the treatments in use today were used from the 1890s to the 1930s, albeit with cruder techniques. The first nail for internal fixation of SCFE is attributed to Sturrock in 1894, but due to infection the device had to be removed at 2 days so early effectiveness was not defined (247). The mention of nailing was in an article on traumatic separation of epiphyses, and the proximal femoral epiphyseal separation was either a displaced acute slip or a pure traumatic fracture-separation. The first osteotomy to correct deformity was reported by Keetley in

SECTION II ~ Slipped Capital Femoral Epiphysis

1888; he performed subtrochanteric wedge osteotomy of the femur (139). A later paper by Keetley on coxa vara in 1900 outlined evolving concepts of the disorder, most of which were cases of SCFE, and the surgical approach of valgus osteotomy for correction (140). Of note is the great care he took in outlining his claims to priority for surgical correction by osteotomy and for reporting the first case "in which the nature and seat of the disease [coxa vara] were diagnosed correctly during life and before operation." There was early awareness of the corrective value of osteotomy, and procedures were done in the neck, intertrochanteric region, and shaft followed by positioning of the distal fragment in abduction, internal rotation, and flexion. Helbing illustrated several of the osteotomy procedures described by his contemporaries (106). Whitman was an early advocate of manipulation either with general anesthesia or in traction followed by hip spica immobilization (269). Closed manipulative reduction with the goal of complete reduction still was being recommended by some decades later (175). The tendency to attempt repositioning of the head as though the patient had a displaced fracture was common, but many poor results were recognized early. b. Reviews of Treatment Methods and Results Treatments used as well as their benefits and features were reviewed by Kleinberg and Buchman in 1936 (152). It was felt that conservative treatment could be effective but that it was advisable primarily in the preslipping or very mild slipping stages. Rest with prevention of all weight beating in their opinion would produce excellent results in those instances. The manipulative treatment had certain problems because it was based on an incorrect assumption, namely, that the disorder in reality was an incomplete fracture of the neck of the femur. Reduction by manipulation into extreme abduction and internal rotation was used by Whitman but the results were felt to have been unsatisfactory. The authors pointed out that even in mild slip situations reduction rarely resulted in perfect repositioning. Even when the surgeon considered that the manipulation was gentle, extreme force was placed on the femoral head and neck region causing extensive damage to the circulation to the femoral head. They indicated that "once the reduction is completed the possibilities for re-establishment of the circulation of the femoral head are rather meager." At multiple arthrotomies the head was always firmly fixed to the neck and displacement was possible only by means of osteotome and mallet. The only possible results of manipulations were further crushing of the head and increased damage of the blood supply. Manipulative treatments would be of value only in cases of acute slipping, and even then a manipulative reduction would have to be executed deliberately and with gentleness. For all other forms, the manipulative treatment was felt to be inadvisable because "1) the reduction may be impossible; 2) the reduction is often incomplete and illusory; 3) the circulation of the head is likely to be disturbed as evidenced by late deformity of the head; 4) traumatic arthritis and even ankylosis may

405

ensue; and 5) no change is effected in the pathologic epiphyseal plate." Though recommending operative treatment, they felt that open reduction and instrumentation were inappropriate because there was no provision for the revascularization of the deranged epiphyseal plate tissue. There was potential value of drilling of the epiphyseal growth plate "for the purpose of re-establishing circulation and causing premature ossification of the part." That approach was limited to the preslipping or mildly slipping stages but had been used in several cases with good results. Most of their attention surgically was directed to the correction of moderate to severe deformity and to the establishment of circulation across the physis to produce early fusion between the head and neck. This was best accomplished by an operative resection of the epiphyseal plate and realignment and contact between the cancellous bone of the head and neck of the femur. Open reduction involved a wedge resection of the physis and the adjacent femoral neck, curettage of cartilage from the head region, and placement of the head onto the reshaped neck followed by closure and cast immobilization. They actually dislocated the head and clearly relatively little attention was paid to maintaining vascularity, although comments on the need for it were made. Postoperative treatment was prolonged until revascularization of the head occurred. In summary, many of the principles of current treatment were recognized. Slipped epiphyses were not fractures and could not be treated as such. The treatment should be atraumatic so as to correct deformity and establish circulation between the head and neck of the femur. Treatment of the preslipping stage was by bed rest or by brace, although transphyseal drilling of the head and neck could be effective. In a report on 44 cases published in 1945, Moore stressed the need to avoid trauma to the epiphysis by operation or manipulation as much as possible and strongly felt that the risk of damage to the blood supply to the epiphysis following surgery outweighed the advantage of earlier mobilization (190). Virtually all of his patients were treated conservatively either with bed rest alone or with hip spica immobilization in which the limb was positioned in wide abduction and internal rotation but without efforts at specific reduction of deformity. The limb was protected until transphyseal ossification of the growth plate particularly in its central regions was noted radiographically. AVN occurred in only 1 of 29 patients with minimal displacement and 4 of 23 patients with moderate displacement, but was seen in 2 of 3 epiphyses following complete separation. Moore felt that his results were much better than those being reported following manipulative or surgical reposition of the epiphysis. By the 1940s, there was good awareness of the fact that the results of treatment in many cases of SCFE were poor, leading to arthritis in early adult life. Howorth (116, 119) and Klein et al. (149, 151) also summarized the attempts of the previous decades, recognizing in particular that correction of the deformity by forceful closed manipulation followed by fixation in plaster casts led to poor results, as did subsequent

406

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

A

11

j.~< !i,i

~ -:~ ~-~

-\

F I G U R E 9 (A, B) The surgical technique for the transphyseal bone graft or bone pegging operation is outlined. After curettage, three cortical grafts were placed from neck to head, and the bone block was reinserted. [Reprinted from (111), with permission.]

attempts at open reduction followed by fixation in plaster. Use of open reduction and osteotomy at the neck and peritrochanteric regions were resorted to frequently. c. Transphyseal Drilling Simple transphyseal drilling was shown to induce premature growth plate fusion in cases of slipped capital femoral epiphysis. The method was used in 6 cases by Pomeranz and Sloane in 1935 (213). Mayer reported on the procedure in 20 patients diagnosed early when slipping was extremely slight (182). Under radiographic control a drill bit was placed through the greater trochanter and neck across the physis and into the femoral head. The plate was penetrated at 7 or 8 different points. The limb then was immobilized in a hip spica cast for approximately 10 weeks until radiographic transphyseal fusion was identified. The average period was 3 months. Mayer indicated that 20 cases had been treated with normal function of the hip and an excellent gait. A similar approach was reported by Kiaer (145) and by Mathiesen (179). Mathieson reported that between 1939 and 1954 a total of 36 cases of minimal slipped epiphysis less than 1.5 cm had been treated by transphyseal drilling. Three to 4 drill holes were placed through the physis into the head following which the patient was treated with either traction or hip spica. Most patients had fused the physis within 3 months. There were no instances of avascular necrosis and results were considered excellent. d. Transphyseal Bone Graft Howorth, in an article with Ferguson in 1931, first described the open transphyseal bone graft to enhance head-neck fusion in slipped capital femoral epiphysis (73). Howorth was the chief proponent of the transphyseal bone graft or bone-pegging operation. He performed several hundred of these procedures for mild and moderate slips. No open reduction or wedge osteotomy of the neck was done. A small entry point was made in the anterior surface of the femoral neck, after which a drill or curette was passed up the neck and across the physis cen-

trally into the femoral head. Three small bone grafts were then cut from the ilium, which were largely cortical and averaged 1.25 inches long and 3~6 in. wide. The bone pegs were then placed into the hole in the neck, through the physis, and into the bone of the secondary ossification center. After a period of several days in bed until the soft tissues healed, the patient then proceeded either to chair or crutches without weight bearing. Transphyseal fusion generally occurred within 12 weeks. In a 1966 report on 200 bone-pegging operations, Howorth claimed near universal excellent results with no documented episodes of chondrolysis or avascular necrosis. He felt the operation to be vastly superior to nailing regardless of the method of fixation (120). He also claimed that good or excellent results had been reported for the bonepegging operation done by others in 151 out of 152 cases. Heyman and Herndon provided an excellent review of the literature in relation to experiences with pinning in the early phases (111). As a response to the less than perfect results with other treatments, they reported that an iliac crest cancellous bone graft placed from the femoral neck across the physis into the center of the femoral head allowed for rapid fusion, thus eliminating the concern about further slippage (Figs. 9A and 9B). Reduction was not done. They commented on 19 operations noting the average time to fusion of the epiphysis as defined by X ray to be 2.3 months. The clinical and radiologic results were excellent in all cases. There was no postoperative plaster immobilization, and patients were ambulatory on crutches within a few weeks of surgery. In no case did they note any increase in displacement, and there was no evidence of chondrolysis, avascular necrosis, or other degenerative changes. They exposed the anterior surface of the neck of the femur and removed a rectangularshaped section of bone cortex about 3 cm long and 1 cm wide along the long axis of the neck of the femur. They then used a drill or reamer 1 cm wide directed across the growth

SECTION II ~ Slipped Capital Femoral Epiphysis

407

FIGURE 10 (A, B) The osteoplastyprocedureremovesthe bonyobstacleto flexionand abduction caused by the prominent superolateral portion of the femoral neck rather than attempting to reposition the head by more complicatedosteotomyprocedures.

plate and into the center of the femoral head. Small pieces of cancellous bone obtained from the iliac crest were then impacted into the depths of the defect, thus enhancing fusion. e. Internal Fixation without Reduction and Pinning in Situ Utilization of the three-flanged Smith-Peterson nail was beginning to give improved results with in situ stabilization with postoperative rehabilitation simplified by the fact that there no longer was any need for hip spica immobilization (149, 151, 271). Results with the large Smith-Peterson nail, however, though improved from previous reports, were far from perfect and many observers reported that, although the device stabilized the head on the neck, it did little to increase or enhance the rate of fusion. Fusion still was quite slow, appearing to take anywhere from 4 to 18 months, whereas in many instances fusion never occurred and the femoral head grew off the tip of the nail. Wilson commented on the value of pinning in situ in 1938 (271). He stressed the importance of early diagnosis and early treatment and also indicated that his earlier experience using weight beating caliper braces and ambulatory plaster spicas as protective devices was highly ineffective. Recumbent treatment in plaster hip spica also was criticized as being uncertain and necessary until fusion was obtained, which could be 1 year or more. Even after many months of immobilization, subsequent slippage was noted along with major muscle and bone atrophy. Even at this early date, there was controversy between those using the drilling method from the femoral neck through the epiphyseal plate into the head with insertion of small bone grafts and those who performed metallic stabilization. Wilson indicated that the open epiphysiodesis method "requires arthrotomy and seems unnecessarily complicated." He favored insertion of the Smith-Peterson nail by transtrochanteric pinning under radiographic control without arthrotomy. Reduction was not attempted. He clearly documented the value of the nail in that it stabilized the epiphysis, preventing further displacement, and penetrated the cartilaginous growth plate, allowing the ingrowth of reparative elements and thus promoting early fusion of the epiphy-

sis. Obliteration of the growth plate occurred in 4-6 months. Bilaterality was common. Serious shortening of the involved extremity did not develop following pinning because of the age of the patient, the fact that most growth had occurred already, and the fact that relatively little femoral growth was at the proximal end. f. Cervical Osteoplasty Heyman et al. subsequently described a conservative operative procedure for severe slips of the femoral capital epiphysis in which transphyseal stabilization had occurred, but the displacement led to severely disabling limitation of motion (113). Rather than proceeding with open reduction or cervical osteotomy, they simply shaved the prominent superolateral portions of the neck to enhance the range of motion, a procedure referred to as an osteoplasty (Figs. 10A and 10B). The procedure was recommended for extreme limitation of flexion as well as extensive loss of abduction and intemal rotation. At operation, they noted severe downward and backward displacement of the head of the femur with solid union. The obstruction to motion was a large bony prominence at the anterosuperior aspect of the neck of the femur at its junction with the displaced epiphysis. This was seen to impinge against the rim of the acetabulum. Rather than performing an osteotomy with its inherent risk to the vascularity, the bony prominence was removed with an osteotome. This served to improve markedly the motion of the hip without affecting the position of the head and neck. In many cases, they noted that an almost complete range of motion was obtained by removing the bony prominence, which had impinged against the acetabular rim. The patients were rehabilitated with early postoperative motion and weight beating with crutches. It is the proximal portion of the femoral neck that appears as the angular or rounded ridge of bone at the anterosuperior aspect. The osteoplastic operation consists of removing the angular or rounded prominence of bone at the neck of the femur where it is adjacent to the femoral head. If complete fusion had not occurred, then bone graft epiphysiodesis was performed at the same sitting.

408

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o]r the Femur

Heyman et al. subsequently reviewed 28 patients having the procedure, which was done on those with severe slippage and external rotation of the femur with marked limitation of internal rotation and abduction. Flexion of the hip also was limited and was most free only in extremes of external rotation. The coxa vara and shortening were unaffected by the operation, although improvements in the range of motion, in particular involving increased abduction and internal rotation, were felt to be highly beneficial. Pain was eliminated. The gait improved markedly with the knee directed straight forward and the residual limp slight. There was no evidence of avascular necrosis or early osteoarthritis. The procedure had been described several decades previously by Poland, by Whitman, and by Vulpius and Stoffel, but its relatively wide adoption was due to the work of Heyman et al. (113, 119). A study by Wilson et al. on 300 hips treated at the Hospital for Special Surgery in New York City from 1936 to 1960 summarized well the positive and negative features of the clinical approaches at that time (272). The primary apparatus used for pinning in situ was the Smith-Peterson nail. This study began to focus on the problems with this nail, in particular instances of distraction of the epiphysis with insertion of the device. This was seen in 12 of 222 fixations in situ. This complication also was noted by Wiberg, leading him and others to use thinner nails (270). There were markedly better results with in situ fixation, however, than in those instances with moderate or severe deformity in which efforts to correct the deformity were made. In hips treated with pinning in situ, most of which were dsplaced only slightly, 13% were rated radiographically fair or poor and 9% clinically fair or poor. However, when correction of the deformity followed by fixation was used, and most of these were moderate to severe in extent, the radiographic fair and poor results were 45% and clinical fair and poor 38%. Each of the complications that came to characterize works of the later decades were recognized, including pin penetration, avascular necrosis, acute cartilage necrosis, and osteoarthritis. When hips were treated other than by fixation in situ, the operative approach used involved wedge osteotomy of the femoral neck and either open or closed reduction. Wedge osteotomy of the femoral neck in this era was highly problematic, with the combined results from many published end result studies being good in 164, fair in 15, and poor in 72. The incidence of poor results, which involved primarily osteonecrosis or severe arthritis, was 29% in the combined reported series and 23% in their own series. Wiberg in particular noted a 27% incidence of AVN (23 of 84 cases) from his institution (270). Correction of marked displacement by either open or closed reduction also resulted in a large percentage of poor results, with a report by Moore on open reduction of the growth plate in 87 hips with marked displacement showing the incidence of poor results at 50% (190). Closed reduction either failed to correct the deformity or if successful was followed by a high incidence of necrosis. This was emphasized by Jerre, who in 117 hips treated by

closed manipulation noted an incidence of AVN of 41% in 24 hips that were actually shown to have the displacement corrected and 7.5% in which no reduction followed (131). The authors concluded that early diagnosis when the slipping was minimal led to the best results with treatment by fixation in situ; when efforts were made to correct deformity, the complication level rose dramatically. Excellent summations of treatments at that time also were provided by Durbin (64), Hall (100), Howorth (116, 119), and Morscher (193). Manipulation, spica immobilization, and open reduction without cervical shortening all were recognized to have a high incidence of avascular necrosis and poor results. The review by Howorth (119) is the most comprehensive by far referring to virtually all papers published on treatment from the 1890s to 1960. Morscher reviewed the earlier European approaches to both nailing and transphyseal bone grafting for slipped epiphysis. He reported generally good to excellent results with these techniques in mild to moderate slips, reserving cervical wedge osteotomy only for the most severe cases (193). Virtually all patients with a mild to moderate slip are now treated by pinning in situ. It is essential to differentiate whether a displaced femoral head of moderate to severe to complete magnitude represents an acute, acute on chronic, or chronic slip. Inherent in this assessment also is consideration using the stable-unstable classification. We define the truly acute slip with no previous symptoms as essentially an acute fracture-separation in an epidemiologically predisposed patient. Its treatment varies from the chronic slip (we will discuss the acute on chronic separately), but it is important to recognize that all references to an acute slip in the literature do not use strictly similar criteria of definition. The question of whether displacement of the femoral head, be it acute or chronic, should be reduced as well as the timing of any reduction remains controversial in many centers. 3. CURRENT APPROACHES TO TREATMENT OF THE

ACUTE SLIPPED CAPITAL FEMORAL EPIPHYSIS Acute slips are far less frequent than chronic slips. Loder et al., in their large multinational study of 1993 slips, noted 85% to be chronic (more than 3 weeks of symptoms) and 15% acute (168). Some surgeons attempt "gentle" closed reduction in acute instances (in the temporal classification group) or in unstable cases (in the stable-unstable classification of Loder et al.) with moderate to severe deformity. The problem with attempting reduction in these instances is that it is not possible to monitor vessel function or to know how much reduction force is safe and how much is dangerous in relation to the vascularity. Two types of closed reduction can be done. One approach is to place the patient on bed rest with skin traction with medial or internal rotation. Once position has improved, pinning in situ is performed. A second approach is to perform "gentle" closed reduction by manipulation under general anesthesia followed by pinning in situ. Of great importance in the assessment of results is

SECTION II ~ Slipped Capital Femoral Epiphysis

the need to recognize that spontaneous partial or full reduction of the acute slip can occur in association with the unavoidable moving of a patient between the time of radiologic diagnosis and the actual performance of the surgical pinning operation. Most series define acute slippage as occurring with some form of trauma, discomfort of less than 3 weeks, and instability characterized by the inability to bear weight. In most there are also reports of mild symptoms of several weeks duration. Rather than calling this acute on chronic, some groups, like Peterson et al., focus on the acute pain and usually moderate to severe displacement and suggest that "duration of prodromal symptoms is an unreliable subjective indicator of acute SCFE" (210). In their series of 91 acute slips over a 40-year period, the incidence of AVN was 14% (13/91). They felt that manipulative reduction under general anesthesia did not increase the risk of AVN. Results were better if reduction was performed within 24 hr of presentation (AVN, 7%) compared with those reduced after 24 hr (AVN, 20%). A 14% AVN rate (5/35) also was reported by Casey et al., although they felt that reduction would be safer if done by skin traction with internal rotation followed by in situ fixation (45). They treated 35 acutely slipped epiphyses with manipulation only or with manipulation and traction (followed by pinning in situ), but the time from onset to manipulation ranged from 1 to 34 days. Fahey and O'Brien reviewed 75 cases of acute slip from the literature (70). Closed reduction and cast (40 cases) led to 47.5% satisfactory results, closed reduction and internal fixation (23) led to 65.2% satisfactory results, and open reduction (12) led to 83.3% satisfactory results. They had satisfactory results in all 9 with closed manipulative reduction under general anesthesia with in situ pinning. The relative safety of closed manipulative reduction and epiphysiodesis or pin fixation in 50 cases of acute slipped capital femoral epiphysis was shown by Aadalen et al. (1). The patients were placed into four groups based on treatment. The best results were in patients treated by manipulative reduction within 24 hr of the onset of acute symptoms, although this only involved 8 patients with no episodes of AVN. There was a 15% rate of AVN (7 of 47) in which treatment was by manipulative reduction followed by open epiphysiodesis, pin fixation, or both. In various additional subgroupings, no significant difference between epiphyseal stabilization by pin fixation or epiphysiodesis was noted. Hall, in a study of 173 hips treated by many methods, concluded that manipulation was a relatively safe and effective method of reducing deformity in patients seen soon after an acute episode but that it should be reserved for this acute group only (100). Reduction was performed not so much by straight longitudinal traction but rather by the medial rotation component. Aronson and Loder take a cautious, nonmanipulative approach even to the acute or unstable form of slipped epiphysis (14). Their priorities are to avoid avascular necrosis, avoid chondrolysis, and prevent further slip. In relation to

409

the correction of deformity, they recognize that it is associated with a high incidence of complications "so manipulative reduction under anesthesia or an acute corrective osteotomy is not recommended." They recommend preoperative bed rest to decrease the synovitis and intra-articular effusion followed by operative stabilization with a single pin, with careful positioning of the patient on the operating room table with no attempt to perform a manipulative reduction. Even with great care incidental reduction could occur on occasion, but they stress that it was never actively sought. No preoperative traction was used because that alone could increase the intra-articular pressure or damage the posterior vasculature by effectively bringing about a relatively forceful reduction. Stanitski also noted incidental reduction in approximately one-half of a small series of his own cases (246). He also stressed the risks of formally attempting closed reduction in acute slips. In summary, the acute slip, strictly defined, is the only variant of slipped capital femoral epiphysis deemed capable of reasonably safe management by attempted gentle reduction prior to pinning in situ within 24 hr of presentation. The risk of AVN still appears to be in the 15% range, although much of this may be caused by the initial injury rather than the treatment. Even in this subgroup, many opt strictly for pinning in situ. 4. CURRENT APPROACHES TO TREATMENT OF THE CHRONIC SLIPPED CAPITAL FEMORAL EPIPHYSIS

Mild to moderate slippage is treated generally with pinning in situ. If the deformity is severe with posterior slip of the head onto the back of the neck in a chronic state, pinning in situ, though preventing further displacement and alleviating pain, cannot return the head-neck-shaft relationship to a normal position. These patients then would continue with considerable shortening, external rotation positioning of the lower extremity, and a Trendelenburg gait. Attention has been directed in some instances of moderate and severe slippage to correcting the head-neck relationship with the acetabulum and shaft. The major danger in this approach is the possible causation of femoral head avascular necrosis and an increased risk of chondrolysis. The slippage has occurred gradually and the periosteum and vessels on the posterior medial concavity of the head-neck junction are shortened. In addition, newly synthesized fibro-osseous tissue also deposited in the concave regions risks being torn by reduction efforts, closed or open, further damaging the femoral head blood supply. There are several methods used to reposition the femoral head in a chronic slip, and differing levels and times at which correction can be done. Reduction of femoral head displacement can be by closed or open means. Excellent reviews of overall management have been written by Crawford (54, 55). a. Ways o f Inducing Growth Plate Fusion Prematurely in Situ to Prevent Further Slippage with Chronic or Stable Lesions Transphyseal Pinning in Situ, without Reduction:

410

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

FIGURE 11 Slippedcapital femoral epiphysis treated with pinning in situ. Parts (A-D) show an excellent result with pinning in situ with three Knowlespins. Slight medial head displacement is noted on the anteroposteriorpreoperative film. (A) The superior neck is in line with the lateral border of the femoral head bone rather than overlapping. (B) The lateral view shows slight widening of the physis and slight posterior displacement of the head relative to the neck. Pinning in situ led to rapid transphyseal fusion and excellent appearance at skeletal maturity in anteroposterior (C) and lateral projections (D).

The most common approach to treatment at present for stable or chronic slips is to stabilize the displacement in situ without attempts at reduction (Figs. 11 and 12). This prevents further slippage of the head-neck relationship on a mechanical basis initially due to the stabilizing effects of the fixation and shortly afterward due to the induction of premature growth plate fusion. It is the induction of growth plate fusion that remains the primary goal of therapy. Perforation of the epiphyseal growth plate one or multiple times by transphyseal drilling in the course of pinning allows the epiphyseal circulation with its osteoprogenitor cells, to relate intimately to the metaphyseal circulation with its osteoprogenitor cells forming a series of small bone bridges that ultimately lead to full fusion. Physeal damage also is caused by the slippage itself and the continued weight bearing prior

to diagnosis such that the plate already is damaged when diagnosis is made. The greater the degree of slippage, the longer the time to diagnosis, and the more unstable the hip, the greater the growth plate damage will be. Once fusion has occurred, further slippage is prevented. Over 50 years ago, the Smith-Peterson nail was used to treat this condition. This form of intervention had been developed for femoral neck and intertrochanteric fractures in adults. The technique soon was abandoned in children because the extreme hardness of the epiphyseal and femoral neck bone and the relatively tenuous attachment of the head to the neck in slipped capital femoral epiphysis were such that the large, bulky nail itself frequently furthered the displacement of the head, often with causation of AVN. Shortly thereafter, thinner screws were used to affix the head to the

SECTION II 9 Slipped Capital Femoral Epiphysis

F I G U R E 12 (A-D) These figures demonstrate how the entry point for the single screw must be progressively more proximal and more anterior the greater the degree of posterior displacement because the tip of the pin should be at or close to the center of the head in both projections. The arrow illustrates the approximate point of entry in parts (A) and (B) (different patients). On the frog lateral views, displacement is much greater in part (D) [lateral view of (B)] than in part (C) [lateral view of (A)]. (E) Drawing of CT scan shows the position of posterior displacement of the head relative to the neck in a severe slip. Anterior part is at the top with posterior at bottom.

411

412

C H A P T E R 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

F I G U R E 12 (continued) (F) Anteroposterior (Fi) and lateral (Fii) radiographs of a severe slip show Ace screw position following pinning in situ. A series of CT images (Fiii-Fviii) shows the progression of pin tip from the anterolateral entry point on the neck (Fiii) to the eventual central position in the head (Fviii). Note head displacement relative to neck (Fv-Fvi) compared to the normal side. The tip of the pin enters physis in part (Fv) and the femoral head in (Fvi).

SECTION il ~ Slipped Capital Femoral Epiphysis

413

FIGURE 13 Pin protrusion through the femoral head articular cartilage can escape detection by standard radiographic views. Anteroposterior and lateral radiographs at 90~relationship to each other (arrows) in this experimentalillustration would show the tip of the pin to project as appearing to be within the bone. [Reprinted from Nguyen and Morrissy (1990), J. Pediatr. Orthop. 10:341346, 9 LippincottWilliams& Wilkins, with permission.] neck. At Children's Hospital, Boston, a single Wood screw was the treatment of choice for several decades. The patients were protected on crutches for as long as 6 months. The desired central placement of the Wood screw in both anteroposterior and lateral radiographic projections frequently required the passing of guide wires and drill holes through the growth plate several times to assure appropriate positioning. The combination of the single screw, rest on crutches, and multiple perforations of the physis during the process of positioning the screw all enhanced premature fusion. Multiple transfixing screws then began to be used in many centers to enhance the mechanical stabilizing effect, although this was a clinically intuitive assumption (200). The multiple pins led to some major complications recognized only after a several-year delay. Steinmann, Hagie, or Knowles pins provided an effective mechanical stabilization, shortening the period of crutch use and still allowing for premature fusion. Many types of transphyseal fixation screws have been used over the past few decades depending on time and country, but the treatment principle remains the same. A single screw currently is favored again in many centers for reasons listed next.

Danger of Pin Penetration through Articular Cartilage (Walters and Simon): Waiters and Simon performed a landmark study when their assessment of results following pinning in situ of slipped caoital femoral epiphysis led to the observation that many patients suffered from unrecognized pin penetration following treatment (259). They noted that the fixation nail or screw position may appear on anteroposterior and lateral X-ray evaluation to be entirely within the bone of the femoral head, but in reality it might be protrud-

ing through it into or beyond the articular cartilage because radiographs postsurgery were unable in every instance to evaluate adequately the position of the tip of the pin (Fig. 13). Waiters and Simon pointed out that, if the tip of the internal fixation device relative to the X-ray beam did not lie in one of the planes of assessment, it might well be through the cartilage and in the joint, whereas its projection in the plane of the X-ray film showed it to be within the femoral head bone. The accuracy of this observation was confirmed by producing experimental models on cadaver femurs in which some degree of protrusion of the pin through the head was obtained and X-rays in varying positions were shown not to reveal this. Clinical study then was performed retrospectively on 102 patients treated at Children's Hospital, Boston, between March 1971 and October 1977. The treatment was by a single screw in two with the others treated by multiple Knowles or Hagie pins. Walter and Simon's findings indicate how dramatically the concept of pinning in situ has changed since the 1970s. In 90% of the patients, evaluation of anteroposterior lateral and frog lateral radiographs showed no visual evidence of any pin protrusion beyond the bony surface of the femoral head, whereas in 10% a single pin, usually noted on a single view, actually was found to be at the bony surface or protruding 1-2 mm beyond the head into the joint. Pins were frequently left in place even when they were shown on clinical radiographs to protrude if it was felt that they were not directed onto a "weight-beating surface" and if they were asymptomatic. When the "true distance" of pin location was calculated, only 40 of 102 patients could be classified in category I in which all pins were within the femoral

414

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

head. A further 40 patients (40%) were found to be in category II in which maximum pin protrusion beyond the bony femoral head was 5 mm or less. Twenty-two of the 102 patients were in group III in which maximum pin protrusion beyond the femoral head was more than 5 mm. When postoperative films were assessed, no patients in group I had evidence of joint cartilage damage. Sixty-eight percent of patients in group II showed some evidence of chondrolysis or subchondral bone changes; all patients (100%) in group III showed some evidence of chondrolysis or bone changes. All patients in this study had only mild or moderate deformity initially. Waiters and Simon pointed out that, from a theoretical standpoint, devices close to the surface of the head may protrude through it and may not appear to be outside it on either anteroposterior, true lateral, or frog lateral radiographs. Their model in vitro demonstrated that unrecognized pin protrusions could occur and not be detected by standard radiographic projections. They established a template used to calculate the safety limits of pin placement in relation to proximity to the surface of the head. The closer to the periphery of the head the pin tip was positioned on either anteroposterior or lateral projections, the less the margin of safety. Waiters and Simon had identified a major problem with pinning in a retrospective review, noting that many pins had been passed beyond the subchondral bone plate of the femoral head into and frequently through the femoral head cartilage, causing damage to both femoral and acetabular hip joint cartilages. The deep position of the pins either was not recognized or, if recognized, was not felt to be of concern. Independent confirmation of this finding soon followed from other centers. Biplanar radiographic views can fail to detect articular cartilage penetration due to geometric configuration of the head, but measures to minimize or eliminate misinterpretation have been suggested. The increased awareness should minimize one of the complications of treatment, cartilage necrosis or chondrolysis, although not all cartilage necrosis is due to pin penetration. Shortly after the work by Waiters and Simon, retrospective reviews of pinning in situ from other centers confirmed the high incidence of unsuspected pin penetration. Lehman et al. reviewed 63 hips that were pinned as treatment for SCFE and noted a 36.8% incidence of pin penetration not previously recognized (165). Bennet et al. assessed 148 hips using a three-dimensional mathematical model of the femoral head and pin position and noted pin protrusion in 41 hips, an incidence of 27.7% (23). Of these 22 had protrusion sufficiently great to have entered the joint space, with the others showing the tip of the metallic implant to be in articular cartilage. These studies confirmed that standard radiographs taken at right angles to each other often missed evidence of pin protrusion. Models showing this occurrence then were generated by several groups (Figs. 13 and 14). Swiontkowski reviewed 66 pinning procedures and noted a 30% incidence of pin penetration (249). These three studies, published shortly after the report by Waiters

and Simon and thus with surgery done prior to the time of awareness of their findings, confirmed the high incidence of pin-related problems. A retrospective study by GonzalezMoran et al. showed frequent pin penetration when multiple Kirschner wires (1-6, average number 3.8) were used for fixation; the rate was 34% (40 of 188 wires used) (89). Continuation of Physeal Growth Following Pin Insertion: An additional problem with pinning was continuation of growth following insertion of the pin. Many instances of this resulted in no problems, but in others increased slippage occurred due to loss of the protective effect on the one hand and failure to induce physeal fusion on the other. Laplaza and Burke performed a study in 71 hips that had been treated with Steinmann pins, Knowles pins, or cannulated screws (164). Evidence of continuing epiphyseal growth was seen in 29% of hips treated with a Steinmann pin, 18% of hips treated with Knowles pins, but in no hips with treatment by cannulated screws. They recommended the use of one cannulated screw in the treatment of mild and moderate slipped capital femoral epiphysis. No cases of chondrolysis or AVN were noted in the entire series. Progression of the slippage did not happen in any of the hips, although in some repositioning of the pins was performed. Time to Physeal Fusion Following Pinning: There is relatively rapid fusion of involved physes following pinning, which shows the effect of transphyseal perforation as well as the fact that the physes undergoing displacement already are damaged and predisposed to early fusion. Gonzales-Moran et al. showed a mean time to closure of 7.86 months in 31 hips treated with Kirschner wire fixation and 7.12 months in 31 hips treated with single cannulated screws (89). Although closure time is variable, many papers comment on the 6-9 month range. Difficulties with Removal of Pins after Physeal Fusion: An additional problem with pinning in situ involves difficulties and complications related to removal of the pins once physeal fusion has occurred. The problems are (1) inability to remove the pin because of breakage with the deeper part becoming relatively inaccessible, (2) removal of the pin with such great difficulty that excessive bone must be removed and the patient requires lengthy protection with crutches, and (3) fracture through the lateral pin entrance point. These problems are increased when multiple pins are used. Swiontkowski documented difficulty with pin removal (249). In 11 of 18 cases in which pin removal was performed, the removal took longer with larger blood loss than the original procedure. Pin fracture or sheafing of the ends occurred in 5 of these 8 cases requiting the procedure to be halted. Large amounts of lateral cortical bone removed in 3 cases required extra protection postoperatively. Studies on the Number of Pins Needed for Effective Stabilization: Studies have been performed in an effort to provide some documentation of the clinical impression that a single screw provided adequate treatment for a slipped capital epiphysis as long as it was placed centrally in anteroposterior

SECTION II ~ Slipped Capital Femoral Epiphysis

and lateral projections. Chung and Hirahata characterized the biomechanical features of multiple pin fixation (48). Kruger et al. demonstrated experimentally in dogs that single pin fixation was 83% as strong and 78% as stiff as the normal femoral physis but then proceeded to recommend the use of two pins (159). Subsequent investigations, however, supported the use of single compared with a double screw fixation treatment. Biomechanical analysis by Karol et al. was done following the creation of SCFE in bovine femurs (137). One side was repaired with a single screw and the other with two screws. Specimens were reloaded to failure and double pin fixation yielded only a 33% increase in the stiffness compared with single pin fixation. The stiffness, however, of neither double nor single screw fixation approximated that of the intact physis. Single screw fixation was recommended because the small gains in stiffness by the second screw did not offset the risk of complications. A similar study was performed by Kibiloski et al. in 12 pairs of bovine femora with single pinning on one side and double on the other (146). The specimens then were subjected to physiological shear loads across the epiphysis. The rate of creep was decreased by 23% with double screws compared to single screws at slow walking and 30% at fast walking. The results were not statistically significant, and the authors again recommended single screw fixation because the small gains in resistance to cyclic creep at physiological loading were not statistically different and did not offset the complications to be expected with multiple screws. Single screw fixation for acute and acute on chronic SCFE has been reported in 21 hips treated from 1990 to 1993 by Goodman et al. (90). All hips were fixed with a single cannulated screw with no attempt made at reduction. Results were excellent with no cases of AVN or chondrolysis. There was no loss of position during the healing phase and physeal closure occurred at a mean of 9.6 months. It was concluded that single screw fixation was adequate for uncomplicated acute and acute on chronic SCFE. In this series, 9 of the hips were defined as acute and 12 acute on chronic. An acute slip was defined as one in which there was a sudden onset of usually severe symptoms with less than 3 weeks of symptom duration. There were no radiographic signs of remodeling or new bone formation at the epiphyseal-metaphysealjunction. The acute on chronic designation was made when there was a sudden worsening of symptoms with hip pain longer than 3 weeks and no radiographic signs of recent remodeling or new bone formation. Single percutaneous pin fixation for chronic SCFE also has been reported as a valid technique by Samuelson and Olney (227). In a review of 24 chronic slips treated with a single Knowles pin, results were excellent with all patients experiencing complete closure of the growth plate within 12 months and no evidence of AVN, chondrolysis, slip progression, pin penetration, hardware failure, or intertrochanteric fracture. The first 7 patients in this series were treated with two pins, whereas the subsequent 17 were treated with single

415

pins without problems. The single pin was placed directly at the center of the femoral head in all radiographic projections. Ward et al. reported on 53 hips treated with a single screw (261). After a mean follow-up of 32 months, 92% demonstrated physeal fusion with no cases of chondrolysis or AVN reported. They also found that single pin placement provided excellent stability in each of 5 hips with acute or acute on chronic slips. Aronson and Carlson reported excellent or good results in 36 of 38 hips with mild slips, 10 of 11 hips with moderate slips, and 8 of 9 hips with severe slips with single screw fixation (16). AVN developed in only 1 patient with no evidence of chondrolysis. There was loss of position after single screw treatment in only 1 of 8 patients with acute slips. Technique f o r Percutaneous Fixation: Griffith pointed out that the slipped epiphysis was positioned posteriorly directly behind the neck such that it was important in performing internal fixation to introduce the pin through the anterolateral or even the anterior aspect of the proximal end of the femur so that it could be directed posteriorly (93). Colton also illustrated this point (52). Morrissy clarified use of the single pin technique for percutaneous fixation (191, 192). The pins must be somewhat more anterior in entrance position than most appreciate such that a direct lateral approach can lead to difficulty with central positioning of the pin in the femoral head. The technique, which utilizes the cannulated screw, points to the need for a more anterior point of entry into the shaft to allow for direction of the pin into the central aspect of the femoral head in its posteriorly displaced position. Morrissy emphasized the importance of radiographic control for accuracy of pin placement. It is essential for the pin to be in the central axis of the femoral head and that it not penetrate the joint. One pin or screw is recommended and there is only one correct location for it: the central axis of the femoral head (Figs. 12A-12D and 12F). The operative procedure is done on a fracture table with the affected leg abducted 10-15 ~ and internally rotated without force as far as it will go to bring the femoral neck as nearly parallel to the floor as possible. Fluoroscopic image intensification with a free range of motion is essential. A single screw now is recognized by most as adequate and appropriate for the chronic slip. Evidence continues to increase that a single screw also is effective for the acute and acute on chronic slip. If two screws are to be used in the latter situation, the first should be in the central axis of the femoral head with the second below it. Morrissy indicates that the second screw should be at least 8 mm from the subchondral bone to prevent penetration, which is undetected radiographically. The screw in the central axis on all projections can go 2 mm from the subchondral bone on the true lateral. The more anterior entry point allows central perpendicular fixation in relation to the growth plate axis. Nguyen and Morrissy stressed the importance of the anterior femoral neck as the starting point for the in situ fixation device, with the exact location depending on the amount

416

CHAPTER 5 " Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

FIGURE 14 Theseillustrations showthe importanceof appropriatepin placement. Althoughit is important for the pin to be placed centrally within the head on both AP and lateral projections, it is also important that it not protrude through the posteriorcortex of the neck where the vascularityis most prominent. [Reprinted from Nguyen and Morrissy(1990),J. Pediatr.Orthop. 10:341-346, 9 LippincottWilliams& Wilkins,withpermission.] of slippage (197). The greater the posterior slippage, the more anterior and proximal the pin entry point must be. The entry point is on the femoral neck rather than the lateral cortex. CT scans are particularly effective in showing the need for altering the position of entry with the changing degrees of posterior displacement. Another valuable concept in determining the point of placement of the pin is that not only should the pin be in the central axis of the femoral head but it also should pass at fight angles to the physis. It is extremely important not to place the pin into the superior and lateral segment of the femoral head in which the lateral epiphyseal arteries are most prominent. It also is important not to have the entry site on the lateral cortex of the femur. The need to have the pin in the central axis of the femoral head and perpendicular to the epiphyseal plate mandates that the starting point for pin insertion would vary according to the degree of slipping. The more severe the slip, the further anterior the starting point. For those used to pinning a hip in an adult, a meaningful change in orientation is needed in relation to the slipped capital femoral epiphysis, and this change in orientation is greater with a greater degree of posterior slippage. Basically the pinning is done from anterior to posterior because it is essential both to attain position in the central axis of the femoral head and to do so by passing at right angles to the epiphyseal line. If the pin winds up in the superolateral aspect of the femoral head, the dangers of pin penetration are greater as are the dangers of damage to the lateral epiphyseal artery. If, on the other hand, the screw passes out of the posterior surface of the neck and then into the head, the danger of damage to the retinacular vessels on the posterior lateral aspect of the neck also is great (Fig. 14).

Complications of treatment often can be worse than the natural history of the disorder left untreated. Howorth (116, 199) and Hansson et al. (102), among others, have noted that avascular necrosis does not occur in the untreated hip regardless of the degree of displacement and that hips allowed to heal with moderate and, in some instances, severe deformity can continue to do well into mid or late adult life. Brodetti related the vascular anatomy of the femoral head to the possibility of bad results with pinning in the superolateral aspect (36). Even if the nail or pin remains within the neck and head, placement into the superolateral aspect can damage the lateral epiphyseal arteries, which are immediate intraosseous branches of the lateral ascending cervical vessels. Brodetti's experimental work recommended the placement of nails in the central zone of the femoral head to minimize or completely prevent interference with the blood supply. More recent studies have shown that the lowest complication rates occur when the pins are in the varus position and below the superior quadrant of the head. Morrissy has pointed out that "in situ pinning is a radiographic technique" (192). The direction in which the displacement of the femoral head occurs is determined by muscular forces and the particular anatomy of a proximal femur. In SCFE, the femoral neck rotates externally and the femoral head slides posteriorly around the axis of the femoral neck but remains in the acetabulum. Inferior displacement of the head is prevented by the shape of the femoral neck, although on plain radiographs the femoral head appears to have slipped inferiorly. CT scans show that it really is the posterior displacement that is being defined. Griffith pointed this out several years ago (93). With severe slips, the

SECTION II ~ Slipped Capital Femoral Epiphysis

femoral neck may begin to move proximally in relation to the femoral head so that the metaphysis comes to the lateral edge of the acetabulum. This occurs with complete slips. CT scan is quite helpful to assess the amount of posterior slip along with the often extensive remodeling of both the anterior and posterior femoral neck surfaces. The physis in SCFE is damaged, and relatively benign interventions lead to relatively rapid fusion. Even utilization of a single transphyseal screw, which provides relatively little mechanical stability, in the vast majority of cases leads to fusion. Separate reports by Pomeranz and Sloane (213), Mayer (182), Kiaer (145), and Mathiesen (179) indicated that multiple drillings across the physis alone lead to growth plate closure. Results from More Recent Series on Pinning in Situ:

Stambough et al. analyzed 80 patients with chronic SCFE in whom pinning was performed (245). Serious complications occurred in 10 patients. The severity increased as the number of pins increased. The fewest complications occurred with varus pin position with tip placement inferiorly in the head. Most problems occurred when the pin tip was within the superior and anterior quadrants. All of their cases were chronic with symptoms for more than 3 weeks. Their group I patients were treated with three or more pins and group II patients had two or one pin. Those in group I tended by necessity to have more valgus or lateral positioning of the pins and those in group II more horizontal or varus placement. Complications were highly concentrated in group I in which there were three cases of AVN, five of chondrolysis and one of subtrochanteric fracture; none of these complications occurred in group II, with only one pin fracture seen. When complications were assessed in relation to pin position, most complications occurred when the pin was in the superior-anterior region of the head and when it was left in 2.5 mm from the subchondral bone surface. The authors clearly favor the use of two or one pin as complications markedly were diminished in those situations. They felt it mandatory to avoid the superior and anterior quadrant because of the higher incidence of AVN, pin penetration, and chondrolysis. Single or two screw placement also proved to have favorable results in a review by Herman et al., even in hips with grade III slips of greater than 50% (109). Four were acute, 11 acute on chronic, and 6 chronic. Stabilization without progression of slip was achieved in all patients, and screw placement was satisfactory by the criteria of Stambough in all patients. Complications in this group still occurred, however, with three cases of AVN and one with chondrolysis. The chondrolysis was present in the patient preoperatively. One of the potential problems was unintentional reduction of the displacement, which occurred in 9 of 15 patients just based on their movements within the hospital because attempted reduction was not part of the management scheme. AVN developed in 3 of these 9 hips. The hips thus were unstable.

417

Although earlier studies indicated that black children might be more susceptible to complications with treatment of SCFE (203), in particular chondrolysis and AVN, more recent studies have indicated that this is not the case. One of the earliest studies showing an absence of increased complications in the black patient was the study by Bishop et al. who noted a satisfactory result in 87% of 70 hips with slipped epiphyses in 50 black children (27). Chondrolysis developed in only 6% and then only when there was persistent intra-articular protrusion of the fixation pin. AVN was present in 7% but was always attributable to overreduction with an acute slip. Aronson et al. (15) concluded in a study of 55 children, 89% of whom were black, that black children were not more susceptible to chondrolysis, a conclusion also reached by Stambough et al. (245). Aronson and Loder, in a larger review of 74 black children with 97 slips, reiterated that conclusion showing satisfactory results with only 3 cases of chondrolysis in multiple or single pinning in situ (13).

The study by Aronson et al. (15) reviewing patients treated from 1977 to 1983 showed good results in general with pinning in situ, although the surgical approach was generally from the lateral aspect of the cortex in terms of pin placement and the slipped epiphysis was stabilized by 2 pins in 48 hips, 3 in 28, and 4 in 4. Eighty hips were assessed with excellent or good results in 86% of mild slips, 55% of moderate, and 27% of severe. Poor pin position was subsequently felt to be associated with 60% of the 20 poor results. The complication of chondrolysis developed in 3 hips (4%) with AVN in 2 hips (3%). The authors concluded that pinning in situ still was the advisable approach but that it would be improved by anterolateral pin entry points and also that one screw would be sufficient for stabilization and would minimize complications. This study also shows the value of early diagnosis because the results were progressively worse the greater the severity of displacement. In the large majority of patients, the induced premature closure of the proximal femoral capital epiphysis leads to insignificant limb length discrepancy problems and no meaningful change in proximal femoral structural orientation at maturity. In the subgroup of patients, however, who develop the slip at the early end of the time spectrum in the juvenile age group, premature closure can lead to limb length discrepancy and perhaps more importantly to a relative coxa vara due to overgrowth of the greater trochanter. In a review of 33 hips that had pinning of juvenile SCFE, growth disturbances in 64% of the hips were noted including trochanteric overgrowth, coxa vara, and coxa breva. Segal et al. defined the juvenile or younger age grouping as those with the onset of slip at least 1 year less than the reported mean age for the disorder, which in boys was less than 12.5 years and in girls less than 10.5 years of age (234). The growth changes were defined in this subset of patients. This group of patients also was felt to show a higher incidence of bilaterality and a higher incidence of endocrine disorder. In an effort to minimize

418

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormalities of the Femur

growth problems, differing approaches have been suggested in this particular age group. One involves the use of a hip spica to allow for stabilization of the pathological process. Although immobilization in the spica itself can induce premature fusion, a high number of patients worsen the slip following discontinuation of immobilization. Surgical approaches also have been advised. These involve stabilizing the periphyseal region and bone of the secondary ossification center with a central pin that is nonthreaded, thus allowing growth to continue while providing some mechanical support. Segal et al. showed a schematic diagram of an apparatus they developed that used two nonthreaded or smooth pins passed from trochanter through neck and physis into the secondary center (234). The two smooth pins then were bent distally from the trochanter and were stabilized against the side wall of the cortex with a one-third semitubular plate with two screws that served to buttress the pins. Stabilization in Situ Using Open Femoral Head-Neck Epiphysiodesis with Bone Graft: Open epiphysiodesis with iliac crest bone can be effective in causing head-neck union but is a more extensive procedure and is not used widely today. The open bone graft epiphysiodesis was developed by Ferguson and Howorth in 1931 (73). In the bone graft epiphysiodesis procedure, the growth plate is drilled extensively from a window made in the femoral neck. Curettage then is done to damage the physeal area further. The core site is enlarged to accommodate at least three corticocancellous grafts taken from the adjacent ilium. Each graft is at least 0.5 cm wide. Postoperative management consists of 12 weeks of bed rest followed by gradual walking with crutches and increased weight bearing as indicated by X-ray assessment. Heyman reported on 42 cases in 1949 (112) with good results using a similar technique, and Howorth (116, 119, 120) was reporting on over 200 cases without major complication. At the time, surgical stabilization of the headneck junction was difficult to obtain and use of the SmithPeterson nail had many complications associated with it. The rationale for developing the procedure thus was clear. In brief, the hip capsule was opened and several holes were drilled through the neck and the epiphyseal line into the head. Bone slivers then were inserted into the drill holes. The operation led to relatively rapid growth plate obliteration and subsequent fusion, generally within 8-10 weeks, and gained considerable prominence. Those groups developing extensive experience also reported good results. Weiner et al. reported a 30-year experience with bone graft epiphysiodesis in 1984, assessing 207 hips with very few complications (265). There was only one case of AVN in the chronic group and no cases of acute cartilage necrosis. The authors felt that the procedure was as least as good as multiple pin fixation for SCFE due to rapid growth plate closure, the avoidance of pin penetration or hardware removal, and the extremely low incidence of overall complications. They also described the procedure in greater detail in an earlier report in which 106 hips were available for examination (185).

There was no evidence of acute cartilage necrosis or avascular necrosis, and the physis closed following bone grafting in all but four hips. There was occasional graft resorption and further slipping in two hips, graft resorption alone in one hip, and one instance in which the graft was placed inappropriately. At that time they indicated that over 500 of these procedures had been reported in the world literature. Many skeptics remained, however, because the procedure was relatively complicated, requiring opening of the hip joint as well as surgical intervention at a deformed headneck region, which itself was somewhat fragile. Reports have indicated a relatively high level of complication in these procedures, although the number of cases done per surgical unit would have some bearing on the outcome. The records of 43 patients who underwent 64 open bone peg procedures were reviewed by Rao et al. (222). Healing occurred at an average of 17 weeks after surgery. At time of healing, however, 42% of the hips showed a change in the degree of slip. The average operating time was slightly over 2 hr per hip, which is considerably greater than most pinnings in situ. Complications involved 4 hips with AVN, 3 with chondrolysis, 3 with infections, 4 with delayed wound healing, and 44 with heterotopic ossification. Rao et al. concluded that "because of the potential morbidity of this procedure, we no longer perform it as the primary operation for stable slipped capital femoral epiphysis." Ward and Wood also disparaged routine use of open bone graft epiphysiodesis (260). They noted physeal fusion in only 12 of 17 cases due to resorption, movement, or fracturing of the graft. Femoral head position changes occurred along with one case of chondrolysis, one of myositis ossificans, and 10 of anterolateral thigh hypesthesia. Zahrawi et al. performed a comparative study of pinning in situ and open epiphysiodesis over a 12-year period from the same institution (275). Pinning was performed in 61 hips and open arrest in 33. At follow-up evaluation, which averaged 6-7 years, 91.7% of patients treated by pinning in situ had good or excellent results as compared with 71.6% of the patients treated by epiphysiodesis. Further analysis showed that, in those treated by pinning, 3.3% were considered failures, whereas in the epiphysiodesis group 25% were considered failures. Their comparative results showed wound infection of 3% for pinning and 12% for open epiphysiodesis. Cartilage necrosis occurred in 2 patients after pinning in situ. AVN occurred but only when manipulation was performed prior to pin insertion. Immobilization Treatment with Hip Spica Casting without Operative Stabilization: Immobilization treatment with either hip spica casting or prolonged traction was the treatment of choice in the first part of the twentieth century prior to routine use of transphyseal pinning. Overall results were fair to poor but the method continued to be used by some practitioners for several decades. Griffith reported on 67 patients treated by immobilization with traction or a plaster hip spica (93). Further slipping occurred in 24% of these hips, and in

SECTION II 9 Slipped Capital Femoral Epiphysis

10 of the hips the slipping occurred even while immobilization was being performed. Chondrolysis developed in 19 hips (28%). In only 31 (46%) of the hips treated by immobilization was further slipping prevented without complications affecting the viability of the joint. Betz et al. have reported that immobilization in a bilateral hip spica without operative intervention can prevent further slippage in a majority of cases while allowing the persistence of growth after stabilization of the pathological process (24). Some relatively older patients close to skeletal maturity will proceed to growth plate fusion in cast because the physeal area has been damaged by the slipping and continued weight bearing. The size of the large majority of patients with slipped capital femoral epiphysis plus the awkwardness of hip spica treatment in general argue against this treatment, but it can be used on occasion when operation is contraindicated by systemic illness or the patient is so young that transphyseal fusion must be avoided. They reported on 32 patients with 37 slipped hips treated in hip spica. The mean time of immobilization was 12 weeks, and whereas there were no cases of AVN, slip progression occurred in 2 and chondrolysis in 7 (19%).

b. Treatment Leading to Change of Femoral Head Position Closed Reduction: At present, efforts at closed reduction always are considered to be unwise in chronic slips but are needed for acute unstable slips as noted above where improvement of position is sought. Early attempts at improving head-neck position used strong skeletal traction through a femoral K-wire to reduce the head physically as though the displacement had been caused by an acute fracture. In general this approach did not succeed in reducing displacement, and often served to tear the posterior vessels such that femoral head avascular necrosis was a common occurrence. As noted earlier in the section on pathoanatomy, the slippage occurs slowly such that the capsule, retinacular vessels, and periosteum posteriorly tend to shorten and thicken and epiphyseal-neck stability can be enhanced by associated periosteal new bone formation. The femoral neck periosteum would have to be stretched and torn in reducing any displacement, and in many instances AVN was caused. Attention then turned to surgical intervention to reduce and reposition the head, better its geometric relationship, and hopefully protect the blood supply under direct vision. At present, awareness of whether the slippage is stable or unstable is leading to more physiological management. If a hip is categorized as stable, it implies that the head-neck bond, most of which is mediated by the periosteum, is relatively intact such that efforts at reduction might likely tear and damage the vascularity. If a hip is unstable, efforts at reduction are more likely to succeed in changing position but still carry some risk for further tearing the vascularity. The "bottom line" is that closed reduction is considered unwise by many in either instance; at best it is recommended only for the acute slip within 24 hr of presentation. At pres-

419

ent closed reduction is not recommended at all by many because of the risks of AVN. The futility of manipulation was illustrated by Griffith (93). In his large series a subset of 15 hips with gradual slipping were manipulated, but none showed any improvement in the position of the epiphysis. Of the 44 hips with acute slipping, 29 were manipulated with improvement in the position of the epiphysis in 11 (38%). However, avascular necrosis subsequently developed in 8 of these 11 reduced hips (73%). Of the 15 acute slipped epiphyses that were not manipulated, only 1 developed avascular necrosis and even that had been reduced by traction. Griffith thus demonstrated a highly significant association between closed reduction and avascular necrosis. Traction also had negative sequelae with few benefits. Seventy-nine hips showing gradual slipping were treated with longitudinal traction and internal rotation of the leg from 2 weeks to 12 months. None of the hips showed any improvement in the position of the epiphysis, but chondrolysis developed in 20 hips. Twenty of the 44 acute slipped epiphyses were treated by longitudinal traction for at least 7 days. The one epiphysis that was effectively reduced subsequently developed avascular necrosis. Six of the hips later were reduced by manipulation. Chondrolysis developed in 8 of the 20 hips. Griffith also addressed the issue of "gentle" manipulation and indicated that each surgeon considered his manipulation to have been gentle, and the fact that only 38% of the acute slipped epiphyses were reduced supports this contention. In spite of the care exercised the blood supply was frequently disrupted. He concluded that "manipulation is too hazardous to be recommended." In spite of this concern there have been some who regard manipulation in chronic slips as worthy of attempt. Fairbank reported a small series of 16, concluding that the dangers of manipulative reduction properly performed may have been overestimated (72). The value of correction was in obviating the need for extensive operative procedures. His only contraindication was bony fusion of the epiphysis. He stressed the need for "gentle" manipulation that was followed by pinning. Such concepts as "without undue force" and "firm but never forceful" indicate awareness of concerns in regard to vascularity; as noted previously, however, it simply is not possible to know by gross unmonitored reduction how the vascular supply is responding. Open Reduction: The next approach taken for moderate to severe deformity was open reduction, which was done initially with no removal of the metaphyseal femoral neck bone. Problems similar to those described with closed reduction occurred in that stretching and tearing damage to the posterior capsule and vessels proved to be extensive and subsequent avascular necrosis was unacceptably high. Open reduction for slipped capital femoral epiphysis without neck shortening via cuneiform osteotomy is never warranted in current management schemes. Open Reduction and Cuneiform Osteotomy: The next approach in treatment evolution schemes was to perform an open reduction in what is referred to as a gentle fashion after

420

CHAPTER 5 ~ C.oxa Vara in Developmental and Acquired Abnormalities o f the Femur

removal of a proximal wedge of femoral neck bone (cuneiform osteotomy) to shorten the relatively prominent anterior, superior, and lateral portions of the neck and allow the head to be reduced onto the remaining neck with no tightening or tearing applied to the posterior thickened and shortened capsule and periosteum. The open reduction with cuneiform osteotomy had been described as early as 1898 by Alsberg (7) and 1909 by Whitman (269). Whitman noted that it sometimes was possible to reduce the femoral head onto the neck without removing any bone, which represents a pure open reduction, whereas in other instances a wedge-shaped section of bone from the femoral neck was required to be removed before a satisfactory position could be obtained, which represents one of the earliest reports of a cuneiform osteotomy. This technique was referred to by Green in 1945 (92) and Martin in 1948 (178), both of whom stressed the importance of removing a sufficient portion of the neck to allow the head to be repositioned and to gently elevate the periosteum from the neck without damaging the inferior cervical vessels in the retinacula of Weitbrecht (Figs. 15A and 15B). The procedure continued to be done over the next few decades, although relatively poor results were obtained by many because of the high incidence of AVN. Gage et al. reviewed their results with cuneiform osteotomy for moderately or severely slipped femoral capital epiphysis done over a 35-year period beginning in 1938 and confirmed the highly problematic nature of the approach, with AVN at 28.5% and cartilage necrosis or chondrolysis at 37.6% (84). They performed a detailed review of the literature in relation to AVN with cuneiform osteotomy, assessing 393 hips in which the number with AVN was 83, giving a similar incidence of 21.1%. They felt, in fact, that the incidence of poor results would have been higher because many series did not comment either on chondrolysis or on lateterm degenerative arthritis. In the 71 cuneiform osteotomies done in their study prior to 1968, 60 were subcapital adjacent to the physis, 10 mid-neck, and 1 in the base of the neck. All hips had moderate or severe chronic slipping and the average age at surgery was 13 years 7 months, with a range from 10 years 9 months to 17 years 2 months. Gage et al. defined the displacement as moderate if it was between 25 and 50% and severe if it was more than 50%. The amount of slipping was determined by the amount of displacement of the epiphysis on the femoral neck expressed as a percent of the diameter of the neck as seen on the lateral radiograph. Cartilage necrosis was seen in 29 of 77 hips for an incidence of 38%. Thirteen of the 29 hips also had AVN, leaving 16 with chondrolysis alone. Twenty-one of 29 showed some improvement of cartilage joint space with time. AVN developed in 22 hips (28.5%). The average time from surgery to diagnosis was 8.8 months with some evidence of necrosis as early as 2 months. The AVN was felt to be complete in 14 hips and partial in 8. Preoperative manipulation contributed meaningfully to the rate of AVN. Gage et al. had done a few procedures, 6, with a basilar neck osteotomy, and

in fact if these were excluded from the analysis the rate of AVN went up to 31% and that of chondrolysis to 41%. However, they did not feel that a documented relationship existed between chondrolysis and postoperative immobilization. Although a chondrolysis diagnosis was not desirable, there was some evidence for recovery presumably because some of the basal chondrocytes had survived. The complications of necrosis of either bone or cartilage were related to the severity of the preoperative slipping or the amount of correction obtained. To a great extent it appeared that problems were related to the procedure itself. Their conclusion, not surprisingly, was that cuneiform osteotomy of the femoral neck in either the subcapital or mid-neck region had such a high incidence of severe complications that it should be abandoned. They indicated their preference for either basal neck or subtrochanteric osteotomy. The inherent attractiveness of the repositioning approach at this level, however, continued to encourage surgeons to perform it while making great efforts to minimize the negative sequelae. Particularly effective in this regard was Dunn of England, who reported on his approach in 1964 (62). He clearly described the six possible ways in which a displaced slipped upper femoral epiphysis can present. These involved an acute traumatic displacement in a hip that previously was normal, an early chronic slip, an acute on chronic slip, a severe chronic slip in which the epiphyseal line was still open, a severe chronic slip in which the epiphyseal line was closed, and the finding of secondary arthritis with deformity in early adult life. The physiological considerations are illustrated in a series of drawings in Fig. 16 derived from his work. His treatment plan followed in logical fashion with the acute slip, which essentially was a type I epiphyseal fracture-separation treated as an emergency with reduction perforrhed within a matter of hours and the femoral head fixed by closed pinning. In the early chronic slip with mild to moderate displacement, the femoral head could be stabilized without manipulation by pinning in situ. In the acute on chronic slip, the retinacular vessels will have been shortened so no attempt should be made to bring about closed reduction, with the deformities best treated by open operation with shortening of the neck by trapezoid osteotomy. If the epiphyseal line was open, Dunn felt that repositioning osteotomy was valuable and that open reduction and cervical osteotomy were preferable to a subtrochanteric osteotomy. He stressed the importance of careful technique, entering the hip by a lateral approach, elevating the trochanter, protecting the retinacular vessels by careful elevation of the synovium from the back of the femoral neck, and cervical shortening before replacing the head on the end of the neck. If the epiphyseal line was closed, cervical osteotomy and open reduction had no part to play with deformity to be corrected by either subtrochanteric osteotomy or cervical osteoplasty. Dunn thus proposed the open reduction and proximal cervical osteotomy for the acute on chronic slip in which neck shortening with the trapezoidal osteotomy was essential and

SECTION II ~ Slipped Capital Femoral Epiphysis

F I G U R E 15 Illustrations showing the principles of open reduction and cervical wedge osteotomy to minimize negative sequelae. (A) The periosteum gently is freed~fromthe bone and the neck is shortened by removing metaphyseal bone (hatched areas in D and E) to allow the head to be repositioned without stretching the shortened posterior periosteum. The same principles are outlined in part (B). An AO cancellous screw now is used for stabilization. [Part A reprinted from Green, W. T. Arch. Surg. 50: 19-32, copyrighted 1945, American Medical Association.]

421

422

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

FIGURE 16 Vascularconsiderations in performance of open reduction and cervical osteotomy, derived from the works of Dunn, are shown here. In these illustrations the proximal end of the femur is shown in lateral projection. Displacement of the head, when present, is posterior. The two illustrations at fight show the negative consequence of closed or open reduction in chronic SCFE without shortening and reshaping the neck. At lower fight the black femoral head represents an intact vascular supply even with displacement. At upper fight, reduction without surgical shortening of the neck causes rupture of the posterolateral vessels as they enter the head. The white femoral head represents avascularity. The three diagrams at left illustrate the technique of open reduction in severe chronic SCFE with shortening and posterior trimming of the neck, periosteal elevation to protect the blood supply of the head, and gentle reduction stabilized with three pins. The stippling of the femoral head represents an intact blood supply at each stage. Arrows at lower left represent incision and gentle elevation of the periosteum to protect posterior vessel source. Arrows in middle left diagram illustrate segments of neck bone resected to allow for shortening and debulking of the posterior neck new bone formation. Reduction of head in upper fight diagram is then safely accomplished with vessel supply intact. Stabilization is with three pins. [Reprinted from Dunn, D. M. (1964). J. Bone Joint Surg. [Br] 46B:621-629, with permission.]

for the severe chronic slip in which the epiphyseal line was still open. For the late chronic slip in which the epiphyseal line had closed, cervical osteotomy clearly would transect the intraosseous vessels such that, even if the posterior retinaculum remained intact, the risk of AVN was far too high. The primary object of the operation was to place the femoral head on the end of the neck without stretching the retinacular vessels. This could be achieved in one of two ways: first, the neck can be shortened by a cut just distal to the epiphyseal line and at a fight angle to the long axis of the neck so as to include about 88 in. of the posterior aspect of the neck, with the segment removed being trapezoidal rather than cuneiform as often indicated because the posterior bony beak also was removed. The second point involved care such that vessels on the posterior surface of the neck and head would not be damaged. Dunn suggested an anterior approach to the neck with sub-periosteal dissection maintaining the

posterior vascularity such that the head could be eased off the neck and the synovium raised from the back of the neck right down to the base of the capsule. The back of the neck could now be seen, the bony beak trimmed, and the superior surface of the neck trimmed. It then was possible for the head to be replaced with the entire posterior synovium of the neck free of tension. Stabilization was performed with a nail and the patient kept on bed rest for 1 month, at which time crutch walking began. Union of the epiphysis to the metaphysis generally occurred around 3 months. Dunn reported on 23 open reductions, 19 of which did very well going on to become clinically normal. Four hips developed complications involving one segmental necrosis of the femoral head, one complete AVN, and two instances of chondrolysis. He compared his approach to the intertrochanteric osteotomy and supported the former because it fully corrected the anatomical abnormality rather than compensating for it. Dunn and Angel further reported their detailed results in 1978 (63). They had also reviewed the literature in relation to the open reduction and cuneiform osteotomy for severe slips. Their review table reported only those operations that were undertaken as a primary procedure and in which particular attention was paid to the blood supply to the femoral head. They noted documentation of 94 cases in reports from 1945 to 1972 with an AVN rate of 14% (13 cases). They separated out the acute slip with a violent injury and the severe chronic slip treated after growth plate closure because they represented different entities. Dunn and Angel stressed that open reduction and cuneiform osteotomy should not be performed in the severe slip with growth plate closure because of the unacceptably high risk of AVN due to intraosseous vessel communication between the metaphyseal and femoral head vessels. They reviewed their assessments using the three-type categorization of an early chronic slip, an acute on chronic slip, and a chronic slip with the physis open. The term "acute on chronic" slip was used when there was evidence of new bone on the posterior aspect of the metaphysis, indicating that the acute episode had been preceded by a minor degree of chronic slip. Even though some minor trauma might have occurred, the severity of injury would not have been expected to fracture the neck of the femur or displace the epiphysis in an otherwise healthy person. Thus, Dunn and Angel were reluctant to use the term "acute" even as others used it in cases in which there was severe pain less than 3 weeks prior to treatment and frequent inability to bear weight. They reserved the term acute for what essentially is a type I epiphyseal fracture-separation. The open reduction and cuneiform osteotomy were reserved for the acute on chronic slip with severe displacement or for the chronic slip with severe displacement with open physis. Seventy-three procedures performed over a 23-year period were assessed. No cases of acute traumatic slips were included. There were 25 patients with acute on chronic slips and 48 with severe chronic slips. Dunn and Angel reduced the femoral head anatomically. It was necessary to remove

SECTION Ii ~ Slipped Capital Femoral Epiphysis

the posterior bony beak otherwise the retinacular vessels would be stretched over it. A lateral approach was preferred. The greater trochanter was elevated through its growth plate. The capsule was incised in the long axis of the neck with the incision extended around the anterior and posterior edges of the acetabulum. The subluxated femoral head was visible along with the reddish posterior surface of the femoral neck. The front of the femoral neck was pale and avascular. The synovial membrane on the neck was incised in front of the vascular area around the anterior margin of the head. The posterior soft tissues then were stripped sub-periosteally to the margin of the head and down to the base of the neck, using extreme caution. The growth plate was visualized peripherally and a wide gouge was inserted between the cleavage plane through the growth plate between the head and neck of the femur. Dissection was performed on the remains of the growth plate, levering the head off the neck. Two osteotomies were performed. The first was along the long axis of the neck to remove the bony beak from its posterior aspect. The second shortened the neck by a few millimeters and was made with a slightly curved sweep transverse to the top of the neck. This removed the remains of the growth plate from the neck and the adjacent metaphyseal neck bone. The remains of the growth plate then were removed from the femoral head. The head at this stage was reduced onto the neck without tension on the posterior structures. If there was any degree of tension, the neck was shortened further. Three pins were drilled up the neck to emerge at different positions. The deformity was reduced and the pins driven farther into the head for stabilization. In the lateral view, the head was reduced fully. In the anteroposterior view, it was important not to overposition the head into varus but rather to leave it with a slight valgus position of some 20 ~ The trochanter was reduced and held with pins, and soft tissue closure followed. Skin traction was maintained for 4 weeks after which crutch walking began. Assessments were made by subjective, clinical, and radiological criteria with results expressed as good, fair, or poor. The major complications assessed were AVN, chondrolysis, and osteoarthritis. In 73 hip operations, the entire series, clinical results showed 75.3% good, 8.2% fair, and 16.4% poor. Radiologically, the results were less effective but still registered 56.2% good, 16.4% fair, and 27.4% poor. In a review by diagnostic category, 7 operations were done with a severe chronic slip in which the growth plate was closed. It soon became evident that results in this category were poor, and indeed 6 were listed as poor with only 1 good. This group subsequently was not felt to be appropriate for open reduction and toward the latter part of the series these individuals were treated with trochanteric osteotomy. The other two groups had improved results. Analysis was performed on 40 cases of severe chronic slips with the physes open in white children only and 23 acute on chronic slips in white children only. The relatively few black patients in the series were not assessed in the subgroupings because of the then widely held

423

belief that blacks did not do as well with the SCFE disorder. Indeed, of the 4 black patients in the series, all did poorly. In the patients with severe chronic slip with plate open, clinical assessment indicated 92% good, 3% fair, and 5% poor, whereas radiologically there were 75% good, 15% fair, and 10% poor. In the acute on chronic group, clinically there were 70% good results, 4% fair results, and 26% poor resuits, whereas radiologically there were only 43% good resuits, 26% fair, and 30% poor. The group that did the best, however, was that with the severe chronic slip with physis open. AVN occurred in only 1 case of 40 with open growth plate and no acute displacement. With acute on chronic slips, vascular changes were much more frequent. Dunn and Angel felt that this was not due to technical reasons because in this group the operation technically was easier, but rather was due to the fact there had been damage to the vascular pedicle at the time of the acute slip or kinking of the vessels before the lesion was able to be repaired surgically. Chondrolysis was seen more throughout the spectrum of disorders; there were 13 cases with severe chronic, 3 with acute on chronic slip, 3 in 4 of the black patients, and 4 in the 7 patients with closed growth plates. A review of the relatively few radiographs shown in the article, using the benefit of knowledge gained subsequently based on the work of Waiters and Simon, shows that some of the pins were too long and might have compromised the articular cartilage. In addition, some patients had a wide, rigid, triflanged single nail inserted, which may have caused distraction of the vessels posteriorly at the time of passage of the nail from the relatively soft neck into the harder head. Broughton et al. reported a later series in which cuneiform osteotomy had been performed in 115 hips (37). At a mean follow-up of 12 years there were 3 cases of AVN alone, 2 of AVN and chondrolysis, and 8 of chondrolysis alone. Rey and Carlioz reported good results with the Dunn procedure (225). One of the main proponents of cuneiform osteotomy of the femoral neck in the United States has been Fish (76, 77). He assessed 42 hips with sufficiently severe displacement to require surgical correction by means of a cuneiform osteotomy of the neck of the femur just distal to the physis (76). The purpose, similar to that of Dunn, was to restore the normal anatomical relationship of the head to the neck. His results were excellent in 40 of 42 cases, with only 1 instance of avascular necrosis and 1 of osteoarthritis. Fish noted that reviews from the earlier literature reporting the worst results were those in which manipulation was performed prior to surgery and a Smith-Peterson nail was used for stabilization. Fish's report involving 42 hips with cuneiform osteotomy indicated that all had slips of greater than 30 ~ of displacement. Seven patients had an acute slip on a preexisting chronic slip. Cuneiform osteotomy was considered to be the only procedure that could fully correct the deformity anatomically, with the basicervical procedure allowing correction for a slip of 50 ~ and the biplanar trochanteric osteotomy for one up to 70 ~ Surgery was done within 24-48 hr after

424

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

presentation. The patient had the limb elevated on pillows in flexion, abduction, and external rotation with no effort at traction. Sufficient bone had to be removed to allow easy anatomical reduction of the head on the neck. The base of the curved wedge must be in the plane of anticipated correction of the epiphysis. The wedge is removed so that the curved contour of the physis will match the corresponding curved cancellous surface of the neck of the femur. The bone in the posterior aspect usually is removed by a small curette or hand-held curved osteotome. When anatomical reduction of the epiphysis has been accomplished, fixation is obtained by 2 - 4 pins. Fish concluded that it was possible to perform cuneiform osteotomy of the femoral neck with only slight danger of avascular necrosis of the head of the femur. Patients whose physes had already closed were not candidates for cuneiform osteotomy because of increased damage to the intraosseous vessels. Gentle preoperative positioning without traction, the need for no manipulation, and surgery of a meticulous type done shortly after presentation and certainly within 24 hr were essential. Fish reviewed his series 10 years later at which time a long-term clinical and X-ray assessment of cuneiform osteotomy was possible in 61 patients with a slip of greater than 30 ~ (77). The results were excellent in 55 hips, good in 6, fair in 2, and poor in 3. Pin penetration was noted in all 6 of the patients who had osteoarthritis and in 1 patient who had chondrolysis. Complete AVN developed in 2 patients with segmental AVN in 1. Each of these 3 had an acute on chronic slip. Fish thus documented excellent results in 83.3% of 66 patients, good in 9.1%, fair in 3%, and poor in 4.5%. By using the grading categories of Dunn and Angel, good results (excellent + good) totaled 92.4%. The value of the open reduction and cuneiform osteotomy was great in the severe chronic slipped epiphysis, but the riskier group was the acute on chronic slip. Because the AVN only occurred in those patients, the precipitating factor appeared to be the injury to the capsular blood supply at the time of the slip and not the osteotomy. Clarke and Wilkinson modified the Dunn approach to cervical osteotomy and reported improved results (50). They utilized Muller's anterolateral trochanter sparing approach. Following a T-shaped incision in the capsule, the large amount of callus protruding from under the anterior edge of the acetabulum was seen (Fig. 17). The attachment of the periosteum to the posterior edge of the epiphysis was preserved. The callus was removed with a curved osteotome followed by curettage. Once all of the repair callus had been removed with the original femoral neck remaining intact, the epiphysis was noted to slip forward into normal place. The femoral neck was rotated laterally to expose the posterior lip of repair at its proximal end, and this was carefully removed. The epiphysis was gently reduced by internal rotation of the shaft and neck following which a cannulated compression screw was inserted for stabilization. The patient was protected in a plaster spica for 8 weeks. Results with this approach were considered superior to the classic Dunn

FIGURE 17 The gross appearance of a severe chronic slip at open operation is shown. The hypertrophic call~s between the head and neck must be excised at openreductionalong with someof the adjacentneckbone. [Reprinted from Clarke, H. J., and Wilkinson, J. A. (1990). J. Bone Joint Surg. [Br] 72B:854-858, with permission.]

procedure. They reported excellent results in 13,good in 2, and only 1 poor result with avascular necrosis. In the 16 patients, there were 2 episodes of segmental AVN, 2 of chondrolysis, and 1 of osteoarthritis. DeRosa et al. supported the value of cuneiform osteotomy of the proximal neck of the femur in severe SCFE (59). They reviewed 27 severe grade III slips over a 10-year period treated this way. No hips were rated excellent because they felt that such a rating was warranted only for normal hips unaffected by disease or surgery. The results were 19 good, 4 fair, and 4 poor, with each of the poor results associated with AVN. The AVN rate was 15%. In percentages, the results were good in 70.4%, fair in 14.8%, and poor in 14.8%. They concluded that the cuneiform osteotomy continues to have a place in the treatment of severe SCFE and in particular with severe grade III slips of 60 ~ or greater and open physes. The shortening resection of the neck particularly posteriorly is needed to remove reactive bone. Velasco et al. performed open reduction and cervical cuneiform wedge resection according to the technique of Dunn in 66 hips with moderate to severe slippage. Avascular necrosis occurred in 7 cases and chondrolysis in 8. In 48 hips followed for more than 10 years (mean of 20.6 years), results were classified as good in 22, moderate in 16, and poor in 10. They felt that their results were better than studies from pinning in situ with moderate and severe slips (255). The major point of contention between those favoring open reduction and cuneiform osteotomy and those recommending moredistal procedures at either the base of the neck or the inter- or subtrochanteric region is the question of whether full anatomic restoration is needed to prevent future osteoarthritis. Although the basicervical and trochanteric

SECTION II ~ Slipped Capital Femoral Epiphysis

procedures are, to an extent, compensatory, it has not been shown that the compensation leads to a change in the headacetabular relationship sufficient in itself to induce the osteoarthritis. A detailed study in this regard would be of extreme importance. Compensatory Osteotomies: Early reports with the open reduction approach documented a 30-40% incidence of avascular necrosis. Although this has diminished in more recent reports, it has not been eliminated. At this stage in the evolution of the treatment of slipped epiphyses, individuals have resorted to compensatory osteotomy in safer regions more distal to the epiphyseal vascularity. Crawford has reviewed the various levels and types of corrective procedures used (55). This approach was popularized by Southwick (241) in North America and Imhauser (121,122) in Europe, both of whom developed similar triplanar subtrochanteric osteotomies. The osteotomy removes an anterolateral wedge of bone, which allows for correction of varus, extension, and external rotation deformities. The principles are summarized in Fig. 18A from several operative approaches. The Imhauser, often performed as the Imhauser-Weber (263), osteotomy is intertrochanteric. The final geometric change involves internal rotation of the distal fragment to correct the fixed external rotation positioning of the shaft, flexion of the distal segment to correct the proximal extension deformity position of the head, and valgus repositioning to correct for the varus position of the head. Subsequently, surgeons have performed the triplanar intervention at increasingly more proximal levels. An intertrochanteric approach can give excellent correction and is performed closer to the site of the deformity than a subtrochanteric approach, although still in the safe range in terms of the proximal femoral vasculature. Kramer et al. have performed a basicervical osteotomy (156). In the Kramer procedure, which he referred to as a compensating osteotomy at the base of the femoral neck, a section of bone at the base of the neck is removed through the anterior trochanteric region. The osteotomy is proximal to the greater trochanter of the femur such that the abductor musculature is restored to its physiological position. The procedure serves to correct both the varus, on the basis of the laterally based wedge, and the extension deformity on the basis of the slightly anterior position of the wedge. Kramer's group selected for operation patients with 40 ~ or more of deformity on either the anteroposterior or lateral radiograph. The capsule of the hip joint is opened anteriorly, and the amount of bone to be removed is determined primarily by visualization at the time of surgery. The osteotomy is considered by some to be extracapsular technically because anteriorly the capsule of the hip joint and the retinacular blood supply extend to the intertrochanteric line, but posteriorly the capsule of the hip joint and the blood supply extend only to the junction of the middle and distal one-thirds of the femoral neck. Compensatory osteotomy thus lies distal to the more important posterior blood supply. The more distal of the two osteotomy lines is made

42S

first and is perpendicular to the femoral neck along the anterior trochanteric line. The osteotomy is extended to the posterior cortex, which is left intact. The second osteotomy is oblique. The wedge itself is anterolateral to correct both the varus and extension components. The osteotomy has been held with 2-3 pins between the trochanter and neck. The greater trochanter then is reattached. The initial report was on 56 compensating osteotomies, although the results were not presented in detailed fashion. There was need for additional surgery in 2 either to increase the valgus position of the head or to change the rotational component. There were 4 instances of avascular necrosis, but these were felt to be due to falls subsequent to the surgery and compression fracturing of the femoral head. A few instances of chondrolysis also were reported, although in one case in particular it was felt to be due to pin penetration. A detailed long-term review of an extracapsular base of neck osteotomy by Abraham et al. reviewed 36 hips with moderate-severe SCFE (2). They felt that 90% had excellent or good results and documented no instances of AVN. The procedure involved a two-plane wedge osteotomy based anteriorly and superiorly on the anterior surface of the base of the neck. The distal osteotomy line starts from the lesser trochanter and passes through the growth plate of the greater trochanter. The extent of the wedge at its widest part generally is in the 15-mm range. The cuts converge posteriorly to make a single osteotomy on the posterior cortex. The external rotation component also is corrected as the osteotomy is closed with an intemal rotation maneuver. The maximum bone wedge removed is 20 mm. The osteotomy then is fixed with 3 - 4 cannulated screws. The postoperative care involves partial weight bearing with crutches for 6 - 8 weeks. Interpretation of the results of Abraham et al. indicated 21 hips excellent, 11 good, 2 fair, and 2 poor, leading to 90% excellent and good results and 12% fair and poor results. There were no cases of AVN. Chondrolysis developed in 5 hips in 3 patients. The extracapsular base of neck osteotomy was a safe way to correct moderate to severe AVN, and AVN had not been shown to develop in any of the 36 hips treated. The bony cuts are performed distal to the medial circumflex artery as it courses along the posterior edge of the hip capsule of the femoral neck. With a severe slip, the amount of correction of the femoral head varus and posterior tilt was somewhat limited with this procedure, and with the most severe slips, a perfectly normal head-shaft angle could not be restored. Removal of a wedge greater than 20 mm in width at its greatest extent was not warranted. Osteotomy also has been done in the intertrochanteric and immediate subtrochanteric regions to correct the head-neck deformity in a compensatory fashion while minimizing or avoiding the severe complication of avascular necrosis. Southwick popularized osteotomy through the lesser trochanter for SCFE in which a triplanar correction was achieved (241). He proposed use of the procedure when the head had slipped from 30 to 70 ~ in any plane. His initial report

426

CHAPTER 5

~

Coxa Vara in Developmental and Acquired Abnormalities of the Femur A Triplanar Osteotomies

1

1 2 ~ ~ , , ,"

~~

closing

~ ntenor wedge

1. Anterior Based Wedge 9 to correctposteriorheadtilt (flexionosteotomy) 2. Lateral Based Wedge 9 to correctvarusheadposition (vaigus-abductionosteotomy) 3. Internal Rotation 9 to correctextemalrotation deformity(derotationosteotomy)

SECTION II ~ Slipped Capital Femoral Epiphysis

was based on 28 hips. Southwick wrote on the detailed preoperative planning with templates to mark the wedge angle and size needed. The operation essentially is an anterolateral closing wedge, with that part based laterally allowing for correction of the varus deformity into valgus and that part based anteriorly allowing for correction from extension of the head to a relatively more flexed and normal position. The triplanar aspect is fully achieved with internal rotation of the distal fragment to correct the external rotation deformity. Fixation was by a relatively crude external fixation device rather than with the blade plate, which increasingly came to be used by others with time. The distal fragment was rotated internally, flexed, and abducted to bring about the triplanar correction. Once the osteotomy healed and the limb was placed in the weight bearing position, varus deformation would have been corrected to a valgus orientation, the posterior angulation or extension of the head would have been flexed into a normal position, and the external rotation deformity of the lower extremity would have been corrected by the internal rotation mechanism. Southwick noted that, if posterior slipping of the head was greater than 60 ~, it was corrected incompletely by the wedges because tilting of the osteotomy site of more than 60 ~ would produce excessive shortening of the femur. The internal rotation of the distal shaft partially rectified the positioning of the head. Southwick did not specifically pin the physis and noted that, by the time the osteotomy had healed, in virtually all instances the physis had undergone fusion as well. A hip spica was used postsurgery for 8 weeks primarily to protect the epiphysis, which had not been stabilized by pinning in situ. In rating his own results in the 28 hips, 21 were excellent, 5 good, and 2 fair. There were no poor results and no cases of avascular necrosis in the 28 hips. In an additional note, Southwick added that 55 patients had been treated by this method (the later ones not being included in the report) and avascular necrosis still had not been documented. The study assessed patients who had been followed for 5 years or more. There were a few instances of joint space narrowing, which would be referred to as chondrolysis, although considerable improvement with time was noted consistent with other reports. Supporters of the intertrochanteric and subtrochanteric osteotomy point to its value in not causing avascular necrosis

427

and indicate that the compensatory nature of the correction leaves the head and neck in an appropriate anatomic relationship to the acetabulum. The intertrochanteric procedure became the standard procedure for severe and in some instances moderate SCFE in many centers over the next three decades. Southwick discussed the level of osteotomy for severe slippage in an editorial in the Journal of Bone and Joint Surgery in 1984 in response to two papers, one of which described excellent results following the cervical osteotomy and the other which described excellent results following the biplanar osteotomy at the lesser trochanter (242). Southwick reminded the readers that "high femoral neck osteotomy has proved to be very dangerous in the hands of most surgeons." He repeated the rationale for the triplanar approach stressing again the absence of avascular necrosis in the procedure, felt that chondrolysis was not inherent to the procedure, and supported the value of the basicervical procedure of Kramer. Ireland and Newman reviewed 35 intertrochanteric compensatory osteotomies from their unit (125). They pointed out that Perkins (209) in 1932 and Newman (196) in 1956 had performed the procedure as well. The operative corrections achieved were intended to be approximate being based on two plane radiographic appearances of varus and extension. The osteotomy itself, therefore, combined valgus and flexion with a third component, which was a final derotation to neutral of the distal femoral shaft on the proximal fragment giving the triplanar correction. They did not pin the involved epiphysis, noting no tendency to re-displacement after osteotomy. Correction was held with a blade plate. The surgical description indicated only that an appropriate anterolateral wedge was removed after which the femoral shaft was flexed and abducted slightly to close the wedge, at which time the leg was rotated to the neutral position and the side plate stabilized to the femoral shaft. Hip spica was used for 6 weeks. Ireland and Newman's results included both clinical and radiological assessments. The clinical results were 28 good, 5 fair, and 2 poor, and the radiological results were 21 good, 10 fair, and 4 poor. There were no instances of avascular necrosis, although chondrolysis was noted in 4 all of which went on to a radiologically poor result. There was no evidence of further epiphyseal slip postsurgery. Ireland and Newman concluded that, although three-fourths of the hips

FIGURE 18 The principles of the compensatory osteotomy for moderate to severe slipped capital femoral epiphysis are illustrated in part (A). They are used at the basicervical and intertrochanteric regions. Each of these approaches addresses the posterior slippage or hyperextended position, the varus malformation, and the external rotation deformity. Clinical examples are shown in parts (B-D). (Bi) Anteroposterior radiograph shows bilateral slipped capital femoral epiphysis following initial treatment. A mild slip on the right was pinned in situ with an Ace screw and healed uneventfully. A more severe slip on the left was pinned initially with two AO compression screws. (Bii) Frog lateral view shows central neck and head position of the single pin fixation on the right. On the left, the posterior displacement of the head relative to the neck is seen as is the position of the two compression screws. Note the relatively more proximal and anterior entry points of the screws on the left in relation to the severe displacement compared with positioning on the fight in which there is virtually no displacement. (Ci) Repositioning osteotomy later was performed on the left hip after physeal fusion had occurred. Intertrochanteric osteotomy fixed with an AO blade plate improved the varus, external rotation, and extension of the head and neck fragment. (Cii) Frog lateral view shows improved position of the head particularly in relation to the flexion component of the osteotomy. Note the changed orientation of the two persisting compression screws. (Di, Dii) More magnified prints of the left hip region showing AP and lateral projections to better demonstrate the extent of the correction with the triplanar osteotomy.

428

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormalities of the Femur

had clinically good results, some of these were downgraded because of the radiologic appearance, which was based on inadequate correction of deformity. In all cases, this was felt to be either an undercorrection as seen on the lateral radiograph with not enough flexion built into the procedure or a tendency to overcorrection on the anterior posterior radiograph with too much valgus created. Imhauser has written extensively on his intertrochanteric osteotomy (121, 122). A report of 55 patients followed 1122 years post three-dimensional intertrochanteric osteotomy indicated that 40 of the 55 hips examined clinically and radiologically had full ranges of motion, with 10 hips showing minimal and 5 hips significant limitation of motion. Radiologically 73% of the 55 hips were rated excellent or good with 27% showing early degenerative arthritis. Imhauser felt that the intertrochanteric osteotomy prevented or at least delayed the development of degenerative changes. Weber supported the value of the Imhauser osteotomy for major slips between 20 and 50 ~ (263). The procedure was combined with transphyseal pinning. Fineschi and Guzzanti described a linear intertrochanteric osteotomy for severe slips in which deformity was corrected by manipulation of the proximal fragment to position the head in a normal centered position in the acetabulum rather than by removal of specific wedges, followed by insertion of a blade plate (74). They followed 21 cases from 4 to 12 years postsurgery and noted no episodes of avascular necrosis. The purpose of the procedure was to align the head appropriately with the acetabulum with no concern about the compensatory deformity of the neck and trochanteric regions. They reported excellent functional results in 18 of 21 cases. Of the other 3, 2 had slight restriction of movement and 1 was graded as poor because of stiffness. Their favorable impression of the inter- or transtrochanteric osteotomy was based on its ability to recenter the epiphysis efficiently with no or minimal risk of avascular necrosis. In centers performing relatively large numbers of realignment procedures and having experience in several types of operative approach, good to excellent results are frequently demonstrated. Ballmer et al. reported on 63 slipped capital femoral epiphyses treated either by femoral neck osteotomy or by intertrochanteric osteotomy (19). At an average follow-up of 10 years, 90% were rated good to excellent clinically. In 33 hips with a follow-up of more than 10 years, mild degenerative arthritis was present in 36%. Ballmer et al. felt that slipping of 60 ~ or less was best treated by the intertrochanteric osteotomy, with the femoral neck osteotomy, which had a higher complication rate, reserved for slipping of 60 ~ or more. Szypryt et al. also compared two operative approaches for moderate or severe slips (250). They assessed 23 patients who had undergone a Dunn's open reduction and 30 hips treated by epiphyseal arrest and osteoplasty as advocated by Heyman and Herndon. The 11 hips with moderate slip (30- 50 ~ treated by the Heyman-Herndon procedure did significantly better than 18 hips with severe slip (>50 ~)

treated by the same method. The Dunn procedure was more effective in those with severe slips displaced greater than 50 ~ The Kramer osteotomy provides correction of 50 ~ and the Southwick procedure correction of 60 ~ Carlioz et al. reviewed operative intervention for SCFE in 80 cases (43). They detailed their guidelines, from earlier studies which involved in situ fixation for displacement less than 30 ~ in situ fixation followed by corrective osteotomy at the intertrochanteric level for those in which there was 3060 ~ displacement; osteotomy through the femoral neck (cervical osteotomy, Dunn procedure) for displacement between 60 and 90 ~, and closed reduction and screw fixation for acute slipping. In no instances was casting used. When the open reduction and cuneiform osteotomy were used, they reported 20 good results, 3 fair, and 4 poor. There was relatively little use of the intertrochanteric osteotomy, but results in 5 cases were 4 good and 1 poor. In the open reduction cases, of the 4 poor results, 3 involved cases of chondrolysis. Carlioz et al. concluded that the open reduction with cervical osteotomy was "difficult and dangerous" and that it should be used rarely and only in those with extreme displacement. They ultimately recommended the triplanar trochanteric procedures as being much safer. They revised their earlier protocol recommending the open reduction only for slipping of 90 ~ or greater with the growth plate open. With such a degree of severe deformity with the growth plate closed, intertrochanteric osteotomy was mandatory. Carlioz et al. supported the belief that AVN never occurred spontaneously in the natural history of slipped epiphysis and that it essentially was always a complication of intervention. Chondrolysis could occur spontaneously but generally was associated with management involving such criteria as closed reduction, penetration of the screw into the joint cavity, excessive immobilization in cast, or excessive degrees of valgus or flexion during corrective osteotomy. There are two possible times of intervention for any compensatory osteotomy. One approach is to do both head-neck stabilizing and repositioning procedures at the same time. The other is to perform an in situ stabilization pinning and then wait several months before performing the corrective osteotomy. The latter approach is preferable in the view of some for two reasons. Most importantly, the complication of hip stiffness has been reported when osteotomy is done at the same time as pinning. In addition, some patients do not find the slight external rotation position of the lower extremity either cosmetically or functionally troubling and elect not to proceed with repositioning osteotomy. It will be important for clinical studies to determine the long-term sequelae of leaving moderate and even severe slips without corrective osteotomy. Examples of osteotomy to correct severe deformity are shown in Figs. 18B-18D. Conflicting Approaches to Treatment in Moderate and Severe Slipped Capital Epiphysis: No universal agreement has been reached concerning the treatment of moderate and severe slipped capital femoral epiphysis. Many still recommend pinning in situ of a slipped epiphysis regardless of its

SECTION II 9 Slipped Capital Femoral Epiphysis

extent followed by observation of the patient and performance of compensatory osteotomy at a later date and sometimes only with meaningful symptoms as an indication for intervention. Many patients handled under this approach, therefore, wouldhave the pinning alone with no osteotomy ever performed. This is particularly true in moderate slips. The tendency to need surgical realignment is greater with the severe slips, but even here some centers have allowed patients to go without correction and noted relatively minimal long-term symptoms. Postpinning remodeling: Pinning alone particularly in moderate slips relies on the fact that patients can compensate on their own for the mild to moderate deformity, often with the position somewhat improved by remodeling of the headneck region in association with repair and continued use. O'Brien and Fahey showed how femoral neck remodeling improved the radiographic appearance of the proximal femur several years after pinning in situ (201). A large number o f their patients were treated only by pinning in situ. In a subgroup of 12 patients with moderate and severe displacement assessed 2-17 years postpinning, all but 2 had satisfactory remodeling of the head and neck and were asymptomatic. Even the 2 with minimal remodeling were asymptomatic. They also noted some spontaneous correction of the external rotation deformities. In those in which failure of remodeling would become a problem, O'Brien and Fahey recommended the cervical osteoplasty of Heyman et al. to improve motion. This approach continues to be used and described even in papers published within the past few years. Bellemans et al. reviewed 59 hips in 44 children with SCFE, all treated by pinning in situ (22). The average clinical and radiologic follow-up was 11.4 years, and they noted 53 hips (90%) to be excellent or good. Their study assessed postoperative remodeling, which was accomplished by local resorption and apposition of bone and also, they felt, by correction of the disturbed anatomic axes in proportion to the severity of the slip along with global thickening of the femoral neck. Resorption of the superolateral prominent portion of the metaphysis of the femoral neck was noted in 54% of cases, and apposition of new bone at the posteroinferior aspect of the neck was seen in 59%. The average frog leg head-shaft angle of Southwick decreased an average of 13.5 ~ from 25 ~ immediately postoperatively to 12.5 ~ at latest follow-up. The AP head-shaft angle decreased an average of 7 ~ from 16~ on the first postoperative radiograph to 9 ~ at latest follow-up. The average thickness of the neck also was increased significantly with time compared with the contralateral normal side. Favorable results could be obtained with pinning in situ due to the global remodeling process because the remodeling was more extensive than had been reported previously. Due to the fact that pinning was felt to have fewer complications than redirectional osteotomies, the simpler procedure provided very satisfactory long-term results due to the remodeling processes. The Southwick lateral head-shaft angle was calculated on the frog leg lateral radio-

429

graphs by subtracting the head-shaft angle on the normal side from that on the affected side. The average slip in this study according to the Southwick criteria was 25 ~ with 56% of the slips mild (<30~ 39% moderate (30-60 ~ and 3% severe (>60~ Jerre et al. studied hip motion at an average of 32.7 years after SCFE in 128 hips without signs of osteoarthritis (129). They concluded that hips with no treatment or those treated with fixation in situ only showed no clinically significant loss of hip motion as compared with normal hips. The greatest loss of motion of the hips in those treated with fixation in situ was diminution of internal rotation. They concluded that the loss of hip motion after fixation of the epiphysis in situ over the long term was "very slight and hardly clinically relevant" and that there was no indication for early surgical intervention with osteotomy. Siegel et al. studied 39 patients 2 years after pinning in situ for slipped capital femoral epiphysis (238). The study was done to assess range of motion of the hip and femoral remodeling. Although there was considerable increase of motion from the preoperative state, they felt that this motion returned despite minimum bony remodeling. The greatest percentage of motion of the hip returned within 6 months after treatment. This served in particular to increase the amounts of flexion, abduction, and internal rotation. They attributed this increased motion to relief of pain, spasm, and synovitis and to subsequent soft tissue stretching. Plain radiograph and CT imaging studies assessed bone remodeling itself. In spite of smoothing of the contours of the head-neck axis with resorption of bone from the superior surface of the neck and synthesis of bone inferiorly and posteriorly, only minimal change occurred in the relationship of the femoral head to the shaft and no change occurred in the angle between the femoral neck and the shaft after fixation in situ. Remodeling and resorption thus led to a smoothing of the contours of the proximal femur without a change in the axes of the deformed bones. Complications versus long-term values of realignment procedures: Aronson and Karol strongly support the value of pinning in situ for the stable slip regardless of the severity (12). Stabilization of the slip and prevention of AVN and chondrolysis are paramount, and there is no role for early realignment procedures. They stressed that the stable slip will not reduce with intraoperative manipulation and that open reductions still are associated with too many complications. This is the case particularly if osteotomy is done at the same time that the slip is stabilized. Hall found AVN to be the most common cause of a poor result. The percentage with AVN was calculated for each of several treatment groups. The pattern of AVN worsened with the incidence of manipulation and the more proximal positioning of osteotomy in relation to the physis. When pinning was performed with the narrow Moore pin with or without manipulation there was no AVN. It increased to 5% with a Smith-Petersen nail without manipulation and was 9.1%

430

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

with nonoperative treatment, which did, however, include manipulation. Subtrochanteric osteotomy had a 10.9% incidence of AVN, nonoperative treatment without manipulation 12.5%, Smith-Petersen nail plus manipulation 37.5%, and all types of cervical osteotomies 38.1% (100). Realignment operations, though widely practiced for slipped capital femoral epiphysis, still show the possibility of short-term complications, although it is expected that the position closer to the anatomic norm should minimize osteoarthritis later. Convincing and definitive studies in this regard would be helpful. Jerre et al. assessed realignment procedures in 37 hips at an average follow-up of 33.8 years (130). They noted serious short-term complications in 7 of 22 hips treated by subcapital osteotomy, 3 of 11 hips treated by intertrochanteric osteotomy, and 3 of 4 hips treated several years previously by manipulative reduction. They concluded that the natural history of slipped epiphysis was "probably not improved by any of the treatments used in our study. We therefore discourage the use of subcapital and intertrochanteric osteotomy as well as manipulative reduction for the primary treatment of chronic slipped upper femoral epiphysis." The realignment procedures had been performed from 1946 to 1959. That results following severe slippage even with intertrochanteric osteotomy often were far from perfect was demonstrated in a long-term follow-up study of 26 patients by Maussen et al. (181). They reviewed 26 of their own cases of moderate to severe SCFE treated by the intertrochanteric approach. Where slippage was less than 40 ~ in 10 hips, subsequent arthritis was present only in 1. In 16 cases, however, in which slippage exceeded 40 ~ osteoarthrosis was present in 15 of 16 even though correction was adequate. They concluded that the intertrochanteric approach did not prevent degeneration in cases with the most severe slip and recommended only fixation in situ without realignment with correction only of the rotational component during the adolescent years. Follow-up was longer than in most studies with a mean of 9 years and a range from 3 to 26 years. The surgical approach had involved either a valgus derotation osteotomy or a formal Southwick procedure. Maussen et al. used a relatively rigorous grading system for osteoarthrosis of the hip, which was based on radiographic rather than clinical criteria. The inference is clear that many of these patients would go on to symptomatic hips, although at the time of assessment the problem was primarily radiographic. This report is somewhat disconcerting about the long-term results with 40 ~ slippage being the border below which good results could be achieved and above which osteotomy resulting in good correction did not seem to prevent some degree of degenerative change. The report also provided an excellent overview of results in other studies following in situ fixation alone without reorientation of the head and neck region using either metal pins or transphyseal bone grafting. Although a large number of papers are referenced, the gradation involves mild, moderate, and severe cases with the analysis

involving only the percentage classified as excellent or good. Because in all series those with in situ pinning tend to be extensively populated by mild to moderate slips, this analysis may not be valid for the entire spectrum of the disorder. Maussen et al. interpreted the results, however, to indicate that internal fixation by pinning or by bone graft epiphysiodesis without realignment produced long-term results that were good, even in cases with moderate to severe slippage. Review of the intertrochanteric osteotomy assessments from the literature, however, indicated that the question of subsequent arthrosis had not been assessed in detail. The cervical osteotomy did have many instances of poor results documented due to arthrosis and AVN. Maussen et al. present the case, therefore, that even moderate to severe cases of SCFE in the adolescent period should be treated without realignment using only pinning or bone graft epiphysiodesis. Subsequent correction would be performed only for symptomatic states rather than attempting realignment for all to prevent such problems. Differing views continue to be expressed. Another longterm study by Schai et al. assessed 51 patients with unilateral severe SCFE of 30-60 ~ treated by intertrochanteric osteotomy and examined an average of 24 years postsurgery (229). They concluded that 55% showed neither radiographic nor clinical signs of degenerative hip disease, with 28% having moderate disorders and 17% severe osteoarthritis. Intertrochanteric osteotomies were performed with stabilization by either AO blade plates or AO condylar plates. Stabilization of the epiphysis preosteotomy with 2-3 Steinmann pins was important. The principle of reestablishment of hip anatomy indirectly by compensatory means through the intertrochanteric region had been introduced by both Imhauser and Southwick. They concluded that the results were superior to pinning in situ alone. They felt that angles greater than 30 ~ warranted corrective osteotomy. c. Prophylactic Pinning o f Contralateral Side at Initial Presentation The relatively high incidence of bilaterality

noted in patients followed to skeletal maturity plus the fact that most patients at presentation have only a unilateral slip have led to the practice in many centers of pinning the contralateral normal hip at the same time that treatment is undertaken for the slipped epiphysis so as to eliminate immediately any chance of contralateral slipping. The advantages of the approach are the ability of the patient to resume full activity at all levels once both physes have fused with no risk of subsequent slippage. Some of the second side slips are asymptomatic, and the possibility exists that the second slip will not be recognized until late in its evolution such that a moderate or severe deformity is presented for treatment. Because the results are best in those with minimal to no slippage the value of pinning in situ is high. Disadvantages have resulted from prophylactic pinning, however, for two reasons. In some series, such as our own, although 50% of patients had bilateral involvement, 25% of them were bilateral at the time of presentation and only

SECTION I! ~ Slipped Capital Femoral Epiphysis 25% experienced the second side slip over the ensuing period prior to skeletal closure. If all patients in such a group had been pinned prophylactically, 3 of 4 operations would have proven to be unnecessary. Jerre et al. recommended that prophylactic pinning of the contralateral hip should not be standard (128). In reviewing 61 patients treated for unilateral slipped upper femoral epiphysis, there were 14 (23%) who had evidence of bilateral slipping at initial primary review, whereas 11 (18%) subsequently slipped prior to skeletal maturity. They concluded that, if all 61 contralateral hips had been pinned prophylactically at primary admission, 36 of the operations (59%) would have been unnecessary. Jerre et al. thus recommended that radiographs be done every 3-4 months until growth plate closure with only those hips where a slip occurred to be pinned. The reason for X rays was that in only 2 of 25 patients with bilateral involvement (8%) was slipping of the contralateral hip symptomatic. Of greater concern, however, was the finding in some retrospective series that the contralateral asymptomatic hip had a complicated pinning such that damage was caused even though none was present initially. With greater awareness of the dangers of inappropriate pin placement these complications have minimized. In many centers the patients and families are presented with two options. One simply involves pinning in situ at the same time that the primary slip is treated. The other is a recommendation to follow the child closely throughout the remaining years of growth and to intervene surgically only if a slip develops. Instruction is given on the importance of immediate orthopedic assessment for warning signs of early slippage with the development either of limp, however slight, or hip, thigh or knee discomfort. Some also take the further precaution of recommending radiographs every 3-4 months (because some second side slips are asymptomatic) and limiting sporting activities. Not surprisingly, those clinics showing the highest extent of bilaterality are the most supportive of the values of the prophylactic approach. Jensen et al. recommended bilateral pinning at initial treatment in all patients with a SCFE (127). Engelhardt strongly recommends prophylactic treatment and quotes many papers from the past few decades showing the incidence of bilaterality ranging from a low of 19% to a high of 80% in the Billing and Severin study (67). He prefers CT assessment when the child is approaching the age of skeletal maturity because growth plate closure is seen more clearly and thus earlier by CT than by plain radiography. Hagglund recommends prophylactic pinning of the contralateral hip in all cases of SCFE, with the proviso that the technique used should have a low complication rate (97). d. Complications of Slipped Capital Femoral Epiphy. sis Five major complications can be associated with SCFE. Avascular Necrosis: Avascular necrosis almost always is a complication of treatment rather than a complication of the disease itself; it rarely is seen in association with acute, acute on chronic, or chronic slips in continuity even if the slip has

431

proceeded to complete posterior displacement. Closed reduction of the chronic slip never is warranted because any correction gained risks excessive trauma, which tears the posterior vessels. Although a definite change in the relationship of the femoral head to the femoral neck has occurred in SCFE, it is important to realize that this process has been occurring for several weeks to several months and that spontaneous attempts at stabilization posteriorly and medially by fibrous, fibrocartilaginous, or osseus tissue have led to thickening and shortening of the posterior periosteum within which the retinacular vessels are situated. Many also will preclude the use of closed reduction in the acute on chronic case because it is not possible to know when only the acute component has been reduced (Fig. 19A). Avascular necrosis as a complication of SCFE began to gain wide recognition in the late 1920s. Axhausen recognized the entity referred to as "aseptic" necrosis initially to distinguish it from problems widely known to occur after infection. Moore described the histopathology well in a 12-year-old boy who had suffered an injury 1 year previously, which had responded reasonably well to treatment in bed for 4 weeks after which walking was resumed (189). There was vascular fibrous tissue invasion of the marrow spaces, which contained areas of necrotic marrow debris. The dead trabeculae, whose lacunae were empty, were being absorbed and replaced by living bone. Numerous osteoblasts, osteoclasts, and multinuclear giant cells were seen throughout the living connective tissue. The superficial zone of the articular cartilage was normal but the deeper zones were necrotic and being replaced by bone from below. Hall reviewed a series of 173 hips with SCFE noting 27 cases of AVN and 3 of chondrolysis (100). The complication is recognized as occurring due to damage to the blood vessels, which are concentrated on the superolateral and posterior aspects of the neck. Because displacement occurs in a posterior and medial direction, there is a tendency for the vessels to theihead to be stretched. Spontaneous episodes of AVN in the absence of therapy are virtually unknown. The vasculature, though slightly at risk, appears to stretch gradually~jn relation to the slowly evolving chronic slip. When treatment, however, is done in too harsh a fashion damage and subsequent AVN follow. The factors relevant in the consideration of AVN involved delay in diagnosis, the amount of displacement at time of diagnosis, and the type of treatment. Factors predisposing one to AVN include moderate to severe displacement requiring efforts at improving position, a relatively longer time prior to a diagnosis, and treatments that are characterized by manipulation and efforts at reduction be they closed or open. Due to the age of the patient and the relatively small time of growth remaining for remodeling, avascular necrosis in SCFE is a major problem and almost invariably leads to osteoarthritis sometimes in young adult life. AVN always is associated with deformation of the femoral head. Avascular necrosis also can occur following a pinning of the slip. The problem relates to pins that are placed in the

432

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

FIGURE 19 Radiographicexamples of avascular necrosis (A) and chondrolysis (B), the two major severe complications of slipped capital femoral epiphysis, are shown.Both of these can lead to early adult osteoarthritis.

anterosuperior and posterosuperior quadrants in the area in which the lateral epiphyseal vessels enter the femoral head. Though there can be some damage with a pin remaining totally within head and neck bone, major problems could occur if the pin exits the neck and then reenters the head thus causing extensive damage to the vascular leash, which lies on the surface of the neck prior to entering the head between the physeal and articular cartilage regions. This region is very difficult to assess radiographically, and thus the recommendation stressed by Aronson and Loder is to place internal fixation strictly in the central zones of the femoral head and neck to minimize chances of injuring the blood supply (14). Damage also can occur if vessels traverse the neck in the anteroinferior quadrant in which the inferior metaphyseal vessels enter the head and neck and also supply some regions. Lowe summarized both his own experience with early AVN and chondrolysis and other previous assessments (172). In a study of 100 cases of SCFE, he noted 21 hips that developed necrosis of the bony epiphysis or chondrolysis following treatment. There were 6 cases of AVN and 15 of chondrolysis. Although the feeling was widespread that AVN was a complication of treatment, there were some instances, for example, those reports by Moore, that epiphyseal necrosis could occur naturally even in cases in which there was minimal displacement. The AVN could be variable and in some instances recovery was possible. All hips had had major displacement and had been treated. The incidence of AVN was much higher after "successful" reduction,

which implied movement of the head-neck junction with its implication of tearing of the associated vessels. The AVN was analogous to Perthes disease in which the articular cartilage survived. Chondrolysis: Chondrolysis refers to a destruction of the articular cartilage of the femoral head, which is associated radiologically with joint space narrowing and clinically with discomfort, decreased range of motion, and on occasion actual fusion of the joint (123, 173) (Fig. 19B). Waldenstrom pointed to chondrolysis as a complication of SCFE in 1931 (256). He reported on 3 cases of necrosis of the joint cartilage after a slipped epiphysis. All were chronic, having showed symptoms from 6 months to 1 year, and subsequently were treated with closed reduction under anesthesia followed by hip spica immobilization with the involved lower extremity in abduction and internal rotation. Mobilization was continued for 2 months prior to cast removal and rehabilitation. After some months, the joints became more and more stiff and finally all mobility ceased. The earliest finding radiographically after a few months was thinning of the joint cartilage. The joint continued with increased stiffness and often a formal ankylosis. Waldenstrom clearly differentiated the disorder from avascular necrosis of the bone, which involved variable parts of the femoral head bone but never the joint surface alone. He felt that cartilage necrosis was not due to avascularity of the bone especially because the cartilage on the acetabulum also was affected. He felt that the necrosis in his described cases involved only the joint cartilage. He felt that damage to the capsule and syn-

SECTION II ~ Slipped Capital Femoral Epiphysis

ovium was the cause of the joint surface necrosis because that was the source of nutrition. Because all of his cases occurred in hips that had been reduced, which meant manual reduction under anesthesia in a fairly forceful fashion by current standards, he felt it was the tearing of the capsule that induced fibrosis and subsequent poor nutrition. Therefore, he felt that it was the reduction of the epiphyseal slip by relatively vigorous treatments that caused the cartilage necrosis. He then treated 24 subsequent cases without vigorous reduction but rather by slow traction with the patient in bed. Chondrolysis can occur in patients with SCFE who have not undergone treatment. The large majority of instances, however, are associated with treatment, with most of those following internal fixation in which the fixation pin is either into or through the articular cartilage leading to mechanical destruction with movement. The chondrolysis entity, however, also has been reported after immobilization in a hip spica cast and after intertrochanteric osteotomy. Though early papers documented a higher incidence in black patients, more recent and detailed studies have not shown any predisposition in blacks to the disorder. The series of Lowe was unusual in that more cases of chondrolysis were present than of AVN (172). The result with chondrolysis almost invariably was poor due to stiffness of the hip and malposition. Diminution of the joint space on radiographs remained the primary radiologic sign, and clinically there was discomfort and usually markedly diminished ranges of motion. The highest positive associations were with immobilization in plaster spica or prolonged traction greater than 7 weeks. The prolonged immobilization also was a feature in other series, notably those of Moore (190), Jerre (131), Hall (100), and Waldenstrom (256). Although claoadrolysis and AVN can coexist on occasion, in general they ~ e separate entities. Heppenstall et al. noted a 26% incidence of ~chondrolysis in a series of 65 patients (17 involved), but oally 3 of 21 hips with chondrolysis had associated AVN (108). Ingrain et al. reviewed the literature on chondrolysis from Waldenstrom's initial report of 3 cases in 1930 to their own review of 79 cases in 329 hips with SCFE in 1981 (123). Many series were showing very high levels of chondrolysis, the greatest being a 55% incidence in 116 patients by Orofino et al. (203), with other high incidences being Boy~l et al. (16%) (32), Howorth (41%) (123), Maurer and Larsen (28%) (180), TiUema and Golding (40%) (253), Hartman and Gates (16%) (105), Heppenstall et al. (26%) (108), Gage et al. (38%) (84), and Ingram et al. (24%) (123). When all series, including their own, were averaged, the rate of chondrolysis was 19% (332 of 1746). The disorder was best prevented from occurring because treatment was nonspecific and essentially symptomatic, involving bed rest, crutches, traction, and various physical therapy modalities. Antiinflammatory drugs were helpful. The feeling was prevalent that there was a predisposition to the disorder, which involved some immunologic processes, never clearly defined, which led to a synovitis and subsequent degenerative cartilage changes. There was increasing evidence that

433

the disorder was higher with certain types of therapy in particular when the joint was penetrated by internal fixation devices (51%), after open reduction (55%), after cervical osteotomy (37%), after trochanteric osteotomy (59%), and with increased immobilization in cast posttreatment. The disorder was felt to be less frequent with mild slips and acute slips, although this may reflect the more benign treatment methods used and the lack of complications associated with them. Most chondrolysis complications are related to treatment rather than to an inherent predisposition of the patient to the disorder. Ingram et al. reported in detail on radiologic and biopsy assessments of the hip joint from 16 previous cases from the literature plus 23 of their own (123). The radiographic changes involved progressive joint space narrowing, usually superiorly, which often proceeded to entire concentric diminution. There tended to be a generalized demineralization of adjacent femoral and acetabular bone. The situation frequently resolved and stabilized, but on occasion proceeded within a few years to a formal osteoarthritis with marked joint space narrowing, osteophytes, subchondral cysts, and subchondral sclerosis. Persistence of synovitis greater than 6 weeks postsurgery should lead to concern about a developing chondrolysis (262) In one of the larger series 41 cases of chondrolysis were reported by Lance et al. (162). A wide variety of treatments had been used to treat the primary slip, including plaster casts alone (7), closed reduction and plaster (3), intertrochanteric osteotomies (10), open reduction (6), wedge osteotomies of the neck (5), transphyseal bone graft (3), extra articular epiphyseal arrest (4), and pedicle graft (1). Two patients had not been treated but still developed the disorder. One of the major principles of treatment for the chondrolysis is to diminish weight beating using either crutches or, if possible, bed rest with traction. It is important to keep the hip region extended and positioned along the neutral axis to prevent abduction, adduction, flexion, or rotation deformities. Gentle range of motion exercises and occasional use of antiinflammatory drugs usually are helpful. Repair tends to be very slow and symptoms can last several months to years before full or at least clinically useful motion is regained. Many patients, however, do recover fully or close to fully. In those cases not healing well, there is radiographically evident progressive diminution of joint space, osteophyte formation, and a triangular appearance of the head in association with adjacent bone and cartilage collapse. The authors felt that the best treatment was suspension-traction to enhance articular motion and joint lubrication. In those that went on to further joint destruction, a hip arthroplasty often was needed preceded by osteotomies in many instances. The authors felt that only 44% of those affected obtained a good or excellent result with absence of pain, a normal gait, and reasonably good motion. In summary, there are four contributing factors leading to the disorder. (1) Mechanical: Chondrolysis often is associated with pin penetration into the joint. With resumption of

434

CHAPTER 5 9 Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

walking the pin tip scarifies the adjacent articular cartilage. (2) Nutritional: Inability to receive synovial nutrition often is a cause particularly in those patients immobilized in cast for several months. (3) Ischemia: Ischemia of the bone does not necessarily lead to avascular necrosis alone but also to chondrolysis on occasion. (4) Intra-articular pressure: Osteotomies of the neck and intertrochanteric region often tighten the hip joint capsule in the process of the valgus and extension repositioning, which increases the intra-articular pressure and secondarily limits the ability of nutrients to diffuse into the cartilage. Stiffness: Stiffness occurs only as a complication of AVN or chondrolysis. External rotation left untreated diminishes the functional range of motion, although the actual arc of motion numerically is unchanged. Shortening on the Involved Side: Shortening on the involved side must be assessed in a patient with SCFE. If the disorder turns out to be bilateral, clinically significant limb length discrepancy rarely occurs. Even in those patients who have unilateral involvement, there is infrequent clinically significant limb length discrepancy. Any shortening is due to a combination of factors, including treatment, which induces a premature fusion of the proximal femoral capital growth plate, removal of excessive amounts of bone in association with compensatory osteotomy, and loss of length caused by the slippage in particular if this is not corrected by a compensatory osteotomy. Lower extremity length discrepancy, however, rarely is a problem because of the age at which the disorder occurs and the fact that most patients have only mild to moderate deformation. Because only 30% of the growth of the femur and 15% of the growth of the lower extremity occur at the proximal femoral capital epiphysis, and because most patients are 11 years of age or older at the time of the disorder, there rarely is sufficient growth remaining to account for a discrepancy of greater than one-half to threefourths of an inch. Howorth noted that most patients with minimal to moderate slip who had the transphyseal bone pegging procedure almost always had shortening limited to one-fourth to one-half inch and often less (116). The fact that most patients either have only mild to moderate slippage or are anatomically corrected if they have greater displacement minimizes the extent of shortening. A contralateral distal femoral epiphyseal arrest is needed on occasion. Adult Osteoarthritis: There is no question but that some patients with slipped capital femoral epiphysis will develop osteoarthritis in middle to late adult years. Those who suffer either chondrolysis or avascular necrosis may have moderate to severe arthritic symptoms, even in early to mid-adult life. The importance of minimizing and if possible completely eliminating the occurrence of these disorders thus is obvious. What is less certain is the amount of deformation of the head-neck region that contributes to eventual adult degenerative change. The longer range studies indicate that mild slippage is not causative in this regard and that even moderate slippage left untreated other than by fixation in situ has

an extremely low incidence of osteoarthritic change well into the fifth and sixth decades. Most but not all agree that with severe and complete slips corrective osteotomy or even primary open reduction and internal fixation are warranted because of gait and functional considerations. Some will stabilize any slip, regardless of degree, in situ and perform osteotomy later only for troublesome symptoms, which are the external rotation deformity of the lower extremity and sometimes the Trendelenberg gait. A matter of further importance, however, is whether correction must be done at the head-neck junction for optimal long-term results or whether the compensatory osteotomies at the basicervical, intertrochanteric, and even subtrochanteric regions provide sufficient correction to render treatment at those sites preferable. Krahn et al. reviewed 36 patients out of 264 with SCFE who had developed AVN (155). Twenty-four hips were assessed at an average follow-up of 31 years. AVN indeed was a complication with many negative long-term sequelae. Nine of 22 patients had already undergone reconstructive surgery, 4 during adolescence and 5 during adulthood. In the remaining 13 patients (15 hips), there had not yet been any operative intervention but all showed degenerative changes on radiography. This report did not give the treatment methods in the patients without the AVN complication, but certain characteristics were seen in those who had the AVN complication. In the 4 hips that had early deformation requiting reconstructive surgery, 2 had closed reduction with pinning and 2 had open reduction, 1 of whom also had cuneiform osteotomy. In the 5 patients with progressive changes requiring surgery in adulthood, 2 had closed reduction with pinning, 2 had closed reduction with casting, and 1 also had cuneiform osteotomy. In those patients with progressive degenerative changes that had not at that time required surgery, cuneiform osteotomy had been performed in 5 of the 13, closed reduction and pinning in 4, and open reduction in 1. In the series, therefore, the large majority of patients had either closed or open reduction and many with cuneiform osteotomy as well. These three techniques in particular almost always have been implicated in the patient with AVN. e. Long-Term Follow-up Studies Boyer et al. performed a detailed long-term study of 149 hips with SCFE assessed 21-47 years postdiagnosis at a mean of 31 years (33). They assessed treatment methods performed between 1915 and 1952 when many of the surgical interventions were relatively crude by current standards. They confirmed the previous belief that "the mild slip has an excellent prognosis when pinned in situ and if no realignment procedure is attempted." This report, written in 1981, appeared before the work of Waiters and Simon (259), which indicated that some cases of pinning were problematic even if not so recognized by the surgeon. In other words, the long-term results of mild slips pinned in situ without pin penetration would remain excellent, perhaps even somewhat better than the assessment in the Boyer paper. Their conclusion remains intact today, namely, that pinning in situ is safer than any type of reduc-

SECTION II ~ Slipped Capital Femoral Epiphysis

tion or realignment in that it presents fewer technical problems and also requires a minimum of immobilization. Their work also concluded that malunion of the moderate slip should be accepted and pinned in situ. Boyer et al. noted a considerable number of technical complications with the operative procedures then in use. At present, it would be anticipated that surgical correction techniques for the moderate slip would have been improved such that complications from the surgery would be fewer, but there are relatively few definitive studies in this regard. They also concluded that pinning in situ was a safe and reliable method of treating the moderately slipped capital femoral epiphysis. They recognized that severe slips would benefit from realignment procedures, although results in their series were not good, again because of technical surgical problems. They had a high incidence of chondrolysis or AVN. Boyer et al. were, however, able to study 7 patients with severe uncorrected slips and noted the long-term results to be "remarkable." Six of the 7 had good clinical results, marred by some limping and diminution of abduction and internal rotation, but overall showing good, painless hip function. The 7th patient had a poor result. They pointed again to the concern about AVN and chondrolysis in particular after a closed manipulation of chronic SCFE and also the risk of femoral neck osteotomy. In assessing their cases, they used the Southwick method of measurement in some, but where they only had radiographs of the affected hip a more general measurement was used, defining a mild slip as one with displacement of the head on the neck of less than one-third of the diameter of the femoral neck, moderate displacement of one-third to one-half the diameter, and severe displacement of more than one-half the diameter. A long-term study by Ordeberg et al. published in 1984 assessed 49 cases of SCFE without primary treatment 2060 years after diagnosis (202). Assessment was by questionnaire in all and clinical and radiographic examination in 44 of the 49. Patients originally seen between 1910 and 1960 were assessed, involving 57 hips with a mean observation time of 37 years (20-60 years). As a general conclusion, they indicated that "2 of these 49 (still living) cases have required surgery because of secondary arthrosis, far fewer than were found in a comparable group of cases treated with closed reduction and hip spica." A positive Trendelenburg test was noted in approximately one-third, and an even fewer number noted some limping after walking considerable distances. Limb length discrepancy of 2-5 cm was noted in 15. The index of pain correlated with the degree of displacement, and marked functional restrictions were noted mainly in cases with severe slip. Ordeberg et al. felt that deterioration in the later years of life was "comparatively slight and can in part be explained by age." They concluded that clinical observations regarding pain, walking capacity, and range of motion showed a much better function than expected. The cases with severe clinical problems almost invariably were among those with severe slipping, but some even in this

43~

group were doing well. Studies in the patients with a 35-year interval did not support the expected finding of deterioration with time. Ordeberg et al. concluded that "with few ex~ ceptions, coxarthrosis developed only in hips with severe displacement." Carney et al. continued a long-term follow-up study of SCFE from the Iowa group (44). In this study, 155 hips were assessed at a mean follow-up of 41 years after onset of symptoms. Of these, 42% were mild, 32% moderate, and 26% severe. With chronic slips, symptomatic treatment only was used in 25%, a spica cast in 30%, pinning in 24%, and osteotomy in 20%. The study indicated that degenerative joint disease as classified radiographically worsened with increasing severity of the slip and also when reduction or realignment had been done. AVN of 12% and chondrolysis of 16% also were common with increasing severity of the slip and when reduction or realignment had been performed. Their presence almost always indicated a poor result. Deterioration with time was noted to be most marked with increasing severity of the slip. The study of Carney et al. confirmed that the natural history of a malunited slip was mild deterioration related to the severity of the slip, and additional complications. Techniques of realignment, however, were felt to be associated with a risk o f appreciable complications and thus adversely affected the natural history of the disease. Their conclusion, out of step with much of the pediatric orthopedic world, was that "pinning in situprovided the best long-term function and delay of degenerative arthritis with low risk of complications." Their findings supported those of Boyer et al., even though this series was reviewed 12-15 years after the former. Studies reported by Hagglund et al. (98) and Hansson et al. (102) in a southern Sweden SCFE population noted that, in 57 cases with no primary treatment, no AVN (segmental collapse) or chondrolysis was seen. At long-term follow-up 12/53 had "severe arthrosis," but only 1 had hip replacement surgery and clinically "most patients had a good hip function with at least tolerable pain and a good walking capacity." Symptomatic treatment or pinning in situ resulted in high clinical ratings with only 2% of hips needing a secondary reconstruction procedure. When closed reduction and spica casting had been used, the combined rate of AVN and chondrolysis was 13% with reconstructive procedures sought in 35% of the hips. Cervical osteotomy had a combined rate of AVN and chondrolysis of 30% with reconstructive procedures needed in 15% (96). The group recommended pinning in situ for treatment because either open or closed reduction clearly increased problems. In their Series for 39 hips in which the slips had been reduced, osteonecrosis developed in 12 (31%) and chondrolysis in 11 (28%). In 116 hips that had not been reduced, AVN occurred in 7 (6%) and chondrolysis in 14 (12%). They concluded that the natural history of the malunited slip was one of mild deterioration related both to the severity of the slip and to complications of treatment. Realignment, however, risked

436

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

"substantial complications and adversely affects the natural course of the disease." They supported pinning in situ regardless of the severity of the slip as providing the best longterm function associated with a low risk of complications and, thus, the most effective method of delaying the development of degenerative arthritis. Hagglund et al. studied long-term results after nailing in 204 slipped epiphyses evaluated at an average of 28 years postprocedure (98). The only early complication noted was segmental collapse of the femoral head seen in 4 of 179 hips nailed in situ and 4 of 25 hips operated after reduction. Subsequent osteoarthritis was twice as frequent after reduction approaches than after fixation in situ. They concluded that "nailing or pinning in situ is the method of choice when possible, regardless of the degree of slipping." Prophylactic pinning of the contralateral hip was indicated because!of tlaeir high incidence of bilaterality. The second longzterm study by this group of femoral neck osteotomy assessed 33 patients with a severe slipped capital femoral epiphysis treated primarily with wedge osteotomy of the femoral neck (96). The patients were assessed at an average of 28 years postsurgery. Segmental collapse and or chondrolysis developed in 10 of the 33 patients. In 9 of these available for reassessment, all had severe arthritis with poor function. The conclusion of Hagglund et al. was that the value of realignment by wedge osteotomy of the femoral neck was questionable. These long-term studies are essential to help determine which adolescent situations require compensatory osteotomy or open reduction with cuneiform osteotomy and whieh can be treated effectively by pinning in situ alone.

K. Coxa Vara Due to Other Acquired Causes These entities are discussed under their more specific sections in Chapter 3, Developmental Dysplasia of the Hip, Chapter 4, Legg-Calve-Perthes Disease, Chapter 9, Skeletal Dysplasias, and Chapter 10.

III. D E V E L O P M E N T A L OF THE FEMUR

ABNORMALITIES

\

1 FIGURE 20 Illustrationsfrom the work of Drehmann (61) show good awareness of the underlying pathoanatomy. (Compare with Figure 2 lB.) groupings within these general terms lend themselves to specific treatment approaches. Part of the confusion in terminology comes from the fact that these four categorizations can occur as isolated deformities in some, whereas in others two or even three of the descriptive abnormalities can be found in the same femur. Some authors have concentrated primarily on the proximal femoral abnormalities, some have focused on the coxa vara, which is a feature of the relatively milder proximal femoral focal deficiency cases but also exists in acquired disorders and in isolated fashion as infantile coxa vara, and some have considered congenital short femur as a specific entity, although many femurs with this abnormality also have a proximal coxa vara. Distal femoral developmental abnormalities are described infrequently but are present fairly often and can lead to deformities that are clinically symptomatic.

A. Terminology Developmental abnormalities of the femur comprise an extremely wide spectrum of disorders from complete absence of the femur to its presence with a normal structure and only a mild degree of shortening. Considerable attempts at classification have been made, but the disorders are so variable that no single universal approach has been acceptable to all. The overall pattern of categorization, however, is reasonably well-defined. One approach is to divide the abnormalities into proximal femoral focal deficiencies, coxa vara, congenital short femur, and distal femoral abnormalities because

B. Proximal Femoral Focal Deficiency Wmually all of the severe developmental abnormalities of th~ femur are concentrated at its proximal end (5, 8, 79, 83, 86, 134, 205). It took several decades before reasonably accurate classifications of the nature and type of the variable deformities were outlined. These truly congenital disorders often were discussed with and confused with infantile coxa vara, which now is recognized as an isolated disorder of postnatal onset in the large majority of cases. Drehmann (61) (Fig. 20)'and Golding (88) showed examples of coxa vara,

SECTION III ~ Developmental Abnormalities of the Femur which included but did not differentiate congenital and infantile varieties. Although the entity now is known as proximal femoral focal deficiency (PFFD), the several variants were recognized early under the term congenital coxa vara (61, 106, 182, 199). The deformities are quite variable in extent, and several classifications have been presented in an effort to better understand the entity and categorize it for treatment. Several classifications are presented here because each provides information on the types of deformity that have been observed. Some are strictly pathoanatomic, whereas others categorize disorders based on varying treatment approaches.

1. CLASSIFICATIONS a. Aitken Classification Aitken divided the entity into four classes, A-D. Class A: The head of the femur is present along with an adequate acetabulum and a very short femoral segment. Initially there is no bony connection noted between the femoral segment and the head of the femur, but at skeletal maturity bone continuity is seen although in most instances there is a subtrochanteric pseudarthrosis. Class B: The femoral head is present and the acetabulum is adequate, but the femoral shaft is short and deformed with a small bony tuft on its proximal end and no bone or cartilage continuity between the shaft and head and neck segment at any time. Class C: The acetabulum is severely dysplastic and there is neither a bone nor cartilage model of the femoral head. The shaft of the femur is short with an ossified tuft at the proximal end. Class D: Both the acetabulum and the femoral head are completely absent, there is a deformed shortened femoral shaft, and there is no proximal tufting of the shaft of the femur. This class of deformity frequently is bilateral. b. Amstutz Classification Amstutz defined proximal femoral focal deficiency as "the absence of some quality or characteristic of completeness of the proximal femur, including stunting or shortening of the entire femur" (8). In his study, and that of Aitken, a portion of the distal femur always was present, even if only represented by a misshapen ossicle. Amstutz also brought attention to the fact that a coxa vara deformity in addition to shortening was characteristic of many cases of proximal femoral focal deficiency. His classification defined five morphological groups identifiable radiographically at birth and also included developmental changes with time. Congenital bowed femur with coxa vara, which had not previously been included with PFFD entities, was represented (Figs. 21A-C). Type l: Congenital bowed femur with coxa vara. The anterolateral bowing of the femoral shaft, most apparent in the proximal half, is associated with medial femoral cortical sclerosis. The capital femoral epiphysis ossifies, and because it is well-positioned in the acetabulum, it is not associated

437

with acetabular dysplasia. There may, however, be a delay in appearance of the secondary ossification center. Type 2: There is a subtrochanteric pseudarthrosis with lack of bone continuity between the head-neck-trochanteric region and the rest of the shaft. This is characterized clinically by a progressive varus of the proximal femur and delayed development of the head in particular. Two possible patterns of development follow. In one, there is progressive varus and lack of union of the two fragments, whereas in the other, there is either complete bone repair or a rigid pseudarthrosis with close apposition of the bone fragments. Type 3: The hip region is formed in that there is a cartilaginous femoral head present and the acetabulum has no evidence of dysplasia. There is no initial radiologically defined bone continuity between the head, neck, and trochanteric region of the femur and the shaft. The shaft is shortened and has a variable degree of proximal bulbousness. Ossification of the femoral capital epiphysis often is markedly delayed. Subsequent development can be variable leading to four subgroups. In type 3A, there is union between the proximal and distal fragments with a coxa vara persisting. In type 3B, there also is union but the coxa vara is much more marked and the greater trochanteric epiphysis greatly overgrows the superior surface of the head. In type 3C, there is only a fibrocartilaginous union between the two fragments with marked proximal displacement of the distal fragment. In type 3D, no continuity is achieved. Type 4: The hip joint components are formed with an acetabulum and capital femoral epiphysis present in all such that acetabular dysplasia is not seen or is only minimal, although ossification of the capital femoral epiphysis may occur as late as 2.5 years of age. The proximal end of the distal femoral shaft tapers sharply, almost to a point, which differentiates it from type 3, which has a bulbous proximal end. The tapering represents an unfavorable prognostic sign as union of the two fragments never occurs and is followed by proximal migration of the distal fragment. The acetabulum eventually becomes dysplastic because of the failure of union. Type 5: Dysgenesis is severe such that none of the normal precursor hip components involving either the capital femoral epiphysis or the acetabulum develop. The classification of Amstutz has been widely used. Panting and Williams reviewed their cases in relation to his approach (204). They also pointed out, as had others, that radiographic evidence of an acetabulum in the first year of life indicated the presence of a well-located femoral head and neck, even if they were not seen due to delayed ossification of the secondary center. Additional classifications for proximal femoral focal deficiency: Over the course of the next several years, the PFFD syndrome became better recognized and treatment for the most part was intensive. Efforts were made to refine the classifications of Aitken and Amstutz to define disorders into a more practical and clinically oriented mode in relation to

438

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

A

TYPE IA

EARLY

TY

9 Q E II EARL TYPE II LATE

/

TYPE III EARLY TYPE IliA

TYPE III

TYPE I

TYPE III D LATE

C

F I G U R E 21

(A-C) The proximal femoral focal deficiency classification of Amstutz (8) is shown.

specific diagnosis, projected outcomes, and treatment approaches. These multiple approaches also reflect the findings that virtually no two cases are identical because the disorder represents a true spectrum in terms of extent and that with time the radiographic appearance changes in ways that also can differ depending on whether there is worsening or improvement, which can be spontaneous or based on treatment. c. Fixsen and Lloyd.Roberts Radiologic criteria were described by Fixsen and Lloyd-Roberts, which would allow one to determine whether a proximal femoral focal dysplasia was evolving to a stable or unstable state (79). In many children with this disorder, radiographs showed a short femur with a seemingly absent proximal one-third of the femoral shaft, head, neck, and trochanter. Clinical findings, however, often demonstrated a stable hip in association with the short-

ening and the femur flexed and externally rotated. The stability implied continuity between the femoral head and the proximal end of the shortened femoral shaft such that the intervening radiolucent area would be occupied by a cartilage model in which ossification was delayed. Two possible outcomes were seen clinically. In the favorable state, the cartilage model of the proximal femur gradually ossified with maintenance of stability, and eventually an entire femur was seen although the shaft remained short. The unfavorable outcome was associated with the development of one or more pseudarthroses at the osteocartilaginous junction or within the cartilaginous model so the continuity between the hip and femoral shaft was lost. The resulting instability led to proximal migration of the femoral shaft in relation to the head and neck because a breakdown of the pseudarthrosis pre-

SECTION III ~ Developmental Abnormalities of the Femur C

EARLY

~

~1"

(3 E V EARL

TYPE V LATE

FIGURE 21 (continued)

disposed to further upward displacement of the femoral shaft, although the proximal segment remained in the acetabulum. Fixsen and Lloyd-Roberts' criteria for distinguishing those that would proceed to spontaneous healing from those that would not were based on a study of 30 hips in 25 patients. All were observed until final definition of a stable or unstable state had been established. (1) In the stable hips, the shaft length was greater than one-half of the normal side in 5 of 6 cases, whereas the shaft length was less than one-half the normal side in 6 stable and 8 unstable cases. When the acetabulum resembled the normal, the femoral head always was present although ossification might be delayed. If there was no acetabulum, the head was absent. If there was acetabular dysplasia, then the head might well dislocate with time. In general, the shorter the ossified part of the shaft, the less the likelihood of spontaneous healing. The distance of the proximal end of the ossified shaft from the acetabulum was an important factor. When the distance was greater than the normal side, the hip was ultimately stable in 10 and unstable in 5, and when the distance was equal to or less than the normal side, the hip was stable in 1 and unstable in 4. (2) All unstable hips showed progressive migration of the femoral shaft upward, which indicated an impending dissolution of the pseudarthrosis. When the proximal end of the ossified

439

shaft was bulbous, all 12 hips were stable. When there was a tuft or cap, only 3 were stable with 10 unstable, and when there was a tapered point, none were stable and 5 were unstable. (3) The proximal end of the ossified shaft either was blunt and irregular or pointed. Sclerosis of the proximal shaft was related either to the site of angulation or to a pseudarthrosis. In those hips that became stable, the sclerosis essentially was in the mid-shaft region, well below the proximal end and associated with angulation. In the unstable defects, the sclerosis almost always was immediately distal to the site of the pseudarthrosis or angulation and had the appearance of an inverted " V " and was more proximal. In a retrospective study, several of the classifications in this section were reviewed, and it was the feeling of Sanpera et al. (228) that the radiologic parameters described by Fixsen and LloydRoberts were the most reliable factors for predicting future outcome of the femur from the time of birth onward. d. Lange, Schoenecker, and Baker Lange et al. classified their 42 patients into four categories (163). They often had difficulty assigning some patients to the specified Aitken or Amstutz group and formulated their categorization to conform to treatment approaches. In the less severely involved cases, these proximal shaft signs and stable-unstable concepts of Fixsen and Lloyd-Roberts were incorporated. Class 1: There is a coxa vara with the apex of angulation at the subtrochanteric region and a mild to moderate shortening of the femur associated with anterolateral bowing. Class 2: The acetabulum is present on the earliest films, but there is delayed (8-18 months) ossification of the femoral head. The proximal femoral shaft is displaced laterally, but the head and shaft are joined by a cartilage bridge. This bridge tissue eventually will ossify, although a pseudarthrosis might be present. The continuity is evident by stability with passive motion on clinical examination. Class 3: The acetabulum is intact but there also is late ossification (12-18 months)of the femoral head. The femoral shaft is separated from the head radiographically, but neither bone nor cartilage bridges are seen either initially or with time. There is detectable instability with independent motion between the shaft and the head. Class 4: There is severe proximal bone and cartilage deficiency with absence of the acetabulum, femoral head, and most of the shaft. At the distal end, there often is only a small segment of bone. On occasion, this bone fragment may be completely missing or the distal femoral bone is fused to the proximal tibial epiphyseal secondary ossification center, forming one tissue mass without a joint. e. Gillespie and Torode The patients were divided by Gillespie and Torode into two groups, which could be differentiated on clinical grounds and led to markedly different treatment options (86). In group 1, the congenital hypoplastic femur had sufficient development that the hip and knee could be made functional and lower extremity length equalization in many would be possible. In group 2, there was a proximal femoral focal deficiency in which the hip joint

440

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

never was normal and the knee joint always was useless. The lower extremity length discrepancies in group 1 rarely were as great as those in the more severely involved group 2. In addition, the flexion and abduction deformities at the hip in group 1 were less marked. In group 1, the radiographic characteristics showed the femur to be 40-60% of normal length with proximal to distal continuity, coxa vara usually in the subtrochanteric region, lateral bowing of the shaft, and a hypoplastic knee. In group 2, the femur was markedly shortened and a deficiency in bone always was noted. The head and neck often were absent, the shaft was markedly deficient, and the knee was hypoplastic. f. Kalamchi, Cowell, and Kim In the approach of Kalamchi et al., type 1 included patients with congenital shortening of the femur with a normal hip joint and no other femoral defects (134). Type 2 included those with a congenital short femur and coxa vara, generally with bowing and medial proximal femoral shaft sclerosis. The acetabulum was normal and the femoral head was well-positioned, although the secondary ossification center often formed late. In type 3, the proximal femur was deficient but the acetabulum was normal, indicating presence of the femoral head. This also tended to ossify late. There was shortening of the shaft and sclerosis of the proximal and middle one-thirds. If the proximal metaphysis of the affected femur was dysplastic, showing marked broadening and irregularity, the limb was designated type 3. Two patterns occurred with growth. In one, the defect went on to ossify in various degrees of varus, whereas in the other pattern the defect did not ossify, leading to a pseudarthrosis and lack of continuity between the two segments. Type 4 included limbs with no acetabulum and no femoral head and, thus, with an unstable distal segment. The distal segment was short and in the form of a tapered spike. Type 5 included limbs with no hip joint and no evidence of a separate femoral segment except perhaps a small distal fragment of bone, which usually was adjacent to the tibia. g. Haminishi Haminishi reviewed a large series of 70 patients with 91 congenital short femurs encompassing the entire spectrum of femoral developmental abnormalities (101). He compared the structural differences caused by the drug thalidomide with those of spontaneous occurrence and noted no essential anatomic difference between the two groups, although the whole complex of abnormalities differed in that the thalidomide group tended to show radius anomalies whereas the non-thalidomide group had femur-fibula-ulna anomalies. In relation to the spontaneous type, there were 67 affected femurs in 56 patients. His all-inclusive classification divides the entity into five types with varying subtypes. Type I. Simple hypoplasia of the femur: (a) normal shape; (b) slightly angulated shaft and cortical thickening. Type II. Short femur with angular shaft: (c) marked lateral angulation and cortical thickening resulting from transverse subtrochanteric ossification defect; (d) decreased neck-shaft angle. Type III. Short femur with coxa vara: (e) type IIIa (straight shaft), stable coxa vara with marked cortical thick-

ening at the lesser trochanter; (f) type IIIb (angulated shaft), progressive coxa vara with thickened cortex. Type IV. Absent or defective proximal femur: (g) absent or fibrous neck and trochanter; migration of the upper shaft, short shaft-head distance, and diaphyseal transverse ossification defect; (h) absent neck and trochanter and small femoral head connecting directly to the tapered shaft; (i) all the proximal femur is absent. Type V. Absent or rudimentary femur: (j) rudimentary distal femur, which is ossified later. h. Pappas A detailed study primarily based on patients followed longitudinally was published by Pappas, who developed a nine-class categorization (205). Pappas defined the percent of femoral shortening in each of the nine classes, detailed the femoral and pelvic abnormalities, assessed associated abnormalities of the tibia, fibula, patella, and feet, and defined treatment objectives. The large number of patients available for this study demonstrated a continuum of abnormalities. Class I refers to the situation in which the femur is entirely absent and the acetabular region of the pelvis is markedly hypoplastic. Class II: the proximal 75% of the femur is absent. Class III: there is no bony connection between the femoral shaft and head although the femoral head, which has delayed ossification, is present in the acetabulum. Class IV: the femur is present to approximately onehalf its length, but the proximal abnormalities show the femoral head in the acetabulum with the head and shaft joined by irregular calcification in a fibrocartilaginous matrix. It is these four disorders that are generally referred to as proximal femoral focal deficiency. In class V, the femur diaphysis and distal end are incompletely ossified and hypoplastic. In class VI, the proximal two-thirds of the femur is perfectly normal and the hypoplasia is in the distal one-third with an irregular distal femoral region and no evident distal epiphysis. Classes V and VI are examples of what could be described as distal femoral focal deficiency. Class VII is congenital coxa vara with a hypoplastic femur, which is shortened and somewhat bowed and also demonstrates lateral femoral condylar deficiency. Class VIII is seen infrequently but involves a proximal femoral coxa valga, a hypoplastic femur, and abnormality of the distal femoral condyles, with the lateral condyle being somewhat flattened. Most would include congenital short femur in this category, which perhaps most represents class VIII, although it characteristically has anterolateral bowing, which Pappas does not demonstrate. The class IX femur is essentially normal and might be defined by others as having only shortness referred to as hemiatrophy or anisomelia. Pappas also demonstrates the frequently seen underdevelopment of the lateral femoral condyle predisposing one to both a valgus deformity at the knee and a tendency toward lateral patellar subluxation. 2. CLINICAL CHARACTERISTICS The more severe abnormalities are recognizable at birth with a markedly shortened thigh, which is bulky with the hip flexed and abducted and the extremity externally rotated.

SECTION III ~ Developmental Abnormalities of the Femur From one-half to two-thirds of patients also have associated musculoskeletal abnormalities. The most common associated irregularity is ipsilateral fibular hemimelia, but a large number of variable abnormalities of either upper or lower extremities and also of the axial regions have been described (38, 101). One of the most detailed studies of associated congenital deformities was reported by Hamanishi in 70 patients with 91 affected femurs. No genetic basis, predisposing factors, or specific causal factors (other than thalidomide) have been identified. 3. PATHOANATOMICFINDINGS a. Soft Tissue Anatomy Magnetic resonance imaging studies have proven useful in assessing the soft tissue anatomy in association with PFFD. A detailed study by Pirani et al. of Aitken types A, B, C, and D in 6 patients with 7 affected hips demonstrated that all muscles appeared to be present (212). Most of the muscles were smaller than their normal counterparts, including the gluteus maximus, medius, and minimis complex, quadriceps, adductor brevis and longus, adductor magnus, pectineus, semimembranosus, semitenclinosus, and biceps femoris. The sartorius muscle, however, was hypertrophied, indicating a possible causative factor in the characteristic flexion, abduction, and external rotation deformity of the hip. The obturator externus muscle was elongated and remained muscular almost to its insertion. In type A PFFD, the short external rotators of the hip were larger than normal in terms of diameter and inserted into the posteromedial aspect of the greater trochanter. The abductors were smaller than normal. Hip musculature appeared to be positioned to contribute to both the normal and abnormal ranges of motion seen and clearly played a role in providing hip stability. b. Gross and Histopathologie Findings in Proximal Femoral Focal Deficiency One detailed study from a 21week-old fetus with unilateral PFFD has been reported by Boden et al. (29). Their examination indicated a unilateral PFFD with normal skeletal development of the contralateral limb and the rest of the skeleton. The radiographic appearance and the specimen photographs including the hip joint indicate that the abnormal side was developing as an Aitken type A deformity, sometimes referred to as congenital short femur with coxa vara. The femoral head was well-located in the acetabulum, there was complete tissue continuity between the head-neck trochanteric regions and the shaft, and a proximal metaphyseal-diaphyseal varus was seen with the tip of the greater trochanter lying higher than the most superior portion of the femoral head. Histologic sections clearly revealed structural abnormalities in the proximal part of the affected femur with failure of formation of a normal growth plate. On the normal uninvolved side, cartilage canals in the femoral epiphysis were noted as was the orderly array of cell changes in the physis progressing from the reserve to proliferating to hypertrophic zones, following which normal metaphyseal bone formation was seen. On the involved side, the patterning of the epiphyseal cartilage appeared normal as

441

did the vascular canals. There was slightly less organization of the proliferating zone of chondrocytes, although flattening of the cells was seen. The characteristic column formation did not occur. Most striking was the markedly decreased size of the hypertrophic zone, the lack of linear columnization of the hypertrophic zone, and the irregular formation of bone in the metaphysis due to the lack of an appropriate cartilage scaffold upon which the bone could be synthesized. The mineralization front of the hypertrophic zone was altered, and there was abnormal persistence of glycogen in chondrocytes deep into the growth zone. Many-of the clinical reports contain gross descriptions of the proximal femoral region, particularly in relation to those patients who have undergone surgical exploration of the proximal femoral bowing, either to place bone graft in the region to enhance bone development or to treat the pseudarthroses that frequently develop. Thus, there is good clinical correlation based on gross anatomic exam in the living in many instances. In those proximal femurs that are defined as clinically stable, the femoral head, neck, and trochanteric regions are present and continuous with the developing diaphysis, although tissue continuity often is maintained by persistence of the cartilage model in which ossification has been delayed. Another characteristic feature of the disorder, even when the femoral head is round and appropriately positioned in the acetabulum, is the fact that there is delayed vascular invasion of the cartilage model of the head, leading to a delay in formation of the secondary ossification center. This is described as appearing anywhere from 6 to 18 months, whereas in the vast majority of patients, the secondary center is present by 6 months and begins forming as early as 3-4 months. These clinical and radiographic observations correlate well with the 21-week fetal study by Boden et al. of an abnormal histologic appearance of the physeal region of the developing bone. In many instances, therefore, the cartilage model of the developing femur has formed, but in the proximal half there is inappropriate bone formation in relation to the proximal metaphysis and also to the secondary ossification center. The structural appearance of the physeal cartilage can be markedly abnormal with diminution of size and poor organization of the proliferating zone of the cartilage, which subsequently leads to a markedly shortened and disorganized hypertrophic cell zone. Even though there is some vascular invasion from the diaphyseal side, the abnormal structure of the hypertrophic zone and its poor mineralization contribute to diminished vascular invasion. The causes of these abnormalities remain unclear, but structural studies begin to show which part of the developmental sequence is interrupted. 4. TREATMENT OPTIONS Management in the proximal femoral focal deficiency group addresses six possible concerns: (1) the lower extremity length discrepancies; (2) the establishment of bony continuity of the femur by correcting any proximal pseudarthrosis;

442

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

(3) establishment of the most stable proximal femoral-acetabular relationship; (4) weakness of the proximal musculature; (5) correction of rotational or angular deformities of the femur; and (6) the effect on the extremity of other ipsilateral deformities primarily fibular hemimelia. Management of length discrepancies will be described in detail in Chapter 9. Treatment options have been detailed in several studies (5, 8, 10, 11, 38, 79, 86, 134, 154, 226). In the more severe variants of the PFFD deformity, such as the Aitken types C and D, there is little to no role for surgical intervention and treatment is directed to prosthetic management. In the less severe A and B types, considerations involve the presence or absence of bone continuity in the proximal femur between the head and neck segment, which generally is in good relationship to the acetabulum and the distal two-thirds of the shaft. In the newborn and in the first few years of life there may be no evident bony continuity between proximal and distal segments, although clinical testing may indicate stability. In these instances there is structural continuity mediated by cartilaginous tissue, which will eventually ossify. Prosthetic management and close observation are used initially to minimize proximal migration of the femoral shaft, which is indicative of a developing pseudarthrosis at the nonossified regions of the proximal femur. If the bone remains straight, it is indicative of cartilage continuity and stability and observation alone is sufficient. With time ossification of the cartilage fragment occurs. If a pseudarthrosis is developing, operative intervention with bone graft application is warranted. Location of the femoral head in a well-shaped acetabulum is essential to hip stability as with any hip abnormality. Attention to reduction of the femoral head deep into the acetabulum follows accepted principles of proximal femoral and acetabular management. On occasion coxa vara must be corrected with a proximal femoral valgus osteotomy, but only if there is sufficient depth of the acetabulum.

C. Congenital Short Femur Congenital short femur is a relatively common cause of limb length discrepancy and frequently is mentioned as an independent entity in relation to causes of lower extremity length discrepancy. It fits, however, into the spectrum of femoral developmental disorders. In the simplest variants of congenital short femur, the involved bone is shorter than that on the opposite side and demonstrates anterolateral bowing primarily in the proximal one-third of the shaft, although the headneck trochanter regions are normal and there is no coxa vara. This entity was clearly pointed out by Ring (226). In addition to being slightly bowed, the shaft of the femur shows both medial and lateral cortical sclerosis, and the clinical presentation is characterized by a slight hip flexion contracture and external rotation positioning of the femur with internal rotation markedly limited sometimes with no movement beyond neutral.

The second variant of congenital short femur is that associated with a congenital coxa vara in addition to the anterolateral bowing of the proximal diaphyseal region. This disorder often is mentioned as an Aitken type A variant of the PFFD syndrome. The coxa vara is relatively mild and only infrequently is it associated with the triangular fragment of the neck, which is far more characteristic of infantile coxa vara. A congenital short femur generally leads to shortness of the lower extremity greater than 5 cm, although most are manageable by femoral lengthening. The rate of increase in the discrepancy with time is the same throughout growth such that, if the femur is 84% of the length of the normal side at 1 year of age, it almost invariably will be 84% of the length of the opposite side at skeletal maturity. Planning for surgical intervention thus is straightforward in terms of timing. Contralateral distal femoral epiphyseal arrest is performed for discrepancies projected to be less than 4-5 cm, with ipsilateral femoral lengthening done for discrepancies greater than 4-5 cm. Management is detailed in Chapter 9.

D. Distal Femoral Developmental Abnormalities Although the major and far more severe developmental abnormalities of the femur occur in the proximal one-half, there is a small subset of associated distal femoral developmental abnormalities that can be troublesome clinically. Even in the most severe proximal abnormalities, there is almost always some presence of the distal femur, even though it is as small as a structureless cartilage mass and ossicle in the most severe forms of PFFD. Distal femoral developmental abnormalities can be present where there is an intact femur; the closer to normal the proximal half of the femur, the less dramatic the distal changes, but in those instances in which such changes are present, they may reach clinical significance. The distal femoral developmental abnormalities tend to affect the distal femoral epiphysis in two ways. One leads to slight underdevelopment of the lateral condylar region of the femur such that it is not as long as that on the medial side, tending to a valgus deformation, and the other leads to a less well-developed anterior-lateral segment, causing malformation of the patellar groove region and thus predisposing one to lateral patellar subluxation and full dislocation of the patella and quadriceps mechanism. The function of the quadriceps mechanism can be worsened further by a slight tendency to external rotation positioning of the distal femur, even in those situations in which abnormalities of the proximal one-half or two-thirds predominate. Abnormalities of the distal femur also are seen in more serious developmental disorders of the leg but can be associated with either fibular hemimelia or tibial hemimelia. Tsou has described a rare congenital abnormality of the distal femoral epiphysis in which the proximal part of the femur, including the shaft, otherwise was normal (254).

SECTION IV ~ Infantile Coxa Vara

FIGURE 22 An established infantile coxa vara lesion is shown. The triangular fragment on the inferior surface of the neck is characteristic. Note the vertical radiolucent defect. Hilgenreiner's epiphyseal angle is increased. [Reprinted from Weinstein et al. (1984), J. Pediatr. Orthop. 4: 70-77, 9 Lippincott Williams& Wilkins, with permission.] IV. I N F A N T I L E C O X A V A R A

A. Terminology Infantile coxa vara is a condition in which deformity is isolated to the proximal femur without abnormalities of the middle or distal femur or the rest of the skeleton. It is characterized by a pathognomonic bony discontinuity of the inner surface of the femoral neck, appearing as a triangular-shaped bone fragment with its base along the inferior surface of the neck bordered by the physis medially and superiorly and a vertical radiolucent fissure laterally. The disorder is limited almost exclusively to the physis and neck, with the femoral head epiphysis, greater trochanter, and acetabulum otherwise normal initially (Fig. 22). It is not congenital in that many patients have been described in whom the hip radiographs in the first year of life were normal but coxa vara deformity developed shortly afterward. There is relatively little shortening of the femur in infantile coxa vara, unlike the situation in truly congenital coxa vara associated with dysgenesis of the entire proximal femur. The limb shortening with infantile coxa vara is caused by the varus deformation and the decrease in length of the femoral head and neck region only.

B. Clinical and Radiographic Presentation of Infantile Coxa Vara It is well-established that, in some patients with infantile coxa vara, radiographs taken during the first year of life can be normal with the radiographic appearance changing during the second year of life and beyond. Infantile coxa vara generally presents around 2 years of age with a painless awkward gait with a waddling component, which may be unilat-

443

eral or bilateral. In bilateral cases lumbar lordosis is present. The gait worsens progressively over the next few years. The Trendelenburg test is positive and hip abduction and internal rotation are limited. The disorder has an equal sex incidence and is bilateral in 33-50% of patients (66, 176, 208, 220). The children do not experience discomfort but tend to fatigue readily. On radiographs, the physis is more vertical than normal, and it appears inferiorly and medially to be branched like an inverted Y. This vertical fissure is the characteristic radiographic feature of infantile developmental coxa vara. It rarely if ever is seen in the first 2 years of life. Some of the pathological findings are felt to be consistent with trauma, but there have been only rare instances describing recognizable trauma. Genetic factors have been implicated because several cases of infantile coxa vara have been reported in siblings or in twins. Fisher and Waskowitz documented 16 reports of familial developmental coxa vara in the literature and added their own (78). The initial clear description of what we now refer to as infantile coxa vara was by Hoffa in 1905 (114). The disorder also had been described by Hofmeister under the name of coxa vara adducta (115). Elmslie described infantile coxa vara with great clarity (66). In his report the sex incidence was the same, and 8 of the 20 cases were bilateral. A waddling gait had been noted when the child first began to walk in one-half of the patients, and most of the other symptoms had become clearly evident by the ages of 6 - 8 years. The greatest limitation of hip motion was in abduction, which often was abolished completely. Lordosis was extremely marked in particular with bilateral involvement. At 5 years of age, there is "a downward displacement of the head of the femur carrying with it the adjoining portion of the base of the neck." The slippage of the head and its retention in apposition with the base of the femoral neck by the periosteal coveting occurred as a result of sudden accident or by a process of gradual slipping. With displacement of the head, the downward pressure of the body weight transmitted to the upper margin of the acetabulum becomes directly transverse to the line of the infraction which has been produced and nearly transverse to the epiphyseal line. With increasing age, the displacement of the head would increase owing to this mechanical factor. Moreover, the neck of the femur which develops from this epiphyseal line chiefly after the fifth year, will be developing in the wrong direction, owing to the line of growth now being nearly vertical instead of horizontal. Sequential X rays showed that (1) displacement of the head tends to increase with age, (2) the neck does develop pointing transversely or even downward, and (3) the infraction in the upper border of the neck frequently is visible. Elmslie felt that direct trauma could cause the lesion but that birth injury, although possible, was difficult to prove. The disorder was not tickets because depression of the femoral neck is not common as a result of infantile tickets. It is now widely considered that the disorder develops initially in the first and second years of life and that the

444

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

femoral neck is short and largely cartilaginous. Shortening can occur in unilateral cases with one report referring to about a 4-cm discrepancy. In virtually all series, the infantile coxa vara cases are isolated abnormalities with no other femoral or skeletal irregularities seen. In perhaps 15% of cases, this appearance is seen with skeletal dysplasias, generally cleidocranial dysostosis, or spondyloepiphyseal dysplasia congenita. Blockey clearly has recorded his impressions that infantile coxa vara is an acquired lesion (28). He notes that "half or more of the femoral neck is naked and the shaft looks laterally rotated. This appearance could be produced by the capital epiphysis with its epiphyseal plate and triangular fragment of metaphysis sliding distally until the superior aspect of the head abuts on the inferior surface of the proximal part of the neck. The line of cleavage remains visible and the displacement is with the femoral neck." Blockey thus likens infantile coxa vara to distal movement of the head as though an infantile slipped epiphysis with a triangular fragment of bone had occurred. He feels this is due either to unappreciated extrinsic trauma in the early months of life or to slipping through the metaphyseal side of the physis through bone that is somewhat softened or pathologic. He thus feels that "infantile coxa vara is likely to be due to distal movement of the head fragment relative to the shaft and neck." Infantile coxa vara usually presents clinically in the second and third years of life.

C. Pathoanatomy of Infantile Coxa Vara The earliest pathoanatomic study was that of Hoffa in 1905 in which he described the gross and histologic findings in both proximal femurs in a 4-year-old boy who had undergone resection of the femoral head-neck region (114). Gross anatomic drawings show the coxa vara deformity in his two cases. He also described other cases. He found no evidence of rickets or trauma. The growth region of the capital femoral epiphysis had abnormal signs of endochondral growth with a lack of the proliferating chondrocyte zone. In the cartilage, the typical proliferating zone was lacking, the trabeculae did not exhibit any growth by apposition, and the bone marrow did not resemble that of a normal child but showed regressive changes. This region was characterized by vascular invasion. Hoffa illustrated a femoral head and neck specimen cut in the coronal plane, which shows the coxa vara deformity, the secondary ossification center in its normal position, and early bone formation at the medial and inferior surface of the physeal and epiphyseal regions. Helbing drew a similar histologic conclusion in 1906 from a specimen from a 4-year-old child (106). The epiphyseal line was irregular and the cartilage cells were in complete disorder with no trace of a columnar arrangement. The bone trabeculae were thin and surface osteoblasts were not seen. Similar histologic irregularities were described by Delitala (57), Barr (20), and Zadek (274), who all noted irregularities in the physeal region of the most medial part of

the growth plate. Camitz ruled out any resemblance to rickets in a detailed histologic study of three cases (42). He noted deformity of the femoral head similar to others, but few would now accept his suggestion of similarity to Perthes disease. Vascular tissue often invading the vertical fissure ruled out a hypovascularization etiology. Although he felt the disorder was developmental, he recognized that it was postnatal in occurrence. Trauma was felt to play no role in causation. Camitz' work is most valuable for demonstrating the irregular junction between physeal cartilage and metaphyseal cervical bone. The physeal cartilage itself was normal in thickness as assessed macroscopically. The neck adjacent to the physis medially contained cartilage islands, which led to a fragmented appearance radiographically. Burckhardt also studied histologic specimens from the medial portion of the femoral neck, which showed substitution of a wellorganized cartilage region by connective tissue and poorly organized cartilage (39, 40). Babb et al. summarized this early work and concluded that there was "nothing characteristic in the microscopic appearance of tissue removed from the femoral neck" (17). By this they appeared to indicate that no primary pathognomonic cartilage or bone irregularity was seen because there are evident, although secondary, irregularities. Barr examined histologic specimens and noted that the cartilaginous junction with the metaphyseal bone spicules was abrupt (20). The endochondral sequence was not occurring in anywhere close to its normal fashion. Invasion of cartilage by blood cells was absent. The cartilage was hypocellular and appeared more like nonendochondral hyaline cartilage. Zadek reported examination of tissue removed from the superior border of the neck of the femur in a patient only 5 years of age, showing an epiphyseal cartilage plate similar to one that might be undergoing early closure rather than one actively involved in cartilage endochondral growth (274). On occasion, fragments of growth plate cartilage in the metaphysis were surrounded by compact bone. Zadek also provided a detailed translation of the histologic report by Delitala. The neck of the femur had considerable fatty tissue rather than hematopoietic marrow. It also contained isolated cartilaginous nodules surrounded by bone, indicating a disordered endochondral sequence. The physis was abnormal, being characterized more by an undifferentiated hyaline-type cartilage than by a characteristic stratified physeal cartilage. The cartilage tissue was invaded irregularly by vascular tissue, which normally would not be seen in this age group. In some sections he described vessels passing through the physeal cartilage completely and linking diaphyseal (metaphyseal) bone of the neck with bone of the secondary ossification center of the epiphysis. Endochondral ossification was seen irregularly along the physis, which clearly was abnormal. Due to the vertical nature of the physis the line of ossification was in an abnormal direction, leading to slowness of bone transformation and irregularity of form and arrangement of the spongy trabeculae. Johanning also

SECTION IV ~ Infantile Coxa Vara reported similar histologic findings (133). The physeal cartilage was wider than normal but showed scattered islands of bone adjacent to it. There were few areas showing normal transition from cartilage to bone. Pylkkanen reviewed the histological studies done prior to 1960 (220). Among the best investigations were those of Hoffa and Helbing, which were based on resection of the whole proximal part of the femur, a procedure done at the time because of concern about tuberculosis. Pylkkanen then summarized the interplay of three major factors in the pathogenesis of infantile coxa vara: growth phenomena, staticmechanical relationships, and circulatory relationships in the hip joint. He examined core biopsy tissue from the medialinferior region of the metaphysis of the femoral neck in 25 patients with the biopsies taken from the vertical fissure region during osteotomy. The epiphyseal plate itself was not studied. The biopsies were from patients ranging in age from 3 to 18 years but clustered between 6 and 10 years of age (18 of 25). Endochondral ossification was defective. Histologically each of the specimens showed a "striking uniformity," although the patients were at different ages and the lesions were of different degrees of severity. Common features involved (1) cartilaginous tissue, which in the majority of cases formed a uniform plate corresponding to the zone of rarefaction visible in the radiograph, (2) metaphyseal bony tissue immediately adjacent to the cartilage, and (3) connective tissue invading both bone and cartilage. The cartilage plate consisted of cartilage of the same type as in the epiphyseal plate, but it demonstrated pathological changes in all instances characterized by markedly disturbed cell arrangements. On occasion the growth plate structure was the same as in the normal epiphyseal plate, whereas in others the cartilage cell arrangements were completely irregular. Both hypocellular and hypercellular areas were seen. Connective tissue of a fibrovascular nature was commonly interspersed. Junction between cartilage and bone showed endochondral ossification but in all specimens the process was much slighter than normal. The cell columns were short and in some sites totally irregular. The metaphyseal bone was characterized by osteoporosis. Connective tissue was common throughout the cartilage region. Pylkkanen produced an extensive analysis of the changes dividing those in each study into slight, moderate, and marked and commenting on (1) structural changes and the nature of endochondral ossification in the cartilage regions, (2) structural changes and disturbances of the bone marrow and the bone segments, and (3) connective tissue as it related to both cartilage and bone. The radiolucent fissure in most instances was composed of cartilage, which on occasion appeared to be structured almost as though it were an epiphyseal growth plate. Major structural changes were noted, however, in all cases. "The process of endochondral ossification at the junction between this cartilage and the metaphyseal bone was considerably disturbed." Interference with ossification, which in some sites was partial, in others complete,

445

was observed in all cases. Changes in the metaphyseal bony tissue were considerable, similar from case to case, and involved an osteoporosis. The amount of connective tissue within the areas of the bone or cartilage varied considerably from case to case although it tended to be more marked in more severe cases. Changes of rickets, bone necrosis, or inflammation were not seen. The zone of rarefaction running across the neck of the femur in the X rays consisted of cartilage resembling that of the epiphyseal plate, but mostly with a markedly disturbed cell arrangement. The process of ossification was severely disturbed, and the adjacent metaphyseal bone was atrophic and sometimes contained large islands of cartilage. Large amounts of connective tissue were seen within the areas of cartilage and bone. The bone marrow was scanty and fibrotic. Tissue from the zone of rarefaction showed a cartilage tissue plane across the metaphyseal spongiosa. Prior to 6-7 years of age, a cartilage pseudo-physis was seen. Much of the tissue was fibrocartilage, but some elements of a disorderly endochondral sequence were seen. In older patients beyond 8 years of age, cartilage islands were embedded in fibro-osseous tissue. Osteoporosis and fibrotic marrow were seen to either side of the cartilage. In patients 11-15 years of age, an essentially fibrous pseudoarthrosis was seen. Serafin and Szulc obtained 19 specimens for histologic investigation from the femoral neck during the process of valgus corrective osteotomy (235). Tissue from the vertical fissure was primarily cartilaginous in nature with some similarity to physeal tissue. Those parts of the physis itself available for review showed evidence of impairment of growth particularly of the endochondral ossification sequence. There were hypocellular regions, irregular arrangements of cartilage columns, and a tendency to a cartilage-bone interface in which spicules of calcified cartilage normally would extend more deeply into the metaphyseal bone and be surrounded by new bone formation on them. Once the infantile coxa vara deformity was well-established, the physeal and in particular vertical defect tissues were more fibrous and calluslike than truly cartilaginous. Bos et al. briefly referred to the appearance of lateral physeal cartilage obtained by core biopsy in two patients 4 and 9 years old at the time of valgus osteotomy of the proximal femur (31). The cartilage lacked any physeal characteristics of proliferating and hypertrophic chondrocytes and appeared as a relatively hypocellular cartilage mass interspersed between epiphyseal and metaphyseal bone but not contributing to growth by any evident endochondral mechanism. Chung and Riser published a case report of a 5-year-old boy with a unilateral infantile coxa vara who died of unrelated causes (49). He had had an intertrochanteric osteotomy on the involved femur 2 years previously. The authors reviewed previous studies of total head specimens in which it was felt that coxa vara had resulted from a defect in endochondral ossification with large amounts of fibrous tissue

446

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

rather than cancellous bone noted in the femoral neck metaphysis. There has long been concern that coxa vara was due to impaired blood supply to the femoral head-neck region particularly due to defects in the medial ascending cervical artery. Chung and Riser were able to perform vessel perfusion studies in coxa vara with comparative studies done on the contralateral normal proximal femur. The neck-shaft angle on the affected side was 85 ~ compared to the normal 135 ~. The neck was shorter and the growth plate on the affected side was wider (0.3 versus 0.2 cm). Assessments of the vertical growth plate noted an absence of orderly columnization in the hypertrophic cartilage layer, sparse bony trabeculae in the metaphysis, and little evidence of endochondral bone formation. The cartilage of the involved growth plate was pulled apart at several points with slits appearing along the long axis. Cystlike cavities were left in a number of places. The perichondrial ring, however, appeared to be normal. The cartilage formed regions of clones on the metaphyseal side, but none resembled a normal endochondral sequence. Cartilage cells near the metaphyseal border did not hypertrophy nor form a characteristic calcified matrix as a template for bone trabeculae formation. The normal vascular invasion from the metaphyseal side also was irregular and abnormal. The growth plate appeared to have split medially into two parts and to have isolated a triangular fragment of cancellous bone. The vertical fissure contained cartilage cells similar to those in the abnormal growth plate but they were even more disordered. The cartilage regions of the trochanter also were abnormal, although not as markedly involved as the medial femoral growth plate regions. The vascular patterns of the capital femoral epiphysis on the involved side were normal. The lateral ascending cervical arteries or lateral epiphyseal arteries appeared to be normal outside the bone and also within in relation to the secondary ossification center. The medial ascending cervical arteries, however, were fewer in number and smaller than normal both extraosseously and intraosseously. Chung and Riser concluded that the major problem appeared to be a defect in endochondral ossification medially. The absence of orderly endochondral ossification columns at the metaphyseal border of the head-neck growth plate resulted in decreased production of metaphyseal bone. The vascular perfusion studies demonstrated that intraosseous arteries supplying the metaphyseal side of the growth plate and the extraosseous medial ascending cervical arteries on the surface of the femoral neck were fewer in number and smaller in diameter than normal.

D. Evolution of Radiographic Change Radiographs during the first year of life, if obtained, appear to be normal (9, 28, 66). The developmental changes are relatively slow thereafter and indeed the initial radiographs, even after development of the lesion, can appear normal because the changes are occurring in the nonossified cartilage regions of the physis. By 3 years of age, however, diagnostic changes are seen. These involve increased obliquity or more

vertical positioning of the physis and a developing varus deformity of the head and neck region relative to the shaft. The neck tends to be somewhat shortened. In virtually all studies it also is considered to be retroverted. There is a characteristic radiolucent vertical fissure giving an inverted V or inverted Y appearance to the physis. The vertical fissure is inferior to the physis, being within the upper and inner part of the femoral neck. The adjacent metaphysis shows irregular ossification rather than being a smooth, slightly curvilinear line. The width of the radiolucent vertical defect is generally greater than the physis. The greater trochanter continues to grow normally and tides progressively higher than the femoral head. It tends to develop a beaking at its tip. In those situations in which a considerable number of years have passed, the head appears to slip inferior to the trochanter. The trochanter may form a facet adjacent to the ilium. Initially, the acetabulum develops normally because the femoral head is situated within it. With increasing varus tilt and inferior displacement of the head in relation to the acetabulum, acetabular dysplasia can occur. Frequently there is a delayed appearance and delayed development in size of the secondary ossification center on the involved side, although the head itself tends to remain spherical. At later stages the secondary ossification center of the femoral head is osteopenic, presumably due to diminished effective weight bearing. In many instances, increased radiolucency appears between the head and the neck due to the delay in ossification. Eventually, this will be replaced by bone. Early in the second decade premature fusion of the head-neck physis occurs usually beginning at the inferior-posterior region. On occasion, in patients not treated and followed into adulthood, a true pseudarthrosis develops in which the neck-shaft angle is most abnormal and often as little as 40-60 ~. Presently, assessment of the underlying structure in a coxa vara can be better appreciated with MR imaging. Two radiographic indices have been used to measure the extent of the coxa vara deformity. The most common is simply the head-neck-shaft angle measured from an anteroposterior radiograph. This sometimes is referred to as the angle of inclination of'the neck. When this is less than 110 ~ many consider the coxa vara deformity to be established. In a relatively large study of 42 hips by Catonne et al. the average angle was 88 ~ with a range from 65 to 110 ~ (46). Similar ranges for the head-shaft angle are reported in other series. Desai and Johnson reported an average head-shaft angle initially of 96 ~ in 20 hips with a range from 85 to 115 ~ (60). Schmidt and Kalamchi noted an abnormal neck-shaft angle of 94 ~ in 22 hips with a range from 74 to 120 ~ (231). Weinstein et al. studied several patients with isolated congenital coxa vara (by which was meant the infantile coxa vara deformity), although these were grouped with other patients with congenital coxa vara associated with shortened or bowed femurs. In this group as well the average preoperative neck-shaft angle was 90 ~ with a range from 44 to 120 ~ (266). The older the patient and the longer the period of time before initial valgus osteotomy, the greater the diminution in

SECTION IV 9 Infantile Coxa Vara the neck-shaft angle. The clinical and radiographic progression of deformity was documented clearly by Pouzet (217). This also was well-documented in a large series of 130 affected hips in 106 patients from Poland studied by Serafin and Szulc (235). The average neck-shaft angles in groups aged 2-6, 7-11, and 12-16 years and post-skeletal maturation were 85 ~ 71 o, 67 o, and 55 ~ respectively. The second angle used to document the coxa vara deformity in the infantile variant is that described by Weinstein et al. and referred to as Hilgenreiner's epiphyseal (HE) angle (266) (Fig. 22). This uses Hilgenreiner's line as a horizontal axis and a line through the metaphyseal side of the physeal defect as the vertical axis. The average HE angle in their group of 100 normal hips was 16~ with a range from 0 to 25 ~ The average HE angle that developed in their coxa vara patients was 82 ~ with a range from 66 to 120 ~ Desai and Johnson also utilized this measurement in assessing 20 hips and noted an average HE angle of 66 ~ at initial evaluation with a range from 45 to 90 ~ (60).

E. Pathomechanics of Deformity in Infantile Coxa Vara There is a clear dissociation of growth between the greater trochanter, which is normal or relatively normal, and the head-neck regions, which clearly are affected. There is a uniform crescentic growth plate under the head-neck and greater trochanter until 4 years of age, at which time divergence of the head-neck axis leaves two separate physes. The trochanteric portion of the growth plate contributes to the length of the shaft, whereas that of the head-neck region is responsible for shaping of that area alone particularly due to its oblique position. Once there is a relative change in growth rate and, thus, in position of these two structures, the head and neck region is placed at a mechanical disadvantage. With the physis increasingly directed toward the vertical plane, the tendency to its slippage is markedly greater. When the head and neck angle is less than 100~ worsening is inevitable. With time the medial physeal cartilage fuses, thus allowing for further and more rapid worsening of the condition. The clinical and radiographic appearances remain consistent with a lateral rotation deformity of the neck and shaft with the head being positioned in a medial and relatively retroverted position. The radiographs indicate that one-half or more of the femoral head is uncovered and the shaft appears laterally rotated. This appearance could be produced by the capital epiphysis including the physis and the triangular fragment of the metaphysis sliding distally until the superior aspect of the head comes against the inferior surface of the proximal part of the neck. Presentations of the pathogenesis of the deformity along these lines would indicate that the patient essentially has suffered a type II fracture-separation. Due to the existence of a large number of bilateral and familial cases infantile coxa vara cannot be attributed simply to trauma alone, although the worsening displacement appears to have a mechanical cause on the basis of growth in

447

relation to the disadvantageous early varus position. Pouzet felt that the primary site of abnormality was the physis and the femoral neck, with the femoral head and its secondary ossification center and the diaphysis remaining normal (216, 217). For the first few years the femoral head was round with good density of the secondary ossification center and positioning of the head in the acetabulum. Deformity occurred initially because of insufficient biological growth of the medial physis and adjacent femoral neck followed by secondary biomechanical changes due to the altered position of the head and orientation of the physis in relation to weight bearing. In the most advanced form, the neck is extremely shortened and the head essentially is plastered against the medial diaphysis adjacent to the lesser trochanter with extreme relative overgrowth of the greater trochanter. In a biomechanical sense, deformity in infantile coxa vara in a growing child worsens with time because the region of greatest stress in a proximal femur positioned into coxa vara is concentrated around the vertical radiolucent fissure and triangular fragment. The possibility has been raised that the etiology of the vertical fissure is traumatic, being analogous to a type II fracture-separation (28). It would appear, however, that rather than being acutely traumatic the vertical fissure represents a stress fracture or pseudo-fracture. Pauwels stresses that the varus malposition not only hinders the new formation of bone but as a consequence also serves to break down any newly formed trabeculae shortly after their synthesis (207). Pauwels indicates, therefore, that infantile coxa vara is a weight bearing deformity. The triangular fragment of bone on the medial surface of the neck is limited medially by the epiphyseal line and laterally by the fissure. The latter is truly vertical, whereas the epiphyseal line has increased obliquity from the normal. The radiographic abnormalities in the neck change fairly dramatically after subtrochanteric osteotomy, which again supports the stress and trauma causation. Walter postulated that relatively excess body weight in coxa vara was present, with the vertical fissure representing not a congenital deformity but a physiological reaction of the bone to excess shearing stresses of weight beating (258). The rapid healing of the lesion that occurred following valgus osteotomy was further confirmation of this finding. Chung and Riser feel that the vertical fissure is indeed a stress phenomenon and that the varus worsens from the increasing load on the femoral head of the growing individual, who is becoming progressively heavier (49). A vertical fissure tends to persist for several years because continued and increasing stress is concentrated in the area osteogenic activity decreases. When the stress is relieved by valgus osteotomy alone, the vertical fissure almost always disappears by healing, being replaced by bone. Both sides of the vertical fissure contain cartilage cells resembling those in the disorganized growth plate. Three separate events appear to occur in the development of an infantile coxa vara deformity. (1) The first event involves a decrease in function of the medial and inferior part

448

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired AbnormaBties of the Femur

of the growth plate. This appears evident from the histologic studies in which there is a lack of cartilage proliferation, the growth plate becomes disorganized, there is vascular and fibrous invasion, and bone islands form. It does not appear that the entire proximal femoral capital growth plate is involved initially. The involvement is of the inferior medial portion with both histologic and radiographic evidence seen for this in that considerable vertical obliquity of the growth plate develops. If the entire proximal femoral growth plate was involved initially, the physis would remain horizontal and that of the greater trochanter would continue to grow. In fact, however, there is a tilting of the entire head-neckgreater trochanter complex and the plate itself becomes oblique. This must indicate that the tethering or nongrowth effect is at the medial and inferior margin of the plate with the lateral two-thirds of the head-neck physis functioning well. (2) The second stage in the pathogenesis involves a slippage of the head and physeal region and the adjacent portion of the metaphyseal bone of the neck in relation to the rest of the upper end of the femur. Once the varus reaches a certain degree, the stresses on the vertical physeal region are sufficiently great that mechanical slippage occurs. This would seem to be gradual as the patient does not appear to experience any discomfort and as such is more along the lines of a stress fracture or even analogous to an in situ slipping of the proximal femoral capital epiphysis. This would appear to represent the etiology of the inverted triangular fragment at the inferior surface of the neck. Bone and degenerate cartilage remain present in the inferior region and the radiolucent line seen medially represents the epiphyseal cartilage and the remnant of the physeal cartilage, with the other line more lateral toward the femoral neck bone representing the vertical fissure or stress fracture. This region too shows cartilage histologically although it is even more disorganized than that of the physis. Others have postulated that there is a lateral rotation of the neck and shaft region, which also is analogous to the situation in slipped capital femoral epiphysis. There are some similarities to a type II epiphyseal fracture-separation, although the upper end of the femur is entirely cartilage at this age (except for the secondary ossification center of the head) with a single growth plate composing the head, intertrochanteric, and greater trochanteric cartilage areas. (3) Final resolution of the deformity occurs when a subtrochanteric valgus osteotomy is performed. This serves to reposition the head deeply into the acetabulum and to reposition the vertical growth plate close to its normal horizontal axis. Once the hip has been positioned in this way, there is almost invariable healing of the triangular fragment, which is consistent with it being a stress fracture rather than a cartilaginous embryonic maldevelopment. This viewpoint was defined first by Elmslie in his reports in 1907 and 1913 and appears to be accurate today. His views were reviewed more recently by B lockey, who supports a traumatic etiology. On occasion, episodes of trauma do lead to the appearance of infantile coxa vara as described by Blockey (28),

Elmslie (66), and Joachimsthal (132). The worsening of the disorder with time on a biomechanical basis was defined by Elmslie. As the vast majority of the cases of infantile coxa vara develop prior to 4 years of age, at which time there is still cartilage continuity from the head and neck region and the greater trochanter, the pathogenesis of the deformity cannot be defined clearly by radiographic studies. Whether the initial varus deformity is caused by trauma or by failure of growth in the medial and inferior portion of the neck, the varus itself will predispose one to further slipping. When the epiphysis of the head has been moved into the vertical plane, the neck of the femur develops in a horizontal rather than a linear and oblique direction because the line of growth now is nearly vertical instead of horizontal. When the head-neck-shaft axis is less than 90 ~ the femoral head does not rest appropriately in the acetabulum and the weight bearing is done on the most lateral aspect of the head, which has the thinnest cartilage surface. The waddling gait is worsened by the high-tiding greater trochanter, which serves to shorten the point of attachment of the gluteus medius and minimus muscles, leaving them relaxed and unable to perform at their normal level of function. The gait is worsened further by shortening of the involved femur if the condition is unilateral. The entire spectrum of growth deformities in infantile coxa vara was well-reviewed by Serafin and Szulc (235). The primary abnormality leads to a reduction in the neckshaft angle, which also can be documented as an increase in the Hilgenreiner epiphyseal angle. There is retroversion in the femoral neck in 85% of cases, upward overgrowth of the greater trochanter in virtually all cases, diminution in the size of the femoral head in virtually all cases, and some shallowness and underdevelopment of the acetabulum in most cases. The earlier the head is repositioned by valgus osteotomy into its normal relationship to the acetabulum and the less severe the growth abnormalities of the neck, then the better the acetabular contours.

F. Clinical-Radiographic Correlations Fairbank listed the specific radiographic criteria for a diagnosis of infantile coxa vara as (1) decreased neck shaft angle, (2) wide, vertically aligned physis or proximal femoral epiphyseal plate, (3) irregular metaphyseal ossification, (4) shortened femoral neck, (5) triangular osseus fragment adjacent to the inferior margin of the physis, (6) normally shaped but osteoporotic femoral head, and (7) straight femoral shaft (71). Johanning assessed several cases with long-term followup X rays (133). In patients with infantile coxa vara, the initial radiographs were characterized by the varus deformity of the head and neck axis, a short and poorly developed neck, and widening of the angle formed by the epiphyseal line in relation to the horizontal. He also noted the earliest formation of calcific deposits at the inferior surface of the neck in

SECTION IV ~ Infantile Coxa Vara relation to the widened and more vertical physeal space between the secondary ossification center and the metaphysis. He commented on the increasing inferior slippage of the head and neck fragment in relation to the rest of the proximal femur with time. Indeed, in the age group between 10 and 15 years, the slipping of the head became much more marked. This is illustrated particularly well in an untreated but well-documented case showing the typical development of infantile coxa vara at 3, 8, and 13 years. At 3 years of age, the superior surface of the head is at the level of the tip of the greater trochanter and the characteristic abnormalities can be seen. At 8 years of age, there is considerable overgrowth of the greater trochanter, which is now above the lip of the lateral acetabulum, whereas at 13 years, the greater trochanter approached the iliac spine and the lower part of the femoral head is well below the lesser trochanter. One of the problems is the part played by trauma in infantile coxa vara. As Johanning indicates, the trauma need not come from a single event but rather can be the outcome of abnormal stresses over time interfering with nutrition of the parts involved. The good results achieved by valgus osteotomy with appropriate healing of the vertical fissure clearly point to trauma and shear stresses as causative because no other generalized or systemic problem would respond that well to a simple change in position. Magnusson showed that the average age at which symptoms set in was 3.3 years and that the major mode of presentation was limping (176). There were no problems prior to the onset of the limping. He describes the gradual onset of changes in the radiographic picture, with the primary abnormality being the varus position and the vertical fissure occurring secondarily and not until the varus has reached a certain degree. He feels that the fissure is an "insufficiency fracture," which at a later stage may proceed to a real pseudarthrosis. Magnusson presents a series of 85 hips with infantile coxa vara treated in Stockholm, Sweden, between 1927 and 1941. The average age at subtrochanteric cuneiform osteotomy was 10.9 years and the average age at long-term follow-up was 27.3 years. The average age at first examination, at which time the diagnosis was made, was 6 years so that there was an average of 4.9 years between initial diagnosis and surgery. Some of the long-term changes involved shortening of the neck, which was particularly marked in those operated at a later age. On occasion, there also were changes in the shape of the head. During the course of the disease, there was obvious bone atrophy and deformation. Deformation of the head rarely was seen when the neckshaft angle was reestablished surgically. In virtually all cases, the acetabulum tended to be more shallow than normal. Magnusson thus felt that "to complete the picture of a fully developed coxa vara infantum, the following should be added to the roentgenological symptoms already k n o w n - the varus position and possibly the vertical fissure: 1) a shortened collum (neck); 2) a more or less deformed caput (head); and 3) a shallow acetabulum." He clearly indicated that the

449

longer the process continued undisturbed, the more characteristic the changes. The only correctable deformity was the neck-shaft angle. He stressed that "the earlier this restitution takes place during the course of the pathological process, the less characteristic will the changes be in the parts concerned of the hip joint and the better the functional results will be in the long run." The major long-term problem in those treated late was a persisting pseudarthrosis. When present, virtually complete resorption of the neck had occurred. The head had been deformed and the acetabulum was quite shallow. His conclusion was straightforward-the subtrochanteric cuneiform osteotomy was the logical form of treatment and the earlier it was performed, the less progressive the long-term changes and the better the longterm result.

G. Management of Infantile Coxa Vara Coxa vara responds well to surgical repositioning of the proximal end of the femur by valgus osteotomy at the intertrochanteric or subtrochanteric level. The aim of treatment is 2-fold: to correct deformity and to enhance repair of the vertical defect by ossification. Since the 1890s when coxa vara was formally defined, proximal femoral valgus osteotomy has been performed to correct the deformity and remains the mainstay of therapy today. Nonoperative therapy plays no role. Observation remains important as the timing of any surgical intervention often is unclear. Although coxa vara will not correct spontaneously, the rate of progression can vary greatly. Due, however, to the fact that the growth plate is not normal, it is not unusual for any operation to require repetition during the growing years. The degree of correction frequently must be extensive, depending of course on the severity of deformity. Those operations that undercorrect the deformity are doomed to a far earlier repetition than would otherwise be the case. Weighill has pointed out the value of adductor tenotomy along with the valgus osteotomy (264). He felt that 10 of 10 patients with the combined approach did well, whereas only approximately three-fourths of 22 osteotomies without adductor release did well. Operative treatment involves a proximal femoral valgus osteotomy, which accomplishes the correction of each of the three major clinical abnormalities by (1) repositioning the head into the weight beating position in the acetabulum, (2) lengthening the lever arm, thus making the abductor muscles function more effectively, and (3) lengthening the limb to compensate for the loss of growth. Le Mesurier noted, as have many others, that the valgus osteotomy also allows the radiolucent vertical neck defect to heal (166). Healing of the defect was seen in 11 of 12 cases within 1 year of surgery. Babb et al. recommended subtrochanteric osteotomy at 6-8 years of age with "wide" abduction of the distal limb (17). In a mechanical sense, the negative shear forces on the head and neck would be converted to compression forces, healing the vertical osseous defect and stabilizing the

450

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

proximal femoral anatomic relationships. Although some surgeons recommended drilling across the triangular cervical defect or even bone grafting, neither of these are needed after valgus osteotomy to ensure healing of the defect by bone. Guidelines for therapy were defined as early as 1938 by Pouzet of Lyon, France (216). He recognized that the goals of treatment were to straighten the head and neck relationship to a normal range, to obtain ossification of the cartilaginous fissure of the neck, and, if possible at the same time, to obtain premature fusion of the growth plates of the proximal femur to prevent recurrence. This triple goal was best obtained by corrective subtrochanteric osteotomy. Previous treatments felt to be neither effective nor necessary involved closed reduction of the deformity, bone graft to the neck to allow the independent fragment to unite, and resection performed in the neck region to remove the pseudarthrosis. The subtrochanteric valgus osteotomy not only corrected the deformity but allowed for the rapid ossification of the neck as well as for premature transphyseal fusion such that the varus would not recur. Pouzet referred to several other reports of excellent results with osteotomy that had been described in the previous decade. Correction often was enhanced by adductor myotomy. The appropriate time for surgery varied from case to case but appeared to be best between 5 and 8 years of age. He also pointed out the frequent premature fusion of the femoral head-neck growth plate cartilage following osteotomy. This was not a complication of the surgery but rather indicated that the physeal cartilage was not normal as part of the disease state. Pylkkanen has also indicated that conservative measures were of no avail and that surgery was warranted (220). Subtrochanteric valgus osteotomy was needed, and "the earlier the surgical intervention, the larger was the ratio of good results in each age group." He concluded that "the lesion should be corrected as early as possible." This generally means intervention in the 4- to 8-year-old range. Additional procedures can involve a distal transfer of the greater trochanter to further lengthen and thus strengthen the abductor muscle group, although this should not be done if there is still meaningful trochanteric growth remaining. On the other hand, if done toward the end of skeletal growth the trochanteric transfer also will serve to obliterate the greater trochanteric growth plate and prevent recurrence of the deformity. The need to transfer the greater trochanter distally is dependent on its position after completion of the valgus osteotomy. In some, rotation of the proximal end of the femur alone is sufficient to appropriately position the head-neck and greater trochanter areas, whereas in others the greater trochanter still tides too high and must be separately and further displaced. Growth of the proximal femoral capital epiphysis in infantile coxa vara is, by definition, less than that of the greater trochanter. In addition there is a marked tendency for the physis to close prematurely, further worsening the prognosis. The closure may be evident radiographically or by MR imaging. If osteotomy is done early in the second decade

and the physis appears to be closed, in situ greater trochanteric epiphyseal arrest should be considered. There is no need to treat the nonunited fragment of the femoral neck; with correction of the coxa vara deformity this fragment almost invariably will heal itself spontaneously. Valgus osteotomy converts the vertical fissure subject to sheafing forces to a more horizontal plane where compression, favorable to healing, occurs. The positive Trendelenberg sign, which is always seen, is converted to normal with effective valgus repositioning. Pylkkanen noted this in 96 of 114 cases of widely varied severity. Lower extremity length discrepancy can be managed by contralateral distal femoral epiphyseal arrest. There is relatively little shortening of the femur in infantile coxa vara, unlike the situation in truly congenital coxa vara associated with dysgenesis of the proximal femur. The limb shortening with infantile coxa vara is caused by the varus deformation and the decrease in length of the femoral neck. Amstutz noted shortening in 10 patients with unilateral developmental coxa vara to be not more than 4 cm, which he felt was consistent with developmental irregularity at the head and neck region exclusively (10, 11). Pouzet referred to 3-5 cm of shortening at skeletal maturity in severe untreated cases (217). Even the mildest cases of proximal femoral focal deficiency lead to much more extensive shortening than could be accounted for by the varus of the hip. Virtually all studies note that the end results in patients with infantile coxa vara are dependent on the initial angle of inclination, the age at the time of surgery, and the extent of valgus correction at the time of surgery. Poor results are seen in those in whom the neck-shaft angle is less than 90 ~, the patients are operated initially after 9 years of age, and valgus correction is less than 130 ~ To a great extent, the timing for surgery is dependent on age at initial presentation of the disorder (over which the orthopedic surgeon has relatively little control), but it also is dependent on the extent of the deformity and its rate of progression. Schmidt and Kalamchi noted that coxa vara with mild deformity with neck-shaft angles greater than 110 ~ tends to heal the metaphyseal fragments early and subsequently improves the angles (231). In these hips only observation is performed. Similar observations have been made by others (266). Deformities less than 100 ~ on the other hand, tend to progress and surgery generally is recommended at that range. Hips with angles between 100 and 110 ~ are observed initiallywith the timing of intervention dependent on progression of the deformity. Catonne et al. strongly indicated that anatomical results were better if the osteotomy was made before the age of 9 years (46). Most recurrences in the study by Desai and Johnson were in those older than 5 years of age at surgery (60). Serafin and Szulc showed a clear temporal relationship in the extent of subsequent good results (235). In those operated between 2 and 9 years of age there were 80% good results with clear diminution after that, showing surgery at 10-11 years of age

SECTION IV ~ Infantile Coxa Vara

yielding 62% good results, 12-16 years 52% good results, and 17 years and older only 33% good results. Thus, there appears to be little to be gained by waiting if the angle of deformity is less than 100 ~ and certainly if the patient is older than 8 or 9 years of age. Correction appears to be held best with initial valgus repositioning greater than 130 ~ In terms of the Hilgenreiner epiphyseal angle measurement, invariable progression appeared with the angle greater than 60 ~ a tendency to spontaneous correction when the angle was less than 45 ~, and variable response when the angle was between 46 and 59 ~ such that especially close assessment was needed in the latter group (266). Several types of valgus osteotomy have been reported with little apparent difference in result. Weighill found no differences in results between subtrochanteric and intertrochanteric sites of osteotomy (264). In virtually all instances, healing of the vertical metaphyseal defect occurs. In addition, there is almost always premature closure of the proximal femoral growth plate. This must be followed closely in all patients, and if it appears to be occurring with a few years of growth remaining, it may be necessary to perform elective epiphyseal arrest of the greater trochanteric physis to maintain the appropriate alignment. In those undergoing surgery after 10 years of age, fusion of the greater trochanteric epiphysis at time of surgery generally is recommended. Serafin and Szulc recommend significant correction of the neck-shaft angle, advising that postsurgically it be 10-15 ~ greater than

451

normal (235). A postoperative HE angle of 35 ~ or less and a head-shaft angle of 130 ~ or more almost always correlated with consistently satisfactory results. With appropriate attention to surgical detail and relatively early intervention the aims of surgery generally can be achieved: (1) to create a normal neck-shaft angle, (2) to promote healing of the bony defect, and (3) to reorient the physis into a more horizontal position. Careful follow-up is essential after surgery because in some there can be recurrence and in almost all there is premature closure of the proximal femoral physis and a tendency to relative overgrowth of the greater trochanter after that. Premature physeal closure is an almost invariable component of the disorder and should not be attributed to the valgus repositioning itself. It is seen in virtually 90% of cases and is due to the primary deformity and not specifically to the treatment. Coxa vara can occur in association with many of the generalized skeletal dysplasias. It is common particularly with spondyloepiphyseal dysplasia congenita and cleidocranial dysostosis and also has been described in variants of multiple epiphyseal dysplasia and metaphyseal dysostosis. In the skeletal dysplasias it is common for the acetabulum to be abnormal as well. Some cases of coxa vara in the skeletal dysplasia group have the isolated neck fragment, but most do not. Coxa vara in the specific dysplasias will be discussed in depth in Chapter 9. Examples of cases of coxa vara with varying etiologies treated by valgus osteotomy are shown in Figs. 23 and 24.

452

CHAPTER 5 ~

Coxa Vara in Developmental and Acquired Abnormalities of

the

Femur

F I G U R E 23 Examples of two cases of coxa vara are shown. Parts (A-F) represent the patient with bilateral coxa vara with spondyloepiphyseal dysplasia congenita. Parts (G-I) represent a unilateral case of coxa vara. (A) Anteroposterior radiograph of the pelvis in a 7-year-old girl with SED congenita. She presented with a distinct waddling gait. There is no ossification of the secondary ossification centers of the femoral head in spite of her age. Note also the characteristic triangular fragment of the medial neck region on the left. The acetabulae are markedly underdeveloped with no subchondral bone seen on either side. (B) Frog lateral view of pelvis also demonstrates the absence of secondary ossification centers and underdeveloped acetabulae. (C) Proximal femoral valgus osteotomy on the left served to reposition the head and neck into a more favorable weight bearing position. Healing of the triangular fragment soon followed. (D) Frog lateral view shows the central part of the neck indicated by the screw now pointing toward the triradiate cartilage in the depths of the acetabulum. (E) Localized view of the left hip preosteotomy shows the considerable degree of coxa vara with the tip of the greater trochanter (white arrow) well above the head and neck region. The characteristic triangular fragment of the inferior medial neck is seen. (Fi) Arthrogram and anteroposterior view of the left hip serves to outline the spherical contours of the femoral head and markedly shortened adjacent neck. The head is located in the acetabulum but the latter is extremely poorly developed. In addition, the head points medially and inferiorly rather than medially and upward to the weight bearing surface. (Fii) Frog lateral view of the arthrogram shows the spherical femoral head in appropriate relation to the acetabulum. The head and neck are markedly shortened and there is no secondary ossification center even at 7 years of age. (G) Unilateral coxa vara in a 6-year-old boy is seen. The head is spherical and is well-situated in the acetabulum, but the neck is markedly shortened and there is extreme relative overgrowth of the greater trochanter, whose tip rises above the level of the superior acetabulum. (I-I) The head now is positioned nicely in the acetabulum and the tip of the greater trochanter is displaced inferiorly following proximal femoral valgus osteotomy. (I) Frog lateral view postosteotomy shows good position of the femoral head in relation to the acetabulum.

SECTION IV ~ Infantile Coxa Vara

F I G U R E 23 (continued)

453

F I G U R E 24 The patient with undefined skeletal dysplasia and bilateral coxa vara. (A) Anteroposterior radiograph of the pelvis shows bilateral proximal femoral coxa vara. The tip of the greater trochanter bilaterally is at the same level as the most superior part of the femoral head. Both femoral heads are well-located and the acetabulae are well-developed. The necks are shortened, however, even though the physeal regions appear widened. Proximal femoral valgus osteotomy was performed with Wagner blade plate fixation. There was excellent correction of the varus deformation at the time of healing. (C) High-power view of the right hip shows intraoperative positioning after osteotomy. Note that the proximal femoral head-neck physis is now almost horizontal, the tip of the greater trochanter is well below the most superior surface of the femoral head, and there has been excessive valgus angulation at the osteotomy site. Three years later at age 8 there had been recurrence of the varus deformation (not shown). (D) Proximal femoral valgus osteotomy again was repeated with even more extensive valgusization performed. Anteroposterior radiograph shows the position at healing after the second series of osteotomies. (E) Frog lateral view shows the excellent position of the femoral heads postosteotomy. (F) Anteroposterior X ray 3 years after the second osteotomy at 11 years of age shows retained position of the correction. Both femoral heads are well-located in the acetabulae. The trochanters remain normally positioned and the proximal femoral growth plates appear in their normal position with continued growth. (G) Frog lateral view shows persisting physeal function and excellent structural development of the acetabulae, femoral heads, femoral necks, and shafts. The child continues to walk without waddling and has as full range of hip motion at 14 years of age.

References

References 1. Aadalen RJ, Weiner DS, Hoyt W, Herndon CH (1974) Acute slipped capital femoral epiphysis. J Bone Joint Surg 56A: 1473-1487. 2. Abraham E, Garst J, Barmada R (1993) Treatment of moderate to severe slipped capital femoral epiphysis with extracapsular base-of-neck osteotomy. J Pediatr Orthop 13:294-302. 3. Agamanolis DP, Weiner DS, Lloyd JK (1985) Slipped capital femoral epiphysis: A pathological study I. A light microscopic and histochemical study of 21 cases. J Pediatr Orthop 5:40-46. 4. Agamanolis DP, Weiner DS, Lloyd JK (1985) Slipped capital femoral epiphysis: A pathological study II. An ultrastructural study of 23 cases. J Pediatr Orthop 5:47-58. 5. Aitken GT (1969) Proximal femoral focal deficiency--deftnition, classification, and management. In: Proximal Femoral Focal Deficiency. A Congenital Anomaly, ed GT Aitken, pp. 1-22, Washington, DC: National Academy of Sciences. 6. Alexander C (1966) The etiology of femoral epiphysial slipping. J Bone Joint Surg 48B:299-311. 7. Alsberg A (1899) Anatomische und klinische betrachtungen uber coxa vara. Zeitschr f Orthop Chir 6:106-172. 8. Amstutz HC (1969) The morphology, natural history, and treatment of proximal femoral focal deficiencies. In: Proximal Femoral Focal Deficiency. A Congenital Anomaly, ed GT Aitken, pp. 50-76, Washington, DC: National Academy of Sciences. 9. Amstutz HC (1970) Developmental (infantile) coxa vara--a distinct entity: Report of two patients with previous normal roentgenograms. Clin Orthop Rel Res 72:242-247. 10. Amstutz HC, Freiberger RH (1962) Coxa vara in children. Clin Orthop Rel Res 22:73-92. 11. Amstutz HC, Wilson PD, Jr (1962) Dysgenesis of the proximal femur (coxa vara) and its surgical management. J Bone Joint Surg 44A: 1-23. 12. Aronson DD, Karol LA (1996) Stable slipped capital femoral epiphysis: Evaluation and management. J Am Acad Orthop Surg 4:173-181. 13. Aronson DD, Loder RT (1992) Slipped capital femoral epiphysis in black children. J Pediatr Orthop 12:74-79. 14. Aronson DD, Loder RT (1996) Treatment of the unstable (acute) slipped capital femoral epiphysis. Clin Orthop Rel Res 322:99-110. 15. Aronson DD, Peterson DA, Miller DV (1992) Slipped capital femoral epiphysis. The case for internal fixation in situ. Clin Orthop Rel Res 281:115-122. 16. Aronson DD, Carlson WE (1992) Slipped capital femoral epiphysis: A prospective study of fixation with a single screw. J Bone Joint Surg 74A:810-819. 17. Babb FS, Ghormley RK, Chatterton CC (1949) Congenital coxa vara. J Bone Joint Surg 31A:115-131. 18. Balensweig I (1926) Femoral osteochondritis of adolescents and sequelae. Surg Gyn Obstet 43:604-614. 19. Ballmer PM, Gilg M, Aebi B, Ganz R (1990) Ergebnisse nach subkapitaler und Imhauser-Weber osteotomie be epiphyseolysis capitis femoris. Z Orthop 128:63-66. 20. Barr J S (1929) Congenital coxa vara. Arch Surg 18:1909-1119. 21. Barrett IR (1985) Slipped capital femoral epiphysis following radiotherapy. J Bone Joint Surg 5:268-273.

455

22. Bellemans J, Fabry G, Molenaers G, Lammens J, Moens P (1996) Slipped capital femoral epiphysis: A long-term followup, with special emphasis on the capacities for remodeling. J Pediatr Orthop Part B 5:151-157. 23. Bennet GC, Koreska J, Rang M (1984) Pin placement in slipped capital femoral epiphysis. J Pediatr Orthop 4:574-578. 24. Betz RR, Steel HH, Emper WD, Huss GK, Clancy M (1990) Treatment of slipped capital femoral epiphysis: Spica cast immobilization. J Bone Joint Surg 72A:587-600. 25. Bianco AJ (1966) Treatment of slipping of the capital femoral epiphysis. Clin Orthop Rel Res 48:103-110. 26. Billing L, Severin E (1959) Slipping epiphysis of the hip. A roentgenological and clinical study based on a new roentgen technique. Acta Radiol (Supp) 174:1-76. 27. Bishop JO, Oley TJ, Stephenson CT, Tullos HS (1978) Slipped capital femoral epiphysis: A study of 50 cases in black children. Clin Orthop Rel Res 135:93-96. 28. Blockey NJ (1969) Observations on infantile coxa vara. J Bone Joint Surg 51B:106-111. 29. Boden SD, Fallon MD, Davidson R, Mennuti MT, Kaplan FS (1989) Proximal femoral focal deficiency. Evidence for a defect in proliferation and maturation of chondrocytes. J Bone Joint Surg 71A:l 119-1129. 30. Bone LB, Roach JW, Ward WT, Worthen HG (1985) Slipped capital femoral epiphysis associated with hyperparathyroidism. J Pediatr Orthop 5:589-592. 31. Bos CFA, Sakkers RJB, Bloem JL, vd Stadt RJ, vd Kamp JJP (1989) Histological, biochemical, and MRI studies of the growth plate in congenital coxa vara. J Pediatr Orthop 9:660-665. 32. Boyd HB, Ingram AJ, Bourkard HO (1949) The treatment of slipped femoral epiphysis. South Med J 42:551-560. 33. Boyer DW, Mickelson MR, Ponseti IV (1981) Slipped capital femoral epiphysis: Long-term follow-up study of one hundred and twenty-one patients. J Bone Joint Surg 63A:85-95. 34. Brailsford JF (1933) Slipping of the epiphysis of the head of the femur. Lancet i: 16-19. 35. Brenkel I, Dias J, Finlay D, Gregg P (1988) Mechanical factors in patients with slipped upper femoral epiphysis. J Bone Joint Surg 70B:852. 36. Brodetti A (1960) The blood supply of the femoral head and neck in relation to the damaging effects of nails and screws. J Bone Joint Surg 42B:794-801. 37. Broughton NS, Todd RC, Dunn DM, Angel JC (1988) Open reduction of the severely slipped upper femoral epiphysis. J Bone Joint Surg 70B:435-439. 38. Bryant DD, III, Epps CH, Jr (1991) Review: Proximal femoral focal deficiency: Evaluation and management. Orthopaedics 14:775-784. 39. Burckhardt E (1946) Zur klinik and pathologischen anatomie der coxa vara infantum. Helv Chir Acta 13:28. 40. Burkhardt E (1946) Zur histologie der coxa vara infantum. Helv Chir Acta 13:123. 41. Burrows H (1957) Slipped upper femoral epiphysis: Characteristics of a hundred cases. J Bone Joint Surg 39B:641-658. 42. Camitz H (1934) Etude comparee sur la coxa vara dite congenitale et l'osteochondrite coxale juvenile (coxa plana). Acta Chir Scand 73:520-575. 43. Carlioz H, Vogt JC, Barba L, Doursounian L (1984) Treatment of slipped upper femoral epiphysis: 80 cases operated on over 10 years (1968-1978). J Pediatr Orthop 4:153-161.

456

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

44. Carney BT, Weinstein SL, Noble J (1991) Long-terrn followup of slipped capital femoral epiphysis. J Bone Joint Surg 73A:667-674. 45. Casey BH, Hamilton HW, Pobechko WP (1972) Reduction of acutely slipped upper femoral epiphysis. J Bone Joint Surg 54B:607-614. 46. Catonne Y, Dubousset J, Seringe R, Conard JP, Dintimille H, Gottin M, Rouvillain JL (1992) Infantile coxa vara: A review of 28 cases. J Orthop Surg 6:181-191. [Rev Chir Orthop 78: 153-163, 1992.] 47. Chapman JA, Deakin DP, Green JH (1980) Slipped upper femoral epiphysis after radiotherapy. J Bone Joint Surg 62B: 337-339. 48. Chung SMK, Hiram TT (1981) Multiple pin repair of the slipped capital femoral epiphysis. In: Clinical Biomechanics: A Case History Approach, eds J Black, JH Dumbleton. pp. 94-115, New York: Churchill Livingstone. 49. Chung SMK, Riser WH (1978) The histological characteristics of congenital coxa vara: A case report of a 5 year old boy. Clin Orthop Rel Res 132:71-81. 50. Clarke HJ, Wilkinson JA (1990) Surgical treatment for severe slipping of the upper femoral epiphysis. J Bone Joint Surg 72B:854-858. 51. Cohen MS, Gelberman RH, Griffin PP, Kasser JR, Emans JB, Millis MB (1986) Slipped capital femoral epiphysis: Assessment of epiphyseal displacement and angulation. J Pediatr Orthop 6:259-264. 52. Colton CL (1987) Sipped upper femoral epiphysis. In: Recent Advances in Orthopaedics. pp. 61-77, Edinburgh: Churchill Livingstone. 53. Cowell HR (1966) The significance of early diagnosis and treatment of slipping of the capital femoral epiphysis. Clin Orthop Rel Res 48:89-94. 54. Crawford AH (1988) Slipped capital femoral epiphysis. J Bone Joint Surg 70A: 1422-1427. 55. Crawford AH (1989) The role of osteotomy in the treatment of slipped capital femoral epiphysis. Instructional course lectures. AAOS 273-279. 56. Cruess RL (1963) The pathology of acute necrosis of cartilage in slipping of the capital femoral epiphysis: A report of two cases with pathological sections. J Bone Joint Surg 45A: 1013-1024. 57. Delitala F (1913) Sulla coxa vara congenita. Arch Di Ortop 30:382-396. 58. DeQuervain F (1898) De la coxa vara. Semaine Med 18:41-48. 59. DeRosa GP, Mullins RC, Kling TF (1996) Cuneiform osteotomy of the femoral neck in severe slipped capital femoral epiphysis. Clin Orthop Rel Res 322:48-60. 60. Desai SS, Johnson LO (1993) Long-term results of valgus osteotomy for congenital coxa vara. Clin Orthop Rel Res 294: 204-210. 61. Drehmann G (1911) Die coxa vara. Ergeb Chir Orthop 2: 452-487. 62. Dunn DM (1964) The treatment of adolescent slipping of the upper femoral epiphysis. J Bone Joint Surg 46B:621-629. 63. Dunn DM, Angel JC (1978) Replacement of the femoral head by open operation in severe adolescent slipping of the upper femoral epiphysis. J Bone Joint Surg 60B:394-403. 64. Durbin FC (1960) Treatment of slipped upper femoral epiphysis. J Bone Joint Surg 42B:289-302.

65. Duverney GI (1751) Traite des Maladies des Os, volume 1, chapter 8. pp. 355-375, Paris. 66. Elmslie RC (1907) Injury and deformity of the epiphysis of the head of the femur: Coxa vara. Lancet i:410-417. 67. Engelhardt P (1990) Epiphyseolysis capitis femoris: Uberlegungen zur therapie der gegenhufte am wachstumsende. Z Orthop 128:262-265. 68. Exner GU (1986) Growth and pubertal development in slipped capital femoral epiphysis: A longitudinal study. J Pediatr Orthop 6:403-409. 69. Fabry G, MacEwen GD, Shands AR (1973) Torsion of the femur. A follow-up study in normal and abnormal conditions. J Bone Joint Surg 55A:1726-1738. 70. Fahey JJ, O'Brien ET (1965) Acute slipped capital femoral epiphysis. J Bone Joint Surg 47A:1105-1127. 71. Fairbank HAT (1928) Infantile or cervical coxa vara. In: The Robert Jones Birthday Volume. A Collection of Surgical Essays, ed H Milford. pp. 225-241, London: Oxford University Press. 72. Fairbank TJ (1969) Manipulative reduction in slipped upper femoral epiphysis. J Bone Joint Surg 51B:252-262. 73. Ferguson AB, Howorth MB (1931) Slipping of the upper femoral epiphysis: A study of seventy cases. J Am Med Assoc 97:1867-1872. 74. Fineschi G, Guzzanti V (1986) Linear intertrochanteric osteotomyfor chronic slip of the upper femoral epiphysis. Internat Orthop 10:83-88. 75. Fiorani G (1881) Sopra una forma speciale di zoppicamento. Gazz d Osp 2:717-749. 76. Fish JB (1984) Cuneiform osteotomy of the femoral neck in the treatment of slipped capital femoral epiphysis. J Bone Joint Surg 66A:1153-1168. 77. Fish JB (1994) Cuneiform osteotomy of the femoral neck in the treatment of slipped capital femoral epiphysis. J Bone Joint Surg 76A:46-59. 78. Fisher RL, Waskowitz WJ (1972) Familial developmental coxa vara. Clin Orthop Rel Res 86:2-5. 79. Fixsen JA, Lloyd-Roberts GC (1974) The natural history and early treatment of proximal femoral dysplasia. J Bone Joint Surg 56B:86-95. 80. Floman Y, Yosipovitch Z, Light A, Viskoper RJ (1975)Bilateral slipped upper femoral epiphysis: A rare manifestation of renal osteodystrophy. Isr J Med Sci 11:15-20. 81. Frangenheim P (1909) Zur pathologie der osteoarthritis deformans juvenilis des huftgelenks, uber coxa vara und traumatische epiphysenlosung am oberen femurende. Beitr Klin Chir 65:19-53. 82. Frangenheim P (1911) Weitere untersuchungen uber die pathologie der coxa vara adolescentium. Beit Klin Chir 72: 239-255. 83. Freund E (1936) Congenital defects of femur, fibula and tibia. Arch Surg 33:349-391. 84. Gage JR, Sundberg AB, Nolan DR, Sletten RG, Winter RB (1978) Complications after cuneiform osteotomy for moderately or severely slipped capital femoral epiphysis. J Bone Joint Surg 60A:157-165. 85. Gelberman RH, Cohen MS, Shaw BA, Kasser JR, Griffin PR, Wilkinson RH (1986) The association of femoral retroversion with slipped capital femoral epiphysis. J Bone Joint Surg 68A: 1000-1007.

References 86. Gillespie R, Torode IP (1983)Classification and management of congenital abnormalities of the femur. J Bone Joint Surg 65B:557-568. 87. Glorion Ch, Norotte G, Rigault P, Padovani JP, Touzet Ph, Finidori G (1992) Caput valgum in children: Natural history and treatment in a series of 17 hips at skeletal maturity. J Orthop Surg 6:120-127. [Rev Chir Orthop 78:82-89, 1992.] 88. Golding C (1939) Congenital coxa vara and the short femur. Proc Roy Soc Med 32:641-646. 89. Gonzalez-Moran G, Carsi B, Abril JC, Albinana J (1998) Resuits after preoperative traction and pinning in slipped capital femoral epiphysis: K wires versus cannulated screws. J Pediatr Orthop 7:53-58. 90. Goodman WW, Johnson JT, Robertson WW (1996) Single screw fixation for acute and acute-on-chronic slipped capital femoral epiphysis. Clin Orthop Rel Res 322:86-90. 91. Grashey R (1910) Coxa vara retroflexa traumatica. Beitr Klin Chir 70:186-196. 92. Green WT (1945) Slipping of the upper femoral epiphysis. Arch Surg 50:19-32. 93. Griffith MJ (1976) Slipping of the capital femoral epiphysis. Ann Roy Coil Surg Eng 58:34-42. 94. Guzzanti V, Falciglia F (1991) Slipped capital femoral epiphysis: Comparison of a roentgenographic method and computed tomography in determining slip severity. J Pediatr Orthop 11:6-12. 95. Haedke M (1902) Zur aetiologie der coxa vara. Deutsche Zeit Chir 66:89-96. 96. Hagglund G, Hansson LI, Ordeberg G, Sandstrom S (1986) Slipped capital femoral epiphysis in southern Sweden: Longterm results after femoral neck osteotomy. Clin Orthop Rel Res 210:152-159. 97. Hagglund G, Hansson LI, Ordeberg G, Sandstrom S (1988) Bilaterality in slipped upper femoral epiphysis. J Bone Joint Surg 70B:179-161. 98. Hagglund G, Hansson LI, Sandstrom S (1987) Slipped capital femoral epiphysis in southern Sweden: Long-term results after nailing/pinning. Clin Orthop Rel Res 217:190-200. 99. Hagglund G (1996) The contralateral hip in slipped capital femoral epiphysis. J Pediatr Orthop Part B 5:158-161. 100. Hall JE (1957) The results of treatment of slipped femoral epiphysis. J Bone Joint Surg 39B:659-673. 101. Hamanishi C (1980) Congenital short femur. Clinical, genetic and epidemiological comparison of the naturally occurring condition with that caused by thalidomide. J Bone Joint Surg 62B:307-320. 102. Hansson LI, Hagglund G, Ordeberg G (1987) Slipped capital femoral epiphysis in southern Sweden. Acta Orthop Scand 58:1-67. 103. Harte RH (1897) Epiphyseal separation of the head of the femur, with report of a case. Trans Am Surg Assoc 15: 215-223. 104. Hartjen CA, Koman LA (1990) Treatment of slipped capital femoral epiphysis resulting from juvenile renal osteodystrophy. J Pediatr Orthop 10:551-554. 105. Hartman JT, Gates DJ (1972) Recovery from cartilage necrosis following slipped capital femoral epiphysis. A 7 year study of 166 cases. Orthop Rev 1:33-37. 106. Helbing C (1906) Die coxa vara. Zeit f Orthop Chir 15: 502-631.

457

107. Henrikson B (1969) The incidence of slipped capital femoral epiphysis. Acta Orthop Scand 365-372. 108. Heppenstall RB, Marvel JP, Chung SMK, Brighton CT (1974) Chrondrolysis of the hip. Clin Orthop Rel Res 103: 136-142. 109. Herman MJ, Dormans JP, Davidson RS, Drummond DS, Gregg JR (1996) Screw fixation of grade III slipped capital femoral epiphysis. Clin Orthop Rel Res 322:77-85. 110. Heyerman W, Weiner D (1984) Slipped epiphysis associated with hypothyroidism. J Pediatr Orthop 4:569-573. 111. Heyman CH, Herndon CH (1954) Epiphyseodesis for early slipping of the upper femoral epiphysis. J Bone Joint Surg 36A:539-554. 112. Heyman CH (1949) Treatment of slipping of the upper femoral epiphysis. A study of the results of 42 cases. Surg Gyn Obstet 89:559-565. 113. Heyman CH, Herndon CH, Strong JM (1957) Slipped femoral epiphysis with severe displacement: A conservative operative treatment. J Bone Joint Surg 39A:293-303. 114. Hoffa A (1905) Die angeborene coxa vara. Deut Med Woch 31:158-159. 115. Hofmeister F (1898) Zur pathologie und therapie der coxa vara. Beitr Klin Chir 21:299-364. 116. Howorth B (1957) Slipping of the upper femoral epiphysis. Clin Orthop Rel Res 10:148-173. 117. Howorth B (1966) History: Slipping of the capital femoral epiphysis. Clin Orthop Rel Res 48:11-32. 118. Howorth B (1966) Pathology: Slipping of the capital femoral epiphysis. Clin Orthop Rel Res 48:33-48. 119. Howorth B (1966) Treatment: slipping of the capital femoral epiphysis. Clin Orthop Rel Res 48:53-70. 120. Howorth B (1966) The bone-pegging operation. For slipping of the capital femoral epiphysis. Clin Orthop Rel Res 48:79-97. 121. Imhauser G (1957) Zur pathogenese und therapie der jugendlichen huftkopflosungz. Orthop 88:3-41. 122. Imhauser G (1977) Spatergebnisse der sog. Imhauser-osteotomie bei der epiphysenlosung. Z Orthop 115:716-725. 123. Ingram AJ, Clarke MS, Clark CS, Marshall WR (1982) Chondrolysis complicating slipped capital femoral epiphysis. Clin Orthop Rel Res 165:99-109. 124. Ippolito E, Mickelson MR, Ponseti IV (1981) A histochemical study of slipped capital femoral epiphysis. J Bone Joint Surg.63A: 1109-1114. 125. Ireland J, Newman PH (1978) Triplane osteotomy for severely slipped upper femoral epiphysis. J Bone Joint Surg 60B:390-393. 126. Jacquemier M, Noca P, Dick R, Bollini G, Moulia-Pelat JP, Migliani R, Faure F (1992) A study of femoral anteversion in slipped upper femoral epiphysis: A review of 25 cases. J Orthop Surg 6:9-15. [Rev Chir Orthop 77:530-536, 1991.] 127. Jensen HP, Steinke MS, Mikkelson SS, Thomsen PB (1990) Hip physiolysis. Bilaterality in 62 cases followed for 20 years. Acta Orthop Scand 61:419-420. 128. Jerre R, Billing L, Hansson G, Wallin J (1994) The contralateral hip in patients primarily treated for unilateral slipped upper femoral epiphysis. J Bone Joint Surg 76B:563-567. 129. Jerre R, Billing L, Karlsson J (1996) Loss of hip motion in slipped capital femoral epiphysis: A calculation from the slipping angle and the slope. J Pediatr Orthop Part B 5:144-150.

458

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities o f the Femur

130. Jerre R, Hansson G, Wallin J, Karlsson J (1996) Long-term results after realignment operations for slipped upper femoral epiphysis. J Bone Joint Surg 78B:745-750. 131. Jerre T (1950) A study in slipped upper femoral epiphysis: With special reference to the late functional and roentgenological results and to the value of closed reduction. Acta Orthop Scand Supp 6:1-157. 132. Joachimsthal G (1909) Ueber coxa vara. Bildung nach reposition der angeborenen Huftverrenkung. Munch Med Woch 1: 641-643. 133. Johanning K (1951) Coxa vara infantum: I. clinical appearance and aetiological problems. Acta Orthop Scand 21: 271-297. 134. Kalamchi A, Cowell HR, Kim KI (1985) Congential deficiency of the femur. J Pediatr Orthop 5:129-134. 135. Kallio PE, Mah ET, Foster BK, Paterson DC, LeQuesne GW (1995) Slipped capital femoral epiphysis. Incidence and clinical assessment of physeal instability. J Bone Joint Surg 77B: 752-755. 136. Kallio PE, Paterson DC, Foster B K, LeQuesne GW (1993) Classification in slipped capital femoral epiphysis. Sonographic assessment of stability and remodeling. Clin Orthop Rel Res 294:196-203. 137. Karol LA, Doane RM, Cornicelli SF, Zak PA, Haut RC, Manoli A (1992) Single versus double screw fixation for treatment of slipped capital femoral epiphysis: A biomechanical analysis. J Pediatr Orthop 12:741-745. 138. Keenan WNW, Clegg J (1994) Idiopathic bilateral slipped upper femoral epiphyses in a child under six years of age. J Bone Joint Surg 76B:495-496. 139. Keetley CB (1888) A case of rachitis adolescentium. Illus Med News 1:7-8. 140. Keetley CB (1900) Coxa vara. Lancet i:1115-1118. 141. Kelsey JL, Keggi KJ, Southwick WO (1970) The incidence and distribution of slipped capital femoral epiphysis in Connecticut and southwestem United States. J Bone Joint Surg 52A:1203-1216. 142. Kelsey JL, Acheson RM, Keggi KJ (1972) The body build of patients with slipped capital femoral epiphysis. Am J Dis Child 124:276-281. 143. Kelsey JL (1973) Epidemiology of slipped capital femoral epiphysis: A review of the literature. Pediatrics 51:1042-1050. 144. Key JA (1926) Epiphyseal coxa vara or displacement of the capital epiphysis of the femur in adolescence. J Bone Joint Surg 8:53-117. 145. Kiaer S (1948) Epiphyseolysis capitis femoris treated with inforation. Acta Orthop Scand 17:81-92. 146. Kibiloski LJ, Doane RM, Karol LA, Haut RC, Loder RT (1994) Biomechanical analysis of single- versus doublescrew fixation in slipped capital femoral epiphysis at physiological load levels. J Pediatr Orthop 14:627-630. 147. Kinoshita J, Kaneda K, Matsuno T, Hosokawa Y, Nagashio (1995) Slipped capital femoral epiphysis associated with hyperparathyroidism: A case report. Internat Orthop 19:245-247. 148. Kirmisson ME (1918) Coxa vara et obesite. Bull Acad Med 79:183-193. 149. Klein A, Joplin RJ, Reidy JA (1943) Treatment in cases of slipped capital femoral epiphysis at the Massachusetts General Hospital. Arch Surg 46:681-686.

150. Klein A, Joplin RJ, Reidy JA, Hanelin J (1953) Management of the contralateral hip in slipped capital femoral epiphysis. J Bone Joint Surg 35A:81-87. 151. Klein A, Joplin RJ, Reidy JA, Hanelin J (1953) Slipped Capital Femoral Epiphysis. Springfield, IL: Charles C Thomas. 152. Kleinberg S, Buchman J (1936) The operative versus the manipulative treatment of slipped femoral epiphysis. J Am Med Assoc 107:1545-1551. 153. Kocher (1894) Ueber coxa vara, eine berufskrankheit der wachsthumsperiode. Deut Zeit f Chir 38:521-548. 154. Koman LA, Meyer LC, Warren FH (1982) Proximal femoral focal deficiency: Natural history and treatment. Clin Orthop Rel Res 162:136-143. 155. Krahn TH, Canale ST, Beaty JH, Warner WC, Lourenco P (1993) Long-term follow-up of patients with avascular necrosis after treatment of slipped capital femoral epiphysis. J Pediatr Orthop 13:154-158. 156. Kramer WG, Craig WA, Noel S (1976) Compensating osteotomy at the base of the femoral neck for slipped capital femoral epiphysis. J Bone Joint Surg 58A:798-800. 157. Kredel L (1896) Coxa vara congenita. Zentralbl f Chir 23: 969-973. 158. Krempien B, Mehls O, Ritz E (1974) Morphological studies on pathogenesis of epiphyseal slipping in uremic children. Virch Arch Path Anat Histol 362:129-143. 159. Kruger DM, Herzenberg JE, Viviano DM, Hak DJ, Goldstein SA (1990) Biomechanical comparison of single- and doublepin fixation for acute slipped capital femoral epiphysis. Clin Orthop Rel Res 259:277-281. 160. Kulick RG, Denton JR (1982) A retrospective study of 125 cases of slipped capital femoral epiphysis. Clin Orthop Rel Res 162:87-90. 161. Lacroix P, Verbrugge J (1951) Slipping of the upper femoral epiphysis: A pathological study. J Bone Joint Surg 33A: 371-381. 162. Lance D, Carlioz A, Seringe R, Postel M, Lacombe MJ, Abelanet R (1981) La chondrolyse ou coxite laminaire juvenile apres epiphysiolyse femorale superieure. Rev Chir Orthop 67: 437-450. 163. Lange DR, Schoenecker PL, Baker CL (1978) Proximal femoral focal deficiency: Treatment and classification in 42 cases. Clin Orthop Rel Res 135:15-25. 164. Laplaza FJ, Burke SW (1995) Epiphyseal growth after pinning of slipped capital femoral epiphysis. J Pediatr Orthop 15:357-361. 165. Lehman WB, Menche D, Grant A, Norman A, Pugh J (1984) The problem of evaluating in situ pinning of slipped capital femoral epiphysis: An experimental model and a review of 63 consecutive cases. J Pediatr Orthop 4:297-303. 166. Le Mesurier AB (1948) Developmental coxa vara. J Bone Joint Surg 30B:595-605. 167. Lewin P (1928) An unusual roentgenographic finding in the hip. Am J Roent 19:290-291. 168. Loder RT, et al. (1996) The demographics of slipped capital femoral epiphysis. Clin Orthop Rel Res 322:8-27. 169. Loder RT, Arbor A, Richards BS, Shapiro PM, Reznick LR, Arbor A, Aronson DD (1993) Acute slipped capital femoral epiphysis: The importance of physeal stability. J Bone Joint Surg 75A:1134-1140.

References

170. Loder RT, Farley FA, Hernzenberg JE, Hensinger RN, Kuhn JL (1993) Narrow window of bone age in children with slipped capital femoral epiphyses. J Pediatr Orthop 13:290-293. 171. Loder RT, Wittenberg B, DeSilva G (1995) Slipped capital femoral epiphysis associated with endocrine disorders. J Pediatr Orthop 15:349-356. 172. Lowe HG (1961) Avascular necrosis after slipping of the upper femoral epiphysis. J Bone Joint Surg 43B:688-699. 173. Lubicky JP (1996) Chondrolysis and avascular necrosis: Complications of slipped capital femoral epiphysis. J Pediatr Orthop Part B 5:162-167. 174. Lucas RC (1883) Form of late tickets associated with albuminutia, tickets of adolescents. Lancet i:993-994. 175. Macausland AR (1935) Separation of the capital femoral epiphysis. J Bone Joint Surg 17:353-369. 176. Magnusson R (1954) Coxa vara infantum. Acta Orthop Scand 23:284-308. 177. Mann DC, Weddington J, Richton S (1988) Hormonal studies in patients with slipped capital femoral epiphysis without evidence of endocrinopathy. J Pediatr Orthop 8:543-545. 178. Martin PH (1948) Slipped epiphysis in the adolescent hip: A reconsideration of open reduction. J Bone Joint Surg 30A:9-19. 179. Mathiesen FR (1958) Slipping of the proximal femoral epiphysis. Acta Orthop Scand 27:115-125. 180. Maurer RC, Larsen IJ (1970) Acute necrosis of cartilage in slipped capital femoral epiphysis. J Bone Joint Surg 52A:39-40. 181. Maussen JPGM, Rozing PM, Obermann WR (1990) Intertrochanteric corrective osteotomy in slipped capital femoral epiphysis. Clin Orthop Rel Res 259:100-110. 182. Mayer L (1937) The importance of early diagnosis in the treatment of slipping femoral epiphysis. J Bone Joint Surg 19: 1046-1051. 183. McAfee PC, Cady RB (1983) Endocrinologic and metabolic factors in atypical presentations of slipped capital femoral epiphysis. Clin Orthop Rel Res 180:188-197. 184. Mehls O, Ritz E, Krempien B, Gilli G, Link K, Willich B, Scharer K (1975) Slipped epiphyses in renal osteodystrophy. Arch Dis Child 50:545-554. 185. Melby A, Hoyt W, Weiner DS, (1980) Treatment of chronic slipped capital femoral epiphysis by bone graft epiphysiodesis. J Bone Joint Surg 62A:119-125. 186. Milch H (1937) Epiphysiolysis or epiphyseal coxa anteverta. J Bone Joint Surg 19:97-116. 187. Milgram JW, Lyne ED (1975) Epiphysiolysis of the proximal femur in very young children. Clin Orthop Rel Res 110: 146-153. 188. Mirkopulos N, Weiner DS, Askew M (1988) The evolving slope of the proximal femoral growth plate relationship to slipped capital femoral epiphysis. J Pediatr Orthop 8:268-273. 189. Moore RD (1945) Aseptic necrosis of the capital femoral epiphysis following adolescent epiphyseolysis. Surg Gyn Obstet 80:199-204. 190. Moore RD (1945) Conservative management of adolescent slipping of the capital femoral epiphysis. Surg Gyn Obstet 80: 324-332. 191. Morrissy RT (1989) Principles of in situ fixation in chronic slipped capital femoral epiphysis. In: Instructional Course Lectures, American Academy of Orthopedic Surgeons. p. 257-262, St. Louis: CV Mosby.

459

192. Morrissy RT (1990) Slipped capital femoral epiphysis. Technique of percutaneous in situ fixation. J Pediatr Orthop 10: 347-350. 193. Morscher E (1960) Die operativ therapie der epiphyseolysis capitis femoris. Z Orthop 92:153-174. 194. Muller E (1888) Ueber die verbiegung des schenkelhalses im wachstumsalter. Ein neues krankheitsbild. Beitr Klinische Chir 4:137-148. 195. Muller E (1966) On the deflection of the femoral neck in childhood. A new syndrome. Clin Orthop Rel Res 48:7-10 (in English, partial article). 196. Newman PH (1960) The surgical treatment of slipping of the upper femoral epiphysis. J Bone Joint Surg 42B:280-288. 197. Nguyen D, Morrissy RT (1990) Slipped capital femoral epiphysis: Rationale for the technique of percutaneous in situ fixation. J Pediatr Orthop 10:341-346. 198. Nixon JR, Douglas JF (1980) Bilateral slipping of the upper femoral epiphysis in end-stage renal failure: A report of two cases. J Bone Joint Surg 62B: 18-21. 199. Noble TP, Hauser ED (1926) Coxa vara. Arch Surg 12: 501-538. 200. Nonweiler B, Hoffer M, Weinert C, Rosenfeld (1996) Percutaneous in situ fixation of slipped capital femoral epiphysis using two threaded Steinmann pins. J Pediatr Orthop 16: 56-60. 201. O'Brien ET, Fahey JJ (1977) Remodeling of the femoral neck after in situ pinning for slipped capital femoral epiphysis. J Bone Joint Surg 59A:62-68. 202. Ordeberg G, Hansson LI, Sandstrom S (1984) Slipped capital femoral epiphysis in southern Sweden: Long-term result with no treatment or symptomatic primary treatment. Clin Orthop Rel Res 191:95-104. 203. Orofino C, Innis JJ, Lowrey CW (1960) Slipped capital femoral epiphysis in Negros. J Bone Joint Surg 42A: 1079-1083. 204. Panting AL, Williams PF (1978) Proximal femoral focal deficiency. J Bone Joint Surg 60B:46-52. 205. Pappas AM (1983) Congenital abnormalities of the femur and related lower extremity malformations: Classification and treatment. J Pediatr Orthop 3:45-60. 206. Pare A (1572) Fractures of the neck of the femur. In: Cinq Livres de Chirurgie, Paris, and (1840) In: Oeuvres Completes, volume 2, book 13, chapter 21. Paris. 207. Pauwels F (1936) Zur therapie der kindlichen coxa vara. Ver Deut Orthop Ges 30:372. 208. Pavlov H, Goldman AB, Freiberger RH (1980) Infantile coxa vara. Pediatr Radiol 135:631-640. 209. Perkins G (1932) Treatment of adolescent coxa vara. Brit Med J i:55-56. 210. Peterson MD, Weiner DS, Green NE, Terry CL (1997) Acute slipped capital femoral epiphysis: The value and safety of urgent manipulative reduction. J Pediatr Orthop 17:648-654. 211. Petit JL (1709) In: Maladies des Os, p 354, Leyden, and (1837) In: Traite des Maladies des Os, Oeuvres Completes. p. 169. Paris. 212. Pirani S, Beauchamp RD, Li D, Sawatzky B (1991) Soft tissue anatomy of proximal femoral focal deficiency. J Pediatr Orthop 11:563-570. 213. Pomeranz MM, Sloane MF (1935) Slipping of the proximal femoral epiphysis. Arch Surg 30:607-624.

460

CHAPTER 5 ~ Coxa Vara in Developmental and Acquired Abnormalities of the Femur

214. Ponseti IV, McClintock R (1956) The pathology of slipping of the upper femoral epiphysis. J Bone Joint Surg 38A:71-83. 215. Ponseti I, Barta CK (1948) Evaluation of treatment of slipping of the capital femoral epiphysis. Surg Gyn Obstet 86: 87-97. 216. Pouzet F (1938) Le traitement de la coxa vara congenitale. La Presse Med 46:1095-1096. 217. Pouzet F (1934) L'evolution anatomique des aplasies du col femoral (coxa vara congenitales, a fissure verticale). Lyon Chir 31:712-724. 218. Pritchett JW, Perdue KD (1988) Mechanical factors in slipped capital femoral epiphysis. J Pediatr Orthop 8:385-388. 219. Purl R, Smith CS, Malhotra D, Williams AJ, Owen R, Harris F (1985) Slipped upper femoral epiphysis and primary juvenile hypothyroidism. J Bone Joint Surg 67B: 14-20. 220. Pylkkanen PV (1960) Coxa vara infantum. Acta Orthop Scand Supp 48:7-120. 221. Rammstedt C (1900) Uber die traumatische losung der femur kopfepiphyse und ihre folgeerscheinungen. Arch Klin Chir 61:559-583. 222. Rao SB, Crawford AH, Burger RR, Roy DR (1996) Open bone peg epiphysiodesis for slipped capital femoral epiphysis. J Pediatr Orthop 16:37-48. 223. Reikeras O, Bjerkreim I, Kolbenstvedt A (1982) Anteversion of the acetabulum in patients with idiopathic increased anteversion of the femoral neck. Acta Orthop Scand 53:847-852. 224. Rennie AM (1960) The pathology of slipped upper femoral epiphysis: A new concept. J Bone Joint Surg 42B:273-279. 225. Rey JC, Carlioz H (1975) Epiphysiolyses a grand deplacement. Reduction sanglante par la technique de Dunn. Rev Chir Orthop 61:261-273. 226. Ring PA (1961) Congenital abnormalities of the femur. Arch Dis Child 36:410-417. 227. Samuelson T, Olney B (1996) Percutaneous pin fixation of chronic slipped capital femoral epiphysis. Clin Orthop Rel Res 326:225-228. 228. Sanpera I, Sparks LT (1994) Proximal femoral focal deficiency: Does a radiologic classification exist? J Pediatr Orthop 14:34-38. 229. Schai PA, Exner GU, Hansch O (1996) Prevention of secondary coxarthrosis in slipped capital femoral epiphysis: A long-term follow-up study after corrective intertrochanteric osteotomy. J Pediatr Orthop 5:135-143. 230. Schlesinger A (1905) Zur aetiologie und pathologischen anatornie der coxa vara. Arch Klin Chir 75:629-642. 231. Schmidt TL, Kalamchi A (1982) The fate of the capital femoral physis and acetabular development in developmental coxa vara. J Pediatr Orthop 2:534-538. 232. Schreiber A (1963) Epiphyseolysis capitis femoris: Beitrag zur frage der beidseitigkeit-gleichzeitiges vorkommen von Wirbelsaulenveranderungen. Z Orthop 97:4-11. 233. Segal LS, Weitzel PP, Davidson RS (1996) Valgus slipped capital femoral epiphysis. Clin Orthop Rel Res 322:91-98. 234. Segal LS, Davidson RS, Robertson WW, Drummond DS (1991) Growth disturbances of the proximal femur after pinning of juvenile slipped capital femoral epiphysis. J Pediatr Orthop 11:631-637. 235. Serafin J, Szulc W (1991) Coxa vara infantum, hip growth disturbances, etiopathogenesis, and long-term results of treatment. Clin Orthop Rel Res 272:103-113.

236. Shapiro F, Simon S (1984) Slipped capital femoral epiphysis: Patient profile and treatment results. Orthop Trans 8:373. 237. Shea D, Mankin HJ (1966) Slipped capital femoral epiphysis in renal tickets. J Bone Joint Surg 48A:349-355. 238. Siegel DB, Kasser JR, Sponseller P, Gelberman RH (1991) Slipped capital femoral epiphysis. A quantitative analysis of motion, gait, and femoral remodeling after in situ fixation. J Bone Joint Surg 73A:659-666. 239. Skinner SR, Berkheimer GA (1978) Valgus slip of the capital femoral epiphysis. Clin Orthop Rel Res 135:90-92. 240. Sorensen KH (1968) Slipped upper femoral epiphysis: Clinical study on etiology. Acta Orthop Scand 39:499-517. 241. Southwick WO (1967) Osteotomy through the lesser trochanter for slipped capital femoral epiphysis. J Bone Joint Surg 49A:807-835. 242. Southwick WO (1984) Editorial: Slipped capital femoral epiphysis. J Bone Joint Surg 66A:1151-1152. 243. Speer DP (1982) Experimental epiphysiolysis: etiologic models of slipped capital femoral epiphysis. In: The Hip Society: The Hip. Proceedings of the 10th Open Scientific Meeting of the Hip Society. pp. 68-88, St. Louis: CV Mosby. 244. Sprengel (1898) Ueber die traumatische losung der kopfepiphyse des femur und ihr verhaltniss zur coxa vara. Arch Klin Chir 57:805-840. 245. Stambough JL, Davidson RS, Ellis RD, Gregg JR (1986) Slipped capital femoral epiphysis: An analysis of 80 patients as to pin placement and number. J Pediatr Orthop 6:265-273. 246. Stanitski CL (1994) Acute slipped capital femoral epiphysis: Treatment alternatives. J Am Acad Orthop Surg 2:96-106. 247. Sturrock CA (1894) The after results of simple separation of epiphyses. Edinburgh Hosp Rep 2:598-608. 248. Sutro CJ (1935) Slipping of the capital epiphysis of the femur in adolescence. Arch Surg 31:345-360. 249. Swiontkowski MF (1983) Slipped capital femoral epiphysis: Complications related to intemal fixation. Orthopaedics 6: 705-712. 250. Szypryt EP, Clement DA, Colton CL (1987) Open reduction of epiphysiodesis for slipped upper femoral epiphysis. J Bone Joint Surg 69B:737-742. 251. Teinturier P, Dechambre H (1968) Etude d'anteversion de la hanche de l'enfant. Rev Chir Orthop 54:545-551. 252. Terjesen T (1992) Ultrasonography for diagnosis of slipped capital femoral epiphysis. Acta Orthop Scand 62:653-657. 253. Tillema DA, Golding JSR (1971) Chondrolysis following slipped capital femoral epiphysis in Jamaica. J Bone Joint Surg 52A:1528-1540. 254. Tsou PM (1982) Congenital distal femoral focal deficiency: Report of a unique case. Clin Orthop Rel Res 162:99-102. 255. Velasco R, Schai PA, Exner GU (1998) Slipped capital femoral epiphysis: A long-term follow-up study after open reduction of the femoral head combined with subcapital wedge resection. J Pediatr Orthop Part B 7:43-52. 256. Waldenstrom H (1930) On necrosis of the joint cartilage by epiphysiolysis capitis femoris. Acta Chir Scand 67: 936-946. 257. Walker SJ, Whiteside LA, McAlister WH, Silverman CL, Thomas PRM (1981) Slipped capital femoral epiphysis following radiation and chemotherapy. Clin Orthop Rel Res 159: 186-193.

References 258. Walter H (1931) Zur genese der coxa vara. Med Klin 27: 1071-1073. 259. Waiters R, Simon R(1980) Joint destruction: A sequel of unrecognized pin penetration in patients with slipped capital femoral epiphyses. In: The Hip: Proceedings of the Eighth Annual Open Scientific Meeting of the Hip Society. pp. 145-164, St. Louis: CV Mosby. 260. Ward WT, Wood K (1990) Open bone graft epiphyseodesis for slipped capital femoral epiphysis. J Pediatr Orthop 10:14-20. 261. Ward WT, Stefko J, Wood KB, Stanitski CL (1992) Fixation with a single screw for slipped capital femoral epiphysis. J Bone Joint Surg 74A:799-809. 262. Warner WC, Beaty JH, Canale ST (1996) Chondrolysis after slipped capital femoral epiphysis. J Pediatr Orthop Part B 5: 168-172. 263. Weber BG (1965) Die Imhauser-osteotomie bei floridem gleitprozess. Z Orthop 100:312-320. 264. Weighill FJ (1976) The treatment of developmental coxa vara by abduction subtrochanteric and intertrochanteric femoral osteotomy with special reference to the role of adductor tenotomy. Clin Orthop Rel Res 116:116-124. 265. Weiner DS, Weiner S, Melby A, Hoyt WA, Jr (1984) A 30year experience with bone graft epiphysiodesis in the treatment of slipped capital femoral epiphysis. J Pediatr Orthop 4: 145-152. 266. Weinstein JN, Kuo KN, Millar EA (1984) Congenital coxa vara. A retrospective review. J Pediatr Orthop 4:70-77.

461

267. Wells D, King JD, Roe TF, Kaufman FR (1993) Review of slipped capital femoral epiphysis associated with endocrine disease. J Pediatr Orthop 13:610-614. 268. Whitman R (1894) Observations on bending of the neck of the femur in adolescence. NY Med J 59:769-774. 269. Whitman R (1909) Further observations on injuries of the neck of the femur in early life; with reference to the distinction between fracture of the neck and epiphyseal disjunction as influencing positive treatment. Med Rec 75:1-8. 270. Wiberg G (1959) Considerations on the surgical treatment of slipped epiphysis with special reference to nail fixation. J Bone Joint Surg 41A:253-261. 271. Wilson PD (1938) The treatment of slipping of the upper femoral epiphysis with minimal displacement. J Bone Joint Surg 20:379-399. 272. Wilson PD, Jacobs B, Schecter L (1965) Slipped capital femoral epiphysis: An end-result study. J Bone Joint Surg 47A: 1128-1145. 273. Wolf EL, Berdon WE, Cassady JR, Baker DH, Freiberger R, Pavlov H (1977) Slipped capital femoral epiphysis as a sequela to childhood irradiation for malignant tumors. Radiology 125:781-784. 274. Zadek I (1935) Congenital coxa vara. Arch Surg 30:62-102. 275. Zahrawi FB, Stephens TL, Spencer GE, Clough JM (1983) Comparative study of pinning in situ and open epiphysiodesis in 105 patients with slipped capital femoral epiphyses. Clin Orthop Rel Res 177:160-167.

CHAPTER

6

Epiphyseal Disorders of the Knee Distal Femur, Proximal Tibia, and Proximal Fibula

I. II.

Normal Developmental Variability

VI.

Osteochondritis Dissecans of Distal Femur

VII.

III. Infantile Tibia Vara (Biount's Disease) IV. Adolescent Tibia Vara V. Osgood-Schlatter Disease (Tibiai Tubercle Chronic Traumatic Apophysitis)

VIII.

I. N O R M A L D E V E L O P M E N T A L VARIABILITY

Congenital Dislocation of the Knee

ValgusAngulation Following Proximal Tibial Metaphyseal Fractures in Childhood Disorders of the Proximal Fibular Epiphysis

both femurs and tibias can be done, preferably in the standing position if the patient is old enough, to measure the degree of bowing and to rule out any underlying disorder. Radiographic findings in those with more marked but still physiologic genu varum include medial beaking of the distal femoral and proximal tibial metaphyses, thickening of the medial tibial cortex, and a slightly underdeveloped wedgeshaped medial secondary ossification center of both distal femur and proximal tibia (Figs. 2A and 2B). The two most common pathological causes of childhood bowing are tickets and the skeletal dysplasias. Even in severe cases of physiologic bowing improvement should be seen between 2 and 3 years of age; if this is not occurring, concern is raised about a pathologic disorder or a developing tibia vara (Blount's disease).

A. Physiologic Genu Varum and Genu Valgum in Childhood Bow legs are a normal feature of early childhood development and are commonly referred to as physiologic genu varum. Salenius and Vankka (143), in a study involving 1480 radiographic examinations in growing children, documented diminution of the femoral-tibial diaphyseal angle from a mean of 15~ varus at birth, to neutral at 24 months, to 10~ valgus at 36 months, and to an eventual 5 - 6 ~ valgus by 6 years (Fig. 1). The varus can be as high as 30-40 ~ before age 2 years and still undergo spontaneous correction. The clinical appearance is worsened by internal tibial torsion in many. Heath and Staheli (69) established the normal limits of knee angulation in 196 white children from 6 months to 11 years of age. Varus and valgus were measured from clinical photographs taken in a standardized position with markers over the anterior superior iliac spines and the centers of the patellae. The children were maximally bowlegged at age 6 months, progressed toward neutral knee angles by 18 months, showed maximal knock-knees of 8~ around 4 years, and then decreased gradually over the next few years to less than 6 ~ at 11 years. They considered bowlegs after age 2 years to be abnormal and knock-knees of more than 12~ abnormal. This developmental sequence of genu varum at birth, its disappearance during the 1st or 2nd year of life, a change to genu valgum during the 3rd or 4th year, and establishment of the final alignment of slight valgus at about 6 years of age has been well-recognized by many for some time. Due to the high likelihood of spontaneous correction, no specific treatment is needed. Anteroposterior radiographs of

B. Normal Radiographic Developmental Variants of the Distal Femoral and Proximal Tibial Epiphyses There are many developmental irregularities in the radiologic appearance of the distal femoral epiphysis, particularly of the secondary ossification center, which are well within a normal range and must not be mistaken for pathological processes. (Figs. 3A-3C) The irregularities include (1) rough or serrated margins of the secondary ossification center, (2) thin bony protuberances from the margins of the secondary center, and (3) small accessory ossification centers (30, 114, 145, 157, 230). At 1-2 years of age, there often is a serrated border of the secondary ossification center particularly at its medial, lateral, or distal margins. This border on occasion is present circumferentially and may be concentrated and particularly irregular at the medial aspect. In the older child, in whom the secondary ossification center is 462

SECTION i ~ Normal D e v e l o p m e n t a l V a r i a b i l i t y

+20 ~

463

DEVELOPMENT OF THE TIBIO-FEMORAL ANGLE DURING GROWTH

+15 ~

VALGUS

Or)

rr +10o

+ 5~ _

13 yrs

o o

.5 ~

(.9 ..J X.lo ~

FIGURE

1

Development of the tibiofemoral angle during growth with varus angulation shown above the horizontal line

and valgus angulationbelow the line. [Reprintedfrom Salenius and Vankka(143), withpermission.]

developed more extensively, there often are X-ray changes at the distal margins of the secondary center suggestive of an osteochondritis dissecans lesion (Fig. 3C). These irregularities are asymptomatic, disappear with time, and should not be considered as pathologic lesions. Caffey and associates defined three variants of these developmental irregularities in the older child (23). Group 1 femurs show varying degrees of roughening of the margins of the secondary ossification centers, occasionally with separate small foci of calcification within the cartilage just beyond the roughened edge. These often are present medially and laterally and are particularly well-demonstrated on the tunnel views. Group 2 lesions are larger localized marginal irregularities in the form of concave indentations. These generally tend to be on the inferior surface of the bone adjacent to the condyles or at their more lateral aspects and are not seen in the characteristic medial regions, as is the classical osteochondritis dissecans lesion. Group 3 changes feature irregularities similar to type 2 but with an independent island of bone in the marginal crater. Caffey et al. concluded that "uneven marginal ossification of the femoral condyles is a feature of healthy chondral bone formation." These essentially are transitory and the incidence of marginal irregularities falls progressively with age from 85% at age 4 years to 10% at age 12 years. Roughly 30% of all males and 17% of all females had defects sufficiently marked to place them in the group 2 and 3 categories. Many of these irregularities are

present on the posterior aspect of the bone adjacent to the femoral condyles and are detected only with tunnel views. Irregularities of the type 1 group also have been described by Sontag and Pyle and were almost invariable in a study of 220 children from 1 to 72 months of age in the sense that some irregularity of outline or texture of the distal femoral epiphysis was seen at some time (157). In the more marked groups 2 and 3, the X-ray changes were similar to the changes observed in symptomatic osteochondritis dissecans. Irregularities in marginal smoothness were detected quite early by some but were not recognized by all. Both Ludloff (114) and Christ (30) demonstrated that there was a period, generally extending from age 2 to 6 years, in the development of the femoral epiphysis when its margins normally had considerable unevenness, described as roughness in general with occasional spiny projections or protuberances in particular. These initially appeared on the medial side, which almost invariably was rough during the early years, and in the 5th year there often were irregularities although not as pronounced on the lateral margin. The protuberances occurred infrequently on the articular margins of the secondary ossification center, and after the 6th year such roughness generally had disappeared. Ludloff, as early as 1903, pointed out the existence of irregular condylar margins up to the age of 4 years, noting a serrated outline during the 1st and 2nd years of life and the more prominent protuberances from age 2 to 4 years, both medially and laterally.

464

CHAPTER 6 " Epiphyseal Disorders of the Knee

A

Physiologic Genu Varum 1. Diaphyseal Varus (Femoral- Tibial)

Metaphyse 5,d~ Beaking /

Epiphyseal~<~ Wedging" - ?

_ . Medial Cortical "1/ ~Thickening//

2. Medial Metaphyseal Beaking (d. femur, p. tibia) 3. Medial Epiphyseal Wedging (d. femur, p. tibia) 4. Medial Cortical Thickening (tibia)

d = distal p =proximal

F I G U R E 2 Plain radiographic characteristics of physiologic genu varum are shown. (A) The major characteristic radiographic features of physiologic genu varum are illustrated. (B) A plain radiograph with normal alignment of the right knee and physiologic genu varum of the left. Note the increased thickness of the medial tibial cortex on the bowed side compared to the normal side and the internal tibial torsion on the bowed side in comparison to the normal (proximal fibula superimposed on proximal tibia).

An extremely detailed radiographic study on epiphyseal growth and ossification in the knee was presented by Scheller in 1960 (145). His investigation involved radiographic studies on 876 children, 473 boys and 403 girls, from the first year of life to 15 years of age. He paid particular attention to the ossification pattern of the secondary ossification centers and carefully documented irregularities from the smooth pattern. He defined marginal irregularities as (1) rough or uneven, (2) spiny processes referred to as protuberances, which measured at least 1 mm in height, and (3) bony nuclei of varying size referred to as accessory ossification centers, which were encountered immediately adjacent to the secondary ossification center. At the distal femur at birth and during the first year of life, the secondary ossification center of the epiphysis was elliptical in shape in both frontal and lateral projections. In the lateral projection it retained this shape throughout the first year, but toward the end of the first year

in the frontal projection it began to present an oval configuration with the blunt end directed laterally and the more or less pointed end medially. In the elliptical stage, therefore, the epiphysis showed no signs of an intercondylar fossa and lacked any demarcation between the femoral condyles; with the first indication of the intercondylar fossa a shallow depression in the distal margin between the lateral and medial condyles could be distinguished. It was not, however, until after the first year of age that this distinction invariably was seen. From the 2nd to the 4th years the medial condyle secondary center was almost triangular in shape, whereas the lateral was roughly rectangular. By the 4th year in most instances the bone of both condyles was on the whole rectangular and remained so throughout skeletal growth. Scheller's detailed study at the distal femur indicated that the medial and lateral margins were more or less rough during most of the first decade of development with the medial margin more uneven than the lateral and for a longer period of time. Medial margin protuberances exceeding 1 mm in length were more frequent and larger than those on the lateral margin. The articular margins were never as rough as the medial or lateral margins. In the first two years of life, there was virtual universal roughness or serrated appearance of the margins. The protuberances were most common medially in boys from 2 to 8 years of age and in girls from 1 to 4 years of age, whereas laterally the protuberances were most common in boys from 2 to 6 years of age and in girls from 1 to 4 years of age. The protuberances were quite common and could be seen in anywhere from 50 to 100% of those studied in the years of high incidence. The accessory ossification centers, which often were only 1-2 mm in size and scarcely detectable, were seen in boys at the distal femur in 17% of cases, with most concentrated between 1 and 10 years of age, and in girls in only 5% of cases again concentrated between 1 and 10 years of age. The accessory centers tended to be medial and lateral in position up until 4 - 5 years of age, after which they tended to be present adjacent to the articular surfaces at the margins of the secondary center. With further growth of the secondary center the accessory centers were incorporated into the main bony mass. The accessory centers tended to be situated between protuberances or in marginal depressions of the serrated edges. Scheller clearly pointed out, however, that no relationship could be demonstrated between accessory ossification centers in the knee and osteochondritis dissecans (OD). He felt that his study established that the accessory centers did not occur in parts of the femoral condyles corresponding to the sites of prediliction of the OD lesions. Proximal tibial epiphyseal development also was studied. In the frontal projection the shape of the proximal secondary center of the tibia was rounded during the first half of the first year and elliptical with a more or less flattened distal surface during the second half of the first year. Epiphyses of elliptical shape tended to persist for 2 - 3 years. Between the 3rd and 10th years a somewhat triangular shape to the secondary center was present with the base adjacent to the

SECTION II ~ Osteochondritis Dissecans of Distal Femur

465

A

9-'.:.

,i,

FIGURE 3 Characteristicdevelopmentalirregularities of the distal femoral secondary ossification center are illustrated. (A) An example of a thin bony protuberance is shown on both the medial and lateral sides of the secondary center. [Reprinted from Christ, N. (1929). Arch. Orthop. Unfall.-Chir. 27:610-630, copyright notice of Springer Verlag, with permission.] (B) Serrated medial borders of the distal femoral secondary ossification center are shown [derived from Ludloff (114)]. (C) Radiograph from a slightly older patient shows a characteristic irregularity on the inferior margin of the secondary ossification center bone, which on occasion has been mistaken for an osteochondritisdissecans lesion.

physis. Its prevalence diminished beginning with the 6th and 7th years, and by 10 years of age the secondary center had a rectangular structure. Shaping irregularities of the medial and lateral margins of the proximal tibial secondary ossification center were frequently seen, but the articular margins rarely were irregular to any appreciable extent. The medial margin was consistently more uneven than the lateral margin in the first 2 years of life and serrated or rough edges were seen in virtually all individuals, a pattern that continued through most of growth particularly medially. Protuberances, however, were less common and less extensive than in the femur. They were seen in roughly one-fourth of boys between the ages of 4 and 10 years, with appearance before or after that virtually unseen. In girls medial protuberances were seen in only 10-20%, concentrated between the ages of 1 and 8 years. Lateral tibial margin protuberances were seen as well in the same age groups at a slightly higher level of frequency. Articular margin protuberances were rare. Ac-

cessory ossification centers were seen in approximately 10% of boys concentrated between 1 and 10 years of age and only 3% of girls concentrated mainly between 1 and 5 years of age. The large majority of proximal tibial accessory ossification centers were present medially or laterally. Scheller observed that the distal femoral and proximal tibial secondary ossification centers initially were narrower than the adjacent metaphysis but that the two became equally wide in the 5th year in girls and in the 7th year in boys.

II. O S T E O C H o N D R I T I S DISTAL FEMUR

DISSECANS

OF

A. Disease Profile Osteochondritis dissecans is an epiphyseal disorder in which a localized segment of subchondral bone undergoes necrosis

466

CHAPTER 6 ~ Epiphyseal Disorders o[ the Knee

and demarcation from the surrounding normal bone, and with fracture and failure of repair it can become separated, along with the overlying articular cartilage, partly or completely from the joint surface forming an osteocartilaginous loose body or joint mouse. The disorder is seen most commonly in the knee (distal femur), but also in the elbow (capitellum) or ankle (talus), and even more rarely in the proximal humerus and proximal femur. As many as 85% of reported cases have involved the knee. Usually it is unilateral, but bilateral symmetrical lesions have been described. The large majority of cases are nonfamilial, but isolated cases of familial involvement have been described. Other cases are associated with endocrinopathies or certain skeletal dysplasias, such as multiple epiphyseal dysplasia and diastrophic dysplasia. The male:female ratio is approximately 2.5:1 in most series with both sides affected equally. Bilateral cases are seen in approximately one-third of patients. The symptoms involve mild to moderate knee discomfort, usually of several months duration, and quadriceps muscle atrophy. Loose body formation can occur spontaneously, but on occasion a traumatic event leads to breaking free of the fragment. Locking and buckling then can occur. The lesion has been recognized for well over a century, and the more general recognition of the occurrence of loose bodies in the knee extends back over 300 years (57, 136). There was early recognition of the fairly frequent occurrence of loose osteochondral fragments within the knee joint and much discussion of the underlying causes. Studies from the mid-nineteenth century were usually unclear regarding the traumatic component of any episode of loose body formation. During the late nineteenth century, there was awareness of acute osteochondral fractures as well as osteochondromatosis of synovial origin; in the course of assessment of these disorders the osteochondritis dissecans entity was recognized and defined separately. The first clear-cut description is widely attributed to Paget (134) in 1870, and the disorder was defined as osteochondritis dissecans by Koenig (92) in 1887. Barth (11), however, reviewed cases from the French literature as early as 1834. Detailed reviews were provided by Wolbach and Allison (175) in 1928 and by Wagoner and Cohn (170) in 1931 at a time when osteochondritis dissecans was a clinically and radiologically well-defined entity.

B. Original Descriptions by Paget, Teale, and Koenig Paget described two patients in whom he defined what is now referred to as osteochondritis dissecans of the knee (134). Both presented with knee discomfort and a loose body noted on arthrotomy. He referred to the focal disorder as "quiet necrosis." Paget presented the cases to indicate that "necrosis of bone may take place, and the dead bone may be exfoliated without the usually attendant suppuration and other signs of destructive inflammation." The fragment removed contained both cartilage and bone, and Paget com-

mented that "this body looked exactly like a piece of the articular cartilage of one of the condyles of the femur." Histologic sections were made in which the cartilage surface appeared normal and the cell and matrix appearances were "exactly like those of articular cartilage." He indicated that "this loose body was a piece of the cartilage, together with a very small portion of bone of one of the condyles of the femur." His description of the histologic structure of the cartilage of the loose body left no doubt that it was articular cartilage. Paget felt that acute trauma did not account for the disorder but the cause was quiet necrosis and that the loose bodies "are sequestra, exfoliated after necrosis of the injured portion of cartilage, exfoliated without acute inflammation." He attributed the lack of inflammation to the fact that the lesion was in articular cartilage. Because "its substance being without blood vessels," it was not surprising that inflammatory changes did not occur. Injuries to articular cartilage distinct from acute trauma went on to necrosis and exfoliation with no signs of destructive inflammation. The disorder occurred such that "after injury to a previously healthy joint, a loose body is found in it having shape and general aspect and texture of a piece of articular cartilage with or without some portion of sub-adjacent bone." Paget's brief paper differentiated these loose bodies from those derived from "abnormal overgrowths of cartilage whether in the changes of chronic rheumatoid arthritis or in those of the dendritic growths of synovial fringes." Paget subsequently wrote that "Since my paper on loose cartilages in the knee was printed, I have learned that the late Mr. Teale published an explanation of their occurrence similar to mine in the 39th Volume of Medico-Chirurgical Transactions. I regret to have overlooked a paper with which I ought to have been acquainted." Teale, 15 years earlier in 1855, published a report "Case of Detached Piece of Articular Cartilage Existing as a Loose Substance in the Knee Joint" (161). The disorder was in a 37-year-old man who injured his knee and was unable to work for 3 weeks, following which he was fine other than having symptoms of a knee disorder. At arthrotomy 1 year later, a loose body was removed that was "flattened, circular in form, and irregular or ragged at its border." One of its surfaces had the appearance of cartilage and was smooth and slightly convex; the other was concave and rough from a layer of bone. The patient developed an infection and died, allowing for a postmortem knee examination. "At the under surface of the inner condyle, the articular cartilage showed a circular depression about 88 in depth having a rough surface of bone at its base." On comparison of this breach in the articular cartilage with the fragment that had been removed, they were found to correspond accurately with each other; on placing the detached fragment in the cavity in the condyle, the continuity of the articular surface was restored. It was evident that the loose body was a portion of the articular cartilage along with a thin layer of bony substance. Teale concluded that a "portion of articular cartilage and of the adjoining layer of

SECTION II ~ Osteochondritis Dissecans of Distal Femur

467

F I G U R E 4 Illustration of the common site of involvement of a distal femoral osteochondritis dissecans lesion. (A) The distal femoral lesion described by Teale (161), as illustrated in the work of Fisher (57), is shown. This represents a relatively large lesion. (B) The characteristic OD lesion is illustrated on the outer aspect of the distal femoral medial condyle adjacent to the point of attachment of the cruciate ligament [derived from (115)].

bone had been injured by the a c c i d e n t . . , and that by a slow process of exfoliation extending through a period of about 12 months, the injured part was cast off and became loose in the joint." The distal femur and loose body described by Teale were retained in a pathology collection and were illustrated by Fisher in his 1921 article on loose bodies (Fig. 4A) (57). Koenig reasoned that the slow nature of the onset and the mildness of the early symptoms were inconsistent with the idea of a single definitive trauma and that the more likely etiologic feature was a minor original trauma whose effects were intensified by a continuance of exercise (92). He postulated that a subchondral fracture was the initial relatively minor injury. The articular cartilage is devoid of nerve supply and there is very little sensation in the cancellous bone, both of which led to the fact that subchondral injury could be accompanied by little or no pain. Koenig felt that the disorder was associated with a poorly defined pathological process, which led to the gradual separation of the bone fragment. He implied that it was inflammatory in nature, hence the use of the term osteochondritis, but indeed no evidence was presented at that time nor has any been seen subsequently to define it as a primary inflammatory disorder. Nevertheless, his clinical observations clearly defined the entity as different from traumatic or osteodegenerative disorders. He indicated that the loose bodies formed were "an entirely circumscribed disease of the joint ends which has been described as osteochondritis dissecans." The loose bodies were formed without any injury and they separated from the joint ends "in consequence of a process as yet unexplained." Koenig pointed out the focal nature of the disorder and that

other than a slight fluid effusion the joints looked perfectly sound.

C. Three Stages of the Disorder Three stages of the disorder as described by Conway and others are outlined here (33). 1. STAGE I The lesion is present and there may be a fairly welldemarcated prominence of the articular surface with the articular cartilage covering the elevation continuous with the rest of the cartilage surface but of a different color. This prominence can be separated easily and beneath it there is an excavation of the cancellous bone and chondral portion of the articular end of the bone. 2. STAGE II The fragment has become separated more distinctly but still lies within the excavated area of the articular surface, being held only by some adhesion bands. The fragment can be moved easily and the cartilage over it is discolored compared with the rest of the normal cartilage. In addition, the cartilage and underlying bone are not firmly attached to the main body of the cancellous bone but give the appearance of having been dissected off, which led Koenig to use the term osteochondritis dissecans. 3. STAGE III This stage is characterized by the complete sequestration of the fragment from its place on the articular surface and its

468

CHAPTER 6 9 Epiphyseal Disorders of the Knee

displacement into the joint cavity. The cavity itself is lined by a thin layer of reddish-gray tissue with no pathological cell profile seen on histologic section. The cartilage in many instances continues to be nourished through synovial fluid or by blood vessels within the stalk of a synovial pedicle to which it may become attached. Degenerative changes in the fragment may be slight. In general, however, "both the articular bone and cartilage tend to a general necrosis." The tendency is to form "fibrocartilage along the surface of separation." In some cases, the proliferative changes may be considerable with evidence of new bone formation seen. The old bone has become necrotic. It is often the case that an injury allows for displacement of the fragment from its crater, but most now feel that the entire process cannot be explained by a single injury. Joint exploration at surgery with examination for the loose body and the crater make it clear that the condition is one of long duration.

D. Age of Occurrence The primary age at onset of symptoms is variable but is seen between 5 and 25 years of age. The two largest age groups are those between 10 and 15 years of age prior to physeal closure, referred to as the juvenile form, and those between 15 and 20 years of age after physeal closure, which evolve into the adult forms. The average age at occurrence in the juvenile form in major series has ranged from 11.3 to 13.4 years and the disorder is seen infrequently under 8 years of age.

E. Regions of Involvement of Distal Femur The distal femur is the most common site of occurrence. The most common lesion is shown in Fig. 4B. The regions affected were well-described by Aichroth and in most other series the findings fit into the distribution he outlined (Fig. 5) (1). He defined six areas of occurrence, with the most common being the lateral aspect of the medial condyle of the distal femoral epiphysis. Lesions of the medial condyle accounted for 85% of the disorders and those of the lateral condyle 15%. A total of 69% of the lesions was at the classical lateral aspect of the medial condyle, 6% were defined as "extended classical" being somewhat larger than those normally seen, and 10% were on the inferocentral surface of the medial femoral condyle. On the lateral side, 13% were on the inferocentral part of the lateral condyle and 2% were anterior on the lateral condyle. A relatively large number of patellar lesions occur but these will not be discussed here. Mubarek and Carroll documented 122 juvenile OD lesions, 62% of which were on the lateral aspect of the medial femoral condyle, 22% were on the central aspect of the lateral condyle, and 16% were patellar (124). Virtually all femoral lesions are located in the middle to posterior one-third of the articular surface as defined on lateral radiographs. Linden

documented 80% of the femoral lesions in the medial condyle with no difference in position between childhood and adult groups (109).

F. Etiology Although there is no universally accepted cause for osteochondritis dissecans, three primary causes have been suggested. 1. TRAUMATIC The most commonly accepted cause of osteochondritis dissecans is trauma, with most feeling that a single acute episode is not at fault but rather a series of repetitive injuries to the osteochondral region that cause a crushing or fissuring of the subchondral bone with subsequent bone necrosis. The subchondral bone fails to heal to the adjacent normal bone due to the continuing stresses upon it. The articular cartilage, which initially is normal because it receives its nutrition by diffusion from the synovial fluid, eventually demarcates from the surrounding normal cartilage due to the failure of support from the necrotic and collapsing subchondral bone. This allows for the osteochondral fragment to be defined, and this fragment then may be partially or fully extruded into the joint to form the loose body. The school supporting trauma or injury was defined early by Paget (134), Koenig (92), and Fairbank (54) among others, although varying impressions of the nature of the trauma have been presented. These include acute trauma and repetitive subclinical trauma. Some liken the injury to an osteochondral fracture in which there is injury to a focal area of cartilage and bone in association with a single traumatic episode. Should this not be sufficiently great to incapacitate the individual, continued activity would lead to additional damage and inability of the focus to heal. Two types of repetitive trauma have been suggested as causative of the disorder. The first involves impingement of the medial articular facet of the patella against the lateral aspect of the medial condyle particularly in association with stressful and high-level athletic activities. Aichroth demonstrated the feasibility of this mode of occurrence in a cadaver study in which blue dye was placed on the medial articular facet of the patella and, with the normal patellofemoral articulation intact, the knee joint was fully flexed. Increasing flexion of the joint "demonstrated the areas of contact between the patella and the medial femoral condyle" such that in full flexion the classical site for OD was stained. With an anterior blow on the flexed knee, osteochondral fractures could be expected from the impact of the patella upon the medial condyle. Osteochondral fractures could be sustained in any area of the joint in which areas of osteochondritis dissecans had been identified. The second source of repetitive trauma particularly to the medial condyle at its lateral aspect is a prominent tibial spine, which can be shown to impinge against the region in question particularly with the internal rotation of the tibia in the final few degrees of knee extension. This idea was presented by Fairbank (54)

SECTION II 9 Osteochondritis Dissecans o f Distal Femur

MEDIAL CONDYLE / 85 PER CENT

469

|

CLASSICAL

EXTENDED CLASSICAL

69 PER CENT

6 PER CENT

INFERO - CENTRAL I0 PER CENT

LATERAL CONDYLE 15 PER CENT

INFER0 - CENTRAL 13 PER CENT

ANTERIOR 2 PER CENT

FIGURE 5 Distribution of sites of occurrence of distal femoral osteochondritis dissecans lesions as outlined by Aichroth [reprinted from (1), with permission]. Most series show a similar distribution.

and strongly supported by Smillie (155). Aichroth created experimental osteochondral fractures in the distal femur of skeletally mature rabbits and stabilized them with varying degrees of effectiveness (2). Those that failed to heal had a radiographic and histologic appearance similar to that seen in human OD. He thus concluded that "the fragment in osteochondritis dissecans follows an osteochondral fracture which remains ununited."

2. GENETIC Families with a predisposition for osteochondritis dissecans have been described. Bernstein reported OD in three family members each with bilateral knee involvement (16). These would appear to represent subtle examples of skeletal dysplasias. Most studies concerning a familial incidence of osteochondritis dissecans document an autosomal dominant pattern of heredity with a very high degree of expression. The large majority of patients have a short stature syndrome, although the radiographs shown, other than for the osteochondritis dissecans, do not show evidence of malformation of the long bones other than that in the subchondral bone region. Most of the osteochondritis dissecans in these cases are multijoint and short stature is a common finding, both of which point to an underlying skeletal dysplasia as being the inciting cause. A study by Mubarak and Carroll assessed 31 members of a family in which there were 12 proven and 8 possible cases detected (123). The most frequent site of origin was the patella, but lesions also were noted on both medial and lateral femoral condyles. Of the 20 patients, 10 had evidence of short stature, which was defined as a height less than the 5th percentile. Similar findings were defined by White (172). His three patients were a woman aged 28 years

who was only 4'4" tall and developed lesions in the medial femoral condyle of the knee and hip, a man aged 21 years who was 4'6" tall and developed osteochondritis dissecans in the capitellum and both knees with the lesion appearing to arise in the lateral condyle of the femur, and a woman aged 21 years who was 4'5" tall and had lesions in both elbows as well as one knee and ankle. Hanley et al. presented a family with OD and documented 10 additional families that had been reported in the literature from 1925 to 1963 (67). Virtually all of the patients reported had knee involvement and many had elbow involvement as well. These are also the two major joints affected in isolated OD apparently not connected to familial disorders. The familial osteochondritis dissecans has been associated with other structural malformations in addition to the short stature, such as an abnormal facial appearance, congenital ptosis of the eyelids, peculiar external ears, bony fusion of the manubriosternal joint, and short fifth fingers and fifth toes. The existence of well-documented familial cases, although infrequent in relation to the large number of patients with OD, indicates either a multifactorial etiology for OD or at least the likelihood that some underlying disorder predisposes one to the development of lesions. 3. CIRCULATORY DISTURBANCES The third theory considers the osteonecrosis of the subchondral bone as being primary due to circulatory disturbances with the subsequent articular cartilage loosening secondary to the bone necrosis. This theory was quite common in the early decades of the century, but current interpretations indicate the bone necrosis to be secondary rather than primary. Ludloff postulated a focal area of bone necrosis due

470

CHAPTER 6 9 Epiphyseal Disorders of the Knee

to injury of a small arterial vessel entering the bone adjacent to the posterior cruciate ligament (115). Axhausen clearly illustrated the avascular necrosis of bone and postulated that embolic phenomena, benign infected emboli, caused the necrosis (5-9).

G. Pathogenesis and Pathoanatomic Findings 1. EARLYDESCRIPTIONS Detailed understanding of osteochondritis dissecans began with the early observations of Paget (134) and Koenig (92). Koenig, in an article published in 1887 based on clinical and anatomic findings in three cases, recognized that many of the loose joint bodies originated with pieces becoming displaced from the subchondral bone and cartilage regions of the epiphyses. He developed the term osteochondritis dissecans, by which he implied an inflammation of the bone and cartilage. The origin was not specifically traumatic, although he did feel that secondary loosening of the abnormal bone and cartilage focus could occur after injury. Koenig proposed a gradual mode of formation in situ through reactive dissection processes in the vicinity of a previously damaged region. Experiments by Hildebrand (1895) following the creation of free cartilage fragments and free bone and cartilage bodies in the joints of goats showed that avascular cartilage tissue remained viable via synovial fluid nutrients, although he was skeptical about the assumption of similar nutritional patterns for bone (74). Over the next few years, the fact that osteochondritis dissecans was a separate and specific condition became evident. Cases were reported by Riedel in which he indicated that foreign bodies were caused not only by trauma or as a consequence of arthritis deformans but also through the osteochondritis dissecans mechanism (140). An early case was described in which there was a narrow cleft approximately 1 cm long from the articular surface of the capitellum inward ending within the bone. He felt that with further time separation of bone and cartilage would have occurred. There was no trauma reported. Schmeiden found no evidence of a spontaneous dissecting joint inflammation but noted that not all joint mice came from the synovial tissue (147). Ludloff felt that in osteochondritis dissecans the same region of bone always was affected, secondary to traumatic vascular damage (115). The segment of bone died and was loosened gradually, following which the overlying cartilage became affected. Further mild trauma allowed for cracking into the necrotic segment, which eventually broke free into the joint and formed a loose body. The pathogenesis of the disorder was gradually coming into focus with recognition that it involved the loosening of cartilage and bone fragments in otherwise healthy knee joints without the initial occurrence of any evident trauma. With time, however, the fragments displaced leading to the formation of free osteochondral bodies. The disorder occurred preferentially at the distal femur at a typical site along the outer side of the inner or medial femo-

ral condyle in the intercondylar fossa adjacent to the insertion of the posterior cruciate ligament. 2. AXHAUSEN Axhausen began to define his views on osteonecrosis in relation to disorders of the skeleton in 1912 and continued to write extensively on this matter over the next several years (5). The focus of his investigations was on defining aseptic avascular bone necrosis at a clinical and histopathological level. He developed experimental models for bone necrosis in the lower epiphysis of the femur of the dog using electrolytic needles to produce focal necrosis. His histologic reproductions from his 1912 article clearly define the various cell and tissue events in bone necrosis and repair with extraordinary clarity from the dog model and a human case with bone necrosis due to syphilitic infection. Necrotic bone was defined by intact matrix and empty osteocyte lacunae. Repair served to allow for synthesis and deposition of new bone on a scaffold of persisting dead bone. The subsequent reaction originated in the marrow of the subchondral layer which became transformed into proliferating connective tissue. The cartilage became necrotic and was partially absorbed and partially was dissected, much like Koenig has suggested. Bilateral cases of osteochondritis dissecans began to be described, although it was only when a piece of cartilage and bone had become detached that symptoms arose. Axhausen wrote an important paper entitled "The Occurrence and Significance of Epiphyseal Nutritional Interruptions in Man" (8). Additional formation of loose bodies came from sources other than circumscribed traumatic lesions. He indicated that a diagnostic sign of this typical cartilage-bone body is that a layer of closed, nonfissured joint cartilage firmly attached covers a generally wedge-shaped or oval necrotic dead piece of bone, which includes marrow. Often there is no interruption at all in the continuity of the joint cartilage, whereas the piece of dead bone is isolated completely by a remnant of young connective tissue. Axhausen decisively rejected causation by an initial traumatic event. He explained the origin of osteochondral loose bodies by necrosis due to the embolic obstruction of epiphyseal terminal arteries due to bacterial emboli in which infection or osteomyelitis did not occur but there was a mechanical effect obliterating the local vessels. The subsequent events were as described by Koenig. Axhausen recognized that the primary lesion was a focal necrosis of the epiphyseal bone. This was followed by proliferation of the surrounding tissues, which began to penetrate the dead bone. A compression fracture of the necrotic bone interrupted the regenerative invasion and prevented further healing. The necrotic bone fragmented with further pressure and began to undergo slow resorption. The necrotic segments were demarcated by granulation tissue, and a slight injury could free the necrotic part to become a loose body in the joint. Axhausen also recognized and described this mechanism as underlying Perthes disease of the femoral head.

SECTION II ~ Osteochondritis Dissecans o f Distal Femur

3. KONJETZNY

Konjetzny felt that the osteochondroses such as Perthes and osteochondritis dissecans presented a disease picture based on necrosis of the subchondral epiphyseal bone marrow with no or slight damage to the joint cartilage (93). He felt that the disorder had nothing to do with any primary fracture and that the subchondral epiphyseal necrosis had to be attributed to acute or subacute vascular blocks. He felt there was no acute fracture but that traumatic insults in a broader sense might play an etiologic role. Much of what happened clearly was speculative, although it was suggested that traumatic damage to the joint capsule with its vessels could be an etiologic cause. 4. BARTH AND KAPPIS AND THE MECHANICAL SCHOOL OF CAUSATION

Barth proposed direct traumatic causation for osteochondritis dissecans, feeling that these were essentially osteochondral fractures (11). He felt that Koenig's explanation of a spontaneous necrosis with subsequent displacement of the fragment had not been proven. Direct traumatic causation in this era was accepted by many, but others could not reproduce the lesion experimentally with single traumatic episodes and the impression that repetitive trauma was causative came into prominence. Kappis indicated that the lesions were "consistent eventually with well-healed fractures" (88). No evidence was found for any pathological fracture, tickets, or osteitis fibrosa. He felt that mechanical conditions play a large role in the origin of the condition. He pointed out that tangential and rotational forces could act on the convex surface of the condyle and traumatize the cartilage and adjacent bone. He did postulate some predisposition to the disorder as well as the fact that the cartilage and bone away from periosteum were insensitive. Mechanical pressure (repetitive) also was suggested as causative by Phemister (136), Hellstrom (70), Fisher (57), yon Dittrich (168), and Wolbach and Allison (176). On a more specific note, Freiberg observed that the tibial spines were elongated in three of his cases; using radiographic and cadaver studies he noted tibial spine impingement at the classical medial condylar-lateral aspect site when the knee was flexed and the tibia externally rotated (58). Bernstein noted that the tibial spine was elongated, deformed, and approximated the intercondylar notch (16). Burckhardt and Schmidt (as quoted by Wagoner and Cohn) separately demonstrated patellar pressure at the medial and lateral condyles with knee flexion greater than 60 ~ and the leg rotated internally and externally, respectively (170). If these mechanical positions were present in stressed circumstances, the likelihood of traumatic impingement sequelae was greater. Kappis disagreed with Axhausen, feeling that the existence of bone necrosis was not necessary in the earliest stages and that the cases of loosening of joint bodies from the articular surfaces that did not have traumatic or arthritic antecedents were caused by "cartilage or bone factors which occur as a result of unknown injury or by me-

471

chanical means." He continued to write on "the significance of traumatic mechanical influence in the origin of circumscribed epiphyseal diseases." Kappis felt strongly that the disorders could originate in a bone break and that fracture was capable of explaining all of the difficulties that arise in these disorders at sites particularly exposed mechanically especially during the growth period. Bone could break even in the presence of ordinary demands. Gebele articulated the theory, much of which appears reasonable today, that "repeated mechanical incidents with mild contusion and compression of the part of the joint on which the greatest demand is placed lead to an injury to the bony trabeculae of the spongy tissue with hemorrhage, displacement of the vessels, and ischemia. This is followed by necrosis of the bone-cartilage section, aseptic inflammation and demarcation" (59). It thus appears that most investigators identified the same two or three etiologic features but disagreed as to which was primary and which secondary in order of importance. Many observers thus felt that mechanical stress on the bone trabeculae was the cause of subsequent necrosis. In summary, differing views were presented on the causation of the disorder, including an osteochondritis (Koenig) (92), tearing fracture [Barth (11) and Kappis (88) and others], and vascular obstruction caused by mycotic embolism [Axhausen (12) and Konjetzny (93)], or injury to a small end artery entering the bone adjacent to the posterior cruciate ligament (Ludloff) (115). Excellent summaries of loose body formation and pathology were published by Phemister (136) and Fisher (57). 5. FURTHER CLINICAL AND PATHOLOGICAL CORRELATIONS

Wagoner and Cohn described the gross and microscopic appearance of an osteochondritis dissecans lesion of the distal femur quite consistent with other reports (170). The loose body from the classic site on the lateral aspect of the medial femoral condyle was ovoid, glistening white, and with smooth rounded edges. It was composed of a thick layer of articular cartilage and a thin layer of subchondral bone. A corresponding oval crater on the femoral condyle near the point of attachment of the posterior cruciate ligament extended through the articular cartilage into the bone. Its size and shape corresponded to those of the loose ovoid fragment. The margins of the crater were rounded. The adjacent synovium was hypertrophied slightly, inflamed, and hypervascular. The loose body consisted histologically of cartilage and bone. The cartilage appeared viable and structurally normal. The bone was necrotic without evidence of osteoblastic activity. Evidence of bone resorption was seen but osteoclasts were not demonstrated. Liebman and Iseman observed that the medial femoral fragment often was held in place by a small bridge of intact articular cartilage or synovium (108). Both the crater bed and the adjacent surface of the lesional bone often were covered

472

CHAPTER 6 ~ gpiphyseal Disorders of the Knee

with fibrocartilage. If the loose body remained attached to the synovium, it often experienced continued growth of both cartilage and bone segments. An excellent review of the theories and early pathoanatomic studies on osteochondritis dissecans by Nagura is detailed particularly in relation to the German literature (125). He described a case of osteochondritis dissecans of the capitellum in a 23-year-old with onset of the disorder at 16 years of age. A job involving repetitive heavy lifting with the elbow extended was felt to be causative. Histology demonstrated a thick zone of separation between the cartilage-bone fragment and underlying spongy bone tissue. On the normal side of the defect, the bone was viable and showed continuing evidence of new bone formation. On the fragmented side, there were regressive changes involving extensive deterioration of the bone with fissures, bruising, and osteonecrosis. The intervening zone of demarcation was composed of cartilage referred to as a cartilage callus. Nagura experimentally produced an interruption in the continuity of subchondral bone that was not in contact with the periosteum and subsequently noted the formation of a zone of embryonic cartilage or mature cartilage in the gap. As far as osteochondritis dissecans was concerned, the disorder was "neither a dissecting inflammation--that is a gradual formation of joint mice in loco by reactive dissecting processes in the vicinity of a primarily damaged region without primary fracture (Koenig); nor an immediate detachmentmthat is a splitting or tearing fracture (Barth); nor yet a primary osseous necrosis caused by bacterial embolism induced vascular obstruction (Axhausen)." He defined a dynamic but chronic destructive-reparative process "resulting from an original not very significant interruption in the continuity of the subchondral bone tissue and its filling with cartilaginous callus." He defined an initial causative destructive lesion, which was not very significant but led to a primary fracture following which a reparative response was mounted characterized by a filling by cartilaginous callus of the defect site and its ossification by the endochondral sequence. The next or secondary destructive element was mediated by continuing mechanical stress, which led to repeated breaks in continuity and resulted in further changes such as fibrillation of the cartilaginous callus zone and resulting dissection. This then triggered the second reparative or restorative process, which attempted to further repair the cartilage callus that was repeatedly being injured. A relatively insignificant interruption to the subchondral spongy bone occurred, which could be complete or incomplete, although the incomplete types later progressed to a complete break within the subchondral bone. This subchondral bone cleft then was filled with embryonic cartilage causing a demarcation zone seen radiographically, which emanated from the spongy bone marrow. Continued use of the joint led to pressures on the joint surface, which led to increased cartilage callus in the tissue demarcation zone. This led to protrusion of the bone-cartilage fragment from the joint surface, and the subsequent lack of congruity led

to further structural damage. The greater the mechanical stresses, the more the embryonic cartilage tissue developed into cartilage tissue and even fibrocartilage. With continued use of the joint, further destructive processes set in, which finally led to bone death and gradual detachment of the fragment, now referred to as a joint mouse, into the joint space. Nagura thus feels that trauma occurred first and necrosis only later. The reverse view would imply that vascular changes came first leading to necrotic bone, and then necrotic bone itself was subject to microfractures with repeated stress due to its inability to repair itself. Nagura felt that "the cartilaginous zone originates as a consequence of an interruption in the continuity of bone tissue." He summarized that on the basis of histologic and experimental investigations osteochondritis dissecans could be attributed to a "succession of secondary destructive and constructive processes resulting from an original not very significant interruption in the continuity of the subchondral bone tissue and its filling with cartilaginous callus" (125). Jaffe felt that the subchondral osteonecrosis occurs as a consequence of traumatic squashing of the articular bone end (82). The excessively large medial tibial spine may act as the traumatizing force. Even if the site is intact on gross inspection, the osteochondritis dissecans lesion is outlined by an encircling furrow in the articular cartilage, which indeed may show some cracks and tears. The cartilage may be abnormally soft and discolored. The cartilage may be of normal thickness but more often is excessively thick. The subchondral bone usually is yellowish-white due to its lack of vascularity and may be covered by fibrous and fibrocartilaginous connective tissue if it has broken away from the underlying bone. Microscopically, the overlying cartilage tends to be normal at least in the early stages. Chondrocytes are present. There can be some fibrillation and degeneration of the surface and perhaps some calcification. "The subchondral osseous tissue shows substantial or complete aseptic necrosis." The marrow too is completely or partially necrotic and the intertrabecular marrow spaces usually contain at least some granular debris. "The subchondral spongy trabeculae are usually abnormally thick whether the constituent osseous tissue is partially viable or completely necrotic." "One gains the impression that the trabeculae have been thickened through periodic deposition of new bone." This gives the appearance of "periods of partial restoration of the local blood supply with consequent deposition of new bone on the existing trabeculae which had been partially or completely devitalized." Milgram studied 50 operated cases of OD of the distal femur in an effort to correlate radiologic findings, observations at the time of surgery, and histopathological studies of the excised specimens (121). He concluded that the etiology of OD lay in a "traumatic episode," although this often was not recognized clinically, and that the avascular necrosis was a secondary occurrence. Specimen radiographs also were used in the study. Most of his patients were young adults whose distal femoral epiphyses were closed. The fact that

SECTION II ~ O s t e o c h o n d r i t i s Dissecans o f Distal Femur

the study population involved those having surgery and the fact that many were skeletally mature indicate the more serious nature of the adult disorder than those prior to skeletal maturity, as described in studies by Green and Banks (64), Lofgren (113), and Wiberg (173). In 22 cases the fragment remained in the condylar defect, although it was only partially attached in most. There were 7 transitional cases with some fragments in the condyle and some free loose bodies, and there were 21 cases with loose bodies and empty femoral condylar defects. The radiographic findings of the lesions in situ were described. The following findings in different cases were noted: (1) generally there was subchondral bone attached to the articular fragments; (2) degenerative calcification of the articular cartilage occurred such that the radiodensity was not completely due to the subchondral bone fragment alone; (3) the lesion was displaced from the base of the persisting femur; and (4) frequently there was revascularization and healing of the fragment to the underlying bone. Milgram stressed that in approximately half of the cases radiologic density was due to degenerative calcification of the adjacent articular cartilage and not solely due to the attached subchondral bone. Often there were areas of new bone formation in the lower parts of the articular cartilage fragment. At surgery the osteochondral lesion was noted to be composed of a smooth convex cartilaginous surface and a roughened inner side that was the site of bony separation. When the fragment separated completely and became a loose body, there was a tendency to rounding or smoothing of the entire fragment. The thickness of bone on the inner surface of the cartilage was quite variable. In some instances no bone was seen at all, whereas in others the thickness of the bone was greater than the thickness of the articular cartilage. Milgram felt that the variation in thickness of the bony portion suggested fracture clefts at different distances from the articular surface. The histologic appearance of the bone in the OD lesion also was variable. In many instances the bone was necrotic and showed no changes indicative of prior remodeling. In several instances there was necrotic bone, but evidence that this had been remodeled implied several insuits prior to fragment separation. In some specimens central necrotic trabeculi were covered by new layers of bone containing viable osteocytes including a viable bone marrow, indicating that repair had occurred prior to separation. Occasional groups also were seen in which no bone necrosis was present with the bone appearing viable. Articular cartilage invariably was present in each specimen. The absence of avascular necrosis in some specimens also speaks of a frankly traumatic occurrence. If the specimen occurred adjacent to the synovial soft tissues at the inner margins of the intercondylar notch, the likelihood was greater that some blood supply to the fragment persisted and either prevented avascular necrosis or allowed for its early repair if it occurred. Milgram reviewed the literature in relation to the question of whether persisting osteonecrosis existed in the crater, by which is meant that bone that persisted adjacent to the defect site. In five specimens in which the bone of the

473

base of the lesion of OD was examined histologically, evidence of bone infarction invariably was absent. This too indicates a traumatic lesion for the OD defect with the osteonecrosis secondary. In some lesions the bony fragment was lined by a zone of avascular fibrocartilage, the tissue termed callus by Nagura. The chondrocytes appear viable. Milgram concluded that trauma, often unnoticed because of the insensitivity of cancellous bone, seems to be the initiating factor for the production of the cleft in the subchondral bone. Subsequent trauma, often repetitive, causes progressive disruption of the defect site and may interrupt the repair process perhaps several times. Motion induces the formation of Nagura's cartilaginous callus in the cleft. Some cases go on to union and revascularization, but in others a loose body eventually is formed. The prolonged nature of the series of events leads to the variable findings, which at one end of the spectrum have only necrotic bone and also intermediate levels of necrotic bone covered by new bone on the trabecular surfaces. Milgram also notes that some of the specimens consisted only of articular cartilage. The free osteochondral bodies that arise in association with the disease can grow and progressively calcify by means of surface proliferative changes nourished by the joint fluid. Indeed half of the specimens of OD contain no subchondral bone but only articular cartilage. There may be a fragmentation within the OD lesion. Generally there is sclerotic bone at the base of the crater, indicating repair phenomena in the normal epiphysis adjacent to the defect. Chiroff and Cooke studied six excised fragments from OD lesions from the lateral aspect of the medial femoral condyle (29). Each of these was composed of viable bone and articular cartilage overlying viable bone, which rested in the bed of fibrocartilage. Endochondral ossification was seen histologically. The authors felt that the OD lesion indicated a reparative process. The cartilage was felt to be viable and intact at all levels, and the bone of the lesion was not necrotic but viable with a normal population of osteoblasts, osteoclasts, and marrow cells. Fibrocartilage between the lesion and existing normal bone of the distal femoral epiphysis appeared to be actively engaged in endochondral ossification. Chiroff and Cooke concluded that repair was well underway even in patients up to 30 years of age and concluded that, "Surgery for the classical, attached lesion appears to be contra-indicated on the basis of the pathologic changes encountered in the present series of patients." This series, although small, had one patient 13 years of age with each of the five others between 18 and 29 years. Two of the patients had loose bodies, whereas in the others the cartilage was intact. The loose bodies or the OD lesions in situ were removed because of continuing symptoms. This report is valuable in showing that the lesions that are stable and undisplaced are undergoing a repair phenomenon, which should be encouraged by conservative management. Green and Banks presented photomicrographs from a girl 15 years of age with a medial femoral condylar osteochondritis dissecans and a large sequestrated bone fragment in situ

474

CHAPTER 6 ~ Epiphyseal Disorders of the Knee

(64). At arthrotomy "the cartilage at the site of the lesion showed a greater color than normal and was cracked along one margin of the subchondral defect." Microscopic sections showed the articular cartilage to be well-preserved except in the immediate area of the fracture. Most of the subchondral bone was not viable. They considered the basic process to be aseptic (avascular) necrosis of the subchondral bone with all other changes secondary. In many areas, there was fibrous tissue replacement, whereas in others there was osteoid tissue and new bone. The bone was of a woven conformation with areas of early lamellation indicating a repair process. In some areas, dead bone was being resorbed and replaced with fibrous tissue. If the patient continued to walk on the involved knee, the overlying cartilage cracked and a loose body was thrown into the joint. If immobilization was instituted, the dead bone was repaired and the living articular cartilage remained intact. The histopathology is not considered to differ greatly in juvenile as compared with early adult cases, although the healing of lesions while the physes are still open is welldocumented to be much better than healing after physeal closure. An excellent study using surgically excised cylindrical specimens of 14 young adult patients with osteochondritis dissecans from the distal femur has been reported by Linden and Telhag (110). They found that the lesions were composed of normal hyaline articular cartilage, which showed no pathological changes, whereas the underlying subchondral bone showed evidence of a fracture either immediately adjacent to the cartilage or as far as 1 cm from it. The cartilage and its chondrocytes appeared to be quite normal. Staining using both hematoxylin-eosin as well as Safranin-O showed a virtually normal pattern. There were well-defined fissures within the subchondral bone, scattered areas of osteonecrosis as defined by empty lacunae within the bone, and disorganized trabeculae in great disarray. Some of the lacunae were empty but in others normal osteocytes could be seen. There was osteoblastic and osteoclastic activity, indicating both repair and resorption occurring simultaneously. In some preparations, an increased number of osteoid seams was seen, which is a sign of regeneration. One generally saw areas of "newly formed connective tissue, at various stages of differentiation," with some demonstrating a fibrocartilage appearance. Although osteonecrosis clearly is seen, it is unclear whether this is due to ischemia as a prime initiating cause of the disorder or due to a subchondral fracture with viability impaired by continuing activity, which served to damage any repair tissue ingrowth and lead to a vicious cycle of weak necrotic bone. Both the structured appearance of the cells and matrix of the articular cartilage and its thickness were normal in all preparations. They felt that "the sequence of events is primarily in the bone and not in the cartilage." They concluded that "osteochondritis dissecans in adults seems to represent some form of subchondral fracture with intact articular cartilage initially. If the healing process is restricted, buildup of the fragment gives

rise to cracks in the cartilage and its eventual degeneration. With further damage, the bone and cartilage fragment separates first partially and then completely at which time it becomes loose in the joint and the knee has problems referable not only to the crater left in the articular and subchondral region but also due to the loose body itself."

H. Current Understanding of the Disease Entity The pathogenesis of an OD lesion occurring due to chronic repetitive subclinical trauma is outlined in Fig. 6. The delimited osteochondritis dissecans lesion is composed of articular cartilage, which is thicker than the neighboring cartilage. The cartilage appears viable, having chondrocytes in its lacunae, and maintains its normal structure. Fragments of subchondral bone persist with the cartilage, but the marrow spaces contain granular debris and fatty nonviable tissue, whereas the lacunae in the bone are empty. The deep surface of the lesion is sealed off from surrounding normal bone by a layer of reparative tissue, which can be fibrous, fibrocartilaginous, cartilaginous, or fibro-osseous in structure. In OD lesions that repair either spontaneously or with conservative management, ingrowth of vascularized bone tissue reestablishes normal bone and marrow tissue. Those loose bodies that show new bone on dead bone imply a repair process that was underway but ultimately ineffective. In many lesions, there is no evidence of effective repair. With progression of the disorder an avascular segment of bone separates from normal condylar bone, and the remaining crater consists of fibrous tissue and vascularized bone in its bed. The initial inciting cause appears to be an osteochondral or subchondral fracture caused by focal injury to the joint surface, which remains nonunited and with repetitive trauma leads to eventual separation of a fragment of bone with the overlying cartilage included. In Aichroth's series, the size of the fragment was quite variable with the average size being 2 x 2 cm and one reaching 4 x 4 cm. The fragment usually is single but in some cases there are two or three bony pieces. The cartilage of the femoral condyle is intact relatively early on, although it is somewhat soft to manual palpation due to collapse of the underlying bone. In general, however, there is a slight roughening of the articular cartilage surface. On occasion, a clear line of demarcation between the cartilage overlying the bony defect and the adjacent cartilage can be seen. This can be increasingly marked, leading to tears in the cartilage, partial separation of the fragment, and, on occasion, complete separation of the fragment leading to a loose body within the joint. The abnormality first becomes evident radiographically as an elliptically shaped rarefaction of the peripheral or lateral border of the medial distal femoral condyle, and after a time this lesion becomes free at the outer peripheral border. Thus, the articular surface generally is normal or close to normal appearing when the disorder is first detected, even though the bone defect can be seen clearly radiographically. It is important to determine whether

SECTION I! ~ O s t e o c h o n d r i t i s Dissecans

of Distal Femur

475

Pathogenesis of Osteochondritis Dissecans Lesions I~ . . . .

9 Subchondral

Fracture

IV

~S~~,

~ 9 Continued Micromotion

Occult Fracture as Initial Event

II

sc

ar

Loose Body Extrusion 9 Imperfect Repair of Fracture with Reactive Tissue (RT) RT = cartilage, fibrocartilage, fibrous or fibro-osseous tissue 9 Sclerotic Bone (Sc) 9 Micromotion Continues

i

9 N = necrotic bone V = viable cartilage

OD Stable- In Situ

9 Cartilage Surface Softened +_ Discoloured

III j,

Break in Cartilage Surface

Loose Body Migration

9 Separation of Fragment from Epiphyseal Bed

OD Unstable- In Situ FIGURE 6

The pathogenesis of an OD lesion occurring due to chronic repetitive subclinical trauma is illustrated.

the OD lesion is stable or unstable. This term applies only to lesions seen to be completely nondisplaced in situ. A stable lesion implies that the articular cartilage is intact; an unstable lesion implies that there are cracks in the articular surface and that motion occurs at the radiolucent repair tissue interface, often with a true gap or space created between the lesion and the base of the crater in epiphyseal bone.

I. Radiographic and Other Imaging Findings The radiographic appearances reflect the "phenomena of delimitation, loosening, and detachment of the osteochondral body" (82) (Figs. 7A-7C). The most common region of involvement is the lateral portion of the medial femoral condyle. If the articular cartilage is intact and the defect is both stable and non-displaced, the elliptical bone lesion often is outlined by a faint narrow zone of increased radiolucency. The adjacent bone may be sclerotic, flattened, or interrupted by small radiolucent cysts. With time, the density of the lesion increases and the radiolucent margin becomes better defined. With displacement, a loose body may be found. The tunnel view in the anteroposterior plane often is helpful to

reveal the lesion more clearly, but information is obtained from each of the anteroposterior, lateral, and skyline views. More exact definition of the lesions is available by tomogram, arthrogram, CT scanning, or MR imaging (43, 44, 117). The choice of imaging modality is dependent on available facilities plus the type of information needed. Arthrograms can detect partial displacement by determining either an irregular surface or dye beneath the cartilage surface between the bone fragment and the adjacent epiphyseal bone. Plain tomography or computerized tomography (CT) more clearly defines the extent of the lesion, the radiolucent cleft surrounding the lesion, and sclerotic, cystic, or flattening reactions of the adjacent epiphyseal bone. Magnetic resonance imaging also outlines the structural changes and provides evidence of vascularization patterns especially with gadolinium enhancement. Correlative studies have been done to analyze the mechanical stability of OD lesions using combinations of plane radiographs, scintigraphy, MR imaging, and arthroscopic intervention. Some clear parameters have emerged in terms of assessing the underlying stability without the need for direct arthroscomy or arthrotomy. Although these investigations

476

CHAPTER 6 ~ Epiphyseal Disorders o f the Knee

(A-C) Radiographic examples of osteochondritis dissecans of the distal femur are shown (arrows). Lesions in situ characteristically are in a subchondral position surrounded by a zone of radiolucency and also usually by a region of relative sclerosis of the persisting epiphyseal bone bed. FIGURE 7

are not absolute, a high degree of certainty is attached to them. Mesgarzadh et al. noted that all lesions smaller than 0.2 cm 2 were stable, whereas those that were 0.8 c m 2 or larger were loose (119). In addition, all lesions with a sclerotic margin of 3 mm or greater are loose. These two assessments can be made by plane radiography. MR imaging is particularly helpful in resolving the stability issues. A reliable sign of loosening is felt to be the presence of fluid at the interface of the fragment and its parent bone, a finding noted in all loose OD lesions subsequently confirmed by arthroscopy. This finding is analogous to the previous observation made by arthrograms in which an articular cartilage defect

allowed fluid to position itself within the subchondral bone, implying looseness. Scintigraphy or bone scanning can be used but it is felt to be less reliable as a specific indicator. Uptake is increased at the periphery of the lesion, which is indicative of the sclerotic or reactive bone. It remains difficult to quantify, however. Studies by DeSmet et al. also found the MR imaging to be highly useful in assessing stability (43, 44). They established four criteria for instability that had a high degree of correlation with arthroscopic findings and clinical outcome. These signs were (1) presence of a line of high signal intensity at the interface between the OD fragment and adjacent bone, (2) articular fracture indi-

SECTION II ~ O s t e o c h o n d r i t i s Dissecans o f Distal Femur

cated by high signal joint fluid passing through the subchondral bone plate, (3) focal OD defect with joint fluid, and (4) a 5-mm or larger fluid-filled cyst deep to the lesion. Any one of these four findings was indicative of an unstable lesion, which then warranted more aggressive and earlier therapy. The high signal image between the persisting bone and the OD defect is variably interpreted either as fluid, indicating continuity with the joint cavity, or as fibrovascular tissue, indicating a lack of bony repair. The size of the lesion also has been found to correlate well with results. The smaller the lesion the better the result. Guidelines are available from published studies. Cahill et al. found an average area of 309 mm 2 in successfully treated lesions and an average of 436 mm 2 in lesions that failed clinical management (24). Hughston et al. documented an average OD of 424 mm 2 for knees with good or excellent results and an average area of 815 mm 2 for those with poor results (78). The final study by DeSmet et al. found good outcomes when the lesion average was 194 mm 2 and poorer outcomes in those that averaged 647 mm 2 (44).

J. Age at Occurrence, Treatment, and Relation to Healing In those individuals who develop the osteochondritis dissecans disorder prior to closure of the distal femoral epiphyseal growth plate, the incidence of healing either spontaneously or with conservative treatment is significantly higher than in those who develop the disorder after skeletal maturation. Spontaneous healing in childhood osteochondritis dissecans was shown to occur by Decker (40), Wiberg (173), Hellstrom (70), Hellstrom and Ostling (71), van Demark (167), Lofgren (113), and Green and Banks (64) among others in juvenile cases in which only immobilization was used. These observations were interpreted as supportive of a traumatic etiology. Lofgren specifically studied OD in young patients (113). He observed that repair of OD lesions in children and adolescents was possible without surgical operation and indeed that repair invariably occurred. In addition the repair phenomenon was greatly quickened by immobilization of the limb for a period of time. He described nine cases with the age at time of diagnosis varying between 6 and 19 years. He concluded that "Spontaneous repair in children at adolescence is not only possible, but there is actually a tendency to spontaneous recovery." Treatment was by use of a long leg cast for a variable period of time. The recognition was becoming more widespread that OD in children and adolescents with the growth plates open had a more favorable prognosis in terms of repair than in those in young adults following growth plate closure. Both Wiberg and Lofgren clarified use of the term "spontaneous." In some instances authors were referring to healing of the defect in the absence of any treatment, whereas others used it to refer to healing in association with specific conservative management such as diminution of weight beating, which generally involved use of the long leg cast and/or crutches. Both of these authors

477

pointed out that the rate of repair was greatly enhanced by the use of immobilization, particularly early in the course of the symptomatic phase of the disorder. Clear delineation of the disorder in the pediatric population was made by Green and Banks in 1953 (62). They assessed 36 lesions in 27 patients with the age at onset from 4 to 15 years. There were 32 knees, 3 elbows, and 1 ankle affected. They demonstrated excellent results in the large majority of patients undergoing nonoperative treatment in which the limb was protected with either a cylinder cast or a patten-bottom brace. In 17 of 18 joints treated conservatively, results were excellent clinically and radiologically at an average of 4.5 years posttreatment and a minimum of 1 year follow-up. Their work led to the wider recognition that osteochondritis not only had a significant incidence in childhood but that prior to skeletal maturity nonoperative treatment "may be expected to lead to spontaneous healing without residual deformity," such that "in the majority of instances, surgical intervention is not indicated." Healing with protection from direct weight bearing was relatively quick and could "occur even in patients in whom there is apparent sequestral formation within the confines of the cavity of the osteochondritis lesion." Linden reported a long-term study of childhood onset osteochondritis dissecans averaging 33 years after the time of diagnosis (109). In 23 joints involved when the patients were children, no complications were observed later in life that could with certainty be referred to as osteochondritis dissecans. Those developing the disorder in the early adult years, however, had a worse prognosis and only 10 of 48 adult onset patients did not show signs of osteoarthrosis. Localization of the lesion was comparable to that of Aichroth, with 80% of the lesions located in the medial femoral condyle. Of the 23 joints in 18 patients in the childhood group, with the average age at diagnosis 12.7 years, there was no osteoarthrosis in 21 patients and only mild osteoarthrosis in 2. In this series, the most common time of presentation of the disorder was between 10 and 15 years. Linden concluded that the "prognosis of osteochondritis dissecans was vastly different for children--patients with open epiphyseal lines at the time of diagnosismcompared with adults" and that "complications usually do not develop in individuals diagnosed as having osteochondritis dissecans before closure of the epiphysis of the distal end of the femur, regardless of treatment." J. P. Green also concluded that "conservative treatment is indicated in adolescents and children" (62). Bradley and Dandy have reported excellent results with drilling of the osteochondritis dissecans lesion performed before skeletal maturity (21). They discussed 8 boys and 2 girls with an average age at operation of 12 years 11 months. Radiologic healing occurred within 12 months in 9 of the 11 knees. The mean age at onset of symptoms was 11 years 6 months. All lesions were on the medial femoral condyle at the classical site. They drilled a minimum of six holes through the overlying articular cartilage into the fragment and

478

CHAPTER 6 9 Epiphyseal Disorders o f the Knee

the underlying epiphyseal bone using a 1.5-mm Kirschner wire. Drilling through the fragment into the epiphysis was followed by restoration of a normal radiographic appearance in 10 of the 11 knees. They felt that the same procedure after skeletal maturity was less effective. This series allowed no time for nonoperative management. Similarly good results prior to epiphyseal closure were defined by Hughston et al. in 1984 (78). They found that "knees with osteochondritis dissecans of the femoral condyles that had no other abnormal physical findings or functional disability responded well to conservative treatment before epiphyseal closure. Those with objective evidence of looseness of a necrotic fragment or functional disability were best treated surgically." Aichroth and colleagues have continued their detailed studies of a large series of patients and reached results somewhat contradicting the reports mentioned previously, particularly in relation to those with childhood onset of the disorder (165). In 22 knees with osteochondritis dissecans diagnosed before skeletal maturity and followed into middle age, 32% had radiographic evidence of moderate or severe osteoarthritis at an average follow-up of 33.6 years. Only one-half had a good or excellent functional result. Poor resuits were found if the defect was large or affected the lateral femoral condyle. Their review indicated that, although 50% of patients showed some radiological signs of osteoarthritis, one-third of these changes were very minor and the rest had moderate lesions only. They concluded, however, that lesions in the classical position, which was felt to be nonweight-bearing, generally did well as long as any loose fragments were excised. In a separate study, Aichroth et al. assessed the degree of healing of stable lesions, which were treated conservatively with the expectation that they would heal. Stable fragments were left in situ following assessment by arthroscopy with stability assessed with the use of a hook. Unstable fragments were removed and the resulting crater was curetted. Thirteen of 21 stable lesions had healed, but virtually all of these except 1 were at the inferocentral lateral or medial condylar regions or anterior. Of the classical lesions, only 3 of 10 healed. Aichroth et al. concluded that stable osteochondritis dissecans in the classical position on the medial femoral condyle often did not heal spontaneously. The value of surgical drilling in such cases is an important consideration. The various management approaches have been well-reviewed by Clanton and DeLee (31), with particular concentration on the multiple surgical interventions in the adult.

K. Summary of Treatment Approaches in Childhood OD It is quite clear that cases of childhood osteochondritis dissecans, by which is meant those disorders that occur prior to physeal closure, have much better results both in terms of relatively early healing and the absence of degenerative knee

problems decades later than those patients who develop the disorder in adulthood following physeal closure. Many and actually most patients, therefore, in the childhood age group can be treated effectively without surgical intervention. Close assessment is essential, however, and the following approach appears justified based on the knowledge of the pathoanatomy, the healing process of the disease, and extensive reports from the literature. 1. OBSERVATION ONLY In those patients who are particularly young, by which is meant those from 5 to 7 or 8 years of age, and who are asymptomatic, observation alone is sufficient. There indeed may be a subset of patients in whom the differentiation between a normal developmental irregularity of the distal femoral epiphysis and a minimal osteochondritis dissecans cannot be made. 2. CONSERVATIVE TREATMENT The mainstay of conservative treatment is a combination of crutches and a long leg cast to immobilize the knee both for comfort and to enhance stability during the healing phase, allowing translesional bone repair to occur. Reports as early as that of Decker in 1938 accurately reported healing of OD lesions in childhood variants with cast immobilization only (40). The later works of Lofgren (113) and of Green and Banks (64) were particularly informative of the value of this conservative approach in childhood osteochondritis dissecans. In many centers, efforts are made to determine whether the OD lesion is stable or unstable, an approach that seems desirable. The conservative treatment is used when there has been no displacement of the OD segment from its site in the distal femur. Plain radiographs, however, and even plain tomography or CT scanning can still leave unclear the question as to whether the lesion is inherently stable. Three ways are used commonly to make this assessment. (1) An arthrogram can be performed, and if the dye remains external to the articular cartilage surface overlying the lesion, then the implication is that stability is present. If, on the other hand, the dye dissects between the bone of the lesion and the persisting bone of the femoral epiphysis, clearly there is both a tear of the articular cartilage surface and a free space around much or all of the OD fragment and it is deemed relatively unstable. (2) The issue of stability can be assessed more directly at arthroscopy during which time a blunt hook can be inserted through a separate portal and used to probe the softened OD region. The presence or absence of motion at the OD site leads to the designation of the type of stability. (3) MR imaging is the least invasive way to assess stability, although it does so indirectly primarily by identifying high signal intensity between the lesion and the surrounding epiphyseal bone, which implies an unstable OD lesion. Conservative treatment is warranted particularly when the lesion is stable and also is resorted to more frequently when the defect is relatively free of the weight beating surface.

SECTION III ~ Infantile Tibia Vara (Blount's Disease)

3. OPERATIVE CONSIDERATIONS Operative intervention can be related to stable lesions in situ or unstable lesions in situ and to lesions that are partially or completely displaced, at which time they are referred to as loose bodies. a. Stable Lesions in Situ In the childhood age group the widespread tendency is to treat stable lesions in situ with a conservative cast immobilization approach. Reports are present, however, on drilling of the osteochondritis dissecans site in childhood to hasten the repair process by allowing vascularization to pass from the normal bone through the region of surrounding fibrocartilage or fibrous tissue into the necrotic OD bone segment. There are two ways to perform the drilling. Under fluoroscopic control one can drill through the side wall of the epiphysis from the healthy epiphyseal bone into the bone of the OD lesion. The advantage of this approach is that it leaves the articular cartilage intact. The second approach is to drill directly through the articular cartilage into the necrotic segment and then across the zone of rarefaction into the healthy epiphyseal bone. Multiple punctures are made to enhance vascular ingrowth and repair. Although good results are reported, this method is not overly attractive because of the damage it can do to the surface articular cartilage. b. Unstable Lesions in Situ Drilling from normal epiphyseal bone into the bone of the OD defect can be performed cautiously in this lesion with care being taken to make certain that the drilling itself does not cause displacement of the lesion. Casting to enhance healing also is a good option because relief from weight beating alone may be sufficient to allow translesional vascularization and bone repair. The third option in this group is to enhance stability with a small compression screw. c. Partially Displaced OD Lesion The tendency here is to open the joint at arthrotomy and curette the base of the crater, removing fibro-osseous, fibrocartilaginous, or fibrotic tissue from the bed and allowing a bleeding surface to be exposed to the OD bone fragment. The OD fragment is trimmed of frayed tissue and then replaced in position and held with fine AO screws. These can be removed once healing has occurred. d. Loose Body Once a fragment has displaced completely from the OD site, surgery is warranted in virtually all instances. The loose body simply can be removed at arthrotomy and the crater examined. If it is off the weight beating surface, no additional management is needed. If it is in a weight bearing region, two considerations are possible. If the loose body was freed relatively recently and appears to fit the crater, then both the loose body and the crater itself can be cleaned of excess repair tissue and the free body can be repositioned and held with AO screws. If the loose body itself is in great disrepair or if it has increased in size greatly due to nutrition either from the synovium or from an attached synovial pedicle, then the site of the crater can be debrided of repair tissue and drilled multiple times to allow

479

for mesenchymal tissue ingrowth and differentiation to a form of hyaline cartilage. If the individual is close to skeletal maturity, consideration also can be given to chondrocyte implantation much as is being done in adult forms.

4. ASSESSMENTFOR STABLE AND UNSTABLELESIONS The importance of assessing (1) stability and instability, (2) the size of the lesion, and (3) the position of the lesion remains essential. Stability allows for conservative management, whereas instability often benefits from a more interventional approach. Size correlations have been shown to favor conservative management for smaller lesions and interventional management for larger lesions. In those areas that are non-weight-beating, there is less likelihood of developing subsequent osteoarthritis. It has been suggested, however, that "if there is a large segment of denuded bone in a weight beating position, the outlook in the long term is poor. Even a young child does not have the ability to repair such defects fully. Lesions in the classical position will generally do very well so long as any loose fragments are excised" (165). Crawfurd et al. made a particular study of whether stable lesions had healed (36). They felt that stable OD in the classical position on the medial femoral condyle often does not heal spontaneously, whereas elsewhere in the knee, for example, at the inferocentral lateral and inferocentral medial as well as anterior regions, healing occurred almost invariably. The range of options is well-summarized by Hughston et al. in a review of 95 knees in 83 patients followed from 2 to 31 years. Their series comprised both childhood and early adult patients. Their conclusions, based on their extensive study plus a review of the literature, indicated that when osteochondritis dissecans of the femoral condyles is diagnosed before epiphyseal closure, the physical findings are negative, and there is no functional disability, the lesion is best treated conservatively by continuing normal activities supplemented by quadriceps-strengthening exercises rather than by immobilization and rest. When the lesion is diagnosed either before or after epiphyseal closure and there is objective evidence of looseness of the fragment as well as functional disability the defect is best treated by open reduction and temporary internal fixation of the fragment until healing occurs. In knees with a long standing lesion in which the loose fragment has become smaller than the crater and in which fixation and healing of the fragment would not restore a congruous joint surface the lesion is best treated by excision of the fragment.

III. I N F A N T I L E T I B I A VARA

(BLOUNT'S DISEASE) A. Terminology Infantile tibia vara is a growth disturbance of the medial part of the proximal tibia involving the physis primarily and shortly thereafter the epiphyseal cartilage, the secondary

480

CHAPTER 6 ~ Epiphyseal Disorders of the Knee

ossification center, the slope of the articular cartilage, and the adjacent metaphysis. The disorder leads to a bowing of the lower extremity with the tibial deformation characterized by a varus angulation at the proximal metaphyseal level, often with internal tibial torsion of the leg and hyperextension of the knee.

B. Clinical and Radiographic Profile of Infantile Tibia Vara The disorder is slightly more common in females than in males and slightly over 50% of cases are bilateral. Obesity is an associated factor, particularly in the North American population. Dietz et al. noted obesity in 12 of 18 patients in one series (45). Blacks appear to be particularly prone to the disease. Infantile tibia vara (Blount's disease) refers to bowing of the proximal tibia due to abnormal structural changes in the proximal medial tibial growth plate. The disorder only rarely can be diagnosed before 2 years of age and is felt to be initiated by events superimposed on a physiologic genu varum that does not resolve. 1. BLOUNT Blount provided the first detailed coherent description of the disorder of the proximal medial tibial epiphysis, which he termed tibia vara (19). He described 13 of his own cases and reviewed 23 with similar characteristics from the literature, 15 of which were also considered to be examples of the specific disorder tibia vara. The disorder quickly became recognized as a discrete disorder (41, 153). The first welldocumented case is attributed to Erlacher (1922), who reported a deformity with involvement of the medial half of the proximal tibial epiphysis (52). F. Langenskiold (102) and Nilsonne (127) also described individual cases in 1929. Blount's classic description recognized that the disorder was not limited to the growth plate but was an abnormality of growth also involving the epiphyseal cartilage, the secondary ossification center of the epiphysis, and the adjacent metaphysis. He felt that both cartilage and bone were involved. His 13 patients included 10 with the infantile variant and 3 with the adolescent variant. In the infantile type, "in each case there was a history of normal development to the age of from 1-2 years, usually with some overweight. Then the exaggerated physiological bow leg, instead of developing gradually into the normal knocked knee, became more marked." Many of the patients had bilateral involvement. No likely etiological factors were currently in evidence then, but Blount noted that the X rays were so "uniform in appearance as to suggest a common cause." He readily recognized the abrupt varus angulation just below the proximal tibial epiphysis, the medially expanded and sometimes irregular epiphyseal line, the medial wedge-shaped epiphyseal secondary ossification center, and the prominent beaklike recurving medial metaphysis. There were cartilage islands within the metaphyseal beak, and hyaline cartilage frequently covered the me-

dial metaphyseal bony prominence, which allowed for growth and continuing enlargement medially. B lount also recognized a separate adolescent group with onset between 6 and 12 years of age. The disorder became apparent with increasing bowing without apparent cause and without the symptoms of rickets. The deformity often was bilateral, particularly in infantile cases, and frequently was associated with subsequent spontaneous disappearance of the bowleg on one side. In the adolescent group, the angulation usually was unilateral. Limp generally was seen in unilateral cases with a more prominent waddling in bilateral cases. Blount noted the abrupt angulation of the metaphysis with the apex laterally just below the knee joint. In addition to genu varum, other deformities involve a recurvatum (hyperextension) at the knee, internal rotation of the tibia on the femur, shortening of the leg, and abnormal mobility of the knee on mediallateral stresses. In his description of the radiographic findings, he referred to sloping medial epiphyses with beaklike metaphyses, irregularity of the medial contours of the proximal tibial epiphyseal line with abrupt angulation at the metaphysis with the apex laterally, lacy enlargement of the metaphysis medially with a beaklike angulation below the proximal tibial epiphysis, marked bowing of the proximal tibia with a thickened medial cortex, a narrow and dense epiphyseal plate with premature closure medially, and deficient secondary ossification centers with sloping and hypodevelopment of the medial epiphyseal region. Blount's initial description, highlighted by X rays, case studies, and tracings from his own radiographs, defined the entire spectrum of tibia vara. The 49 radiograph tracings provide a good demonstration of the underlying pathogenesis. In his early summation of the pathology, he felt that in the infantile form the changes consisted essentially of faulty growth of the epiphyseal cartilage and delayed ossification of the medial portion of the proximal tibial epiphysis. A beaklike projection of the metaphyseal bone formed secondarily as a buttress under the epiphysis. Blount recognized that episodes of spontaneous correction could occur. He felt that braces or other types of support did not increase the incidence of spontaneous repair, although they did provide symptomatic mechanical relief. His earliest thoughts on therapy were primarily noninterventional. He felt that "conservative measures should be continued during a period of observation of several years. When the deformity remained stationary, osteotomy should be performedmin adolescence for cosmetic reasons and in adults when function is greatly disturbed." Blount proposed waiting before proceeding to operative treatment, which in the adolescent phase would occur only when the epiphysis was closed. He felt that the infantile type seemed to remain stationary after 3-4 years. Lasting correction was possible following early osteotomy, but deformity often recurred when the osteotomy was performed later. Recurrence after osteotomy was recognized in several cases. The abnormal mobility of the knee on medial-

SECTION III ~ Infantile Tibia Vara (Blount's Disease)

lateral strain had to be factored into the timing and type of surgery because that abnormal mobility often contributed to as much as half of the varus angulation. Blount thus stressed the need to slightly overcorrect the deformity rather than leaving it undercorrected when surgery was done.

C. Clinical-Radiographic Grading Scheme of Langenskiold, Types I-VI The clinical and radiographic changes in tibia vara have been well-classified by Langenskiold (97) and Langenskiold and Riska (99), who defined types I-VI. There is some interobserver variability (160), but this remains the most useful classification for assessment and clinical management of the disorder. It was initially based on 23 cases. The categories merge into one another, leading to the interobserver variability, but have proven to be an accurate clinical-radiographic categorization. A common feature of the untreated cases of tibia vara, which Langenskiold noted independent of the age at which they were seen, was the abrupt medial angulation at the upper end of the tibia. We have modified the illustration of types I-VI to include the shape of the articular surface, which plays a major role in determining long-term function (Fig. 8A). Stage I: The initial changes are noted at 2-3 years of age. Irregular metaphyseal ossification is seen immediately adjacent to the entire physis, with the most proximal and medial part of the metaphysis characterized by a distal and medial pointing spur or beak of bone rather than being curvilinear as in the normal. The medial part of the secondary ossification center is not as well-developed as the lateral. Stage II: The proximal medial portion of the metaphysis is depressed further in a distal and medial direction, and the smooth curvilinear proximal metaphyseal-physeal curve shows a shallow depression immediately proximal to the bony spur, which is not as dense as the adjacent bone due to the presence of physeal cartilage. The medial aspect of the secondary ossification center has less height than the normal at its most medial aspect, and the superior part of the secondary center medially has a downward obliquity. These changes tend to be concentrated between 2.5 and 4 years. Stage III: The bony metaphyseal spur medially is more extensive, the medial stepoff in metaphyseal bone is deeper due to the presence of physeal cartilage, and the most medial part of the bone of the secondary ossification center has a slightly stippled appearance. This stage tends to occur between ages 4 and 6 years. Stage IV: The metaphyseal spur and metaphyseal depression are more extensive, and the most medial bone of the secondary ossification center can be seen to pass distal to the physeal line toward the metaphyseal depression. Often, this bone occurs as islands of ossification separate from the main secondary center of ossification. This stage tends to be seen from ages 5 to 10 years. Stage V: The metaphyseal spur and secondary ossification center depression are greater yet, and the most medial part of the secondary ossification center is

481

now depressed into the metaphyseal region well below the physis and is separated from the bulk of the secondary center by a horizontal cleft. The obliquity of the superior part of the medial secondary ossification center is marked and the adjacent articular surface is deformed and depressed. This stage tends to be seen from 9 to 11 years. Stage VI: A welldemarcated transphyseal bone bridge now joins the medial part of the secondary ossification center with the adjacent metaphysis. Union has occurred across the depressed oblique medial part of the proximal tibial physis, and the horizontal radiolucency separating the normal and medially depressed parts of the secondary center of ossification also is obliterated by bone. There is marked obliquity of the superior aspect of the medial secondary ossification center. Additional features characterize the tibia vara condition and are marked particularly from stage III onward: the femoral-tibial diaphyseal angle increases in a varus direction; the obliquity of the proximal medial tibial surface leads to a progressive lateral subluxation of the tibia in relation to the femur; the femur responds to the tibial abnormality in many instances by a relative overgrowth of its medial condyle and in severe instances by developing a lateral curvature of the distal metaphyseal region, and the knee deformity is worsened by a tendency to hyperextension and internal torsion of the tibia in relation to the femur. Radiographs showing the various stages of tibia vara are seen in Figs. 8Bi-8Bvi.

D. Pathogenesis of Varus Deformity Due to increased stress on the proximal medial tibial physeal cartilage, physeal damage occurs and a cycle is established in which decreased medial growth combined with normal proximal lateral tibial and fibular growth leads to further varus, which further increases the compressive forces on the medial physis. The growth retarding stresses initially affect the radiolucent physeal cartilage and are not visible on plain radiographs. The initial plain radiographic indications of a developing pathological tibia vara by current determinations therefore are secondary changes in the adjacent bone of the metaphysis and secondary ossification center. By the time such changes are seen, some damage to the physeal cartilage already has occurred. The basic underlying cause of the disorder remains unclear, although it appears to be based on altered mechanical forces during the vulnerable early growth period. The high incidence of tibia vara in the Caribbean has been related to earlier times of walking (61), although a study in South African blacks noted no difference in time of walking between those affected and the normal population (13). The disorder is localized to the proximal tibia so no systemic chondrodystrophic process has been identified. In North America there is a high correlation between obesity and tibia vara, but in other populations this is not always the case. At present, it appears that Blount's disease represents a physiologic genu

,a,

F I G U R E 8 (A) This illustration outlines the Langenskiold classification of tibia vara from type I, which is the mildest variant, to type VI. We have included the shape of the proximal tibial articular surface and epiphyseal cartilage. These cannot be appreciated

SECTION III ~ Infantile Tibia Vara (Blount's Disease)

483

F I G U R E 8 (continued) on plain radiographs but play a major role in determining the eventual clinical result. Cartilage shape is currently best illustrated by either arthrogram or magnetic resonance imaging. (B) The radiographic appearance of stages I-VI tibia vara is illustrated. The right proximal tibia is shown in each instance. (Bia) A type I lesion is barely able to be differentiated from normal. It is difficult to make the diagnosis of a type I tibia vara by clinical and radiographic criteria at less than 2 years of age. The bowing is accompanied by a medial metaphyseal beak or spur of bone and an irregular surface of the medial metaphyseal bone adjacent to the physis (arrow). In (Bib), the metaphyseal beaking medially, underdevelopment of the medial tibial secondary ossification center, and internal tibial torsion (represented by tibial-fibular overlap) are greater than normal. (Bii) In stage II the proximal medial portion of the metaphysis is more depressed in association with greater varus deformation. There also is a shallow depression in the medial proximal physeal-metaphyseal bone margin. The medial part of the proximal tibial secondary ossification center is developing a wedge shape and is less uniformly developed or oval than the lateral segment. (Biii) In stage III, changes from stage II are somewhat more exaggerated. There is further irregularity of the medial tibial epiphysis with obliquity of the superior medial margin of the secondary center and a tendency now to early irregular fragmentation of the most medial portion of it. There is further depression and scalloping of the medial physeal-metaphyseal bony junction. (Biv) In stage IV, the medial bone of the secondary ossification center now passes into the depression below the line of the more lateral aspect of the physis as bone begins to form in the medial epiphyseal component within the metaphyseal depression (arrow). (This radiograph was made after healing of a valgus osteotomy.) (Bva, Bvb) In stage V, the metaphyseal spur and secondary ossification center depression into the metaphyseal region are greater and the most medial part of the secondary ossification center now is well beyond the physis into the metaphyseal region. The obliquity of the superior part of the secondary ossification center is greater, as would be the articular surface if visualized by either MR imaging or an arthrogram. Both left and fight images show some persisting radiolucency between the medial metaphyseal bone and the depressed spur of epiphyseal bone. (Bvi) In stage VI, a transphyseal bone bridge has now formed medially, linking the epiphyseal and metaphyseal bone masses and totally obliterating the medial physis. Note the extreme obliquity of the medial tibial epiphysis and the lateral tibial subluxation.

v a r u m that does not s p o n t a n e o u s l y correct by 2 years of age and w h o s e w o r s e n i n g into a p a t h o l o g i c state is increased by obesity. Studies in J a m a i c a n children by B a t e s o n also suggest that severe b o w l e g s d e v e l o p f r o m p h y s i o l o g i c b o w l e g s that do not s p o n t a n e o u s l y correct (12). T h e negative effects of e x c e s s i v e pressure on e p i p h y s e a l g r o w t h h a v e b e e n d e m onstrated. O n c e w e i g h t bearing continues on a k n e e in varus position, excess m e d i a l p h y s e a l stress is accentuated. A finite

at varus of 30 ~ with changes m o r e m a r k e d in the o b e s e child (34). It was calculated that, in a 2-year-old child, 20 ~ of varus led to forces sufficient to retard physeal growth, and as little as 10 degrees varus in an o b e s e 5-year-old child led to forces e x c e e d i n g those n e c e s s a r y to retard growth. Histologic studies also d e m o n s t r a t e that increased stresses d a m a g e the medial physis on a m e c h a n i c a l basis.

e l e m e n t analysis of the p r o x i m a l tibia has indicated c o m pressive stress in the medial physis to be s e v e n times n o r m a l

dial part of the p r o x i m a l tibial physis. This finding, described by G o l d i n g and M c N e i l - S m i t h (61), was investigated further

In many, the deformity is concentrated at the p o s t e r o m e -

484

CHAPTER 6 9 fpiphyseal Disorders of the Knee

at arthrotomy by Siffert and Katz (151). In each of five knees undergoing osteotomy, intra-articular exam revealed a prominent depression of the posterior half of the medial tibial articular surface. With slight flexion of the knee, the medial femoral condyle slipped into the defect. The medial collateral ligament then became lax allowing for the demonstration of increased medial-lateral instability.

E. Pathoanatomy Tissue samples have been studied to help assess the pathologic appearance of physeal and periphyseal tissue in infantile tibia vara. 1. BLOUNT Blount used histologic sections from a biopsy of the medial tibial metaphyseal beak in an effort to understand the disorder (19). There was a marked distortion of trabeculae of bone at the point of junction with the cartilage. The distortion of the epiphyseal cartilage with the bending of the bone trabeculae had pushed the periosteum outward, and it had arched over the bone and cartilage. There were trabeculae of bone on the outer surface of the cartilage, which appeared to be independent of the bone trabeculae at the epiphyseal line. The cartilage had become irregular in appearance in places and granulation tissue had invaded some of these areas. The photomicrographs show cartilage hypertrophy and lobulation, which is quite abnormal. In certain areas, however, the parallel alignment of the physeal cartilage cells persisted and served to synthesize metaphyseal bone by the endochondral mechanism. In each case, several common denominators were seen radiographically, including the abrupt angulation just below the proximal medial tibial epiphysis, an expanded sometimes irregular epiphyseal line, the wedge-shaped epiphysis, and the prominent beaklike recurving medial metaphysis. Histologic sections identified cartilage islands within the beak. Blount reviewed the various possible known causes of bowing and indicated that none appeared to be present in his patients. He concluded from his 13 cases and the 15 from the literature that "there is no evident cause." He considered the epiphyseal region to be the primary site of deformity in the infantile type. In reviewing the pathology, he concluded that the changes "consist essentially in faulty growth of the epiphyseal cartilage and delayed ossification of the medial portion of the proximal tibial epiphysis. A beak-like projection of the metaphysis forms secondarily as a strut under the epiphysis." The areas of rarefaction seen on the X ray in the metaphyseal beak were felt to correspond to islands of cartilage seen in the histologic sections. 2. LANGENSKIOLD Langenskiold assessed the medial beak region of the proximal metaphysis and the adjacent physis of the tibia with narrow core biopsies in 10 patients (97, 98). All'histological

sections were performed in the frontal plane to correspond with the anteroposterior radiograph. In the lower Langenskiold grades, cartilage columns and hypertrophic cells can be recognized although these tend to be at right angles to the axis of the diaphysis. With increasing deformity, the cartilage sections showed progressively poorly defined columnar regions with interspersed fibrovascular areas. Areas of nearly acellular fibrous cartilage often were seen. The perichondrial ossification groove of Ranvier was quite irregular and generally not recognizable. The bone itself showed no distinct pathological changes. In the most advanced cases, only fibrocartilage was seen in the previous physeal area. There were no organized cartilage cell columns and signs of cellular activity, which would have provided growth in length of the bone, were practically absent from the region in question. In other cases of severe involvement no traces of the true epiphyseal plate could be seen in the histologic sections. Langenskiold summarized the histological findings in his cases of the infantile-type tibia vara, which ranged in age from 2.5 to 11 years. "The most characteristic histological findings in the pathological area in the medial part of the upper end of the tibia are the changes seen in the zone of the resting cartilage: 1) islands of densely packed cells showing a greater degree of hypertrophy than that corresponding to their topographical position; 2) islands of almost acellular fibrous cartilage; 3) abnormally large groups of capillary vessels." Pathological areas of tissue often were scattered within resting cartilage of normal appearance. Changes were seen in all stages of the disease. In the most severe older cases, signs of proliferation of the cartilage in the medial plate were "practically absent." The cartilage layer covering the medial aspect of the epiphyses joining the articular cartilage to the epiphyseal plate was abnormally thick but appeared as typical resting cartilage. The borderline of this cartilage layer with the bone tissue was irregular. Langenskiold felt that the histological findings suggested disturbances of growth, maturation, and ossification of the epiphyseal cartilage, with changes similar to those seen after the influence of abnormal pressure. Langenskiold detailed the histologic findings, reflecting the worsening of the disorder beginning with slightly abnormal cartilage in the youngest patient and ending with medial fibrous cartilage and bone bridge formation in the oldest. In case 1 at 2.25 years of age, "the sections consisted principally of normal resting cartilage with its overlying perichondrium. In an area of about 2x3 mm surrounded by normal cartilage, some 50 cross-sections of capillary vessels were seen. Between them, the cartilage cells were hypertrophied. This area showed a greater cellularity than the surrounding cartilage and scattered portions of fibrillated matrix." The presence of capillary vessels could be interpreted as an abnormal finding, which ultimately would predispose one to either fibrosis or new bone formation or both. The fact that the cartilage cells in relation to the capillary vessels were hypertrophied also is consistent with premature and ectopic

SECTION III ~ Infantile Tibia Vara (Blount's Disease) new bone formation as an alteration of the normal endochondral sequence. In case 2 at 3.25 years, the proximal metaphyseal beak region showed "an area of fibrous cartilage." Although the area was far from the epiphyseal plate, "there were cartilage cell columns arranged at right angles to the axis of the diaphysis." This cartilage obviously was a vestige of the epiphyseal cartilage. In the zone of the resting epiphyseal cartilage, slight irregularities in the distribution of the cells were seen. The cartilage cell columns were of normal appearance. This would appear to refer to the abnormal position of the posteromedial portion of the physis and also to the fact that growth is occurring even though it is at right angles to the long axis. The fact that growth continues also is shown by the fact that the zone of cartilage cell columns and the metaphyseal bone are of normal appearance. In case 4 at age 2.5 years, tissue from the apex of the beak was obtained with both physeal cartilage and adjacent metaphyseal bone. The ossification zone and the zone of cartilage cell columns showed a "fairly normal condition, with the exception of some columns extending several millimeters into the metaphyseal bone." A lower power photomicrograph shows overlying cartilage but clear fibrovascular invasion adjacent to it. At the periphery, vessels were penetrating both the epiphyseal cartilage and the metaphysis particularly at a region of cartilage cell columns in an area of nearly acellular fibrous cartilage. Both the acellularity and the vascular proliferation are signs of being associated with the degeneration of cartilage and its replacement by nongrowth tissue. In case 8 at age 3.5 years, tissue at the extreme periphery of the physeal cartilage is hypocellular and shows no evidence of an endochondral sequence, even though it is adjacent to both endochondral bone and intramembranous bone. In case 9 at 4 years of age, the peripheral ossification groove showed a normal amount of cells, but in some sections irregularities and increased capillary invasion were seen. The zone of cartilage cell columns of the physis showed some irregularity, but the number of cells in the columns seemed normal. The resting cartilage above showed "islands of highly cellular tissue and others of nearly acellular fibrous cartilage. These islands were scattered in areas of normal resting cartilage in which groups of some dozens of cross sectioned vessels were seen. The ossification zone showed irregularities which were closely related to those seen in the adjacent zone of cartilage cell columns. Otherwise the bone tissue showed no distinct pathological changes." These findings also are indicative of normally functioning tissue in an aberrant position, which gradually is beginning to lose its appropriate specialization. Even at the periphery of the medial physis, at its most medial and inferior portion, cartilage tissue persists and some columnar organization of cells is seen along with the beginnings of an endochondral sequence. In case 14 at 9.5 years, which would appear to represent a type IV deformity, a fairly extensive amount of tissue was obtained medially again from the inferior portion of the beak adjacent to the oblique primary physis. This included the

485

peripheral part of the distal or primary branch of the double medial part of the epiphyseal plate, a part of the triangular bone fragment lying between the branches of the double part of the plate, and a portion of the resting cartilage covering the medial aspect of the triangular fragment. The peripheral part of the physis almost exclusively is fibrous with no real cartilage columns seen. "Signs of cellular activity providing growth in length of the bone were practically absent in the region in question" and much bone was seen. In case 15 at 10.25 years, a type V categorization, sections included the entire metaphyseal beak, the layer of resting cartilage covering the medial aspect of the proximal end of the bone and a part of the diaphysis merging into the beak. No traces of the true epiphyseal plate could be seen in the sections.., the cartilage consisted partly of normal hyaline cartilage of the same type as articular cartilage, but in it were scattered numerous islands of more cellular cartilage. True cartilage cell columns were not seen. The bone tissue included in the sections showed some fibrosis of the marrow but no other distinct pathological changes. Although cartilage persisted, there was no evidence of growth whatsoever. It also should be noted that this section is from the inferior beak region and also at the most peripheral part of the physis. In the final case 17 at age 11 years, a type V bordering on type VI example, sections from the most medial, inferior, and peripheral part of the physis showed that the "epiphyseal plate consisted of fibrous cartilage showing no signs of proliferation."

3. SLOANE,SLOANE,AND GOLD; JAFFE An excellent histologic section from the medial epiphysealmetaphyseal beak region from a 7-year-old girl with bowing due to tibia vara was presented (154). The article appeared 1 year before Blount's description, but the photographs, radiographs, and clinical descriptions are all diagnostic of tibia vara, although the article was titled "Dyschondroplastic Bow Legs." An excellent photomicrograph of the biopsy is shown along with a detailed pathologic report by Jaffe. The line of the proximal medial tibial epiphyseal plate sloped sharply downward in the direction of the shaft. The epiphyseal cartilage in this region was rather pale staining particularly the nuclei, and the cells appeared to be either dead or dying. Nowhere did the cartilage even remotely approach the normal epiphyseal architecture. There had been an extremely feeble attempt in group line-up (columnation) but there was not even a suggestion of a well-defined zone of clear cells (hypertrophic cells) that characterizes the normal epiphyseal plate, although a few of these clear cells were seen. The reciprocal alignment of resolving cartilage and growing bone trabeculae also was absent. In its place was a very thin irregularly demarcated zone of poorly calcified cartilage, which penetrated in a most uneven fashion with disorderly arrangement of shapeless bone trabeculae. Many, if not most, of the cartilage cells were devoid of nuclei. Marrow spaces showed

486

CHAPTER 6 9 Epiphyseai Disorders of the Knee

considerable fibrosis. Orderly deposition of osteoblasts was not to be seen. The disorder was characterized by a failure of growth and differentiation of cartilage cells, with even the "resting" cartilage appearing to be afflicted. Examination of the photomicrograph indicates that the metaphyseal bone medially from the beak is of a reactive intramembranous bone formation rather than emanating from a physeal endochondral sequence as would normally be the case. 4. LAMY AND WEISSMAN Lamy and Weissman reported on two cases of tibia vara and reviewed the pathological anatomy from other studies (96). Their principal observation was the degeneration of physeal cartilage characterized by the development of hypertrophic cells within the cartilage and the transformation of the hyaline cartilage with its replacement by granulation tissue. Numerous vessels and small centers of ossification came to be found within the cartilage matrix medially. These findings were particularly marked on the metaphyseal side of the physis, whereas the epiphyseal or upper regions of the cartilage were flattened. Within the substance of the physeal cartilage were numerous clefts or cracks, which were more or less regular and filled with a homogeneous noncellular substance. There also were aberrant islands of cartilage among the bone trabeculae far from the physis itself. The bony trabeculae of the metaphysis were thin and had lost their parallel orientation, being disorderly and oblique.

5. GOLDING AND MCNEIL-SMITH Biopsy specimens from six cases of tibia vara showed no evidence of avascular necrosis or infection (61). The medial physeal cartilage adjacent to the metaphyseal beak was "grossly disorganized in pattern with irregular cartilage columns, the cells in some areas being hypertrophic, in others almost non-cellular fibrocartilage." F. Imaging Assessments in Relation to Tibia Vara Patients with tibia vara are assessed routinely with plain radiographs of the knee in anteroposterior and lateral projections. The plain radiologic appearances of physiologic bowlegs (Fig. 2A) and early pathological (simple severe) bowlegs were shown by Bateson to have several features in common, including (1) bowing of the femur and tibia at the knee (diaphyseal angles in varus position), (2) thicker medial than lateral tibial cortices, (3) internal tibial torsion, and (4) distal femoral and proximal tibial medial metaphyseal beaking (12). Additional information is provided by the following techniques and measurements.

1. KNEE JOINT ARTHRoGRAPHY Arthrography outlines well the cartilaginous surfaces of the proximal tibial and distal femoral condyles (39, 53). It is particularly useful in those up to 5 years of age in which there is relatively less development of the secondary ossifi-

FIGURE 9 Arthrogramoutlining the contours of the proximal tibial articular surface in a stage III disorder. A slight downward obliquity to the proximal medial tibial articular surface is beginning to form, but the structure of the cartilage is far closer to normal than that of the adjacent secondary ossificationcenter and metaphysis. cation center than in the older age groups, particularly in those with a moderate to severe tibia vara in which development of the secondary ossification center is delayed in its medial part (Fig. 9). Evensen and Steffensen demonstrated considerable sloping of the proximal medial tibial articular surface in a 7-year-old girl with tibia vara in an early arthrography report (53). 2. MAGNETIC RESONANCE IMAGING Magnetic resonance imaging now plays a role in assessing the structure and to a certain extent function of the proximal medial physis. It also outlines the shape of the epiphyseal and articular cartilages. Plain radiographic assessments, which only demonstrate bone tissue, indicate the growth sequelae of the physeal damage, whereas the MR image has the ability to assess the physeal, epiphyseal, and articular cartilages. Varying signal intensities along the physeal cartilage as well as transphyseal vascular communication are signs of concern about developing deformity. 3. FEMORAL-TIBIAL DIAPHYSEAL ANGLE It remains difficult to quantify the amount of varus deformity. The femoral-tibial diaphyseal angle is an index commonly used in attempts to define the lower extremity varus

SECTION III 9 Infantile Tibia Vara (Blount's Disease)

487'

directly. The diaphyseal measurement is imprecise, however, in young patients with tibia vara as frequently there is associated bowing of the distal femoral and proximal and distal tibial metaphyses and adjacent diaphyses, which makes the long axes of the bones difficult if not impossible to determine. Standing anteroposterior films of both femurs and tibias are taken. Many with tibia vara have knee joint ligamentous laxity, which worsens the varus deformity and must be considered in documenting angular deformity and in planning for and carrying out surgical correction. Measurement of this angle is outlined in Fig. 2.

4. METAPHYSEAL-DIAPHYSEAL ANGLE OF THE PROXIMAL TIBIA Due to the difficulties inherent in measuring the femoraltibial diaphyseal angle, Levine and Drennan proposed the tibial metaphyseal-diaphyseal angle to focus on the deformity of the proximal tibia itself (106). This angle can help to distinguish between physiological bowing and a true developing tibia vara in most instances. One line is drawn perpendicular to the longitudinal axis of the tibia. The second line is drawn joining the beaks of the metaphysis on the medial and lateral sides at their most distal points (Fig. 10). The metaphyseal-diaphyseal angle is the angle bisected by these two lines. In 29 of 30 extremities with an initial metaphyseal-diaphyseal angle of more than 11 ~ radiographic changes of tibia vara later developed. Only 3 of 58 extremities with an angle of 11 ~ or less later had a tibia vara diagnosis eventuate. It is important to recognize that, although this angle measures secondary bony changes in the infantile tibia vara syndrome, it can provide an early indication of developing change.

5. MEASUREMENTOF THE MEDIAL PHYSEAL SLOPE Another radiographic indicator of varus deformity is the medial physeal slope (91). This is an angle formed by the intersection of a line parallel to the lateral half of the tibial physis and a line parallel to the medial deformed physis (Fig. 11). The latter measurement is along the most proximal border of the medial sloping proximal tibial metaphysis because the physis itself is radiolucent. A medial physeal slope greater than 60 ~ always was associated with recurrent varus deformity.

G. General Management Considerations We cannot yet define exactly when a limb with physiologic genu varum demonstrates a failure to correct the bowing such that the pathological condition infantile tibia vara is present. There are, however, some reasonable clinical guidelines. The physiologic varus should be maximal between 12 and 18 months of age after which it starts to diminish, reaching a mean value of 0 ~ at 24 months of age. An example of spontaneous correction is illustrated by a series of radiographs in Figs. 12A-12E. In our study, we were unable to define plain radiographic criteria for an infantile tibia vara

O

FIGURE 10 Measurementof the metaphyseal-diaphysealangle at the proximal tibia is shown. Care must be taken to determine accurately the most prominent beaking points of the proximal tibial medial and lateral metaphyseal bone. [Reprintedfrom (106), withpermission.]

before 2 years of age (56). The important time frame during which one is concerned about an evolving tibia vara is between 2 and 3 years of age. Any persistence of varus after 2 years of age, especially if it is of the same extent or becoming worse, leads to a concern about infantile tibia vara. After 2 years of age treatment considerations become important. If varus persists after 24 months, a presumptive diagnosis of tibia vara is warranted, even in the absence of plain radiographic changes. Concern about an evolving tibia vara is heightened by increased varus-valgus instability, increased internal tibial torsion, and an overweight child. Close assessment every 3 - 4 months is recommended. Above-knee orthoses or bivalved long leg night casts are used by some to hasten resolution without operative intervention. Their use and the responses of involved limbs to their use must be followed closely as there is little documented evidence of their effectiveness. Oyemade, however, has pointed out that serial plaster cylinder casts combined with wedging can lead

488

CHAPTER 6 ~ Epiphyseai Disorders of the Knee

6

gically should be made only after the patient is 2 years of age and is based on worsening or failure of correction of the varus in spite of conservative measures and worsening of the radiologic appearance of the proximal tibia to Langenskiold grade II or greater. Indeed, it is almost impossible to make a definitive diagnosis of tibia vara before 2 years of age because that represents the time period when transition from the physiological to pathological state is occurring. The presence of massive obesity should further hasten the decision to perform osteotomy. More recent summaries of treatment are in general agreement with the approaches just discussed (22, 27, 42, 56, 65, 68, 86, 148, 156).

H. Recurrent Deformity Following Osteotomy When the recurrences are studied in relation to weight, Langenskiold grade, and age at initial surgery, prognostic features can be identified (56). 1. RECURRENCES IN RELATION TO WEIGHT

F I G U R E 11 Measurement of the medial physeal slope as outlined by Kling et al. (91) is illustrated.

to good correction of some angular proximal tibial deformities of differing causation (133). Loder and Johnston felt brace treatment for patients in the earliest phases, stages I and II, was effective in only one-half of cases. Once a true infantile tibia vara is diagnosed, valgus osteotomy of the proximal tibia and associated fibular osteotomy are the treatment of choice. Valgus osteotomy realigns the mechanical weight bearing axis at the knee and decreases the pressure on the medial tibial physis. If physeal damage has been mild, it appears to be reversed and recurrences are eliminated. If the damage has been sufficiently extensive, full reversal does not occur, even with appropriate valgus repositioning, and recurrences are seen. As our concept of the pathogenesis of infantile tibia vara is a physiologic genu varum that does not resolve, there is no clear-cut radiologic definition as to when a limb is moving from a physiological to a pathological state. The danger with waiting until the pathological state clearly is evident radiographically, such as in grade III or greater lesions, is that physeal damage has occurred to such an extent that recurrence following osteotomy is highly likely and possibly inevitable. No recurrences occurred in our study following osteotomy with grade I lesions. Even in grade II lesions we noted four recurrences in eight procedures. The danger with operating on those with grade I lesions, on the other hand, is the fact that some of the patients might have shown spontaneous resolution of the deformity, especially as radiologic identification of a grade I lesion is somewhat subjective. The decision to intervene sur-

Massive obesity clearly worsens prognosis. Each of the massively obese patients in our series had bilateral involvement, and deformity recurred in 12 of 14 limbs undergoing osteotomy. This phenomenon of massive obesity differentiates our patient population from the European [Finland (12,14) and Sweden (24)], Black African [South Africa (3)], and Caribbean [Jamaica (9) and Martinique (6)] groups. Obesity was not a major contributing factor in these reports, and Bathfield and Beighton (33) specifically noted no increased obesity in 110 Black South African patients. 2. RECURRENCES IN RELATION TO LANGENSKIOLD GRADE

None of 9 Langenskiold grade I lesions required repeat osteotomy, whereas all 9 limbs with grade V and VI lesions at initial surgery required repetition. The effectiveness of osteotomy also progressively diminished from grade II to grade III to grade IV. Even in grade II, considered by most to be an early stage, 4 limbs required no repetition of surgery, but 4 did. When grades IV, V, and VI were considered, only 1 tibia remained straight whereas 13 required repeat osteotomy. Thus, it is evident that detailed presurgical radiographic assessment by tomography, CT scanning, or MR imaging is warranted at least in grades I V - V I to determine the type of additional intervention, other than metaphyseal osteotomy. 3. RECURRENCES IN RELATION TO AGE A clear differentiation based on age also was noted, with a recurrence rate of 76% in limbs operated on initially after 4.5 years, whereas those operated on before that time had a recurrence rate of only 31%. Although both increasing grade and increasing age correlate with increased recurrences, the causal factor in the recurrence is directly related to the radiologic grade, which is closest to reflecting actual physeal damage.

F I G U R E 12 A series of standing anteroposterior lower extremity radiographs documents spontaneous resolution of a physiologic genu varum. No treatment was given. The varus gradually diminishes with time, as does the medial tibial cortical thickening, the internal tibial torsion, the proximal medial metaphyseal beaking, and the relative underdevelopment of the proximal medial tibial secondary ossification center. Ages: (A) 16 months, (B) 22 months, (C) 2.5 years, (D) 3 years, and (E) 10 years.

490

CHAPTER 6 9 Epiphyseal Disorders o f the Knee

In a review of 158 osteotomies for tibia vara from the literature and 37 from his own study, Zayer found a similar pattern of increasing recurrence with increasing age at initial osteotomy (178). Of 96 cases from the world literature operated on from 2 to 8 years of age, there was a recurrence rate of 21%, whereas in 51 operated on from 8 to 14 years, the recurrence rate was 76%. In his own group, those operated on between 2 and 7 years had a 29% recurrence rate and those from 7 to 15 years had a 47% rate. Studies summarized from the first two decades after recognition of the disorder led to the belief that surgery when the patient was less than 8 years of age generally led to a favorable result with recurrence not a problem. Other and more recent patient reviews, however, indicate that much earlier intervention, before 4 - 5 years of age, is needed to minimize long-term sequelae. Medbo concluded in 1964 that more favorable resuits were likely when surgical correction of the varus position was done at an early age (3-4 years) (118), an observation with which we agree. The need for 1 or 2 additional osteotomies was greater the later the age at initial osteotomy beyond 4 years. Schoenecker et al. noted that only 1 patient in 12 operated on initially before 5 years of age needed additional surgery, whereas 14 of 21 (67%) operated on initially after 5 years of age required additional procedures (148). Roy and Chaise reported definitive correction with surgery in 5 patients at an average age of 3 years 4 months or 40 months (range = 22-52 months) whereas recurrence was seen in each of 3 patients operated initially at 6 years 9 months or 81 months (range = 80-84 months) (142). Loder and Johnston noted that poor results increased in proportion to higher stages at presentation (111). Single osteotomy before 4 years of age gave good results in 85% of tibias. Early tibial osteotomy appears to be the treatment of choice with the child less than 4 years of age for predictably good results.

I. Spontaneous Correction On occasion infantile tibia vara corrects spontaneously. In our series we noted two patients with bilateral involvement in whom one limb improved on its own whereas the other limb worsened. Positive responses to bracing have been documented, but detailed definitive studies on nonoperative treatment have yet to appear. Thirteen patients in our series who subsequently had osteotomy had some form of bracing or night casting, but the regimens and compliance were so variable that conclusions about the value of nonoperative approaches cannot be made.

J. Surgical Approaches to Tibia Vara 1. PROXIMAL TIBIAL-FIBuLAR OSTEOTOMY The mainstay of surgical treatment for tibia vara is a proximal tibial-fibular valgus osteotomy. The procedure serves to correct the varus deformity and to decrease the stress on

the proximal medial tibial epiphysis so that uniform growth can occur. The proximal tibial deformity is such that the valgus correction often must be accompanied by external tibial-fibular rotation (to counteract internal tibial torsion) and flexion (to counteract slight hyperextension). The ideal time for surgery is prior to the time that irreversible medial physeal damage has occurred or recurrence of the deformity might follow. At present we cannot determine the potential for physeal function in grades I-IV tibia vara, but studies of results postsurgery indicate a progressively higher rate of recurrence at each higher stage and each greater year of age at initial surgery. Even some stage II deformities seemingly well-corrected are followed by recurrence. The stage I deformity, however, is difficult to differentiate by clinical and plain radiographic criteria from a physiologic genu varum, which will correct spontaneously. Most observers, ourselves included, rarely will diagnose a tibia vara disorder under 2 years of age. If varus deformity persists and worsens after 2 years of age, if radiographic changes worsen, and if the patient is obese, the likelihood of the value of early intervention increases. The concept of waiting until a defined age, such as 8 years, or until late in childhood so as to need only one operation rarely is accepted now because the epiphyseal, physeal, and articular cartilage deformities define joint relationships more than eventual diaphyseal alignment. The latter can always be corrected by osteotomy, but the cartilage modeling defects are much more difficult to overcome. The level of tibial osteotomy is metaphyseal. It must be below the tibial tubercle to prevent damage, which could lead to genu recurvatum. The closer the osteotomy is to the physis, the better the alignment correction. Room must be left for fixation devices, and the physis itself should not be entered. The distal fragment should maintain its relationship to the proximal medial cortex; if it is positioned medial to it, the effect of the weight beating realignment gained by the valgus ~position is lessened. Correction of varus deformity should be to a position of slight valgus. Examples of osteotomies for tibia vara are shown in Figs. 13 and 14A-14E. Many techniques have been used to hold the valgus, external rotation and extension positioning until healing. Our preference is an AO linear, Y, or T plate along with a long leg cast, but crossed Kirschner wires, staples, and other internal devices along with long leg casting have been used with good results. Price et al. have reported on the effectiveness of external fixators such as the Orthofix apparatus in holding correction (Fig. 15) (139). They allow for more proximal osteotomy than linear AO plates. 2. COMPLETION OF PROXIMAL TIBIAL PHYSEAL ARREST If there is relatively little growth remaining or if the bone bridge is too extensive, it is advisable to complete the proximal tibial growth plate arrest laterally in conjunction with the valgus osteotomy to prevent postsurgery loss of correction. Any shortening that might occur is preferable to the

SECTION III ~ Infantile Tibia Vara (Blount's Disease)

491

medial tibial joint surface above the physis in tibia vara with a good result 16 years postsurgery. Siffert reported on an intra-epiphyseal osteotomy for marked tibia vara to elevate the medial articular plateau, a procedure which does not cause further damage (Fig. 16) (152). It attempts to correct residual medial epiphyseal and articular surface obliquity rather than relating to the metaphyseal deformity itself. Siffert also reported on an intra-epiphyseal osteotomy performed on a female 6 years 7 months old with a 13.5-year follow-up (152). A good long-term result was demonstrated, with the X ray at 16 years of age showing congruent articular surfaces with no evidence of shortening. Metaphyseal valgus osteotomy is added. Accurate positioning of cuts is needed or the osteotomy can crack into the articular weight bearing surface. Roy and Chaise (142), Sasaki et al. (144), and Gregosiewicz et al. (66) performed the medial elevating osteotomy below the physis (or through the bone bridge), carefully directing the proximal cut into the intercondylar region (Fig. 17). This would be done only for type V and VI lesions. Full physeal obliteration is warranted in such situations, and iliac crest or other bone graft may be needed to support the medial elevated segment. K. Adult Sequelae of Childhood Tibia Vara FIGURE 13 Tibiavara with low-gradeLangenskioldlesionresponsive to early proximal tibial-fibular valgus-derotation osteotomy. Correction held throughoutremainderof growth. (A) Preoperative, (B) Postoperative. worsening of tibial surface obliquity, which invariably accompanies recurrent varus. 3. FOCAL PHYSEAL BONE BRIDGE RESECTION If there are several years of growth remaining and a bone bridge has formed medially, good results are reported on occasion with bone bridge resection, use of an interposition material, and valgus osteotomy (100, 101). Beck et al. reported good results following bridge resection and replacement by fat or Cranioplast (14). Osorio and Costa reported good results in two with methyl methacrylate used as the interposition material (132). Bone bridge resection and the application of transphyseal distraction techniques, chondrodiatasis, also can be attempted (25). Focal bone bridge resection is discussed in Chapter 8, Section X. 4. ELEVATIONOF DEPRESSEDMEDIAL TIBIAL ARTICULAR SURFACE In those cases with grade IV, V, or VI tibia vara, generally there is clinically significant depression or tilting of the proximal medial tibial articular surface. Surgical correction of this deformity is an integral part of management but must be done accurately or damage can be worsened. Storen was the first to report on intra-epiphyseal proximal tibial osteotomy in tibia vara with the operation performed at the end of skeletal growth (159). He reported on operative elevation of the

There have been relatively few detailed studies on the longterm effects of a childhood tibia vara. If the disorder is corrected by a proximal tibial and fibular valgus osteotomy in the early years of life without the recurrence of deformity, a good to excellent long-term prognosis would be expected. In those with progressively higher Langenskiold grades, throughout the growth years modeling of the medial proximal tibial epiphysis, including the shape of the articular surface, is progressively abnormal. Lateral tibial subluxation, overgrowth of the medial femoral condyle, and asymmetric weight bearing all lead to increased wear and tear on the articular and meniscal cartilages. The long-term result thus is dependent more on articular relationships than on femoraltibial diaphyseal alignment (Figs. 18A-18C). Medbo noted 2 excellent, 5 good, and 3 fair results when assessments were done at skeletal maturity (118). Fifty valgus proximal tibial osteotomies had been needed in 30 affected knees. Zayer noted no degenerative change in 133 knees with tibia vara under the age of 30 years, but degenerative change was observed in 11 of 27 affected knees over 30 years of age (179). There was no absolute relationship between degenerative change and the degree of varus. Hofmann et al. noted early degenerative arthritis in more than one-half of the knees in their small series (75). Twelve patients, 19 knees, who had osteotomies for tibia vara were evaluated after skeletal maturity at an average of 22.4 years. The average age at initial osteotomy was 7.5 years (range = 4-11 years). Twelve of the 19 knees already were symptomatic and 8 showed early degenerative changes by arthrotomy

F I G U R E 14 A series of five radiographs shows how proximal tibial osteotomy done early and repeated as needed can lead to good joint structure at skeletal maturity even in the presence of massive obesity. (A) Bilateral bowing and proximal tibia vara were wellestablished at 3 years of age when the patient weighed 100 lb. Proximal tibial and fibular valgus osteotomies straightened both limbs well. Correction on the right has held throughout growth with the patient now 12 years of age. (B) On the left there was recurrent deformity by age 6 years when the patient weighed 150 lb. Repeat osteotomy was done. (C) Recurrence on the left again by age 12 years led to a final osteotomy at age 13 years (D, E) with weight 250 lb. Note the excellent joint alignment as well as femoral-tibial alignment maintained by early and repeat osteotomies, even with the clinical persistence of massive obesity.

SECTION IV 9 A d o l e s c e n t Tibia Vara

493

F I G U R E 17 In those instances in which a type VI tibia vara exists with bone bridge formation medially, better geometric and anatomic correction often can be achieved with osteotomy through the physis and outer region of the metaphysis as shown here. Following elevation of the proximal medial tibial articular surface and adjacent epiphysis and metaphysis the lateral physis must be obliterated to prevent any recurrence of deformity. [Reprinted from Roy and Chaise (1978), Rev. Chir. Orthop. 65:187-190, 9 Masson Editeur, with permission.]

or arthroscopy. These studies support the value of early osteotomy in tibia vara, by which is meant osteotomy shortly after diagnosis is made at 2-3 years of age, especially if the patient is obese and the Langenskiold grade is II or more.

IV. A D O L E S C E N T T I B I A VARA F I G U R E 15 A unilateral external fixator can be adapted for the stabilization of proximal valgus and derotation osteotomies for tibia vara. An illustration of the use of the Orthofix apparatus by Price e t al. is shown here. [Reprinted from Price e t al. (1995), J. Pediatr. Orthop. 15:236-243, @ Lippincott Williams & Wilkins, with permission.]

,

!a~L-,3

vz

..:; r

A ~~~176176

%~

~ ,;,A~.vuu o o UUOU ,

:..

"e/ '

o C ,0~ ,

"/

)~. ' "~ D(

I ,(I,i!;'! "/

F I G U R E 16 Siffert reported on intra-epiphyseal osteotomy for types V and VI deformities. This procedure does not damage the physis further and can be adopted for use with grade IV or V lesions. [Reprinted from Siffert (1982), J. Pediatr. Orthop. 2:81-85, 9 Lippincott Williams & Wilkins, with permission.]

A. Terminology Adolescent tibia vara refers to bowing due to diminished growth of the proximal medial tibial physis that develops after 6-8 years of age. This variant was recognized by Blount (19) in his initial paper as well as by Langenskiold (98). More recent studies show no association with trauma, a predominance of black patients in American studies, a male predominance, and a high incidence of massive obesity (17, 112, 162, 171). The deformity appears to worsen markedly during the adolescent growth spurt. The patient often presents not only with deformity but also with a limp, discomfort, and a lower extremity length discrepancy. This disorder is representative of a premature asymmetric arrest of epiphyseal growth rather than a dysplasia. The term late-onset tibia vara has been used to define this disorder by Thompson and Carter, with subdivision into juvenile (4-10 years) and adolescent (11 years or older) subsets (163). The infantile, juvenile, and adolescent forms appear to represent a continuum of medial physeal problems with the more severe forms manifesting themselves earlier.

B. Clinical Profile A review of studies reveals an increasingly clear and uniform clinical profile. These studies comprised assessments

494

CHAPTER 6 ~

FIGURE 18

Epiphyseal Disorders of the Knee

(A-C) Examples of severe deformity in tibia vara due to extreme obesity, relatively late initial osteotomy, and subsequent inattention to proximal medial tibial articular surface obliquity and persisting lateral physeal growth with diminished to absent medial physeal growth. Although femoral-tibial diaphyseal alignment can be readily addressed even at skeletal maturity, it is articular surface obliquity, tibial subluxation, and, to a lesser extent, patellar subluxation and internal tibial torsion that lead to adult osteoarthritis. Each of the examples here shows severe deformity at skeletal maturity. Most of these patients were treated in the 1960s at a time when little to no attention was paid to the medial proximal tibial articular surface obliquity or to the marked diminution of physeal growth in type IV-VI lesions. Note the occurrence of proximal medial tibial articular surface obliquity, relative overgrowth of the medial condyle of the distal femur, lateral subluxation of the tibia in relation to the distal femur, and severe persisting internal tibial torsion as indicated by the relationship of the proximal fibula to the tibia.

SECTION IV ~ Adolescent Tibia Vara

of 45 patients with 64 knees involved with the adolescent tibia vara. The incidence of bilateral to unilateral involvement was 19 to 26, but in the two largest series each with 15 patients there was equal involvement overall between bilateral and unilateral cases. There was a clear male predominance in each study, with a 3.2:1 ratio documented. There was an extremely high incidence of black patients in each of four studies, all from the United States. The incidence was 41 of 45 patients (91%). The age at onset and age at diagnosis also were in the adolescent range. Children were noted to have the peak age of onset between 10 and 11 years of age whereas the first osteotomy generally was performed between 12 and 14 years of age. Obesity, which was generally massive, was an almost invariable finding. In 37 of 38 patients obesity was documented, with the weight either beyond the 2nd standard deviation above the mean or beyond the 95th percentile in relation to age on the weight charts. The degree of obesity often was massive with the patients not only beyond the normal range but beyond by a considerable margin. The truly massive nature of the obesity was such that, in the New Orleans series, every patient was greater than the 95th percentile in weight and most exceeded that value by 30-50 kg (17). All heights were in the normal range. The weight exceeded the 95th percentile in every case as reported by Thompson et al., and again the massive nature was indicated by the fact that their weights exceeded the 95th percentile by a mean of 20 kg per child (range = 10-61 kg) (162). The mean values of varus deformity at the time of initial surgery were similar in each of the four studies, ranging between 13 and 22 ~ The final valgus correction was good, ranging in terms of mean values from 0 to 3 ~ valgus. Each series tended to report some undercorrections leaving the patient in a varus range. Recurrence was not commonly seen, and when present usually was in the younger aged patient who still had a few years of growth remaining. In spite of surgical correction, the results rarely were defined as excellent and indeed only 56% were described as good. Much of the long-range problem appeared to be due to the massive obesity as much as to the specific alignment of the bones. The overall length discrepancy problem was not extensive, primarily because so many of the patients had bilateral involvement. In addition, the deformity tends to occur when relatively little growth remains. The younger patient and the presence of unilateral involvement can lead to lower extremity length discrepancies at skeletal maturity, although only isolated instances of discrepancies greater than 2.0 cm are reported.

C. Variable Opinions on Whether Late-Onset Tibia Vara Is Superimposed on a Preexisting Varus Deformity Beskin et al. noted that a clear history of preexisting varus in earlier childhood was obtained in less than one-half of

495

7 patients had significant physiologic varus in early childhood that spontaneously improved, although not completely (171). In adolescence the deformity again worsened. Loder et al. noted only 1 patient with a history of persistent childhood genu varus (112). Patients in the Ohio groups had no evidence of significant physiological bowleg during early childhood (162, 163). Henderson and Greene documented 2 patients with late-onset tibia vara in whom neutral mechanical alignment of the femurs and tibias was clearly documented in an extremity that subsequently developed the tibia vara within 19 months (72). They concluded that preexisting varus alignment was not invariably a prerequisite for development of the late-onset tibia vara. Most patients noted progressive bowing over a short period of time, usually 6-12 months, in association with the adolescent growth spurt.

D. Physeal Height and the Question of Distal Femoral Varus Tilt Beskin et al. noted persistent widening of the lateral femoral physis when compared to the ipsilateral medial femoral physis or the opposite normal lateral side (17). They felt that this contributed to lateral overgrowth due to an unbalancing of forces and less stress on that region. In the early stages of the disorder, the medial tibial physis also was widened but had a fragmented appearance that subsequently proceeded to a focal growth diminution with narrowing. Growth inhibition appears to be occurring medially and growth stimulation laterally. Beskin et al. felt that relative overgrowth of the lateral condyle occurred, leading to a tilt into varus of the distal femoral articular surface. Thompson and Carter also noted increased lateral physeal height of the proximal tibia in 8 knees and increased medial height in 4 but reported the alignment of the femoral condyles to be normal (163). Knee deformity was due to angulation between the proximal tibial articular surface and the diaphysis. Often there is a mild posteromedial depression of the articular surface of the tibia. Most series report minimal to absent proximal medial tibial metaphyseal beaking and little tibial torsion in the adolescent form. Recurrences are relatively infrequent and tend to occur in those presenting and operated on under 10 years of age. Currarino and Kirks noted lateral widening of the distal femoral and proximal tibial physes, as well as varus deformity in an obese 12-year-old black male with onset at 11 years (37). Henderson et al. performed a prevalence study of lateonset tibia vara in the group at risk (73). They assessed both white and black adolescents who weighed at least 210 lb and who were screened for varus alignment. Their chart illustrates that not only were the patients greater than the 95th percentile in weight but most were exceeding the 95th percentile by an average of 50 lb. Radiographs of the 7 boys who screened positive showed that 2 had late-onset tibia vara. In thei~ study, 80 adolescent black males weighing at 1

,

pwlk I

f'~

11

1

11

.

.

.

.

496

CHAPTER 6 9 Epiphyseal Disorders o f the Knee

tibia vara. Sixty obese adolescent white males were screened in an identical fashion and no cases of tibia vara were identified. They also noted that the prevalence of late-onset tibia vara had increased significantly in their practice over a relatively short time period, a fact that they attributed to the overall increase in morbid obesity in the adolescent population in America. From 1985 to 1993 they indicated that they were diagnosing the same number of cases of adolescent tibia vara as of infantile tibia vara, which defined a marked change from studies a few decades earlier when the adolescent form was considered to be relatively rare.

E. Association of Femoral Varus with Tibia Vara in Late-Onset Blount's Disease Kline et al. documented the fairly frequent existence of femoral varus as a significant deformity in late-onset tibia vara (90). They assessed six patients with onset older than 6 years of age. Five of their six patients were black and obese. They felt there was an average of 10~ of varus deformity of the distal femur such that more than half (range = 56-76%) of the knee varus deformity in late-onset patients resulted from femoral varus. The femoral component is not metaphyseal but appears to be due to relative undergrowth of the medial femoral distal condyle and relative lateral femoral overgrowth as defined by others, showing widening of the physis laterally. These findings must be taken into consideration in planning realignment osteotomies for tibia vara in the adolescent phase. The distal femoral varus leads to a tilting of the joint axis, and the ideal alignment is to have the joint axis parallel within 2-3 ~ to the ground in the stance phase. On occasion, therefore, distal femoral valgus correction is warranted as part of total orthopedic management for adolescent tibia vara because both distal femoral and proximal tibial vara are present. The alignment of each of the distal femur, distal femoral and proximal tibial articular surfaces, and proximal tibia must be assessed on standing radiographs in determining levels of surgical correction.

F. Radiographic Assessments Because much of the epiphysis has developed by the time that late-onset tibia vara occurs, the radiographic changes in the disorder are less marked than in the infantile variants. The proximal medial tibial epiphyses are somewhat wedgeshaped due to mild to moderate flattening of the medial portion of the epiphysis. Both Langenskiold (102) and Thompson et al. (162) have noted the narrowing of the middle portion of the medial half of the epiphysis in most instances. The varus is present between the articular surface of the medial tibia and the diaphysis, but medial beaking of the metaphyseal area is not a major feature. The physis itself has a variable plain radiographic appearance. During the developmental stages of the disorder there tends to be some widening of the physis medially, although bone on the adja-

cent metaphyseal side is irregular rather than linear and often is described as ragged in appearance. The lateral physis may be of normal height or on occasion somewhat widened due to decreased compressive stresses there. On occasion this is associated with slight lateral condylar overgrowth of the distal femur. With time the physis on the medial side narrows. Although bone bridges can be formed late, they are not as characteristic a feature as in the severe infantile variants. Kline et al. have observed that on occasion a considerable part of the varus deformity at the knee is due to the bowing of the distal femur as well as that of the proximal tibia (90). Thompson et al., however, noted that the alignment of the femoral condyles was normal in all knees such that the varus deformity in their patients was due exclusively to angulation between the proximal tibial articular surface and the diaphysis (162). There was minimal, if any, metaphyseal widening in the anteroposterior or lateral projections. Tomography, on occasion, revealed a mild posteromedial depression of the articular surface of the tibia. Beskin et al. also noted consistent widening of the lateral femoral physis when compared to the medial femoral physis of the same knee (17). They felt that the lateral tibial physis was not significantly affected in their patients. The medial tibial physis demonstrated a characteristic irregular widening as well. They also documented relative overgrowth of the lateral femoral condyle, with a tilt in the transcondylar axis from a normal value of 90 ~ to a tibia vara value of 96.3 ~

G. Pathoanatomy of Adolescent Tibia Vara The histopathology of adolescent tibia vara is documented more extensively than that of infantile tibia vara even though its occurrence is less frequent (26, 162, 171). Biopsy samples can be obtained more readily because surgical treatment is performed near the end of skeletal growth and frequently involves osteotomy along with a complete proximal tibial epiphysiodesis to prevent recurrence. The biopsy samples show a marked narrowing of the proximal medial growth plate in comparison to the lateral side. The physis shows the changes characteristic of increased pressure, which are similar to what one notes in association with physiologic growth plate closure. Chondrocyte turnover ceases in the germinal and proliferating zones and the region of columnar cells is shorter than normal. The appearance of the physis is disorganized. There is a paucity of hypertrophic chondrocytes. Vascular invasion occurs both from the epiphyseal side into the resting or germinal zones and from the metaphyseal side through the hypertrophic to the columnar zones. The appearance is quite consistent with the presumed pathophysiology of an asymmetric premature medial growth plate closure due to the increased pressure from obesity often superimposed on a slight preexisting varus position. Histologic sections from the medial growth plate region of a 12-year-old male with adolescent tibia vara were described in detail by Thompson et al. (162). Within the physis

SECTION V ~ Osgood-Schlatter Disease (Tibial Tubercle Chronic Traumatic Apophysitis)

itself were scattered islands of active but irregular ossification, which were primarily but not exclusively on the metaphyseal side. At the regions of endochondral ossification were irregular trabeculae of woven bone rimmed with osteoblasts together with persisting cartilage. In some sections were several small collections of necrotic cartilage surrounded by a zone of proliferating chondrocytes, and in other sections transphyseal bone bridges were noted. Many of the cartilage regions were hypercellular and several of the cells appeared necrotic. The normal endochondral ossification mechanism was structurally irregular and the bone in the metaphysis was both woven and lamellar, appearing to represent intramembranous repair bone rather than endochondral bone. There was no evidence of an inflammatory process or avascular necrosis of bone. The results of most studies are consistent with the finding of cartilage necrosis with subsequent failure of growth, fibrovascular invasion, and in terminal stages transphyseal bone bridge formation. Abnormal pressure on the physeal cartilage appears to be the principal etiologic factor causing its degeneration and necrosis. Medial physeal tissue from two patients with adolescent tibia vara was described by Wenger et al. (171). The histopathologic studies demonstrated numerous clefts and fissures throughout the hypertrophic zone of the physis with marked structural disorganization. Regions of cartilage that persisted were seen extending deep into the metaphysis. There was irregular alignment of the trabeculae and large amounts of fibrovascular repair tissue in the metaphyseal region. There was loss of the normal orderly cartilage columnation. The chondrocytes were in disorganized clusters with extensive intercellular matrix at the cartilage-bone junction. Metaphyseal bone itself also was abnormal with the medial trabeculae aligned transversely rather than longitudinally. Large amounts of fibrovascular repair tissue also were noted as were islands of fibrocartilage in the metaphysis, suggesting an injury and repair response. Wenger and associates attributed the histologic changes to mechanical disruption of the growth plate as evidenced by the irregularity and fissuring of the cartilage and by the repair tissue at the junction of the physis and metaphysis. The cartilage also appeared hypocellular although areas of chondrocyte clustering were seen. Pitzen and Marquart described a histologic study of a proximal medial tibia in adolescent tibia vara in which they noted strictly localized degeneration of the growth plate (137).

497

worsening of deformity and possibly eliminate the need for osteotomy. Any correction, however, is dependent on continuing function of the proximal medial tibial physis. If, however, there is sufficient growth remaining that the deformity would worsen, one can elect to complete the growth plate arrest by performing a proximal tibial epiphysiodesis across the entire extent of the physis. If the deformity is minimal this may be sufficient; if not, proximal tibial and fibular valgus osteotomy would be needed as well. In those cases in which the varus deformity is moderate to severe, the best approach is to complete the proximal tibial epiphyseal arrest across the entire physis in definitive fashion and correct the deformity with the proximal tibial and fibular valgus osteotomy. As noted in the preceding sections, attention must be given to the distal femur, which may have varus deformity as well and may need correction by osteotomy in addition to the tibia. There appears to be a relatively minor role for preserving growth by bone bridge resection. In adolescent tibia vara the involvement tends to be more extensive than in the infantile variant where often it is localized to a relatively small area of the medial peripheral physis.

V. O S G O O D - S C H L A T T E R D I S E A S E (TIBIAL TUBERCLE CHRONIC TRAUMATIC APOPHYSITIS)

A. Terminology and Outline Osgood-Schlatter disease refers to pain at the tibial tubercle region in adolescence due to partial tearing of the anterior part of the tubercle bone and cartilage by forces exerted on it from the patellar tendon and the extensor knee mechanism. Tibial tubercle apophysitis was defined as a distinct entity by both Osgood (131) and Schlatter (146) in 1903. Clinical awareness of the specific symptom complex along with early revelation of the characteristic radiographic changes on lateral films of the proximal tibia led to characterization of the disorder. Both Osgood and Schlatter in their separate papers differentiated the condition from the previously recognized complete traumatic avulsion of the tibial tubercle. It is accompanied by variable discomfort and a prominence on the anterior surface of the tibia at and just below the insertion of the patellar tendon due to soft tissue swelling and reactive bone and cartilage proliferation.

H. Treatment Adolescent tibia vara develops in a slow and relatively painfree fashion initially, but almost always is well-established in terms of damage to the medial proximal tibial physis when increasing pain and deformity lead to diagnosis. For this reason preventive or conservative measures are difficult. On occasion the varus is sufficiently mild that no treatment is needed particularly if the patient is close to skeletal maturity. Lateral hemi-epiphysiodesis can be done to prevent the

B. Pathophysiology The condition is characterized by repeated strain placed on the tibial tubercle by the quadriceps muscle acting through the quadriceps tendon, patella, and patellar tendon during the final years of skeletal growth. The tibial tubercle originally is formed in cartilage, which is continuous as an anterior and distal projection from the proximal tibial epiphyseal cartilage. Repeated stresses during the final years of skeletal

498

CHAPTER 6

9

Epiphyseal Disorders of the Knee

growth can lead to injury (partial tearing) of the cartilage tubercle and an irregular ossification pattern characterized by separate ossification regions distinct from the normal advancing ossification front. Repetitive trauma remains widely accepted as the cause of the Osgood-Schlatter disorder (32, 48, 50, 51,129, 166, 174). The quadriceps femoris muscle is the largest muscle in the body and inserts via the patellar ligament and patellar tendon onto the relatively small tibial tuberosity. The development of the tuberosity and its relation to this disorder has been well-described by several workers. Extensive studies of the disorder were performed by Ehrenborg and associates (46-51), who also concluded that the problem represented an avulsion of a portion of the patellar tendon from the tibial tuberosity. The skeletal age in 26 patients with Osgood-Schlatter disease was within the normal range in all but one, which led Yashar et al. to conclude that no general abnormality of physeal development was associated with the disorder (177). A positive correlation with the Osgood-Schlatter disorder is the finding of a lower than normal patellar angle. Sen et al. measured the patellar angle between the posterior articular surface and the line from the end of the inferior articular cartilage to the lower patellar apex (149). The angle formed by these two lines averaged 33 ~ in 68 joints with the disorder and 47 ~ in 71 age-matched and 198 adult controls. The authors reasoned that the smaller angle in Osgood-Schlatter patellae required greater quadriceps muscle contraction to perform the same amount of work, thus increasing the likelihood of a traction apophysitis.

C. Tibial Tuberosity Development The proximal epiphysis of the tibia consists of a cartilage mass at birth. In the front of this is a cartilaginous segment or tuberosity, which projects anteriorly and distally to rest in front of the metaphysis. One or more centers of ossification will develop toward the tip of this projection distally and then merge with bone from the main body of the proximal tibial epiphyseal secondary ossification center. 1. HULTING Hulting has divided development of the tibial tuberosity into three stages: cartilaginous, apophyseal, and bony (79). The cartilage stage persists in girls until a mean age of 11 years and in boys until 13 years. The ossification of the tibial tuberosity takes place about 2 years earlier in girls than in boys. The apophyseal stage of development in girls occurs between 8 and 12 years and in boys between 9 and 14 years. There are, therefore, wide variations in ossification patterns of the tibial tuberosity. The bony stage refers to development after the tibial tubercle ossification center and the proximal tibial epiphyseal center have united into one bony mass. This third stage occurs in girls from the age of 10 years on and in boys from 11-12 years on. Definitive fusion of the tibial

tubercle to the underlying metaphysis and diaphysis occurs at about 18 years of age. 2. EHRENBORG AND ENGFELDT Ehrenborg and Engfeldt described four stages of tibial tuberosity development (Fig. 19), a description that is slightly more accurate than the three-stage interpretation (46-48). The tuberosity forms from a tongue of cartilage growing from the upper tibial epiphysis over the anterior aspect of the tibial metaphysis. A single ossification center usually forms in this cartilaginous tongue, although occasionally the center may be multiple. Formation of the tibial tuberosity can be divided into four stages: (1) the cartilaginous stage; (2) the apophyseal stage in which ossification centers appear distally on the tongue of cartilage; (3) the epiphyseal stage in which the centers have coalesced to form a tongue of bone, which is continuous with that of the proximal tibial epiphysis; and (4) the bony stage in which the epiphyseal growth plates have closed. As a general rule, the cartilaginous stage is still present at 11.5 years, the apophyseal stage from 11.67 years, the epiphyseal stage at 13 years, and bone union after 15 years in the female and 17 years in the male. Initial bone of the tibial tubercle forms during the apophyseal stage from a separate center (separate, that is, from the secondary ossification center bone of the proximal tibial epiphysis). The bone nucleus is in the lower anterior quarter of the cartilaginous tubercle. In the epiphyseal stage, the bone nucleus of the tibial tubercle has extended proximally and is beginning to establish bony communication with the secondary center of the upper tibial epiphysis. Symptoms generally arise when the tibial tuberosity is in the apophyseal or epiphyseal stage. Most of the fibers of the patellar ligament at these stages are inserted into the cartilage anteriorly. 3. OGDEN, HEMPTON, AND SOUTHWICK The evolution of the tuberosity is described in seven stages from the prenatal period to skeletal maturity (Table I) (128, 129). During the cartilage phase, the tibial tubercle growth plate region initially is fibrocartilaginous with a columnar physeal-like region only found proximally in the region continuous with and adjacent to the proximal tibial growth plate. As the tuberosity matures through the apophyseal and epiphyseal stages, the most important cellular changes are the progressive modulation of the growth plate from fibrocartilage to a true cell columnation. Once the ossification center has formed completely (epiphyseal stage), early in the second decade of life, the columnated growth plate extends almost to the end of the tuberosity. During the late fetal period, the hyaline cartilage outgrowth of the proximal tibial epiphysis is progressively distally displaced to become situated adjacent to the anterior portion of the tibial metaphysis. Four to 6 months postnatally, a growth plate develops under the tibial tuberosity. The growth plate of the tibial tuberosity has three distinct regions: (1) a zone of endochondral bone formation; (2) a zone

SECTION V ~ Osgood--Schlatter Disease (Tibial Tubercle Chronic Traumatic Apophysitis)

Stages of Tibial Tubercle Development Post

Ant

Cartilage Stage

Apophyseal Stage

Epiphyseal Stage

Bony Stage

Rare to O-S

Relation to Osgood-Schlatter Development F I G U R E 19 The four stages of tibial tuberosity development and its relationship to Osgood-Schlatter disease are illustrated. These were defined by Ehrenborg and Engfeldt. The lower part of the illustration demonstrates the mechanism of formation of the Osgood-Schlatter lesion, which occurs almost exclusively in the apophyseal and epiphyseal stages.

TABLE I

D e v e l o p m e n t o f Tibial T u b e r o s i t y a

Prenatal Phase

Stage 1

No tibial tuberosity is present. The growth plate of the proximal tibia is oriented transversely.

Stage 2

Development of an anterior outgrowth from the tibial chondroepiphysis at same time as fibrovascular ingrowth and vascularization of the chondroepiphysis.

Stage 3

Relative distal displacement of the tuberosity by longitudinal growth at the proximal tibial physis and anatomical separation from the proximal tibial physis by continued fibromesenchymal-vascular ingrowth.

Postnatal Phase

Stage 4

Development of a separate growth plate associated with the tibial tuberosity and subsequent coalescence with the primary proximal tibial growth plate.

Stage 5

Development of a secondary ossification center in the distal portion of the tuberosity.

Stage 6

Coalescence of the ossification centers of the tuberosity and the proximal tibial epiphysis.

Stage 7

Closure of the contiguous growth plates of the proximal tibia and tuberosity.

aFrom Ogden and Southwick (129).

499

500

CHAPTER 6 ~ Epiphyseal Disorders of the Knee

of intramembranous bone formation through fibrocartilage; and (3) a zone of intramembranous bone formation through fibrous tissue (128, 129). The Ogden and Southwick study included tissue from fetus and stillborn from 40 to 360 mm crown-rump length and from children ranging in age from 1 week to 12 years, as well as from several young adults. There is no specific delineation of the tuberosity at the 85-mm stage of development. At 120 mm, there is the beginning of tuberosity formation as an outgrowth of the hyaline cartilage of the chondral epiphysis, below which a region of fibroblastic-mesenchymal-vascular (FMV) tissues starts to grow into the chondroepiphysis. The patellar tendon begins to send fibers into the developing tuberosity. By 150 mm, the tuberosity is more distally situated and structurally separated from the remainder of the chondroepiphysis by the FMV region. Even by 180 mm, the tuberosity has no growth plate. It is, however, level with the hypertrophic zone of the proximal tibial growth plate. The anterior border of the growth plate grades into the FMV region as does the tibial tuberosity cartilage. The perichondrial ring still readily demarcates the growth plate from the FMV region and tuberosity. The demarcation between the hyaline cartilage of the tuberosity and the fibrous tissue of the FMV region is distinct. Immediately prior to birth, there is cartilage continuity between the epiphyseal cartilage anteriorly and the tuberosity cartilage. The proximal tibial growth plate ends abruptly, however, prior to reaching the posterior border of the developing tuberosity. At this junction, blood vessels cross the metaphysis into the epiphysis effectively separating the proximal tibial physis from the tuberosity. There still is no evidence of a growth plate between the tuberosity and the metaphysis. This region is separated by the groove of Ranvier including the bone bark and the adjacent FMV region just posterior to the tuberosity. No evidence of growth plate on the posterior surface of the tuberosity adjacent to the FMV region is seen. By 7 months of age, the most proximal part of the tibial tuberosity has developed an early growth plate continuous with that of the proximal tibia but that still, however, does not show a finely ordered endochondral sequence. Just distal to this is the fibrocartilaginous zone of the tuberosity growth plate with immediate transition to bone without an intervening hypertrophic cartilage zone. At the lowest part of the tuberosity, a fibrous zone continues to exist between the tuberosity cartilage and the metaphyseal cortical bone. The transition proximal to distal, therefore, is from undifferentiated hyaline cartilage to fibrous tissue to bone. At 15 months, a vascular and fibrous region still separates the proximal tibial growth plate from the tibial tuberosity growth plate. The patellar tendon inserts primarily into the distal region of the tubercle. It has an extensive insertion into the tuberosity from the level of the fibrocartilaginous region to the distal end of the tuberosity. The three regions described previously are still evident at 24 months. There is still a steplike gradation from proximal to distal of endochondral tissue, fibro-

F I G U R E 20 Tissue types are shown at the junction of the tibial tubercle with the proximal tibial metaphysis as outlined by Ogden et al. This tissue orientation is present throughout most of the growth period. PT, patellar tendon; EC, epiphyseal cartilage; SOC, secondary ossification center; P, physis; M, metaphysis.

cartilaginous tissue, and fibrous tissue (Fig. 20). Even at 12 years, the same orientation persists. At 11 years, there is beginning ossification in the distal end of the tibial tuberosity. At the same time, the proximal tibial ossification center is beginning to extend distally into the tuberosity cartilage by the endochondral sequence. By 13 years, there is a continuous ossification center between the proximal tibial epiphysis and that of the tibial tuberosity. The tubercle growth plate is changing its structure with the columnated physeal region, being found under a larger amount of the tuberosity ossification center. The columns, however, still are quite narrowed and stretched. The fibrocartilaginous region is localized to the most distal portion of the tuberosity. The patellar tendon inserts into the ossification center via a region of fibrocartilage with minimal intervening hyaline cartilage anteriorly. 4. STRUCTURE OF THE MATURE TIBIAL TUBEROSITY The tuberosity is divided into a smooth proximal part and a more roughened distal section (Fig. 21). At maturity, the tibial tuberosity is a fairly prominent elevated region of bone on the anterior proximal surface of the tibia (77, 107). A groove separates it from the bone above at its superior surface. The groove outlines the superior and lateral margins of

SECTION V ~ Osgood-Schlatter Disease (Tibial Tubercle Chronic Traumatic Apophysitis)

501

In summary, the strongest part of the insertion of the ligamentum patella is into the groove on the anterior surface of the tibia and just proximal to the tubercle. In addition, tendon fibers are attached weakly to the upper smooth part of the tuberosity below the groove and firmly more distally to the superior margin and inferior angles of the roughened area (Hughes and Sunderland). The largest muscle of the body, the quadriceps femoris, passes into the quadriceps tendon and then via fibers, which mostly are inserted into the patella. Some fibers run anterior to the patella and join up at the apex with other fibers of the patellar ligament before continuing to reach the tibial tuberosity.

F I G U R E 21 Relationships of the tibial tubercle to insertion of the patellar tendon in adulthood are shown. Lewis has defined these two patterns. At left the ligamentum patella is attached to both epiphyseal and diaphyseal segments with (a) being the smooth epiphyseal part of the tibial tubercle and (b) the lower ridged and roughened diaphyseal part terminating below in a crest or ridge. At right the area of attachment of the ligamentum patella is almost entirely to the epiphysis with (d) representing the smooth epiphyseal prominence of the tibial tubercle limited below by the crest (e). [From Lewis, O. J. (1958). J. Anat. 92:587-593. Reprinted with the permission of Cambridge University Press.]

the tuberosity and begins at the proximal medial angle of the tuberosity, where it frequently expands into a shallow trough. From that point, it curves downward and outward. In about 75% of bones, there is a localized area proximal to the groove that is smooth. The tuberosity itself is subdivided into two approximately equal areas, a proximal smooth part and a distal rough part. The line of demarcation is obliquely arranged and somewhat lower to the lateral side. The strongest insertion of the ligamentum patella is into the groove that delineates the tuberosity proximally and laterally. The portion attached to the commencement of this groove at the proximal medial angle extends for a few millimeters along the medial border of the tuberosity. Distal to this insertion, some fibers are attached to the smooth area but easily can be stripped from the bone. Below this, the tendon fibers are firmly attached to the superior margin and inferior angles of the roughened areas of the tuberosity. The epiphyseal line cuts obliquely across the tuberosity along the lines separating the smooth and rough areas. Both medially and laterally it turns sharply upward and descends before turning at right angles to run posteriorly. Medially, it ascends to the groove region but laterally it does not ascend so high. The patellar tendon inserts proximally into the epiphyseal prolongation and then continues distally, where it fuses with the periosteum covering the diaphysis in the region of the rough part of the tuberosity (Fig. 21). The periosteum along the anterior border of the tibia separates into two parts at the apex of the epiphyseal beak: one portion continues anterior to the beak where it fuses with the ligamentum patellae and the other ascends posteriorly as far as the base of the beak. This leads the epiphyseal extension, therefore, to be covered both anteriorly and posteriorly by fibrous tissue.

D. Pathoanatomic Changes in Osgood-Schlatter Disease 1. GENERAL CONCEPTS

There have been many reports of the pathoanatomy. Most studies, including those of Ehrenborg and Engfeldt, do not demonstrate separations along the growth plate of the tuberosity itself; rather, the separations occur from the anterior surface and midportion of the tuberosity (47) (Fig. 19). The absence of specific growth plate tearing also is evidenced by the rarity of genu recurvatum due to premature closure of II the physis of the tibial tuberosity. Ogden and Southwick's II studies "seemingly would support the concept of avulsion through the early ossification center either as it is transforming from cartilage to bone or once the bone has formed while the normal tuberosity growth plate which appears to be a specific adaptation to tensile stress would be minimally involved if at all" (129). Ogden and Southwick feel that "Osgood Schlatter's disease is an avulsion of a portion of the developing ossification center and overlying hyaline cartilage.., the fibrocartilaginous growth plate appears to be a structural adaptation to prevent avulsion of the tibial tuberosity away from the anterior tibial metaphysis." The disorder thus is due to an inability of the developing secondary ossification center of the tubercle to withstand tensile forces, resulting in avulsion of segments of the ossification center and the eventual formation of extra bone ossicles between the fragments. Once understanding of the normal structure has been clarified, the changes noted are relatively consistent and straightforward. Virtually all of the studies reported are from excised surgical specimens rather than being assessments in situ. 2. LAZERTE AND R A P P

Lazerte and Rapp found no evidence of inflammation or primary aseptic necrosis of the tibial tuberosity (104). They felt that the disorder was of traumatic origin with the avulsed bone fragment showing some necrosis but primarily new bone formation. Each of nine patients described were males ranging in age from 13 to 17 years. Abridged descriptions from some of their cases follow.

502

CHAPTER 6 9 Epiphyseal Disorders of the Knee

Case 1: The anterior surface of the tibial tuberosity was irregular with a separate fragment noted a few millimeters away. Histologically there was a defect at one point in the anterior cortex of the tuberosity filled with highly cellular and vascular granulation tissue. The separate ossicle in the tendon seemed to fit into this defect. The broken spicules of the tuberosity were viable and covered by thick layers of osteoid lined with osteoblasts. The bone of the separate ossicle was mostly necrotic although the marrow was viable. There was active new bone formation about the necrotic spicules as well as osteoclastic resorption. Case 2: Skeletal maturity had been reached and there was a loose ossicle in the tuberosity separate from the main bone. All bone of the tibial tuberosity was viable. The bone defect was filled with moderately vascularized fibrous connective tissue. There was callus and new bone formation at the margins of the defect and fibrous and fibrocartilaginous tissue between the tuberosity and the loose ossicle. The surface of the ossicle facing the underlying bone was covered by callus. Case 3: This specimen included the entire lower tip of the tibial tuberosity, the epiphyseal plate, and the thin layer of underlying metaphysis. Near the distal tip of the tubercle where the epiphyseal plate blended with the tendinous tissue, the matrix of the cartilage showed some degeneration but no necrosis. There was no evidence of bone necrosis in the tuberosity itself. Case 4: Radiographs showed irregularity on the anterior surface of the tibial tuberosity and separate irregular bodies in the tendon. The epiphyseal line was open and orderly and the cartilage of the growth plate was well-preserved. The metaphysis was normal. The tuberosity showed an irregular anterior surface with marked osteoblastic activity and osteoid deposition. There was some fibrocartilaginous callus. Some of the loose spicules of bone were necrotic, but surrounded by repair tissue. Case 5: The cartilaginous epiphyseal plate was wellpreserved. The tuberosity, however, showed a large defect bordered by callus, the epiphyses were closed, and loose ossicles were seen. The ossicles in the tendon were relatively mature, but the margin closest to the tuberosity showed evidence of new bone formation plus proliferation of cartilage with direct transformation into bone. The tissue between the ossicles and underlying bone was composed of dense cartilaginous tissue and fibrocartilage. Case 6 (7 in text): Radiography showed the open growth plate, irregularity on the anterior surface of the tuberosity, and small separate densities in the tendons. By histology, the metaphysis was normal as was the epiphyseal plate. The portion of the bony epiphysis of the tuberosity next to the epiphyseal plate was intact and normal. The anterior surface of the tuberosity was irregular. The cortical bone anteriorly exhibited small defects filled with cellular and vascularized connective tissue and callus. The bone spicules were covered by osteoid seams and prominent osteoblasts. Some parts of the ossicle were necrotic.

Lazerte and Rapp summarized as follows. A defect in the anterior cortical bone of the tibial tuberosity was found in every case. A zone of proliferating connective tissue filled and surrounded the defect. Histologic sections showed irregularly arranged osteoid tissue and occasionally cartilage. The bone spicules showed osteoblastic proliferation and new bone formation. There was no evidence of necrosis in either the epiphyseal bone of the tuberosity or the epiphyseal plate. There was no evidence of infection. The loose ossicle in the tendon was composed of bone surrounded by callus and areas of necrosis in some cases. In some cases, a cleft resembling a pseudarthrosis was found in the fibrocartilage tissue between the tuberosity and the ossicle or between adjacent ossicles. The epiphyseal plate did not seem to be involved primarily in these cases. There was no necrosis and no lines of cleavage through the cartilage. Polarized light showed the defect to be filled with poorly oriented fibrovascular tissue. Lazerte and Rapp concluded that "the intra-tendinous ossicle in Osgood-Schlatter's disease was derived originally from an avulsed fragment of the tibial tuberosity." The ossicle was enlarged by the proliferation of callus about it, and the defect in the anterior cortex of the tibial tuberosity was gradually filled by scar and callus tissue. The avulsed fragment was occasionally necrotic but was quickly replaced by ew bone. Lazerte and Rapp concluded that "the best explanation for the pathogenesis of the condition was that the force of quadriceps contraction concentrated on a small portion of incompletely developed bone produced an avulsion fracture." The line of separation of the fragment occurred not through the cartilage plate but through the tibial tubercle ossification center itself (Fig. 19). 3. UHRY Uhry noted that the ossification center of the tibial tuberosity appears at about 11 years of age (166). Several months later, union of the tibial tuberosity bone center and that of the advancing center from the proximal tibial epiphysis occurs. Fusion of the tuberosity with the shaft of the tibia lags behind union of the remainder of the upper epiphyseal complex and often does not occur until around 18 years of age. Gross anatomy shows that the lateral and medial patellar retinacula have central portions attached to the patella, which then pass into the patella ligament. There are two thinner and less axial portions that attach to the tibia along the oblique ridges as far from the midline as the site of attachment of the medial and lateral collateral ligaments. The patellar ligament is attached to the tibial tubercle in the upper part of the tibial crest in somewhat oblique fashion, longer on the lateral side. This ligament is continuous medially and laterally with the retinacula. The collateral attachments of the patella complex are able to preserve the greater part of the extension power of the knee, even when the central patellar ligament attachment to the tibial tubercle has been removed. Uhry described the pathological changes in OsgoodSchlatter disease from 23 specimens. His interpretation

SECTION V ~ Osgood-Schlatter Disease (Tibial Tubercle Chronic Traumatic Apophysitis)

was that "the fundamental pathologic change in OsgoodSchlatter disease is separation of 2 or more structures comprising the tibial tubercle complex with the interposition of scar tissue between them." There appeared to be no local disease process antedating the separation. The scarring and callus formation were generally seen between the patellar ligament and the apophysis on the anterior face of the tibial tubercle. This was seen in 13 of 23 specimens. In 3, the defect was present on the posterior surface of the apophyseal cartilage plate, and in 4 others, it was in the zone of growth at the junction between the columns of proliferating cartilage and the newly formed endochondral bone. In others, variable areas were seen. Uhry felt that hemorrhage had occurred early in the evolution of the disorder. At the time of the examination, however, it either had disappeared or was wellorganized. Osteoblasts and osteoclasts were seen. Necrotic bone in the loosened fragments anteriorly was seen occasionally and was being replaced by creeping substitution. Uhry also performed a detailed assessment of the pathological changes described in the literature, which were consistent with interpretation of the condition as a repair reaction to trauma. This region was predisposed to injury based on the specific anatomic developments in the early part of the second decade and the extreme forces exerted on it by the powerful quadriceps muscle. The majority of separations occur at the anterior face of the apophysis rather than at the growth plate itself. Uhry concluded that the disorder develops on the basis of a minor separation of structures, one from another, in the complex comprising the tibial tubercle and patella ligament and that the characteristic pathologic changes represent scar-callus repair at the site or sites of separation. The separation may be a clean one or may entail the inclusion of bone fragments in a softer part as it is torn free. . . . the majority of separations occur at the ligament-apophysis face, the anterior apophyseal surface. Uhry also listed 14 histopathologic contributions to the literature between 1927 and 1937. 4. EHRENBORG AND ENGFELDT As part of their large study, Ehrenborg and Engfeldt wrote on the histologic changes, which were then correlated with clinical and radiographic features (47). Both Osgood and Schlatter indicated in 1903 in their initial papers that the lesion arose by direct or indirect violence and resulted in a fracture through the tuberosity of the tibia. Several subsequent authors felt that the disorder was an osteochondritis, but little evidence had accumulated to accept this opinion. Surgical material was available from each of 17 patients operated out of 170 in their series, a 10% incidence of surgical intervention. Ehrenborg and Engfeldt noted that, in the normal development of the tibial tuberosity, a center of ossification occurs at the distal tip and then grows or enlarges more proximally until it fuses with the secondary ossification center of the proximal tibial epiphysis. This phase of

503

development is found in girls between 8 and 12 years of age and in boys between 9 and 14 years. Their histologic observations showed that the changes essentially were localized to the area of the actual tendon attachment with "no lesions at the epiphyseal junction with the tibia or within the spongiosa of the tibial tuberosity." No degenerative changes were found that were felt to be primary to the disorder, and there were no changes indicative of any underlying pathologic process. Ehrenborg and Engfeldt concluded that "all the observed changes seemed to be explainable by trauma with avulsion in the attachment of the ligamentum patellae to the tuberosity of the tibia." They felt that the complex of changes would be difficult to attribute to a single injury but rather was due to a series of repetitive mild traumas. In some early cases, there were tears between the ligament and the tibial tuberosity anteriorly. They felt that the avulsion took place within the anterior cartilage. In a companion study of 53 patients based on X rays, much interesting information was accumulated. In 41 cases, new bone formation was anterior to the tibial tubercle ossification center, and in 6 cases, it appeared where the ossification center eventually appeared. 5. ADDITIONAL HISTOPATHOLOGIC STUDIES AND INTERPRETATIONS Several other histologic studies outlining the associated pathoanatomy have been performed, although these have been subject to varying interpretations. The observations are generally in the same range and similar to those described in the preceding sections, although the interpretations as to causation vary. Asada and Kato (4) studied several cases with good radiographic and histologic assessments. They considered the findings to be consistent with repetitive trauma mediated by the patellar tendon in relation to causing traction injuries of the tibial tubercle tissues. Cole (32), describing similar radiographic and histologic findings, placed causation as due to rapid growth during adolescence, which tended to place the quadriceps muscle under great physiologic strain resulting in changes within the patellar tendon. These changes within the tendon tended to alter the blood supply, and Cole felt that the radiographic appearance in the tibial tubercle was based upon the altered circulation within the patellar tendon and its attachment. Fibrocartilaginous regions appeared in the patellar tendon due to chronic repetitive trauma within the tendon, and these areas subsequently became calcified and later ossified. The constitutional predisposition or hereditary cause was postulated by yon Lutterotti (169). Jentzer and Perrot (84) focused on the circulatory disturbance within the tendon-tubercle junction. Jentzer and Perrot also interpreted the histopathology to indicate that the tuberosity of the tibia became incapable of fully resisting the microtraumatic insults transmitted by the patellar ligament and thus gave way partially or became otherwise altered. They proposed the term apophysiolysis of the tuberosity of the anterior tibia. By this they referred to a

504

CHAPTER 6 9 Epiph~tseal Disorders of the Knee

pathologic fracture of the tuberosity, which itself had been altered by a dystrophic process under the influence of repeated traction from the patellar ligament. Jentzer and Perrot reviewed the structure and development of the tibial tuberosity and concluded that it provided a region of resistance to the tendinous forces upon it. They pointed out that because the vast majority of adolescents did not develop an OsgoodSchlatter disorder there had to be some additional factor other than mechanical forces to cause the disorder. That additional factor involved abnormalities of the circulation. They felt that all of the histopathologic findings from tissue necrosis at the more extreme spectrum to evidence of bone and cartilage repair could be attributed to the mediation of abnormal circulation at varying phases of the disorder. By apophysiolysis Jentzer and Perrot were defining a clinical state in which the diverse radiographic and histopathologic findings varying from tissue degeneration to tissue repair were due to an interplay of circulatory, functional, traumatic, and perhaps humoral factors. They also pointed out that on occasion the disorder did not fully resolve at the termination of adolescence and presented with symptomatic knee discomfort at the tibial tubercle site well into adulthood.

E. Clinical and Radiologic Features of Osgood-Schlatter Disease 1. EHRENBORG AND ENGFELDT

Their study comprised 170 patients, 68 girls and 102 boys (48). The lesion was bilateral in 48 (28%) patients. The mean age at the commencement of symptoms was 10 years 7 months in girls and 12 years 7 months in boys. The mean age at initial medical consultation was 11 years 6 months in girls and 13 years 3 months in boys. Ehrenborg and Engfeldt divide the clinical assessment of the disorder into stages based on the radiographic developmental appearance of the tibial tuberosity. Stages are (A) the cartilaginous stage before any ossification centers are seen; (B) the apophyseal stage in which the separate tubercle ossification centers appear in the tongue of cartilage; (C) the epiphyseal stage in which the bone centers have coalesced to form a tongue of bone, which in turn has fused with the main tibial secondary ossification center; and (D) the bony stage in which the epiphyseal line has closed. The active Osgood-Schlatter lesion occurs in developmental stages A-C, with stage D referring only to residual deformity (Fig. 19). At the time of the initial radiographic exam, one-third of the cases were in stage B and two-thirds in stage C with less than 1% in either stage A or stage D. The final radiological results were divided into four groups. Group 1: Normal radiographic appearance of the tibial tubercle region. Group 2 (the most common appearance): Fragmentation had healed, leaving only a small deformity of the tibial tubercle contour with a slight anterior bony prominence. Group 3: Healing of fragmentation took the form of accretion on the tibial tuberosity, which was very prominent in some cases. Group 4: The displaced fragment

did not unite firmly with the tibial tuberosity but instead gave rise to large or small separate ossicles, which tended to be symptomatic for a considerable period of time. Approximately 25% showed this final picture with 57 of 218 damaged knees demonstrating it. The age of occurrence is relatively narrow with symptoms not seen before the age of 8 years in girls and 10 years in boys. Fresh Osgood-Schlatter lesions are not found in those older than 14 years. As noted previously, the disorder results from a partial tear in the region of attachment of the ligamentum patella to the cartilage at the anterior part of the tibial tubercle with a part of the cartilage accompanying the avulsed tendon fibers. The entire appearance can be described as based on the "dislocation of fragments of cartilage." A common finding in one-fourth of the cases was the separate ossicle best explained by a pseudarthrosis mechanism.

2. UHRY Similar clinical numbers were shown by Uhry in his study of 79 cases (166). The age at onset of symptoms (when known) was concentrated in the 11- to 13-year group, and the age of presentation for treatment showed most patients between 11 and 15 years of age. The most prominent time of onset of symptoms in females was in the 10- to 11-year age group in females and in the 12- to 13-year-old group in males. Radiographic abnormalities are shown in Figs. 22A-22D.

F. Clinical Symptoms and Management The clinical situation was well-described by Osgood, who indicated that the "avulsions of a small portion and partial separation of the tubercle are common. They do not cause complete loss of function, but without treatment, are a long continued serious annoyance (131)." 1. GENERAL TREATMENT APPROACH The general treatment approach is threefold. (1) The initial approach involves a decrease in physical activity to allow the injured region to repair itself. This then should be followed by physical therapy to strengthen the quadriceps muscles, which almost always will have weakened during the period of discomfort and decreased function. (2) Immobilization is sometimes required if symptoms continue. In milder cases this is done with a knee immobilizer, which may be removed for bathing and gentle range of motion exercises. If the disorder is more marked or the patient is noncompliant, a cylinder cast is worn for a 3- to 4-week period. In the vast majority of cases, decreased activity or immobilization will lead to repair without recurrence. (3) In less than 5-10% of cases, repair does not occur and symptoms persist or recur after each attempt at immobilization. The lateral film shows the development of a separate ossicle, which has not united with the tibial tubercle. If the patient is 1 year or more beyond skeletal maturity and symp-

SECTION V ~ Osgood-Schlatter Disease (Tibial Tubercle Chronic Traumatic Apophysitis)

505

F I G U R E 22 Osgood-Schlatter radiographs in the apophyseal (A) and epiphyseal (B) stages of the disorder are shown. Part (C) shows the bony prominence at skeletal maturity. White arrows point to the bony reaction. Part (D) shows intact physis with a clear defect with tibial tubercle bone.

toms persist, it is advisable to remove the ossicle surgically and pare down the adjacent bony reactive sites with a rongeur (Fig. 23). 2. SURGICAL APPROACHES Surgical approaches to the Osgood-Schlatter condition have been fairly extensive and have changed through the years. At present, surgery is considered only after skeletal maturity when symptoms persist due to a nonunited bone fragment (ossicle). The initial approaches sought to bring about a union of the loose ossicles with the underlying tibial tuberosity and proximal tibia. The surgery was fairly extensive, the time of immobilization was considerable, and there were a number of failures. Bosworth described several operations that had been used to enhance stability and mini-

mize pain, varying from pegging with ivory, metal, and wood screws to autogenous bone grafting (20). Thomson reported on several approaches involving drilling of the tubercle, use of autogenous bone pegs, excision of the tubercle, and complete excision of the tubercle with adjacent periosteum (164). DePalma also drilled the tibial tubercle (164). More recently, the simpler approach of removing any loose ossicle at skeletal maturity has become accepted, leading to excellent results with a minimal period of immobilization (Fig. 23). This was reported in 1980 by Mital et al. (122). Glynn and Regan have discussed their results involving 44 procedures equally divided between an earlier series treated by drilling of the tibial tubercle without removal of the prominent tubercle and their more recent approach in which the loose pieces of cartilage or bone were excised

506

CHAPTER 6 ~ Epiphyseal Disorders of the Knee

F I G U R E 23 A loose ossicle in a patient with Osgood-Schlatter disease at skeletal maturity is shown. This was symptomatic and was removed in association with trimming of the prominent bone edges. The preoperative film is shown at left and the postoperative result at right.

without removal of the tubercle or any drilling intervention (60). They concluded that there was a much higher incidence of good or excellent results in the latter and recommended the simpler operation, which allowed for rapid mobilization and return to full activity. Glynn and Regan indicated that approximately 70% of patients with the disorder improved with simpler measures and did not require surgery. In most centers, however, a surgical approach would be required in a far smaller proportion of patients. The general feeling is that there are no indications for operating in stages A, B, or C but only in stage D when the likelihood of spontaneous repair is dramatically diminished. Many nonfused ossicles are asymptomatic and require no intervention. Binazzi et al. (18) also strongly supported the effectiveness of excising painful intratendinous ossicles with or without excision of the prominent parts of the adjacent tibial tubercle in comparison with efforts designed to bring about bony fusion of the nonincorporated fragments. 3. COMPLICATIONS OF O S G O O D - S C H L A T T E R DISEASE There are relatively few severe long-term complications of Osgood-Schlatter disease. Even though the disorder occurs close to a major growth region, the fact that there are relatively few long-term growth sequelae indicates, consistent with the pathoanatomic findings, that it is

the bone and cartilage of the anterior part of the tibial tubercle that are affected pathologically rather than the epiphyseal growth plate itself. The lack of complications also is consistent with observations that the tibial tubercle-tibial junction involves a true physis only in the upper one-third of the posterior aspect of the tibial tubercle, below which the tissue is fibrocartilaginous and fibrous such that the eventual final union of the tubercle to the anterior metaphysis and diaphysis is via an intramembranous mechanism rather than a true endochondral sequence. The major long-term growth problem, although it rarely is reported, is genu recurvatum. Description of this complication was provided initially by Stirling, who reported premature fusion of the anterior part of the growing epiphysis with continued growth of the posterior part leading to genu recurvatum (158). Two formal case reports in patients 11 years old at disease onset by Jeffreys (83) and Zimbler and Merkow (180) observed that both cases with genu recurvatum eventually required correction with proximal tibial flexion osteotomies. Lynch and Walsh also have reported two cases of genu recurvatum complicating the Osgood-Schlatter disorder (116). 4. ADULT SEQUELAE OF OSGOOD--SCHLATTER DISEASE

a. Woolfrey and Chandler. A review of 272 consecutive knee radiographic examinations performed in males com-

SECTION VI ~ Congenital Dislocation o f the Knee

plaining of pain showed approximately 10% with changes in the tibial tuberosities (176). The abnormal tuberosities were felt to be due to adolescent Osgood-Schlatter disease. Three types of change were noted on lateral proximal tibial radiographs: a prominent, irregularly surfaced tibial tubercle; a prominent, irregularly surfaced tubercle with a free particle of bone anterior and superior to the tuberosity; and a smooth tuberosity with a free bone particle anterior and superior to the tuberosity. On occasion discomfort continued after conservative management and responded to surgical removal of the loose bone fragment. The fragments were embedded in the posterior aspect of the patellar tendon. b. K r a u s e et al. A series of 69 knees in 50 patients with no specific treatment in the acute phase was reassessed at a mean of 9 years postdiagnosis (95). At review, 38 patients (76%) had no limitation of activity except for difficulty kneeling. Anterior knee pain, usually directly over the prominent tibial tubercle, was the problem. In the 69 affected knees there were separate ossicles in 28, an abnormally shaped tuberosity in 20, and 21 showing a smooth tibial outline. An interesting correlation was noted between initial radiographs and radiographs and clinical findings at maturity. Those with no bony change initially, only soft tissue swelling, all had normal tibial radiographs at maturity and were asymptomatic, whereas those with fragmentation of apophyseal bone at presentation all showed abnormal tuberosities at review, many of which were symptomatic. There were no cases of premature epiphyseal closure.

VI. C O N G E N I T A L THE K N E E

DISLOCATION OF

A. Definition and Clinical Profile Congenital dislocation of the knee (CDK) refers to a fixed hyperextension of the leg at the knee joint present at birth. The deformity is apparent clinically at birth with the affected knee hyperextended 10-20 ~ with further extension often possible such that the toes sometimes may be placed against the chest. The initial clinical description was by Chatelain in 1822, who reported on a newborn with her leg bent onto the anterior part of the thigh (28). The child presented head first and there was no evidence of swelling, inflammation, or trauma to account for the defect. Reduction without pain was achieved readily by manipulation, but deformity recurred with the release of manual pressure by an immediate contraction of the thigh extensor muscles. Treatment was successful after use of a splint holding the knee in the flexed position for 15 days with full correction noted at 23 days. Chatelain reasoned that the disorder developed in utero well before birth, that it was not a birth injury, and that it was due to the inertia or dysfunction of the flexors of the thigh and the strong action of their antagonists.

507

Potel (1897) reviewed congenital knee deformities in great detail and illustrated the congenital knee subluxation and dislocation variants (Fig. 24) (138). Congenital dislocation of the knee is rare in comparison to congenital or developmental dislocation of the hip. Bensahel et al. estimated that it was 40-80 times as rare as the hip disorder (15). Curtis and Fisher had a CDK:CDH ratio of 1:40 in a 25-year period at Newington, CT (38). Kopits indicated that congenital dislocation of the knee occurred once for every 100 cases of congenital dislocation of the hip (94). Jacobsen and Vopalecky also calculated a CDK: CDH ratio of 1:100 for their Scandinavian population (81). There is a female predominance of approximately 2:1. In seven series from 1960 to 1989 a female predominance always was noted; the female:male ratio was 98:44 or 2.2:1. In a Shriner's Hospital survey the female:male ratio was 99: 56 or 1.8:1 (89). When all of these series are assessed together, a female:male ratio of 197:100 or 1.97:1 is determined. Bilateral cases are more frequent than unilateral, with 60-70% being bilateral. In single side involvement both are affected equally. Breech presentations are more common than in the normal population and range from 10 to 41% but still do not account for the majority of disorders. Unlike CDH or DDH in which the large majority of cases are present in an otherwise normal child, isolated cases of congenital dislocation of the knee are relatively rare. The large majority of patients have additional connective tissue disorders, with the two most common being congenital dislocation of the hip, which can be present in as many as 50% of patients, and clubfoot, which can present in as many as 33% of patients. A relatively high number of patients also are syndromal with such disorders as Larson's disease with multiple joint dislocations, Ehlers-Danlos disease, or arthrogryposis, defined as multiple joint contractures present at birth. In many series each of the patients has an associated disorder, and when large numbers of groups are collated it is generally felt that anywhere from 80 to 90% of patients have additional and often severe musculoskeletal abnormalities (38, 85, 89, 103, 126). The deformity develops slowly in utero because the popliteal vessels and nerve are invariably intact.

B. Classification The classification of Leveuf and Pais has gained wide acceptance (Fig. 25) (105). Type 1 involves severe genu recurvatum or hyperextension of the knee joint at birth without displacement of the joint surfaces of the femur in relation to the tibia and with the long axes of each bone crossing at the joint line. Type 2 presents with subluxation with the tibial epiphysis displaced anteriorly in relation to the distal femoral articular surface, and in type 3 there is a complete anterior dislocation of the tibial epiphysis so that it lies in front of the distal femoral condyles, leading to a marked hyperextension deformity.

508

CHAPTER 6 9 Epiphyseai Disorders of the Knee

Extension normale.

Genu recurvatum.

Luxation en avant.

FIGURE 24 Illustrationfrom the monograph by Potel (138) shows his definition of congenital knee dislocation, proceeding from normal extension to genu recurvatum to full anterior dislocation.

C. Pathoanatomy Characteristic soft tissue abnormalities are present with the disorder. The anterior part of the knee shows transverse creases with the skin posteriorly tending to be smooth. Soft tissue abnormalities vary in extent with those in type 1 being the mildest and those in type 3 the most severe. The greater the deformity, the more prominent the posterior femoral condyles in the popliteal space. Within the knee, there is either stretching or (rarely) congenital absence of the anterior and posterior cruciate ligaments, but the menisci and cruciates tend to be relatively normal in most. External to the knee joint are fibrous contractures of the extensor quadriceps muscle of the knee, the fascia lata, and the anterior joint capsule. The anterior capsule is tight, shorter than normal, and attached to the femur at the suprapatellar region. The tight contracture of the fascia lata also leads to lateral rotation and occasionally lateral subluxation of the leg as it thickens and adheres to the aponeurosis of the vastus lateralis. The extensor quadriceps mechanism is invariably fibrous, atrophic, and contracted. Both Potel (138) and Middleton (120) attributed the hyperextension deformity to contracture of the quadriceps muscle during the latter part of intrauterine life. Middleton showed histologic sections of quadriceps with fat and fibrous tissue replacing and surrounding otherwise normal appearing muscle groups. The scarring is more developed inferiorly and laterally with the vastus medialis muscle relatively spared. Usually there are multiple fibrous adherents to firmly fix the quadriceps muscle and patella to the distal femur. The extensor mechanism, including the patella, often is displaced laterally. The hamstring tendons are occasionally displaced anteriorly to further worsen the hyperextension. The patella is occasionally absent and usually is hypoplastic and more proximally placed than normal when present. The epiphyseal changes appear to be secondary to growth in abnormal positions and in relation to abnormal pressures. Assessments, therefore, in the newborn period tend to show relatively normal appearing distal femoral and proximal tibial epiphyseal regions. The articular cartilage appears normal. The bone shaping changes worsen with time in the subluxed or dislocated position. Femoral changes appear greater than tibial. The femoral condyles tend to be

flattened or squared and the intercondylar region is hypoplastic. The tibial plateau is hypoplastic with flattening of the tibial spines and posterior tilting of the posterior articular surface. Shattock (150) described the pathology of genu recurvatum in a fetus at term with excellent detail and reviewed the entire entity as well. Both knees presented with the hyperextension deformity in a child born at full term who appeared otherwise normal but who died from respiratory problems. The left knee could not be manipulated beyond full extension, whereas the right allowed only a few degrees of flexion. Dissection indicated that "the resisting structures (to flexion) were the anterior portions of the c a p s u l e . . . ; these became extremely tense." Even with cutting the capsule there was still difficulty in obtaining flexion of the knee with the next obstacle being the cruciate ligaments. The anterior fibers of the anterior cruciate ligament were particularly tight. There was some deformity of the end of the femur, which was described as markedly misshapen, but the articular surface of the tibia was relatively normal. The lower surface of the femur was short and flattened anteriorly. Shattock stressed

'",,,,,,,,,,%

A FIGURE 25 The classification of Leveuf and Pais as presented by Curtis and Fisher (38) is shown. The disorder thus ranges from genu recurvatum, in which the long axes of the femur and tibia still meet within the intra-articular region, to anterior subluxation, in which the distal femoral and proximal tibial articular surfaces still relate to each other but the long axes are misaligned, to a full dislocation, in which there is no longer any appropriate orientation of the two articular surfaces.

SECTION VI ~ Congenital Dislocation of the Knee

509

assessment early during the course of conservative treatment to make certain that true reduction is occurring and not a false or imperfect correction. Parsch and Schulz feel that ultrasound enabled them to control the progress of conservative treatment and more accurately find the time when conservative treatment was no longer of value, thus allowing for early surgical intervention. F I G U R E 26 An example of congenital dislocation of the knee is shown. In (B) (right) the lateral radiograph shows the displacement of the distal tibia in relation to the distal femur. The ultrasound study (A) outlines well the displacement of the articular surfaces in relation to each other. The patella is the oval structure centrally, the tibia is to the left, and the distal femur is at lower fight with its articular surface barely touching the posterior tibial surface. Generalized underdevelopment of the knee region is further shown by absence of distal femoral and proximal tibial secondary ossification center.

that the most immediately tight structure was the capsule in front of the lateral ligaments. Additional abnormal cases that he had dissected were described. He stressed that, after all of the muscles had been dissected free in the specimens in question, flexion of the knee still was not possible and "it was only after dividing the whole of the capsule in front of the lateral ligaments that the joint could be flexed: the articular surface of the femur had undergone alterations of form in adaptation to the overextended position in which the joint was fixed." He felt that the findings were consistent with a condition that was due to prolonged malposition in utero. Shattock was able to note that, although many cases were bilateral, there also were reports of unilateral involvement. He also noted that some cases appeared to be curable with manipulation, whereas others were so rigid that surgical intervention either was required or would have been required had the patient survived. It was well-recognized that the deformity was frequently accompanied by other malformations, with those of the hip and clubfoot being the most common. The disorder also was noted to occur, however, as the only malformation in some instances.

D. Diagnostic Considerations Clinical exam and plain radiographs usually allow for early diagnosis. The distal femoral and proximal tibial secondary ossification centers should be present at birth, which further aids in the determination of relationships at the knee by radiographs. Ultrasonography, however, has proved to be of value in early assessments of CDK (Fig. 26); in a study by Parsch and Schulz, the positioning of the cartilage models of the distal femur and proximal tibia was well-revealed and the classification of Leveuf was confirmed as accurate (135). Information now obtainable by ultrasound was previously determined by some using an arthrogram, but that procedure involved some form of anesthesia as well as injection of dye. An additional value of sonography is the ability to repeat the

E. Treatment Approaches Initial treatment involves range of motion exercises to enhance the flexion position and nighttime splinting with either plaster casts or plaster splints to allow for stretching of the tightened anterior quadriceps and capsular mechanism. Fulltime splinting also can be used. Operative intervention is used only if there is a failure of conservative therapy. The vast majority of cases with the type 1 or genu recurvatum diagnosis can be managed conservatively, but it may take several months for final stretching to occur. Ultrasound is of increasing value in assessing the effectiveness of management. Radiographically, often there is delayed ossification of the proximal tibial and distal femoral secondary ossification centers, and this should be assessed carefully both initially and during the course of treatment. Conservative and manipulative splint therapy also can be quite effective with the subluxed type 2 knee. For the formally dislocated type 3 knee, open surgical release generally is needed. Even in the type 2 and type 3 categories, however, the gentle manipulative therapy and splinting usually is warranted as soon as diagnosis is made. When manipulative therapy is performed, particularly in the type 2 subluxation group, it is important to assess the patient carefully either by radiologic or by ultrasonographic means to make certain that false correction is not occurring. The treatment also must be gentle to prevent epiphyseal growth plate fracture-separation or metaphyseal buckle fractures. Fractures are a real concern during manipulative therapy and can involve either distal femoral or proximal tibial epiphyseal growth plate fracture-separations or adjacent metaphyseal fractures. Surgical intervention is designed to remove the tightness associated with the extensor mechanism anteriorly. The surgical procedures performed are dependent on abnormalities found in the extensor apparatus. In most of the instances surgical treatment will involve (1) release of the tensor fascia lata in a transverse fashion, (2) freeing of the quadriceps muscle and tendon, particularly the vastus lateralis component, from abnormal attachment to the internal muscular septum and the anterior part of the femur, (3) a quadricepsplasty to lengthen the extensor mechanism by any of several V - Y approaches, (4) anterior capsule release in transverse fashion from medial to lateral aspects, (5) replacement of the hamstring tendons to a posterior position if they have subluxed or dislocated anteriorly, and (6) realignment of the patella in the midline as it frequently has been displaced laterally and proximally.

510

CHAPTER 6 ~ Epiphyseal Disorders o f the Knee

The obstacles to reduction present in virtually all cases are a tight anterior capsule and a contracted quadriceps mechanism. Curtis and Fisher have provided an excellent diagrammatic historic review of many of the operative methods of quadriceps lengthening, presenting 9 different approaches (38). As with most orthopedic deformities due to soft tissue contractures or looseness, the earlier the age at operation the better the result. Laurence reported excellent results in patients operated as early as 2 weeks of age (103). In his series, 15 infants responded completely within 8 weeks with conservative treatment, and those who were recalcitrant responded quite well to surgical intervention within the first year of life. Treatment of the CDK often is complicated by the need to treat adjacent joints, particularly hip and foot, and by the frequent presence of systemic disorders such as arthrogryposis or spina bifida. Johnson et al. also noted that results were better in unilateral cases and when surgery was performed before 2 years of age (85). If necessary, the adductor muscles and the lateral collateral and cruciate ligaments must be sectioned. Early surgical intervention, if needed, also was favored by Ferriss and Aichroth, who operated on several patients at an average age of 6 months, with a range of 1-15 months (55). Long-term results generally are good to excellent, although much is dependent on the nature of the disorder causing the CDK as well as the involvement of adjacent joints.

VII. V A L G U S A N G U L A T I O N F O L L O W I N G PROXIMAL TIBIAL METAPHYSEAL F R A C T U R E S IN C H I L D H O O D

A. Description and Clinical Profile One of the most troubling childhood fractures to treat in the first decade of life is an un-displaced or minimally displaced fracture of the proximal tibial metaphysis. A large percentage of patients develop a valgus angulation distal to the fracture site over the course of the first year or two following fracture. This entity was first defined in 1953 by Cozen (35), and many papers have been written reiterating its occurrence since then (10, 63, 80, 87, 130, 141, 181). Many etiologies have been proposed for the deformity. It has been noted to occur even in those situations in which the fracture was undisplaced and remained un-displaced throughout healing or in which healing occurred following reduction with no apparent loss of position while the fracture healed. A profile of the injury has emerged over the past several years. It appears to occur in from one-third to almost onehalf of such fractures, even when the treating physicians are completely aware of the likelihood of its occurrence and take every possible precaution to prevent it. The valgus angulation particularly is prone to occur when the patients are from 3 to 10 years of age at the time of injury and particularly when they are less than 5-6 years of age. The fracture heals

without malposition but is followed by increasing angular deformity over the next year or two, with the deformity remaining unchanged 2 years after injury. The valgus deformity expressed as the femoral-tibial diaphyseal angle generally ranges from 8 to 15~

B. Etiological Considerations Underlying Valgus Deformation The most convincing cause of deformity appears to be overgrowth of the proximal medial tibial physis, leading to progressive valgus angulation. The overgrowth appears to be related to the more increased blood supply adjacent to the medial periphyseal area than is seen laterally. Zionts et al. demonstrated increased radionuclide uptake in the medial portion of the proximal tibial growth plate on the fractured side with a bone scan done 5 months postfracture in a 6-year-old who was developing a tibia valga (181). Aronson et al. created an experimental model in the 8-week-old rabbit tibia to show the effects of asymmetric proximal tibial physeal stimulation (3). In 11 rabbits, the medial periosteum was removed from the left proximal tibial metaphysis and apartial medial metaphyseal osteotomy was made 5 mm distal to the physis. All 11 rabbits developed valgus deformity averaging 12.2 ~ A similar lesion was created laterally and led to varus deformation averaging 9.8 ~ in 10 of 11 animals. Many theories have been proposed to explain the posttraumatic tibia valga. These have been reviewed by Zionts et al. (181), Jordan et al. (87), Robert et al. (141), Green (63), Balthazar and Pappas (10), and Ogden et al. (130). The bulk of current evidence supports asymmetric growth stimulation of the medial proximal tibia secondary to increased vascular response. The disorder thus appears to represent an overgrowth phenomenon, which leads not only to angular deformity but also to a slight increase in the length of the involved tibia. It is not likely that a single cause underlies all cases. It is important to understand all of the theories proposed because some are well-documented and treatable and awareness on the part of the physician should decrease the likelihood or extent of some sequelae. Theory 1: Inadequate Reduction. Examples are seen in which the patient suffers a greenstick fracture of the proximal tibial metaphysis, which remains open medially with continuity of the lateral proximal tibial cortex. The best approach for this fracture is treatment in a long leg cast following closed reduction holding the distal fragment in a varus position and also keeping the knee maximally extended because flexion lessens control of the relatively short proximal fragment. An inadequately reduced greenstick fracture is not seen, however, in the large majority of cases because there are many well-documented instances of healing of complete tibial and fibular fractures in anatomic position with tibia valga occurring over the subsequent 1-2 years. Theory 2: Early Weight Bearing. The injury often appears to be relatively mild such that prolonged immobilization

SECTION VIII ~ Disorders o f the Proximal Fibular Epiphysis

may not be maintained until definitive radiologic evidence of full healing. Early weight beating in association with a loose cast can lead to progressive deformity.

Theory 3: Loss of Periosteal Restraint on the Medial Growth Plate. Rupture of the periosteum medially tends to lessen somewhat the transphyseal force and minimize this mechanical controlling feature of the periosteum. Houghton and Rooker demonstrated nicely that division of the periosteum in the metaphyseal region can lead to valgus deformity in rabbits (76). Others have postulated that it is the increased blood supply in relation to the repair phase following periosteal injury that is the stimulus rather than the fibroelastic membrane effect across the physeal region. Theory 4: Soft Tissue Interposition. Evidence has been found in which the periosteum is inverted into the medial fracture gap, thus preventing full reduction. Others who have explored these wounds routinely do not find soft tissue invariably to be interposed. Even in those situations in which it is interposed and subsequently is removed tibia valga can occur after healing. Theory 5: Tethering by the Intact Fibula. This theory tends to have been disproved by the frequent evidence of increased fibular growth in those patients with tibia valga. Theory 6: Salter-Harris Type V Fracture of the Lateral Physis. Evidence is more supportive of increased vascularity on the medial side, as can be shown by bone scan, rather than a limited growth phenomenon laterally. The fact that the involved bone usually is longer than that on the uninvolved side also limits the validity of the lateral proximal tibial physeal growth abnormality theory. Theory 7: Asymmetric Overgrowth. Currently this is supported by most authors with excellent evidence primarily based on increased bone scan activity medially and the appearance of tibial growth arrest lines, which indicate more growth medially than laterally. The overgrowth is due to asymmetric fracture-induced hyperemia.

C. Guidelines for Treatment Guidelines for treatment are becoming clear. It remains important to reduce accurately any fracture of the proximal tibial metaphysis and to hold it until definitive radiologic healing has occurred. Immobilization is recommended in a long leg cast with the knee fully extended and a varus force exerted at the fracture site. Completion of the proximal tibial fracture with or without the production of a proximal fibular fracture has been recommended by some, but the large majority of authors currently do not recommend that approach. Even with un-displaced fractures or with those that are perfectly reduced, the overgrowth phenomenon with valgus angulation still can occur. It also is clear that proximal tibialfibular varus osteotomies do not represent totally benign interventions. Although most doubt the spontaneous correction of valgus with growth, the large majority of instances show no further worsening of the deformity after 2 years.

511

The deformity can be somewhat unattractive cosmetically but rarely is it of functional significance. Osteotomy several years after the fracture still carries a high risk of recurrence and is best delayed until the time of skeletal maturity. For those desiring slightly earlier and less radical treatment, asymmetric proximal medial tibial epiphyseal arrest close to the time of skeletal maturity can be done. The negative cosmetic features of the valgus angulation are limited somewhat by increased muscle bulk about the limb and also stimulation of distal lateral tibial growth to minimize the valgus deformation.

VIII. DISORDERS OF THE PROXIMAL FIBULAR EPIPHYSIS Disorders of the proximal fibular epiphysis can occur in the following settings.

A. Congenital Proximal Tibial-Fibular Synostosis Congenital proximal tibial-fibular synostosis can occur. There is true bony union of the proximal secondary ossification centers of the tibia and fibula, but the overall femoraltibial articulation is not affected. Because there is little meaningful movement between these two bones their presence as a single fused mass has no clinical implications.

B. Proximal Fibular Elongation Relative overgrowth of the proximal fibular epiphysis in relation to that of the proximal tibia is a feature of some skeletal dysplasias. Those in which it is found in particular include achondroplasia, hypochondroplasia, mesomelic dysplasia, metaphyseal chondrodysplasia, and spondyloepimetaphyseal dysplasia. The slightly elongated fibula usually is apparent clinically but rarely causes functional problems. Peroneal function should be tested at examination, but it seldom is affected. On occasion there is an associated varus deformity of the knee, and some have felt that one causative mechanism could be the relative overgrowth of the fibula forcing the knee to the varus malposition. Due to the different sizes of the bones, however, it would appear more reasonable to assert that primary undergrowth of the tibia was the causative feature. On occasion, those favoring the former etiology have recommended the excision of the prominent fibular head in an effort to control the varus deformation. No results of this procedure have been reported. If used it is important to be extremely careful not to damage the peroneal nerve in the process.

C. Hypoplasia of Fibula On occasion the proximal end of the fibula is relatively underdeveloped in relation to that of the tibia. This too represents

512

CHAPTER 6 9

Epiphyseal Disorders of the

Knee

a radiographic finding without clinical significance. The peroneal nerve function should be assessed clinically to make certain that function is preserved. Hypoplasia is seen most commonly with relatively rare skeletal dysplasias such as campomelic dysplasia, chondroectodermal dysplasia (EllisVan Creveld), rare chromosomal abnormality syndromes, and the de la Chapelle dysplasia.

D. Hereditary Multiple Exostosis Prominent exostoses of the proximal fibular metaphysis are seen in hereditary multiple exostosis. These have the clear potential to damage by stretching the peroneal nerve with growth. If this occurs, careful excision of the exostosis and freeing of the nerve should result in the restoration of function. The proximal fibular exostosis can be associated with either varus or valgus deformation in association with extososis involvement of the proximal tibia. Should proximal tibial and fibular osteotomy be performed, it is extremely important to monitor peroneal nerve function as this can be worsened particularly with any instance of varus osteotomy. The risk of nerve stretching with varus osteotomy is increased in the presence of fibular exostoses.

E. Proximal Fibular Overgrowth Secondary to Damage to the Proximal Tibial Physis If there is growth damage to the proximal tibial epiphysis with continuing growth of the proximal fibular epiphysis, a change in the relationship of the two bones occurs. If the proximal fibular overgrowth is becoming a problem functionally or cosmetically, then proximal fibular epiphyseal arrest can be performed. In most patients having a proximal tibial epiphyseal arrest before the age of 10 or 11 years, an associated proximal fibular arrest also is recommended. After that age the fibula frequently is not treated surgically. There is always the risk of a peroneal palsy with any proximal fibula epiphyseal arrest, and that procedure should be done only with a meaningful clinical indication.

References 1. Aichroth P (1971) Osteochondritis dissecans of the knee: A clinical survey. J Bone Joint Surg 53B:440-447. 2. Aichroth P (1971) Osteochondral fractures and their relationship to osteochondritis dissecans of the knee: An experimental study in animals. J Bone Joint Surg 53B:448-454. 3. Aronson DD, Stewart MC, Crissman JD (1990) Experimental tibial fractures in rabbits simulating proximal tibial metaphyseal fractures in children. Clin Orthop Rel Res 255:61-67. 4. Asada T, Kato S (1927) Zur atiologie der sogenannten Schlatterschen krankheit. Zeit f Orthop Chir 48:191-216. 5. Axhausen G (1912) Ueber einfache, aseptische knochen- und knorpelnekrose, chondritis dissecans und arthritis deformans. Arch Klin Chir 99:519-574.

6. Axhausen G (1914) Die entstehung der freien gelenkkorper und ihre beziehungen zur arthritis deformans. Arch Klin Chir 104:581-679. 7. Axhausen G (1920) Zur entstehung der freien solitarkorper des kniegelenks. Deut Med Wochens 46:825-826. 8. Axhausen G (1922) Uever vorkommen und bedeutung epiphysarer ernahrungs-unterbrechungen beim menschen. Munch Med Wochens 69:881-884. 9. Axhausen G (1924) Uber der abgrenzungsvorgang au epiphysaren knochen (osteochondritis dissecans Konig). Virch Arch Path Anat 252:458-518. 10. Balthazar DA, Pappas AM (1984) Acquired valgus deformity of the tibia in children. J Pediatr Orthop 4:538-541. 11. Barth A (1898) Die entstehung und das wachstum der freien gelenkkorper. Arch Klin Chir 56:507-573. 12. Bateson EM (1968) The relationship between Blount's disease and bowlegs. Br J Radio141:107-114. 13. Bathfield CA, Beighton PH (1978) Blount disease: A review of etiological factors in 110 patients. Clin Orthop Rel Res 135:29-33. 14. Beck CL, Burke SW, Roberts JM, Johnston CE, II (1987) Physeal bridge resection in infantile Blount disease. J Pediatr Orthop 7:161-163. 15. Bensahel H, DalMonte A, Hjelmstedt A, Bjerkreim I, Wientroub S, Matasovic T, Porat S, Bialik V (1989) Congenital dislocation of the knee. J Pediatr Orthop 9:174-177. 16. Bernstein MA (1925) Osteochondritis dissecans. J Bone Joint Surg 7:319-329. 17. Beskin JL, Burke SW, Johnston CE, II, Roberts JM (1986) Clinical basis for a mechanical etiology in adolescent Blount's disease. Orthopaedics 9:365-370. 18. Binazzi R, Felli L, Vaccari V, Borelli P (1993) Surgical treatment of unresolved Osgood-Schlatter lesion. Clin Orthop Rel Res 289:202-204. 19. Blount WP (1937) Tibia vara: Osteochondrosis deformans tibiae. J Bone Joint Surg 19:1-29. 20. Bosworth DM (1934) Autogenous bone pegging for epiphysitis of the tibial tubercle. J Bone Joint Surg 16:829-838. 21. Bradley J, Dandy DJ (1989) Results of drilling osteochondritis dissecans before skeletal maturity. J Bone Joint Surg 71B: 642-644. 22. Bradway JK, Klassen RA, Peterson HA (1987) Review article: Blount disease: A review of the English literature. J Pediatr Orthop 7:472-480. 23. Caffey J, Madell SH, Royer C, Morales P (1958) Ossification of the distal femoral epiphysis. J Bone Joint Surg 40A: 647-714. 24. Cahill BR, Phillips MR, Navarro R (1989) The results of conservative management of juvenile osteochondritis dissecans using joint scintigraphy. Am J Sports Med 17:601-605. 25. Canadell J, DePablos J (1985) Breaking bony bridges by physeal distraction. Internat Orthop 9:223-229. 26. Carter JR, Leeson MC, Thompson GH, Kalamchi A, Kelly CM, Makley JT (1988) Late-onset tibia vara: A histopathologic analysis. A comparative evaluation with infantile tibia vara and slipped capital femoral epiphysis. J Pediatr Orthop 8:187-195. 27. Catonne Y, Dintimille H, Arfi S, Monchet A (1983) La maladie de Blount aux Antilles. A propos de 25 observations. Rev Chir Orthop 69:131-140.

References

28. Chatelain A-C (1822) Observation d'une luxation congenitale du tibia en artier. Bibliotheque Med 75:103-105. 29. Chiroff RT, Cooke CP, III (1975)Osteochondritis dissecans: A histologic and microradiographic analysis of surgically excised lesions. J Trauma 15:689-696. 30. Christ N (1929) Uber die enchondrale ossificationszone der distalen femurepiphyse. Arch Orthop Unfall-Chir 27:610-630. 31. Clanton TO, DeLee JC (1982) Osteochondritis dissecans: History, pathophysiology and current treatment concepts. Clin Orthop Rel Res 167:50-64. 32. Cole JP (1937) A study of Osgood-Schlatter disease. Surg Gynecol Obstet 65:55-67. 33. Conway FM (1937) Osteochondritis dissecans: Description of the stages of the condition and its probable traumatic etiology. Am J Surg 38:691-699. 34. Cook SD, Lavernia CJ, Burke SW, Skinner HB, Haddad RJ, Jr (1983) A biomechanical analysis of the etiology of tibia vara. J Pediatr Orthop 3:449-454. 35. Cozen L (1953) Fracture of the proximal portion of the tibia in children followed by valgus deformity. Surg Gynecol Obstet 97:183-188. 36. Crawfurd EJP, Emery RJH, Aichroth PM (1990) Stable osteochondritis dissecansmDoes the lesion unite? J Bone Joint Surg 72B:320. 37. Currarino G, Kirks DR (1977) Lateral widening of epiphyseal plates in knees of children with bowed legs. AJR 129:309-312. 38. Curtis BH, Fisher RL (1969) Congenital hyperextension with anterior subluxation of the knee. J Bone Joint Surg 51A: 255-269. 39. Dalinka MK, Coren G, Hensinger R, Irani RR (1974) Arthrography in Blount's disease. Radiology 113:161-164. 40. Decker P (1938) Guerison d'une osteo-chondrite dissequante bilaterale du genou. Medizin Wochenschr (J Suisse Med) 68: 221-223. 41. DeMoraes F, Perricone G (1959) Tibia vara (genu varum par osteochondrose tibiale). Acta Orthop Belg 25:285-316. 42. DeSanctis N, Della Corte S, Pempinello C, DiGennaro G, Gambardella A (1995) Infantile type of Blount's disease: Considerations concerning etiopathogenesis and treatment. J Pediatr Orthop 4:200-203. 43. DeSmet AA, Fisher DR, Graf BK, Lange RH (1990) Osteochondritis dissecans of the knee: Value of MR imaging in determining lesion stability and the presence of articular cartilage defects. AJR 155:549-553. 44. DeSmet AA, Ilahi OA, Graf BK (1997) Untreated osteochondritis dissecans of the femoral condyles: Prediction of patient outcome using radiographic and MR findings. Skelet Radiol 26:463-467. 45. Dietz WH, Jr, Gross WL, Kirkpatrick JA, Jr (1982) Bount disease (tibia vara): Another skeletal disorder associated with childhood obesity. J Pediatr 101:735-737. 46. Ehrenborg G, Lagergren C (1961) Roentgenologic changes in the Osgood-Schlatter lesion. Acta Chir Scand 121:315-327. 47. Ehrenborg G, Engfeldt B (1961) Histologic changes in the Osgood-Schlatter lesion. Acta Chir Scand 121:328-337. 48. Ehrenborg G (1961) The Osgood-Schlatter lesion: A clinical study of 170 cases. Acta Chir Scand 124:89-105. 49. Ehrenborg G, Lagergren C (1961) The normal arterial pattern of tuberositas tibiae in adolescents and in growing dogs. Acta Chir Scand 121:500-510.

513

50. Ehrenborg G, Engfeldt B (1961) The insertion of the liagmentum patellae on the tibial tuberosity: Some views in connection with the Osgood-Schlatter lesion. Acta Chir Scand 121: 491-499. 51. Ehrenborg G (1962) The Osgood-Schlatter lesion: A clinical and experimental study. Acta Chir Scand 288:1-36. 52. Erlacher P (1922) Deformierende prozesse der epiphysengegend bei kindern. Archiv f Orthop Unfall-chir 20:81-96. 53. Evensen A, Steffensen J (1957) Tibia vara (osteochondrosis deformans tibia). Acta Orthop Scand 26:200-210. 54. Fairbank HAT (1933) Osteochondritis dissecans. Br J Surg 21:67-82. 55. Ferris B, Aichroth P (1987) The treatment of congenital knee dislocation. A review of nineteen knees. Clin Orthop Rel Res 216:135-140. 56. Ferriter P, Shapiro F (1987) Infantile tibia vara: Factors affecting outcome following proximal tibial osteotomy. J Pediatr Orthop 7:1-7. 57. Fisher AGT (1921) A study of loose bodies composed of cartilage or of cartilage and bone occurring in joints. With special reference to their pathology and etiology. Br J Surg 8: 493-523. 58. Freiberg AH (1923) Osteochondritis dissecans. J Bone Joint Surg 5:3-16. 59. Gebele H (1936) Zur frage: Unfall und osteochondritis dissecans. Zentralbl Chir 63:2293-2299. 60. Glynn MK, Regan BF (1983) Surgical treatment of OsgoodSchlatter's disease. J Pediatr Orthop 3:216-219. 61. Golding JSR, McNeil-Smith JDG (1963) Observations of the etiology of tibia vara. J Bone Joint Surg 45B:320-325. 62. Green JP (1966) Osteochondritis dissecans of the knee. J Bone Joint Surg 48B:82-91. 63. Green NE (1983) Tibia valga caused by asymmetrical overgrowth following a nondisplaced fracture of the proximal tibial metaphysis. J Pediatr Orthop 3:235-237. 64. Green WT, Banks HH (1953) Osteochondritis dissecans in children. J Bone Joint Surg 35A:26-47. 65. Greene WB (1993) Infantile tibia vara. J Bone Joint Surg 75A:130-143. 66. Gregosiewicz A, Wosko I, Kandzierski G, Drabik Z (1989) Double-elevating osteotomy of tibiae in the treatment of severe cases of Blount's disease. J Pediatr Orthop 9:178-181. 67. Hanley WB, McKusick VA, Barranco FT (1967) Osteochondritis dissecans with associated malformations in two brothers: A review of familial aspects. J Bone Joint Surg 49A: 925-937. 68. Hansson LI, Zayer M (1975) Physiological genu varum. Acta Orthop Scand 46:221-229. 69. Heath CH, Staheli LT (1993) Normal limits of knee angle in white childrenmGenu varum and genu valgum. J Pediatr Orthop 13:259-262. 70. Hellstrom J (1922) Beitrag zur kenntnis der sogenannten osteochondritis dissecans in kniegelenk. Acta Chir Scand 55: 190-221. 71. Hellstrom J, Ostling K (1934) Ein klinischer beitrag zur kenntnis der osteochondritis dissecans. Acta Chir Scand 75: 273-318. 72. Henderson RC, Greene WB (1994) Biology of late-onset tibia vara: Is varus alignment a prerequisite? J Pediatr Orthop 14: 143-146.

514

CHAPTER 6 ~ Epiphyseal Disorders oJr the Knee

73. Henderson RC, Kemp GJ, Hayes PRL (1993) Prevalence of late-onset tibia vara. J Pediatr Orthop 13:255-258. 74. Hildebrand (1895) Experimenteller beitrag zur lehre von den freien gelenkkorpern. Deut z Chir 42:292-308. 75. Hofmann A, Jones RE, Herring JA (1982) Blount's disease after skeletal maturity. J Bone Joint Surg 64A: 1004-1009. 76. Houghton GR, Rooker GD (1979) The role of the periosteum in the growth of long bones: An experimental study in the rabbit. J Bone Joint Surg 61B:218-220. 77. Hughes ESR, Sunderland S (1946) The tibial tuberosity and the insertion of the ligamentum patellae. Anat Rec 96:439-445. 78. Hughston JC, Hergenroeder PT, Courtenay BG (1984) Osteochondritis dissecans of the femoral condyles. J Bone Joint Surg 66A: 1340-1348. 79. Hulting B (1957) Roentgenologic features of fracture of the tibial tuberosity (Osgood-Schlatter's disease). Acta Radiol 48:161-174. 80. Jackson DW, Cozen L (1971) Genu valgum as a complication of proximal tibial metaphyseal fractures in children. J Bone Joint Surg 53A:1571-1578. 81. Jacobsen K, Vopalecky F (1985) Congenital dislocation of the knee. Acta Orthop Scand 56:1-7. 82. Jaffe HL (1972) Osteochondritis dissecans. In: Metabolic, Degenerative, and Inflammatory Diseases of Bones and Joints. pp. 584-598, Philadelphia: Lea and Febiger. 83. Jeffreys TE (1965) Genu recurvatum after Osgood-Schlatter's disease: Report of a case. J Bone Joint Surg 47B:298-299. 84. Jentzer A, Perrot A (1941) Remarques sur la maladie d'Osgood-Schlatter. Rev Orthop 27:176-191. 85. Johnson E, Audell R, Oppenheimer WL (1987) Congenital dislocation of the knee. J Pediatr Orthop 7:194-200. 86. Johnston CE, II (1989) Infantile tibia vara. Clin Orthop Rel Res 255:13-23. 87. Jordan SE, Alonso JE, Cook FF (1987) The etiology of valgus angulation after metaphyseal fractures of the tibia in children. J Pediatr Orthop 7:450-457. 88. Kappis M (1920) Osteochondritis dissecans und traumatische gelenkmaus. Deutsch Z Chir 157:187-213. 89. Katz MP, Grogono BJS, Soper KC (1967) The etiology and treatment of congenital dislocation of the knee. J Bone Joint Surg 49B:112-120. 90. Kline SC, Bostrum M, Griffin PP (1992) Femoral varus: An important component in late-onset Blount's disease. J Pediatr Orthop 12:197-206. 91. Kling TF, Jr, Volk AG, Dias L, Morgan RC, Jr, DeRosa GP (1990) Infantile Blount's disease treated with osteotomy: Follow-up to maturity. Orthop Trans 14:634-635. 92. Koenig F (1887) Ueber freie korper in den gelenken. Deutz fur Chir 27:90-109. 93. Konjetzny (1924) Diskussionsbemerkung zum vortrage Lehmann uber osteochondritis. Zentral bl f Chir 51:2451-2453. 94. Kopits E (1925) Beitrage zur pathologie und therapie der angeborenen kniegelenkssubluxationen. Arch Orthop UnfallChir 23:593-609. 95. Krause BL, Williams JPR, Chir AC (1990) Natural history of Osgood-Schlatter disease. J Pediatr Orthop 10:65-68. 96. Lamy L, Weissman L (1946) Tibia vara. J Radiol Electro127: 409-414. 97. Langenskiold A (1952) Tibia vara (Osteochondrosis deformans tibiae): A survey of 23 cases. Acta Chir Scand 103:1-22.

98. Langenskiold A (1955) Aspects of the pathology of tibia vara (Osteochondrosis deformans tibia). Ann Chir Gynaec Fenn 44:58-63. 99. Langenskiold A, Riska EB (1964) Tibia vara (osteochondrosis deformans tibiae): A survey of 71 cases. J Bone Joint Surg 46A: 1405-1420. 100. Langenskiold A (1975) An operation for partial closure of an epiphysial plate in children, and its experimental basis. J Bone Joint Surg 57B:325-330. 101. Langenskiold A (1989) Tibia vara: A critical review. Clin Orthop Rel Res 246:195-207. 102. Langenskiold F (1929) Demonstration eines mit genu-varumbildung einhergehenden dunklen leidens in der oberen tibiaepiphyse, sowie uber die technik der bogenformigen osteotomie. Acta Chir Scand 64:193-197. 103. Laurence M (1967) Genu recurvatum congenitum. J Bone Joint Surg 49B:121-134. 104. Lazerte GD, Rapp IH (1958) Pathogenesis of OsgoodSchlatter's disease. Am J Path 34:803-815. 105. Leveuf J, Pais C (1946) Les dislocations congenitales du genou (genu recurvatum, subluxation, luxation). Rev Orthop 32:313-350. 106. Levine AM, Drennan JC (1982) Physiological bowing and tibia vara: The metaphyseal-diaphyseal angle in the measurement of bowleg deformities. J Bone Joint Surg 64A:1158-1163. 107. Lewis OJ (1958) The tubercle of the tibia. J Anat 92:587-593. 108. Liebman C, Iseman RM (1940) Osteochondritis dissecans. Am J Roent 43:865-870. 109. Linden B (1977) Osteochondritis dissecans of the femoral condyles: A long-term follow-up study. J Bone Joint Surg 59A:769-776. 110. Linden B, Telhag H (1977) Osteochondritis dissecans: A histologic and autoradiographic study in man. Acta Orthop Scand 48:682-686. 111. Loder RT, Johnston CE, II (1987) Infantile tibia vara. J Pediatr Orthop 7:639-646. 112. Loder RT, Schaffer JJ, Bardenstein MB (1991) Late-onset tibia vara. J Pediatr Orthop 11:162-167. 113. Lofgren L (1954) Spontaneous healing of osteochondritis dissecans in children and adolescents. Acta Chir Scand 106: 460-478. 114. Ludloff K (1903) Uber wachstum und architektur der unteren femurepiphyse und oberen tibiaepiphyse. Beitr Klin Chir 38: 64-75. 115. Ludloff K (1908) Zur frage der osteochondritis dissecans am knie. Arch Klin Chir 87:552-570. 116. Lynch MC, Walsh HPJ (1991) Case report: Tibia recurvatum as a complication of Osgood-Schlatter's disease: A report of two cases. J Pediatr Orthop 11:543-544. 117. McGill JJ, Demos TC, Lomasney LM (1995) Osteochondritis dissecans: Imaging modalities. Orthopedics 18:1180-1186. 118. Medbo I (1964) Tibia vara (Osteochondrosis deformans tibiae or Blount's disease): Treatment and follow-up examination. Acta Orthop Scand 34:323-336. 119. Mesgarzadeh M, Sapege AA, Bonakdarpour A, Revesz G, Moyer RA, Maurer AH, Alburger PD (1987) Osteochondritis dissecans: Analysis of mechanical stability with radiography, scintigraphy, and MR imaging. Radiology 165:775-780. 120. Middleton DS (1935) The pathology of congenital genu recurvatum. Br J Surg 22:696-702.

References 121. Milgram JW 91978) Radiological and pathological manifestations of osteochondritis dissecans of the distal femur. Radiology 126:305-311. 122. Mital MA, Matza RA, Cohen J (1980) The so-called unresolved Osgood-Schlatter lesion: A concept based on fifteen surgically treated lesions. J Bone Joint Surg 62A:732-739. 123. Mubarak SJ, Carroll NC (1979) Familial osteochondritis dissecans of the knee. Clin Orthop Rel Res 140:131-136. 124. Mubarak SJ, Carroll NC (1981) Juvenile osteochondritis dissecans of the knee: Etiology. Clin Orthop Rel Res 157:200-211. 125. Nagura S (1960) The so-called osteochondritis dissecans of Konig. Clin Orthop Rel Res 18:100-121. 126. Niebauer JS, King DE (1960) Congenital dislocation of the knee. J Bone Joint Surg 42A:207-225. 127. Nilsonne H (1929) Genu varum mit eigentumlichen epiphysenwer-anderungen transfer. Acta Chir Scand 64:187-192. 128. Ogden JA, Hempton RF, Southwick WO (1975) Development of the tibial tuberosity. Anat Rec 182:431-446. 129. Ogden JA, Southwick WO (1976) Osgood-Schlatter's disease and tibial tuberosity development. Clin Orthop Rel Res 116:180-189. 130. Ogden JA, Ogden DA, Pugh L, Raney EM, Guidera KJ (1995) Tibia valga after proximal metaphyseal fractures in childhood: A normal biologic response. J Pediatr Orthop 15:489-494. 131. Osgood RB (1903) Lesions of the tibial tubercle occurring during adolescence. Boston Med Surg J 148:114-117. 132. Osorio F, Costa EB (1985) La desepiphysiodese associee a l'osteotomie tibiale dans le traitement de la maladie de Blount: A propos de 2 observations. Rev Chir Orthop 71: 167-171. 133. Oyemade GAA (1981) The correction of primary knee deformities in children. Internat Orthop 5:241-245. 134. Paget J (1856) Article I: On the production of some of the loose bodies in joints. St. Bartholomew's Hospital Reports 6: 1-4. 135. Parsch K, Schulz R (1994) Ultrasonography in congenital dislocation of the knee. J Pediatr Orthop [B] 3:76-81. 136. Phemister DB (1924) The causes of and changes in loose bodies arising from the articular surface of the joint. J Bone Joint Surg 6:278-315. 137. Pitzen P, Marquardt W (1939) O-beinbildung durch umschriebene epiphysenwachstumsstorung. Zeit f Orthop 69: 174-186. 138. Potel G (1897) Etude sur les malformations congenitales du genou. Lille, L Danel. 139. Price CT, Scott DS, Greenberg DA (1995) Dynamic axial external fixation in the surgical treatment of tibia vara. J Pediatr Orthop 15:236-243. 140. Riedel (1890) Einige gelenkpraparate: Osteochondritis dissecans. Verhandl Deutsch Gesellsch Chir 19:399-417. 141. Robert M, Khouri N, Carlioz H, Alain JL (1987) Fractures of the proximal tibial metaphysis in children: Review of a series of 25 cases. J Pediatr Orthop 7:444-449. 142. Roy L, Chaise F (1979) Maladie de Blount: Revue de huit cas. Rev Chir Orthop 65:187-190. 143. Salenius P, Vankka E (1975) The development of the tibiofemoral angle in children. J Bone Joint Surg 57A:259-261. 144. Sasaki T, Yagi T, Monji J, Yasuda K, Kanno Y (1986) Transepiphyseal plate osteotomy for severe tibia vara in children: Follow-up study of four cases. J Pediatr Orthop 6:61-65.

515

145. Scheller S (1960) Roentgenographic studies on epiphysial growth and ossification in the knee. Acta Radiol (Stock) Supp 195:1-303. 146. Schlatter C (1903) Verletzungen des schnabelformigen fortsatzes der oberen tibiaepiphyse. Beitr Klin Chir 38:874-887. 147. Schmieden (1900) Ein beitrag zur lehre von den gelenkmausen. Arch Klin Chir 620:532. 148. Schoenecker PL, Meade WC, Pierron RL, Sheridan JJ, Capelli AM (1985) Blount's disease: A retrospective review and recommendations for treatment. J Pediatr Orthop 5:181-186. 149. Sen RK, Sharme LR, Thakur SR, Lakhansal VP (1989) Patellar angle in Osgood-Schlatter disease. Acta Orthop Scand 60: 26-27. 150. Shattock SG (1891) Genu recurvatum in a fetus at term. Trans Pathol Soc Lond 42:280-292. 151. Siffert RS, Katz JF (1970) The intra-articular deformity in osteochondrosis deformans tibiae. J Bone Joint Surg 52A:800-804. 152. Siffert RS (1982) Intraepiphyseal osteotomy for progressive tibia vara: Case report and rationale of management. J Pediatr Orthop 2:81-85. 153. Simon L (1948) Tibia vara epiphysarea. Paediatr Danub 6: 93-97. 154. Sloane D, Sloane MF, Gold AM (1936) Dyschondroplastic bowlegs. J Bone Joint Surg 18:183-187. 155. Smillie IS (1957) Treatment of osteochondritis dissecans. J Bone Joint Surg 39B:248-260. 156. Smith CF (1982) Tibia vara (Blount's disease). J Bone Joint Surg 64A:630-632. 157. Sontag LW, Pyle SI (1941) Variations in the calcification pattern in epiphyses: Their nature and significance. Am J Roent 45:50-54. 158. Stirling RI (1952) Complications of Osgood-Schlatter's disease. J Bone Joint Surg 34B:149-150. 159. Storen H (1970) Operative elevation of the medial tibial joint surface in Blount's diseasemOne case observed for 18 years after operation. Acta Orthop Scand 40:788-796. 160. Stricker SJ, Edwards PM, Tidwell MA (1994) Langenskiold classification of tibia vara: An assessment of interobserver variability. J Pediatr Orthop 14:152-155. 161. Teale TP (1856) Case of detached piece of articular cartilage: Existing as loose substance in the knee joint. Med-Chir Trans 39:31-33. 162. Thompson GH, Carter JR, Smith CW (1984) Late-onset tibia vara: A comparative analysis. J Pediatr Orthop 4:185-194. 163. Thompson GH, Carter JR (1990) Late-onset tibia vara (Blount's disease): Current concepts. Clin Orthop Rel Res 225:24-35. 164. Thomson JEM (1956) Operative treatment of osteochondritis of the tibial tubercle. J Bone Joint Surg 38A:142-148. 165. Twyman RS, Desai K, Aichroth PM (1991) Osteochondritis dissecans of the knee: A long-term study. J Bone Joint Surg 73B:461-464. 166. Uhry E (1944) Osgood-Schlatter disease. Arch Surg 48: 406-414. 167. Van Demark RE (1952) Osteochondritis dissecans with spontaneous healing. J Bone Joint Surg 34A:143-148. 168. Von Dittrich K (1925) Uber osteochondrolysis traumatice (osteochondritis dissecans gen Konig). Eine klinische u. histologische studie. Virch Arch Path Anat 258:795-819. 169. von Lutterotti M (1948) Beitrag zur genese der schlatterschen krankheit. Zeit f Orthop 77:160-175.

516

CHAPTER 6 9

Epiphyseal Disorders of the

Knee

170. Wagoner G, Cohn BNE (1931) Osteochondritis dissecans: A resume of the theories of etiology and the consideration of heredity as an etiologic factor. Arch Surg 23:1-25. 171. Wenger DR, Mickelson M, Maynard JA (1984) The evolution and histopathology of adolescent tibia vara. J Pediatr Orthop 4:78-88. 172. White J (1957) Osteochondritis dissecans in association with dwarfism. J Bone Joint Surg 39B:261-267. 173. Wiberg (1943) Spontaneous healing of osteochondritis dissecans in the knee joint. Acta Orthop Scand 14:270-277. 174. Willner P (1969) Osgood-Schlatter's disease: Etiology and treatment. Clin Orthop Rel Res 62:178-179. 175. Wolbach SB, Allison N (1928) Osteochondritis dissecans. Arch Surg 16:1176-1186. 176. Woolfrey BF, Chandler EF (1960) Manifestations of OsgoodSchlatter's disease in late teenage and early adulthood. J Bone Joint Surg 42A:327-369.

177. Yashar A, Loder RT, Hensinger RN (1995) Determination of skeletal age in children with Osgood-Schlatter disease by using radiographs of the knee. J Pediatr Orthop 15:298-301. 178. Zayer M (1973) Natural History of Osteochondrosis Tibial (Mb. Blount). Lund, Sweden: Lagerblads. 179. Zayer M (1980) Osteoarthritis following Blount's disease. Intemat Orthop 4:63-66. 180. Zimbler S, Merkow S (1984) Genu recurvatum: A possible complication after Osgood-Schlatter disease: Case report. J Bone Joint Surg 66A:1129-1130. 181. Zionts LE, Harcke HT, Brooks KM, MacEwen GD (1987) Case report: Posttraumatic tibia valga: A case demonstrating asymmetric activity at the proximal growth plate on technetium bone scan. J Pediatr Orthop 7:458-462.

CHAPTER

7

Epiphyseai Growth Plate Fract u re-S epa ration s I.

II.

V.

Introduction: Pre-radiographic Era, Pathoanatomic Approaches, and Pathophysiologic Approaches Clinical and Experimental Investigations of Growth Plate Fracture-Separations in the Pre-radiographic Era

VI. VII.

III.

Clinical Approaches to Growth Plate Fracture-Separations in the Radiographic Era IV. Pathophysiologic Approaches to Growth Plate FractureSeparation

VIII.

General Clinical Profile of Growth Plate Fracture-Separations Clinical Features of Acute Epiphyseal Fracture-Separations Traumatic Damage to Growth Plates by Pathologic, Chronic Repetitive, and Indirect Effects Management of Negative Sequelae of Growth Plate Fracture-Separations

approaches to growth plate fracture-separations. We will also present the pathophysiologic approach to growth plate fracture-separations and review its development as an extension of the work from the preradiographic and pathoanatomic eras. Projected therapeutic benefits of the pathophysiologic approach involve earlier detection and more biological treatment of negative growth plate sequelae.

I. I N T R O D U C T I O N : P R E - R A D I O G R A P H I C ERA, PATHOANATOMIC APPROACHES, AND PATHOPHYSIOLOGIC APPROACHES Epiphyseal growth plate fracture-separations have been recognized and studied for over 200 years. Early comments on their occurrence were made in the 1500s and 1600s, but it was not until the early 1700s that meaningful work on the subject appeared. They were considered first in specific detail by Reichel in 1759 in a thesis published in Leipzig in which he distinguished between traumatic and spontaneous (pathologic) separations and indicated that following separation the growth plate remained adherent to the epiphysis (268) (Fig. 1). Extensive numbers of articles on these injuries were published in the nineteenth century in the preradiographic era of growth plate fracture assessments, many of which remain valuable today. With the discovery and widespread clinical use of radiographs 100 years ago, each fracture could be assessed easily prior to, during, and after treatment. From the 1930s to the 1990s, several classifications of epiphyseal fracture-separations were proposed. These are referred to as pathoanatomic because they are based on the pattem of fracture as determined on plain radiographs. One of the therapeutic benefits of the pathoanatomic approach was the recognition that anatomic open reduction and rigid internal fixation were helpful in minimizing growth problems in certain fracture-separations. The current focus in relation to growth plate injuries is moving toward more dynamic assessment of these injuries and utilization of pathophysiologic approaches at the time of initial assessment, during the healing phase and in the months immediately following healing. In this chapter we outline the varying

II. C L I N I C A L A N D E X P E R I M E N T A L INVESTIGATIONS OF GROWTH PLATE F R A C T U R E - S E P A R A T I O N S IN T H E PRE-RADIOGRAPHIC ERA

A. Early Clinical Descriptions Case descriptions establishing the reality of the entity were published in the early 1800s. Prior to this time and for several decades afterward, many surgeons did not recognize the existence of epiphyseal fracture-separations or considered them to be very rare and to occur only in the very young. In 1818 Laurent (188) referred to three cases of Dr. Champion of Bar-le-Duc, France: a proximal humeral fracture-separation in an 11-year-old child confirmed at postmortem examination, a distal humeral injury in a 13-year-old child, and a distal tibial fracture-separation caused by a complicated delivery in a newborn. Laurent also discussed an example of multiple epiphyseal separations noted at autopsy in a 7-month-old fetus with hydrocephalus. Roux, in his 1822 thesis "Dissertation on Separation of the Epiphysis," reported epiphyseal fracture-separations of the distal femur in one patient and multiple separations of the distal radius, distal ulna, and three proximal phalanges studied by gross inspection following 519

520

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations atomic specimen obtained at the time of amputation (107). In 1836 Goyrand discussed a distal radial injury in an 11year-old child whom he saw shortly after injury and prior to the onset of severe swelling (128). Clinical examination indicated displacement near the wrist joint, but the normal relationship of the radial styloid to the wrist ruled out dislocation. Other clinical features such as the ease of reduction, a sensation on reduction similar to cartilage surfaces of a joint relocating rather than the rubbing of fracture fragments, and the occurrence of healing after only 20 days of immobilization led him to distinguish a distal radial epiphyseal separation from a metaphyseal fracture.

B. Important Studies Establishing the Validity of the Entity of Epiphyseal Growth Plate Fracture-Separations Two extensive multipart papers on epiphyseal fractures by Rognetta (275) and Gueretin (130) appeared in the French literature in the 1830s. They will be summarized in some detail as they demonstrate not only the then current conceptions of the entity but also the beginning of the development of a coherent body of investigational and clinical information. 1. R O G N E T T A

F I G U R E 1 Illustration from Reichel's thesis (268) published in 1759 shows a characteristic type II fracture-separation of the distal femoral epiphysis. The epiphysis, physis, and metaphyseal fragment are shown at left. Reichel recognized that the physis remained with the epiphyseal segment after most fractures and that adjacent segments of metaphyseal bone frequently were separated as well.

amputation (276). Roux, as quoted by Poland, insisted upon the following: "1) the danger of ignoring epiphyseal separation when its reduction ought to be possible; 2) the progress of the phenomena which result from an error of judgment; 3) the possibility of separation of the epiphyses occurring up to the 18th year; and 4) the possibility of such separation taking place on the largest cylindrical bones as well as on the medium sized and smallest" (258). In a later article he added additional cases (some of questionable validity) and noted that in his thesis he had discussed the differentiation between internal (pathologic) and external (traumatic) causes of epiphyseal separations; the differential diagnosis between epiphyseal separations, joint dislocations, and fractures, and the relative brevity of treatment time needed for traumatic separations (276). Fontenelle reported on a distal femoral epiphyseal fracture-separation in an 11-year-old child based on the an-

Rognetta (1834) wrote a four-part article in which he stressed the need to recognize epiphyseal fracture-separations as a definite entity and provided an excellent summary of the earliest examples of the recognition of epiphyseal injuries from the sixteenth, seventeenth, and eighteenth centuries (275). Many, including Pare, Severin, Eysson, Platner, A1binus, Hailer, Kerkeringius, Spingeler, Coyter, Reichel, van Swieten, Morgagni, Weiss, Duverney, Petit, Verduc, Palletta, Ruysch, Bertrand, and Monteggia, had commented on the entity and presented usually isolated case reports of traumatic or spontaneous (pathologic) epiphyseal separations. Ruysch commented in 1713, according to Rognetta, that it was possible to separate the epiphysis in the living as soon as the perichondrium surrounding the growth cartilage was torn. Rognetta pointed out the importance and high degree of organization of the growth cartilage and stressed the observations of Howship, which were aided by crude early forms of microscopic assessment. These observations included the following: (1) the growth cartilage formed an integral part of the bone itself; (2) the cartilage was continuous both with the epiphysis and with the body of the entire bone rather than merely being placed between them; (3) the growth cartilage had within its interior and adjacent to it a specific organization with arteries, veins, circulatory canals, and a mucus substance; (4) the epiphyseal bed, far from just being resorbed with time, formed an outline at its base upon which bone was deposited; and (5) the perichondrium that surrounded the growth cartilage exteriorly was closely adherent to it. Much of the initial article was devoted to a description of the differences between the various epiphyses

SECTION I! ~ Clinical and Experimental Investigations in the Pre-radiographic Era in terms of their structure, ossification, and fusion times to the adjacent diaphyses. He noted that the epiphyseal cartilage was strongly adherent to the perichondrium that covered it and that this membrane actually formed the principal bond between the epiphysis and the diaphysis. The joint capsules and ligaments were stronger than the periosteum in children, which explained why traumatic dislocations in the young were very difficult if not impossible to produce experimentally and why separation of the epiphysis invariably resulted. In a study of two fetuses at term he was not able to dislocate any of the joints because the epiphyses separated first. In the clinical state, cases of traumatic dislocation rarely were described before the ages of 5 or 6 years, whereas many examples of epiphyseal detachment were reported. He also pointed out the work of Wilson, who in his book On the Bones and Joints, published in London in 1820, reported that, when longitudinal traction alone was applied, it took a force of 550 lb when the periosteum was intact to detach the distal femoral epiphysis in a childhood specimen, whereas when the periosteum was removed a force of only 119 lb was required (331). Rognetta felt that far less force was needed to produce epiphyseal separation in the newborn. The point was made, however, that when the surrounding periosteal-perichondrial sheath was removed the only intrinsic support that the growth plate provided was from the mammillary processes from the cartilage plate, which interdigitated for a short distance into the diaphysis. Rognetta, quoting Portal, observed that ossification of the distal femoral epiphysis always began during the final 15 days of fetal life, an observation that is valuable in determining the age of the newborn concerning whether it was fully at term. He reviewed the importance of the growth cartilages in long bone growth on the basis of the work of Hunter, Duhamel, and others and the differing mechanisms of bone formation from the periosteum and via the epiphyseal cartilage plate and the bony nucleus of the epiphyseal cartilage (our secondary ossification center). Rognetta pointed out how epiphyseal separation of the proximal femur was produced by trauma from the newborn period and for several years into childhood (and even beyond) rather than traumatic dislocation or femoral neck fracture. He stressed that there was a clear pathological difference in infants of very young age between an epiphyseal separation and fractures adjacent to the growth plate. In the epiphyseal separation there was a clean division of the cartilage line without marrow involvement, whereas many childhood fractures up to the age of 3 or 4 years involved a crushing of one side of the metaphyseal-diaphyseal bone with partial tearing of the periosteum, partial cracking of the cortical bone, and damage to the marrow contents. Separation of the epiphysis and bone fractures in infants formed "two distinct disorders." Epiphyseal separations were recognized as complications of difficult deliveries and frequently were detected at postmortem examination. Rognetta referred to several clinical mechanisms by which proximal humeral epiphyseal separa-

521

tions occurred both in the newborn period and through the growing years. The frequency of distal radial epiphyseal separations was recognized as was the fact that proximal radial or ulnar separations were rare. Two distal radial separations were described: one in a 15-year-old male seen at autopsy and another in a 10-year-old girl based on clinical assessment. Fracture-separations also were identified at the distal femur, proximal tibia, and distal tibia. Sibley of London had even diagnosed a separation of the tibial tubercle following strong muscular action (288). Valgus deformation was noted to follow fracture-separations of the distal tibial epiphysis. 2. GUERETIN Gueretin (1837) wrote a rigorous three-part comprehensive review of growth plate fractures (130). He detailed the age at growth plate fusion, the relative growth rates of the various ends of the long bones, his human cadaveric experiments on reproducing dislocations, growth plate separations, and metaphyseal-diaphyseal fractures, the distinction between spontaneous (pathologic) and traumatic separations, a discussion of injuries at each major long bone epiphysis, treatments, and results including complications. The growth in length of bones at their diaphyseal ends through the functions of the cartilage growth plates was well-recognized, as was the fact that if healing following a growth plate fracture was associated with a bony callus ("un cal osseux") the bone would remain much shorter than that on the opposite side, which was an early reference to transphyseal bone bridge formation. The surrounding periosteum and perichondrium were recognized as having a major supporting role for the epiphysis. Experimental studies showed that in the newborn epiphyseal separations invariably occurred rather than joint dislocations and that proximal humeral separations readily were produced. In the newborn, the entire epiphysis separated from the diaphysis. Separation in the newborn and for several months thereafter tended to be completely through the cartilage plate with no bone attached. After 2-3 years of age the structure of the epiphyses changed somewhat and the growth plate fractures were less uniform in appearance; some occurred slightly away from the epiphyseal plate, whereas those involving the epiphyses often dragged with them fragments of "the extremity of the diaphysis." (The term metaphysis seldom was used in the nineteenth and early twentieth century orthopedic and anatomic literature.) Different patterns were also seen after 10 years of age not associated with pure structural separations but rather with less regular snapping or bursting patterns. Distinction was made between external or traumatic separations and internal or spontaneous separations caused by such pathological conditions as tickets, infection, scurvy, congenital syphilis, and tuberculosis. A review was made of epiphyseal separations of each long bone. The most common epiphyseal fractures were those of the proximal humerus especially in association with difficult births. At the distal humerus, complete transverse separations were noted as were separations that passed

CHAPTER 7 ~

522

Epiphyseal Growth Plate Fracture-Separations

between the separate ossification centers to involve internal or external condylar fractures. It was recognized early that the distal radial epiphysis very commonly was separated in childhood but that the proximal radial separation was rare. Proximal femoral separations were difficult to produce in experimental postmortem studies and only rarely were verified in clinical situations. The distal femur and proximal tibia were known sites of epiphyseal separation. The distal tibia and fibula readily were separated experimentally in the newborn but infrequently recognized clinically. Gueretin made the interesting observation that, in his experimental approaches to the distal tibia, he sometimes noted the periosteum to be intact whereas the epiphysis itself separated. Others subsequently noted this sequence of events, with cartilage separation occurring first and displacement occurring only with secondary rupture of the periosteum. The end results of epiphyseal fractures were characterized by uneventful healing, healing with deformity and shortening, nonunion, or amputation or death. Malunions commonly were reported, and shortening of the involved limb often was apparent. There was recognition that growth deformities could occur even after healing of the epiphyseal fracture. Gueretin reported his interpretation of a case shown by Reichel that epiphyseal healing after separation occurred "by means of a bony bridge (cal)." He went on to indicate that "in the humerus illustrated by Reichel in addition to the shortening caused by the collapse of the humeral head and its faulty position one still could establish or ascertain a perceptible difference in length for the rest of the bone." He thought that the bone had lost its growth in length because it no longer had at one of its extremities the epiphyseal cartilage plate where this growth occurred. These cases proved the need to observe shortening in any limb that had suffered a growth plate separation previously whether or not union had occurred. Treatment of spontaneous (pathologic) separations was directed primarily at the underlying cause. In the straightforward traumatic separations, treatment involved (1) reduction, (2) maintenance of reduction for 8-30 days in the group from 0 to 4 years of age and for 15 days to 3 months in those older than 4 years of age, and (3) prevention of complications of an infectious or vascular nature.

3. SMITH Smith presented a very clear description of epiphyseal separation of the distal end of the humerus, demonstrating not only its existence but the way of differentiating it from what are now recognized as supracondylar fractures of the distal humerus and from elbow dislocations (295). He pointed out the basic anatomical fact "that the lower epiphysis of the humerus does not include the condyles which belong entirely to the shaft of the bone." Canton illustrated two cases of fracture-separation of the distal femoral epiphysis, both of which came to amputation (54). The first example is what we would identify today as a

F I G U R E 2 Illustration of a distal femoral epiphyseal fracture-separation from an article by Canton (54) published in 1861. The fracture line passed along the physis of the distal femur at left and then split into the metaphysis, demonstrating the type II pattern. Pathoanatomic studies of gross specimens during that era also allowed for demonstration of multiple and divergent fracture patterns subsequently not appreciated on plain radiographs. In this specimen, the metaphyseal segment C is shown clearly and was described as a separate fragment.

type II fracture-separation and the second appears as a type I (Fig. 2).

C. Relatively Slow Acceptance by Many of the Existence of Epiphyseal Fractures For the next several decades debate continued over the existence of epiphyseal fracture-separations, with acceptance slowed by failure of many of the leading surgeons to recognize them or to consider them only as rare and not particularly important. Salmon performed extensive experimental work, inducing fracture-separations in human bones, and was credited by Poland as being the first to prove that tearing of the periosteum around the growth plates was accompanied by separation (258). Much awareness of the occurrence of growth plate fracture-separations came from studies on stillborn or neonatal infants with extensive limb trauma accompanying difficult deliveries. Poland documented well the fact that widespread acceptance and accurate description of these fractures dated only from the 1870s. Many surgeons and investigators focused on the fracture fragments associated with these injuries and the fact that the transverse lines of separation were often in the spongieux layer (or within the outermost parts of the metaphysis "below" the hypertrophic

SECTION II ~ Clinical and Experimental Investigations in the Pre-radiographic Era zone in our terminology). Thus, they felt comfortable considering these injuries as fractures because pure epiphyseal fractures, with the line of injury totally within the cartilage, were felt to be rare or nonexistent. The treatments were also the same. Although there was still widespread uncertainty within the profession as to the nature and extent of the injuries, the overall tendency to accept their existence increased. Poland quotes Holmes as saying in 1869, "the conclusion to which my experiments of this injury would lead me is that fracture occurs not very rarely at or in the immediate neighborhood of the epiphyseal line; that the line of fracture coincides in these cases partially with that of the epiphyseal cartilage but seldom completely, that the general symptoms are therefore the same as those of fracture, while the special symptoms must be sought from the anatomy of each joint." In a special discussion group of the Surgical Society of Paris in 1865, many prominent surgeons of the day spoke about the entity (204). Marjolin, in a review of 600 fractures in children from his hospital over a 7-year span, reported that no diagnoses of separation of the epiphyses had been made. He commented that others also considered complete epiphyseal separations to be rare. Still others focused on the associated fractures of the diaphysis that adhered to the epiphyseal fragment and thus led to their interpretation as fractures rather than growth plate injuries. Richet then described a case of a separation of the distal radial epiphysis in a 15-year-old male that was irreducible by closed manipulation and was reduced following resection of the distal radial diaphyseal (metaphyseal) segment. The diaphyseal part of the radius jutting out had "a smooth surface with rounded processes (mammillary projections) that obviously represented the diaphyseal portion of the radius separated from the epiphysis." Subsequent examination of the resected segment clearly established the reality that epiphyseal separation and not a bone fracture had occurred. Lefort described another resected fragment of the proximal tibia showing partial epiphyseal separation with passage of the fracture line into the adjacent diaphyseal bone. Broca mentioned elsewhere that epiphyseal separation could be produced in the postmortem state, which led to his belief that it was also common clinically. Smith further expanded upon the reality of the entity noting the existence of epiphyseal fracture-separations, which he referred to as "epiphysary disjunction," at either end of the humerus, lower epiphysis of the tibia, and lower epiphysis of the radius. He continued to note in 1867 that epiphyseal fracture-separations were "either but slightly noticed or all together omitted in our systematic works on surgery" and that there continued to be considerable ignorance about them by the profession (296). He felt that they were a clear class of injuries that could be readily diagnosed and should be considered in any child with injury around a joint. As well as discussing the distal humeral epiphyseal fracture-separation, he pointed to the occurrence of such injuries at the distal tibia in which the internal malleolus preserved its relation-

523

ship to the foot and ankle joint but not to the leg or proximal tibia above. The proximal humerus was also the site of growth plate fracture-separation. Smith again stressed knowledge of the anatomy, in that the epiphysis comprised not only the head of the bone but also both of the tuberosities. He felt that fracture at this region was quite common. Separation of the lower epiphysis of the radius was also a frequent occurrence. He felt that it could be distinguished quite readily from a fracture of the bones of both the distal radius and the ulna, in the sense that the radial epiphyseal radiation usually was isolated and displaced in a dorsal direction with no tendency to supination.

D. Pathoanatomic Studies of Epiphyseal Growth Plate Fracture-Separations The morbidity and mortality rate from trauma in the nineteenth century was so high that there were many pathoanatomic studies of human amputation or autopsy specimens. The authors were able to assess bone, cartilage, and soft tissue damage directly, whereas now we rely almost exclusively on radiographic approaches because the large majority of growth plate injuries are treated by closed means. Even when open reduction and internal fixation are used, exposure of the physis is kept to a minimum. Their reports can be studied with interest today, especially in view of the crucial importance to the eventual outcome of the path of fracture lines within the radiolucent growth plates. These are undetectable by plain radiographs but help to determine in a major way which fractures will heal uneventfully and which will predispose are to focal transphyseal bone bridge deformity.

1. BROCA Broca had further described the structure of the growth region, demonstrating that between the cartilage of the epiphysis and the spongy tissue of the diaphysis there were two layers of variable thickness: one of which was no longer plain cartilage and the other not yet fully spongy tissue. The layers were referred to as the chondroid and spongioid, respectively (37). The thickness of these two layers varied according to age and the particular bone. Broca, as well as providing macroscopic terminology for the epiphyseal region, also pointed out that the epiphyseal growth plate cartilage remained with the epiphysis after fracture.

2. FOUCHER Foucher focused more attention on epiphyseal growth plate fracture-separations in 1860 (111, 112). Foucher felt that he had established a direct relation between the thickness of the chondroid layer and the ease of epiphyseal separation and between the appearance of bony tissue in the epiphysis and the type of separation. He felt that with trauma there was a true interruption in continuity through the plate and not a simple separation of adjacent or contiguous parts. The layer of union between the diaphysis and epiphysis was

524

CHAPTER 7 9 Epiphyseal Growth Mare Fracture-Separations

TABLE I Foucher 1860 Classification- First Detailed Pathoanatomic Classification of These Injuries (Macroscopic Examination )

Type I Type II Type III

Divulsion epiphysaire Pure epiphyseal separation within couche chondroide Fracture epiphysaire Fracture between spongioide and spongieux layers Fracture preepiphysaire Fracture within metaphyseal bone

not level but often alternatively concave and convex, especially in the older patient. At birth each of the growth plates was relatively level with the deviations more apparent after the age of 1-2 years. The shape of the growth plate varied from bone to bone and had no relationship to that of the adjacent articular surface. Foucher presented the first detailed pathoanatomic classification of these injuries in 1860 based on extensive examination of specimens at the macroscopic level. Between the cartilage of the epiphysis and the bony tissue of the diaphysis, which was referred to as the "tissue spongieux," there were two distinct layers of the growth plate, as defined initially by Broca, referred to as the couche chondroid and the couche spongioid. These can be seen by examining a growth plate macroscopically (or with a magnifying glass), with the couche chondroid essentially referring to the germinal, proliferating, and upper hypertrophic layers and the couche spongioid referring to the lower part of the hypertrophic layer and the zone of provisional calcification, including the upper reaches of bone deposition on the calcified cartilage cores. Foucher defined three types of fracture depending on the transverse level of injury (Table I). Type I was a pure epiphyseal separation (divulsion or decollement epiphysaire) in which the transverse line of injury was completely within the growth plate cartilage at the point of union of the chondroid and spongioid layers. Type II was a fracture-separation (fracture epiphysaire) with some fragments of bone present on the epiphyseal side separation surface and separation through the point of union of the spongioid and spongieux layers. Type III was a fractureseparation within metaphyseal bone or the tissue spongieux layer just below the growth cartilage, referred to as a juxtaepiphyseal fracture (fracture preepiphysaire). He never noted transverse separation solely within the cartilage tissue (although subsequently others have noted such pathways). The presence of a secondary fracture was not the important consideration, it was the epiphyseal separation that was paramount. He indicated that the line of separation could be variable in any single injury involving each of the chondroid, spongioid, and spongieux layers partially because the growth plate often was not a straight line but increasingly concave with age. This observation concerning a wandering line of separation was also made by Bret and Curtillet (34) and by Poland (258). Most separations from 1 month to 1 year of

age were type I, those between 1 and 4-5 years were type II, and those from 5 to 10 plus years were type III. Foucher pointed out the importance of associated lesions with epiphyseal separation, namely, the separation of the periosteum. It was widely recognized that the periosteum was only minimally adherent to the underlying bone in children. Epiphyseal separations caused by direct traction were not accompanied by a separation of periosteum, but tearing of the periosteum was an invariable occurrence with displaced epiphyseal fracture-separations. The age of the subject was not the only factor that influenced the frequency and ease of epiphyseal separation; there were diverse results in different epiphyses. He then referred to the mechanisms that produced these lesions, a feature that was felt to be extremely important. He had never been able to displace an epiphysis, regardless of age, by a force directly applied on the epiphysis. Traction along the axis of the limb sometimes produced epiphyseal separation in the very young, but he felt that massive force greater than what would be seen in the human was needed. Very rarely did direct traction alone prove to be sufficient in the living to produce an epiphyseal separation. It was with the flexion, abduction, rotation, and extension movements that the displacements occurred. Extension at the knee was able to produce epiphyseal separation usually of the distal femur but occasionally of the proximal tibia. At the elbow this produced separation of the lower humeral epiphysis and very rarely that of the olecranon. At the proximal humerus and proximal femur yet another mechanism was involved, namely, with abduction forces and with rotation. He summarized by indicating that it was torsion that favored epiphyseal separation in each bone. The abnormal positions of extension and abduction combined always with torsion or rotation predisposed one to epiphyseal separation. The effect of severe muscular contraction on the separation also was felt to be important. Foucher summarized with seven conclusions: (1) The epiphysis could be separated from the diaphysis either by trauma or spontaneously in association with other disorders. In the latter case separation was only an epiphenomenon following after other affections in particular those of the periosteum, which was swollen and weakened, for example, by sub-periosteal abscesses. (2) The traumatic epiphyseal separation was easiest to produce when the infant was youngest. (3) The level of epiphyseal and diaphyseal separation varied according to age and cause with three specific levels of separation noted: (a) the level where the chondroid layer met the spongioid layer; (b) the level where the spongioid layer met the spongy (spongieux) layer; and (c) within the spongy bone tissue itself. (4) Regardless of the level of lesion, they involve complete loss of cartilage continuity, with associated fractures dependent on the mechanisms of production. (5) Exaggerated motion is the most common cause of epiphyseal separations with muscular action having secondary influence. (6) the surface of the separated fragments is alternatively convex and concave and the periosteum is largely separated

SECTION II ~ Clinical and Experimental Investicjations in the Pre-radiocjraphic Era from the diaphysis. (7) An important prognostic point is whether the separations are intra-articular or extra-articular. 3. VOGT AND BRUNS

In Germany, Vogt (319) and Bruns (41) each wrote detailed articles concerning epiphyseal separations. Bruns described a macroscopic classification similar to that of Foucher, differing mainly in his feeling that on occasion transverse separations occurred higher within the cartilage plate such that both fragments (epiphyseal and diaphyseal) clearly showed a cartilage coveting. He summarized in detail and completely referenced 81 epiphyseal separations (78 from the literature and 3 of his own) that had been authenticated by gross inspection of open wounds or by pathology examination either at autopsy or following amputation. His work was done to assess the nature of growth plate damage following these injuries. A cardinal question in terms of outcome was the specific tissue layer in which the separation occurred. Bruns distinguished between two common patterns of fracture: that within the lower part of the growth cartilage, although 5 of 81 cases were within the higher levels of cartilage, and that within the bony substance of the adjacent diaphysis. Vogt, performing animal experimentation, observed pure epiphyseal separations only in very young animals, with injuries of the growth cartilage in older animals usually being mixtures of cartilage separations and bony fractures. The developmental state of the epiphysis was felt to contribute to the pattern of fracture. In chondroepiphyses, before the appearance of the secondary ossification center, pure cartilage separation was more common, whereas in osteoepiphyses where bone had replaced cartilage the cartilage separations usually were associated with bone fractures. Experiments in goats and lambs injuring the epiphyseal growth plate and adjacent regions were done to assess where the greatest growth disturbances occurred. Vogt found that the greatest disturbance took place after injury to the epiphyseal cartilage plate. His work indicated that after sub-periosteal resection of the physeal cartilage in growing bones, regeneration of bone itself took place but growth in length ceased. Diaphyseal resection caused no growth disturbance, but removal of the cartilage between the diaphysis and the epiphysis did cause a growth disturbance. Separation of the epiphyseal cartilage from the adjacent diaphysis caused growth disturbance if there was damage to the cartilage substance. 4. OLLIER

In an effort to assess limb shortening, Oilier produced several lesions in the area of the growth plate and demonstrated that considerable growth arrest occurred not because of premature growth plate fusion but rather due to diminished growth over a prolonged period, with trouble due to diminished nutrition slowing growth not by premature ossification but rather by arrest of proliferation of the cartilage

525

cells (240). When he produced major direct growth plate injuries, however, by excising growth cartilage of either the distal femur or the proximal tibia, he noted a massive stoppage of growth that was asymmetric if he had made the excision of the growth plate cartilage focally and complete if the lesion had extended across the entire transverse diameter of the bone. He showed clear examples of growth arrest in lesions made in the 1-month-old rabbit. Destruction of the lateral region of the distal femoral growth plate led to considerable shortening and extensive valgus deformation with continued growth of the medial and central parts of the plate. Considerable shortening and varus deformity of the proximal tibia were demonstrated with incision of the medial part of the growth cartilage also in the 1-month-old rabbit. Oilier studied the influences of diaphyseal fractures and periosteal irritation on long bone growth and clearly demonstrated that both led to increased growth on the involved side. He specifically studied growth plate fracture-separations in terms of subsequent growth. His experiments illustrated that growth plate separations influenced subsequent growth in a variable fashion. There was no cessation of growth if reduction had been made immediately and inflammation avoided, but growth arrest was considerable when the reduction had not been performed. The latter point soon was recognized to be inaccurate; nonreduced fractures did not invariably proceed to growth arrests. Oilier also wrote a detailed work on traumatic lesions in the juxtaepiphyseal region of diaphyseal bone, which he referred to as l'entorse juxta-epiphysaire (juxtaepiphyseal sprains) (241). Injuries in the area of a joint in the adult led to either ligamentous tears, intra-articular fractures, or dislocations, whereas in the young child the weakest areas were either the epiphyseal growth plate or the region at the extreme end of the diaphysis immediately adjacent to the plate. It was often this latter juxtaepiphyseal region that suffered the effects of trauma. The lesions Ollier was describing generally were non-displaced and thus poorly understood clinically. He referred to a case in which a fall had caused a seemingly mild shoulder injury in a 7- or 8-year-old girl, which subsequently presented as a growth arrest with shortening of the proximal humerus by 7 cm years later. This and other clinical examples were used as evidence to confirm his previous experimental findings on the possibility of growth arrest in a clinical setting. Occult trauma in some instances might lead to subsequent growth arrest even though the initial injury appeared benign and any evidence of infection had been excluded. The juxtaepiphyseal sprain was considered to represent the first degree of what would become an epiphyseal separation if greater force were applied. They were particularly common where the bone was osteopenic with systemic involvement. Those most clinically significant occurred in the first years of life up to about 6 or 7 years of age. The bending of bone, incomplete infractions, and subperiosteal fractures were not seen at later ages. He defined the variety of these lesions on the basis of the changing

526

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

consistency of the metaphyseal tissue. An example is an abduction injury. The fracture was within the diaphysis only 810 mm from the lowest part of the growth plate. The cortex on the concave side cracks and bends first, and the adjacent spongy bone and its associated trabeculae has cracks. The marrow is disturbed, but there is no lesion of the growth cartilage and the intact periosteum hides the view of the cracked bone. A benign form of this is the metaphyseal buckle fracture. If the traumatic force continued, more pronounced squeezing of the juxtaepiphyseal cortical and trabecular bone occurred on the concave side, with separation of segments of spongy bone and perhaps separation of a part of the growth cartilage on the convex side with the periosteum remaining intact. The final stage involved further force: a trans-diaphyseal spongy bone fracture. He concluded that the juxtaepiphyseal sprain was a group of lesions produced in the juxtaepiphyseal region of the diaphysis of long bones. Poncet, from the same school as Oilier, demonstrated how damage to radial or ulnar epiphyses had a negative growth effect on the other nonaffected bone (259). 5. BIDDER Bidder produced focal growth plate defects in the rabbit by several methods, leading to a bone bridge between epiphysis and metaphysis and causing growth retardation (25). 6. NOVE-JOSSERAND

Nove-Josserand performed experiments to study the mechanisms of growth arrest histologically following small partial cartilage growth plate injuries and to disprove the then widespread belief that growth cartilage irritated by trauma underwent a generalized acceleration of its development to fuse prematurely (228). The studies were done in 2week-old and 1-month-old rabbits using aseptic techniques with no infections (which had compromised previous studies) seen. He created variably sized focal vertical lesions through the growth plate cartilage and elicited transphyseal bone bridge formation between epiphysis and diaphysis well-shown histologically (Fig. 3). In 8 of 10 instances he induced a localized transphyseal ossification between the epiphysis and the diaphysis. He concluded that it was this bony communication between epiphyseal and diaphyseal bone that led to growth arrest and deformity. The "bony wedge" between epiphysis and diaphysis established a "true synostosis fixing one to the other like a solid and inextensible nail." He also felt that the bony bridge formation could be rapid or slow. 7. BRET AND CURTILLET

Bret and Curtillet examined human specimens microscopically to assess histologic structure using the Foucher classification as a reference point (34). This represents one of the earliest microscopic studies; the work of Foucher and Broca was macroscopic even though microscopic techniques were available. (Broca did correlate his macroscopic finding

F I G U R E 3 The fact that some epiphyseal growth plate fractureseparations could be followed by physeal damage and transphyseal bone bridge formation was studied by several French investigators in the late nineteenth century. This illustration in a rabbit model from the work o f Nove-Josserand (228) shows epiphyseal bone above, a break in the physeal cartilage centrally, and passage of bone into the gap.

of normal growth plates at a microscopic level.) In newborn specimens the line of fracture was not felt to be within the chondroid layers of the growth plate because microscopically small fragments of bone were seen attached to the cartilage side, thus representing a "fracture epiphysaire" and not a true separation (divulsion epiphysaire). The separation thus had occurred in what was referred to as the lower part of the spongioid layer. In an older child 3 years of age the fracture lines were farther from the epiphyseal cartilage totally within spongy bone, and in two teenage specimens the fractures were even farther away from the growth cartilage (fracture preepiphysaire). They concluded that the small spicules of bone adherent to the epiphyseal fragments indicated that pure epiphyseal separations did not exist and that the Foucher classification at the macroscopic level was inaccurate to assessments at the microscopic level. They continued to recognize the value of a separate definition of juxtaepiphyseal fractures, however, and termed traumatic separation to exist for any fracture produced between the growth cartilage and the bulbar region of the diaphysis. This article, although interesting from the point of view of the questions being asked about the precise histologic level of injury, involved epiphyses from only four patients in whom the fractures were created manually at the time of autopsy. 8. CORNIL AND COUDRAY Cornil and Coudray performed particularly noteworthy studies in demonstrating, at the histologic level, the varying

SECTION II ~ Clinical and Experimental Investigations in the Pre-radiographic Era

527

FIGURE 4 Extensiveexperimental studiesby Cornil and Coudray(70) usingthe rabbit modelidentifiedeach of~th~l~arda0anatomic patterns of epiphyseal fracture-separation that later came to characterize classificationsthroughoutthe twentieth century. In part (A) a type IV pattern is showninvolvingpassage of the injuryfrom the articular cartilage (a) throughthe epiphyseal secondaryossification center bone and then throughthe physis and metaphysealbone. (B) Illustrationfrom Cornil and Coudrayshowinga transphysealbone bridge linkingepiphysealand metaphysealbone with the rest of the physisremainingopen. Theypointedout that frequentlythere were longitudinal tears within the fracturedphysealcartilage openingcommunicationbetween epiphysealand metaphysealbone.

types of epiphyseal fracture-separation patterns producible in immature rabbits (70, 71). The initial studies involved angular and rotational forces applied to 54-day-old rabbits immediately postsacrifice with assessment histologically of fracture patterns at the distal radius and ulna, distal femur, and distal tibia and fibula. They noted the similarity of patterns produced in humans and rabbits to the extent that they reproduced all varieties of lesions seen. Epiphyseal separations were produced along with mixed lesions involving partial epiphyseal separation and associated diaphyseal or epiphyseal fractures. The term metaphyseal was not widely used in that era, and the authors are referring to what we now call the juxtaepiphyseal or metaphyseal regions. The diaphyseal fractures were either continuous with the epiphyseal cartilage fractures (as in the Salter-Harris II terminology), at fight angles to the growth plate passing into the diaphysis (metaphysis) as longitudinal clefts, or transverse in the diaphysis seemingly separate from the physeal separation. Many more fractures then were created in the living, with sacrifice and histologic studies done after several days or weeks to determine the sequelae of injury, particularly in relation to growth problems. They defined multiple fracture patterns, including (1) pure transverse epiphyseal separations; (2) complete transverse fractures through the spongioid and diaphyseal spongy bone layers; (3) mixed lesions involving partial epiphyseal separations with deviation of the line of injury causing a diaphyseal fracture; (4) complex fractures involving oblique planes from the diaphysis

through growth plate, epiphysis, and articular cartilage; and (5) longitudinal fractures of the growth plate cartilage. The longitudinal fractures of the growth plate in particular warranted close attention because they frequently were demonstrated histologically in isolation or coexisting with each of the type 2-4 injuries described previously even without displacement (Fig. 4A). The cartilage often was broken by one or several longitudinal tears, opening a communication between epiphyseal and diaphyseal bone. The loss or separation of cartilage was filled in by solid bony spans directly continuous with bone of the epiphysis and diaphysis (Fig. 4B). Multiple histological preparations were shown demonstrating "un cal osseux" or bony callus replacing growth plate cartilage. Cornil and Coudray summarized by indicating that longitudinal fractures of growth cartilage always repaired themselves by "un cal osseux" (bony callus) quickly developed and thrown out like a bridge between the epiphysis and the diaphysis. They felt that the longitudinal fractures were frequent and that they often led to shortening of the involved bone because shortening could be seen even in those cases without displacement. The general conclusion of the work was somewhat flawed in that they felt that virtually all growth plate fractures led to growth inhibition. Although this is inaccurate it should not cloud the central importance of the concept of the longitudinal growth plate cartilage cleft allowing epiphyseal and metaphyseal vessel and bone communication followed by bone bridg~formation. The detailed

528

CHAPTER 7

9

Epiphyseal Growth Plate Fracture-Separations

drawings of the histologic sections and the descriptions in the text illustrate other important features of growth plate fractures: (1) the line of injury within the physis tends to wander through the various physeal layers; (2) the cartilage can be fragmented at focal points into several pieces; and (3) the adjacent bone on the diaphyseal and occasionally epiphyseal side often is fractured in various planes such that mixed or complex lesions are prominent. Studies also demonstrated displacement of cartilage plate fragments into the diaphysis and repair of segments of damaged physeal cartilage by fibrous tissue rather than bone. 9. THE LEVEL OF TRANSVERSE FRACTURE The level of fracture came to dominate the discussion of these injuries. Most studies of this era, in which many clinical samples were studied, generally considered the transverse injury to be a fracture of the outermost reaches of the diaphysis (metaphysis) because a thin layer of ossifying tissue (spongioid) clung to the cartilage. The level of injury is of prognostic significance because fractures through the lower parts of the hypertrophic zone or the metaphysis have little likelihood of producing growth arrest, whereas those through the germinal, proliferating, and upper hypertrophic zones can be problematic. Pure cartilage separation was conceded to exist but was felt to be concentrated in the very young (less than 1 year) and rather rare. Bruns considered lesions within the cartilage to be common, however (41). Most studies defined some bone fragments with the cartilage fracture, with involvement of the spongioid and spongy diaphyseal bone layers most common. The work of Bret and Curtillet spurred this view, Oilier pointed out the common juxtaepiphyseal lesion, and Rollet opted for a "juxtaepiphyseal detachment" terminology. Coulon focused on the rarity of the pure transcartilage pathway of separation, commenting that "we have found in most of those in which the accident was recent that the fracture if it traversed the epiphyseal line at one part of the bone generally left it at another part." Oilier felt that epiphyseal separations never occurred within the cartilage, but rather within or at the level of spongioid bone (240). 10. MECHANISMS OF INJURY

Many studies were done to assess the mechanism of injury that leads to epiphyseal fracture-separations. Enormous amounts of traction applied along the longitudinal axis of the bone were needed to bring about epiphyseal displacement particularly when the periosteum remained intact. Virtually all studies from Rognetta (275) and Gueretin (130) in the 1830s and onward showed that torsional and rotational movements were far more effective in bringing about separation with less force. Tapret and Chenet described a distal femoral epiphyseal fracture-separation assessed postamputation from a 989 male. The fracture pattern was along the physis and then into the diaphysis. The ligaments

were intact and they postulated an extension-rotation mechanism (306). Many experiments had shown the extreme importance of the periosteum in supporting the attachment of the epiphysis to the diaphysis. Internal support was relatively weak and was due to the interdigitation of the epiphyseal cartilage columns with diaphyseal bone from below. These tissues are in continuity and, thus, no specific anatomic structure is defined by them although, when the physis is ruptured away from the metaphysis-diaphysis, examination of the undersurface of the cartilage shows small cartilage processes protruding from the surfaces. These have been referred to as mammillary processes. 11. EPIPHYSEAL FRACTURE-SEPARATIONS ACCOMPANYING TRAUMATIC BIRTHS The occurrence of epiphyseal separations accompanying difficult deliveries was well-documented. Extreme manual traction and rotation on the limb during difficult deliveries led to epiphyseal separations rather than to ligamentous injuries, as had been suspected previously. Simpson documented one case in which the child died during birth, and subsequent analysis showed epiphyseal separation at the upper end of the femur, the lower end of the tibia, and the lower end of the fibula and a partial separation of the epiphysis of the distal femur (291). His review of the literature also showed increasing awareness of this injury, which previously had been unsuspected. Experimental work also showed that epiphyseal separation could be produced with extremes of traction. He reported the work of Hofmakl, who indicated that attempts to produce dislocations in fetal cadavers at hip, elbow, and shoulder joints always failed and ended only in epiphyseal separations. He also quotes Kuestner as indicating that his experiments showed that, in the lower extremities, the epiphyses that yielded most easily were those at the lower end of the tibia and fibula. The observation was also made, consistent with many reports including those at the present time, that the epiphysis initially gives way from within with displacement passing outward with further pressure.

E. Understanding of Growth Plate Fracture-Separations at the End of the Pre-radiographic Era 1. HUTCHINSON

Hutchinson presented a general overview of epiphyseal injuries, including a review of fractures at each anatomic area (150-152) in 1893-1894. In his introduction he referred to the possibility of growth arrest with these injuries and to the fact that intra-articular joint involvement of the fractures often worsened prognosis. He pointed out that epiphyseal separations were associated with significant amounts of periosteal elevation and stripping, which led to subsequent hematoma formation and abundant new bone repair along the

SECTION II ~ Clinical and Experimental Investigations in the Pre-radiographic Era

pRON&'r~ QU~OR,

,'rzuq.

PH'f$1$

FIGURE 5

Illustration from the work of Hutchinson (150) demonstrates a type I distal radial fracture-separation. This was recognized as one of the most common of injuries. Note the continuity of the periosteum on the dorsal or concave surface and the rupturing of the periosteum on the volar surface. The physeal tissue is noted to be displaced with the epiphyseal fragment.

adjacent shaft. He also pointed out that fractures led to a separation of the lowermost parts of the epiphyseal growth plate such that most of the plate remained with the epiphysis. The presence of epiphyseal separations, however, in many instances showed elevation but no tearing of the periosteum, which confirmed the belief that the strongest means of union between the shaft and the epiphyses was via the periosteal and perichondrial sleeve. Although epiphyseal fractures were relatively common, they showed "a remarkable power of resistance" by which he meant that, following healing, they continued to function rather than being obliterated. Hutchinson referred to six separations of the lower epiphysis of the radius that he followed for a considerable period of time, only one of which showed some shortening (Fig. 5). His feeling, however, consistent with that of Oilier, was that displaced epiphyses had a higher incidence of growth arrest. Hutchinson was one of the earliest to note the importance of prompt and accurate reposition of the displaced bone in particular with "careful avoidance of rough manipulation in so doing."

2. JOUON Jouon presented a clear view of all aspects of growth plate fracture-separations in 1902 (163-166). He still felt it necessary to state, however, that traumatic separation of the epiphyses constituted a loss of continuity in bones that was very different from fractures in all viewpoints, including causes, symptoms, treatment, and prognosis. The three most common fractures were those of the proximal humerus, distal radius, and distal femur. His focus, as with many preceding works from the nineteenth century, was on an understanding of the pathological anatomy of the fractures, their mecha-

529

nisms of occurrence (generally forceful movement beyond the normal range combined with torsion), and the methods of treatment required both in general and in reference to each specific epiphysis. General treatment approaches involved approximately 3 weeks of plaster cast immobilization for non-displaced injuries. Displaced fracture-separations required reduction under general anesthesia and maintenance of reduction or continuous traction in a plaster cast for 1 month. Irreducible fractures were treated by open reduction and removal (resection) of the diaphyseal segments of bone. As early as 1878, Oilier recommended and practiced cuneiform osteotomy, segmental bone resection, and growth plate cartilage removal (chondrectomy) in cases in which deformity followed growth plate arrest. 3. POLAND In 1898 John Poland published a classic treatise close to 1000 pages long, Traumatic Separation of the Epiphyses, reviewing a vast number of articles pertaining to growth plate fracture-separations (257, 258). His work summarized the extensive knowledge on growth plate injuries that had accumulated during the pre-radiographic era. His several chapters dealt with history, anatomy, etiology, age and sex, frequency of the various separations, experiments in pathological anatomy, mechanism, prognosis, and results, and a discussion of fractures at each specific epiphysis. He listed and clearly illustrated a classification of epiphyseal fractureseparation patterns which included five types although a numbering system was not applied (Figs. 6A and 6B). These included pure and complete transverse separation, partial transverse separation with fracture of the diaphysis, partial transverse separation with fracture of the epiphysis, complete transverse separation with fracture of the epiphysis, and fracture passing from diaphysis through the growth plate to epiphysis and articular cartilage. Poland thus indicated the great frequency with which the transverse epiphyseal separation deviated obliquely into the diaphysis, carrying with it a variably sized diaphyseal fragment. He also defined the much less frequent association of the epiphyseal separation with an epiphyseal fracture passing into the joint itself. Three additional pathoanatomic findings remain valid today: (1) The periosteum is widely detached from the diaphysis in displaced fractures but remains attached to the epiphysis on the concave side of the injury. The periosteum has its strongest attachment to the developing bone beyond the growth plate into the epiphysis. (2) In many instances, there can be separation of the epiphysis but no displacement of the fragments due to the stability provided by the intact periosteum. The absence of displacement with epiphyseal separations due to maintenance of the periosteum was noted by Salmon in studies on newborn human cadavers. The periosteum can be lifted or separated from the diaphyseal bone but not torn. These separations without displacement were

530

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations

A

~'m. 6.--Pure and complete separation

Fxo. 8.--Partia! separation, with fracture of the epiphysis

Fro. 7.--Partial separation, with fracture of the diaphysis

Fro. 9.--Complete separation, with fracture of the epiphysis

FIGURE 6 Parts(A) and (B) from the work of Poland (258) show his representations of epiphyseal growth plate fracture-separations. His Figure 6 refers to the pure and complete separation commonly listed as a type I injury now. Figure 7 refers to a partial separationwith fracture of the diaphysis, now recognized as a type II injury. Figure 8 refers to a partial fracture-separation of the epiphysis, now recognized as a type III injury. and Figure 9 refers to a complete separation with fracture of the epiphysis, which is not classified currently although examples occasionally are seen. Part (B) from Poland's work illustrates the type IV fracture-separation of the distal lateral humeral epiphysis.

considered to be common and often were felt to be ligamentous sprains. Poland notes that "the history of many examples of arrest of growth of bones is that the injury had been considered as that of a sprain." Poland clarified that the age of greatest occurrence of these injuries was often after 11 or 12 years, rather than early in childhood. He presented detailed pre-radiographic era tables on the ages of appearance of the epiphyseal ossification centers and their time of union to the diaphyses. In a summary of a major series of cases, the frequency of long bone epiphyseal injuries was as follows: (1) lower epiphysis of femur; (2) lower epiphysis of radius; (3) upper epiphysis of humerus; (4) lower epiphysis of tibia. Of his personal series of 693 separations, the upper extremity:lower extremity ratio was 1.6:1 (426:267), including distal femoral (125, 18%), proximal humeral (120, 17%), distal radial (112, 16%), distal humerus (81, 12%), and distal tibia (46, 7 %). (The internal epicondyle of the distal humerus was 61 or 9%.) Systematic clinical reviews of traumatic epiphyseal separations were also written by Wolff in Germany in 1900 (332) and by Kirmisson in France in 1904 (176). 4. EARLY SURGICAL EFFORTS IN RELATION TO GROWTH PLATE INJURIES Surgical intervention in relation to growth plate fractureseparations was increasingly bold as problems referable to certain types of fracture became apparent clinically. Diaphyseal resection was done to allow for reduction in otherwise nonreducible fracture-separations (204). Delens reported on open reduction and surgical resection of the diaphyseal (our metaphyseal) fragment of what we would now refer to as a displaced Salter-Harris type II distal femoral growth plate fracture (83). This allowed for the removal of interposed periosteum and muscle, easy reduction, and relief of vessel constriction and did not compromise subsequent growth. Open reduction and diaphyseal resection were done increasingly to achieve reduction without excess pressure on the swollen adjacent tissues and skin. Jetter frequently operated on proximal humeral fracture separations, observing clean intracartilage separations, intracartilage separations with associated longitudinal fractures, partial epiphyseal separations with diaphyseal fractures ( S - H II), and juxtaepiphyseal fractures (within the metaphysis close to the growth plate) (161). As early as 1873 there were clinical reports of growth plate excision and stimulation by periosteal irritation in the treatment of leg and forearm problems caused by growth damage to one of the paired bones. By the end of the nineteenth century, which coincided with the end of the pre-radiographic era, there was good experimental and clinical awareness of the epiphyseal growth plate fracture-separation entity and of the general range of treatments needed. Shortening and angular deformity were recognized as potential negative sequelae. The problems also were understood to be due to the mode of repair postinjury as well as due to the initial degree of displacement.

SECTION II! ~ Clinical Approaches in the Radiographic Era

POLAND

BERGENFELDT

AITKEN

SALTER-HARRIS

STEINERT

~~ ',

FIGURE 7 Severalpathoanatomicclassificationsof epiphysealgrowth plate fracture-separations are shown: Poland, 1898; Bergenfeldt, 1933; Aitken, 1936; Salter and Harris, 1963; Steinert, 1965. The similarity in categorization is evident. [Reprinted from Shapiro (1992). Adv. Orthop. Surg. 15:175-203, 9 LippincottWilliams& Wilkins,with permission.]

III. C L I N I C A L A P P R O A C H E S T O G R O W T H P L A T E F R A C T U R E - S E P A R A T I O N S IN T H E RADIOGRAPHIC ERA

531

companying the epiphysis on its diaphyseal side; group III, epiphyseal separation combined with a diaphyseal fracture; group IV, epiphyseal separation combined with an epiphyseal fracture; group V, epiphyseal separation combined with epiphyseal and diaphyseal fracture; and group VI, juxtaepiphyseal fracture. The epidemiological patterns were documented for such factors as the number of fractures at each epiphysis, the group distribution at each epiphysis, and the age-group relationship. The frequency distribution was similar to that seen now, with distal radial separations most common followed by those at the distal humerus and distal tibia. Group I fractures accounted for 7.4% of all injuries, group II 31%, group III 50%, group IV 6.1%, group V 4.2%, and group VI 1.3%. In one case Bergenfeldt felt that fracture occurred between the epiphysis and the upper part of the growth plate. The incidence of growth inhibition problems overall was quite small and was attributed to a direct lesion of the proliferating cartilage or, in completely intra-articular epiphyses, to marked damage to vessels with loss of nutrition to the growth plate cartilage. There were no instances of increased growth following epiphyseal separation. He concluded that growth damage was significantly less in humans than in animals subjected to experimental growth plate injury, an impression that must be borne in mind in interpreting animal work. The epiphyseal separation in many of the injuries was localized to the zone of ossification and did not interfere with the growth cartilage itself. Even when damage was directly to the growth cartilage, it infrequently led to growth damage. He recognized that negative growth sequelae resulted especially from group V injuries in which both epiphyseal and diaphyseal fractures occurred, with 33% of cases in this group developing shortening.

A. Pathoanatomic Classifications The advent of radiography immediately removed any lingering doubts about the existence of epiphyseal fractureseparations and allowed vastly more accurate diagnosis and treatment to occur. In this section we will describe the classifications that were formulated on the basis of plain radiographic assessments and indicate how these classifications came to dominate the perception and management of growth plate fracture-separations. We refer to these classifications as pathoanatomic by which we mean that they are based on the pattern of fracture as determined by plain radiographs. Several of these classifications are illustrated in Figure 7 and described in more detail next. 1. BERGENFELDT In 1933, Bergenfeldt presented a large study involving 310 patients from Stockholm, Sweden, seen between 1919 and 1928 (23). His work represents the first definitive study from the radiographic era. His pathoanatomic classification defined six types of epiphyseal fracture-separation based on the radiographic appearance as follows: group I, epiphyseal separation with no demonstrable bone involvement; group II, epiphyseal separation with traces of thin bone fragments ac-

2. AITKEN Aitken wrote extensively on epiphyseal growth plate fracture-separations, with articles on injuries at the distal radius (5, 6), proximal humerus (8), distal femur (9), and proximal (10) and distal tibias (7). He indicated that "in discussing the end results of epiphyseal fractures, each epiphysis must be considered individually" as they varied greatly in anatomical structure and physiological function. Such factors as size, shape, and position of each epiphysis as well as the amount of growth contribution had to be considered. Each epiphysis was subject to different types of fractures, the end results of which differed considerably. He first defined his classification of three types in relation to distal tibial epiphyseal fractures (1936) and then related to it distal femoral and proximal tibial fractures. Type I: The fracture line runs parallel to the cartilage plate and just within the newly formed bone on the diaphyseal side prior to passing obliquely outward through a diaphyseal segment. The epiphysis is displaced en masse with little or no damage to the cartilage plate because the fracture line appears to pass through the zone of degenerating cartilage cells and newly formed osteoid tissue and bone.

532

CHAPTER 7 ~ Epiphyseai Growth Plate Fracture-Separations

Type II: The fracture line crosses the bony epiphysis from the joint surface to the cartilage plate, crushing the plate in just under half the cases. In his initial work, Aitken described the fracture line as passing only to the growth plate; in later work, he showed it continuing along the growth plate. If the fracture line emerged between the bony epiphysis and the cartilage plate no growth damage would occur; but if it passed through the plate and emerged between the plate and the diaphysis growth damage would occur. Type III: The fracture line runs from the joint surface through the bony epiphysis, crossing the growth cartilage and passing out the metaphysis. Types II and III were seen particularly at the distal femur and proximal and distal tibias. Mechanism was also implicit in his approach. The more favorable prognosis in type I fractures was due to the fact that they were avulsion injuries, with a less favorable prognosis in types II and III injuries being due to the fact that they were compression injuries with the fracture line crossing and crushing the growth plate. Aitken recognized, however, that even type I avulsion injuries of the distal femoral epiphysis on occasion led to serious growth deformity. Over the next 30 years, large numbers of studies were performed assessing the results of epiphyseal fracture-separations in relation to the radiographic and clinical findings at specific epiphyses. The pathoanatomical classifications pointed out by Bergenfeldt and Aitken commonly were referenced. Fractures of the distal tibia commonly were assessed primarily because they were quite frequent in occurrence and also had a variable prognosis because of the fact that varying fracture patterns were seen there. The works by Buttner (1956) (48), Bartl (1957) (17), and Ehalt (1958) (96) were notable. Bartl in particular reviewed in excellent detail 235 cases of epiphyseal fracture-separations of the distal tibia in a study between 1928 and 1953. Rehbein and Hofmann pointed out the negative effects of axial injuries to the growth plates in 1963 (267) (Fig. 8). They commented on complete fusion of the functioning growth plate with longitudinal forces affecting the complete physis, unilateral asymmetric growth plate arrests with axial forces concentrated on one part of the periphery of the physis, and central arrests in which the axial force was limited to the central part of the physis. 3. S A L T E R AND H A R R I S Over the past 40 years the classification of Salter and Harris has been used extensively to define growth plate injuries (281). Their work defined five types. Type I: There is complete separation of the epiphysis from the metaphysis with no bone fracture and with the growing cells of the epiphyseal plate remaining with the epiphysis. Type II: The line of separation extends along the growth plate then exits through a portion of the metaphysis. The periosteum is torn on the convex side of any angular deformity but is intact on the concave side where the metaphyseal fragment is present.

( a

ttttttt

b

!

I/

/ !

I A ~

F I G U R E 8 Illustration from the work of Rehbein and Hofmann (267) shows their concept of growth plate damage in association with axial forces. In part (A) the small arrows at left indicate damage over the entire extent of the epiphysis, leading to complete premature growth plate closure. Forces in part (B) concentrated at one side of the physis lead to focal growth plate closure and subsequent angular deformity with persisting growth of the rest of the physis. Part (C) illustrates axial pressure centrally leading to a focal bone bridge formation and a tethering effect, allowing for the characteristic fishtail appearance with medial and lateral physeal growth continuing. [Reprinted from (Lang.) Archiv. f. Klin. Chir. 304:539-562, 1963, 9 Springer Verlag.]

Type III: An intra-articular fracture extends from the joint surface through the secondary ossification center to the hypertrophic zone of the growth plate and then along the growth plate to the periphery. Type IV: An intra-articular fracture extends from the joint surface through the secondary ossification center, across the growth plate, and exits through the metaphysis. Type V: A severe crushing force is "applied through the epiphysis to one area of the epiphyseal plate." Rang has added a type VI to the Salter-Harris classification encompassing injury to the periphery of the growth plate region with removal of the overlying periosteal-perichondrial tissues of the groove of Ranvier (262, 263). Removal of this tissue component allows bony bridging between epiphysis and metaphysis. The article briefly referred to the earlier works of Bergenfeldt and Aitken. Salter and Harris added the type V injury to their classification. The occurrence of growth plate arrest after a seemingly benign injury for which no fracture was

SECTION III ~ Clinical Approaches in the Radiographic Era

diagnosed had been known or suspected since the works of Oilier but was clearly delineated by the type V category. This injury can be diagnosed by plain radiographs only retrospectively after a premature growth plate fusion has occurred, whereas the other types, excepting some undisplaced type I injuries, are seen on the initial postinjury film. Rehbein and Hofmann also drew attention to growth plate damage following axial force (267). They illustrated (Fig. 8) how axial force across the entire width of a physis could lead to complete destruction of the growth cartilage, how asymmetric forces could produce partial growth damage and angular deformity, and how local centralized axial damage could produce a central focal physeal arrest and a funnel-shaped withdrawal of the physis. The type V injury has been documented infrequently in other series (216, 248, 252), and the proposed mechanism of occurrence, a compression force along the long axis of the bone, has been doubted by some, whereas others (175) postulate varus, valgus, and shear forces as causative factors. The Salter and Harris paper described the applied anatomy and histology referable to growth plate injuries, reviewed their own experimental work, and cautioned that prognosis depended not only on the type of classification of the injury but also on age, blood supply, method of reduction, and open injury contamination. General principles of treatment and treatment at each specific growth region were detailed. Subsequent works have pointed out some limitations of the Salter and Harris approach in that it serves as a poor prognostic indicator at certain epiphyses. The numerical grading scheme, particularly types I-IV, has been widely adopted and has served as a baseline for describing reported cases of fracture-separation for the past four decades. 4. STEINERT Steinert assessed growth plate injuries in 1965 in an article entitled "Epiphyseal Separations and Epiphyseal Fractures" (304). He differentiated five forms of epiphyseal damage: (1) epiphyseal separation without displacement (epiphyseal loosening); (2) chondroepiphyseal separation with displacement; (3) osteochondroepiphyseal separation (the "osteo" referring to the metaphyseal fragment); (4) epiphyseal fracture; and (5) crush fracture of the diaphysis (metaphysis) with contusion of the epiphyseal junction. The combined epiphyseal, transphyseal, and diaphyseal fracture is illustrated elsewhere in the article. He indicated that prognosis in types 1 and 2 was good, in 3 relatively good, and in 4 and 5 often unfavorable. 5. WEBER

Weber (1980) attempted to provide a prognostic focus based on the pattern of fracture and mechanism of injury (324). The various mechanisms of injury had been discussed extensively in the preradiographic era but were mentioned much less frequently when the radiographic appearance of the fracture pattern became the focus of categorization and management. Weber drew on the work of both Steinert and

533

Aitken for his approach. He also subdivided the injuries as (1) epiphyseal separations, occurring in the layer of hypertrophic cells, being associated with shear forces, leaving the germinal and proliferating cell layers of the plate intact, and having a good prognosis, and (2) epiphyseal fractures, occurring through the articular surface, secondary ossification center, and germinal proliferating cell layers before exiting through the hypertrophic zone or metaphysis, being produced by axial compression and bending, and having a worse prognosis. Among the epiphyseal separations, Weber included (1) loosening of the epiphyseal plate with or without dislocation and (2) partial separation of the epiphysis together with a metaphyseal fracture fragment. Among the epiphyseal fractures, he included (1) fracture of the epiphysis with extension of the fracture line through the epiphyseal plate and then along the interface between the growth plate and the metaphysis and (2) fracture of the epiphysis with extension of the fracture line through the epiphyseal plate and then through the metaphysis. The lesions associated with epiphyseal separation had a relatively benign fate because the injury was through the layer of hypertrophic or degenerating cartilage cells or within the outer reaches of the zone of primary ossification such that the germinal and proliferating layers of the physis, which were responsible for longitudinal growth, remained intact. Even if a metaphyseal bone segment was displaced with the epiphyseal fragment, there was no growth arrest phenomenon because the physis immediately adjacent to the metaphyseal fragment was not fractured and retained its blood supply. Epiphyseal fracture, however, whether of the Aitken type II or III or the SalterHarris type III or IV category had a poorer prognosis because the germinal and proliferating zones of the epiphyseal growth plate cartilage were of necessity damaged. Weber thus felt that "abnormal growth can be predicted on the day of the accident if the injury is sufficiently severe." The mechanism of injury could be deduced, he felt, from consideration of the pathological anatomy as seen in the radiograph. Separation of the epiphysis was felt to be relatively benign because it was caused by an avulsion or shear force that was relatively harmless in comparison with the compression fracture, which crossed the epiphyseal bone and the epiphyseal plate, damaging the crucial cell proliferation region in the process. Although there was no ability for the damaged physis to correct itself, many of the sequelae of epiphyseal fractures subsequently could be considered to be due to transphyseal bone through gaps in the normal epiphyseal growth plate anatomy. Anatomic reduction and rigid internal fixation served to minimize the gap and allow the best opportunity for normal growth to occur. The purpose of operative treatment was to return the anatomical segments of metaphysis, growth plate cartilage, and epiphyseal bone to their normal relationships with each other. Weber's classification defined prognosis as being either good, type A, or doubtful, type B, depending on the line of fracture. The good prognosis or type A injuries were Aitken type I or Salter-Harris types I and II and the doubtful

534

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations

prognosis or type B injuries were Aitken type II or III or Salter-Harris types III and IV. Accurate reduction and rigid internal fixation in the type B group were felt to be essential to minimize negative sequelae. 6. M O R S C H E R

Morscher also approached the injuries from a mechanistic viewpoint. He made the point that the hallmarks of a good classification are simplicity, completeness, and general applicability (220). A classification should help determine the appropriate form of treatment. He felt that neither the Aitken nor the Salter-Harris approach fully satisfied these criteria. Both classifications were felt to be incomplete. The Salter-Harris classification, even though somewhat more extensive than that of Aitken, was felt to concentrate too much on the distal tibial epiphysis and to leave out other important categories. He included among these types a varus crush injury of the distal femoral growth plate region with disruption of the femoral origin of the lateral collateral ligament, which led to premature fusion laterally and a resultant genu valgum. He also pointed to the triplane injury of the distal tibia. Morscher also asserted that neither paid attention to the mechanism of injury, although Aitken did use the mechanistic terms avulsion and compression. He felt that it should be possible to determine the type of lesion and prognosis of an epiphyseal injury from the mechanism causing it. Four factors were felt to be determinants, including (1) the direction of the force producing the injury, (2) the size of the force, (3) the mechanical resistance of the epiphyseal cartilage plate, and (4) the mechanical resistance of the adjacent epiphyseal and metaphyseal bone. Morscher indicated that compression type growth plate injuries were associated with fractures that passed through the bony epiphysis and produced mechanical growth plate damage particularly to the germinal and proliferating cell layers, predisposing them to growth disturbance. Traction and shear forces on the other hand led to epiphyseal separation tending to the metaphyseal side of the plate in the already degenerating hypertrophic zone and often associated with prognostically unimportant metaphyseal fractures, such that only rare growth disturbances occurred. Other studies, however, have shown that the fracture line in these injuries can deviate into the germinal and proliferating cell layers, thus damaging the active growing cells. Morscher thus divided injuries into epiphyseal fractures (compression mechanism) and epiphyseal separations (traction-shear mechanism), which included the prognostically unimportant metaphyseal fractures. The epiphyseal fractures were always associated with injuries of the germinal zone of the growth plate, whereas epiphyseal separations were concentrated in the nongrowing or hypertrophic parts of the plate. This classification, originally attributed to Steinert, was helpful in terms of management in that every epiphyseal fracture is "an absolute indication for operative reduction" with "absolute anatomical reduction of the fragments with compression" needed to prevent bony bridge formation between epiphysis and

aA

~

3A

~

4A

4

2eM.__..---._/

z

3

4

$

6

N

F I G U R E 9 Ogden's (232, 233) pathoanatomic classification of growth plate fracture-separations is shown. Types I - V are essentially modifications of the Salter-Harris approach with anatomic variations added showing differences in the level of physeal or periphyseal fracture or the presence of additional fragments making the injuries more complex. As noted in the text, the precise level of fracture can have meaningful effects as to whether negative growth sequelae occur. The type VI pattern was that described by Rang. The type VII fracture does not involve the physis itself but does involve the epiphysis and frequently is referred to as an osteochondral fracture. The type VIII fracture is metaphyseal and similar to that described by Oilier. This injury too can have negative physeal sequelae. The type IX fracture does not involve either the physis or the epiphysis but is included in growth injuries because it involves the periosteum and sub-periosteal cortex. [Reprinted from Ogden (1982), J. Pediatr. Orthop. 2:371-377, 9 Lippincott Williams & Wilkins, with permission.]

metaphysis and a resultant deformity. The simple epiphyseal fracture-epiphyseal separation distinction allowed a prognostic approach from which treatment could be inferred. 7. OGDEN

In the early 1980s, Ogden defined several patterns of fracture presenting 9 basic types with 21 subtypes in an effort to relate specific injury patterns to the risk of growth disturbance (Fig. 9) (232, 233). This classification pointed out the

SECTION !11 ~ Clinical Approaches i n the Radiographic Era

greater variability of fracture patterns even within the I - V categories than was described by the more basic Salter-Harris scheme. Types I-VI are the Salter-Harris plus Rang groups with subtypes defining either the transverse level of injury as Foucher described in the 1860s or additional lines of fracture associated with the basic pattern. Type I injury patterns involve three variants. Type Ia has the fracture line propagating across the physeal cartilage, type Ib across the primary spongiosa with the physeal interface invariably involved, and type Ic disruption of a localized segment of the physis. Ogden indicates that a large majority of the type Ib injuries occur in children with systemic disorders affecting the endochondral calcification patterns within the metaphysis, such as leukemia and thalassemia. Type II injury patterns involve four subtypes. Type IIa involves a partial propagation of the injury transversely across the physis with passage of the fracture path into the metaphysis. Type IIb is the same pattern except that the metaphyseal fragment adjacent to the area of physeal fracture is also fractured and free. Type IIc propagation travels transversely across the primary spongiosa and then exits the bone through the metaphysis. Finally, type IId involves a pattern similar to that of type IIa except that there is localized disruption of the physis at the point of propagation into the metaphysis. Type III also comprises four patterns. Type IIIa involves propagation through the physis with subsequent deviation of the fracture line through the ossification center of the epiphysis and out through the articular cartilage. Type IIIb involves transverse fracture through the primary spongiosa with passage through the physis, secondary ossification center, and articular cartilage. Type IIIc involves a crushing injury through the peripheral cartilage followed by exit of the fracture through the epiphysis. Finally Type IIId is a nonarticular cartilage fracture, for example, of the ischial tuberosity. Type IV injuries are subdivided into three categories. The type IVa pattern involves a combined epiphyseal-physeal-metaphyseal fragment, type IVb involves an epiphyseal-physeal-metaphyseal fragment combined with a type IIIa or IIIb transverse lesion of the physis, and the type IVc fracture involves propagation through a nonarticular epiphyseal region, which is most commonly the intraepiphyseal cartilage of the developing femoral neck. Type V injury patterns involve two variants, which are either eccentric disruption of the physis or a telescoping comminution of the metaphysis into the epiphysis. The type VI injury involves avulsion or crushing of the periphery of the physis at the groove of Ranvier area. Type VII fractures are intraepiphyseal, involving the articular surface analogous to osteochondral fractures. The type VII injuries have two variants. The type VIIa pattern involves an osteochondral fragment of the secondary ossification center physis and trabecular bone, and the type VIIb pattern involves a chondral fragment only passing as deep as the hypertrophic cell layer of the physis of the secondary ossification center. The type VIII injury is a metaphyseal fracture adjacent to the growth plate (and analogous to Ollier's juxtaepiphyseal lesion), which impacts nega-

535

tively on the physis by temporarily cutting off the nutrient artery causing ischemia to the metaphyseal segment between the fracture and the physis. Type IX injury involves damage to the periosteum of the diaphysis and adjacent metaphysis with or without bony injury, which serves to disrupt the normal periosteal intramembranous ossification growth mechanism. The Ogden classification represents the most extensive presentation of the pathoanatomic approach to categorization. 8. PETERSON Peterson has reviewed physeal fractures from the Mayo Clinic and developed a classification encompassing six types (Figs. 10A-10C) (254, 255). He assessed 951 physeal fractures over a 10-year period from one county in Minnesota. On the basis of his study, he found that approximately 16% of the fractures did not fit comfortably into the Salter-Harris classification. Two new fracture types were defined (254). In his review of fractures, he found no Salter-Harris type V injuries and therefore did not include them in his classification. Peterson's numerical classification from types I-VI is based on the increasing degree of subsequent physeal damage found, such that type I has the least amount and type VI the most. His type I fracture refers to transverse metaphyseal fractures that have linear fracture extensions from the metaphysis into the physeal cartilage but do not pass either along or through the physis. He subdivided these into four types (A-D) depending on the number, position, and displacement of the linear fractures passing from the transverse metaphyseal fracture to the physis. Essentially all of the 16% of cases that did not fit into the Salter-Harris classification are defined as Peterson type I injuries. These injuries were concentrated in the distal radius, metacarpals, and finger phalanges, which comprised almost 80% of those found. All were treated nonoperatively with uniformly good results. His type VI physeal fracture refers to major trauma in which part of the physis is missing. These injuries are extremely rare; they require an open or compound injury and are almost always followed by the development of transphyseal bars. Peterson's type II-V patterns are wellrecognized components of all the other pathoanatomic classifications. His classification concentrates on physeal injuries, but he does point out that some epiphyseal fractures do not involve the physis such as those of the ulnar styloid, tibial medial malleolus distal to the physis, tibial spine, osteochondral fractures involving the articular surface and secondary ossification center in several bones, and the distal tip of the fibula. Ogden classified some of these injuries in his type VII group.

B. Mechanistic Considerations 1. BROCA Broca was one of the first to try to relate epiphyseal growth plate fracture-separations to the mechanism of injury and the anatomic structure of the periphyseal ligaments (38, 39). He defined two possible mechanisms: direct, in which the force was applied directly onto the physis, and

536

CHAPTER 7 9

Epiphyseai Growth Plate Fracture-Separations

A I

MAYO

I

II

III

IV

V

METAPHYSIS --'~PHYSIS

METAPHYSIS & PHYSIS

PHYSIS

EPIPHYSIS & PHYSIS

METAPHYSIS & PHYSIS & EPIPHYSIS

Poland II Bergenfeldt II & III Aitken I Salter & Harris II

Foucher I Poland I Bergenfeldt I Salter & Harris I

Poland III & IV Bergenfeldt IV Aitken II Salter & Harris III

VI PHYSIS MISSING

Bergenfeldt V Aitken III Salter & Harris IV

C

A

B

C

D

A

B

C

FIGURE 10 The most recent pathoanatomic fracture classification is by Peterson (255). His categories I-VI are illustrated in part (A) along with terminology and patterns from previous pathoanatomic categorizations. Part (B) further characterizes his type I pattern, which essentially involves metaphyseal fractures that pass into the physis. Part (C) illustrates his type VI categorization in which variable segments of the physeal tissue are missing due to the severity of the trauma. [Reprinted from Peterson (1994), J. Pediatr. Orthop. 14:439-448, 9 LippincottWilliams& Wilkins, with permission.]

indirect, in which the physis was ruptured by ligamentous traction. When the mechanism of injury was by a direct force, it was irrelevant whether the physeal ligamentous insertions were above or below the physis because physeal separation was not caused by ligamentous traction. The large majority of injuries were felt to be caused by the indirect mechanism. An essential anatomic feature of epiphyseal separations caused by the ligamentous traction mechanism was the need for the ligament to insert onto the epiphysis between the articular cartilage and the physis (39). In other words, the growth plate had to be extracapsular because all intracapsular growth plates had ligamentous attachments beyond the physis at the metaphyseal level. The tearing away of the physis, or the physeal separation, began at the side where the traction force made the ligament fight because the same force tended to relax the ligament on the other side of the bone. A series of anatomic drawings at each major physis accompanied one of Broca's reports (30). In those regions where ligamentous insertions occurred exclusively onto the epiphysis, a high number of primary and complete separations occurred with the distal radius and distal tibia being good examples. Separations caused by ligamentous traction were also common at the distal

femur, where most (although not all) insertions were onto the epiphysis, but were relatively rare at the proximal tibia and fibula where most insertions were below the physis. The proximal femoral capital epiphysis rarely was the site of fractureseparation and was characterized anatomically by having no major ligamentous or muscle attachments. Broca also pointed out that, in growth plate separations caused by ligamentous traction, the periosteum was elevated from the shaft of the bone often for a considerable distance on the side to which the epiphysis was displaced, whereas it was torn on the opposite side.

2. AITKEN, STEINERT, WEBER, AND MORSCHER Each of the pathoanatomic classifications of Aitken, Steinert, Weber, and Morscher listed in the previous section attempted to incorporate mechanistic features, with the implication being that compression injuries had a far greater negative effect on physeal tissue than traction or shear injuries. It is, of course, difficult to work such dynamic considerations into static or radiographic categorizations, but efforts to consider and document such mechanisms in some way ultimately should prove beneficial.

SECTION IV 9 Pathophysiologic Approaches 3. CAROTHERS AND CRENSHAW Carothers and Crenshaw attempted to derive clinical significance from a classification of epiphyseal injuries and concentrated on those at the ankle due to their frequency and fairly marked differences in outcome (56). In assessing fractures at the ankle, they recognized that anatomically similar fractures and/or growth plate separations had been caused by predictably similar injuries. They felt that the causative mechanism was crucial to a determination of the end result. For this reason, they categorized their ankle fractures into groups dependent on the mechanism of injury and the position of displacement. At the distal tibial and fibular epiphyses, lateral displacement involved abduction type injuries, posterior displacement of the tibial epiphysis was caused by an external rotation type injury, a forcible plantar flexion type injury, axial compression, or direct violence, and medial displacement was seen with adduction injuries. Each of their lateral displacement or abduction injuries involved the entire epiphysis moving as an intact unit with an accompanying lateral metaphyseal fragment (Aitken type I or S - H type II). There were 20 of these injuries in their series. The results were almost uniformly good with few negative sequelae. External rotation epiphyseal fracture-separations were recognized by a tendency toward a posterior displacement of the tibial epiphysis accompanied by a fracture of the fibula. Deformity was not noted in any of the five patients in this group at long-term follow-up. Even in those cases with imperfect reduction, in these groups remodeling occurred well. In those with forced plantar flexion there was posterior displacement of the tibia with a posterior metaphyseal triangular fragment. Long-term growth was good. In the group with posterior displacement due to axial compression and direct violence, there were a few cases but no negative sequelae seen. It was the group with adduction injuries who had the major negative long-term sequelae. There was a marked tendency toward medial displacement of the distal fragments. Second- and third-degree injuries involved fractures through the tibial epiphysis, with first-degree injuries having only minimal medial displacement. In those with second- and third-degree adduction injuries, three of six cases had major negative growth sequelae with deformitites due to medial physeal closure and varus ankle tilting and shortening. They concluded that all of the clinically significant deformities resulting from the injuries of the ankle epiphyses were due to the second- and third-degree adduction fractures. Firstdegree injuries had no negative sequelae. They then invoked mechanistic reasons to explain the long-term prognosis. In external rotation or abduction injuries at the ankle, the deforming force was by a shear mechanism, which involved transverse fracture through the relatively weakened hypertrophic region of the physis and into the metaphysis. With the adduction injuries the medial border of the talus exerted a compressive force through the articular surface and the tibial epiphysis. They referred to their experimental studies in which the force exerted through a bone in compression is

537

many times greater and thus presumably more destructive than that that can be exerted through shear.

C. Negative Sequelae of a Growth Plate Fracture-Separation There are three types of negative sequelae that can follow a growth plate fracture-separation: (1) shortening of a long bone, which occurs due to premature fusion of the growth plate across its entire extent; (2) angular deformity and some shortening, which occur due to a localized transphyseal bone bridge that prevents growth in a part of the physis while continuing growth of the rest of the physis leads to angular deformity; and. (3) joint surface irregularity, which occurs with a failure of anatomic reduction of fractures involving the articular surface or with central transphyseal bone bridges with continuing growth of the physis peripherally. These are illustrated in Fig. 11. The optimal treatment of any growth plate fracture-separation relieves the patient's pain and allows for continuation of normal growth. Treatment is designed to prevent the complications listed or, if they occur, to diagnose them early and minimize their effects by various interventions.

IV. PATHOPHYSIOLOGIC A P P R O A C H E S TO GROWTH PLATE FRACTURE-SEPARATIONS

A. Pathophysiologic Classification--Shapiro 1. CONCEPTUAL CONSIDERATIONS A common thread runs through each of the pathoanatomic classifications presented previously because the patterns of separation and fracture are defined on the basis of plain radiographic appearances. The articles concerning traumatic separation of the epiphyses show that the various categorizations describe the fracture patterns and provide, in some instances, prognostic indications based on the pathoanatomic pattern of injury (283, 286). However, little conceptual progress appeared to have been made from the articles discussing these lesions 50-100 years previously. If one considered the clinical and experimental pre-radiographic era work of Foucher (1860) (111), Cornil and Coudray (1904) (70, 71), and Poland (1898) (258) and the radiographic era study by Bergenfeldt (1933) (22), all subsequent categorizations either were repetitive or showed differences that were not substantial in nature. This is not surprising because the same method of assessment, the plain radiograph, was being used. An extensive literature review by Shapiro (283) of papers reporting fracture-separations from each of the major epiphyses showed that the pathoanatomic classifications often were lacking in prognostic significance. For example, many Salter-Harris type II fractures of the distal femur had extremely negative sequelae, whereas the vast majority of

538

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations

Negative Sequelae of Growth Plate Fracture-Separations 1. Shortening Premature physeal fusion across entire extent of physis

2. Angular Deformity with Shortening Localized transphyseal bone bridge with continuing growth of rest of physis

3. Joint Surface Incongruity Failure of anatomic reduction of articular surface fractures

Central transphyseai bone bridge with continuing physeal growth peripherally

= incongruity l = transphyseal epiphyseal-metaphyseal vessel communication with bone bridge formation FIGURE 11 The three types of negative sequelae following a growth plate fracture-separation are illustrated.

Type II fractures of the distal radius and proximal humerus had a benign outcome. It thus was evident that several factors other than the pathoanatomic pattern, as revealed by plain radiographs, played a role in determining the outcome. This led us to consider a different conceptual approach to growth plate fracture-separations to focus on events at the cell and tissue levels that are the ultimate determinants of whether growth postinjury will continue normally or be compromised (Table II). A pathophysiologic approach to growth plate fracture-separations was outlined that sought

to better define the specific factors leading to either full recovery or growth plate damage (283, 286). It was felt to be important to assess each epiphysis specifically and also to take into consideration such factors as the shape of the epiphysis, the age of the patient, and the force and mechanism of the injury. It was our hope that one could superimpose determination of physiologic conditions onto the pathoanatomic categorizations to provide a better prognostic index. The pathophysiologic approach seeks to define whether a growth plate arrest will occur using physiologic indicators as well as

SECTION IV ~ Pathophysiologic Approaches

539

TABLE II C o n c e p t s Underlying the Pathophysiologic Approach to Epiphyseal G r o w t h Plate Fracture-Separations 1. Negative sequelae following growth plate fracture-separations occur due to (1) damage to the blood supply on the epiphyseal side, (2) mechanical crushing damage to cells in the germinal and proliferating cell layers of the physis, and (3) communication between the normally separate epiphyseal vessels with their associated osteoprogenitor cells and the metaphyseal vessels with their associated osteoprogenitor cells, allowing a transphyseal bone bridge to form. 2. Pathoanatomic classifications are based on the pathway of breakage as determined radiologically. The pathophysiologic classification is based on the state of the epiphyseal blood supply and the presence orabsence of epiphyseal-metaphyseal vessel communication. 3. By superimposing the pathophysiologic classification on the pathoanatomic classifications, such injuries can be considered on the basis of the cellular and tissue events that are the main determinants of the end results. 4. Use of the pathophysiologic approach: (1) Knowledge of growth plate anatomy, blood supply leads to expectation of which injuries might produce problems. (2) Knowledge of reported results of injuries at each epiphysis provides a good awareness of possible problem situations. (3) Demonstration at time of injury and/or after reduction of epiphyseal vascularity, epiphyseal-metaphyseal vessel communication in those lesions with a likelihood of growth arrest problems. 5. Pathophysiologic Classification: Type A: The epiphyseal circulation remains intact and there is no communication between the epiphyseal and metaphyseal circulations. Type B: The epiphyseal circulation remains intact, but there is communication between the epiphyseal and metaphyseal circulations, with their associated osteoprogenitor cells, allowing a bone bridge to form. This communication can occur on the basis of gross displacement (B1) or gross microscopic crushing, fissuring, and transverse separation of the cartilage plate (B2). Type C: The epiphyseal circulation is destroyed and the growth plate cartilage dies.

relying on inferences from the radiologic patterns and reported results. Although the various pathoanatomic systems are valuable in describing the fracture patterns within the limits of resolution of plain radiographs, they do not relate to this specific point. On the basis of extensive experimental and clinical evidence, there are two basic ways in which a growth plate fracture-separation can lead to negative growth sequelae. These involve damage to the blood supply on the epiphyseal side, which prevents appropriate chondrocyte proliferation and growth of the plate, and intermixture of the epiphyseal and metaphyseal circulations each with their associated osteoprogenitor cells, thus allowing a focal or extensive transphyseal bone bridge to form.

crushing, fissuring, and longitudinal fracture causing transverse separation of the cartilage plate (BE). Type C: The epiphyseal circulation is damaged and the adjacent growth plate cartilage dies. Descriptions of these basic types are amplified in Figs. 13A- 13D. By superimposing pathophysiologic considerations on the pathoanatomic type it is felt that better prognostic indices can be obtained that will be helpful in terms of treatment of the initial fracture and appropriate follow-up. The experimental and clinical considerations underlying the pathophysiologic approach are described next.

2. PATHOPHYSIOLOGIC CLASSIFICATION A pathophysiologic classification with types A, B~, B2, or C is described to be used in association with the pathoanatomic classifications (Fig. 12) (283, 286). Type A: The epiphyseal circulation remains intact and there is no communication between the epiphyseal and metaphyseal circulations, such that after healing growth proceeds normally. Type B: The epiphyseal circulation remains intact, but there is communication between the epiphyseal and metaphyseal circulations with their associated osteoprogenitor cells, allowing a bone bridge to form. This communication can occur on the basis of gross displacement (B1) or microscopic

The anatomy of the growth plate including its surrounding periosteal-perichondrial sheath and blood supply have been described extensively in Chapter 1. Crucial features in relation to growth plate fracture-separations are reviewed here.

B. Experimental Approaches to Growth Plate Structure, Blood Supply, and Function

1. GROWTHPLATE RELATIONTO CAPSULE The position of the epiphyseal growth plate in relation to the joint capsule and the associated variations in blood supply that occur due to such positions have long been considered to have prognostic significance following injury. Those epiphyses that are entirely intracapsular and completely covered by articular cartilage (proximal femoral capital and

540

CHAPTER 7 9 Epiphyseai Growth Plate Fracture-Separations PATHOPHYSIOLOGIC CLASSIFICATION

E v e s s e l s intact M vessels intact No E - M c o m m u n i c a t i o n

" ~ E vessels intact M vessels intact ~ " = E-M communication

- ~ = E vessels damaged M vessels intact No immediate E - M communication

F I G U R E 12 The pathophysiologic approach is defined here (283, 286). In type A, there are no negative growth sequelae because the epiphyseal vascular supply remains intact and there is no communication between epiphyseal and metaphyseal vessels. In type B, epiphyseal and metaphyseal vessel communication in association with their osteoprogenitor cells occurs, allowing a transphyseal bone bridge to form. In the B~ type with gross displacement, epiphyseal bone rests adjacent to metaphyseal bone. In the B2 type, the growth plate suffers crushing and fissuring, allowing longitudinal fractures to occur that lead to transphyseal vessel interaction between epiphyseal bone and metaphyseal bone. This occurrence is not evident by plain radiographs and accounts for many growth arrest problems after seemingly benign separations. Type C injuries significantly disturb the epiphyseal blood supply, thus rendering all or part of the growth plate nonviable. Bone bridge formation occurs secondarily with vascular invasion of the nonfunctioning growth plate cartilage from the metaphyseal side and eventual linkage of epiphyseal and metaphyseal vessels. [Reprinted from (283), with permission.]

proximal radial) are more likely to suffer damage in association with fracture-separation because of their vulnerable blood supply. The articular cartilage covers the physis completely, passing almost to the level of the physeal cartilage such that vessels enter the epiphysis through a narrow rim of space between the articular and physeal cartilage. Completely extracapsular epiphyses are less vulnerable to complete devascularization following injury, because they obtain their blood supply via vessels that penetrate the side walls of the epiphyses at numerous sites around the entire circumference. There is considerable space between the articular and physeal cartilage through which abundant numbers of vessels can pass. The extracapsular growth plates are at the distal radius, the proximal and distal ulna, the proximal and distal tibia, the proximal and distal fibula, and all epiphyses of the metacarpals, metatarsals, and phalanges. The proximal humeral epiphysis is partially intracapsular (medial part) and partially extracapsular. Those of the distal humerus and distal femur are also partially intracapsular, but primarily extracapsular. The distal femoral epiphysis in particular has a rich circulation penetrating its side walls, which are not covered by articular cartilage. 2. EXTRINSIC PATTERNS OF EPIPHYSEAL BLOOD SUPPLY Dale and Harris defined the two basic patterns of blood supply to epiphyses (75). In type A, the epiphysis is intracapsular and entirely covered by articular cartilage, and in type B, the epiphysis is extracapsular and only partly covered by articular cartilage. When entirely covered by articular cartilage the blood supply enters in a vulnerable narrow zone peripherally between the articular cartilage and perichondrium surrounding the growth plate. When the epiphysis is

only partly covered by articular cartilage, vessels enter through the circumferential borders of the epiphysis between the widely separated articular and physeal cartilage, allowing for an area of supply that is much larger and thus less vulnerable to extensive damage with trauma. 3. NORMAL INTRINSIC EPIPHYSEAL AND METAPHYSEAL BLOOD SUPPLY

Trueta and Morgan reemphasized the dual and separate nature of growth plate blood supply and the fact that the epiphyseal plate itself, with rare exceptions, has no vessels within it (315). There is strict separation of the epiphyseal circulation from the metaphyseal circulation. The epiphyseal vessels supply the epiphyseal cartilage, the bone of the secondary ossification center, and, by diffusion, the germinal, proliferating, and upper hypertrophic cell layers of the epiphyseal growth plate, which are responsible for growth in the length of the bone. The metaphyseal vessels, most of which are terminal ramifications of the nutrient artery and the metaphyseal arteries, pass into the lower regions of the hypertrophic cell layer of the growth plate accompanied by osteoprogenitor cells, which lay down bone on the calcified cartilage cores. Osteoprogenitor cells are present in intimate association with epiphyseal and metaphyseal capillaries such that bone formation occurs in a richly vascularized environment. The state of these two separate circulations posttrauma plays a major role in determining whether normal physealmediated growth will continue. 4. GROWTH PLATE STABILITY The growth plate has both intrinsic and extrinsic support. The columns of cartilage of epiphyseal origin and the bone trabeculae of metaphyseal origin that interdigitate with them

SECTION IV ~ Pathophysiologic Approaches

541

Type BI

Type A ..

.

,~

9 :i :.. . . ". .. .

C

Type B2

D Type C

FIGURE 13 Pathophysiologicfracture types are illustrated [reprinted from (283); with permission]. (A) A type A fracture of the proximal tibia shows the intact epiphyseal blood supply (E) and a fracture line within the growthplate cartilage that still allows for separation of epiphyseal and metaphysealvessels. M, metaphysis;P, periosteal vessels. (B) A type B 1 fracture is shown.Withoutreduction by closed or preferably open methods, bone bridge formationbetween epiphyseal (E) and metaphyseal(M) bone will occurrapidly. (C) A B2fracturepattem is shownin whichcrushing and fissuringof the growthplate allowfor transverse separationof the physealcartilage and longitudinalcommunicationbetween blood vessels and osteoprogenitorcells of epiphyseal and metaphysealbone. (D) The type C fracture pattem is illustratedmost clearly at the proximalfemur due to the tenuous blood supply of this intracapsularepiphysis.

provide intrinsic support. When the undersurface of the cartilage plate is examined carefully following a type I or type II fracture-separation, undulations are seen, which are referred to as mammillary processes. The processes are in essence artifactual, being seen only after fracture-separation has occurred, but they are accurate gross reflections of the histologic cartilage-bone relationships and the slightly varying levels at which they occur. Extrinsic support is provided by the cells and tissues of the circumferential perichondrial ossification groove of Ranvier (47, 285, 298). The circumferential sleeve, which is continuous with the periosteum, completely ensheathes the growth plate, attaching into the epiphyseal cartilage beyond the plate. The supportive function of the periosteal-perichondrial sheath has been welldocumented and provides the large majority of supportive force. Wilson commented in 1820 that the periosteum "strengthened the junction of the epiphysis to the body of the bone" (331). Relatively greater involution of the sleeve in relation to the epiphyseal growth plate in early adoles-

cence (10-15 years of age) may explain, partially, the higher incidence of growth plate fracture-separation in this time period. 5. GROWTH PLATE SHAPE An important factor affecting prognosis following growth plate fracture-separation is the shape of the growth plate (Table III). Not all plates lie in a plane transverse to the long axis of the bone. In those physes that are nonlinear, the irregular shape becomes greater after the first few years of age. The distal femoral and proximal humeral plates in particular have irregular shapes, whereas those of the proximal and distal tibia have less marked, but consistent irregularities. Thus, at certain epiphyses with type I, and in particular type II, fracture-separations the epiphysis and the epiphyseal growth plate do not invariably slide off the metaphyseal fragment along a transverse or gently curvilinear plane. The irregular shapes are often the sites of cartilage crushing and later transphyseal bone bridge formation.

542

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations TABLE III

Epiphyseal G r o w t h Plate Shape

1. Proximal humerus: From birth to 3 years of age, the plate is slightly curved, concave toward the metaphysis. Once

2.

3. 4. 5.

6.

7.

8.

9. 10.

the three ossification centers of the proximal humerus have fused, anteroposterior radiographs demonstrate an inverted V shape to the plate. The epiphysis interdigitates with the metaphysis in a troughlike fashion along the anteroposterior axis. Distal humerus: The plate approaches the transverse plane in the anteroposterior projection. It is somewhat more curved in its lateral projection. In both projections, the concavity is toward the metaphysis. The metaphysis relates to the epiphysis as a cylinder with its long axis in the coronal plane. Proximal radius: The plate approaches the transverse plane in all projections. There is a slight concavity to the plate that relates to the gentle dome-shaped metaphysis. Distal radius: The plate closely approaches the transverse plane in all projections. There is slight concavity relating to a dome-shaped distal radial metaphysis. Proximal femur: From birth to approximately 6 years of age, the plate approaches the transverse plane in all projections. From 6 years onward there is a gentle dome shape to the femoral neck metaphysis with the concavity of the epiphyseal plate toward the neck. Distal femur: The growth plate of the distal femur has an irregular shape. From birth to skeletal maturity in the anteroposterior projection, there is a double concave shape toward the metaphysis. From 8 years onward, the depth of the concavity is relatively greater. On the lateral projection from birth to approximately 7 years of age, the epiphyseal growth plate projects in a transverse plane; during the next few years a slight double concave curve can be noted, and over the final few years, a U-shaped growth plate concave to the metaphysis is seen. Proximal tibia: From birth to approximately 2 years of age in the anteroposterior projection, the plate is slightly curved, concave toward the metaphysis. Over the next several years, the plate approaches the true transverse plane, but again toward the end of growth a slight double concavity is seen. In the lateral projection from birth to approximately 13 years in the male and approximately 10 years in the female, the plate appears level although it is continuous with the apophyseal tibial tubercle plate anteriorly and curves inferiorly at a sharp angle. This becomes readily apparent when anterior ossification commences. Distal tibia: The plate is transverse in the anteroposterior projection except for slight proximal displacement in the medial one-third. The lateral projection shows a gentle curvature of the plate convex toward the metaphysis. The distal tibial epiphysis relates to the metaphysis as a cylinder with the long axis in the coronal plane. Proximal fibula: The plate approaches the transverse plane in all projections. Distal fibula: The plate approaches the transverse plane in all projections.

6. DAMAGE TO BLOOD SUPPLY AND ITS EFFECTS ON GROWTH

Experimental interruption of the epiphyseal blood supply causes cessation of growth in the adjacent physis with the amount of growth plate damage proportional to the extent of epiphyseal devascularization (316). In localized areas the damaged chondrocyte regions are soon invaded by metaphyseal vessels with their associated osteoprogenitor cells, such that a localized bone bridge forms between the metaphysis and the epiphysis. When massive epiphyseal blood vessel damage occurs there is an accompanying large area of physeal necrosis, with metaphyseal vessel and bone cell invasion from the metaphysis leading to premature complete growth plate bone fusion. It is noted that "if the 2 circulations could be kept apart no bridging occurred" (316). Gomes et al. produced growth plate fracture-separations in 112 rats (127). Microangiograms showed the normal dual blood supply of the physis with separate epiphyseal and metaphyseal vessels and an avascular physis. Sequelae of several type III and IV fractures were bone bridges shown by microangiographic illustrations of transphyseal vascular communication.

Haas performed extensive studies on the dog growth plate to assess the relation of the blood supply to the longitudinal growth of bone (132, 133), the localization of the growing point in the epiphyseal growth plate (134), and changes produced in the growing bone after injury to the epiphyseal growth plate (135). These extensive studies both confirmed aspects known previously and added a bulk of new information more precise in detail. His earliest studies demonstrated that "the longitudinal growth of bone is directly dependent upon the integrity of the vascular supply to the epiphysis" (132, 133). Haas performed a detailed assessment of the dog metacarpal and metatarsal blood supply particularly concentrating on the epiphyseal region. He noted a particularly rich blood supply in the peri-epiphyseal region. His growth studies were aided by the fact that in the dog, and also human, metacarpals and metatarsals only one epiphyseal growth plate was situated distally in the second, third, fourth, and fifth bones. In his initial experiment on the relation of blood supply to the longitudinal growth of bone, Haas ligated vessels in varying patterns and varying regions of the metacarpals and metatarsals and also cut tissue in specific regions to

SECTION IV ~ Pathophysiologic Approaches note the subsequent effect on growth. He concluded that no disturbance in growth occurred following an incision through the soft tissue to the level of but not through the periosteum. He also concluded that there was no disturbance of growth when the incision down to the bones also included separation of the overlying tissues. When, however, the entire blood supply to the epiphyseal end of the bone was destroyed by cutting tissues around the entire epiphyseal end of the metacarpal or metatarsal bones, there was an appreciable loss of growth in the bone. When damage was even greater by cutting tissues around the entire distal two-thirds of the metacarpal or metatarsal, there was destruction of both the epiphyseal blood supply and the nutrient artery; there was an even greater disturbance in growth than when only the tissues about the immediate epiphysis were separated. In some of the latter experiments, there was a complete cessation of growth. In another group of experiments, Haas sought to determine the effect of length growth when the nutrient artery alone was destroyed following cutting of tissue about the region of the metacarpal or metatarsal bone where the nutrient artery entered. The disturbances in growth noted were very small in every case, which he interpreted as showing that "diaphyseal or nutrient blood supply is not so important a fact to the length producing property as is the epiphyseal blood supply." He concluded that destruction of the entire blood supply to the epiphyseal end of the metatarsal or metacarpal bone resulted in a marked lessening of its longitudinal growth, even though the nutrient artery remained intact. Conversely, when the nutrient artery was destroyed, there was practically no disturbance in growth length, provided the epiphyseal blood supply remained intact. His evident conclusion was that maintenance of the normal longitudinal growth of bone is dependent upon sufficient blood supply to the region of the epiphyseal cartilage plate. 7. DIRECT DAMAGE TO PHYSEAL CARTILAGE AND SURROUNDING REGIONS AND ITS EFFECT ON GROWTH Haas extended his studies to assess the particular part of the epiphyseal growth plate that was most important in the production of growth in length (135). He summarized well the work of Vogt, who reported on experiments from goats and lambs dealing with injury to the epiphyseal cartilage plate. Both Vogt and Haas found that the greatest disturbance in growth took place after injury to the epiphyseal growth plate. Jahn also performed growth plate experiments using rabbits from 12 days to 4 weeks of age (155). He noted that simple separation of the growth plate was without effect on growth, although excision of the complete cartilage resulted in complete cessation of growth. Injury to the secondary ossification center did not affect longitudinal growth. Jahn felt that the most important part of the epiphyseal cartilage plate in relation to growth was that part of the proliferating cell zone at the beginning of cell columns on the epiphyseal side. Retzius (134) and Leser (193) also concluded from

543

their studies of the cell changes in the epiphyseal cartilage plate that cell proliferation and turnover were greatest in the upper parts of the cartilage columns and that the mitoses continued to decrease as one approached the calcification line. It was evident, therefore, even in the late nineteenth century, that focalization of maximal growth and thus maximal damage to growth was in the uppermost parts of the epiphyseal cartilage plate in the proliferating cell regions. Haas injured the distal metacarpal and metatarsal plates, following which histologic sections were performed. The natural line of cleavage between the epiphysis and the metaphysis occurred at the hypertrophic cell region of the growth plate, which he referred to as the region of "large vesicular cells" (134). It was evident that the epiphyseal cartilage plate remained almost exclusively with the epiphyseal fragment in these separations. Once he was able to demonstrate that the separation through the hypertrophic or vesicular cell layers was seen quite commonly, then localization of points of damage could be performed more readily. Haas' varying excisions therefore were performed after gentle separation of the cartilage plate. Because the physeal line was quite uniform in these epiphyses, virtually all of these separations were of type I or type II with the transverse line of separation through the hypertrophic zone. After removal of a section from the metaphyseal region of growing bones, he concluded that "the metaphyseal region appears to play but little part in the direct lengthening of the bone." In a study of growth effects following simple separation of the cartilage plate, Haas concluded that there was "a definite hindrance to growth," such that the amount of disturbance was dependent upon the amount of injury to the epiphyseal cartilage plate. In clinical practice, however, many growth plate fracture-separations, even with extensive displacement, are not invariably followed by growth damage. In the next subgroup of experiments, Haas excised the epiphyseal cartilage plate and noted only a limited amount of further growth, therefore concluding that the chief growing function was located in the epiphysis. He concluded, in agreement with others, that the greatest growth activity was located in the columns of cartilage as distinct from the hypertrophic regions.Thus, Haas was able to localize growth to the epiphyseal cartilage region. In his various subsets, the greatest growth damage occurred with excision of epiphyseal cartilage plate and metaphysis with, in decreasing order, growth affected by excision of the entire epiphyseal cartilage plate, incision through cartilage at the junction of the epiphysis and metaphysis, and separation in the natural line of cleavage between epiphysis and metaphysis. The least growth disturbance occurred with excision of a metaphyseal segment alone. Thus, he was able to conclude that "the most active and important elements necessary for longitudinal growth are located in the columns of cartilage of the epiphyseal plate" (134). In his next series of studies, Haas produced damage to the growth plate and growth plate region and then sought to

544

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

assess the effects on longitudinal growth (135). Haas established from his previous studies that proper functioning of the cartilage plate required an adequate blood supply, that all growth in length was mediated through the epiphyseal growth plate, that the greatest growth activity was localized in the cartilage columns of the epiphyseal cartilage plate, and that the metaphyseal cartilage (by which is meant the persisting cores of calcified cartilage from the endochondral sequence) related to the transformation of cartilage to bone rather than growth itself. It was the vessels on the epiphyseal side that contributed to longitudinal growth, with the nutrient artery not involved in growth. He then sought to define which injuries to the growth plate led to negative growth sequelae and which did not. Bidder had shown, in a series of experiments upon rabbits, that injuries in the periphyseal area caused growth arrest only when the cartilage itself was injured, after which time the epiphyseal cartilage plate either had degenerated or was missing entirely with tissue substitution by bone trabeculae, which directly connected the diaphysis and the epiphysis (25). Helferich removed the entire cartilage plate from a distal end of the ulna in rabbits and obtained complete cessation of growth. Many additional studies were performed with the general conclusion being that, in some instances, growth arrest occurred following damage to the epiphyseal growth plate, whereas in others it did not. It was becoming apparent that an important point to know was the exact position of the physeal cut because a variation in position gave a variation in results. Haas' work was performed on the epiphyseal growth plates in young dogs or cats from 6 to 8 weeks of age. The lesions involved cross incisions, longitudinal incisions, cross incisions with elevation of the distal segment, longitudinal incisions with removal of the lateral half of the growth plate, cross incisions with removal of the entire distal segment, and cross incisions with removal of the proximal non-epiphyseal growth plate segment. There was only a very slight, if any, disturbance in growth following a transverse incision across the epiphysis (by which is meant the secondary ossification center). When the transverse incision was made through the epiphyseal cartilage plate, in each instance there was a definite disturbance in growth. In the next group, an incision was made through the periosteum around the growth plate dorsally and laterally, after which pressure was made upon the head of the bone. This caused the epiphysis to separate uniformly in the region of the large vesicular or hypertrophic cells of the columns of cartilage. In many instances, there was no subsequent loss of growth. Incisions through the metaphysis did not cause any disturbance in length. In terms of the transverse incisions, it really was only the incisions through the epiphyseal plate that caused a considerable loss of growth. Longitudinal incisions through the distal portion of the bone, including the epiphyseal cartilage plate, the epiphyseal bone, and the articular cartilage, in Haas' work did not pro-

duce growth problems. This would conflict with current views if there were any displacement of adjacent fragments following injury. When there was a transverse incision through the secondary ossification center and then displacement of the distal fragment, there was practically no disturbance in growth unless the incision line had wandered into the epiphyseal cartilage plate. When the transverse incision was through the epiphyseal cartilage plate followed by elevation or displacement of the distal segment, in every instance there was a marked loss of growth but not a complete cessation. Any remaining growth was attributed to the persisting cartilage. When the transverse incision was made through the metaphysis just adjacent to the growth plate and the segment was then elevated and displaced, in practically every instance there was complete loss of growth. When longitudinal incisions were made through the epiphysis, including the epiphyseal growth plate and part of the metaphysis with removal of the segment, in each instance there was a definite decrease in the length of the bone. The length discrepancy increased with time. There was also a growth disturbance with excision of one lateral half of the epiphysis and metaphysis. When a longitudinal half of the entire bone was removed, complete cessation of growth occurred. A transverse incision with removal of the distal fragment through the secondary ossification center showed practically no disturbance in growth or the structure of the epiphyseal cartilage plate. Removal of the entire epiphysis transversely up to the metaphysis led to a failure in growth in length. The transverse incision through the metaphysis and diaphysis with removal of the distal fragment was not definitive. Haas also drilled directly into the epiphyseal secondary ossification center and curetted out a portion of the cancellous bone. There was no hindrance of growth. When, however, the drilling went into the epiphysis and through the epiphyseal cartilage plate, there was a decided hindrance in the growth of the bone following the removal of a portion of the epiphyseal cartilage plate from within. This shows that even with an adequate blood supply from without, there will be a loss of growth if the epiphyseal cartilage plate is injured. If a hole was bored into the metaphysis and the cancellous bone of the epiphysis removed from within by means of a hole made through the epiphyseal cartilage plate, there was loss of growth in almost every instance. This was due to injury of the important cells of the epiphyseal cartilage plate. Haas concluded that "the nearer the injury comes to the cartilage columns, the greater the growth disturbance" and "that there is a relation between the degree of destruction of the cartilage columns and loss of growth." It was these two factors along with disturbances in blood supply that formed the important principles governing epiphyseal bone growth. The overall conclusions from his massive study were that "the result of any injury to the epiphyseal cartilage plate is dependent upon the degree of closeness of the injury to the

SECTION IV 9 Pathophysiologic Approaches

columns of cartilage and to the amount of destruction of the direct blood supply to the region of the epiphyseal cartilage plate" (135). 8. PATHWAY OF FRACTURE LINE THROUGH VARYING LEVELS OF PHYSEAL CARTILAGE

The issue of the fracture pathway through the cartilage growth plate that so intrigued late nineteenth century investigators has received renewed attention over the past few decades. Histologic sections made in our laboratory in rabbit models show the variable levels within the physis and adjacent metaphysis where transverse separation can occur (Fig. 14) (286). Brashear (1959) studied histologic responses in the rat to Aitken type I (Salter-Harris type II) distal femoral growth plate fractures created manually by varus deformation (33). This resulted in distraction on the lateral side of the plate and compression on the medial side. He noted growth disturbances in several animals because the fracture line involved not only the hypertrophic zone but, particularly in the central apical area, frequently passed through all layers of the cartilage due to the "grinding of the metaphyseal bone into the epiphyseal plate." Examples of physeal bone bridge formation were illustrated. Brashear thus demonstrated that not all Aitken type I (or Salter-Harris type II) injuries had a benign prognosis, noting that "the configuration of a particular portion of the epiphyseal p l a t e . . , may make this part more susceptible to injury." Where distracting forces were involved (on the lateral growth plate area in this model), the fracture occurred in the zone of hypertrophic cells. Compression resulted in a fracture medially through the metaphyseal trabeculae. A combination of shearing and compression, however, appeared to result in a grinding of the metaphyseal bone into the epiphyseal plate, damaging all layers of cells and predisposing one to transphyseal bone bridge formation. The configuration of the distal femoral growth plate with its central apical deviation from the linear plane made this part highly susceptible to injury. Brashear also concluded that the level of injury to the epiphyseal plate determined the healing process and the presence or absence of subsequent growth. Variable fracture paths through all levels of the growth plate have been demonstrated by Bright et al. in a rat model, even though all fractures appeared to be Salter-Harris type I (36). There were different patterns of failure with different modes of loading. In only 15% of cases was the fracture line propagated uniformly through the hypertrophic zone, whereas in 85% the crack at least partially passed through the proliferating or germinal cell layers. Peltonen et al. studied the effects of torsional force on segments of bone comprising epiphyseal and metaphyseal bone and the intervening physis. Studies in vitro using cadaveric adolescent sheep proximal tibias demonstrated the physeal fracture line to be through the hypertrophic chondrocyte layer in younger animals, but fracture occurred in a

545

more undulating pattern with metaphyseal bone involvement in older animals (247). Morscher et al. applied traction force in vitro to the proximal tibial epiphysis of rats at varying ages and after growth with varying levels of circulating hormones. The tensile force was applied through the cruciate ligaments following the removal of surrounding soft tissues, including the periosteum. In a carefully controlled environment and using a linear growth plate as a model, fracture occurred at the weakest point, which was always the lower part of the hypertrophic zone within the mineralized component (219). A biomechanical and histologic study of growth plate failure by Moen and Pelker assessed the pattern of growth plate failure when the direction of stress was varied (217). They assessed 1 • 1 • 3 cm 3 central cores of physeal cartilage, which contained none of the surrounding perichondrial tissues, from the proximal and distal femur and proximal tibia of immature cows. Their results showed that the histologic failure pattern varied with each type of applied load. All of the specimens tested in compression failed through the zone of calcification and the distal metaphysis when loaded to the yield point. In this group, there is no evidence for failure of the specimens in the germinal zones. In tension, the majority of the specimens failed in the level of the upper zone of columnization (or proliferation), with the other zones not showing any significant histologic disruption. The shear failure pattern occurred between the upper columnar (proliferating) zone and the lower hypertrophic zone in approximately 80% of the specimens. Portions showed an extremely varied failure pattern with the fracture line being present both in the lower columnar or proliferating zone and in the lower zone of cell hypertrophy. Much of the varying fracture patterns were felt to be due to the complex material nature of the growth plate, in particular the changing orientations of the collagenous matrix. Moen and Pelker concluded that the type of mechanical loading applied to the physis determined the histologic zone of failure in a consistent pattern. The pattern would be even more variable when one considered that, in a clinical situation, the physis was surrounded by the perichondrial tissues that in most instances added a degree of nonlinearity, which would further alter the fracture pattern. Different histologic fracture patterns with different directions of shear loading were also produced in the proximal femoral capital epiphysis of immature 6- to 7-week-old rabbits after removal of the perichondrial sheath (190). Fractures through the physis were not confined to any single zone. Many specimens showed the fracture line weaving through various zones of the plate, often including the resting and germinal zones above and metaphyseal bone below. The level of the physeal injury frequently changed when the physis curved, altering its conformation from the linear plane. Multiple longitudinal cracks perpendicular to the transverse plate also appeared. When similar shear stresses

F I G U R E 14 Photomicrographs ofhistologic sections demonstrate the variable levels of fracture within the epiphyseal growth plate. The fracture-separations were produced by upward manual pressure against the volar surface of the distal metatarsal epiphysis in 2-month-old New Zealand white rabbits. The fracture patterns were Salter-Harris either type I or type II. (A) A section of normal growth plate is shown with epiphyseal bone above and metaphyseal bone below. The layers of the growth plate are seen clearly comprising the germinal or resting zone, the proliferating or columnar zone, and the hypertrophic zone. (B) An example of a SalterHarris type I fracture (as it would be described radiographically) shows the differing levels of separation at left and right that can occur within the same growth plate. (C) A higher power view of B (left side) shows the fracture line passing through the zone of hypertrophic chondrocytes. A higher power view of B (right side) illustrates the fracture line just within the metaphysis below the hypertrophic zone with involvement of a few spicules of bone. (D) Part of the growth plate in another animal shows the fracture line passing through the hypertrophic zone at left but deviating into the zone of proliferating cells and the lower part of the germinal zone at right. (E) Photomicrograph of another Salter-Harris type I fracture. At left there are a few spicules of bone present, indicating the fracture line just below the hypertrophic zone. At right there has been considerable crushing and fissuring of the growth plate within the resting, proliferating, and hypertrophic zones. (F) Higher power view of the right side of E shows the crushing and fissuring of cartilage that can occur. Although this fracture and one solely through the hypertrophic zone, as in region A of B and C, would both appear as type I S - H injuries on plain radiographs, it is not surprising that complications due to mechanical damage might be higher in injuries like the one shown here. [Reprinted from Adv. Orthop. Surg. 15:175-203, 9 (1992) Lippincott Williams & Wilkins, with permission.]

SECTION IV ~ Pathophysiologic Approaches

were applied to progressively older animals, the predominant zone of fracturing changed from the columnar zone in the 2- to 6-week-old group to the hypertrophic zone in the 14-week-old group and from type I fractures in the young to type II in the older animals (278). Gomes et al. performed experimental studies of types I and IV epiphyseal fracture-separations using 112 rats. The fracture line in type I and II lesions often was within the proliferating zone and occasionally reached the germinal zone (127). In one of the few published studies of the physeal fracture pattern in humans in the modem era, Smith et al. assessed a type I fracture in the distal tibia of a 9-year-old boy (293). When they assessed the epiphyseal plate microscopically, the plane of cleavage was found not to be uniform but to involve the zones of resting, proliferating, hypertrophying, and provisionally calcified cartilage cells. Indeed, their quantification showed that, of the transverse diameter of the entire growth plate, 64% had the fracture line through the proliferating zone and 32% had it through the hypertrophying zone, with approximately 2% in each of the resting and provisionally calcified zones. An additional area of continuing study has involved the periosteum and perichondrium in developing bones. It has long been recognized that the periosteum is loosely adherent to the developing diaphysis and metaphysis and yet plays a major role in growth plate support. It is placed circumferentially about the plate and attached by its outer fibrous layer into the epiphyseal cartilage beyond the growth plate. The microanatomic structure of the periphery of the growth plate, referred to as the perichondrial ossification groove of Ranvier, has been well-defined. Chung and colleagues defined the support potential of the periosteal-perichondrial sheath and the histologic plane of fracture in the proximal human femur growth plate from 5 days to 16 years of age (65). The shear strength of the growth plate varied with age and was dependent on the surrounding periosteal-perichondrial complex, especially in infancy and early childhood. They also found that shear failure was characterized by undulating cracks through the metaphyseal, hypertrophic, and columnar zones of the physis in older specimens. Shear force caused a type I fracture through the zone of hypertrophy in the younger specimens, but the older specimens fractured in a type II pattern with an irregular course through the layers of the physis. 9. ADDITIONAL STUDIES Ford and Key (108) in one study and Friedenberg (117) in another, both using the distal femur of the immature rabbit, showed that both size and position of transphyseal defects determine whether growth arrest or deformity would occur and their extent if they did. Friedenberg removed variable amounts of the lateral physeal tissue and noted shortening and valgus deformation, which were greater when larger amounts of physeal tissue, approaching the midpart of the lateral condylar physis, were removed. Transphyseal

547

bone bridges were formed in the gap with no evidence of cartilage growth plate reconstitution. When only a small lateral segment of physis was removed, a narrow bone bridge formed linking epiphyseal to metaphyseal bone, but no shortening or angular deformity was noted. Friedenberg pointed out that relatively small bone bridges did not necessarily lead to growth arrest, with the normal growth force of the remaining physis appearing to overcome any tethering effect, perhaps by elongating the fused area. Ford and Key documented the negligible effects on growth caused by small (88in.) diameter central drill hole defects passing through the physis. When the central defect was markedly enlarged by additional curettage and when circumferential peripheral physeal cartilage defects were made, also by curettage, the growth limiting effects were marked. These studies indicated that there would be a critical amount of physeal tissue needed for growth to continue. Duben and Gelbke showed that two crossed K-wires from the epiphysis through the growth plate to the metaphysis led to growth plate arrest in the dog (93). Gelbke and Ebert also showed that some relatively small transphyseal injuries were still compatible with continued growth, whereas others of greater magnitude led to growth arrest with bone bridge formation (121). A large series of experiments by Campbell and associates involving 175 growth plate alterations in the dog documented variable transphyseal bone bridge formation again dependent on the diameter and physical nature of the defect (51). Of particular interest was the finding that holes drilled longitudinally across the distal epiphyseal plate of the radius and femur, which subsequently were left empty, had maximum growth retardation with transphyseal bone bridge formation when the hole was 88 in. in diameter, but minimal growth retardation when it was diminished in size to 5/32in. When even smaller holes of 0.45 mm in diameter were drilled across the physis, there was no subsequent growth arrest problem. When cortical bone was placed in the defect gap across the physis following drilling of a 5/32-in.-diameter hole, a bone arrest occurred in each of 8 instances, even though that diameter defect, when left to heal on its own, produced only minimal retardation in 13 cases. When the 5/32-in.-diameter drill hole was filled with a smooth Steinmann pin of the same diameter, only minimal growth retardation occurred in 14 cases. When the defect subsequently was filled with a 5/32-in.-diameter threaded metallic pin from the articular surface through the physis into the metaphysis, bone arrest invariably occurred. When growth arrest occurred, therefore, it was due to injury to the epiphyseal plate with the formation of a cancellous bone bridge crossing the defect or occurred when there was effective fixation of the epiphysis to the metaphysis by means of mechanical devices, such as a screw or from bone fusion enhanced by a cortical bone graft. When a smooth pin was placed, however, it prevented transphyseal bone bridge formation and allowed the epiphyseal fragment essentially to slide on it, growing away

548

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

from the metaphysis and not leading to growth arrest problems. Campbell et al. noted that defects across the physis were not repaired by cartilage and that, if the defect was not filled with any substance but allowed to repair on its own, undifferentiated mesenchymal cells flooded the gap and later formed a cancellous bone bridge. If the diameter of the bone bridge was small, the bone continued to grow in length and the bone bridge did not serve any tethering effect. It was also noted that smooth metallic pins of small gauge inserted from the epiphysis into the metaphysis perpendicular to the epiphyseal plate caused less growth retardation than a cancellous bone bridge of equal size. Growth arrest was particularly marked when threaded pins or screws were placed across the epiphyseal plate because they served to mechanically anchor the epiphysis to the metaphysis. Ogden et al. have presented the pathology of acute chondroosseous injury in the child (237). Studies of growth plate fracture models in the rat (127) have demonstrated the sequelae of damage corresponding to the human situation.

C. Pathogenesis of Growth Deformity There are three possible ways in which growth deformity can occur following a growth plate fracture-separation. The pathophysiologic approach is designed to relate to these occurrences at the cell and tissue level. These determinations are based on the extensive investigational and clinical work reported previously. 1. DAMAGE TO BLOOD SUPPLY ON EPIPHYSEAL SIDE One way in which epiphyseal growth plate fractureseparations can lead to growth problems involves damage to the blood supply on the epiphyseal side. As the epiphyseal side vessels are responsible for chondrocyte proliferation and hypertrophy and, thus, trigger the basic growth mechanism of the physis, damage to them brings this process to a halt. With the termination of cell turnover and subsequent cell death, vessels with associated osteoprogenitor cells from the metaphyseal side soon invade the nonfunctioning physeal cartilage and establish transphyseal communication with bone on the epiphyseal side. This can occur over the entire width of the growth plate with massive epiphyseal vessel compromise or locally in less than complete instances of epiphyseal vessel compromise. These mechanisms have been demonstrated experimentally with correlative histologic sections by both Haas (132, 133) and Trueta and Amato (316). The epiphyseal vascular compromise in growth plate fractureseparations is well-documented clinically and tends to occur in intracapsular epiphyses of the proximal femur and proximal radius where the blood supply is quite tenuous. The clinical assumption, backed with extensive clinical examples, is that sufficient blood supply persists after trauma to extracapsular epiphyses because of the more extensive region of blood supply, such that epiphyseal vascular diminution does not occur. It remains extremely difficult to demonstrate par-

tial or focal epiphyseal blood supply loss, and thus it may well be that instances of vessel disruption occur after some injuries to extracapsular epiphyses but simply are not recognized. An article documents epiphyseal vessel disruption in a case of a markedly displaced distal tibial epiphyseal growth plate fracture, even though this is an extracapsular epiphysis (174). Bone scan demonstrated the avascularity well. It will be important in the future to consider focal or regional vascular epiphyseal disruption in high-risk situations, especially when techniques are developed providing higher resolution of vascular supply than current bone scans. 2. INTERMIXTURE OF EPIPHYSEAL AND METAPHYSEAL CIRCULATIONS WITH THEIR ASSOCIATED OSTEOPROGENITOR CELLS

Another mechanism of growth plate abnormality following fracture involves relatively rapid intermixture of the epiphyseal and metaphyseal circulations, which with their associated osteoprogenitor cells allow focal transphyseal bone bridge formation to occur. The bridge serves as a tether minimizing growth of the adjacent persisting physeal tissue. The pathogenesis of such bone bridges in markedly displaced fractures that are allowed to heal with nonanatomic reduction is quite evident and is referred to as a B1 situation. The more common and more worrisome occurrence, however, referred to as a B2 type, involves longitudinal fracturing of the growth plate in association with crushing and fissuring. In such instances no radiographic displacement is seen, although transphyseal vessel communication and subsequent bone bridge formation can occur. 3. DAMAGE TO GROWTH PLATE CHONDROCYTES IN GERMINAL AND PROLIFERATING ZONES BY MECHANICAL CRUSHING LEADING TO DYSFUNCTION AND PREMATURE CELL DEATH

Growth arrests in association with bone bridge formation, although usually rapid, can occur slowly over a several month to year period. Delayed bridge formation may occur by yet another mechanism after a fracture-separation: Epiphyseal vascularity remains intact and separation of the epiphyseal and metaphyseal circulations continues, but the considerable crushing of the germinal and proliferating cell regions of the growth plate leads to mechanical disruption and death of chondrocytes in a portion of the plate. Once the physeal cells die, the stage is set for focal vascular invasion and either bone or fibrous tissue formation. It is this situation that may well lead to the occurrence of a fibrous type bar in the human that then either resolves as it is pulled apart with subsequent growth or turns very slowly over a several month or even few year period into an osseous bridge. We have not been able to produce such fibrous lesions in the rabbit, nor have we seen them referenced in the extensive literature involving damage to growth plates in the experimental animal. The rapidity of response in the rabbit may preclude the occur-

SECTION IV ~ Pathophysiologic Approaches

J In bone

M

cort~Inoge/

M

F I G U R E 15 Possible fracture pathways in what appear radiographically as type III distal tibial epiphyseal growth plate fracture-separations are illustrated. The specific level of transverse fracture through or immediately adjacent to the growth plate cartilage can change the prognosis. At fight the physeal component of the fracture remains totally within the cartilage plate in the hypertrophic zone. Even with the displacement of the fracture through the secondary ossification center of the epiphysis and the articular cartilage, there is still no epiphyseal-metaphyseal vessel communication. This line of fracture would be analogous to that seen in Figs. 14B and 14C on the left. At left the line of fracture is just beyond the hypertrophic zone within the outer reaches of the metaphysis. This is analogous to Figs. 14B and 14C on the fight, although the growth plate orientation is inverted. If there is separation of the epiphyseal fragment, there is communication between epiphyseal and metaphyseal vessels allowing for transphyseal bone bridge formation. Further worsening of prognosis would occur if the growth plate was damaged as seen in Figs. 14E and 14E In the example at left, a more worrisome situation concerning negative growth sequelae exists. The plain radiograph does not have sufficient resolution to delineate the metaphyseal line of transverse fracture when it is this close to the physis. We are beginning to show that MRI studies can make such a delineation. [Reprinted from (283), with permission.]

rence of this lesion that, nevertheless, is an important sequel to some human growth plate injuries. Knowledge of the exact level of the transverse fracture line, which has characterized writings on growth plate injuries from the work of Foucher on, has an important bearing on the possibility of subsequent growth deformity. As we noted in a previous section, although a frequent path of transverse injury is through the hypertrophic zone (132, 281), abundant evidence exists that the transverse injury can involve all levels of the physis (36, 111, 190, 217, 278). One of the eventual advantages of high-resolution MR images or other technologies will be the more exact localization of transverse separation within growth plate cartilage. This is felt to be of prognostic significance because fractures along the germinal and proliferating cell layers can cause mechanical damage to those cells, providing for longitudinal growth. An additional example of the significance of the transverse level of fracture involves the Salter-Harris type III fracture of the distal tibia (Fig. 15). Certain type III fracture-separations have a good prognosis, whereas others are at risk for transphyseal bone continuity. If the fracture line through the epiphyseal growth plate is contained completely within the hypertrophic cell zone, bone block formation should not occur. The fractured bone fragments are intra-epiphyseal and

549

the metaphyseal bone remains in its separate compartment such that there is no epiphyseal-metaphyseal vessel communication. If the fracture line is intra-metaphyseal and the growth plate is displaced slightly, the epiphyseal and metaphyseal circulations with their respective osteoprogenitor cells can come into close association and form a bone block. Plain radiographs do not have sufficient resolution to allow for the subtle but vitally important distinctions concerning the exact transverse level of fracture. They may not demonstrate the presence or absence of longitudinal fractures that allow epiphyseal-metaphyseal vessel communication. All bone bridges do not lead to growth arrests. When the bridge is sufficiently small, forces generated in the adjacent persisting physis may pull apart the bone bridge and minimize or negate any sequelae. This has been shown in both clinical studies and experimental investigations (108, 236). When a bone bridge has formed, the extent and precise location of the bridge must be documented.

D. MR Imaging in Assessment of Growth Plate Fracture-Separations 1. E X P E R I M E N T A L STUDIES A considerable amount of experimental work has been done in our laboratories assessing the effectiveness of MR imaging in relation to growth plate injury. Studies in rabbits have involved the effectiveness of MR imaging (1) in assessing cell and matrix responses to transphyseal focal defects and (2) in assessing the level of transverse fracture in experimental growth plate fracture-separations of the distal femoral and proximal tibial growth plates. When MRI became a practical clinical tool, the possibility of imaging the level of transverse and longitudinal fractures of the cartilage growth plate and transphyseal vascular communication seemed at least feasible. The MRI technique represents a dynamic physiologic indicator because it images soft tissues such as cartilage, blood vessels, and fat as well as a structural indicator differentiating both bone and cartilage. CT scans and tomograms provide more precise resolution of the bony anatomy in particular. To investigate the usefulness of MRI studies in relation to growth plate injury, we performed a series of animal experiments (158).

a. Assessment of Transphyseal Bone Bridge Formation by MR Imaging and Correlative Histology A focal defect was made in the medial proximal tibial growth plate of immature rabbits, which invariably led to transphyseal bone bridge formation (Figs. 16A- 16E). Vessels from the epiphyseal and metaphyseal regions flooded the defect, and mesenchymal cell differentiation led to woven bone and then lamellar bone formation with epiphyseal-metaphyseal bone marrow cavity continuity. Following injury, a series of MRI studies was correlated with plain radiographs, tomograms, and histologic sections. On the day of injury plain radiographs of the growth plate were normal, but MR images clearly showed cartilage damage with absent signal at the

550

CHAPTER 7 "

Epiphyseal Growth Plate Fracture-Separations

F I G U R E 16 A series of MR images and correlative histologic sections from a rabbit model of a focal medial proximal tibial growth plate defect demonstrates the ability of the imaging technique to document bone bridge formation in the early weeks postinjury. (A) MR image immediately after creation of the focal proximal medial tibial growth plate defect in the rabbit model. Signal interruption at the defect site across the physeal cartilage is clearly evident. A plain

SECTION IV ~ Pathophysiologic Approaches

defect (Fig. 16A). The degradation of blood products within the lesion at 1 week registered as increased signal intensity on T 1-weighted, T2-weighted, or gadolinium-enhanced images. Once a bone bridge was formed, the image also varied depending on whether a dense lamellar bar or a marrowfilled trabecular accumulation predominated (Fig. 16B). This work successfully demonstrated the changing transphyseal magnetic resonance images across the defect site associated with the intercommunication of the epiphyseal and metaphyseal circulations and the eventual and invariable bone bridge formation. In the experimental model used, the physeal changes noted on the day of surgery were due exclusively to the drilling. Proton density and T2-weighted images clearly depicted the damage to the cartilage, showing interruption of the cartilage signal at the point of injury at the medial proximal tibial physis. A halo of increased signal intensity on the T2-weighted images appeared to represent edema. There was no contrast material enhancement at the time of surgery. In one animal, the signal intensity in the defect decreased markedly after contrast material enhancement. This was probably due to the T2 shortening effect that predominates at high concentrations of Gd-DTPA, which indicates spillage of contrast material into the lesion because of hemorrhage. The degradation of blood products in the transphyseal component of the lesion and in adjacent hematomas accounts for the increased signal intensity seen on T 1-weighted images obtained 1 week after surgery. The development of early, inhomogeneous contrast material enhancement correlates well with the histologic detection of blood vessel ingrowth in the animals sacrificed 4 days after surgery. The enhancement increases and becomes more homogeneous at 1 week and is maximal at 2 weeks. At this time there is histologic evidence of vascular communication between epiphyseal and metaphyseal vessels. The enhancement decreases and becomes inhomogeneous as bone deposition increases. It is faint and finally disappears as the bone bridge matures and is no longer hypervascular. Deposition of bone by the osteoprogenitor cells accompanying the vessels begins at 1 week and continues thereafter. The ossification in the lesion occurs initially in the periphery, where the newly formed vessels cover the persisting cartilage. This corresponds to the dark rim seen on MR images. At week 5, marrow signal continuity across the physis is dramatic. By week 7, well-formed bone fills the

551

bridge. The signal characteristics of the mature bridge will depend on whether it is composed primarily of dense lamellar bone or trabeculae interspersed with fatty marrow. Changes in the signal intensity from the cartilage not directly related to the drill hole also are seen with MR imaging. The growth plate enhances with GD-DTPA due to diffusion of Gd-DTPA into the cartilage, because the physis is avascular. In one rabbit, significant portions of the plate increased in thickness 1 week after creation of the lesion. This may have represented metaphyseal ischemia, because the interruption of metaphyseal vessels stops ossification. In all the rabbits studied, Gd-DTPA enhancement at the defect site preceded the deposition of bone. The enhancement was distinguishable from the high signal intensity originating from the plate in every case. Thus, enhanced Tl-weighted images can be useful as an indicator of early bony bridge formation. Because we were successful in demonstrating various phases of bone formation histologically in every case, we cannot know whether smaller lesions display enhancement early but never form a clinically damaging bone bridge. MR imaging can be useful in the detection of small disruptions in the growth cartilage and can depict the vascular and osseous changes that lead to formation of a bony bridge. The earlier and most significant developments in this process occur in tissues that are undetectable radiographically. The bone bridge is imaged during the weeks it is actually forming, whereas with plain radiographs it is appreciated only when fully formed and often only after negative growth sequelae, such as shortening and angular deformity, have already occurred.

b. Assessment of Transverse Level of Physeal Fractures by MR Imaging and Correlative Histology MR imaging assessments of the transverse level of fracture within growth plates have also been correlated with histologic studies (287). Experimental physeal injuries were made in distal femoral and proximal tibial growth plates in fourteen 3-month-old New Zealand white rabbits immediately after sacrifice. MR fracture levels were classified along six physeal segments medial to lateral as grade 1 (hypertrophic zone), grade 2 (middle one-third of the physis), or grade 3 (germinal onethird of the physis). The bones were prepared for light microscopy using the JB4 plastic technique. The histologic sections were also graded 1, 2, or 3 in terms of level of transverse fracturing.

F I G U R E 16 (continued) radiograph failed to define the defect because both the physeal cartilage and the defect site are radiolucent. (B) MR image at 5 weeks postinjury shows clear transphyseal signal continuity between the marrow of the epiphysis and metaphysis. (C) Histologic section at 1 week shows early cell response to the transphyseal defect. Persisting physeal cartilage is at right. The blood clot is still present, but mesenchymal cell ingrowth from the epiphyseal and metaphyseal sides is shown (arrows). There is no tendency for physeal cartilage itself to repair the site with chondrocytes. (D) Early mesenchymal cell differentiation to woven bone in the transphyseal region is shown at 2 weeks. There is no tendency to cartilage formation in the gap either from the ingrowing mesenchymal and vascular communication or from the persisting physeal cartilage. (E) Photomicrograph at 8 weeks shows the transphyseal bone bridge, which has now been converted to lamellar bone. The persisting physeal cartilage is shown at the left. The transphyseal MR imaging signal is dependent on the relative amounts of marrow and lamellar bone in the section of the defect being assessed. [Reprinted from Jaramillo, D., et al. (2000). Radiology 15:504-511, with permission.]

552

CHAPTER 7 ~ Epiphyseai Growth Plate Fracture-Separations transverse physeal fracture was shown to be demonstrable by MR imaging and its accuracy was confirmed by histologic study (Fig. 18). This work shows the feasibility of determining the transverse level of the fracture within the physis at the time of initial assessment, a matter that can be of clinical importance in certain physes if it involves large segments of the proliferating and germinal cell layers. 2. C L I N I C A L STUDIES

F I G U R E 17 MR images in parts A and B show how the transverse level of physeal separation can be detected in many instances. (A) A fracture-separation of the femur in an experimental rabbit model assessed by MR imaging shows regions where the fracture line is within the distal femoral physis at left (white curved arrow), through the secondary ossification center bone (two small white arrows), and then at right leaving virtually all the physis on the metaphyseal segment. (B) A type I fractureseparation of the proximal tibia in the rabbit in which all of the physeal cartilage (arrow) is part of the epiphyseal segment. Histologic sections confirmed that the line of fracture was within the hypertrophic zone with essentially all physeal cartilage persisting on the epiphyseal fragment. [Reprinted from Jaramillo, D., et al. (2000). Radiology 15:504-511, with permission.]

In the more horizontal proximal tibial physis, the mean level of fracture by MR imaging was 1.1, indicating a concentration in the hypertrophic zone (mean grade 1.1 _ 0.2; p = 0.02) (Fig. 17A). In the less regular distal femur, the physeal fracture level more often approached the epiphysis in particular in the central angulating portion of the physis (mean grade 2.1 +__ 0.3) (Fig. 17B). When there was MR imaging demonstration of extension into the juxtaepiphyseal segment, histologic study correlated well. The course of

Information and techniques established from the laboratory studies then were applied to a series of 28 human growth plate fracture-separations (159) and an additional 20 human cases (292). MRI was performed in those growth plates in which problematic situations were expected and primarily involved the distal femur, the proximal tibia, and the distal tibia. We have begun to show prognostic evidence of abnormalities that subsequently have gone on to become growth plate defects or to detect evolving growth plate bone bridges earlier than seen with plain radiographs (Fig. 19). MRI has been shown clinically to aid in the assessment of growth plate fracture-separations in four areas. (1) Interruption of the continuity of growth plate cartilage can be demonstrated. These interruptions can be longitudinal fissures, allowing communication between epiphyseal and metaphyseal circulations, or horizontal fractures within the physeal cartilage, the level of which may have a bearing on subsequent growth. Neither of these injuries within the cartilage mass can be detected by plain radiography. (2) The accompanying bone fracture pattern can be demonstrated with much greater resolution than on plain films. Not infrequently, subtle and wandering pathways through adjacent epiphyseal and metaphyseal bone are well-depicted. (3) Bony bridge formation can be seen as it evolves rather than waiting until it is fully formed when plain radiographs demonstrate them. (4) Growth arrest lines, which occur secondarily but clearly point to areas of focal growth arrest, and metaphyseal abnormalities such as avascular segments are revealed earlier and with greater resolution than on plain films. Examples of MR images from some of our clinical cases are presented. Some of these findings are seen within a few days of injury, but others are evident over a several-week to several-month period. We are still in the early phases of interpreting the role of MR in growth plate fracture-separations but expect that advances with clinical MR techniques should enhance resolution capabilities even beyond those presently obtainable. In general, proton density sequences with fat suppression and T2-weighted gradient echo images are best for distinguishing between physeal cartilage and adjacent bone, thereby detecting focal abnormalities in the cartilage such as bony bridges. T2-weighted images are best for distinguishing between the physeal cartilage, which is relatively hyperintense, and the epiphyseal cartilage, which is hypointense. Gadoliniumenhanced Tl-weighted images show an increase in the signal from the physis and, in experimental animals, have shown vascular abnormalities associated with growth disturbances.

SECTION IV ~ Pathophysiologic Approaches

553

FIGURE 18 (Ai)A histologic section of the distal femur whose MR image is shown in Figure 17A shows line of fracture within the physeal cartilage on the right that passes through the epiphyseal bone centrally (two dark arrows and short open arrow) and through the physeal cartilage again on the left. The curved white arrow shows physeal cartilage attached to metaphysis. (Aii) A higher power photomicrograph showing the central intact physeal cartilage (P) above and a fracture line through epiphyseal bone (E) below. [Reprinted from Jaramillo, D., et al. (2000). Radiology 15:504-511, with permission.] (Bi) A histologic section of the proximal tibia (whose MR image is shown in Figure 17B) showing a fracture line separating the hypertrophic zone above from the metaphysis below. This low-powermagnificationshows epiphyseal ossificationcenterbone at top. Upper arrow points to physeal cartilageand lower arrow to metaphysealbone. (Bii) A higher power magnificationshowing the region between the two arrows from Figure 18Bi.

We currently evaluate growth plate abnormalities with MRI by obtaining high-resolution, continuous interleaved thin slices in the coronal and sagittal planes. Coronal T1weighted images identify abnormalities in the hypointense zone of provisional calcification and growth recovery lines. Coronal gradient echo images detect bony bridges as areas of low signal intensity within the substance of the hyperintense physeal cartilage. Sagittal fat-suppressed T2-weighted images separate the physeal cartilage from the adjacent nonossified epiphyseal cartilage and detect abnormalities in the adjacent marrow.

Our first two clinical reports present a preliminary review of our use of MRI to evaluate physeal fractures "at risk" for growth complications in children in the first 4 months following injury. Fractures through epiphyseal and metaphyseal bone are revealed more accurately by MRI than by plain radiography in the acute phase, and MRI scans occasionally allow determination of the transverse fracture level within the cartilaginous physis (if there is some displacement), something that cannot be observed on plain radiographs. This enhanced depiction of subtle fracture anatomy by MRI permits a more accurate analysis and classification of physeal

F I G U R E 19 Plain radiograph and MR images of growth plate fracture-separations and developing sequelae. (A) An anteroposterior knee radiograph in a 1389 male demonstrated, in retrospect, a slight widening of the medial proximal tibial growth plate but no other evidence of fracture. (B) MR image of the same area shown in Fig. 19A identifies a large metaphyseal fragment documenting a type II fracture-separation. This example clearly illustrates the ability of the MR image to detect fracture patterns through adjacent bone that may not be demonstrated on plain radiographs, especially when the standard clinical study in only the anteroposterior and lateral projections is done. (C) MR image of a distal tibia following type I fracture-separation in a 13-year-old female. The physeal region is widened, but of greater importance is the wavy line of light signal through the physeal cartilage indicating the pathway of the intracartilaginous trauma. (D) Sagittal MR image 6 months after a severe type II fracture of the distal radius in a boy 13 years 4 months old demonstrates a lack of cartilage signal centrally replaced by a continuous epiphyseal-metaphyseal signal and an almost obliterated physis in the dorsal one-third, indicating early transphyseal bone bridge formation with bone and vessel communication between epiphyseal and metaphyseal tissues across the physis. 0E) Anteroposterior radiograph of type II distal tibial fractureseparation in a boy 10 years 10 months old. (F) Anteroposterior radiograph demonstrates accurate closed reduction. (G) MR image at 7 weeks shows irregularity in cartilage signal in the central plate area (arrow) suggestive of a developing problem. (H) MR image at 7 months shows the dark growth plate cartilage signal interrupted centrally by a transphyseal signal almost continuous with a signal from epiphyseal and metaphyseal bone. This indicates early bone bridge formation. (I) By 18 months postinjury a much larger bone bridge involving the central one-third of the transverse width of the growth plate has been defined. [Parts A, B, and E - I reprinted from Shapiro and Rand (1992). Adv. Orthop. Surg. 15:175-203, 9 Lippincott Williams & Wilkins, with permission.]

SECTION IV ~ Pathophysiologic Approaches injuries, which should ultimately direct treatment more appropriately as well as provide a more realistic prognosis. MRI scanning in the acute setting (3-10 days after injury) helped to define fracture anatomy in four patients. In three of those patients, the Salter-Harris fracture class based on plain radiographs was found to be incorrect compared to information provided by MRI scans. Although this more accurate designation of the Salter-Harris fracture type did not alter the treatment in these individuals, it provided a better understanding of the potential prognosis of these fractures. MR identification of type III or type IV injuries unappreciated by plain radiographs would have definite treatment and prognostic implications. MRI scans were obtained in 12 patients between 3 and 17 weeks following physeal fractures to assess transphyseal vessel communication and bony bridge formation and, thus, possible early growth arrest. One mechanism for the development of physeal bridges following fracture is the vertical or longitudinal disruption of the physis, which permits mixing of the otherwise separate metaphyseal and epiphyseal circulations. Deposition of bone across a physis has been shown experimentally to start as soon as 7 days following injury. The radiographic diagnosis of physeal bars or bridges may not be possible for months, may require the presence of bony deformity to suggest their existence, and may necessitate further studies such as tomography or CT scanning with additional ionizing radiation exposure to define their location. In this study, MRI scans demonstrated findings not readily apparent by plain radiography, such as physeal narrowing, physeal tethering, and vertical interruption of the growth plate, which in light of subsequent follow-up were highly suggestive of future growth abnormalities. MRI scans are particularly valuable in demonstrating growth arrest lines, which by their presence, absence, or angulation provide important information about physeal function. Harris growth arrest lines are generated by physeal injury, through a temporary slowdown of physeal growth. Upon resumption of physeal growth, thickening or coarsening of the trabeculae serves as a marker, appearing radiographically at the metaphyseal-physeal junction. Growth lines parallel to the physis indicate normal growth, but tilting, angulation, or disruption of an otherwise parallel arrest line suggests physeal abnormality. The association of abnormal physeal arrest lines and traumatic growth plate injuries has been documented previously. Although the physeal arrest lines may appear radiographically as early as 6 weeks after injury, reliable determination of physeal function using these lines often is not possible for 12 weeks after injury. More subtle or partial growth disturbances, especially those causing angular deformity, may not become clinically or radiographically apparent for as long as 18 months to 4 years after injury. The resolution and visualization of arrest lines on MRI scans are significantly better than those on plain radiographs. By 2-3 months after fracture, growth arrest lines were easily identified on coronal and sagittal MRI stud-

555

ies, particularly on spin echo Tl-weighted images. Tilting of arrest lines on MRI scans in several patients was not appreciated on plain radiographs at that stage of recovery. Early physeal bridging also was seen in several MRI studies. Several patients with small focal bridges subsequently had no growth disturbances, which indicates that the growth plate can recover from or "pull apart" small developing or established bony bridges. There is a physeal bar size threshold that the intrinsic physeal recovery capacity cannot overcome, although this appears to vary in different physes at different ages and still must be the subject of study. Due to the marked sensitivity and resolution of MRI, a distinction must be made between transphyseal vessel and new bone formation and clinically significant growth arrest. In some patients, however, the transphyseal bar increased in size on subsequent studies. We caution that the early identification of transphyseal vessel communication mandates repeat MRI studies, rather than immediate surgical excision or application of chondrodiatasis. If the bar increases in size, a growth problem likely will occur. MRI previously has been shown to demonstrate early fibrous physeal bars, which may represent transphyseal vascular communication. In summary, in the acute setting (the first 3-10 days in our study), MRI provides valuable information about fracture anatomy that may more accurately reveal the extent of the injury than plain radiography, even to the point of changing the Salter-Harris class initially designated from the trauma radiographs. During this period, MRI defines cartilaginous injury more clearly than radiography, so that both physeal and articular injuries not appreciated on plain films can be identified. In addition, MRI scans can detect associated soft tissue injuries, such as muscular, tendinous, ligamentous, and meniscal cartilage abnormalities that can be overlooked clinically and radiographically. In the next phase of assessment postinjury, defined as 3-17 weeks, developing bony bars or bridges may be identified or predicted from the MRI assessment of the pattern of physeal disruption. Evaluation of Harris growth arrest lines with MRI before they are apparent on radiographs may disclose developing growth disturbances and allow therapeutic intervention before bony deformity occurs. An additional use of MRI following physeal fractures is to delineate specifically established transphyseal bony bars, usually more than 6 months after injury. In this late period postinjury the advantages of MRI compared to tomography for the mapping of physeal bridges have been documented. 3. CLINICAL USE OF THE PATHOPHYSIOLOGIC APPROACH The pathophysiologic approach can be applied in an increasingly practical way due in part to newer imaging modalities (286). Pathoanatomic classification remains essential because much valuable information has been documented. The pathophysiologic approach is meant to be superimposed on a pathoanatomic classification (Fig. 20). It attempts to

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations

556

PATHOANATOMIC CLASSIFICATION

PATHOPHYSIOLOGIC

( Salter-Harris

)

CLASSIFICATION

- ~--" ~ E essels ntact vessels intact

E vessels intact M vessels intact No E - M c o m m u n i c a t i o n

M

~

= E-M communication

- ~ = E vessels da ag M vessels intact No i m m e d i a t e E - M communication

FIGURE 20 The pathophysiologic approach is meant to be superimposed on a pathoanatomicpattern. One could identify, for example, a S-H type II fracture-separation of the distal femur. Attentionwould then focus on whether one was dealing with a type A, B2, or C situation in regard to events at the cell and tissue level. The pathophysiologiccategorization is given on the basis of epidemiologic features concerning the fracture, an understanding of the literature pertaining to each injury of this epiphysis as to whether significant growth arrests have been documented, the mechanism of injury that might lead to concern about a possible growth arrest, and more specific tomogram, CT, MRI, or bone scan studies to define vascularity and possible longitudinal transphyseal communication. [Reprinted from Shapiro and Rand (1992). Adv. Orthop. Surg. 15:175-203, 9 Lippincott Williams & Wilkins, with permission.]

provide better prognostic considerations for specific injuries by more accurately defining the fracture planes within the cartilaginous growth plate and adjacent bone of the epiphysis and metaphysis. Assessment of vascular and osteoprogenitor cell activities also is necessary. Both of these parameters are beyond the resolution power of plain radiographs. Our approach considers the adequacy of epiphyseal vascularity and the presence or absence of transphyseal vessel communication, which are the two primary determinants of growth postinjury. Use of the pathophysiologic approach is not dependent solely on the application of newer imaging techniques, but is based on a cumulative picture provided by some or all of the following: (1) specific epidemiologic features of each fracture, such as the anatomy of the particular epiphysis fractured, the age of the patient, the force of the trauma, and the displacement that occurred; (2) previous reports concerning the outcome of fractures at each epiphysis; and (3) newer imaging techniques that can provide information with greater pathoanatomic and physiological resolution than that provided by plain radiographs. The modalities currently involve plain tomography (55), CT scanning (55, 80, 103, 260, 320, 334), ultrasound (40, 85), bone scans (140, 174), and MRI (158, 159, 292). Tomograms and CT scans provide clearer resolution of fracture lines through adjacent metaphyseal and epiphyseal bone and enhance the three-dimensional appreciation of the injury, but they are not able to assess cartilage and other soft tissue components. More dynamic physiolog-

ical information is provided by the bone scan, which serves as an index of the vascularity of the fractured segments, and by MRI, which is specific for soft tissue components and thus gives excellent information about the state of the growth plate cartilage and the vascularity of the adjacent marrow in the weeks and months following injury (Table IV). We are not proposing the use of extensive and expensive imaging studies for each growth plate fracture-separation. It is well-known that the incidence of growth problems after fracture-separations of the proximal humeral epiphysis, for example, is minimal. We thus would refer to a common pattern of this injury as a Salter-Harris type IIA fracture, giving the pathophysiologic A on the basis of epidemiologic considerations, literature reviews, and an avulsion mechanism of injury, none of which points to potential long-term problems with this fracture. A different situation exists, however, in terms of Salter-Harris type II, III, or IV fracture-separations of the distal femur. Retrospective studies indicate a fairly high incidence of growth plate damage with each of these injuries. Thus, it is helpful clinically to grade these injuries as A, B1, B2, or C dependent on additional studies. The pathophysiologic approach is used in a dynamic fashion. Initial consideration of a fracture can yield a diagnostic indication of a type IVB~ lesion, which after open reduction and screw fixation is converted to a IVA situation because the epiphyseal blood supply remains intact and we have surgically minimized concerns about epiphyseal-metaphyseal vessel communication. Concern about a IIB2 lesion currently mandates close observation and efforts to document transphyseal vessel communication and growth slowdown, with intervention occurring as warranted clinically. The eventual therapeutic goal would be to convert all clinically worrisome pathophysiologic B1, B2, and C fractures to clinically safe A fractures. This approach thus moves the conceptual and practical management of growth plate fracture-separation beyond the static pathoanatomic designation to the active consideration of events at the cell and tissue level. In the following section, we comment in greater detail on the pathophysiologic approach at the major long bone epiphyses. The pathoanatomic types I-IV represent the Salter-Harris terminology.

V. G E N E R A L C L I N I C A L P R O F I L E OF GROWTH PLATE FRACTURE-SEPARATIONS A. Overview Epiphyseal growth plate fracture-separations account for approximately 15-20% of major long bone fractures in children. They are approximately twice as common in males as in females and approximately three times as common in the upper extremity as in the lower extremity. In those studies that accurately document all physeal fractures, it is those of the hand and in particular the phalanges of the fingers that

SECTION V ~ General Clinical Profile

TABLE IV

557

Pathophysiological Approach to Epiphyseal G r o w t h Plate F r a c t u r e - S e p a r a t i o n s

I.

Pathoanatomic diagnosis Made by plain radiographs II. Pathophysiologic approach Superimposed on pathoanatomic classification to yield A, B1, B2, or C categorization based on following considerations: 9 Fracture profile Age of patient, specific epiphysis injured, anatomy of epiphysis (intracapsular-extracapsular; linear or convoluted shape), amount of displacement 9 Literature review Reported correlation of outcome with pathoanatomic type for specific epiphysis injured 9 Mechanism of injury Extent or trauma; avulsion (epiphyseal separation), compression (epiphyseal fracture) Additional investigations (if warranted in problematic fracture-separations): 9 Adjacent epiphyseal-metaphyseal bone fracture pathway: tomogram, MRI, CT scan 9 Longitudinal growth plate cartilage fracture with epiphyseal-metaphyseal vessel and bone cell communication: MRI (early); tomogram or CT scan (once bone bridge formed) 9 Epiphyseal vascularity: bone scan, MRI 9 Level of fracture within growth plate: MRI 9 Nonossified epiphyseal displacement: ultrasound 9 Growth arrest lines: MRI most sensitive III. General aims of pathophysiologic approach 9 Perform specific treatment to convert B1, B2, or C type fractures to type A, if possible 9 Reveal negative growth plate sequelae early to allow intervention to minimize degree of shortening and angulation 9 Consider events at the cell and tissue levels as well as at plain radiographic level

are the most common. In the major long bones, the fracture of the distal radial epiphysis is most common in all series, with injuries of the distal tibial epiphysis and distal humeral epiphysis next in frequency. The age at occurrence documents peak incidences of female fractures approximately 1.5 years earlier than those in males. The large majority of epiphyseal growth plate fracture-separations occur between 10 and 15 years of age, although fracture patterns at specific epiphyses tend to have characteristic age ranges.

B. Distribution of Physeal and Nonphyseal Fractures in Childhood Physeal fractures constitute approximately 15-20% of all childhood fractures, but studies as high as 30% have been reported (67, 138, 203, 216, 232, 280, 333). In assessments of childhood hand fractures, 34% involve the epiphyses, which represents a higher percentage than in major long bones (141). Data establishing the percentages of physeal and nonphyseal injuries are somewhat variable due to several factors, including the number of patients in a particular study, the upper age limit of patients included in the study, the referral pattern of the institution reporting the profile, and the inclusion or noninclusion of hand fractures. Sakakida reported on the frequency of affected sites of epiphyseal fracture-separations and nonphyseal fractures of the long bones under 16 years of age (excluding hand and foot fractures) (280). He documented a 12.7% incidence of

epiphyseal fracture-separations, with 117 epiphyseal fractures and 800 metaphyseal and diaphyseal long bone fractures. Hanlon and Estes noted a 17% incidence of epiphyseal fracture in 698 patients suffering fractures from birth to 18 years of age (138). Ogden indicated that approximately 15% of all fractures in children involve the physis, a number derived from reference to a large number of articles in the literature over a period of several decades (232). Lichtenberg documented only a 6.5% incidence of epiphyseal fractures in 2058 childhood fractures, but his study included children only to the end of the 12th year of age (194). Compere reported a 14.4% incidence of epiphyseal fractures in 211 children under 15 years of age (67). Three additional detailed studies specifically documented the incidence and pattern of fractures including those of physeal and nonphyseal regions. Worlock and Stower assessed all children 12 years of age and younger who sustained a fracture over a 6-month period in Nottingham, England (333). They documented an 18.5% incidence of physeal injuries, involving 171 epiphyseal injuries in a total of 923 fractures. This study did include hand and foot injuries, but included children only 12 years of age or less. The peak concentration of growth plate fractureseparations is between 10 and 15 years of age. Mizuta et al. noted an incidence of 17.9% physeal injuries based on a study of all acute fractures over a 24-month period at the Adelaide Children's Hospital, the center for pediatric care in the Adelaide metropolitan area of Australia (216). They noted 353 physeal injuries in a total of 1974 fractures. Their

558

CHAPTER 7 ~ Epiphyseal Growth Plate F r a c t u r e - S e p a r a t i o n s

TABLE V

Frequency o f G r o w t h Plate F r a c t u r e - S e p a r a t i o n s a t Major Long Bone Epiphyses ~

Series A

D. radius D. tibia D. humerus D. fibula P. humerus D. ulna D. femur P. tibia P. radius P. ulna P. femur P. fibula n--

Series B

Series C

Summary values A-D

Series D

No.

Percentb

No.

Percent

No.

Percent

No.

Percent

No.

98 59 20 21 22 12 18 6 1

37 22 8 8 8 5 7 2 0.4

47 15 11 6 3 7 0.46 2 7 0.46

3

35 18 17 5 8 3 5 2 2 1 3 1

100 33 24 12 7 16 1 4 16 1

7

114 60 56 15 27 11 17 8 5 3 9 2

170 104 37 68 18 27 13 8 6 4 1 1

37 23 8 15 4 6 3 2 1 1 0.2 0.2

482 256 137 116 74 66 49 26 28 8 17 3

264

327

214

457

aCode: n, No. = number; series A, Peterson and Peterson, 1972; series B, Ogden, 1981" series C, Mizuta et Peterson et al., 1994. bpercent based on value nearest to full number from 0.5% and upward.

study also included hand and foot injuries. Mann and Rajmaira documented a 30% incidence of growth plate injuries in childhood fractures, which is the highest percentage in the literature (203). Their study involved records of patients treated in the pediatric emergency room at Cook County Hospital in Chicago over an 1 I-year period. The study did not include hand or foot fractures, but did follow children from birth to 16 years of age and included 943 growth plate fracture-separations out of a total of 2650 fractures.

C. Incidence of Epiphyseal Growth Plate F r a c t u r e - S e p a r a t i o n s in Males and Females There is an approximately 2:1 male:female incidence in epiphyseal growth plate fracture-separations. Mizuta et al. documented 231 fractures in males and 122 fractures in females (1.9:1.0) (216), Peterson et al. a 637 to 314 male: female incidence (2.01:1.0) (253), Mann and Rajmaira a 505 to 227 male:female incidence (2.22:1.0) (203), and Ehlers and Eberlein a 75:28 male:female incidence (2.68:1.0) (97).

D. Incidence of Epiphyseal Growth Plate F r a c t u r e - S e p a r a t i o n s at Specific Epiphyses The percentage distribution of growth plate fractureseparations throughout the body is greatly dependent on whether the assessment included hand, and to a lesser extent foot, fractures and also the extent of the study in relation to those two regions. Most of the studies of epiphyseal fracture-separations published prior to 1970 concentrated

al.,

Percent 38 20 11 9 6 5 4 2 2 1 1 0.2 1262

1987; series D,

almost exclusively on the major long bones. Four major studies from 1972 on included hand and foot injuries and clearly demonstrated how common they are. In 1972 Peterson and Peterson indicated that finger phalangeal and metacarpal fractures comprised 14.8% of all injuries (248). In 1981 Ogden showed that these hand fractures involved 11.1% of all injuries (232), and in 1987 Mizuta et al. showed that they involved 30% of all growth plate fracture-separations (216). In the most recent study of 1994, Peterson et al. at the Mayo Clinic showed their involvement in 43.8% of all fractures, with 37.4% involving the finger phalanges and 6.4% the metacarpals (253). Table V indicates a compilation of major long bone epiphyseal growth plate fracture-separations from several papers, excluding hand and foot involvement. Almost 70% of long bone growth plate fractures occur at only three areas: distal radius, 38%; distal tibia, 20%; and distal humerus, 11%. The two most common fracture areas in the study by Ehlers and Ebeflein were also the distal radius (38%) and the distal tibia (36.2%) (97). Examples of more unusual epiphyseal fracture-separations were indicated in some papers but not in others, such as those involving the trochanters of the proximal femur or the tibial tubercle apophysis,

E. Age at Occurrence of Physeal Fractures The large majority of epiphyseal growth plate fractureseparations occur between 10 and 15 years of age, although the entire age spectrum is involved from the newborn to those in the process of undergoing final growth plate fusion.

SECTION V ~ General Clinical Profile

559

TABLE Vl Distribution o f Physeal Fractures by Type a ( S a l t e r - H a r r i s Classifications) Year

Author (Reference)

I

II

III

IV

V

Other

Total

1933 1970 1974 1979 1986 1987 1990 1994 Total Percentage

Bergenfeldt (23) Rogers (274) Oh et al. (238)b Mbindyo (206) Worlock and Stower (333) Mizuta et al. (216) Mann and Rajmaira (203)b Peterson et al. (253)

23 7 34 18 30 30 210 126 448 15.6

251 89 73 42 121 257 483 510 1705 59.2

19 9 14 4 5 23 143 104 316 11.0

13 12 12 5 15 42 102 62 248 8.6

0 1 0 2 0 1 5 0 9 0.3

4

310 118 133 71 191 353 943 951 2879 100

149 15 5.3

alncidence series of only one anatomic site not included. Includes only humerus, radius, ulna, femur, tibia, and fibula. bReprinted from Peterson, H (1994). J Pediatr Orthop, Vol. 14, No. 4.

Peterson e t al. documented the age incidence of physeal fractures in males and females from 1 year of age to skeletal maturity (253). The incidence rates were greatest for boys at age 14 with the largest number occurring between the ages of 10 and 15 years, whereas in girls the peak years of incidence were at 11 and 12 years with the ages from 9 to 13 years showing the largest number of fractures. When the fractures at each specific epiphysis are assessed, some clinically meaningful differences are shown. Mann and Rajmaira also noted that female patients tended to have physeal fractures on average about 1.5 years younger than male patients with the same type fracture in the same location (203). In their studies, the mean patient age for most physeal fractures was 10-12 years. Ehlers and Eberlein noted peak age for males at 14 years and for females at 12 years (97). Mizuta e t al. noted the most common occurrence of epiphyseal fracture-separation in boys at age 12 years, with the most common occurrence of injuries from ages 8 to 15 years, and in girls age 11 years was the most common year of occurrence with most of the fractures occurring between 7 and 13 years (216). They also calculated the ages of occurrence between 0 and 5 years of age (infancy), 6 and 11 years of age (preadolescence), and 12 and 17 years of age (adolescence). The percent of physeal fractures in relation to nonphyseal fractures increased at each time period. The results were the same in males and females. Of the fractures that occurred in infancy, 9.3% were physeal, in preadolescence, 15.7% of fractures were physeal, and in adolescence, 29.9% of fractures were physeal. Peterson and Peterson (248) and Mann and Rajmaira (203) broke down the age of the fractures at each specific physis in both males and females. Fractures of the distal humerus occurred at a much younger age than those of the other physes, with girls showing a mean of 4 - 5 years and boys 5 - 8 years (Peterson and Peterson). In the most common distal humeral fracture, the type IV lateral condyle fracture-separation, Mann and Rajmaira also docu-

mented a relatively younger age, with 41 boys averaging 8 years of age and 14 girls averaging 5.3 years of age.

F. Distribution of Specific Salter-Harris Types per Long Bone Region Peterson e t al. have summarized the distribution of physeal fractures by type in each of several studies (Table VI) (253). Even here, the findings in several series over a 60-year period show similar pathoanatomic patterns. Virtually 60% of fractures are of the type II pattern with an additional 15% type I. Type III fractures account for 11% of injuries and type IV 8.6%. Many authors have failed to recognize the occurrence of the type V lesion, whereas others have diagnosed it on occasion but the collected series document only a 0.3% incidence. Peterson, in a pathoanatomic classification, pointed out that a considerable number of fractures do not fit into the Salter-Harris classification, and for that reason he deleted type V and added two additional types (255).

G. Epiphyseal Fracture-Separations with Difficult Births Epiphyseal fracture-separations accompanying traumatic births were well-recognized in the late 1800s in the preradiographic era as described by Simpson and others (291). Due to some difficulty in diagnosis, however, they continued both to occur and to result in delayed recognition even after the onset of radiographs. Truesdell was instrumental in pointing out not only the continued occurrence of these injuries but also ways to recognize them both clinically and radiographically (314). By 1918, he had described 11 cases from his own experience involving 5 epiphyseal separations at the proximal humerus, 3 at the distal humeral epiphysis, 2 at the distal femoral epiphysis, and 1 at the proximal femoral epiphysis. Traumatic neonatal joint dislocations did not

560

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

occur, but due to the major point of weakness at the physeal level extreme trauma would be associated with epiphyseal fracture-separations. These almost invariably were associated with difficult births, most of which were from the breech position or in those instances in which forcible version had been performed. Footling presentations due to the difficult delivery that followed also presented a high likelihood of epiphyseal separation. The characteristic clinical profile therefore involved a difficult and often prolonged delivery characterized by breech or footling presentations, swelling in a joint region and disinclination of the infant to use the limb, and radiographic studies showing no metaphyseal or diaphyseal fractures. As only the distal femoral and proximal tibial secondary ossification centers are present at birth with those of the proximal and distal humerus and proximal femur being absent, plain radiographs often were indistinct in terms of diagnosis. The diagnosis frequently was suspected when radiographs at 7-14 days showed evidence of new bone formation in relation to periosteal elevation. Clinically the infants gradually began to use the affected extremities, and often to clinical exam there was extreme swelling and hardness in the epiphyseal areas. Truesdell stressed, in agreement with opinions today, that operative replacement of a displaced epiphysis was rarely, if ever, indicated and that even when a displacement had been wide and not recognized for some time the outcome had been good. He stressed that "no better evidence of the restorative and corrective ability of Nature during the early years of life can be obtained than in these cases." Ekengren et al. reported on 20 patients with birth injuries to the epiphyseal cartilage and also showed predominant involvement of the upper extremity bones (98). Their series comprised epiphyseal fracture-separations of the proximal humerus in 9 cases, distal humerus 5, proximal femur 1, distal femur 5 (4 patients), and distal tibia 1. They also noted difficulty primarily with breech deliveries, which were reported in 14 cases with 2 other cases having complicated deliveries in other ways. They also concluded that every newborn infant with pain, swelling, or impaired movement of a limb or joint should be suspected of having an epiphyseal fracture-separation and be examined radiographically. At present, ultrasound has been extremely useful in assessment of these disorders. Broker and Burbach described use of ultrasonography in a newborn proximal humeral epiphyseal separation (40). Two cases of separation of the proximal humerus with a review of the diagnostic criteria as listed previously were presented by Lemperg and Liliequist (191). VI. C L I N I C A L F E A T U R E S O F A C U T E

EPIPHYsEAL FRACTURE-SEPARATIONS A. General Principles of Management The pathoanatomic classifications describe the patterns of breakage in epiphyseal growth plate fracture-separations

and are useful in the management and reporting of such injuries. Those of Aitken, Salter and Harris, Ogden, and Peterson are sufficiently detailed that they have been and will be used as reference points for clinical studies. If one superimposes the pathophysiologic classification upon any of the pathoanatomic classifications, an approach centering on consideration of such injuries at the critical cellular level ensues (283, 286). In assessing these lesions, a three-fold approach is assumed: (1) A knowledge of the gross and histologic anatomy of the growth plate involved provides an expectation of which fractures are likely to produce significant problems. Fracture-separations involving displacement of the proximal femoral capital epiphysis and the proximal radial epiphysis, for example, both of which have a relatively tenuous blood supply, have a poor prognosis with concern about a type C pattern on this basis, as is borne out by the reported results. Considerations of growth plate shape, capsular relation, and severity of trauma are also important. (2) The effects of fracture-separations classified by pathoanatomic type at each specific epiphysis are now reasonably wellknown. Knowledge of the accumulated results allows for a good level of awareness concerning which fracture at each specific epiphysis is likely to lead to problems. (3) The most accurate predictions of the outcome follow from consideration and demonstration (at the time of injury and following treatment) of (a) the presence or absence of damage to the epiphyseal blood supply, (b) vertical fissuring and transverse separation of the growth plate, (c) the exact level of the transverse fracture line (in relation to the proliferating zone of cartilage, the hypertrophic zone, or metaphyseal bone), and (d) the extent of communication between the epiphyseal and metaphyseal circulations. These are the factors that are most important, rather than the pattern of breakage as demonstrated radiologically. On the basis of anatomic and empiric knowledge, a very high percentage of excellent results may be expected, for example, with distal radial type II fracture-separations, and specific investigation of the four criteria just listed would be unnecessary. For those epiphyses where the pathoanatomic type of injury can lead to varying results, as in type II injuries of the distal femur, an awareness of the pathophysiologic approach can lead to more intensive investigation and perhaps in time to more specific therapy. It is expected that in the next few decades further refinements of diagnostic techniques will permit early assessment of some or all of these parameters, thus allowing the precise delineation previously attainable only by histologic studies and retrospective review. The current use of bone scanning to assess vascularity of the femoral head after trans-epiphyseal lesions is an example of this approach. We have attempted to show the value of investigating the sequelae of injury at the biologically important cell and tissue levels by use of newer imaging modalities to provide three-dimensional anatomic and physiologic information. The possibility of early diagnosis of a developing problem has been greatly improved by the use of bone scan in relation to epiphyseal vascular injuries, tomograph and CT scanning

SECTION VI 9 Clinical Features o f Acute Epiphyseal Fracture-Separations

to allow a more accurate pathoanatomic designation of the pathway of fracture within epiphyseal and metaphyseal bone, and MRI, which is beginning to provide an indication of the level of fracture within bone and the cartilage growth plate and vascular communication across the plate. The pathophysiologic approach refers not only to classifications of the injuries but also to a more biological treatment of them. At the time of fracture, early recognition of a potential growth arrest problem allows for treatment, usually operative, directed at changing, if possible, a B1, B2, or C situation to a more favorable A type. In displaced type IV lesions, which in the pathophysiologic approach appear as type B1, open operative reduction converts the B~ lesion in many instances to a type A lesion. The open reduction brings into anatomic alignment the articular cartilage surface, the epiphyseal bone, the growth plate, and the metaphyseal bone, thus minimizing the long-term negative growth sequelae. The two areas in which we feel the pathophysiologic approach will have the greatest potential for change relate, however, to the B2 and C type lesions. The situation of the B2 lesion in which there is radiologically undetectable longitudinal crushing and fissuring of the growth plate, thus allowing the epiphyseal and metaphyseal circulations to communicate with each other, has been accepted previously as a part of the injury about which nothing could be done. Reviews of epiphyseal growth plate fracture-separations in which such lesions occurred are filled with examples of premature growth arrests, significant shortening, and major angular deformities. We are beginning to see that this situation can be changed and the negative sequelae at least minimized if not eliminated. The pathophysiologic approach utilizing MRI studies and other modalities may provide very early indication of a developing bone bridge. With close observation of the affected limb one can resort to complete fusion of the growth plate to minimize angular deformity and eliminate the need for corrective osteotomy or apply additional responses to bone bridge formation that allow growth to continue. The pathophysiologic approach to epiphyseal growth plate fracture-separations is outlined in Table IV.

B. Proximal Humerus Proximal humeral growth plate fracture-separations are almost exclusively type I and type II injuries [Aitken (8); Baxter and Wiley (19); Broker and Burbach (40); Burgos-Flores et al. (45); Dameron and Reibel (76); Kohler and Trillaud (179); Neer and Horwitz (223)]. Other types of fractureseparations at this epiphysis are extremely rare to the point of not being described in most series. The metaphysis virtually always is displaced anteriorly. Because the apex of the inverted, V-shaped growth plate is superior with a trough along the anteroposterior axis, permanent damage to this irregularly shaped plate is not great; the large metaphyseal fragment left behind at the posteromedial part of the epiphysis also spares damage to the plate. Angular deformity has not been of clinical significance due to the excellent remod-

561

eling potential of the proximal humeral epiphysis, which accounts for 80% of the growth in length of the bone, and the fact that the shoulder joint has such a wide arc of movement that a perfectly anatomic glenohumeral articulation is not required. Shortening is documented in a fairly high percentage of cases, although this is rarely of clinical significance because it is minimal (less than 2 cm) and because mild length inequality of the upper extremities is not important. Dameron and Reibel noted shortening in 30%, with 15% having 1.5 cm or more involvement (76). Neer and Horwitz documented 16% with 1 cm or more shortening in their entire series (223). In mildly to moderately displaced fractureseparations they noted 11%, and in severely displaced lesions 30% with such shortening. Aitken described a 64% incidence (7/11) of shortening, all in nonoperated fractures, with a range from 0.25 to 1.5 in. (8). Some of the shortening in these cases was felt to be due to malposition with imperfect remodeling, but Aitken commented on premature fusion due to growth plate damage associated with severe trauma and an irregularly shaped plate, even though he noted all injuries to be types I and II (by S-H criteria). The late age at injury also limits the extent of discrepancy. Shortening in those less than 11 years of age at the time of injury was not seen, probably due to the relatively greater thickness of the plate. Baxter and Wiley reported on 57 fractures and noted an age range at occurrence between 8 years 11 months and 15 years 1 month. There was more than 1 cm of shortening in 30% (9/30) with a maximum of 2 cm and insignificant angulation regardless of method of treatment or age at fracture. Excellent results were seen regardless of the degree of displacement or the age at injury. Open reduction did not seem to be warranted because results were so good even when healing was allowed to occur with virtual full displacement. Kohler and Trillaud described 49 cases of epiphyseal growth plate fracture-separation of the proximal humerus in the course of reviewing 136 proximal humeral fractures (179). They too defined only type I and type II fractures, with type I seen in 16.3% and type II in 83.7%. The epiphyseal separations were seen throughout the childhood years; 45% of the fractures occurred between 10 and 14 years of age and 76% occurred between 8 and 16 years of age. Although 45% of the patients (22/49) had surgical intervention, the authors reached the conclusion that surgery was only of marginal benefit and thus indicated infrequently. Length discrepancy was not a problem, being mild to moderate in the 1-cm range with the greatest amount of shortening being 3 cm. In those five cases of epiphyseal separation in which some subsequent growth injury could be identified, four occurred after surgical treatment and only one after nonoperative treatment. The type I epiphyseal separations were seen in all ages, whereas the type II were concentrated in older children. They commented that "the function of these patients is always good or very good regardless of the anatomy of the lesion or the treatment." The appearance of the scar was felt to be a particularly negative aspect of surgical intervention, and they commented that "the appearance of the scar

562

CHAPTER 7 ~ Epiphyseal Growth Plate Fract;ure-Separarions

is very often unattractive" due to its hypertrophic or keloid characteristics. They also raised the possibility that reduction, whether open or closed, and open surgical intervention with internal fixation were as capable of inducing growth damage as they were of minimizing it. They concluded that "surgical treatment presents no advantage over non-operative treatment on the basis of functional results or long term anatomical results" and that "nonoperative treatment should be recommended with surgery being discussed only in special cases." Anatomical reduction was not felt to be essential. Surgery would be considered primarily for major displacement when there was virtually no growth remaining, and even then the simplest form of fixation should be used. Burgos-Flores et al. were somewhat more supportive of a surgical approach with open reduction (45). They analyzed 22 patients who had marked displacement of the proximal humeral fracture-separation from a total group of 65 patients. The group that was more displaced underwent closed or open reduction and Kirschner-wire fixation. All the epiphyseal injuries were of the Salter-Harris type II variety. Following closed reduction, two or three percutaneous Kirschner wires were passed from the lateral cortex of the metaphyseal fragment across the physis into the humeral head. Open reduction was required in three. Two patients suffered final length discrepancies of 4.0 cm (one having had an open reduction and the other closed) with virtually all other patients showing either no discrepancy or discrepancies less than 1.5 cm. It was concluded that patients under the age of 13 years should be treated conservatively regardless of degree of displacement or postreduction angulation. They felt, however, that in certain patients over the age of 13 years, if marked displacement or angulation were seen on initial radiographs, accurate reduction should be obtained usually involving closed reduction and maintenance with K-wire fixation, but on occasion warranting open reduction. Because 80% of humeral growth occurs at the proximal end and fracture at this epiphysis is relatively common, the absence of reports of extensive growth discrepancies is convincing evidence of the benign nature of this injury. The type I and type II injuries are graded, therefore, as A lesions in the pathophysiologic classification. There is no reported evidence of vascular problems involving epiphyseal blood supply. The epiphysis is mainly extracapsular assuring persistence of an excellent blood supply after injury, although parts of the epiphysis are intracapsular. There may be some occasional B2 type injuries, but these occur toward the end of growth and as noted previously do not cause clinical problems. Other than the described use of ultrasound (40) or arthrography to diagnose neonatal fractures, specific radiologic measures other than plain films virtually should never be required with this injury. The vast majority of proximal humeral epiphyseal fracture-separations are treated by conservative means using either a sling alone in nondisplaced or minimally displaced fractures or closed manipulation followed by a sling or long arm cast in those cases that are more

markedly displaced. Immobilization in the abducted position in an upper extremity spica, open reduction, or internal fixation do not seem to be warranted, except with neurovascular compromise or marked deformity in those at or near skeletal maturity. The remodeling potential of the proximal humerus is great because 80% of the growth of this bone occurs at this region. Figure 21 illustrates a type II proximal humeral fracture-separation managed by nonoperative means.

C. Distal Humerus 1. LATERAL HUMERAL CONDYLE The lateral humeral condyle fracture is a type IV fractureseparation. This injury was clearly recognized and diagramed by Poland in 1898 (258). It represents a clear-cut example of the value of operative intervention in particular epiphyseal growth plate fractures with significant beneficial results documented [Badelon et al. (15); Conner and Smith (68); Crabbe (72); Foster et al. (110); Hardacre et al. (139); Jakob et al. (156); Thompson et al. (309)] (Table VIIA). Epiphyseal injuries in the elbow region were recognized in the 1930s to be especially problematic and often poorly responsive to closed reduction and cast techniques (99, 154). With the exception of undisplaced fractures or fractures only minimally displaced by 2 mm or less, it is widely accepted that open reduction and K-wire fixation be done for all of these injuries. Closed reduction followed only by cast immobilization is not recommended because imperfect reduction or subsequent loss of reduction by slippage can lead to severe growth disturbances. Review of more than 125 cases from the literature indicates that open, anatomic reduction with K-wire fixation or periosteal suturing is mandatory for the displaced fracture, and if done the likelihood of an excellent result is very high in the 90-95% range. Malunion, nonunion, avascular necrosis, and deformity due to premature epiphyseal fusion appear to be related almost exclusively to closed reduction with imperfect positioning or delayed or imperfect operative intervention. The open reduction with anatomic repositioning converts a type IVB1 to a type IVA fracture. It reconstitutes the smooth intra-articular cartilage surface, eliminating the risk of degenerative arthritis, and relates epiphyseal bone to epiphyseal bone, growth plate cartilage to growth plate cartilage, and metaphyseal bone to metaphyseal bone, eliminating postfracture growth sequelae. Flynn has commented on his extensive experience with these injuries (106). He described instances of nonunion of minimally displaced fractures in the lateral condyle and advocated open reduction and internal fixation of any fracture displaced greater than 2 mm. He reported on several cases of nonunion of the lateral condyle over a period of 12 years that he had either seen personally or corresponded about with other surgeons. He noted that "the anatomic cause of nonunion is instability of the distal fragment but the most common cause is inadequate treatment of the fresh fracture." The reasons for inadequate treatment were failure to present

SECTION VI ~ Clinical Features of Acute Epiphyseal Fracture-Separations

563

FIGURE 21 Proximalhumeral epiphyseal fracture-separation is shown. (A) Radiograph demonstrates a type II proximal humeral fracture-separation with moderate displacementin a boy 12 years 3 months old shortly after injury. Literaturereviews of these fractures so overwhelmingly demonstrate a benign postinjury course that the pathophysiologic A grading can be assigned without the use of additional imagingprocedures. (B) Anteroposteriorradiograph at approximately4 monthspostinjury shows healing and metaphyseal remodeling medially mediated by periosteal new bone formation. (C) Anteroposterior radiograph at 1 year 2 months postinjury showing completehealing and advancedremodeling. [Parts A and C reprinted from Shapiro and Rand (1992). Adv. Orthop. Surg. 15:175-203, 9 Lippincott Williams & Wilkins, with permission.]

to a physician (3), failure to diagnose the fracture at initial assessment (3), failure to recognize displacement and failing union during the course of management (12), and failure to pin the fracture when there was greater than 2 mm of displacement (3). Late criteria for successful treatment of a nonunion of a lateral condyle fracture are that the fragment is in an acceptable position and the physis of the fragment is still open. The danger of not treating is that the physis of the condylar fragment will grow imperfectly or close prematurely and lead to progressive valgus deformation. The negative end result is the valgus unstable elbow with ulnar neuropathy. A posthealing fishtail appearance of the distal humerus is of no clinical consequence because it causes no symptoms. It represents a central physeal disturbance of growth at the junction between the fractured lateral condyle and the more medial distal humeral physis. Extensive studies continue to be published in relation to these fractures. There is increasing delineation of the grades of fracture, by which is meant the degree of displacement at the time of treatment, and clearer indications of the sequelae. Badelon et al. have defined four grades of this fracture, although all represent a type IV categorization (15). In A lesions (their terminology), the fracture is non-displaced and can be seen as a crack on only one radiographic view. In B lesions there is a visible fracture line with minimal displacement less than 2 mm and generally distinguished by a slide in the lateral cortex; in C there is displacement of greater than 2 mm on all radiographic views and in D major dis-

placement with complete separation of the fracture edges. They reviewed 47 patients. In 10 cases of the first group, all results were excellent and were achieved with either long arm casts or long arm casts with percutaneous pinning. Results were progressively farther from perfect in the remaining three groups. The basic recommendation was to perform open reduction and Kirschner-wire fixation, and they came to the belief that "open reduction and internal fixation are essential" in anything other than a simple non-displaced fracture with less than 2 mm of separation. It is important to preserve condylar vascularization if open reduction is done. There were examples of axial deviation of greater than 10~ in either valgus or varus, which they felt corresponded to incorrect positioning of the condyle at the time of either fracture or immobilization. Although the injuries were type IV, in most cases the fracture line crossed the cartilaginous epiphysis of the distal humeral segment but did not touch the secondary ossification center of the capitellum, minimizing the risk of epiphysiodesis. They went on to note, however, that "nevertheless, we have observed 8 cases of lateral condyle epiphysiodesis without clinical manifestations. The epiphysiodesis mainly concerned the medial part of the lateral condyle." The epiphysiodesis was "relatively central," which "explains the absence of axial deviation and the fishtail aspect of the lower extremity of the humerus after growth is completed." Foster et al. also concurred that closed treatment resulted in satisfactory results if the initial displacement did not exceed 2 mm (110). Percutaneous pinning of non-displaced or

T A B L E VII Author (Cases)
Salter-Harris classifications

Effects on growth

A. Lateral Humeral Condyle Fracture-Separations Jakob et al. (26) (156)

IV 26

In 11/11 patients treated within 8 days of fracture by open reduction and K-wire fixation, the results were excellent with no growth arrest reported. All the poorer results were treated 3 weeks or more after injury. The complications involved premature epiphyseal fusion, malunion, nonunion, and avascular necrosis.

Crabbe (17) (72)

IV

Results were excellent in 16/17 cases treated with open reduction and catgut suture or K-wire fixation. One poor reductioni (10 ~ varus) due to faulty technique. No avascular necrosis, nonunion, epiphyseal arrest, or clinically significant deformity in all other patients.

Conner and Smith

IV 39

Displaced fractures all treated with open reduction, screw fixation. Results almost uniformly excellent except for one nonunion, and only two with mild valgus deformity suggestive of "some interference with epiphyseal growth." "We did not find the development of premature epiphyseal fusion."

IV all

Open reduction internal fixation in 23 yielded 15 excellent, 3 good, and 5 poor results. The 5 poor results were all felt to be technical failures. Closed reduction yielded many fair-good results. "Nonunion, significant angular deformity, avascular necrosis and tardy ulnar palsy are complications that are avoidable."

17

( 3 0 ) (68)

Hardacre et al. (52)( 139)

B. Proximal Radial Fraction-Separations Jeffrey (24) (16~

No differentiation between epiphyseal and metaphyseal fractures

Premature epiphyseal fusion in 2/2 cases with head rotated posteriorly 90 ~ in 3/3 cases with head tilted 60-90 ~ and in 6/18 (33%) with head tilted 30-60 ~ The greater the displacement the worse the prognosis due to "risk of the blood supply to the displaced head particularly in those cases treated by operative reduction."

Tibone and Stoltz (7)(TM)

II 5 III 1 IV 1

Seven epiphyseal fracture-separations and 24 neck fractures in this series of proximal radial fractures in children; 5/5 type II injuries with excellent results but none of the five were displaced. Fair-poor results in the type III and IV lesions.

Jones and Esah (17)( 162)

II

Seventeen epiphyseal fracture-separations and 17 radial neck fractures in this series of proximal radial fractures in children. Results assessed for entire series; not broken down by level of fracture. Many instances of avascular necrosis, malunion, radial head enlargement, and deformity. Results: good 13, satisfactory 10, poor, 11. End result correlated very well with the degree of initial displacement; if 50 ~ or more, 15/22 (68%) with satisfactory or poor rating. 10/24 (42%) with marked displacement (60-90 ~ had premature epiphyseal fusion.

Reidy and Van Gorder ( 3 0 ) (269)

Newman ( 4 8 ) (225)

17

No numerical distinction between epiphyseal and metaphyseal fractures. All epiphyseal fractures were I or II. I 1 II 24

Twenty-five epiphyseal fracture-separations and 20 metaphyseal neck fractures. Avascular necrosis (9); premature fusion of epiphysis (24); abnormal radial head (20). Level of fracture (growth plate versus neck) had only a "minor bearing on the outcome" although plate injuries had slightly better results. Results of growth plate injures: good 14, fair 5, bad 6. Open reduction risks increasing the avascular necrosis.

C. Distal Femoral Fracture-Separations Stephens and Louis (20)( 305)

I 1 II 15 III 1 IV 1 V 2

Angulation occurred in 4/20 (20%) (3 were type II). In all fractures, shortening 1.5 cm or greater in 8/20 (40%); in type II, 5/15 (33%) (average 2.3 cm).

Lombardo and Harvey (34)(199)

I 1 II 24 III 5 IV 3 V 1

Shortening 2.0 cm plus in 36% overall. In type II injuries 1.2 cm plus in 10/24 (42%). Angulation 5 ~ plus occurred in 33% overall and in 8/24 (33%) of type II lesions.

(continues)

TABLE VII (continued)

Author (Cases)
Effects on growth

Salter-Harris classifications

Aitken and Magill (16) ~l~ Czitrom et al. (42) ~74~

II III I II III IV V I II

Roberts (91)(271)

15 1 2 27 2 8 3 28 63

].

Shortening occurred in 3/15 (20% with a range from 88 to 88 in.

J Very careful assessment of 42 patients. Angulation (varus/valgus 3-30 ~ 1/11 38%, IV-V 55%; shortening 0.1 to 3.8 cm 1/11 69%, IV-V 64%; 1.5 cm plus I-II 29%, IV-V 45%. Overall results: excellent-good 66.7%, fair-poor 33.3%. I-II results: excellent-good 72%, fair-poor 28%. "[

Shortening of 1.2 cm plus in 30/50 (60%) evaluated.

J

Neer (17) ~222~

Angulation corrected with growth (presumably due to presence of associated partial growth arrest). Shortening 7/17 (41%) from 0.5 to 3.5 cm; 5/17 (30%) 1.5 cm plus.

Cassebaum and Patterson,59)

Review of several older studies--growth alterations in 26/55 (40%). Shortening 2.0 cm plus in 25%.

D. Proximal Tibial Fracture-Separations Shelton and Canale (39) ~288)

I 9 (23%) II 17 (44%) III 10 (265) IV 3 (8%) V 0

Shortening of 1.0 cm to 2.5 cm in 7/28 (25%). All of these were I or II fractures, none seen in III-IV lesions. Angulation 5-7 ~ 8/28 (29%), also I-II lesions. Only one fracture in entire series in a patient less than 10 years old.

Burkhardt and Peterson (28) ~46~

I II III IV V I II III IV V I II III IV

Only one of 23 patients was less than 10 years of age at time of injury; five patients with exceptional injuries (all lawn-mower trauma) had type IV lesions and were less than 10. Shortening in 5/28 (21%): I(1), II(1), IV(2), V(2). Angular deformity 7/28 (255) from 12~ to 30 ~ I(1), II(1), IV(5).

2 (14%) 3 (22%)

I II III IV V

36(18%) 91 (45%) 49 (24%) 3(1%) 2(1%)

Growth complications of skeletal importance in entire series (shortening 1 cm; angulation 5~ 26/184 (14%). Type I lesions: 1/28 short; II: 11/66 (17%) with shortening and/or angulation; III: relatively few complications; IV: one arrest in three cases; V: two, both with premature arrest.

15(7%) 6(3%)

Worst complication rate (32%) in displaced III/IV lesions, Tillaux, triplane, V.

Nolan et al. (14) <226)

Aitken (14) ~9)

3 (11%) 9 (32%) 6 (21%) 8 (29%) 2(7%) 1 (7%) 6 (40%) 3 (20%) 4 (27%) 1 (7%) 9 (64%)

Population too small for detailed growth sequelae assessment.

No data on overall growth sequelae.

E. Distal Tibial Fracture-Separations Spiegel et al. (202) ~299)

Triplane Tillaux Nolan et al. (90) ~226)

I 9(10%) II 52 (58%) III 14 (16%) IV 11 (12%) V 4(4%)

Shortening 1.0-1.6 cm 6/50 (12% I:0, II:3, III:0, IV: 1, and V:2. Overall results: I good and excellent; V fair and poor; scattered results in II, III, and IV.

Excellent Good Fair Poor Aitken (21)~7)

I 16 (76%) II III 3 (14%) IV 2 (10%)

I-II: III: IV:

6 4 -

II 7 4 5 3

III 3 2 3 1

IV 2 3 0 3

V 1 0 1 2

No deformity; occasional premature fusion, but no clinically significant shortening. No problems. Mild shortening in 1; varus and 1.5 in. shortening in other.

566

CHAPTER 7 ~ Epiphyseai Growth Plate Fracture-Separations

minimally displaced fractures was an acceptable approach to management. They also concluded that those fractures with greater than 2 mm of displacement should be reduced, pinned, and immobilized for 6-8 weeks. Evidence of delayed union after 8 weeks was an indication for internal fixation and possibly bone grafting. Even established nonunions in good position were best treated by open reduction and bone grafting. Valgus deformity was the most common deviation. They stressed that "the functional results were invariably good regardless of the radiographic or clinical findings." Imperfect results radiographically involved an enlarged lateral condyle, delayed union, nonunion, and avascular necrosis. The fracture rarely crosses exactly between the trochlea and capitellum, but it usually includes a variable amount of the lateral trochlea. In the younger patients, therefore, there is very little bone to bone opening to allow for transphyseal bone bridge formation, although other complications can ensue. All eight of their poor results were in fractures widely displaced, rotated, and treated by open reduction. Much of the problem may relate not so much to cartilage destruction at the time of injury but to avascular necrosis induced by the fracture and the surgical intervention. The "fishtail" deformity, which refers to a notch between the trochlea and capitellum on the AP radiographic view, is thought to result from damage to the growth plate centrally with premature fusion of the capitellum to the lower humeral metaphysis. This has been described in the literature from 5 to 32% of cases. This group found the deformity in only four patients without early clinical significance. "We did not encounter any angular deformities as a result of an epiphyseal bridge or bony bar as one might expect from this injury. Our functional results were generally good regardless of the radiologic findings." The fact that the bar would be central would indicate why no angular deformities were seen. The fishtail deformity corresponds to the central bone bridge formation in distal femoral fracture-separations. Rutherford reviewed 39 fractures of the lateral condyle of the humerus and concluded that less than perfect results radiologically rarely led to any clinical problem (279). He took exception with the fact that some had commented on epiphyseal arrest with a transphyseal bone bridge between metaphyseal and epiphyseal bone segments. This bone bridge frequently is present, as noted by the works of others, but tends to appear centrally and thus is rarely a cause of valgus malformation directly. With premature closure of the lateral condylar physis, however, such deformation can occur. The fishtail deformity of the distal part of the humerus occurs commonly, caused by bone bridge formation centrally. The abnormalities of growth of this fracture involve overgrowth of the lateral condyle, formation of a cleft in the trochlea with widening of the distal part of the humerus (fishtail deformity), and premature epiphyseal closure. He indicated that premature epiphyseal closure was seen in one elbow with no associated angular deformity. Thirty-nine fractures were assessed with an average time of injury of 6.3 years

and an average length of followup up 5.5 years. It thus is evident that some patients still had growth remaining and others were at best only in the adolescent period at the termination of the study. It is well-recognized that many of the negative sequelae of lateral humeral condylar fractures, such as tardy ulnar palsy, occur in midadulthood. Rutherford feels that bone bridge formation does not occur in the sense that in his series and others valgus angulation is not detectable. The valgus deformation is described as OCCUlTing due to nonunion, late reduction, avascular necrosis, and malunion. He indicates that "no cases of metaphyseal-epiphyseal cross bridging and subsequent epiphyseal arrest were reported." He concludes that cross-bridging of the epiphyseal cartilage is a rare cause of cubitus valgus after a fracture of the lateral condyle in a child. The fishtail deformity is felt by most to be a central growth arrest or due to a formal separation between the trochlear epiphyseal plate and the trochlear portion of the capitella epiphyseal plate. He concluded that "uncomplicated separation of the 2 portions of the distal epiphyseal plate of the humerus is the real cause of the deformity." Wadsworth reviewed the types of displacement in the lateral humeral condylar entity and commented on the complications such as malunion, premature fusion, nonunion, avascular necrosis, or cubitus valgus with ulnar palsy (321). Malunion refers to the imperfect union between the trochlea and capitellum. It occurs with a type II injury. The fully displaced type III injury results in a nonunion. Premature epiphyseal fusion is mentioned by Wadsworth (321) as well as by Blount (28) and Salter and Harris (281). According to Wadsworth, in the first type, the capitellum epiphysis alone fuses prematurely to the metaphysis and because this epiphysis accounts for ossification of the outer part of the trochlea, disturbance of growth commonly referred to as a fishtail deformity results. In the second type, the normal pattern of fusion is followed with the capitellar and trochlear epiphyses first fusing to each other and then to the metaphysis, but the capitellar epiphysis fuses in advance of that of the trochlea leading to valgus angulation. Avascular necrosis can be seen and is most common with operative intervention, which has led some to be wary of such an approach. Valgus deformity results from a disturbance of the important osseous link between the trochlea and capitellar centers. This may occur in the first type of premature fusion, in malunion, or in established nonunion, all of which result in loss of the outer lip of the trochlea and fishtail deformity. Additional factors are union in malposition, nonunion, avascular necrosis, or premature fusion with diminished growth in the outer half or more of the lower end of the humerus. Ulnar palsy is an additional complication. Two uncomplicated cases treated by open reduction and internal fixation with Kirschner wires are illustrated in Fig. 22. In Fig. 23 an operative approach to a delayed union described by Flynn is shown, and in Fig. 24 a case of growth deformity resulting from incomplete closed reduction is shown.

SECTION VI ~ Clinical F e a t u r e s o f Acute Epiphyseai F r a c t u r e - S e p a r a t i o n s

567

F I G U R E 22 (A) Anteroposterior radiograph of a displaced type IV distal lateral humeral condyle fracture in a 6-year-old male. The pathophysiologic B~ situation exists. (B) Anteroposterior radiograph following open reduction and internal fixation with two K-wires that converted the fracture to type A. This anatomically related metaphysis to metaphysis, growth plate to growth plate, epiphysis to epiphysis, and articular cartilage surface to articular cartilage surface. Results with this intervention have been excellent in terms of minimizing long-term growth sequelae. (C) Anteroposterior elbow radiograph shows the characteristic position of fragment with complete separation and 180~ rotation of the lateral condyle. This is a type IV pattem. The arrow illustrates the rotational component. Open reduction and internal fixation clearly are mandatory to allow for appropriate healing. (D) AP radiograph shows postoperative correction with two K-wires stabilizing the reduced fragment. (E) Lateral radiograph showing wires in place. (F) AP radiograph shows healed fragment following the removal of K-wires. [Parts A and B reprinted from Shapiro and Rand (1992). Adv. Orthop. Surg. 15:175-203, 9 Lippincott Williams & Wilkins, with permission.]

2. DISTAL HUMERAL PHYSIS T h e s e f r a c t u r e - s e p a r a t i o n s , w h i c h occur a l m o s t exclusively in the first 5 years of life, are invariably type I or type II injuries [DeJager and H o f f m a n (81); D e L e e et al.

(82); Dias et al. (85); D o w n s and Wirth (90); H o l d a et al. (145); M i z u n o et al. (215); Siffert (290)]. It is difficult to k n o w w h e t h e r the disorder is e x t r e m e l y rare but very welld o c u m e n t e d or, as one increasingly suspects, represents a

568

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations

,r

FIGURE 23 Illustrationfrom Flynn (106) shows an approach to treating a nonunion of the lateral condyle. Open reduction and intemal fixation serve to remove any intervening fibrous tissue and appropriatelyrealign the fragment linking metaphysisto metaphysealbone, physealtissue to physeal tissue, epiphysealbone to epiphyseal bone, and also articular cartilage to articular cartilage. Metallicpin fixationthen is used with care being takennot to traverse the physis if at all possible. [Reprintedfrom Flynn (1989), J. Pediatr. Orthop. 9:691-696, 9 LippincottWilliams& Wilkins, with permission.]

lesion that is much more common that previously believed. These injuries are being diagnosed more often due to increased clinical awareness and the use of ultrasound [Dias et al. (85)] and MRI techniques. Hutchinson (1893) noted that "before 5 or 6 years [in cadaver experiments] it is easy to detach the lower epiphysis of the humerus en b l o c . . , but that after that age the separation is usually in part a fracture" (150). Smith (295) in 1850 and Farabeuf in 1886 (102) clearly demonstrated distal humeral epiphyseal separations. Neither angular deformity nor shortening due to growth arrest has been reported, although slight persisting, nonprogressive varus due to malposition at the time of healing is relatively common. This fracture-separation conforms to the type A pathophysiologic type, even without more detailed imaging studies. The persisting epiphyseal blood supply after injury in this extracapsular fracture and the relative thickness and gentle curvilinear plane of the cartilage plate in young patients can be inferred to protect against growth arrest problems based on literature reviews showing minimal problems postinjury. Treatments by closed reduction or open reduction with K-wire fixation have been used. Fractureseparation of the distal humeral epiphysis is now an increasingly recognized injury and reports in the literature indicate a fairly consistent pattern. Among the common features of reviews is the fact that the fracture-separation often is misdiagnosed initially either as a dislocation of the elbow or as a fracture of the lateral condyle (6). De Jager and Hoffman

reported on 12 cases (81). The primary problem with the fracture even after its correct diagnosis is the occurrence of cubitus varus after healing. This is reported in many papers and is an example of a malunion rather than asymmetric growth plate damage. The fracture is most common in those under 2 years of age and not infrequently occurs either as a birth injury in association with a difficult delivery or as a result of childhood abuse. The increasing recommendation in those who have seen many cases is for closed reduction to be followed by percutaneous K-wire pinning and long arm casting to ensure appropriate position. Displacement characteristically is posteromedial with only occasional posterolateral or anterior displacements. The injury is almost invariably a type I fracture-separation in the very young and a type II fracture-separation in those older than 2 years of age. De Jager and Hoffman reviewed 6 reports as well as their own totaling 48 fractures. The normal carrying angle was maintained in only 22 patients and it was not known in 2. The cubitus valgus was decreased in 12 patients and was sufficiently decreased to lead to a cubitus varus in an additional 12. They also reviewed the age at occurrence, finding 55% of the fracture-separations occurring under 2.5 years of age and 45% over that age. The largest series reported is of 21 children with fractureseparations of the distal end of the humeral epiphysis [Abe et al. (2)]. Of these, fully 15 developed cubitus varus after treatment with 9 undergoing corrective osteotomy. The study, which used the Ogden classification system, defined one type IA fracture ( S - H I), 16 type IIA ( S - H II), and 4 type IIC ( S - H II). The average age of occurrence in this series was slightly older than that in other groups, averaging 5.1 years and ranging from 1 to 11 years. Accuracy of diagnosis was enhanced by the use of arthrograms in most patients. The abnormal cubitus varus was due to initial malposition and not to a true growth arrest because in virtually all cases the cubitus varus as measured 1 year postinjury did not change from 4 to 15 years postinjury. The common pattern of displacement was posteromedial with only occasional posterolateral displacement. The group currently favors closed reduction and percutaneous pinning followed by long arm cast immobilization for 3 weeks. Cubitus varus again was identified as the most common complication, with 7 of 13 patients treated at the author's own institution developing the deformity and all 8 treated elsewhere developing it. Mizuno et al. reported on six patients with this injury (215). All were between 2 and 18 years old at the time of injury, and, as in other reports, diagnosis of fracture of the lateral condyle and traumatic dislocation of the elbow frequently were made initially. In these fracture-separations, the capitellum maintains its relationship to the head of the radius but not to the humerus. In a fracture of the lateral humeral condyle, the axes of the humerus and forearm remain the same but the relationship of the capitellum to the radial head frequently changes. With a dislocation of the elbow, the relationship between the epiphysis of the capitellum and the radial head is changed. The lesions were type I

SECTION VI ~ Clinical Features o f Acute Epiphyseal Fracture-Separations

569

FIGURE 24 A series of radiographs demonstratingthe long-termnegative sequelae of a slightly displaced lateralhumeral condyle type IV fracture-separation in which the B1 pathophysiologic state was not converted to type A, allowing imperfect growth to occur. (A) The initial injury shows minimaldisplacementof the type IV fracture-separationin a 16-month-oldmale. The position was accepted and the patient was treated with a long ann cast. (B) Anteroposteriorradiograph 10 months followingthe injury shows a well-established nonunion without valgus deformation. (C) Anteroposteriorradiograph 11 years postinjury shows nonunion and valgus deformity that measured 45~clinically. (D) A distal humeral varus osteotomyrealigned the elbow. The nonunited lateral condyle was not painful and operative union was not attempted. [Parts A and B reprinted from Shapiro and Rand (1992). Adv. Orthop. Surg. 15:175-203, 9 Lippincott Williams & Wilkins, with permission.]

(often confused with elbow dislocations) or type II (often confused with lateral condyle fractures). Dias et al. have reported on the value of ultrasound in diagnosing neonatal separation of the distal humerus (185). This modality is used increasingly for epiphyseal growth plate fractures in the neonatal age group, not only with this lesion but also with those of the proximal humeral and proximal femoral epiphyses. The authors concentrated on a review of neonatal separation of the distal humeral epiphysis and were able to report 15 cases, including their own. The disorder generally was associated with a difficult labor, birth trauma, or a footling breech delivery. The tendency in neonatal cases was to manipulate the arm to the correct position and hold it in a collar and cuff or posterior long arm splint.

Long-term results were reported as normal in each of the cases with only one example of 10 ~ varus reported. Barrett et al. (17), reporting on two cases, also noted posteromedial displacement. Treatment was by closed reduction with position maintained by flexing the elbow 90 ~ pronating the forearm, and immobilization in a posterior splint for 3 weeks. 3. MEDIAL HUMERAL CONDYLE These injuries are infrequent but have the potential to cause major growth problems. They are all Salter-Harris type IV injuries. Bensahel et al. (22) described a large number of these fractures, 27 cases, providing a good overview of the problem. The average age of the patients was 6 years 7 months with a range from 6 months to 11 years. Although

570

CHAPTER 7 9 Epiphyseal Growth Plate Fracture-Separations

all of the injuries were type IV in nature, there were three categories seen: (1) un-displaced fractures with the fracture line just visible proximal to the physeal plate (12 cases or 48%); (2) marked displacement medially and proximally without rotation of the fragment (9 cases or 30%); and (3) severe displacement with rotation of the fragment such that the articular surface faces medially and may be incarcerated in the elbow joint (6 cases or 22%). The treatment approach is similar to that used for lateral condylar fractures. Un-displaced fractures are treated in a long arm cast for 4 - 6 weeks. Minimally displaced fractures had either closed reduction or open reduction with K-wire pinning, and all type 3 fractures had open reduction with internal K-wire fixation. The un-displaced fractures gave consistently good results as did the majority of mildly displaced group 2 fractures. Poor results in groups 2 and 3 were seen only in the cases with delay of treatment and diagnosis. There is a relatively high degree of concern with this injury in child abuse, based on the reports in the literature. The youngest case report of fracture of the medial humeral condyle was in a 6-month-old infant reported by De Boeck et al. (79) who was suspected of having suffered from abuse. Open reduction was performed along with K-wire fixation, with normal clinical and radiological development at 3 years of age. Peterson has reviewed each of the three main types of distal humeral epiphyseal fracture separations (249, 251).

D. Proximal Radius It can often be difficult to differentiate pure radial neck metaphyseal fractures and proximal radial epiphyseal fractureseparations, especially as the latter are almost invariably type II with a prominent metaphyseal fragment. There is an equal rate of occurrence between epiphyseal fracture-separations and pure radical neck fractures in most studies, but some stress predominance of the purely metaphyseal fracture. Unless several oblique films are taken it can be difficult to determine at which level fracture has occurred. In four papers that make a clear distinction concerning the level of fracture, there are reports of 17 epiphyseal fracture-separations and 17 radial neck fractures (162), 25 fracture-separations and 20 neck fractures (225), 7 fracture-separations and 24 neck fractures (311), and 16 fracture-separations and 20 neck fractures. The largest published series of 100 fractures documented 79 physeal fractures and only 17 purely metaphyseal (92). The treatment approaches, however, are essentially the same regardless of the level of injury, as are the complications. Prognosis is dependent more on the initial amount of proximal fragment displacement than on the exact level of fracture. The vast majority of fractures of the proximal radial epiphysis are either type I or type II with very few type III or IV patterns reported [Fowles and Kassab (113); Jeffrey (160); Jones and Esah (162); McBride and Monnet (207); Newman (225); O'Brien (230); Reidy and Van Gorder (269);

Steinberg et al. (303); Tibone and Stoltz (311)] (Table VIIB). D' Souza et al. reported 71 type I and II injuries and 8 type IV injuries (92). In spite of this, there is a considerable incidence of avascular necrosis leading to shortening and angular valgus deformity. Premature epiphyseal fusion and avascular necrosis are particularly common in injuries with marked epiphyseal displacement from 60 ~ to 90 ~ Other negative sequelae are limitation of movement (primarily supination and pronation), pain, and radial head overgrowth. The type I injury frequently presents with complete posterior displacement of the radial epiphysis, which, on lateral radiographs, is tilted 90 ~, lying on the posterior surface of the neck. The proximal radial epiphysis is completely intracapsular and analogous to the proximal femur because its blood supply enters the epiphysis from distal to proximal and is subject to rupture with any displacement. There can also be considerable crushing of the growth plate, as the severe valgus stress on the elbow pins the physis between the capitellum and the radial metaphysis. These injuries can have both type C and type B2 pathophysiologic phenomena associated with them. Because the damage from one or both of these occurrences currently is irreparable, accurate repositioning of the radial epiphysis may not prevent premature growth plate arrest, although restoration of the radiohumeral joint is important. With each injury, therefore, care must be taken to discern either a pathophysiologic type A, B, or C situation. mFocal bone scans and MRI studies are most helpful in this regard. Accurate and gentle reduction frequently performed by open means should be helpful in markedly displaced fractures. If the clinical situation still appears relatively unstable, K-wire fixation also is recommended. Many authors, however, have documented the frequently poor results with this injury regardless of the method of treatment. One treatment protocol that seems acceptable to many is immobilization only with displacement 0-30 ~ closed reduction with displacement greater than 30 ~, and open reduction only for complete 90 ~ angular displacement. Steele and Graham have reported on the reduction of fractures with greater than 30 ~ angulation by leverage using a percutaneous Kirschner wire generally placed in the metaphyseal type II fragment (301). D' Souza et al. opted for immobilization only for fractures less than 45 ~ angulation, closed reduction for deformity greater than 45 ~ and open reduction only when closed attempts were unsuccessful.

E. Distal Radius Greater than one-third of epiphyseal growth plate fractureseparations involve the distal radius. The vast majority of these are either type I or type II injuries [Bragdon (32); Funsten (119); Lee et al. (189); Mischkowsky et al. (214)]. The characteristic small dorsal diaphyseal (metaphyseal) fragment in the type II fracture was clearly defined by C. Thurstan Holland in 1929, who used it to confirm the occurrence of epiphyseal separation and to note that the diaphyseal (me-

SECTION VI ~ Clinical Features of Acute Epiphyseal Fracture-Separations

F I G U R E 25 (A) Lateral radiograph shows a characteristic type II fracture-separation of the distal radial epiphysis. The small metaphyseal fragment (Thurstan Holland sign) is shown by the arrow. (B) Anteroposterior radiograph of the distal radius shows an irregular metaphyseal fragment with minimal displacement and the hint of a longitudinal component passing toward the physis (black arrow). The ulnar styloid also is cracked. The lateral radiograph was unremarkable. (C) Oblique radiograph of fracture illustrated in part (B) shows a clear type III pattern as well with epiphyseal and intra-articular involvement. The complexity of many epiphyseal growth plate fracture-separations often increases in terms of adjacent bone involvement whenever more detailed imaging is performed, beginning with simple oblique films and then progressing to tomography, CAT scans, or MR imaging.

taphyseal) fragment was present accompanying (attached to) the epiphysis "in the direction in which the displacement takes place" (146) (Fig. 25A). This plain radiographic finding is sometimes referred to as the Thurstan Holland sign. Werenskiold of Oslo also commented on the presence of a dorsal diaphyseal (metaphyseal) fracture associated with distal radial epiphyseal injuries (327). His radiologic report on epiphyseal injuries published in 1927 concentrated on the most frequent fractures, those at the distal radius. He demonstrated the ability of plain radiographs of high resolution to define even un-displaced pure or "genuine" epiphyseal fractures by the presence of "a thin lamella detached from the diaphysis (metaphysis) at the boundary between bone and cartilage and when displacement is not present this may be the only but sufficient proof of the existence of an epiphyseal separation." Distal radial epiphyseal injuries, however, generally occurred in mixed form with epiphyseal separation being part of the pathway and passage through the diaphys-

571

eal spongiosa the other. This describes the common SalterHarris type II pattern. When the fracture line passed from epiphyseal cartilage into bone "on the dorsal side in the diaphyseal (metaphyseal) spongiosa.., the fracture proceeds as in a typical fracture of the radius so that we have a more or less large 3-cornered diaphysis (metaphysis)-dorsal side." When the bone fragment was large, the epiphysis and fragment were almost always displaced. Lodes reviewed various patterns seen by plain radiographs (198). In a review of 423 distal fracture-separations from one iastitution there was a distribution as follows: 110 type I injuries (26%), 288 type II (68%), 13 type III (3.1%), 10 type IV (2.4%), and 2 type V (0.5%) [Lee et al. (189)] (Figs. 25A-25C). The distal radial growth plate, which is extracapsular, retains its blood supply even with extreme displacement, and the level transverse orientation of the physis generally protects against crushing or fissuring such that the large majority of the injuries can be considered as pathophysiologic A in nature. In type I and II fractures that heal in considerable malposition but in which growth plate continuity persists, extensive remodeling occurs. Although reduction is desirable, repeat reduction carries an increased risk of growth plate injury. The excellent remodeling mechanism, particularly in those with 2 or more years of growth remaining, should be allowed to work. A classic series of papers by Friberg (114-116) documented this remodeling potential of the growth plate itself. Where dorsal tilt predominates, periosteal remodeling involves filling in of bone on the concave side of the deformity, but it is growth plate differential growth that leads to actual correction of the angular deformity of the physis, epiphysis, and joint surface. This occurs by a relative overactivity of the cells at the dorsal half of the plate in relation to those in the volar position. Friberg studied the effect of residual fracture angulation on the distal radial and ulnar epiphyseal plates in children 1-15 years of age. Thirty-eight fractures located in the distal one-fifth of the forearm bones were observed from 1 to 25 months after healing. He clearly demonstrated that an abnormal inclination of the epiphyseal plate induced growth alterations in the plate. The redistribution of growth served to correct the abnormal inclination with the rate of correction following an exponential course. Neither the age of the child at the time of fracture nor the distance from the fracture to the epiphyseal plate (at least within the lower one-fifth of the bone) influenced the capacity for correction. The studies involved angulation of the radius in the dorsovolar plane and the radioulnar plane as well as angulations of the ulna in the dorsovolar plane. The distal epiphyseal plates of both radius and ulna altered their inclinations in relation to the long axis of the bones. Indeed, the only bones that did not tend to correct were those in which the postfracture angular deformities were very small, with a mean of 4 ~. Otherwise, without exception, changes in inclination tended toward normalization of the spatial orientation of the epiphyseal plates. These structures were seen in the radius in the dorsovolar and radial-ulnar planes as well as in

572

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

FIGURE 26 A series of lateral forearm radiographs shows remodeling after an angulated distal radial metaphyseal fracture in a 10-year-oldmale was allowed to heal without anatomic reduction. Note the change in orientation of the physis in relation to the shaft of the bone, indicating that altered physeal growth itself was partly responsible for the remodelingphenomenon. The physeal-diaphyseal angle at time of healing was 30~ (A), at 4 months was 14~ (B), and at 24 months was 7~ (C). The change in physeal tilt is due to more rapid growth in the dorsal part of the physis than in the volar part.

the dorsovolar plane of the distal ulna. It is important to note that the correction did not relate to filling in by the periosteum of the concavity but rather to a change in the angular tilt of the epiphyseal growth plate itself toward the normal plane. The mean correction at the distal radial epiphyseal growth plate was 0.9 ~ per month in the dorsal-volar plane and 0.8 ~ per month in the radial-ulnar plane. The ulnar epiphyseal plate showed a correction of 0.8 ~ per month. The rate of correction was high in the groups with larger primary angulations, and the rate of correction tended to diminish as an angular position closer to normal was achieved. In conclusion, Friberg noted that the epiphyseal plates had a definite and spontaneous ability to change their inclinations in relating to the forearm bones, with change invariably in a direction of correction toward the normal angle (Fig. 26). The greater the angular deviation, the greater the tendency toward correction. The correction was in some way related to the changes in direction and amplitude of the biomechanical forces acting on the plate, although no studies in this regard were done. In assessing fractures in the distal onefifth of the bones, no differences in the rate of correction of

the epiphyseal plate were found in fractures located at different distances from the plate. In a second study of these patients, Friberg assessed the orientation of both the distal and proximal epiphyseal plates of the radius to assess overall forearm bone remodeling in those fractures distally that had healed with residual angulation. The residual fracture angulation induced a change in orientation in both the distal and proximal epiphyseal plates. The tendency was to a normal inclination, with the proximal epiphyseal plate obtaining virtually normal orientation and the distal epiphyseal plate, if anything, tending toward some overcorrection. The upper limit for angulations permitting normalization of the orientation of the distal epiphyseal plate of the radius was 20 ~, because the four cases with angular deformity greater than 20 ~ improved but did not reach the normal state. The final angulation of the distal epiphyseal plate was 2 ~ or less (with a normal at 0) in 33 of the 39 fractures. Complete normalization could not be seen after a distal forearm fracture with a primary angulation of more than 20 ~. The angle of 20 ~ or less, acceptable to allow for spontaneous remodeling, was the same as that described by

SECTION VI ~ Clinical Features of Acute Epiphyseal Fracture-Separations Blount, who also concluded that angulations of more than 20 ~ should not be accepted (28). Nonnemann was a little more liberal in accepting angulation up to 30 ~ with the expectation that spontaneous remodeling would be sufficient (227). Friberg also noted that, when the fracture was of the distal radial epiphyseal growth plate, the residual angulation after healing also influenced the distribution of growth within the proximal epiphyseal growth plate of the radius. The younger the patient at the time of the fracture, the more complete the correction. It is not the age of the child, however, that governs the extent of correction but rather the time remaining to epiphyseal closure. As long as there was growth remaining, the age at fracture was not a governing concern. In the distal radius and ulna, the direction of the angulation of the fracture was of no importance for the corrective result. The wrist joint, however, does have a range of motion that incorporates all directions in which the fracture was measured. In a pure hinge joint, the correction is maximal in the plane of motion of the joint. Friberg stresses that the correction is within the physis itself because only such a mechanism can correct the angular inclination of the physis. Although historically this has been considered to be relatively benign injury in terms of growth sequelae, it is important to recognize that increasing numbers of posttraumatic growth arrests of the distal radial physis are being reported. These are seen in association with three mechanisms of trauma: recognized distal radial epiphyseal fractureseparations; after fracture of the distal radial metaphysis; and with chronic repetitive stress injury to the physis in gymnasts and other athletes. (1) Premature distal radial growth plate closure following recognized distal radial epiphyseal fracture-separations. Although most of the reports are of isolated cases [Hernandez and Peterson (144); Zehntner et al. (337); Valverde et al. (317)], 10 posttraumatic arrests were reported in one of the previously mentioned series [Lee et al. (189)]. Eight of these followed type II fracture-separations, with one type IV and one type V also diagnosed. The authors felt that "the mechanism of injury is of more value than the Salter classification for predicting the severity of growth plate injury," with a compression mechanism leading to greater trouble. Although the vast majority of fracture-separations of the distal radius are type A injuries, the possibility of a B2 situation exists. Concern for a B2 injury is greatest in those fractures that are associated with extensive trauma, open injury, or repeated manipulations. Figure 19D illustrates MR findings in an evolving posttraumatic transphyseal bone bridge. (2) Premature distal radial growth plate closure following fracture of the distal radial metaphysis. Isolated reports are beginning to appear of premature closure of the distal radial physis after fracture of the distal radial metaphysis. Aminian and Schoenecker presented two cases of premature closure of the growth plate secondary to an associated fracture near, but seemingly not involving, the distal radial growth plate

573

itself (14). One was a complete distal radius fracture and the other a toms fracture. The injury to the growth plate was not recognized until 2-3 years later when patients reported symptoms referable to a shortened distal radius with ulnar growth having continued. It was felt that compression injury occurred in association with the original trauma and that ischemia after plaster immobilization was not at fault. Abram and Thompson also reported premature growth arrest after a seemingly simple toms fracture of the metaphysis in a 10-year-old girl with an associated fracture of the ulnar styloid (3). The rarity of this injury is shown by the report of Davis and Green, who noted only 1 case of a patient who had premature closure of the distal radial physis in a review of 547 consecutive fractures of the forearm, including 92 toms fractures of the distal radius or of the radius and ulna (77). (3) Premature distal radial growth plate closure following chronic repetitive stress injury with excessive gymnastic activities. In chronic overuse syndromes, there can be discomfort in the region of the growth plate, and radiographs show the distal radial growth plate to be widened and hazy due to irregularity of the border between the cartilage and the metaphyseal zone of ossification. This represents inability of the growth plate to fully withstand rotational and compressive forces. With decreased sporting activity and 3-4 weeks of immobilization of the involved joint, the symptoms generally disappear and the radiographs return to normal. The injury is particularly common in gymnasts in which compressive and torsional forces with extreme weight beating are prominent, for example, with tumbling or vaulting. Stressrelated widening of the radial growth plate in adolescence has also been noted by Fliegel (105) and Albanese et al. (12). Both Carter and Aldrich (57) and Roy et al. (277) reported 21 cases each in young gymnasts. In the former series, the average age was 13.5 years for the boys and 14 years for the girls. Albanese et al. documented three female patients in their group who developed premature fusion of the distal radial growth plate with continued growth of the distal ulnar plate. Fusion was noted at 12 years, 13 years 10 months, and 14 years 1 month of age. The ulnar positive variance in all three was felt to be due to growth inhibition of the distal radius. Several treatment options are available for premature distal radial growth plate closure. If complete closure has already occurred, then the distal ulnar growth plate can be fused surgically if their relationship is still appropriate. If there is excess ulnar overgrowth, then ulnar shortening can be performed. If the radius has not closed symmetrically and there is angular deformity, then radial osteotomy and on occasion lengthening can be performed with completion of physeal closure. If a focal bone bridge exists and sufficient physeal tissue remains to allow growth to continue, then bone bridge resection can be performed. This may have to be accompanied by distal ulnar growth plate fusion depending on the degree of ulnar variance.

574

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

Waters et al. have pointed out that median nerve compression syndromes can occur following physeal fractures of the distal radius (322). They received eight patients with postfracture median neuropathy, some of whom resolved spontaneously but others of whom had carpal tunnel releases and percutaneous pin fixation. De Pablos et al. attempted to assess postfracture growth of the radius to determine whether clinically significant patterns occurred. The postfracture growth differential was minimal but appeared greatest when there was associated ulnar fracture. In 11 cases with ulnar fracture radial overgrowth was 0.54 cm but in 29 without ulnar fracture (most of which were greenstick or buckle radial fractures) overgrowth was only a mean of 0.13 cm. The rapidity of healing might well have limited excess physeal response. The overgrowth phenomenon does not appear to be of clinical significance in childhood forearm fractures.

F. Distal Ulna Distal ulnar physeal injury in 18 patients and 2 traumatic amputation specimens have been assessed by Golz et al. (126). Patients with simple ulnar styloid fractures were excluded. All clinical patients had associated distal radial fracture. Types I and II predominated, but growth arrest was found in each of the I-IV categories (type I 10, type II 1, type III 6, type IV 1). Excellent histopathological sections demonstrated both unappreciated type I injuries and examples of transphyseal bone bridge formation. Premature physeal closure occurred in 10 patients (55%), and in 7 secondary structural changes (ulnar curvature) in the distal radius occurred due to the tethering effect on medial radial growth. The occurrence of growth plate problems is shown again to relate quite often to the degree of trauma and displacement, with the pathoanatomic class alone being of relatively poor prognostic value. Although the distal ulnar epiphyseal fracture-separation is infrequent, it has a relatively high degree of growth sequelae when it occurs. Lipschultz reported 5 distal ulnar physeal fractures, all with associated distal radial fractures, with 3 type III and 2 type IV patterns and 3 of the 5 causing growth arrest with ulnar shortening of 9, 14, and 29 mm (195). Nelson et al. reported 4 distal ulnar premature growth arrests with 22-39 mm shortening and secondary radial and carpal growth changes (224). Ray et al. reviewed 28 cases of traumatic ulnar physeal premature arrest involving 5 of their own cases and 23 from the world literature (266). They defined three patterns of arrest and noted the compensatory radial changes. In their type 1 distal ulna growth arrest, the physis closed uniformly across its entire diameter leading to symmetric shortening. In the type 2 configuration, the physeal arrest was asymmetric occurring at the lateral part of the physis and leading to predominantly lateral ulnar shortening, and in the type 3 configuration, the distal physeal injury was at the medial side of the physis leading to predominantly medial ulnar short-

ening. The asymmetric growth arrests were most common in the Salter-Harris types III, IV, and V fracture-separation patterns. In their categorization, in those cases in which sufficient radiologic evidence existed for determination, 11 (39%) cases exhibited the type 1 symmetric shortening pattem, 3 (11%) the type 2 lateral ulna shortening pattern, and 6 (21%) the type 3 medial ulna shortening pattern. The ulnar shortening tended to tether the distal radial growth plate primarily on its ulnar or medial side. Thus, there was increased ulnar slope of the distal radial articular surface and ulnar bowing of the distal radial diaphysis and metaphysis. The carpal bones also shifted in relation to the altered radialulnar articulation. Surgical indications for treatment involved cosmetic deformity, progressive carpal subluxation, and decreased ranges of motion. If caught early, treatment of the ulna alone would be appropriate, involving ulnar lengthening with or without epiphyseal arrest of the distal radius and distal ulna if good alignment had been achieved. Radial osteotomy often was needed. Ray et al. also reported the use of the SuaveKapandji procedure for some of these injuries involving a lateral closing wedge radial osteotomy and distal radioulnar arthrodesis with segmental ulnar resection.

G. Metacarpals and Phalanges The large majority of metacarpal and finger phalanx growth plate fractures are type I and II injuries, although occasional type III and IV patterns occur, particularly with ligament avulsion mechanisms [Hastings and Simmons (141)]. In this lengthy review, 34% of hand fractures in children were epiphyseal fracture-separations. The prognosis overall, except with severe mutilating trauma, is good, although open reduction and pin fixation usually are warranted for the rare intraarticular type III and IV fractures.

H. Triradiate Acetabular Cartilage A rare but potentially serious physeal fracture is that involving the triradiate cartilage of the acetabulum (27, 42, 44, 136, 196, 274, 282). The triradiate cartilage is responsible for growth of the acetabulum, deepening it and contributing to lateral growth throughout development. The physis relates to each of the ilium, ischium, and pubis (hence the triradiate conformation) and is bipolar, with physeal tissue relating to each of the three bones. Fractures can occur at any time from shortly after birth to skeletal maturity. Three patterns are seen: types I, II (with the metaphyseal fragment generally medial in the ilium), and V (not appreciated until growth arrest is seen). Prognosis is more favorable with the type I and II injuries. CT scanning allows this injury to be diagnosed more accurately and more often in relation to childhood and adolescent massive pelvic-thigh trauma. Those injured under 10-11 years of age often manifest growth arrest, which

SECTION Vl ~ Clinical Features o f Acute Epiphyseal Fracture-Separations

575

presents as progressive acetabular dysplasia, thickening of the subchondral bone medially, and hip subluxation.

I. Proximal Femur 1. PROXIMAL FEMORAL CAPITAL EPIPHYSIS Fracture-separation of the proximal femoral capital epiphysis is an extremely rare, but potentially serious injury. The first documented case was described by Bousseau secondary to massive trauma in which the 15-year-old patient died and a postmortem exam was performed (30). Childhood fractures of the femoral head and neck region themselves are infrequent, and the trans-epiphyseal fracture-separation, a type I injury, accounts for only 4-8% of these. In a review of 471 fractures of the head and neck of the femur in children (as reviewed from many reported series), only 39 (8%) were trans-epiphyseal fracture-separations (148). The predominant age at fracture clusters in two groups: 2-6 years (9 cases) and 12-15 years (10 cases) of age. The prognosis in terms of avascular necrosis, premature fusion, and coxa vara previously was felt to be almost uniformly bad, but improved results in certain situations have been reported more recently. The blood supply of the intracapsular epiphysis is tenuous. The fractures are associated with major violence and significant displacement. In one series, each of 5 transepiphyseal fractures had dislocation of the femoral head from the acetabulum, all of whom developed avascularity. In a review of 39 fractures in 8 series, Hughes and Beaty noted that approximately half of the fractures were associated with traumatic dislocation of the femoral epiphysis and that it is this factor that predisposes one to avascular necrosis (148). All fracture-separations at this growth plate are type I (Figs. 27A and 27B). Type II-V injuries have not been described. Significant numbers of complications are associated with these separations, involving avascular necrosis of the femoral head, premature growth plate arrest, and coxa vara [Canale and Bourland (53); Lam (184); McDougall (208); Ratliff, Weiner, and O'Dell (325)] even though the type I injury generally is considered to be benign. With each fracture-separation at the proximal capital femoral epiphysis, the question arises initially as to whether one is dealing with a type A or type C situation. The poor prognosis is related to the fact that this proximal femoral growth plate is totally intracapsular and, thus, has a very vulnerable blood supply. The anatomy of the vessels in relation to the femoral head and neck has been well-illustrated by Chung (64). Speedy but gentle reduction, smooth Kirschner-wire pin fixation, and use of a hip spica are desirable to treat the fracture and minimize the likelihood of negative sequelae. Viability of the femoral head can be determined by a bone scan, which can be performed even after metallic fixation. Open reduction generally is favored when the head is dislocated. More recent reports have shown certain types of fractureseparations with a more favorable prognosis. Miller (213) reported no avascular necrosis (AVN) in his study, and Davi-

F I G U R E 27 (A) A displaced type I proximal femoral capital epiphyseal fracture-separation. (B) Treatment involving open reduction and epiphyseal stabilization using two K-wires and hip spica.

son and Weinstein (78) reported AVN in only one of four patients. Better results without avascular necrosis clearly are favored by physeal fracture without dislocation of the femoral head from the acetabulum. When dislocation accompanies the fracture, the likelihood of avascular necrosis is extremely high. It also appears that those suffering the fracture younger than 2.5 years of age have a much more favorable prognosis. Forlin et al. (109) reported on five children with displaced trans-epiphyseal fractures of the femoral neck without dislocation. Ages at fracture ranged from 8 to 26 months, and three were victims of child abuse. All were treated only with hip spica immobilization with no attempt at closed or open reduction. There were no cases of avascular necrosis. In one patient in which diagnosis was delayed for 18 months, nonunion required valgus osteotomy and bone grafting; there was no AVN but premature growth plate closure occurred. In the other patients, two healed with varus, no AVN, and open growth plates, which corrected the angular deformity over the next few years, whereas two healed with no AVN but with premature growth plate closure requiting

576

CHAPTER 7 ~ Epiphyseal Growth Plate F r a c t u r e - S e p a r a t i o n s

valgus osteotomy and attention to limb length discrepancy. Gentle treatment initially using only hip spica immobilization favors union without AVN; subsequent growth problems then are manageable to save a functional hip. Ogden et al. noted that the 5 poor and fair results in 11 patients in his study, by which is meant coxa vara and premature physeal closure, all occurred in children who were more than 2 years old at injury (235). Treatment is characterized by the need for immediate reduction. Gentle closed reduction can be attempted, but if unsuccessful, an immediate resort to open reduction is necessary. Postreduction stabilization also is essential due to a lack of any meaningful periosteal sleeve. Fixation is by smooth K-wires and hip spica in those with growth remaining. All of those in the slipped capital femoral epiphysis age group greater than 11 or 12 years of age can be fixed with cannulated screws. The complications include avascular necrosis, coxa vara, premature physeal closure, nonunion (very infrequent), and rare instances of chondrolysis. 2. NEONATAL PROXIMAL FEMORAL PHYSEAL SEPARATION There are several reports of type I epiphyseal separation at the proximal end of the femur in the newborn or early in infancy (212, 235). The newborn fractures are difficult to diagnose because the ossification centers of both the femoral capital epiphysis and the greater trochanteric epiphysis have not appeared. The characteristic presentation often is confused with septic arthritis or osteomyelitis of the proximal femur. There is marked disinclination to use the hip and the lower extremity lies with external rotation. There is lateral displacement or rotation of the proximal femoral shaft, which can be confused with a developmental dislocation of the hip. Ultrasound is of extreme value for diagnosis especially in the newborn period. X rays become valuable approximately 1 week postinjury when the periosteal new bone formation of the proximal femur can be noted. The fractures, as pointed out by Ogden et al., almost invariably involve the complete physis of the proximal end of the femur, which underlies both the femoral head-neck and greater trochanter (235). A high percentage of these injuries in the newborn are associated with traumatic delivery in particular with either breech or footling presentation. In infancy, it is essential to have a high suspicion of the possibility of child abuse. Ogden et al. (235) reviewed eight proximal femoral epiphyseal separations of whom five were birth injuries and two were victims of child abuse at 5 and 9 weeks of age. Six were unilateral and one bilateral. In their series, the capital femoral epiphysis remained located in all instances. Six patients were treated with Bryant's traction for 6-15 days, following which either hip spica or bilateral abduction splinting was used. Followup varied from 3 to 15 years in five patients and 22 years in one. There were no long-term growth deformities noted. Two had a mild varus that normalized within the first year of life, and another developed anterior bowing of the femur that persisted at 4 years of age.

An experimental investigation was done involving manual trauma in six neonatal stillborn pelvic-femoral units comprising musculoskeletal tissue only to assess the pattern of disruption. In each instance, the hip capsule remained intact. Generally there was a relatively clean break across the physis of the proximal femur and greater trochanter but proximal to the lesser trochanter. The periosteal sleeve was intact posteriorly. Histologic studies showed the fracture invariably to be a type I growth plate injury with the fracture line across the entire proximal femoral epiphysis. In two cases, the fracture line passed into the epiphysis, splitting the germinal zone away from the epiphyseal blood supply. Multiple longitudinal splitting of physeal cell columns also occurred. Ogden et al. reviewed many cases, most of which were isolated single reports from 1898 onward (235). Half of the babies subsequently were described as being "largefor-dates." The frequency of the injury decreased with improved obstetric care. Once beyond the neonatal period of hospitalization, child abuse must be suspected. The fractures not only were through the physis throughout their entire extent but were almost invariably at or below the lower parts of the hypertrophic zone. Medial propagation into the epiphysis was seen on occasion, as were longitudinal splitting and fragmentation of the physis in the experimental model. The germinal cells, however, were not crushed. The posterior periosteum tended to remain intact but often was elevated for several centimeters from the diaphyseal shaft.

J. Distal Femur 1. DISTAL FEMORAL FRACTURE-SEPARATIONS Five types of pathoanatomic pattern occur, but the large majority are type II. The type distribution in five major early reviews (203 cases) was type I 15.4%, type II 71.6%, type III 4.3%, type IV 5.8%, and type V 2.9% (10, 74, 199, 271, 305) (Table VIIC). There is a high incidence of physeal damage, with shortening and angulation occurring in approximately 40% of those cases, even when type II fractures alone are considered (283). The type III and type IV injuries have a similar rate of problems associated with them. The fairly high incidence of growth plate problems involving angular deformity or shortening or both after distal femoral fractureseparations has been noted in many studies [Abbott and Gill (1); Bylander et al. (49); Cassebaum and Patterson (59); Czitrom et al. (74); Griswell et al. (73); Lombardo and Harvey (199); Neer (222); Padovani et al. (243); Riseborough et al. (270); Stephens and Louis (305)]. These problems can occur with type I, II, III, or IV injuries (Figs. 28A-28Q). Bylander et al. (49) concluded that the "Salter-Harris classification is of minor value in predicting growth disturbance after injury to the growth plates of the distal femur and proximal tibia." The extracapsular blood supply to the distal femur is extensive and there is little concern about a type C injury, although as noted previously the epiphyseal vascular state postinjury is not yet amenable to direct assessment in terms of focal diminution. The major consideration, however,

F I G U R E 28 Radiographic appearances, treatments, and sequelae following distal femoral epiphyseal growth plate fractureseparations are illustrated. Parts (A-D) show a common type II pattern treated by closed reduction, stabilization of the fracture by metaphyseal fragment Kirschner wire fixation, and long leg casting. Operative stabilization seeks to minimize continuing postfracture physeal damage by allowing healing to occur in a stable environment. (E) A type III minimally displaced fracture-separation of the distal femoral epiphysis. AP and lateral radiographs were unrevealing but the oblique film showed the separation (arrows). (F) A tomographic radiograph showing epiphyseal bone and articular cartilage interruption and displacement with greater clarity. (G) ACT image from a radiologically similar case. (I-I, I) Reconstitution following open reduction and internal fixation using an intra-epiphyseal AO compression screw. The injury occurred close to skeletal maturity and no growth sequelae followed. (J) Massive direct trauma to the distal femur in a skiing injury resulted in a markedly displaced long metaphyseal fragment and transverse physeal fractures involving both medial and lateral components. The femur was treated with open reduction and internal fixation using multiple compression screws across the metaphyseal fragments, whereas the epiphyseal reduction was held with a long leg cast (Ji). Anatomic healing occurred but

F I G U R E 28 (continued) was associated with virtually immediate complete transphyseal fusion (Jii). (K) An undisplaced type I fracture-separation of the distal femur in an 8-year-old girl in association with a midshaft diaphyseal fracture on the same side. The diaphyseal injury healed without difficulty in traction followed by casting. No specific attention was paid to the distal femoral fracture. Within 1 year of injury it became apparent that there was a distal femoral growth plate focal arrest occurring. (L) A tomogram shows the focal lateral physeal arrest with physeal narrowing to disappearance, increased bone density, angular deformity, and a growth arrest line passing obliquely into the area of the bone bridge. Surgical procedures followed involving varus osteotomy and attempted bone bridge resection. (M) A lateral radiograph 6 years later at skeletal maturity shows a flexion deformity of the distal femur due to premature posterior physeal closure and an extension deformity of the proximal tibia due to unrecognized tibial tubercle area fracture with premature closure. (N) Realignment of the lower extremity was achieved with an extension osteotomy of the distal femur and a flexion osteotomy of the proximal tibia. (O) The corrected positions of the distal femur and (P) proximal tibia are highlighted. Length equality was achieved at skeletal maturity but required an ipsilateral femoral lengthening and a contralateral diaphyseal shortening. (Q) A premature central physeal growth plate closure in a different patient has caused a characteristic fishtail deformity of the distal femur.

SECTION VI ~ Clinical Features o f Acute Epiphyseal Fracture-Separations

is whether the fracture is associated with the type A, B1, or B2 categorizations. The type II fracture at the epiphysis predisposes one to crushing and fissuring of the nontransverse growth plate in association with the great force required for separation. Displacement tends to shear the central and peripheral epiphyseal growth plate regions, frequently causing longitudinal growth plate fracture and communication between metaphyseal and epiphyseal circulations, which allows the formation of a bone bridge. In type II injuries, stabilization of the generally large metaphyseal fragment to the metaphysis using AO compression screws is designed to minimize postfracture micromotion that might further damage the epiphyseal plate. The type II distal femoral injury, therefore, may conform to the type A pattern, but the B2 pattern often prevails. In the type III injury, either the type A or B2 phenomenon can occur. Which type occurs may very well depend on the precise transverse level of physeal fracture. If the transverse fracture is at the metaphyseal level, the likelihood of epiphyseal and metaphyseal communication is increased unless reduction is anatomic. Torg et al. reported specifically on six type III fractures of the medial femoral condyle that they had seen over a several-year period (313). The injury was due to a valgus force applied to the knee. In five of their six cases, the fracture was non-displaced and often a considerable period of time passed prior to its being appreciated as a fracture. They thus stress the importance of the need for cross-table lateral X rays in suspicious cases of knee injury in an effort to detect fat within the joint fluid and confirm the existence of an intra-articular fracture. Oblique and tunnel views were also helpful. At present CT or MR imaging helps to clarify fracture levels. Due to the un-displaced nature of most of their cases and due to the fact that the age at occurrence was between 14 and 16 years, there were virtually no negative sequelae. Treatment was with a cylinder cast. All of the injuries were football- or soccer-related, and in four of the six, the initial routine radiograph did not suffice to allow for the diagnosis. It was the stress radiographs and oblique radiographs that clarified the picture, although at present stress films are not recommended. Thompson et al. reviewed 30 consecutive fractures of the distal femoral epiphysis (310). There were 24 type II injuries, 2 type III, and 4 type IV. Growth-related complications were far more frequent in displaced than in non-displaced fractures. They noted 18 patients (60%) with excellent resuits, 8 fair (27%), and 4 poor (13%). A fair result had a limb length discrepancy of 1-2 cm and an angular deformity of 5-10 ~ whereas a poor result had greater than 2 cm of limb length discrepancy and greater than 10~ of angulation. When considering only the displaced fractures, the quality of the result diminished with 48% excellent, 35% fair, and 17% poor. The poorest results were in the group that demonstrated greater than 50% initial fracture displacement, which correlates with more extensive trauma than in non- or minimally displaced fractures. They noted that 47% of patients developed lower extremity length discrepancies, with 4,

579

however, less than 1 cm, 6 from 1 to 2 cm, and 4 greater than 2 cm. The lower extremity length discrepancy was not related to the Salter-Harris classification, the use of internal fixation, or the chronological age. Thompson et al. strongly recommended that displaced fracture-separations were best treated with early, gentle anatomic reduction under general anesthesia with open reduction if needed. Following reduction, internal fixation with percutaneous Steinman pin fixation was helpful to minimize the risk of displacement. A large percentage of long leg casts used following fracture reduction without internal fixation led to some displacement. The B, pattern generally is present prior to reduction in type IV fractures. Relatively little crushing occurs at the time of type IV injury, and accurate repositioning by an open operation can convert a B~ injury to type A. In the displaced type III and IV lesions, accurate anatomic reduction and AO compression screw fixation closely reoppose the articular cartilage surfaces, epiphyseal bone to epiphyseal bone, growth plate to growth plate, and metaphyseal bone to metaphyseal bone, thus minimizing the likelihood of eventual growth arrest and deformity. 2. NEWBORN FRACTURES OF THE DISTAL FEMORAL EPIPHYSIS Traumatic separations of the distal femoral epiphysis have been described in the newborn. Banagale and Kuhns reviewed 23 cases from the literature, including 1 of their own, and noted that all were Salter-Harris type 1 separations of the distal femoral epiphysis (16). Their review encompassed cases from 1898 to 1983. If we add 5 cases that Riseborough et al. (270) reported from our hospital in 1985, then a series of at least 28 cases has been reported. All were type I injuries. In 7 instances, the cases were bilateral. Where presentation was documented, 22 of the 25 patients were breech with only 3 vertex. Diagnosis frequently was delayed as concern about swelling and disuse of the limb frequently centered about septic arthritis, osteomyelitis, paralysis, or lymphedema. The diagnosis quite frequently came to light only when sequential radiographs showed abundant callus due to the extensive periosteal elevation in association with the fracture. In spite of this, virtually no negative sequelae occurred. With gentle reduction and splinting, the fractures healed uneventfully. The distal femoral physis in the newborn is less sharply angled centrally than in the older patient, and the thickness of the physeal and epiphyseal cartilage is relatively greater in relation to the secondary center bone, both features serving to minimize growth arrest problems. There was occasional angular deformity particularly when diagnosis was unappreciated or late, although this had sufficient time to remodel. Shortening was not a problem. 3. DISTAL FEMORAL FRACTURE-SEPARATIONS STUDIED AT CHILDREN'S HOSPITAL, BOSTON We studied the results following 66 distal femoral physeal fracture-separations (270). Sixteen of the patients were seen initially at Children's Hospital, Boston. Of those who were

580

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

referred secondarily, many already had growth problems, and the complications listed, though accurate, are not necessarily a true reflection of the frequency following such fractures. Lower extremity length discrepancies in excess of 2.4 cm or leading to contralateral distal femoral physeal arrest occurred in 37 patients (56%), and angular deformity of more than 5 ~ or requiring osteotomy occurred in 17 patients (26%). The radiographic sign of central arrest occurred in 13 patients (20%) and was associated primarily with type I and type II injuries. This sign had excellent predictive value for the development of limb length discrepancy. Growth problems correlated well with the severity of trauma and were seen in each of the Salter-Harris types. Fractures in the juvenile age group (2-11 years old) almost invariably were caused by severe trauma and had the poorest prognosis; 19 (83%) of the 23 injuries in juvenile patients resulted in growth problems. Fractures in the adolescent age group (11 years and older) were caused by less extensive trauma, being associated most often with sports injuries; 18 (50%) of the 36 adolescents had growth problems. Birth fractures and those in the first 2 years of life were associated with mild growth disturbances on occasion but did not lead to severe growth problems in any of the 7 patients. We believe that the final results would be improved by anatomic reduction and greater use of internal fixation in type II, III, and IV injuries. The management of these patients after healing of the fracture was not sufficiently intensive, because 19 patients had a length discrepancy of 2.4 cm or more persisting at skeletal maturity. a. Incidence. There were 47 boys and 19 girls, for a 2.5: 1 male predominance. The left femur was involved in 38 patients and the right in 28 (1.4:1). There were no bilateral fractures. b. Type of Fracture. Forty-seven of the fractures were classified by the system of Salter and Harris. There were 7 type I, 25 type II, 7 type III, 6 type IV, and 2 type V fractures. For 19 fractures the type was uncertain, either because the initial radiographs were no longer available for assessment or because the patients were seen initially in other hospitals without appropriate documentation in this regard. c. Age at the Time of Fracture. The ages of the patients at the time of fracture ranged from birth to 16 years 6 months. Five birth fractures and 2 fractures in the second year of life accounted for 10% of the injuries; 23 fractures in the juvenile age range (2 years 3 months to 10 years 11 months) accounted for 35%, and 36 fractures in the adolescent age range (11 years old or more) accounted for 55% injuries. The average age at fracture was 10.8 years for boys and 9.8 years for girls. d. Mechanism of Injury. Three of the 5 birth fractures were associated with breech delivery. Fractures in juvenile patients were associated with much more severe trauma than those in adolescents. Of the 23 children in the juvenile age group, 3 fell from a window, 14 were hit by an automobile, 3 were crushed by a heavy object, 1 sustained gunshot wounds, and only 2 had a sports-related accident. Twenty-

four of the injuries in the adolescents were sports- or playrelated, 11 were automobile- or truck-related, and 1 injury resulted from a fall. e. Closed and Open Fractures. Sixty fractures were closed and 6 were open. The open injuries comprised 4 of the 6 type IV injuries, 1 type V fracture, and 1 fracture whose type was unknown. None of the identified type I, II, or III injuries were open. f. Methods of Treatment. In the 3 type I birth injuries, formal reduction with traction in extension was not used because it was thought that remodeling would correct the deformities that were not controlled by lower extremity splinting. The patient who was a victim of child abuse at the age of 1 year 4 months was treated with a long plaster splint. Two of the 3 older patients with a type I lesion (12 years 5 months, 12 years 6 months, and 12 years 10 months old) had closed reduction and plaster cast immobilization, and 1 had an open reduction with Kirschner-wire pinning and plaster cast immobilization. Each of the 3 latter patients retained the reduction. Twenty of the 25 type II lesions had closed reduction with cast immobilization, 2 required no reduction, and 3 had closed reduction with percutaneous Kirschner-wire or Steinmann pinning of the large metaphyseal fragment under radiographic control. In 1 patient, closed reduction was supplemented with several days of skeletal traction. Of the patients in whom percutaneous pinning was not done, a long leg cast was used for 14, of whom 2 lost some position, and a hip spica cast was used for 8 patients, of whom 1 lost position. One of the 7 type III fractures did not require reduction. Six had closed reduction, 4 of which were anatomical and remained so. The position accepted after reduction remained unchanged in 6 patients in whom a long cast was used and in the 1 patient in whom a hip spica was used. Three of the 6 type IV lesions had open reduction and fixation with Kirschner wires, whereas the other 3, 2 of which were open and required debridement, were reduced by closed manipulation. One of these required open reduction 5 weeks later. One of the 2 type V fractures required closed reduction; both were treated in a long leg cast. In this study, the most important factors associated with growth arrest were not the pathoanatomical types of the injuries but the patient's age at the time of injury and the severity of the injury. None of the five birth fractures, all of which were type I or type II lesions, led to significant growth arrest. All angular deformities corrected with growth even when no specific attempts at reduction were made. A previous a review of the cases of six similar patients also documented the occurrence of such injuries with breech presentation, the uneventful healing, and the tendency for length discrepancies and angular deformities to resolve completely or with minimal discrepancy persisting (48). Virtually all of the lesions in patients in the juvenile age group, however, were associated with massive trauma. This, rather than the specific pattern of injury, correlated with high levels of growth arrest and deformity.

SECTION Vl ~ Clinical Features of Acute Epiphyseal Fracture-Separations This association of significant problems with severity of injury and displacement previously has been noted in physeal fracture-separations and has led to the proposal for consideration and investigation of such injuries on a pathophysiological basis. The majority of the fractures in this series occurred in patients in the adolescent age group, from 11 years to skeletal maturity. Although a considerable number of these patients also had growth arrest problems, the proportion was not as great as that in the juvenile group. Most of the injuries were sports-related, which usually implies a lesser degree of trauma. Eighty-three percent of the juvenile patients compared with 50% of the adolescent patients had a length discrepancy or angular deformity. In terms of the pathophysiological classification, the greater the trauma, the more severe the injury the physis receives, especially damage to the blood supply and physeal cartilage that is not evident radiographically. The majority of injuries that subsequently resulted in difficulty were in patients in the juvenile age group. In these children, the periosteal and perichondrial sheath is quite thick, and its disruption implies massive trauma. In the adolescent age group, the perichondrial and periosteal sheath is thinner and is closer in thickness and strength to that of an adult, and relatively few of these patients had growth arrest problems. The mechanism of fracture in the two age groups indicates that the amount of trauma required to disrupt the plate was relatively less in the older patients. In association with a lesser amount of trauma, one can postulate less crushing of the cartilage plate and less vascular disruption. The principle that high-speed trauma produces more problems in terms of soft tissue damage and length of healing than relatively minor, low-speed trauma in adult diaphyseal fractures appears to apply equally well to physeal fracture-separations in childhood. In addition, the closer the patient is to skeletal maturity, the fewer the sequelae to growth damage. This study documented the central arrest phenomenon that occurred in 20% of the patients. Transphyseal bone continuity was seen across the middle of the growth plate, with the radiolucency that persisted medially and laterally representing apparently normal growth plate cartilage. The central arrest phenomenon was ominous, however, in that following its appearance in all patients either complete radiographic fusion occurred in a matter of months or the central arrest persisted with little or no further contribution of the physis to femoral length. Of the eight central arrest signs in patients whose fracture type was known from the beginning, six were in type I and type II lesions. It has been demonstrated in the rabbit that extensive damage to the central area of the physis results in marked interference with growth. It has also been shown, in a physeal fracture model in the distal end of the femur in rats, that with Aitken type I lesions (Salter-Harris Type II) the fracture line involves the hypertrophic cartilage zone of the physis except in the central region, in which the defect extends more deeply into the zone of resting cartilage cells and sometimes

581

into the bone plate of the epiphysis (33). The central arrest phenomenon is attributable to a similar fracture path in the human femur. Several forms of treatment were used in this series, although the basic principle of obtaining and maintaining as anatomic a reduction as possible was followed. Closed or open reduction should be firm but gentle so as not to aggravate the damage of the physis. There is little advantage in achieving perfect radiographic reduction if the force used to do so causes excessive physeal damage. Many of the patients in this series had immobilization with a long leg cast rather than a hip spica. The results, however, did not demonstrate problems with long leg casts in the type I and type II lesions. Of the 16 patients who wore a long cast, only 2 lost some position, whereas of the 7 who wore a hip spica cast, 1 lost position. If the patient is short or has an obese or muscular thigh, hip spica treatment is safer, but in a child who is relatively tall and thin, immobilization in a long cast with appropriate flexion and molding forces appears to be adequate. Close monitoring is required. In 3 patients in whom a large metaphyseal fragment was associated with a type II fractureseparation, either Steinmann-pin or Kirschner-wire fixation of the metaphyseal fragment to the metaphysis was performed after closed reduction. This makes subsequent displacement even less likely and allows the use of a long leg cast with a greater margin of safety. In the type III and type IV lesions, gentle anatomic reduction by closed or open means and Kirschner-wire or cancellous screw fixation appear to be essential, because the possibility of growth arrest and discontinuity of the articular surface increases in the absence of anatomic reduction. The purpose of operative intervention is fourfold: to restore articular surface continuity, to relate epiphyseal bone only to epiphyseal bone, to restore physeal continuity, and to relate metaphyseal bone to metaphyseal bone. In type III fractures, internal fixation can pass from the epiphyseal fragment to the main body of the epiphysis without traversing the physis. Similarly, in type IV lesions, following accurate reduction, internal fixation passing from metaphyseal fragment to metaphyseal fragment is warranted. Often epiphyseal fragment fixation to the adjacent epiphysis can also be added. Our study highlighted certain problems in the management of patients after healing of the fracture. Nineteen patients had a lower extremity length discrepancy of 2.4 cm or more at skeletal maturity, and 8 of them had a contralateral epiphyseal arrest in an attempt to limit the discrepancy. These results certainly would have been improved by earlier investigation and management. In many patients, the discrepancy problems either were not recognized or were seen and treated too late and not aggressively enough. Tomograms and computed tomography scans are of great value in delineating the presence and extent of bone bridges. Earlier noninvasive vascular studies might help to determine the extrinsic physeal blood supply and possible intraosseous communications between epiphyseal and metaphyseal compartments.

582

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

Magnetic resonance imaging plays an increasing role in our management scheme to assess transphyseal bone bridge formation in the early weeks and months postinjury. Partial inhibition of growth, referred to by Neer as "deceleration of growth," can occur following a physeal lesion. Isolated instances of growth stimulation after fractureseparation also have been reported; the case of one such patient was documented in our series. The importance of following such patients to skeletal maturity even if the physis remains radiographically open following healing is evident. If total growth arrest is found to have occurred, one can readily calculate the eventual limb discrepancy from femoral and tibial length charts and growth remaining charts. If the projected length discrepancy is less than 5.0 cm, we recommend contralateral distal femoral physeal arrest. Such a procedure will not correct any discrepancy that has already occurred but will prevent it from increasing. If further reduction of the length discrepancy is required, contralateral proximal tibial arrest can be done. This procedure was not used often enough or early enough in this series. If it appears that the ultimate discrepancy will be greater than 5.0 cm, one should consider the option of diaphyseal lengthening of the involved limb. Early demonstration of partial arrest can help to prevent or minimize angular deformity by alerting the surgeon either to complete the arrest across the width of the plate or to attempt to resect the bone bridge and interpose a biological or prosthetic implant to allow growth to continue. It is important to remember that not all angular deformities are due to growth arrest phenomena. Certain type I and Type II fractures, if reduced imperfectly, may result in a varus or valgus malunion. We again stress the importance of CT scans or MRI in demonstrating bone bridges or narrowing of the plate. Sufficiently small bone bridges yield to cartilage growth and elongate with continuing activity of the plate; larger bridges, however, lead to growth arrest. Techniques have been developed for resection of such osseous bridges and implantation of a Silastic implant or fat to block the invasion of vascular mesenchymal tissue and reformation of such bridges. These will be described in Chapter 8. K. P r o x i m a l Tibia 1. PROXIMAL TIBIAL GROWTH PLATE Growth plate fracture-separations of the linear horizontal proximal tibial physis are rare, but a considerable number of patients develop growth arrest problems with type II, III, or IV injuries [Burkhart and Peterson (46); Bylander et al. (49); Shelton and Canale (288)] (Figs. 19A and 19B). All five types of pathoanatomic fracture pattern occur. A compilation of 95 cases from 5 major reviews (2, 11, 46, 226, 287) showed type I 13.6%, type II 43.2%, Type III 22.1%, Type IV 18.9% and Type V 3.2% (Table VIID). Approximately 20% of patients develop partial or complete growth arrest with problems occurring following each of the fractureseparation patterns. Except for major machine-type trauma,

proximal tibial fracture-separations do not tend to occur in those less than 10 years of age. Shelton and Canale noted shortening of 1.0-2.5 cm in 7/28 (25%), all types I and II (280). Angulation from 5 to 7 ~ also was frequent (8/28, 29%) and again was limited to the type I and II lesions. Burkhardt and Peterson had similar results (46). The growth plate is extracapsular and has an excellent blood supply. The slight irregularity of plate shape in the transverse plane, the anterior and inferior epiphyseal projection (tibial apophysis), and the extreme amount of force needed to cause fracture-separation can induce considerable growth plate crushing, allowing the epiphyseal and metaphyseal circulations to interrelate and form a bone bridge (type B2), although the majority of lesions still yield a type A situation with no growth arrest. It is important to assess the tibial tubercle carefully because premature fusion of this region can lead to subsequent hyperextension deformities. Figs. 19A and 19B, illustrate how MR imaging can reveal unsuspected adjacent bone fractures. Thompson and Gesler described a rare proximal tibial epiphyseal fracture occurring in a 7-month-old female as a result of apparent child abuse (308). Closed reduction was ineffective in holding alignment and open reduction and internal fixation were used, at which time it was noted that there had been infolding of the pes anserinus and periosteum. There was postrepair tibia valga due to asymmetrical physeal overgrowth, even though reduction had been anatomic. This was corrected subsequently with osteotomy within 1 year of injury. The proximal tibial and fibula varus osteotomy realigned the limbs well, but a mild valgus of 12~ recurred that remained nonprogressive over the subsequent 3 years. 2. ANTERIOR TIBIAL SPINE Tibial spine fractures are intra-epiphyseal fractures in children that do not involve the epiphyseal growth plate or physis. They are included in Ogden's classification as type VII (232, 233). They occur at the point of attachment of the anterior cruciate ligament and thus are the childhood analogue of ruptures of the anterior cruciate ligament in adults. Due to the strength of the ligaments in the childhood age group, an injury that would produce a cruciate tear in the adult avulses a segment of cartilage and bone at the point of insertion in the growing child. With stress, the incompletely ossified tibial eminence in the child fails before the ligament. The failure occurs through the cancellous bone beneath the subchondral plate. The cruciate functional integrity is compromised, and the fracture, if sufficiently large, can actually encroach onto the weight bearing portion of the articular surface of the tibia, particularly the region of the medial tibial plateau. The classification of Meyers and McKeever continues to be used widely: type I is non-displaced, type II is partially displaced or hinged with the posterior segment remaining intact, and type III is fully displaced (95, 211) (Fig. 29). Zaricznyj extended the classification to type IV, which includes comminution of the tibial spine fracture (336). Anterior cruciate ligament laxity has been documented in follow-up

SECTION V! ~ Clinical Features of Acute Epiphyseal Fracture-Separations

I ~

II

. . III ~ /~~ IV~t

I

F I G U R E 29 Classification of anterior tibial spine fractures is shown. [Reprinted from the Journal of the American Academy of Orthopaedic Surgeons, Vol. 3(2), pp. 63-69, 9 1995 American Academy of Orthopaedic Surgeons, with permission.]

studies of these injuries. Even though the ligament itself is intact, if the insertion pulls free and heals in a nonanatomic position, the cruciate laxity can be diagnosed. Wiley and Baxter reported on 82 patients with fractures of the tibial spine in an 8-year period, with 45 available for long-range assessment (328). The average age at occurrence was 11 years with the range from 8 to 16 years. There was a considerable 4:1 male predominance. There were 8 type I, 13 type II, and 24 type III injuries. Forty-two involved the anterior tibial spine and only 3 involved the posterior tibial spine. Type I injuries were treated with simple cast immobilization with the knee in slight flexion following joint aspiration. All healed uneventfully and resulted in minimal residual laxity. In the type I and type II injuries treated with simple immobilization, no difference in laxity between injured and uninjured knees was noted at final examination. In the type II and type III injuries treated with closed reduction compared to the 16 type III lesions treated with open reduction, no meaningful difference was documented in terms of the final knee stability test. The authors concluded that "fractures of the intercondylar eminence of the tibia in children can result in a measurable degree of cruciate ligament laxity . . . . however, the patients are generally asymptomatic. An anatomic reduction of the fracture by arthrotomy does not prevent the cruciate laxity or the loss of full knee extension. Open reduction should be reserved for the completely displaced fracture, particularly the inverted fragment so that a bony surface opposes the fracture bed." The authors concluded that the overall prognosis for fracture of the intercondylar eminence of the tibia was "remarkably good, even when the fracture fragment is completely displaced." Gronkvist et al. assessed 53 cases of anterior tibial spine fractures in patients less than 15 years of age at the time of injury (129). Forty-one patients returned for long-term follow-up. They found a substantial number of patients with residual symptoms and instabilities. They documented 5 type I cases. All of these had good results with conservative cast treatment. Fourteen type II injuries were assessed and the large majority of these had closed reduction followed by casting, and all with the exception of 2 were summarized /

583

as being without symptoms at long-term follow-up. There were 20 type III fractures with 10 described ultimately as being without symptoms, whereas 10 others were symptomatic at follow-up. Almost all of the type III cases had open reduction, with many also undergoing osteosynthesis as well as casting. The authors also mention that, even though the anterior cruciate ligament (ACL) was not torn, there was a concern that it may have been stretched and elongated as well as associated with the anterior tibial spine fracture. This was felt to be the case particularly in those greater than 10 years of age at injury. They concluded that residual symptoms and clinical instabilities were detected at follow-up in 11 of 32 fractures. The symptoms were most marked in severely displaced type III fractures in those initially 10 years of age or greater at the time of injury. Their major concern in treatment was to obtain full reduction of the fracture fragment by hyperextension after aspiration. They were unable to demonstrate in their series, however, "any beneficial effects of open reduction with or without wire fixation." They did not regard this as a contraindication of open reduction in selected cases, but rather as an example of incomplete surgical technique. They even considered the need to countersink the bone fragment to compensate for ACL elongation. Smith reached similar conclusions (294). He assessed 21 cases, of whom 15 were available for detailed follow-up. The 3 with type I or type II fractures had closed treatment, and 12 with type III fractures had open reduction and internal fixation. He found anterior cruciate ligament laxity at long-term follow-up in a large number of cases. Smith also concluded that there was some stretching or elongation of the anterior cruciate ligament even though it remained intact and that the chondro-osseous fracture bore the brunt of the displacement. Experimental study by Noyes et al. demonstrated that, prior to fracture of the tibial spine, the anterior cruciate ligament elongated greater than 50% even though it maintained its continuity (229). Of their 15 patients, 7 were free of symptoms but 8 had varying degrees of pain with associated muscle atrophy, and all had some evidence of anterior cruciate ligament laxity. Willis et al. assessed 97 patients, of whom 50 had a detailed study at least 2 years postinjury (330). There were 8 type I injuries; 42 type II injuries; and 47 type III injuries. The vast majority of type I fractures were treated with a long leg cast in 20 ~ of flexion alone. Virtually all the type II patients were treated with closed reduction, aspiration, and long leg casting. The type III injuries were treated with closed reduction and aspiration plus casting alone or in equal numbers open reduction and internal fixation. They too concluded that no treatment method was measurably superior to another, with closed reduction methods just as effective as open reduction in terms of eventual knee stability. They concluded that anterior tibial eminence fractures should be treated by closed reduction and immobilization in extension. Arthroscopy could be used to insure adequate reduction of the fragment, and open reduction and internal fixation should be reserved for irreducible tibial eminence fracture. The

584

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

long-term prognosis was guarded because of persistent ACL laxity. It is important to recognize that in virtually all studies, even those with good or excellent results, ligamentous laxity was defined but in many instances had not yet resulted in any clinical symptoms. Janarv et al. reported on 61 children with anterior tibial spine (ATS) fractures showing excellent subjective reports with 87% good or excellent and 13% fair (157). In the entire series, there were 8 type I injuries, 28 type II, and 25 type III. All type I injuries were treated in cast only. There were 7 closed reductions in the type II group, with the rest being treated with cast alone. In the type III group, 9 were treated in cast alone, and closed reduction was done in 3, open reduction in 3, and open reduction and internal fixation in 10. They concluded that "ATS fracture in childhood seems to be a relatively benign injury in the long term." In type I fractures, immobilization alone was sufficient. In type II fractures, they recommended closed reduction by hyperextension followed by casting. In type III fractures, they recommended open anatomical reduction and stable internal fixation both to insure optimal repositioning as well as to allow assessment of possible underlying meniscal or osteochondral injuries. McLennan assessed 10 patients with type III intercondylar eminence fracture, with 6 years of follow-up, by repeat arthroscopy based on recurring symptoms (209). He noted long-term morbidity of extension loss, chondromalacia, quadriceps weakness, and ligamentous instability in type III fractures. He felt that a loss of reduction was common after closed reduction and that open reduction and internal fixation, appropriate tensioning of the cruciate mechanism, and aggressive rehabilitation were essential. 3. AVULSION FRACTURES OF THE TIBIAL TUBEROSITY The Watson-Jones classification of tibial tubercle avulsion fractures is used most commonly (323) (Fig. 30A). He classified the disorders into three major categories, types I, II, and III, based on the extent of involvement of the proximal tibial epiphysis as well as the degree of displacement of the fracture fragments. The type I injury is an avulsion of the small separate ossification fragment of the distal portion of the tuberosity, which may be displaced dorsally and proximally. In type II injuries the entire segment formed by the tibial tuberosity epiphysis is hinged and displaced proximally with the angulation axis at the level of the proximal tibial epiphysis without being completely fractured at its proximal base. In type III injuries, the fracture line extends into the proximal tibial epiphysis and passes through the articular surface into the joint and, thus, represents a true Salter-Harris type III fracture-separation. The fractures occur in the narrow range of 13-16 years of age. Slight modifications of this classification have been made by Ogden et al. (Table VIII and Fig. 30B), who described two variants for each type, thus constructing a type IA and IB, type IIA and IIB, and type IIIA and IIIB categorization (231). Closed reduction is warranted in type I and II injuries, but only if

A

/ Type l

Type II

Type III

!

3

3BI

F I G U R E 30 The (A) Watson-Jones and (B) Ogden classifications for tibial tubercle fractures of childhood are shown. [Part A reprinted from Christie and Dvonch (1981), J. Pediatr. Orthop. 1:391-394, 9 Lippincott Williams & Wilkins, with permission. Part B reprinted from (231), with permission.]

hinge displacement is noted. Open reduction is done if bone fragment repositioning cannot be achieved or held (120, 256). Anatomic reduction must be considered the primary objective in type III fractures, and most type III injuries are treated with open reduction and internal fixation. The bone fragment is held with a compression screw and patellar tendon repair is stabilized by tension band wiring (256). There tends to be relatively little deformity due to growth arrest because the age of occurrence is late, around the time that the physis itself is closing and the anterior tibial tubercle is fusing to the anterior tibial metaphysis. An excellent review by Gaudier and Bouret in 1905 (120) demonstrated good awareness of this injury even in the

SECTION Vl ~ Clinical Features of Acute Epiphyseal Fracture-Separations TABLE VIII Type IA Type IB Type IIA

Type IIB Type IIIA Type IIIB

585

O g d e n et, aL Classification of Tibial Tuberosity Fractures"

There is only minimal displacement of the fractured tibial tuberosity. There is greater displacement of the tibial tuberosity although it remains hinged proximally rather than fully displaced. The fracture line is more extensive, involving not only the tibial tubercle but the adjacent anterior and inferior part of the secondary ossification center of the proximal tibia. The fracture is hinged proximally. There is comminution of the fragment although hinging persists proximally. The fracture involves the tibial tuberosity but crosses the proximal tibial growth plate to pass through the secondary ossification center and the articular surface into the joint. The same fracture pathway occurs but the tuberosity fragment is comminuted.

aFrom reference (231).

nineteenth century. The type III fracture was well-known, with involvement of the proximal tibial epiphysis and fracture into the tibial plate and articular joint surface. Open reduction and wire-staple fixation were reported. Henard and Babo reviewed cases in the literature from 1935 to 1980 comprising 76 fractures in 72 patients (143). The male: female ratio was 4.5:1.0. Most type III fractures were observed in 15- to 17-year-old patients, whereas type I and II fractures were noted most from 12 to 14 years of age. For this reason, virtually no reports of growth disturbance of the proximal tibial epiphysis have been reported even after the severe intra-articular type III fractures. Genu recurvatum and lower extremity length discrepancies in particular have not been reported. Bolesta and Fitch recorded a typical profile with 16 cases, all in boys, at an average age of 15.2 years (range = 10-17 years), of whom 4 had underlying OsgoodSchlatter disease (29). There were 3 IB, 6 IliA, and 7 IIIB lesions. All were treated successfully with open reduction and internal fixation. Hand et al. described 7 cases, all in males aged 14-16 years, with 1 type I and 6 type III injuries (137). Each of the type III lesions had open reduction and internal fixation with screws, pins, or wires. All healed uneventfully. In 8 cases, reported by Christie and Dvonch, there was no growth discrepancy or angular deformity in either type I, II, or III injuries (63). The average age of occurrence in that study was 15 years. Chow et al. updated literature reports in 1990 to 150 fractures in 145 patients (62). They reported on 16 patients with good final results in all patients, with full ranges of motion, good strength, and no growth sequelae. Age of involvement was at an average of 13.3 years (range = 10-16 years) with a 13:3 (4.3:1) male:female ratio. Driessnack and Marcus reported a displaced fracture of an unossified tibial tubercle in a 12-year-old boy treated with good early resuits by suturing through adjacent bone and periosteum (91). The matter pertaining to genu recurvatum requires the following observation rarely alluded to in the literature. After extensive lower extremity trauma in the first decade of life, prior to ossification of the tibial tubercle, and in association with femoral and/or tibial diaphyseal or even distal

femoral physeal fractures, there can be nonrecognized proximal tibial tubercle area trauma that only presents 2-3 years later as anterior physeal growth arrest with slowly developing and usually severe genu recurvatum. Pappas et al. (244) and Aitken (9) have alluded to such occurrences. The key feature is the early age at occurrence, first decade of life, and the generally unrecognized primary injury. L. Distal Tibia 1. DISTAL TIBIA FRACTURE-SEPARATIONS The distal tibia is the second most common site for growth plate fracture-separation, after the distal radius. Slightly more than 50% of distal tibial fracture-separations are of the type II pattern, but all five types are represented (Table VIIE). Two specific subtypes of fracture at this physis are now recognized widely: the adolescent Tillaux fracture, a type III pattern involving the lateral half of the distal tibial epiphysis, and the triplane fracture, in which radiographs in the anteroposterior projection show a type III lesion and in the lateral projection a type II lesion. These two fractures are referred to as transitional fractures of the distal tibia because they occur exclusively at the end of the adolescent period when the physis is undergoing its unique pattern of closure and final maturation (320). Spiegel et al. correlated age versus fracture type (Fig. 31) (299). Approximately 10-12% of distal tibial growth plate injuries have growth complications. Most type I lesions result in good to excellent healing, and most type V lesions lead to fair to poor results. Detailed reports of type II, III, and IV lesions show a scattering of results with a fairly uniform distribution between excellent, good, fair, and poor in each type (283, 286). The frequency, variable patterns, and growth arrest problems of distal tibial fracture-separations were well-illustrated in detailed presentations by Giuliani (123) and Bartl (18). Bartl reviewed 235 cases seen over 25 years. Pure epiphyseal separation accounted for 37.5% (S-H type I), separation with diaphyseal fracture 35.8% (S-H type II), separation with epiphyseal fracture 12.8% (S-H

CHAPTER 7

586

16

9 Epiphyseal Growth Plate F r a c t u r e - S e p a r a t i o n s

-

6

15 T

""

6

:

15

14 > bJ r~

"

QI3 I(/)

44

I l

I l

;

t

,

i

i

j./, /

!!

;,

;i

, ,

I I o o

i j , ,

', i o ;

,'

,'

, t o '

; ~ 1 / ;

u

'

~

I

~

I

I I

II

!/

/i

'

t

I I

I ,

o

i

t

J. -

;

,

t

/ 0

!

t

I

~

;

I i

i

,

J

I

' I

~ / ~"

i

~ ~"

I I

I

3 r '. : ,

!

12

z_ t o

T L

~o

3 ..6 / J n , o

< b.I >.

T , o

43 _

I

Is T ,

~

i

',

=

',

"

I

j. I !

W

9 ,L -

.L

TYPE-][ FIBULA

TRIJUVENILE TYPE-Z TYPE-I'g' TYPE-In' TYPE-Z TYPE-X PLANE TILLAUX T Y P E - X TIBIA TIBIA TIBIA FIBULA TIBIA TIBIA TIBIA TIBIA (MEDIAL (CENTRAL) CORNER)

TYPE

OF

FRACTURE

F I G U R E 31

Types of fracture-separation at the distal tibia and their relation to age. [Reprinted from (299), with permission.]

type III), and separation with both epiphyseal and diaphyseal fractures 14.2% (S-H type IV). Of particular interest are the line drawings of radiographs from dozens of cases in both projections showing the variable patterns of fracture with the pathoanatomic gradings. Distal tibial fracture-separations lead to the second highest frequency of bone bridge formation after the distal femur (250). Chadwick and Bentley reported growth retardation in 8 of 28 (26.8%) distal tibial epiphyseal injuries. Dotter and McHolick reported an 18% incidence of negative growth sequelae (89). One of the reasons for this occurrence is the medial physeal irregularity jutting into the metaphysis, which frequently is damaged allowing for transphyseal vessel communication and bone formation. This local anatomic feature has been called Kump's bump or Poland's bump by some in reference to its previous descriptions (180) (Fig. 32). Growth arrests can be seen with each of type II, III, and IV injuries, with especially high complication rates in types III and IV [Cass and Peterson (58); Cooperman et al. (69); Karrholm et al. (167-173); Kling et al. (178); Landin et al. (185, 186); Nolan et al. (226); Spiegel et al. (299, 300)]. Growth plate convexity as seen in the lateral radiograph, and the proximal deviation of a localized segment of the plate in its medial one-third as seen in the anteroposterior radiograph, can result in B2 injuries by predisposition to crushing of the plate with type I and II displacements. In radiologic type III lesions, transverse fractures that histologically are at the metaphyseal level and

type IV injuries with considerable crushing can lead to epiphyseal-metaphyseal bone union of the B2 pathophysiologic type, especially if anatomic reduction is not achieved. Reports on distal tibial fracture-separations have noted the growth arrest problems and the evidence of diminished problems with accurate open reduction and internal fixation. A detailed study of 55 distal tibial and fibular epiphyseal fractures indicated that important prognostic features regarding subsequent growth arrest were the type of treatment, degree of displacement, and age at injury, whereas the Salter-Harris classification system alone "could not significantly predict the growth pattern" [Karrholm et al. (172, 173)]. Both tomography (Fig. 32B) and CT scanning have been extremely helpful in defining more clearly the actual pattern of fracture within epiphyseal and metaphyseal bone [Feldman et al. (103); Spiegel et al. (299); Von Laer (320)]. The MR imaging studies are beginning to demonstrate early transphyseal vessel communication between epiphysis and metaphysis (159, 286, 292) (Figs. 19E-19I and 36H). Hynes and O'Brien have shown that careful examination of plain radiographs following fracture repair to assess growth arrest or disturbance lines can help define the likelihood of future problems (153). 2. P A T T E R N OF C L O S U R E OF D I S T A L T I B I A L PHYSIS Many of the unique fracture patterns of the distal tibial epiphysis relate to the fact that there is not uniform closure of the physis toward the end of skeletal growth but rather medial plate closure occurring prior to lateral closure. The normal fusion pattern in the distal tibial epiphysis has been documented well by MacNealy et al. (Fig. 33) (201). Closure occurs initially at the medial-central portion adjacent to the area of the proximal medial deviation of the physis and then involves the entire medial segment. The earliest closure thus occurs throughout the entire medial segment and is a little more prominent anteriorly than posteriorly. The pattern of closure then moves in a lateral direction, but the entire time for full fusion can take up to 18 months. For a considerable period of time, therefore, the medial physis is closed while the lateral physis is open. Ogden and McCarthy also studied the pattern of growth plate maturation at the distal tibia (234). They too noted closure from medial to lateral sides over an extended several-month period of time; the physiological epiphyseal arrest began medially over the malleolus and then extended laterally. As a general finding, central and medial closure is seen beginning at 12.5 years, completion of medial closure is noted at 13 years, closure moving to the lateral segments occurs at 13.5 years, and complete closure is seen at 14 years. 3. TRANSITIONAL FRACTURES Von Laer has discussed the transitional fractures, the biplane fracture or the juvenile fracture of Tillaux and the triplane fracture, and feels that the injuries are caused by external rotation and eversion (320). The occurrence of a particular type depends on the maturity of the physis and not

SECTION Vl ~ Clinical Features o f Acute Epiphyseal Fracture-Separations

587

F I G U R E 32 (A) Anteroposterior ankle radiographs demonstrate a type III fracture-separation of the distal medial tibial epiphysis in an 11-year-old male. The degree of displacement is difficult to determine. (B) A tomogram of the distal tibial epiphysis shown in A shows the clear-cut separation of the epiphyseal fragment from the adjacent epiphysis. The question is raised as to whether the transverse fracture line passes through the hypertrophic zone, which is less worrisome, or is present within the outer reaches of the metaphysis, which is a more worrisome situation as illustrated in Fig. 15. The normal medial physeal irregularity (arrow) referred to by some as Kump's bump is seen clearly by tomography. (C) Treatment of a similar fracture by intra-epiphyseal AO compression screw to restore articular continuity.

on the m e c h a n i s m of trauma. In his study of the two groups, the average age of females was 13.3 years (range = 11-15 years) and that of the m a l e s was 14.7 years ( 1 2 - 1 6 years). In both types, the m o r e lateral the e p i p h y s e a l fracture line, the m o r e frequently the medial physis was closed. Dias and Giegerich also felt that, in both types of transitional fracture, the pattern of distal tibial physeal closure strongly influenced the type of fracture in the adolescent with external rotation injuries of the foot in relation to the leg (87).

F I G U R E 33 Pattem of closure of distal tibial physis is shown. The dotted areas represent closure on the upper or transverse sections. Closure progresses from medial to lateral sides. [Reprinted from MacNealy et al. (1982). Am. J. Roentgenol. 138:683-689, with permission from the American Roentgen Ray Society.]

a. J u v e n i l e Fracture o f Tillaux. T h e j u v e n i l e or adolescent fracture of Tillaux is a type III epiphyseal f r a c t u r e separation occurring at the distal tibia during the time f r a m e in w h i c h the distal medial tibial physis has closed while the lateral s e g m e n t r e m a i n s open (Fig. 34). Von L a e r refers to

F I G U R E 34 A Tillaux juvenile distal tibial fracture-separation is seen. This pattern occurs when the medial physis has closed while the lateral remains open.

588

CHAPTER 7

9

Epiphyseal Growth Plate Fracture-Separations

this injury as a biplane fracture. CT scans confirm an epiphyseal component only with no metaphyseal involvement. Tillaux performed adult cadaveric experiments in which he found that the tibial fragment of bone was produced by an avulsion force from the pull of the intact anterior inferior tibial-fibular ligament (312). This fracture is referred to by some as the juvenile or adolescent fracture of Tillaux. Kleiger and Mankin reported a large series of patients with this injury, and many subsequent reports have appeared (177). In the large majority of instances, closed reduction is successful but open reduction and AO screw fixation of the displaced fragment lead to anatomic positioning and excellent results. Growth arrest is not a problem because fusion of the physis is already occurring, and indeed the medial half has already fused. These injuries are relatively rare because of the limited time frame during which they can occur. Stefanich and Lozman (302) described 5 cases, Dias and Tachdjian (86) described 3 in a series of 71 ankle fractures, Karrholm et al. (173) reported 17 in 361 cases of distal tibial injury, and Molster et al. reported on 6 fractures as well (218). This fracture represents one of the few that is more common in gifts. The juvenile fracture of Tillaux is an avulsion fracture of the anterolateral portion of the distal tibial epiphysis caused by the pull of the intact anterior inferior tibial fibula or anterior syndesmotic ligament sustained during an external rotation injury to the foot. The displacement of the fragment is either lateral or anterolateral. Rang has indicated that the majority of these fractures are displaced so minimally that no reduction is required and short leg casting is appropriate (263). Long-term studies, however, do show some instances of articular incongruity such that open reduction and internal fixation readily are used. Internal fixation using a compression screw certainly is warranted wherever the degree of displacement is 2 mm or greater on any radiographic projection. CT or tomography is most helpful in providing clearer resolution than plain X rays. b. Distal Tibial Triplane Epiphyseal Fracture-Separation. The triplane fracture also owes its pattern to the unusual configuration of the distal tibial growth plate and, more importantly, the variability of its closure across the transverse diameter. It is essentially a type IV fracture-separation and, as such, almost invariably requires open reduction and internal fixation for optimal results. In its simplest description, it appears like a type III fracture on the anteroposterior ankle radiograph and like a type II fracture on the lateral (Figs. 35A and 35B). Von Laer has shown that the epiphyseal fracture line in biplane or triplane fractures can be intramalleolar, medial, central, or lateral as far as the anterior syndesmosis (320). The wandering fracture line has sagittal, transverse (horizontal), and coronal (frontal) components. Marmor defined the components of this fracture in 1970 (205), although Bartl had alluded to it in 1957 (17). Lynn described two cases and was the earliest to coin the term "triplanar fracture" (200). The triplane fracture is characteristically composed of two, three, and occasionally four frag-

A ~

fracture through epiphysis

po

fracture through metaphysis

~ A

fracture through

growth plate

t

med

B

ant int

med

F I G U R E 35 The triplane fracture appears as a type III fracture on the anteroposterior radiograph and a type II fracture on the lateral radiograph. Because it passes from articular cartilage through epiphysis, physis, and metaphysis it is, in three-dimensional considerations, a type IV fracture. (A) A two-part triplane injury; (B) a three-part injury. (C) The value of CT imaging in defining fracture pattern for the triplane injury is shown. At left, the type IV and type III patterns are seen, in middle, a type IV configuration, and at right, a type II pattern [three sagittal plane images from same ankle injury]. [Reprinted from Karrholm et al. (1981), J. Pediatr. Orthop. 1:181-187, 9 Lippincott Williams & Wilkins, with permission.]

ments with or without fibular fractures. The first large series was reported by Cooperman et al. (69). According to Cooperman, the two-fragment triplane fracture has the tibial shaft, medial malleolus, and anteromedial portion of the epiphysis as one fragment, with the second fragment consisting of the remainder of the metaphysis, lateral epiphysis, and attached

SECTION Vl ~ Clinical Features o f Acute Epiphyseal Fracture-Separations

fibula (Fig. 35A). In the three-fragment triplane fracture, the pattern is similar, but the anterolateral epiphyseal segment is free rather than lying attached to the distal tibial regions. Fragment 1 is rectangular and represents the anterolateral quadrant of the distal tibial epiphysis, fragment 2 consists of the medial and posterior portions of the epiphysis in addition to a posterior metaphyseal spike, and fragment 3 is the tibial metaphysis (Fig. 35B). A four-fragment distal triplane fracture has also been described on occasion. The fourth fragment is the medial tibial epiphyseal segment separated along the physis as well as by the intra-articular component (246, 300). Clement and Worlock showed that the triplane fracture can occur when the distal tibial physis is still completely open (66). Tomograms and CT scanning are the most important for clear delineation of these injuries. Ertl et al. diagnosed 11 three-fragment and only 4 two-fragment fractures using CT, tomography, and observation at open reduction, whereas a group of examiners having no knowledge of these findings and viewing only the plain radiographs diagnosed 11 twofragment and 4 three-fragment patterns (100). Due to the occurrence of this fracture near the end of skeletal growth, angular deformity and limb length discrepancy are not clinical problems (58). Of 13 patients with a triplane fracture studied by Cass and Peterson, none had growth arrest problems. The open reduction and internal fixation are done to restore joint surface anatomy. Karrholm et al. described ankle fractures of the distal tibial epiphysis as comprising four stages: juvenile Tillaux fracture comprising stages 1 and 2; stage 3 is a triplane fracture without a fibula fracture; and stage 4 is a triplane fracture with either two, three, or four fragments along with a fibula fracture (167). Dias and Tachdjian felt that the mechanism of injury was pure external rotation of the foot without either pronation or supination (86). Most observers feel that the juvenile Tillaux fracture is produced by the same mechanism of injury as the triplane form, with the mechanism being supination eversion. Studies of increasing numbers of patients from 1970 on have clarified treatment approaches. The results in general with this fracture are now very good to excellent. Growthrelated problems are extremely rare because of the relatively late age of occurrence in relation to the time of skeletal maturity. The prognosis thus is dependent on the accuracy of the reduction. Both closed and open methods have been used widely, and in each series excellent results in general are reported. In an extensive review of the literature by Rapariz et al., in which they studied several series, the distributions of closed and surgical treatment were noted (264). If one adds their series to that of Landin et al. (186), in a total of 217 cases reported 133 had either no need for reduction or closed reduction only, and 84 had operative intervention with open reduction and internal fixation. A consensus clearly is building for an approach that can be stated as fol-

589

lows: CT imaging is most helpful for defining not only the pattern of fracture and the degree of displacement but the number of fragments. Plain films can be misleading in this regard. If there is no displacement, a long leg cast is appropriate. Closed reduction can be attempted in those situations with displacement. It is important to obtain anatomic or nearanatomic reduction, however, with long-term studies showing that displacement of 2 mm or more in any projection can predispose one to late arthritic changes. If closed reduction does not meet these criteria or if displacement occurs in the first few weeks of immobilization, then open reduction and internal fixation using epiphyseal and metaphyseal compression screws are warranted. Because the long-term results with anatomic reduction appear long-standing, there should be little to no reluctance to resort to open reduction and internal fixation with these injuries. Initial studies showed excellent results with both closed and open reduction. When studies progress, however, beyond 5 years, those ankles with greater than 2 mm of displacement begin to show early degenerative changes. Because this is an intra-articular fracture with the intra-articular component in the central regions well within the major weight bearing areas, the importance of anatomic reduction is evident. In a long-term follow-up of 35 patients with a triplane fracture in which the mean followup was 5 years 2 months (range = 24-162 months), there was no pain in 35, normal walking in 35, full job performance in 35, anatomically perfect radiographs in 33 of 35 (94%), and full ankle joint function in 31 of 35 (88%). There was no varus or valgus deformity, no limb length discrepancy, and only an external rotation deformity greater than 10% in 3 of 35 (8%) (186). 4. CASE ILLUSTRATIONS OF TWO TYPE III MEDIAL FRACTURE-SEPARATIONS A case of growth arrest following a type III distal tibial medial epiphyseal fracture-separation and its management are illustrated in Fig. 36. Figure 32C illustrates a fracture treated by open reduction and AO compression screw fixation, which healed uneventfully. 5. TYPE IIl FRACTURE OF MEDIAL MALLEOLUS WITH INTRA-ARTICULAR OSTEOCHONDRAL FRAGMENT Beaty and Linton presented an interesting case in which a medial malleolar type III distal tibial epiphyseal fracture in a 9-year-old girl was found associated with an intra-articular osteochondral fragment (20). This was seen on an oblique plain radiograph and by tomography. Open reduction served not only to realign the physis and articular surface but to excise the osteochondral fragment. 6. TYPE IV FRACTURES OF THE CLASSIC PATTERN Type IV fracture-separations of the distal tibial epiphysis of the classic or uniplanar pattern can occur. Virtually all of these will involve the medial malleolus. There is a tendency for these injuries to be difficult to diagnose on plain

590

CHAPTER 7 ~

Epiphttseal Growth Plate Fracture-Separations

F I G U R E 36 A sequence of radiographs shows growth arrest following a type III fracture-separation of the distal tibial medial epiphysis. (A) An anteroposterior radiograph with no evidence of injury, but (B) an oblique view clearly shows the minimally displaced type III fracture-separation. (C, D) Treatment was by long leg cast with no reduction. (E) At 3 months postinjury the physis is irregular but no definitive bone bridge is seen. (F) By 6 months mild varus tilt of the distal tibia is seen along with medial tibial physeal closure. Note the distal tibial growth arrest line angling toward the medial bone bridge (arrow). (G) By 9 months the varus tilt is more marked. (I-I) MR imaging defines the medial bone bridge (arrow) with marrow signal continuity across the physis between epiphyseal and metaphyseal bone. (I-L) Distal tibial-fibular valgus osteotomy with completion of physeal closure led to restoration of normal alignment at skeletal maturity.

SECTION VI ~ Clinical Features o f Acute Epiphyseal Fracture-Separations

radiographs. Cass and Peterson noted that, in 18 ankles with a fracture involving the medial malleolus, extension of the fracture into the metaphysis could often be appreciated only on oblique radiographs or with tomography and CT scanning (58). In 9 of 18 tibias with a fracture of the medial malleolus, premature partial closure of the distal physis occurred. This led to angular deformity or limb length discrepancy sufficient to require epiphyseal arrest, osteotomy, or bone bridge excision. There was a marked tendency for the classic type IV fracture separations to occur in a younger and often much younger age group than the triplane fractures. They are also due to inversion and crush injuries and, thus, damage the physeal cartilage to a greater extent. Because of the tendency toward displacement and the high incidence of growth-related problems, the recommendation is for a perfect anatomic reduction, which generally implies open reduction and internal fixation. 7. LATE RESULTS IN DISTAL TIBIAL PHYSEAL FRACTURE-SEPARATIONSWITH AN INTRA-ARTICULAR COMPONENT Landin et al. studied a subset of patients with childhood distal tibial physeal fractures to assess patients with physeal injuries and an intra-articular component (186). They thus separated out 78 patients from 373 who had Salter-Harris lesions types III and IV of the medial malleolus, Tillaux fractures (Salter-Harris type III of the anterolateral part of the distal tibial epiphysis), and triplane fractures. The patients had all been treated during a time frame in which there was an awareness of the fact that these injuries were in a relatively high-risk category and, thus, either accurate closed reduction or open reduction and internal fixation were carefully utilized. The results in general were quite favorable. Part of this was due to the accurate repositioning and part was due to the fact that most fractures occurred at an age when the remaining growth potential of the physis was small, such that premature growth arrest was not clinically significant. In addition, most of the injuries were caused by low-energy trauma. They concluded that the intra-articular component of the fracture should be reduced exactly, preferably with open reduction and internal fixation if needed. The majority of symptomatic ankles had had a Tillaux or triplane fracture. The distribution of fractures numerically showed 14 type III lesions of the medial malleolus, 6 type IV lesions of the medial malleolus, 17 Tillaux fractures, and 28 triplane fractures.

M. Proximal and Distal Fibula Proximal fibula growth plate fractures are rare and unless associated with massive open trauma do not lead to growth problems. Havranek reported on six fractures all caused by automobiles injuring pedestrians (142). There were four girls and two boys involved at an average age of 11.9 years (range = 9.5-15.3 years). Fracture types were three type II,

591

one type III, and two type IV, but all healed uneventfully. One of the six had open reduction and wire fixation; the rest were treated in cylinder casts. The distal fibula frequently is the site of undisplaced type I fractures, with types II-V virtually never seen. Isolated type I lesions, type A by the pathophysiologic classification, all appear to heal uneventfully without growth arrest. Distal fibula fractures associated with fracture-separations of the distal tibia occasionally develop premature fusion. Isolated distal fibula physeal fractureseparations are more common than statistical studies indicate. Due to the generally benign nature of the injury virtually all are treated on an outpatient basis, often symptomatically only, without specific delineation of whether the distal fibular discomfort is a sprain, an epiphyseal type I fracture, or an adjacent metaphyseal injury.

N. Growth Patterns Following Distal Tibial and Fibular Growth Plate Fracture-Separations Using Roentgen Stereophotogrammetry Karrholm and associates performed a series of detailed studies involving the mechanism of injury and growth pattern after epiphyseal growth plate ankle injuries in children (167173). They utilized the Roentgen stereophotogrammetric technique to assess growth. This method, originally developed by Selvik, involved the placement of spherical tantalum balls within the bones, and subsequently the positional relationships between each were followed radiographically with time. At operation or after fracture healing, spherical tantalum balls 0.5 mm in diameter were inserted either at the time of open surgery or percutaneously under local anesthesia. On the fractured side, 3-5 balls were inserted at each side of the growth plate in the distal tibia with 1 ball inserted on each side of the growth plate in the distal fibula as well as in the distal tibia on the intact side. Patients were followed with radiographic examination 1-3 weeks after implantation of markers and then 1 month later. Subsequent exams were performed at intervals of 3-6 months during the first year and 6-12 months in the second and later years if follow-up was indicated. Standardized radiographic technique was used, following which the films were evaluated in a precision instrument for aerial photogrammetry. This method thus allowed for detailed studies of the relative extent of growth of the distal tibial and fibular epiphyses in relation to the normal side. The second aspect of the work by Karrholm and associates involved the assessment of growth pattern in relation to the mechanism of injury. They utilized schemes for interpreting adult ankle fracture mechanisms and applied them in detail to childhood fractures for the first time. They thus related the growth pattern to supination-eversion injuries and supination-adduction injuries, as well as to pronationabduction and pronation-eversion. They were able to define five types of growth pattern following injury. These involved normal growth, initial growth stimulation of varying time length, initial and temporary growth retardation, initial

592

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

and permanent growth retardation, and initial and permanent growth arrest. They felt that the Roentgen stereophotogrammetric method in conjunction with the anatomictraumatological classification permitted early determination of the growth pattern with high accuracy and demonstrated growth disturbances months before they would be evident from plain radiographs. Their summary showed that prediction of a specific growth pattern could be made only with low accuracy but separation of the growth patterns into two groups--one with symmetrical growth, initial and temporary growth retardation and growth stimulation, and the other with progressive growth retardation and growth arrestm could be made. They felt that in 67% of cases, when they compared the predicted outcome with the actual result, the correct classification was made. They felt that the SalterHarris classification alone could not significantly predict the growth pattern.

O. Ligament Damage Following Distal Femoral and Proximal Tibial Physeal Fractures Bertin and Goble performed a retrospective study of 29 cases of physeal fracture about the knee to assess the degree of ligament stability. The average age at injury for the distal femoral fractures was 12 years 5 months, and the average age at injury in the proximal tibial fractures was 14 years 4 months. Follow-up was at a mean of 66 months postfracture. They found that 6 of 16 patients with distal femoral fractures and 8 of 13 patients with proximal tibial fractures had associated ligamentous laxity to static testing at followup evaluation. This represented an incidence of 48%. They concluded that physeal fracture does not preclude ligament damage and that ligamentous assessments, though difficult, should be part of initial assessments as well as posthealing assessments and rehabilitation programs (24).

VII. T R A U M A T I C D A M A G E T O G R O W T H P L A T E S BY P A T H O L O G I C , C H R O N I C REPETITIVE, AND INDIRECT EFFECTS

A. Pathologic Epiphyseal Growth Plate Fracture-Separations 1. MYELOMENINGOCELE Pathologic epiphyseal growth plate fracture-separations are seen in patients with myelomeningocele. These have been likened to a Charcot neuroarthropathy disorder in which repeated damage to a physeal region is allowed to continue because the patient has insufficient ability to appreciate pain and continues to use the affected limb. There have been several studies documenting the occurrence of epiphyseal growth plate fracture-separation in patients with myelomeningocele, stressing in particular the often negative

sequelae. One of the earliest descriptions in the English literature is the paper by Gillies and Hartung in 1938 (122). Gyepes et al. drew clear attention to the entity of marked skeletal changes at the metaphyseal-physeal junctions of the lower extremities in meningomyelocele patients who remained ambulatory (131). The characteristic radiographic changes involved (1) widened cartilaginous epiphyseal plates, (2) irregular, dense, slightly widened metaphyses, and (3) sub-periosteal metaphyseal-diaphyseal new bone formation. To this list can be added (4) minimal to absent epiphyseal displacement, (5) prolonged healing time, and (6) a high incidence of premature growth plate fusion. Repetitive trauma with continuing mobility and limited immobilization were felt to be the keys to the pathogenesis of the clinical and radiological findings. It is evident that muscular paralysis alone does not lead to these changes because they are never seen in such disorders as poliomyelitis or other muscular disorders in which the sensation remains intact. It is the insensitivity to pain that leads to a Charcot-type neuroarthropathy or neurochondrosteopathy. The radiologic changes described will regress once appropriate and relatively prolonged immobilization is instituted. Not infrequently, biopsy of the metaphyseal region has been undertaken with the suspicion that the disorder represented either an osteomyelitis or a sarcomatous process. In the paper by Soutter, a bone marrow metaphyseal biopsy was performed immediately adjacent to the physis, and the marrow space was found to be occupied by callus and evidence of healing fractures (297). Wenger et al. reviewed the records of 244 patients with spina bifida and documented a 23% incidence of at least 1 fracture of a lower limb, including 8 epiphyseal fractures (2.8%) (326). The pattern outlined in their study was similar to that in previous and subsequent papers. The majority of epiphyseal fractures were of either the type I or type II variety. The characteristic findings involved a widening of the radiolucent area of the involved physis, irregularity of the metaphyseal layer of bone, absent to only minimal displacement of the epiphysis, and a prominent metaphyseal periosteal reaction that often led to the concern about an osteomyelitis or neoplastic condition such as osteosarcoma, Ewing's sarcoma, or leukemia. The fractures described all involve the lower extremities because of the nature of the meningomyelocele lesion sparing the upper extremities. Each of the four major lower extremity physes can be affected, involving the proximal and distal femur and proximal and distal tibia. The clinical signs involve swelling and redness about the joint region, although discomfort is minimal. Premature closure of the involved physis occurred in 5 of 9 patients. Because the fractures occur generally in those less than 12 years of age and often early in the first decade, such injuries often lead to significant lower extremity length discrepancies, which complicate an already problematic situation. The fractures generally are late in being diagnosed and tend to take much longer to heal than their counterparts in an otherwise normal child. Because of the delay in union and

SECTION VII 9 Traumatic Damage to Growth Plates the high degree of complication, immobilization in a solid cast for a 2-month period is almost always the recommendation of the various studies. Kumar et al. described 5 physeal fracture-separations in 16 patients with myelomeningocele (181). Four of these involved a distal tibial physis and one the proximal tibial physis. There was no clear-cut history of trauma. X rays revealed an increase in the width of the physis and an irregularity at the physeal-metaphyseal bone junction. The physeal fractures were most common in the lower lumbar level ambulatory group. Healing was much slower in the physeal injuries than in diaphyseal or metaphyseal fractures. The injury should be well-protected in a cast for a period of 8-12 weeks, and weight bearing should not be allowed until early signs of union are clear. Lock and Aronson described seven epiphyseal growth plate fractures in a large group of myelomeningocele patients (197). All metaphyseal and diaphyseal fractures healed satisfactorily whereas the seven fractures that involved the physeal plate were a major problem with three showing delayed union and two premature growth arrest. Four of the fractures were type I and three were type II. Roberts et al. reported five examples of physeal growth plate fracture-separation in four children with myelomeningocele and also noted physeal widening as a prime radiologic indicator of the disorder (272). Each of the injuries was un-displaced and there was no periosteal new bone formation, indicating early diagnosis. Protection from further trauma resulted in rapid clinical resolution. They felt, however, that with early diagnosis the period of immobilization could be shorter than previous studies had indicated, and usually a 4-week period was sufficient. Fromm et al. studied a large number of patients with myelomeningocele, assessing 82 children out of a group of 947 who had sustained a total of 224 fractures (118). Among these were 3 distal femoral epiphyseal separations (1.4%), 3 proximal tibial epiphyseal separations (1.4%), and 6 distal tibial epiphyseal separations (2.7%). In 1 of the 3 cases at the distal femoral epiphysis, there was premature closure of the epiphyseal growth plate, with premature closure also occurring in 1 of 3 proximal tibial and 2 of 6 distal tibial fracture-separations. Edvardsen pointed out the frequency of physeal damage in patients with myelomeningocele (94). He noted the broadening and loosening of the physis due to repetitive trauma. He observed 6 physeal disorders in 50 children from 2 to 7 years of age. Because the observations frequently were made by chance without history of either trauma or discomfort, the number seen probably underrepresented those that occurred. Physeal regions of patients with myelomeningocele could be traumatized during daily walking activities and in association with therapy such as passive joint movements. Quilis reported three cases of fracture-separation of the epiphysis, one of which suffered premature closure of the growth plate (261). The radiographic appearance of these

593

fractures in the epiphyseal and metaphyseal regions can be confused with many disorders, including osteomyelitis, bone sarcoma, bone syphilis, scurvy, neurotropic joints, and battered child syndrome. Multiple case reports of the occurrence of the lesion with its characteristic finding of widening of the epiphyseal line and abundant periosteal reaction with new bone formation were made by Soutter (297) and Golding (125). Golding in particular commented on the similarity of the findings to a Charcot joint. The epiphyseal line was much wider than normal, irregularity of bone appearance occurred in the metaphysis adjacent to the widened physis, and periosteal elevation led to considerable new bone formation in the metaphyseal region. Physeal slippage was almost always minimal in extent. Rodgers et al. reviewed 19 chronic physeal fractures in 13 patients with myelodysplasia. In only 3 cases could the injury be ascribed to a traumatic event (273). Indeed, many of the patients had extreme delay in diagnosis. The most common site was the distal tibia in 10, followed by chronic epiphyseal injury in the distal femur in 4, the proximal femur in 3, and the proximal tibia in 2. All patients were treated with prolonged immobilization averaging 5.8 months (range = 3-18 months) using either braces or casts. Four of the fractures required operative fixation. All injuries had healed at 4.8 years follow-up, but in 4 of the fractures premature growth plate fusion was noted. In each of the many studies reported, fractures are seen most commonly in the distal femur, proximal tibia, and distal tibia, with some also occurring in the proximal femur in higher lesions. In occasional instances, several growth plate fracture-separations occur in the same patient. Thus, it is important to follow patients with myelomeningocele carefully, especially when excessive walking or physical therapy treatments lead to swelling in the joint regions. If an epiphyseal fracture-separation is noted, it would be wise to radiograph each of the physeal areas of the lower extremities to make certain that other injury disorders had not occurred. 2. OTHER PATHOLOGIC DISORDERS Epiphyseal growth plate fracture-separations of a pathologic nature are rare except in myelomeningocele or neurological disorders similar to it in which continuing motor function occurs in the presence of absent or markedly diminished sensation. In most reports of epiphyseal slipping in the literature, there is no clear history of trauma and the disorder generally is attributed to repetitive low-grade trauma with either walking or therapy in which slippage occurs, but with normal pain sensation not being present continuing activity leads to worsening injury. Epiphyseal slipping in relation to other disorders is seen most frequently with slipped capital femoral epiphysis, and the large number of disorders with which proximal femoral slipping can be associated is discussed in detail in Chapter 5. There are two other pathologic entities that can be associated

594

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

with slipped epiphyses in areas other than the proximal femur. When such disorders occur, they are associated almost exclusively with renal osteodystrophy (see Chapter 10), although rare instances of slipping with scurvy have been reported.

B. Premature Physeal Closure Following Seemingly Unrelated Fractures of the Ipsilateral Diaphysis and Metaphysis There is increasing recognition of the occurrence of premature physeal fusion following seemingly unrelated diaphyseal or metaphyseal fractures on the same side. Whereas reports generally have been small with 1-3 patients reported, Hresko and Kasser were able to describe 7 children with this occurrence over a 3-year period from 2 children's orthopedic services (147). The large majority of patients are 11-14 years of age. The limb length discrepancy is often significant because follow-up generally was discontinued with repair and rehabilitation of the shaft fracture because there was no awareness by the treating physician of any physeal injury. In the patients to be reported later from several series, a total of 19 patients with premature physeal closure are described of whom 8 were males and 11 females. The mean age at occurrence in the males was 11 years 10 months (range = 10-14 years) and the mean age in the gifts was 11 years (9-12 years of age). In the entire series, therefore, the mean age at occurrence was 11 years 5 months (range = 9-14 years). Morton and Starr described asymmetrical closure of the upper tibial epiphysis following fracture of the distal onethird of the tibial shaft without apparent injury to the epiphysis (221). The affected epiphyses closed prematurely, resulting in progressive hyperextension deformity of the knee. One patient was an 11-year-old girl and the other a 13year-old boy. As part of their treatment, both had proximal tibial K-wires placed but the wires were not near the tibial tubercle being described as 1 and 1.5 in. or greater distal to the tubercle. The deformity became evident in the second year following injury and in both instances was major in relation to shortening and angular deformity. The authors describe the fractures of the lower one-third of the tibia as being severe although the course of treatment had seemed unremarkable. They also referred to a case of Smillie in a young girl who suffered an open fracture of the tibial shaft complicated in a late stage of recovery by a supracondylar fracture of the femur and then the discovery sometime later of a genu recurvatum caused by closure of the anterior portion of the upper tibial epiphysis. Hunter and Hensinger described multiple premature growth arrests on the ipsilateral side in an 11-year-old girl who, following major trauma, suffered a comminuted spiral fracture of the proximal one-third of the shaft of the fight femur (149). She was treated in a hip spica for 5.5 months. During the year following the injury, she was noted to have early closure of almost all epiphyseal plates of the fight lower extremity (proximal and distal femur, proximal and distal tibia, and proximal fibula) except for the distal fibula.

The growth arrest lines indicated that the closure had occurred shortly after the original injury. The closure phenomenon was so marked that premature fusion of the epiphysis of the greater trochanter and the iliac crest also occurred. Eighteen months following injury, the fight lower extremity was 5 cm shorter than the left. Abram and Thompson noted marked wrist deformity after premature closure of the distal radial physis following what appears to have been an uncomplicated torus fracture of the distal radial metaphysis in a 10-year-old girl (3). Hresko and Kasser described seven patients with long bone diaphyseal fractures who subsequently developed premature physeal closure around the knee (147). In each instance, the growth deformity was marked. There were two femoral neck fractures, two femoral subtrochanteric region fractures, a femoral diaphyseal fracture, and two tibiofibular diaphyseal fractures. Skeletal traction (which can cause premature epiphyseal arrest if the pin is placed directly into a physeal area) was used for only three patients, two of whom had physeal arrests in the bone other than the one in which the traction pin was placed, whereas in the other the traction pin was shown clearly by X ray to be well distal to the proximal tibial physis in which eventual fusion occurred. Beals described three patients who experienced the premature complete physeal closure of the ipsilateral limb following diaphyseal fracture (21). In each instance, a clinically significant length discrepancy occurred. A 14-year-old boy suffered a subtrochanteric fracture of the left femur with subsequent premature closure of the ipsilateral distal femoral and proximal tibial physes. A scanogram at age 20 documented 4.5 cm of lower extremity length discrepancy. A girl 11 years 10 months of age suffered a fracture of the distal fight femur followed by ipsilateral proximal tibial physeal closure with the fight tibia 1.6 cm short 6 months postfracture. A boy 11 years 3 months of age suffered fractures of the distal fight femur and middle fight tibia and experienced premature closure of the proximal femoral physis and the proximal distal tibial physes. Closure of the distal femoral physis was due to direct trauma. Twenty-one months after the fracture, scanograms demonstrated 8.8-cm shortening of the fight limb. Bowler et al. described two cases of premature closure of the anterior portion of the proximal tibial physis with associated genu recurvatum in a 12-year-old boy and an 11-yearold boy, both of whom suffered fight femoral shaft fractures (31). Proximal tibial osteotomies were required in both. Paul et al. reported a 9-year-old girl with a displaced fracture of the distal one-third of the radius and ulna (245). Treatment by closed reduction under general anesthesia appeared unremarkable and healing was uneventful. Three years later, she presented with 5-cm shortening of the ulna and evidence of complete arrest of the distal ulna epiphysis. Aminian and Schoenecker described two fractures of the distal radius apparently with lack of involvement of the distal radial physis, which led to subsequent complete arrest of the adjacent growth plate (14). One injury occurred in a 12-

SECTION VII ~ Traumatic Damage to G r o w t h Plates

year-old girl and the other in an 11-year-old girl who were treated with closed reduction in the first instance and casting without reduction for a torus fracture of the distal radius and a fracture of the ulnar styloid process in the second. The 12year-old girl presented 2 years later with a markedly prominent distal ulna due to premature partial closure of the distal radial physis. The 11-year-old girl presented 2 years later with progressive prominence of the distal left ulna and radiographic evidence of a premature growth arrest of the distal radius. Both eventually required corrective surgery. Although the phenomenon increasingly is well-documented, it still remains relatively rare and unfortunately no definitive understanding of the pathophysiology is available. The possibility exists that the physis is injured in an undetectable way at the time of diaphyseal or metaphyseal fracture, but this alone would appear to be insufficient cause. The age at occurrence of the phenomenon is quite narrow, however, with a mean occurrence at 11 years and virtually all patients between 10 and 13 years of age. It is known that long bone fractures speed up the overall development of the entire bone, presumably in relation to the increased vascularity of the entire bone associated with the reparative response, and the fact that this is occurring during the growth spurt at a time when relatively little growth is remaining might well make the physeal regions vulnerable to premature closure.

C. Physeal Separation Due to the Stress Injury Caused by Chronic Repetitive Activity Over the past few decades, the widespread participation of children and adolescents with open growth plates in intensive athletic training programs, especially where chronic repetitive exercises are done, has led to the recognition of a syndrome that appears compatible with an un-displaced stress separation of the overused physeal area. These appear to be most common at the distal radial physis, but convincing reports have also been presented of occurrences at the proximal humerus and distal femur. An early report by Adams defined the abnormality in the proximal humeral epiphysis of five boys involved in baseball pitching with extensive activity (4). He considered the disorder to be an osteochondrosis, but the symptoms and radiographic descriptions appear compatible with what appears to be a physeal separation. Of the five boys described, two were 13 years old, one 14 years old, and two 15 years old. The principal clinical finding was pain in the shoulder at the end of a hard throwing motion. X rays showed a characteristic widening of the proximal humeral physis and demineralization and fragmentation adjacent to the physis without evidence of avascular bone necrosis. The discomfort quietened almost immediately with rest and avoidance of activity and the X ray shortly regained normal structure. Comparative X ray studies of both shoulders showed demineralization and marked widening of the epiphyseal line in the pitching arm. The symptoms were relieved and the X ray became normal over the few months following discontinuation of the activity.

595

Dotter reported a similar case in a 12-year-old pitcher with similar shoulder symptoms and X-ray changes, and he diagnosed the lesion as a fracture through the epiphyseal cartilage plate (88). Cahill et al. reported six cases of stress fracture of the proximal humeral epiphyseal plate in 11- and 12-year-old male baseball pitchers (50). Symptoms were as described previously with pain and inability to perform. Xray changes showed widening of the epiphyseal line with metaphyseal bone fragmentation and irregularity. Stress fracture through the distal femoral epiphysis in athletes also was reported by Godshall et al. with two cases (124). Chronic, strenuous, and repetitive exercise in relation to sports led to the disorder in two males 14 and 15 years of age. The 14-year-old had pain in his knee for several months on an intermittent basis with acute worsening immediately prior to his assessment. The physis in particular of the distal lateral femur was widened and a diagnosis of stress fracture of the physis was made. This healed uneventfully with rest using crutches for 3 weeks and decreased physical activity for 3 months. Follow up X rays showed complete healing of the stress fracture through the epiphyseal line. In the second case, pain developed over a 4-week period in the knee region. X rays also showed widening of the distal femoral epiphyseal line. Crutch use and diminution of activity for 12 weeks led to complete healing of the physis radiographically. In neither case did premature physeal fusion occur with resumption of more controlled activity. Most reported cases have involved the distal radial physis in gymnasts. Roy et al. reported 21 cases involving stress changes of the distal radial epiphysis in young gymnasts (277). The X ray changes appear consistent with un-displaced stress fractures of the distal radial physis. In this series, no residual growth-related problems were observed. Treatment involved wrist immobilization until the symptoms disappeared followed by increased rest and gradual reinstitution of activity at a less intensive level. In their series, 11 had radiographic changes and in these recovery took at least 3 months; in a second group, 10 had similar symptoms but no radiographic changes and they recovered within an average of 4 weeks. The mean age at time of diagnosis was 12 years with a range from 10 to 17 years. Nineteen of the 21 were female. The amount of activity was extensive, and in the 21 patients 17 worked out at least 6 hr per day, 6 days per week. The characteristic radiographic findings of nondisplaced physeal separation in relation to repetitive stress are widening of the growth plate of the distal radial epiphysis, particularly on the radial and volar aspects, cystic changes usually of the metaphyseal aspect of the epiphyseal plate associated with an increase in irregularity in the physeal-metaphyseal bone margin, and occasional haziness within the usually radiolucent area of the epiphyseal plate. Fliegel reported three cases of stress-induced widening of the distal radial growth plate in adolescent athletes (105). Characteristic radiographic features again were seen involving irregular widening of the growth plate, irregular metaphyseal borders along with flaring of the metaphysis, and,

596

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

on occasion, sub-periosteal new bone formation. Symptoms resolved with diminished activity, although radiographic changes took anywhere from 9 to 24 months to fully resolve. In relatively advanced cases, there is also a tendency to sclerosis of the adjacent metaphysis. Fliegel's cases were in a 14-year-old girl, 14-year-old boy, and 14.5-year-old boy. In two of the instances, X rays returned to normal with rest, but in one, some changes in residual deformity of the metaphyseal area and relative shortening of the radius occurred. Similar stress injury findings of the distal radial growth plate were found in 21 cases by Carter and Aldrich (57). In their group, there were 17 boys (average age 13.5 years) and 4 girls (average age 14 years). Characteristic radiographic findings again involved widening of the distal radial growth plate, particularly on its volar aspect, haziness of the growth plate due to irregularity of the border between the cartilage and the metaphyseal zone, and an increase in the transverse diameter of the metaphysis. The ulnar growth plate can also be irregular. With rest, the haziness and widening of the growth plate resolve, although some irregular metaphyseal scalloping may persist. Physeal injury due to repetitive stress occurs in areas of most rapid growth in which the hypertrophic zone is felt to be most vulnerable. When increased activity is superadded, the physeal separation without displacement occurs. The physis is widened as there is interruption of the normal mineralization process in the metaphysis.

D. Genu Recurvatum after Skeletal Traction Involving Inadvertent Placement of the Proximal Tibial K-Wire through the Tibial Tubercle Physeal Area In using skeletal traction, great attention must be taken not to impinge upon the physeal regions. Skeletal wires through or immediately adjacent to physes can be shown to induce premature bone fusion. The most common area for this complication to occur is the proximal tibia, in which the K-wire usually is placed well distant to the transverse proximal tibial physis but on occasion somewhat anteriorly and thus in the region of the distal extension of the tibial tubercle physis (26, 318).

VIII. M A N A G E M E N T O F N E G A T I V E SEQUELAE OF GROWTH PLATE FRACTURE-SEPARATIONS

A. General Considerations Even with improved diagnosis and treatment of growth plate fracture-separations, some injuries will proceed to growth plate damage and negative growth sequelae. Among the measures to be taken to minimize damage are (1) recognition of non-displaced type I fractures and treatment by cast

immobilization for a minimum of 3 weeks, (2) gentle closed or open reduction of type II distal femoral fractures with metaphyseal fragment-metaphysis wire-pin fixation to minimize postreduction instability, and (3) open anatomic reduction and internal fixation for types III and IV intra-articular injuries with physeal displacement. Once growth plate damage has occurred, transphyseal bone bridge formation serves as a tether that causes either complete or partial cessation of growth. There are three possible major sequelae to a growth plate fracture-separation: shortening, angular deformity, and possible joint surface incongruity (Fig. 11). If there is complete cessation of growth, shortening without angular deformity occurs. If there is only partial or focal bone bridge formation, some shortening usually occurs but the major problem relates to angular deformity as the rest of the growth plate continues to function. Shortening alone is a characteristic of either massive complete bone bridge formation or central bone bridge formation greater than 50% of the diameter of the physis. Angular deformity is characterized by bone bridge formation adjacent to the periphery of the plate. Management of the bone bridge and the sequelae is divided into two time periods, early and late. By late we refer to the situation in which a bone bridge has formed and negative growth sequelae have occurred involving shortening, angular deformity, or joint surface incongruity. The use of the term early refers to the demonstration of a transphyseal bone bridge within the first several months after fracture but prior to the onset of any clinically detectable shortening or angulation. Articular surface irregularity is a problem from the moment it occurs. Management techniques will be described in great detail in Chapter 8 but will be outlined here. Detailed approaches have also been outlined in the literature (236, 250, 284).

B. Management of Early Bone Bridge Formation MR imaging techniques increasingly are able to show the formation of bone bridges early in their evolution within a few months of growth plate fracture. Not all bone bridges are sufficiently large to cause tethering, however. It is wellknown by both clinical and experimental investigations that not all bone bridges are sufficiently extensive to lead to growth plate arrest. The growth force generated by the physis is extremely great, and with many small bridges, such as those with 10% involvement or less, the affected physis can grow away from or stretch out the bone bridge such that no clinical sequelae follow (61,236, 250). Due to the sensitivity of the MR imaging technique it is extremely important to beware this phenomenon and to not operate prematurely for a small physeal bridge that might not cause an arrest. At present it is advisable to follow the patient sufficiently closely that some documentation of actual growth arrest involving slight shortening or early angular deformity is present before transphyseal resection is performed.

References

Once a bone bridge has been defined as clinically problematic, even in the early phase, two possible treatment approaches can be taken. Transphyseal chondrodiatasis even without formal excision of the bridge may be sufficiently effective to allow for this approach. If not, one can perform early transphyseal bone bridge resection and interposition of either fat or other tissues.

C. Management of Late Sequelae of Bone Bridges" Bone Bridge Excision, Physeal Interposition Materials, and Transphyseal Chondrodiatasis The management of the late sequelae includes management of the bone bridge, the angular deformity and any shortening. Management of the bone bridge is dependent on several considerations: the position of the bone bridge, the amount of physeal cartilage replaced by the bone bridge, and the amount of growth remaining in the affected physis. If it is felt that growth is worth preserving, then bone bridge resection can be considered. The general rule for resection of a bone bridge is that it involves only one-third of the affected growth plate or less. Tomography can be used to determine the extent of the bone bridge, although CT scans and MR imaging are used increasingly for more accurate definition. If angular deformity has occurred, osteotomy to correct that deformity as well as bone bridge resection is needed. Bone bridge resection and interposition tissues are discussed in Chapter 8. If it is determined that insufficient growth is remaining to warrant physeal excision, or if the growth plate involvement is too extensive to expect growth to continue following excision, then complete physeal closure is performed surgically to eliminate the chance of worsening angular deformity. Osteotomy is performed to correct angular deformity. One can then either accept the degree of shortening that will occur or elect to perform a physeal arrest on the contralateral side to prevent worsening limb length discrepancy. If there is already a significant limb length discrepancy, then the contralateral physis might have to be fused to prevent worsening of the deformity as well as an additional physis to allow for correction of the discrepancy that has already occurred. If physeal closure will lead to length discrepancies greater than 5 cm, the option of limb lengthening can be considered. On occasion the bone bridge can be excised followed by interposition of a material designed to prevent recurrence or a procedure to regain length using transphyseal chondrodiatasis. Experimental investigations into bone bridge resections, interposition materials, and transphyseal lengthening are discussed in detail in Chapter 8. Langenskiold created defects in rabbit growth plates and demonstrated that fat placed in the defect prevented bone bridge formation because it kept the epiphyseal and metaphyseal vessels separate (187). This procedure has been used

597

clinically with many excellent results [Langenskiold (187); Peterson (250); Williamson and Staheli (329)]. Mallet (202) and Peterson (250) have used methyl methacrylate to bridge the physeal defect, but the method has not gained wide clinical applicability. Many other investigators have confirmed the occurrence of bone bridges with epiphyseal-metaphyseal circulatory communication, generally in the process of describing the effects of various interposition materials in preventing vascular mingling and bone formation. These works include those of Osterman (242) using fat, deep-frozen hyaline rib cartilage, and bone wax, Bright (35) using silastic, Eulert (101) and Lennox et al. (192) using cartilage plugs, and Olin et al. (239) using oriented iliac crest apophyseal cartilage transplants. Some laboratory investigations show possible future clinical promise for vascularized growth plate transplants [Teot et al. (307); Zaleske et al. (335)] and tissue culture growth of chondrocytes prior to their placement into focal defects [Lalanandham et al. (183)]. Another response to a developing growth plate bone bridge involves the use of chondrodiatasis in which pins are placed on either side of the growth plate and gentle distraction is applied with an extemal fixator to pull apart the bone bridge tissue and allow an interposition tissue to reform [A1degheri et al. (13); Canadell and De Pablos (52)]. Caution is warranted, however, at least based on large animal studies because Fjeld and Steen showed growth retardation after lengthening by epiphyseal distraction of 40-70% (104). At the present time this approach is not refined sufficiently to allow for widespread use, but it is beginning to be used clinically in selected instances.

References 1. Abbott LC, Gill GG (1942) Valgus deformity of the knee resulting from injury to the lower femoral epiphysis. J Bone Joint Surg 24:97-113. 2. Abe M, Ishizu T, Nagaoka T, Onomura T (1995) Epiphyseal separation of the distal end of the humeral epiphysis: A follow-up note. J Pediatr Orthop 15:426-434. 3. Abram LJ, Thompson GH (1987) Deformity after premature closure of the distal radial physis following a torus fracture with a physeal compression injury. J Bone Joint Surg 69A: 1450-1453. 4. Adams JE (1966) Little league shoulder: Osteochondrosis of the proximal humeral epiphysis in boy baseball pitchers. Cal Med 105:22-25. 5. Aitken AP (1935) The end results of the fractured distal radial epiphysis. J Bone Joint Surg 17:302-308. 6. Aitken AP (1935) Further observations on the fractured distal radial epiphysis. J Bone Joint Surg 17: 922-927. 7. Aitken AP (1936) The end results of the fractured distal tibial epiphysis. J Bone Joint Surg 18:685-691. 8. Aitken AP (1936) End results of fractures of the proximal humeral epiphysis. J Bone Joint Surg 18:1036-1041. 9. Aitken, AP (1965) Fractures of the proximal tibial epiphyseal cartilage. Clin Orthop Rel Res 41:92-97.

598

CHAPTER 7 ~

Epiphyseal Growth Plate Fracture-Separations

10. Aitken AP, Magill HK (1952) Fractures involving the distal femoral epiphyseal cartilage. J Bone Joint Surg 34A:96-108. 11. Aitken AP, Ingersoll RE (1956) Fractures of the proximal tibial epiphyseal cartilage. J Bone Joint Surg 38A:787-796. 12. Albanese SA, Palmer AK, Kerr DR, Carpenter CW, Lisi D, Levinsohn EM (1989) Wrist pain and distal growth plate closure of the radius in gymnasts. J Pediatr Orthop 9:23-28. 13. Aldegheri R, Trivella G, Lavini F (1989) Epiphyseal distraction. Hemichondrodiatasis. Clin Orthop Rel Res 241:128-136. 14. Aminian A, Schoenecker PL (1995) Premature closure of the distal radial physis after fracture of the distal radial metaphysis. J Pediatr Orthop 15:495-498. 15. Badelon O, Bensahel H, Mazda K, Vie P (1988) Lateral humeral condylar fracture in children: A report of 47 cases. J Pediatr Orthop 8:31-34. 16. Banagale RC, Kuhns LR (1983) Traumatic separation of the distal femoral epiphysis in the newborn. J Pediatr Orthop 3: 396-398. 17. Barrett WE Almquist EA, Staheli LT (1984) Fracture separation of the distal humeral physis in the newborn. J Pediatr Orthop 4:617-619. 18. Bartl R (1957) Die traumatische epiphysenlosung am distalen ende des schienbeines und des wadenbeines. Hefte Unfallheilk 54:228-257. 19. Baxter ME Wiley JJ (1986) Fractures of the proximal humeral epiphysis. Their influence on humeral growth. J Bone Joint Surg 68B:570-573. 20. Beaty JH, Linton RC (1988) Medial malleolar fracture in a child. A case report. J Bone Joint Surg 70A: 1254-1255. 21. Beals RK (1990) Premature closure of the physis following diaphyseal fractures. J Pediatr Orthop 10:717-720. 22. Bensahel H, Csukonyi Z, Badelon O, Badaoui S (1986) Fractures of the medial condyle of the humerus in children. J Pediatr Orthop 6:430-433. 23. Bergenfeldt E (1933) Beitrage zur kenntnis der traumatischen epiphysenlosungen an den langen rohrenknochen der extremitaten. Acta Chir Scand 73:Supp 28, 1-422. 24. Bertin KC, Goble EM (1983) Ligament injuries associated with physeal fractures about the knee. Clin Orthop Rel Res 177:188-195. 25. Bidder A (1873) Experimente uber die kunstliche hemmung des langenwachstums von rohrenknochen durch reizung und zerstorung des epiphysenknorpels. Archiv f Exp Path Pharm 1:248-263. 26. Bjerkreim I, Benum P (1975) Genu recurvatum: A late complication of tibial wire traction in fractures of the femur in children. Acta Orthop Scand 46:1012-1019. 27. Blair W, Hanson C (1979) Traumatic closure of the triradiate cartilage. J Bone Joint Surg 61A: 144-145. 28. Blount WP (1955) Fractures in Children. Baltimore: Williams and Wilkins. 29. Bolesta MJ, Fitch RD (1986) Tibial tubercle avulsions. J Pediatr Orthop 6:186-192. 30. Bousseau M (1867) Disjonction epiphysaire traumatique de la tete du femur et des epines iliaques anterieures. Mort. Autopsie. Bull Soc Anat Pads 12:283-286. 31. Bowler JR, Mubarak SJ, Wenger DR (1990) Tibial physeal closure and genu recurvatum after femoral fracture: Occurrence without a tibial traction pin. J Pediatr Orthop 10:653-657.

32. Bragdon RA (1965) Fractures of the distal radial epiphysis. Clin Orthop Rel Res 41:501-504. 33. Brashear HR, Jr (1959) Epiphyseal fractures. A microscopic study of the healing process in rats. J Bone Joint Surg 41A: 1055-1064. 34. Bret J, Curtillet J (1892) Du decollement traumatique des epiphyses. La Prov Med 6:40-43. 35. Bright RW (1974) Operative correction of partial epiphyseal plate closure by osseous-bridge resection and silicone-rubber implant: An experimental study in dogs. J Bone Joint Surg 56A:655-663. 36. Bright RW, Burstein AH, Elmore SM (1974). Epiphysealplate cartilage. A biochemical and histological analysis of failure modes. J Bone Joint Surg 56A:688-703. 37. Broca P (1852) Sur quelque points de l'anatomie pathologique du rachitisme. Bull Soc Anat Paris 27:141-167, 542-596. 38. Broca A (1904) Le mecanisme des decollements epiphysaires. Assoc Fr Chir 17:626-634. 39. Broca A (1905) Mecanisme des decollements epiphysaires. Press Med 13:140-141. 40. Broker FHL, Burbach T (1990) Ultrasonic diagnosis of separation of the proximal humeral epiphysis in the newborn. J Bone Joint Surg 72A:187-191. 41. Bruns P (1881) Ueber traumatische epiphysentrennung. Archiv f Klin Chir 27:240-264. 42. Bryan WJ, Tullos HS (1979) Pediatric pelvic fractures: Review of 52 patients. J Trauma 19:799-805. 43. Buch P, Folschveiller J, Jenny G (1966) Uber die behandlung von 376 vorderarmschaftbruchen bei kinder. Hefte Unfallheilk 89:51-54. 44. Bucholz RW, Ezaki M, Ogden JA (1982) Injury to the acetabular triradiate physeal cartilage. J Bone Joint Surg 64A: 600-609. 45. Burgos-Flores J, Gonzalez-Herranz P, Lopez-Mondejar A, Ocete-Guzman JG, Smaya-Alarcon S (1993) Fractures of the proximal humeral epiphysis. Intemat Orthop 17:16-19. 46. Burkhardt SS, Peterson HA (1979) Fractures of the proximal tibial epiphysis. J Bone Joint Surg 61A:996-1002. 47. Burkus JK, Ogden JA (1984) Development of the distal femoral epiphysis: A microscopic morphological investigation of the zone of Ranvier. J Pediatr Orthop 4:661-668. 48. Burman MS, Langsam MJ (1939). Posterior dislocation of lower femoral epiphysis in breech delivery. Arch Surg 38: 250-260. 49. Bylander B, Aronson S, Egund N, Hansson LI, Selvik G (1981) Growth disturbance after physeal injury of distal femur and proximal tibia studied by Roentgen stereophotogrammetry. Arch Orthop Traum Surg 98:225-235. 50. Cahill BR (1977) Stress fracture of the proximal tibial epiphysis: A case report. Am J Sports Med 5:186-187. 51. Campbell CJ, Grisolia A, Zanconato G (1959) The effects produced in the cartilaginous epiphyseal plate of immature dogs by experimental surgical traumata. J Bone Joint Surg 41A:1221-1240. 52. Canadell J, De Pablos J (1985) Breaking bony bridges by physeal distraction. Intemat Orthop 9:223-229. 53. Canale ST, Bourland WP (1977) Fracture of the neck and intertrochanteric region of the femur in children. J Bone Joint Surg 59A:431-443.

References 54. Canton E (1861) Two cases of excision of the knee-joint for forcible separation of the lower epiphysis from the shaft of the femur. Dublin Quart J Med Sci 31:74-83. 55. Carlson WO, Wenger DR (1984) A mapping method to prepare for surgical excision of a partial physeal arrest. J Pediatr Orthop 4:232-238. 56. Carothers CO, Crenshaw AH (1955) Clinical significance of a classification of epiphyseal injuries at the ankle. Am J Surg 89:879-887. 57. Carter SR, Aldridge MJ (1988) Stress injury of the distal radial growth plate. J Bone Joint Surg 70B:834-836. 58. Cass JR, Peterson HA (1983) Salter-Harris type-IV injuries of the distal tibial epiphyseal growth plate, with emphasis on those involving the medial malleolus. J Bone Joint Surg 65A: 1059-1070. 59. Cassebaum WH, Patterson AH (1965). Fractures of the distal femoral epiphysis. Clin Orthop Rel Res 41:79-91. 60. Chadwick CJ, Bentley G (1985) Prognosis and classification of epiphysial injuries. J Bone Joint Surg 67B: 148-149. 61. Chadwick CJ (1982) Spontaneous resolution of varus deformity at the ankle following adduction injury of the distal tibial epiphysis. J Bone Joint Surg 64A:774-776. 62. Chow SP, Lam JJ, Leong JCY (1990) Fracture of the tibial tubercle in the adolescent. J Bone Joint Surg 72B:231-234. 63. Christie MJ, Dvonch VM (1981) Tibial tuberosity avulsion fracture in adolescents. J Pediatr Orthop 1:391-394. 64. Chung SMK (1976) The arterial supply of the developing proximal end of the human femur. J Bone Joint Surg 58A:961-970. 65. Chung SMK, Batterman SC, Brighton CT (1976) Shear strength of the human femoral capital epiphyseal plate. J Bone Joint Surg 58A:94-103. 66. Clement DA, Worlock PH (1987) Triplane fracture of the distal tibia: A variant in cases with an open growth plate. J Bone Joint Surg 69B:412-415. 67. Compere EL (1935) Growth arrest in long bones as result of fractures include the epiphysis. J Am Med Assoc 105: 2140-2146. 68. Conner AN, Smith MGH (1970) Displaced fractures of the lateral humeral condyle in children. J Bone Joint Surg 52B: 460-464. 69. Cooperman DR, Spiegel PS, Laros GS (1978) Tibial fractures involving the ankle in children. The so-called triplane epiphyseal fracture. J Bone Joint Surg 60A: 1040-1046. 70. Comil V, Coudray P (1904) Fractures du cartilage de conjugaison, fractures juxta-epiphysaires et fractures des extremities osseuses. Au point de vue experimental et histologique. Arch Med Exp Anat Path 16:257-344. 71. Comil V, Coudray P (1904) Reparation des fractures et decollements du cartilage de conjugaison; frequence des reccourcissments consecutifs aces traumatismes. Assoc Fr de Chir 17:604-617. 72. Crabbe WA (1963) The treatment of fracture-separation of the capitular epiphysis. J Bone Joint Surg 45B:722-726. 73. Criswell AR, Hand WL, Butler JE (1976) Abduction injuries of the distal femoral epiphysis. Clin Orthop Rel Res 115: 189-194. 74. Czitrom AA, Salter RB, Willis RB (1981) Fractures involving the distal epiphyseal plate of the femur. Internat Orthop 5: 269-277.

599

75. Dale GG, Harris WR (1958) Prognosis of epiphysial separation. An experimental study. J Bone Joint Surg 40B:116-122. 76. Dameron TB, Reibel DB (1969) Fractures involving the proximal humeral epiphyseal plate. J Bone Joint Surg 51A: 289-297. 77. Davis DR, Green DP (1976) Forearm fractures in children: Pitfalls and complications. Clin Orthop Rel Res 120:172-184. 78. Davison BL, Weinstein SL (1992) Hip fractures in children: A long-term follow-up study. J Pediatr Orthop 12:355-358. 79. DeBoeck H, Casteleyn PP, Opdecam P (1987) Fracture of the medial humeral condyle. J Bone Joint Surg 69A: 1442-1444. 80. DeCampo JF, Boldt, DW (1986) Computed tomography of partial growth plate arrest: Initial experience. Skelet Radiol 15:526-529. 81. DeJager LT, Hoffman EB (1991) Fracture-separation of the distal humeral epiphysis. J Bone Joint Surg 73B:143-146. 82. DeLee JC, Wilkins KE, Rogers LF, Rockwood CA (1980) Fracture-separation of the distal humeral epiphysis. J Bone Joint Surg 62A:46-51. 83. Delens E (1884) Des decollements traumatiques de l'epiphyse inferieure du femur. Arch Gen Med 1:272-291 and 426-445. 84. DePablos J, Franzreb M, Barrios C (1994) Longitudinal growth pattern of the radius after forearm fractures conservatively treated in children. J Pediatr Orthop 14:492-495. 85. Dias JJ, Lamont AC, Jones JM (1988) Ultrasonic diagnosis of neonatal separation of the distal humeral epiphysis. J Bone Joint Surg 70B:825-828. 86. Dias LS, Tachdjian MO (1978) Physeal injuries of the ankle in children. Classification. Clin Orthop Rel Res 136:230-233. 87. Dias LS, Geigerich CR (1983) Fractures of the distal tibial epiphysis in adolescence. J Bone Joint Surg 65A:438-444. 88. Dotter WE (1953) Little leaguer's shoulder-fracture of the proximal humeral epiphyseal cartilage due to baseball pitching. Guthrie Clin Bull 23:68-72. 89. Dotter WE, McHolick WJ (1953) The results of treatment of traumatic injuries to the distal tibial epiphyseal cartilage. Guthrie Clin Bull 22:165-173. 90. Downs DM, Wirth CR (1982) Fracture of the distal humeral chondroepiphysis in the neonate. Clin Orthop Rel Res 169: 155-158. 91. Driessnack RP, Marcus NW (1985) Fracture of an unossified tibial tubercle. J Pediatr Orthop 5:728-730. 92. D'Souza S, Vaishya R, Klenerman L (1993) Management of radial neck fractures in children: A retrospective analysis of one hundred patients. J Pediatr Orthop 13:232-238. 93. Duben W, Gelbke H (1955) Epiphysendurchnagelungen. Z Orthop 87:108-118. 94. Edvardsen P (1972) Physeo-epiphyseal injuries of lower extremities in myelomeningocele. Acta Orthop Scand 43: 550-557. 95. Edwards PH, Jr, Grana WA (1995) Physeal fractures about the knee. J Am Acad Orthop Surg 3:63-70. 96. Ehalt W (1958) Knochenbruche bei kindem und erwachsenen. Arch Klin Chir 289:391-394. 97. Ehlers PN, Eberlein H (1964) Epiphysenfrakturen. Klinischer beitrag zur frage der spatfolgen. Arch Klin Chir 305:213-230. 98. Ekengren K, Bergdahl S, Ekstrom G (1978) Birth injuries to the epiphyseal cartilage. Acta Radiol Diag 19:197-204.

600

CHAPTER 7 ~

Epiphyseal Growth Plate Fracture-Separations

99. Eliason EL, Ferguson LK (1934) Epiphyseal separation of the long bones. Surg Gyn Obstet 58:85-99. 100. Ertl JP, Barrack RL, Alexander AH, VanBuecken K (1988) Triplane fracture of the distal tibial epiphysis. J Bone Joint Surg 79A:967-976. 101. Eulert J (1979) Transplantation du cartilage de la crete iliaque apres desepiphysiodese: etude experiementale. Rev Chir Orthop 65:65-75. 102. Farabeuf M (1886) Sur 1'epiphyse inferieure de l'humerus et son decollement traumatique. Bull Mem Soc Chir (Pads) 12: 692-698. 103. Feldman F, Singson RD, Rosenberg ZS, Berdon WE, Amodio J, Abramson SJ (1987) Distal tibial triplane fractures: Diagnosis with CT. Radiology 164:429-436. 104. Fjeld TO, Steen H (1990) Growth retardation after experimental limb lengthening by epiphyseal distraction. J Pediatr Orthop 10:463-466. 105. Fliegel CP (1985) Stress related widening of the radial growth plate in adolescents. Eur Soc Pediatr Radiol 29:374-376. 106. Flynn JC (1989) Nonunion of slightly displaced fractures of the lateral humeral condyle in children: An update. J Pediatr Orthop 9:691-696. 107. Fontenelle J (1825) Observation d'un decollement des condyles du femur. Arch Gen Med 9:267-270. 108. Ford LT, Key JA (1956) A study of experimental trauma to the distal femoral epiphysis in rabbits. J Bone Joint Surg 38A: 84-92. 109. Forlin E, Guille JT, Kumar SJ, Rhee KJ (1992) Transepiphyseal fractures of the neck of the femur in very young children. J Pediatr Orthop 12:164-168. 110. Foster DE, Sullivan JA, Gross RH (1985) Lateral humeral condylar fractures in children. J Pediatr Orthop 5:16-22. 111. Foucher (1860) Recherches sur la disjonction traumatique des epiphyses. Mon des Sci Med Pharm 2:713-717. 112. Foucher JTE (1984) The classic. Separations of the epiphyses. Clin Orthop Rel Res 188:3-9. 113. Fowles JV, Kassab MT (1986) Observations concerning radial neck fractures in children. J Pediatr Orthop 6:51-57. 114. Friberg KSI (1979) Remodeling after distal forearm fractures in children. I. The effect of residual angulation on the spatial orientation of the epiphyseal plates. Acta Orthop Scand 50: 537-546. 115. Friberg KSI (1979) Remodeling after distal forearm fractures in children. II. The final orientation of the distal and proximal epiphyseal plates of the radius. Acta Orthop Scand 50: 731-739. 116. Friberg KSI (1979) Remodeling after distal forearm fractures in children. III. Correction of residual angulation in fractures of the radius. Acta Orthop Scand 50:741-749. 117. Friedenberg ZB (1957) Reaction of the epiphysis to partial surgical resection. J Bone Joint Surg 39A:332-340. 118. Fromm B, Pfeil J, Carstens C, Niethard FU (1992) Fractures and epiphysiolyses in children with myelomeningocele. J Pediatr Orthop 1:21-27. 119. Funsten RV (1948) Epiphyseal separation at the lower end of the radius. South Med J 41:501-504. 120. Gaudier, Bouret (1905) De 1'arrachement de la tuberosite anterieure du tibia. Rev Chir 32:305-334. 121. Gelbke H, Ebert G (1953) Tierexperimentelle studie an der verletzten epiphysenfuge. Z Orthop 83:201-219.

122. Gillies CL, Hartung W (1938) Fracture of the tibia in spina bifida vera. Radiology 31:621-623. 123. Giuliani K (1952) Spatzustande nach traumatisch-mechanischen schadigungen der epiphyse am distalen tibiaende. Arch Orthop Unfall-Chir 45:386-394. 124. Godshall RW, Hansen CA, Rising DC (1981) Stress fractures through the distal femoral epiphysis in athletes: A previously unreported entity. Am J Sports Med 9:114-116. 125. Golding C (1960) Spina bifida and epiphyseal displacement. J Bone Joint Surg 42B:387-389. 126. Golz RJ, Grogan DP, Greene TL, Belsole RJ, Ogden JA (1991) Distal ulnar physeal injury. J Pediatr Orthop 11: 318-326. 127. Gomes LSM, Volpon JB, Goncalves RP (1988) Traumatic separation of epiphyses: An experimental study in rats. Clin Orthop Rel Res 236:286-295. 128. Goyrand G (1836) De la fracture, par contre-coup, de l'extremite inferieure du radius. J Hebd des Prog des Sci Med 1:161-183. 129. Gronkvist H, Hirsch G, Johansson L (1984) Fracture of the anterior tibial spine in children. J Pediatr Orthop 4:465-468. 130. Gueretin (1837) Recherches sur le decollement spontane et traumatique des epiphyses. La Presse Med 1:289-294, 297301,305-308. 131. Gyepes MT, Newbern DH, Neuhauser EBD (1965) Metaphyseal and physeal injuries in children with spina bifida and meningomyeloceles. Am J Roent 95:168-177. 132. Haas SL (1917) The relationship of the blood supply to the longitudinal growth of bone. Am J Orthop Surg 15:157-171. 133. Haas SL (1917) The relationship of the blood supply to the longitudinal growth of bone. Am J Orthop Surg 15:305-316. 134. Haas SL (1917) The localization of the growing point in the epiphyseal cartilage plate of bones. Am J Orthop Surg 15: 563-586. 135. Haas SL (1919) The changes produced in the growing bone after injury to the epiphyseal cartilage plate. J Orthop Surg 1: 67-99, 166-179, and 226-239. 136. Hallel T, Salvati EA (1977) Premature closure of the triradiate cartilage. A case report and animal experiment. Clin Orthop Rel Res 124:278-281. 137. Hand WL, Hand CR, Dunn AW (1971) Avulsion fractures of the tibial tubercle. J Bone Joint Surg 53A:1579-1583. 138. Hanlon CR, Estes WL, Jr (1954) Fractures in childhoodmA statistical analysis. Am J Surg 87:312-323. 139. Hardacre JA, Nahigian SH, Froimson AI, Brown JE (1971) Fractures of the lateral condyle of the humerus in children. J Bone Joint Surg 53A:1083-1095. 140. Harcke HT, Zapf SE, Mandell GA, Sharkey CA, Colley LA. (1987). Angular deformity of the lower extremity: Evaluation with quantitative bone scintigraphy. Radiology 164:437-440. 141. Hastings H, II, Simmons BP (1984) Hand fractures in children. A statistical anaylsis. Clin Orthop Rel Res 188:120-130. 142. Havranek P (1996) Proximal fibular physeal injury. J Pediatr Orthop 5:115-118. 143. Henard DC, Bobo RT (1983) Avulsion fractures of the tibial tubercle in adolescents: A report of bilateral fractures and a review of the literature. Clin Orthop Rel Res 177:182-187. 144. Hernandez J, Jr, Peterson HA (1986) Fracture of the distal radial physis complicated by compartment syndrome and premature physeal closure. J Pediatr Orthop 6:627-630.

References 145. Holda ME, Manoli A, LaMont RL (1980) Epiphyseal separation of the distal end of the humerus with medial displacement. J Bone Joint Surg 62A:52-57. 146. Holland CT (1929) A radiographical note on injuries to the distal epiphyses of the radius and ulna. Pro Roy Soc Med 22: 695-700. 147. Hresko MT, Kasser JR (1989) Physeal arrest about the knee assoicated with non-physeal fractures in the lower extremity. J Bone Joint Surg 71A:698-703. 148. Hughes LO, Beaty JH (1994) Fractures of the head and neck of the femur in children. J Bone Joint Surg 76A:283-292. 149. Hunter LY, Hensinger RN (1978) Premature monomelic growth arrest following fracture of the femoral shaft. J Bone Joint Surg 60A:850-852. 150. Hutchinson J, Jr (1893) Injuries to the epiphyses and their results. Lecture I. Br Med J ii:53-56. 151. Hutchinson J, Jr (1893) Injuries to the epiphyses and their results. Lecture II. Br Med J ii:1417-1419. 152. Hutchinson J, Jr (1894) Injuries to the epiphyses and their results. Lecture III. Br Med J i:669-673. 153. Hynes D, O'Brien T (1988) Growth disturbance lines after injury of the distal tibial physis: Their significance in prognosis. J Bone Joint Surg 70B:231-233. 154. Ireland J (1933) Late results of separation of an epiphysis. Ann Surg 97:189-203. 155. Jahn P (1892) Beitrage zur kenntriss der histologischen vorgange bei der wachsthums behinderung der rohrenknochen durch verletzungen des intermediarknorpel. Morph Arb von Schwalbe i:241. 156. Jakob R, Fowles JV, Rang M, Kassab MT (1975) Observations concerning fractures of the lateral humeral condyle in children. J Bone Joint Surg 57B:430-436. 157. Janarv P-M, Westblad P, Johansson C, Hirsch G (1995) Longterm follow-up of anterior tibial spine fractures in children. J Pediatr Orthop 15:63-68. 158. Jaramillo D, Shapiro F, Hoffer FA, Winalski CS, Koskinen MF, Frasso R, Johnson A (1990) Posttraumatic growth-plates abnormalities: MR imaging of bony-bridge formation in rabbits. Radiology 175:767-773. 159. Jaramillo D, Hoffer FA, Shapiro F, Rand F (1990) MR imaging of fractures of the growth plate. Am J Roent 155: 1261-1265. 160. Jeffrey CC (1950) Fractures of the head of the radius in children. J Bone Joint Surg 32B:314-324. 161. Jetter G (1872) Ueber die traumatischen epiphysen-losungen am oberen Ende des humerus. Beitr Klin Chir 9:361-384. 162. Jones ERL, Esah M (1971) Displaced fractures of the neck of the radius in children. J Bone Joint Surg 35B:429-439. 163. Jouon E (1902) Contribution a 1'etude des decollements traumatiques des epiphyses des os longs. Rev d'Orthop 3: 217-245. 164. Jouon E (1902) Decollement traumatique de l'epiphyse superieure de l'humerus. Rev d'Orthop 3:291-325. 165. Jouon E (1902) Decollement traumatique de l'epiphyse inferieure de l'humerus. Rev d'Orthop 3:379-414. 166. Jouon E (1902) Decollement traumatique des epiphyses femorales. Rev d'Orthop 3:475-508. 167. Karrolm J, Hansson LI, Laurin S (1981) Computed tomography of intraarticular supination-eversion fractures of the ankle in adolescents. J Pediatr Orthop 1:181-187.

601

168. Karrholm J, Hansson LI, Selvik G (1982) Roentgen stereophotogrammetric analysis of growth pattern after supinationeversion ankle injuries in children. J Pediatr Orthop 2:25-37. 169. Karrholm J, Hansson LI, Laurin S (1982) Supination-eversion injuries of the ankle in children: A retrospective study of radiographic classification and treatment. J Pediatr Orthop 2: 147-159. 170. Karrholm J, Hansson LI, Selvik G (1982) Roentgen stereophotogrammetric analysis of growth pattern after supinationadduction ankle injuries in children. J Pediatr Orthop 2: 271-279. 171. Karrholm J, Hansson LI, Selvik G (1984) Changes in tibiofibular relationships due to growth disturbances after ankle fractures in children. J Bone Joint Surg 66A:1198-1210. 172. Karrholm J, Hansson LI, Svensson K (1983) Prediction of growth pattern after ankle fractures in children. J Pediatr Orthop 3:319-325. 173. Karrholm J, Hansson LI, Laurin S (1983) Pronation injuries of the ankle in children. Retrospective study of radiographical classification and treatment. Acta Orthop Scand 54:1-17. 174. Kennedy JP, Weiner DS (1991) Avascular necrosis complicating fracture of the distal tibial epiphysis. J Pediatr Orthop 11: 234-237. 175. Keret D, Mendez AA, Harcke HT, MacEwen GD (1990) Type V physeal injury: A case report. J Pediatr Orthop 10:545-548. 176. Kirmisson E (1904) Decollements traumatiques des epiphyses. Assoc Fr Chir 17:572-596. 177. Kleiger B, Mankin HJ (1964) Fracture of the lateral portion of the distal tibial epiphysis. J Bone Joint Surg 46A:25-32. 178. Kling TF, Jr, Bright RW, Hensinger RN (1984) Distal tibial physeal fractures in children that may require open reduction. J Bone Joint Surg 66A:647-657. 179. Kohler R, Trillaud JM (1983) Fracture and fracture separation of the proximal humerus in children: Report of 136 cases. J Pediatr Orthop 3:326-332. 180. Kump WL (1966) Vertical fractures of the distal tibial epiphysis. Am J Roentgenol 97:676-681. 181. Kumar SJ, Cowell HR, Townsend P (1984) Physeal, metaphyseal, and diaphyseal injuries of the lower extremities in children with myelomeningocele. J Pediatr Orthop 4:25-27. 182. Lacombes P, Prevot J, Bardoux J (1988) The prognosis of fractures of the lower end of the femur in children and adolescents. A review of 96 cases. Fr J Orthop Surg 2:364-371. 183. Lalanandham T, Ehrlich MG, Zaleske DJ, Deeney VF, Mankin HJ (1990) Viability and metabolism of cartilage transplanted to physeal regions. J Pediatr Orthop 10:450-458. 184. Lam SF (1971) Fractures of the neck of the femur in children. J Bone Joint Surg 53A:1165-1179. 185. Landin LA, Danielsson LG (1983) Children's ankle fractures: Classification and epidemiology. Acta Orthop Scand 54: 634-640. 186. Landin LA, Danielsson LG, Jonsson K, Pettersson H (1986) Late results in 65 physeal ankle fractures. Acta Orthop Scand 57:530-534. 187. Langenskiold A (1975) An operation for partial closure of an epiphyseal plate in children, and its experimental basis. J Bone Joint Surg 57B:325-330. 188. Laurent (1818) Observations pour servir a l'histoire du decollement des epiphyses des os longs. J Comp Dict des Sc Med 1:317-326.

602

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

189. Lee BS, Esterhai JL, Jr, Das M (1984) Fracture of the distal radial epiphysis. Characteristics and surgical treatment of premature post-traumatic epiphyseal closure. Clin Orthop Rel Res 185:90-96. 190. Lee KE, Pelker RR, Rudicel SA, Ogden JA, Panjabi MM (1985) Histologic patterns of capital femoral growth plate fracture in the rabbit: The effect of shear direction. J Pediatr Orthop 5:32-39. 191. Lemperg R, Liliequist B (1970) Dislocation of the proximal epiphysis of the humerus in newborns: Report of two cases and discussion of diagnostic criteria. Acta Pediat Scand 59: 377-380. 192. Lennox DW, Goldner RD, Sussman MD (1983) Cartilage as an interposition material to present transphyseal bone bridge formation: An experimental model. J Pediatr Orthop 3: 207-210. 193. Leser E (1888) Ueber histologische vorgange an der ossificationsgrenze mit besonderer berucksichtigung des verhaltens der knorpelzellen. Arch Mikrosksp Anat 32:214-222. 194. Lichtenberg RP (1954) A study of 2,532 fractures in children. Am J Surg 87:330-335. 195. Lipschultz O (1937) The end-results of injuries to the epiphyses. Radiology 28:223-231. 196. Ljubosic NA (1967) Poraneni jamky kycelniho kloubu u deti. Acta Chir Orthop Traumatal Cech 34:393-400. 197. Lock TR, Aronson DD (1989) Fractures in patients who have myelomeningocele. J Bone Joint Surg 71A:1153-1157. 198. Lodes R (1949) Handdeformitat nach infizierter traumatischer epiphysenlosung am distalen radiusende mit ulnadiaphysenbruch. Arch Orthop Unfallchir 44:38-72. 199. Lombardo SJ, Harvey JP, Jr (1977) Fractures of the distal femoral epiphysis. J Bone Joint Surg 59A:742-751. 200. Lynn MD (1972) The triplane distal tibial epiphyseal fracture. Clin Orthop Rel Res 86:187-190. 201. MacNealy GA, Rogers LF, Hemandez R, Poznanski AK (1982) Injuries of the distal tibial epiphysis: Systematic radiographic evaluation. AJR 138:683-689. 202. Mallet J (1975) Les epiphysiodeses partielles traumatiques de l'extremite inferieure du tibia chez l'enfant. Leur traitement avec desepiphysiodese. Rev Chir Orthop 61:5-16. 203. Mann DC, Rajmaira S (1990) Distribution of physeal and nonphyseal fractures in 2,650 long-bone fractures in children aged 0-16 years. J Pediatr Orthop 10:713-716. 204. Marjolin, Richet, Lefort, Guersant, Giraldes (1865) Discussion on epiphyseal separations, Dec. 6, 1865, Meeting Societe Imperiale de Chirurgie. Gaz des Hospitaux (Paris), Number 147, Dec. 16, 1865. 205. Marmor L (1970) An unusual fracture of the tibial epiphysis. Clin Orthop Rel Res 73:132-135. 206. Mbindyo BS (1979) Considerations on cases of epiphyseal injury observed at Kenyatta national hospital. East Afr Med J 56:431-435. 207. McBride ED, Monnet JC (1960) Epiphyseal fractures of the head of the radius in children. Clin Orthop Rel Res 16: 264-271. 208. McDougall A (1961) Fracture of the neck of the femur in childhood. J Bone Joint Surg 43B:16-28. 209. McLennan JG (1995) Lessons learned after second-look arthroscopy in type III fractures of the tibial spine. J Pediatr Orthop 15:59-62.

210. Mendez AA, Bartal E, Grillot MB, Liu JJ (1992) Compression (Salter-Harris type V) physeal fracture: An experimental model in the rat. J Pediatr Orthop 12:29-37. 211. Meyers MH, McKeever FM (1970) Fracture of the intercondylar eminence of the tibia. J Bone Joint Surg 52:1677-1684. 212. Milgram JW, Lyne ED (1975) Epiphysiolysis of the proximal femur in very young children. Clin Orthop Rel Res 110: 146-153. 213. Miller WE (1973) Fractures of the hip in children from birth to adolescence. Clin Orthop Rel Res 92:155-188. 214. Mischkowsky T, Daum R, Ruf W (1980) Injuries of the distal radial epiphysis. Arch Orthop Traum Surg 96:15-16. 215. Mizuno K, Hirohata K, Kashiwagi D (1979) Fractureseparation of the distal humeral epiphysis in young children. J Bone Joint Surg 61A:570-573. 216. Mizuta T, Benson WM, Foster B K, Paterson DC, Morris LL (1987) Statistical analysis of the incidence of physeal injuries. J Pediatr Orthop 7:518-523. 217. Moen CT, Pelker RR (1984) Biomechanical and histological correlations in growth plate failure. J Pediatr Orthop 4:180-184. 218. Molster A, Soreide O, Solhaug JH, Raugstad TS (1976) Fractures of the lateral part of the distal tibial epiphysis (Tillaux or Kleiger fracture). Injury 8:260-263. 219. Morscher E, Desualles PA, Schenk R (1965) Experimental studies on tensile strength and morphology of the epiphyseal cartilage at puberty. Ann Pediat 205:112-130. 220. Morscher E (1981) Classification of epiphyseal injuries. In: Fractures in Children, ed G Chapehal. pp. 20-22, New York: Thieme-Stratton. 221. Morton KS, Starr DE (1964) Closure of the anterior portion of the upper tibial epiphysis as a complication of tibial shaft fracture. J Bone Joint Surg 46A:570-574. 222. Neer CS, II (1960) Separation of the lower femoral epiphysis. Am J Surg 99:756-759. 223. Neer CS, II, Horwitz BS (1965) Fractures of the proximal humeral epiphyseal plate. Clin Orthop Rel Res 41:24-31. 224. Nelson OA, Buchanan JR, Harrison CS (1984) Distal ulnar growth arrest. J Hand Surg 9A: 164-171. 225. Newman JH (1977) Displaced radial neck fractures in children. Injury 9:114-121. 226. Nolan RA, Meyers NH, Patzakis MJ, Moore TM, Harvey JR, Jr (1978) Tibial epiphyseal injuries. Contemp Orthop 1:11-23. 227. Nonnemann HC (1969) Grenzen der spontankorrectar fehlgeheilter frakturen bei jugend lichen. Lang Arch Klin Chir 324:78-86. 228. Nove-Josserand G (1894) Des troubles de 1'accroissement des os. Consecutifs aux lesions des cartilages de conjugaison. Rev Chir 14:385-403. 229. Noyes FR, DeLucas JL, Torvik PJ (1974) Biomechanics of anterior cruciate ligament failure: An analysis of strain-rate sensitivity and mechanisms of failure in primates. J Bone Joint Surg 56A:236-253. 230. O'Brien PI (1965) Injuries involving the proximal radial epiphysis. Clin Orthop Rel Res 41:51-58. 231. Ogden JA, Tross RB, Murphy MJ (1980) Fractures of the tibial tuberosity in adolescents. J Bone Joint Surg 62A:205-215. 232. Ogden JA (1981) Injury to the growth mechanisms of the immature skeleton. Skelet Radiol 6:237-253. 233. Ogden JA (1982)Skeletal growth mechanism injury patterns. J Pediatr Orthop 2:371-377.

References 234. Ogden JA, McCarthy SM (1983) Radiology of postnatal skeletal development. VIII. Distal tibia and fibula. Skelet Radiol 10:209-220. 235. Ogden JA, Lee KE, Rudicel SA, Pelker RR (1984) Proximal femoral epiphysiolysis in the neonate. J Pediatr Orthop 4: 285-292. 236. Ogden JA (1987) The evaluation and treatment of partial physeal arrest. J Bone Joint Surg 69A:1297-1302. 237. Qgden JA, Ganey T, Light TR, Southwick WO (1993) The pathology of acute chondro-osseous injury in the child. Yale J Biol Med 66:219-233. 238. Oh WH, Craig C, Banks HH (1974) Epiphyseal injuries. Pediatr Clin NA 21:407-422. 239. Olin A, Creasman C, Shapiro F (1984) Free physeal transplantation in the rabbit. An experimental approach to focal lesions. J Bone Joint Surg 66A:7-20. 240. Oilier L (1867) Traite Experimental et Clinique de la Regeneration des Os et de la Production Artificielle du Tissue Osseux, volume 1, chapters 12 and 13. Paris: Masson. 241. Oilier L (1881) De l'entorse juxta-epiphysaire. De ses consequences immediates ou eloignees au point de vue de l'inflammation des os. Rev Chir 1:785-809. 242. Osterman K (1972) Operative elimination of partial premature epiphyseal closure. An experimental study. Acta Orthop Scand Supp 147:1-79. 243. Padovani JP, Rigault P, Raux P, Lignac F, Guyonvarch G (1976) Decollements epiphysaires traumatiques de 1'extremitie inferieure du femur. Rev Chir Orthop 62:211-230. 244. Pappas AM, Anas P, Toczylowski HM (1984) Asymmetrical arrest of the proximal tibial physis and genu recurvatum deformity. J Bone Joint Surg 66A:575-581. 245. Paul AS, Kay PR, Haines JF (1992) Distal ulnar growth plate arrest following a diaphyseal fracture. J Roy Coil Surg Edinb 37:347-348. 246. Peiro A, Aracil J, Martos F, Mut T (1981) Triplane distal tibial epiphyseal fracture. Clin Orthop Rel Res 160:196-200. 247. Peltonen J, Aalto K, Karaharju E, Alitalo I, Gronblad M (1984) Experimental epiphyseal separation by torsional force. J Pediatr Orthop 4:546-549. 248. Peterson CA, Peterson HA (1972) Analysis of the incidence of injuries to the epiphyseal growth plate. J Traum 12:275-281. 249. Peterson HA (1983) Triplane fracture of the distal humeral epiphysis. J Pediatr Orthop 3:81-84. 250. Peterson HA (1984) Partial growth plate arrest and its treatment. J Pediatr Orthop 4:246-258. 251. Peterson HA (1992) Elbow trauma. Physeal injuries of the distal humerus. Orthopedics 15:799-808. 252. Peterson HA, Burkhart SS (1981) Compression injury of the epiphyseal growth plate: Fact or fiction? J Pediatr Orthop 1: 377-384. 253. Peterson HA, Madhok R, Benson JT, Ilstrup DM, Melton LJ, III (1994) Physeal fractures: Part 1 Epidemiology in Olmsted County, Minnesota, 1979-1988. J Pediatr Orthop 14:423-430. 254. Peterson HA (1994) Physeal Fractures: Part 2. Two previously unclassified types. J Pediatr Orthop 14:431-438. 255. Peterson HA (1994) Physeal Fractures: Part 3. Classification. J Pediatr Orthop 14:439-448. 256. Polakoff DR, Bucholz RW, Ogden JA (1986) Tension band writing of displaced tibial tuberosity fractures in adolescents. Clin Orthop Rel Res 209:161-165.

603

257. Poland J (1897) The diagnosis of traumatic separation of the epiphyses. Pediatrics 4:49-60. 258. Poland J (1898) Traumatic Separation of the Epiphyses. London: Smith Elder and Company. 259. Poncet A (1872) Les deformations produites par l'arret d'accroissement d'un des os de l'avant-bras et des causes de cet arret de developpement; description des pieces pathologiques. Lyon Med 11:594-603. 260. Porat S, Nyska M, Nyska A, Fields S (1987) Assessment of bony bridge by computed tomography: Experimental model in the rabbit and clinical application. J Pediatr Orthop 7:155-160. 261. Quilis AN (1974) Fractures in children with myelomeningocele: Report of 15 cases and a review of the literature. Acta Orthop Scand 45:883-897. 262. Rang M (1969) The Growth Plate and Its Disorders. Edinburgh: ES Livingstone. 263. Rang M (1983) Children's Fractures, 2nd edition. Philadelphia: JB Lippincott. 264. Rapariz JM, Ocete G, Gonzalez-Herranz P, Lopez-Mondejar JA, Domenech J, Amaya S (1996) Distal tibial triplane fractures: Long-term follow-up. J Pediatr Orthop 16:113-118. 265. Ratliff AHC (1962) Fractures of the neck of the femur in children. J Bone Joint Surg 44B:528-542. 266. Ray TD, Tessler RH, Dell PC (1996) Traumatic ulnar physeal arrest after distal forearm fractures in children. J Pediatr Orthop 16:195-200. 267. Rehbein F, Hofmann S (1963) Knochenverletzungen im kindesalter. Arch Klin Chir 304:539-562. 268. Reichel GC (1768) De epiphysium ab ossium diaphysi diductione, Leipzig 1759. In: Sandifort E. Ed. Thesaurus Dissertationum, ed E Sandifort, volume 1. pp. 1-32, Rotterdam: Henricum Beman, volume 1:1-32. 269. Reidy JA, Van Gorder GW (1963) Treatment of displacement of the proximal radial epiphysis. J Bone Joint Surg 45A: 1355-1372. 270. Riseborough EJ, Barrett IR, Shapiro F (1983) Growth disturbances following distal femoral physeal fracture-separations. J Bone Joint Surg 65A:885-893. 271. Roberts JM (1973) Fracture-separation of the distal femoral epiphysis. J Bone Joint Surg 55A:1324. 272. Roberts JA, Bennet GC, MacKenzie JR (1989) Physeal widening in children with myelomeningocele. J Bone Joint Surg 71B:30-32. 273. Rodgers WB, Schwend RM, Jaramillo D, Kasser JR, Emans JB (1997). Chronic physeal fractures in myelodysplasia: Magnetic resonance analysis, histologic description, treatment, and outcome. J Pediatr Orthop 17:615-621. 274. Rodrigues KF (1973) Injury of the acetabular epiphysis. Injury 4:258-260. 275. Rognetta (1834) Memoire sur la divulsion des epiphyses. Gazette Medicale 2:433-438,449-454, 481-457, 513-519. 276. Roux (1837) Lettre sur le decollement epiphysaire. La Presse Med 1:436-438. 277. Roy S, Caine D, Singer KM (1985) Stress changes of the distal radial epiphysis in young gymnasts: A report of twentyone cases and a review of the literature. Am J Sports Med 13: 301-308. 278. Rudicel S, Pelker RR, Lee KE, Ogden JA, Panjabi MM (1985) Shear fractures through the capital femoral physis of the skeletally immature rabbit. J Pediatr Orthop 5:27-31.

604

CHAPTER 7 ~ Epiphyseal Growth Plate Fracture-Separations

279. Rutherford A (1985) Fractures of the lateral humeral condyle in children. J Bone Joint Surg 67A:851-856. 280. Sakakida K (1964) Clinical observations on the epiphysial separation of long bones. Clin Orthop Rel Res 34:119-141. 281. Salter RB, Harris WR (1963) Injuries involving the epiphyseal plate. J Bone Joint Surg 45A:587-622. 282. Scuderi G, Bronson MJ (1987) Triradiate cartilage injury. Report of 2 cases and review of the literature. Clin Orthop Rel Res 217:179-189. 283. Shapiro F (1982) Epiphyseal growth plate fracture-separations: A pathophysiologic approach. Orthopedics 5:720-736. 284. Shapiro F (1987) Epiphyseal disorders. N Eng J Med 317: 1702-1710. 285. Shapiro F, Holtrop ME, Glimcher MJ (1977) Organization and cellular biology of the perichondrial ossification groove of Ranvier: A morphological study in rabbits. J Bone Joint Surg 59A:703-723. 286. Shapiro F, Rand F (1992) Traumatic fracture-separations of the epiphyses: A pathophysiologic approach. Adv Orthop Surg 15:175-203. 287. Shapiro F, Kammen BF, Jaramillo D (1995) Assessment of transverse level of physeal fractures by MR imaging and correlative histology. J Pediatr Orthop 15:860. 288. Shelton WP, Canale ST (1979) Fractures of the tibia through the proximal tibial epiphyseal cartilage. J Bone Joint Surg 61A:167-173. 289. Sibley SW (1853) Fracture of the tubercle of the tibia by the muscular action of the rectus femoris. Med Times Gaz 6: 268-269. 290. Siffert RS (1963) Displacement of the distal humeral epiphysis in the newborn infant. J Bone Joint Surg 45A: 165-169. 291. Simpson AR (1880) On diastases in the bones of the lower extremities of the fetus produced by the accoucheur. Edinb Med J 25:1057-1060. 292. Smith BG, Rand F, Jaramillo D, Shapiro F (1994) Early MR imaging of the lower-extremity physeal fracture-separations: A preliminary report. J Pediatr Orthop 14:526-533. 293. Smith DG, Geist RW, Cooperman DR (1985) Microscopic examination of a naturally occurring epiphyseal plate fracture. J Pediatr Orthop 5:306-308. 294. Smith JB (1984) Knee instability after fractures of the intercondylar eminence of the tibia. J Pediatr Orthop 4: 462-464. 295. Smith RW (1850) Observations on disjunction of the lower epiphysis of the humerus. Dublin Quart J Med Sci 9:63-74. 296. Smith RW (1867) Address in surgery. "Epiphysary disjunctions." Br Med J ii:121-124. 297. Soutter FE (1962) Spina bifida and epiphyseal displacement: Report of two cases. J Bone Joint Surg 44B: 106-109. 298. Speer DP (1982) Collagenous architecture of the growth plate and perichondrial ossification groove. J Bone Joint Surg 64A: 399-407. 299. Spiegel PS, Cooperman DR, Laros GS (1978) Epiphyseal fractures of the distal ends of the tibia and fibula. J Bone Joint Surg 60A: 1046-1050. 300. Spiegel PG, Mast JW, Cooperman DR, Laros GS (1984) Triplane fractures of the distal tibial epiphysis. Clin Orthop Rel Res 188:74-89. 301. Steele JA, Graham HK (1992) Angulated radial neck fractures in children. J Bone Joint Surg 74B:760-764.

302. Stefanich RJ, Lozman J (1986) The juvenile fracture of Tillaux. Clin Orthop Rel Res 210:219-227. 303. Steinberg EL, Golomb D, Salama R, Weintraub S (1988) Radial head and neck fractures in children. J Pediatr Orthop 8: 35-40. 304. Steinert V (1965) Epiphysenlosung and epiphysenfrakturen. Archiv f Orthop Unfall-Chir 58:200-220. 305. Stephens DC, Louis DS (1974) Traumatic separation of the distal femoral epiphyseal cartilage plate. J Bone Joint Surg 56A:1383-1390. 306. Tapret, Chenet (1875) Arrachement de l'extremite inferiecure du femur; decollement epiphysaire et arrachement de parcelles osseuses de la diaphyse. Bull Soc Anat Pads 10:25-26. 307. Teot L, Bosse JP, Tremblay GR (1983) Pedicle graft epiphysis transplantation. Clin Orthop Rel Res 180:206-218. 308. Thompson GH, Gesler JW (1984) Proximal tibial epiphyseal fracture in an infant. J Pediatr Orthop 4:114-117. 309. Thompson GH, Wilber JH, Marcus RE (1984) Internal fixation of fractures in children and adolescents: A comparative analysis. Clin Orthop Rel Res 188:11-20. 310. Thomson JD, Stricker SJ, Williams MM (1995) Fracture of the distal femoral epiphyseal plate. J Pediatr Orthop 15: 474-478. 311. Tibone JE, Stoltz M (1981) Fractures of the radial head and neck in children. J Bone Joint Surg 63A: 100-106. 312. Tillaux P (1897) Traite de Chirurgie Clinique, 4th edition, volume 2. pp. 864-873, Paris: Arseline et Houzeau. 313. Torg JS, Pavlov H, Morris VB (1981) Salter-Harris type III fracture of the medial femoral condyle occurring in the adolescent athlete. J Bone Joint Surg 63A:586-591. 314. Truesdell ED (1918) Further observations upon birthdislocations of the cartilaginous epiphysis. Bull Lying-In Hosp City NY 11:245-257. 315. Trueta J, Morgan JD (1960) The vascular contribution to osteogenesis. I. Studies by the injection method. J Bone Joint Surg 42B:97-109. 316. Trueta J, Amato VP (1960) The vascular contribution to osteogenesis. III. Changes in the growth cartilage caused by experimentally induced ischemia. J Bone Joint Surg 42B:571-587. 317. Valverde JA, Albinana J, Certucha JA (1996) Early posttraumatic physeal arrest in distal radius after a compression injury. J Pediatr Orthop 5:57-60. 318. VanMeter JW, Branick RI (1980) Bilateral genu recurvatum after skeletal traction. J Bone Joint Surg 62A:837-839. 319. Vogt P (1878) Die traumatische epiphysentrennung und deren einfluss auf das langenwachsthum der rohrenknochen. Archiv f Klin Chir 22:343-373. 320. Von Laer L (1985) Classification, diagnosis and treatment of transitional fractures of the distal part of the tibia. J Bone Joint Surg 67A:687-698. 321. Wadsworth TG (1972) Injuries of the capitular (lateral humeral condylar)epiphysis. Clin Orthop Rel Res 85:127-142. 322. Waters PM, Kolettis GJ, Schwend R (1994) Acute median neuropathy following physeal fractures of the distal radius. J Pediatr Orthop 14:173-177. 323. Watson-Jones R (1955) Fractures and Joint Injuries, 4th edition, volume 2. pp. 785-788, Baltimore: Williams and Wilkins. 324. Weber BG (1980) Fracture healing in the growing bone and in the mature skeleton. In: Treatment of Fractures in Children

References

325. 326.

327. 328. 329.

330.

331.

and Adolescents, ed BG Weber, C. Brunner, F Freuler. Berlin: Springer-Verlag pp. 36-57. Weiner DS, O'Dell HW (1969) Fractures of the hip in children. J Traum 9:62-76. Wenger DR, Jeffcoat BT, Herring JA (1980) The guarded prognosis of physeal injury in paraplegic children. J Bone Joint Surg 62A:241-246. Werenskiold B (1927) A contribution to the Rontgen diagnosis of epiphyseal separations. Acta Radiol 8:419-426. Wiley JJ, Baxter MP (1990) Tibial spine fractures in children. Clin Orthop Rel Res 225:54-60. Williamson RV, Staheli LT (1990) Partial physeal growth arrest: Treatment by bridge resection and fat interposition. J Pediatr Orthop 10:769-776. Willis RB, Blokker C, Stoll TM, Paterson DC, Galpin RD (1993) Long-term follow-up of anterior tibial eminence fractures. J Pediatr Orthop 13:361-364. Wilson J (1820) Lectures on the Structure and Physiology of the Parts Composing the Skeleton and on the Disease of

332. 333. 334.

335.

336.

337.

605

the Bones and Joints of the Human Body. London: Burgess and Hill. Wolff O (1900) Ueber traumatische epiphysenlosungen. Deut Zeit f Chir 54:273-324. Worlock P, Stower M (1986) Fracture patterns in Nottingham children. J Pediatr Orthop 6:656-660. Young JWR, Bright RW, Whitley NO (1986) Computed tomography in the evaluation of partial growth plate arrest in children. Skelet Radiol 15:530-535. Zaleske D, Ehrlich MG, Piliero C, May JW, Jr, Mankin HJ (1982) Growth-plate behavior in whole joint replantation in the rabbit. J Bone Joint Surg 64A:249-258. Zaricznyj B (1977) Avulsion fracture of the tibial eminence: Treatment by open reduction and pinning. J Bone Joint Surg 59A:1111-1114. Zehntner MK, Jakob RP, McGanity PLJ (1990) Growth disturbance of the distal radial epiphysis after trauma: Operative treatment by corrective radial osteotomy. J Pediatr Orthop 10: 411-415.

CHAPTER 8

Lower Extremity Length Discrepancies I. II. III. IV. V.

Terminology Clinically Significant Length Discrepancies Limb Length Determination Causes of Lower Extremity Length Discrepancies Developmental Patterns in Lower Extremity Length Discrepancies VI. Lower Extremity Length Discrepancies in Specific Disease Entities: Pathoanatomy, Pathophysiology, Developmental Patterns, and Ranges of Discrepancies

VII. Projection of Limb Length Discrepancies by the Time Skeletal Maturity Is Reached VIII. Use of the Developmental Pattern Classification in Projecting Limb Length Discrepancies IX. Management of Lower Extremity Length Discrepancies X. Direct Operation on Epiphyses to Enhance Growth Potential by Removing Focal Transphyseal Tethers

I. T E R M I N O L O G Y

tal maturity greater than 2.0-2.5 cm warrants treatment, whereas anything under that cannot be expected definitively to have serious negative long-term sequelae. For those wishing to "fine-tune" discrepancy management, there is some evidence that discrepancies as little as 1.25 cm or 0.5 in. predispose one to negative sequelae, but discrepancies less than that are rarely treated surgically by even the most ardent practitioners. The negative sequelae of a lower extremity length discrepancy involve (1) an asymmetric appearance, (2) an awkward gait, (3) the possibility of osteoarthritis of the hip on the longer side due to the associated pelvic obliquity and the uncovering of the femoral head associated with this, and (4) low back pain in association with the compensatory lumbar scoliosis (Fig. 1). It is the general experience, however, that those with discrepancies greater than 2.5 cm show sufficient imbalance to warrant treatment. In those with projected discrepancies of less than 2.0 cm, there is no definitive evidence that leaving these discrepancies untreated will lead to long-term degenerative problems. The gray zone conceming the need for limb equalization is between 2.0 and 2.5 cm difference. The extent of any discrepancy must also be considered in relation to the individual's total height because the effect of the same discrepancy is relatively more marked in someone 5 ft tall than in another greater than 6 ft tall.

Lower extremity length discrepancies refer to differences in length between the two extremities, which can be due to some or all of pelvic, femoral, tibial, and foot height differences. The total length differential in the standing position is not dependent solely on the height of the involved bones but can be altered further by unilateral joint dislocation or subluxation, particularly at the hip but also at the sacroiliac or knee joint; by asymmetric femoral-tibial angular deformity; by asymmetric hip, knee, or ankle contracture, and by fixed pelvic obliquity. The term leg length discrepancy is still used commonly to refer to this broad entity but is imprecise. In strict anatomic terminology, leg refers to the segment between the knee and the ankle rather than to the entire extremity. The terms lower limb or lower extremity length discrepancy are more accurate (69, 192). Lower extremity length discrepancies accompany many epiphyseal disorders. Knowledge of epiphyseal structure and function plays a major role in assessing the effects of abnormalities on growth and in projecting the need for surgical intervention at appropriate times to either minimize or eliminate such discrepancies at skeletal maturity.

II. C L I N I C A L L Y S I G N I F I C A N T LENGTH DISCREPANCIES

B. Percentage of Individuals with Equal Limb Lengths

A. General Guidelines Concerning Extent of Clinically Significant Length Discrepancies

1. CLINICAL-RADIOLOGICAL DATA It has long been recognized that only 25-50% of people have equal lower extremity lengths. Hasse and Dehner (222) over 100 years ago (1893) reported leg length differences in

There are no absolute numbers conceming the extent of length differences that requires treatment. It is the feeling of most practitioners in this field that any discrepancy at skele606

SECTION II 9 Clinically Significant Length Discrepancies

607

Clinical Consequences of a Lower Extremity Length Discrepancy

C3 Lumbar Scoliosis 9 Low back pain 9 Sciatica

Osteoarthritis of Hip 9 Long-side 9 Supero-lateral

Awkward Gait Asymmetric Appearance 9 Flexed knee (long side) 9 T o e walking (sho~ side) 9 Trunk-pelvic tilt

F I G U R E 1 The clinical consequences of a lower extremity length discrepancy are illustrated. Lower extremity length discrepancy with shortness on the left is associated with a compensatory lumbar scoliosis, pelvic obliquity, and a relative adduction positioning of the hip on the longer right side. Note the diminution of the CE angle on the longer right side, which can predispose one to superolateral osteoarthritis.

68% of 5141 soldiers with the left longer than the fight in a proportion of 3.3:1. Lower extremity lengths thus were equal in only 32%. Edinger and Bietermann (148) documented lower extremity lengths in 351 individuals radiologically. Both extremities were of equal length in 178 (51%), had less than 5 mm difference in 27 (8%), and had discrepancies between 5 and 50 mm in 146 (42%). The left was longer than the fight in a proportion of 3.6:1. In one of the few large radiographic studies of limb asymmetry, Rush and Steiner (416) documented the presence or absence of discrepancies in 1000 healthy army recruits referred to the radiology department because of a low back complaint and in 100 general duty soldiers without complaint. An anteroposterior radiograph of the lumbosacral spine, pelvis, and proximal femurs in the standing position was made in standardized fashion. The lower edge of the film was parallel to the standing platform such that any differences in the level of the femoral heads indicated a difference in lower extremity length. Length differences were measured in millimeters. The limbs were equal in length in 23% of the larger mildly symptomatic group and in 29% of the smaller fully asymptomatic group. On the basis of this study only 25% of individuals had equal lower extremity lengths. 2. A N T H R O P O L O G I C A L DATA Anthropological data involving actual skeletal measurements of paired long bones in humans also demonstrate a

tendency to increased length of the left femur compared to the right with a left:right predominance also present but less marked in the tibia. Schultz (427) reported on skeletal measurements from 753 human skeletons involving 232 white, 233 American black, 118 North American Indian, 122 Alaskan Eskimo, 41 Chinese, and 7 Australian aboriginal skeletons. On average the humerus and femur were approximately 8% shorter in females than in males. The measurements were made to the nearest millimeter. In virtually all studies reported, the right upper extremity is longer than the left, usually in the range of 75% of cases. This is felt to reflect the fight-handed dominance in the human species. Virtually all studies, however, also show the length of the femur and tibia to be longer on the left in those in which equal limb lengths are not present. When measurements in 744 individuals were averaged in all races, the left femur was longer than the fight in 50%, the fight was longer than the left in 33%, and both were equal in only 17%. The left:fight predominance therefore was 1.5:1. Differences were less marked in the tibia. In 734 individuals the left was greater than the fight in 45%, the fight was greater than the left in 40%, and the fight equaled the left in only 15%. When both femur and tibia were combined, only 5% had equal lengths in 727 individuals. The left was longer in 54% and the fight was longer in 41%, with a left:fight predominance of 1.3:1. These studies were similar to those clinical measurements made by Haase and

608

CHAPTER 8 ~ Lower Extremity Length Discrepancies

Dehner. Garson (174) measured 70 skeletons noting that the combined femur and tibia lengths were longer on the left in 54.3% and longer on the fight in 35.8%, with equal lengths being present in 10%. The combined femoral-tibial differences when the left was longer averaged 4.8 mm, and when the fight was longer they averaged 3.3 mm. In 124 skeletons from Switzerland, Schwerz (428) found the left femur longer in 52%, the right femur longer in 31%, and both equal in 17%. Almost all studies show smaller discrepancies with the tibia. In all racial groups and in both sexes, asymmetries favoring the left bone of the lower extremity, particularly in relation to the femur, are much more frequent than those favoring the right. The absolute amounts of discrepancy, however, are small with the series in the adult man showing a general average of only 2.5 mm difference in the femurs. The humeral differences were greater, however, averaging 4.1 mm longer on the right. Munter (345) compiled a detailed analysis of 326 (233 male and 93 female) adult Anglo-Saxon skeletons, although all bones were not available from each skeleton. The mean values for femoral and tibial lengths in his series also were greater on the left side for both males and females. The mean femoral lengths in the male were 2.4 mm greater on the left side for the femur and 4.5 mm longer on the left side for the tibia. In females the left side was longer for the femur by 4.8 mm and for the tibia the left side was longer by a mean of 1.5 mm. Munter noted from several other series that there was little to no racial variation in side differences or in proportional differences, such that data could be pooled for more statistical significance. There is little presentation of absolute data in the Munter study because most of the assessments related to correlative values. All of the basic numerical data, however, are presented in the paper's appendix such that the bilateral differences for each individual could be calculated.

C. Clinical Effects of Lower Extremity Length Discrepancies Possible negative sequelae of lower extremity length discrepancies have been studied in efforts to relate symptoms to particular amounts of discrepancy (Fig. 1).

1. OSTEOARTHRITISOF THE HIP Primary osteoarthritis (OA) of the hip is quite common, as is osteoarthritis secondary to the major childhood hip abnormalities of developmental dysplasia, Legg-Calve-Perthes disease, and slipped capital femoral epiphysis. It thus is very difficult to document with statistical certainty whether a length discrepancy alone is the primary determinant of a hip arthritic condition. Morscher (337) has clearly discussed the changes in hip joint mechanics due to leg length discrepancies. He pointed out that studies by Pauwels (375) had shown that a lesser amount of pressure is actually transmitted to the hip joint of the shorter leg due to the pelvic tilt, which serves to increase the area of contact between the femoral head and

the acetabulum. There also is truncal shift over the short side, further minimizing the effort needed for hip abduction. Conversely, the weight bearing characteristics of the longer side hip are worsened because there is both a decrease in coverage of the femoral head by the acetabulum on the longer side due to the nature of the pelvic tilt and an increased load at the joint on the longer side due both to diminution of the area of contact between the femoral head and the acetabulum and also to the increase in tone necessitated for the hip abductor muscles. Morscher likened the action of a lower extremity length discrepancy on the longer leg to a coxa valga deformity. The extent of increased coverage of the femoral head on the short side and decreased coverage on the long side has been documented by Krakovits (278) by a trigonometric series of calculations. When the leg is shortened by 1 cm, the diminution of the CE angle of Wiberg on the longer side is 2.3 ~ 2 cm shortening = diminution of 4.6 ~ 3 cm = 6.8 ~, 4 c m = 9.1 ~, 5 c m = 11.3 ~, 6 c m = 13.5 ~, 7 c m = 15.6 ~ 8 cm = 17.7 ~ 9 cm = 19.8 ~ and 10 cm = 21.8 ~ Gofton (187) also supported the contention that stresses imposed on the longer side hip are greater than normal, with those on the shorter side reduced. Acetabular pressure on the longer side was concentrated laterally due to the adducted position of the proximal femur, leading to superolateral femoral head OA. Goflon and Trueman (188) detected a clear association between idiopathic superolateral osteoarthritis of the hip and lower extremity length discrepancy, with the hip on the longer side involved in 33 of 36 instances. There were 31 of the 36 cases showing OA on the longer side with discrepancies from 5 m m (3~6 in.) and greater with 16 of the 36 having a discrepancy greater than 1.25 cm (0.5 in.). The differences in lower extremity lengths were determined radiographically with strict application of a standardized standing orthoroentgenographic technique. It was the superolateral variant of osteoarthritis that was particularly related to the length discrepancy. The more global or medial forms of OA were not assessed, and none of the patients were determined to have clear predisposing causes of OA such as hip subluxation or dislocation or residual evidence of any of the other childhood hip disorders. When measurements alone were considered, 4 patients had OA on the shorter side, 3 had OA but were level, and 29 had OA on the longer side. When a correction was made to allow for an estimation of the original prearthritic leg length discrepancy by adding 5 mm or 3~6 in. to the side with the OA to represent the amount of articular surface collapse, even more patients were shifted to the group showing length discrepancy with the longer side affected. They concluded that length discrepancy was present in at least 33 of the 36 cases with the longer side developing the OA. 2. L o w BACK PAIN AND SCIATICA The correlation of compensatory lumbar scoliosis with lower extremity length discrepancies also has been studied, as have attempts to equate the shortness with increases in lumbar discomfort. The relationship between shortening,

SECTION II ~ Clinically Significant Length Discrepancies

compensatory scoliosis, and discomfort, however, has been difficult to document. Morscher (337) describes the attempts at correlations well. He noted that Hult (241) found that almost 54% of patients with lower extremity length discrepancies complained of lumbar pain but that 60% of patients without such discrepancies had similar complaints. Electromyographic studies by Taillard and Morscher (469) showed that relatively small leg length discrepancies between 1 and 2 cm could lead to a remarkable increase in muscle activity in several muscle groups. The possibility remains, therefore, that even small discrepancies make it difficult to maintain a complete resting position due to secondary muscle activations. The relationship between lumbar scoliosis and lower extremity length discrepancies is not invariably the same. In the large majority of cases the convexity of the lumbar scoliosis is directed toward the shorter side, but in perhaps 1015% of cases the scoliosis is contralateral to that expected on a purely mechanical basis. Morscher (337) indicates that the development of a lumbar scoliosis may be due more to dynamic forces in association with walking than to static forces as demonstrated in the standing position. Difficulties are associated with attempting to determine whether length discrepancies cause degenerative disk disease. Low back pain in itself is quite common and relatively large numbers of patients very carefully studied would be needed to determine whether there was an increased prevalence in those with lower extremity length discrepancies alone. Many efforts have been made in this regard. In the study by Rush and Steiner (416) involving a larger group of army recruits of 1000 with low back pain and a smaller group of 100 without pain, the percentages of lower extremity length differences in the larger/smaller groups were as follows: 1-5 mm, 39.5%/38%; 6-10 mm, 22.5%/29%; 11-20 mm, 13.3%/4%; and i>21 mm, 1.7%/0%. The combined patients with length discrepancies 5 mm or less were 62.5%/67% and with length discrepancies 10 mm (1 cm) or less were 85%/ 96%. It is evident, therefore, that it was only beyond the 11-mm discrepancy level that the percentage incidence of back symptoms increased beyond the control range. The length discrepancies were then correlated with the detailed quantification of the pathological conditions of the spine seen on both anteroposterior and lateral radiographs. The radiographic abnormalities were present in the same percentage of patients with equal limb lengths and in those with length discrepancies, with 25% of each group showing changes. The authors concluded that in the symptomatic group it could have been length discrepancy itself with the associated compensatory scoliosis rather than the radiographic abnormalities that was responsible for the symptoms. Nichols (348) used clinical tape measurements from the anterior superior iliac spine to the tip of the medial malleolus to document that 7% of 1007 patients without back pain had a length discrepancy of 1.25 cm (0.5 in.) or more, whereas a limb length discrepancy of 1.25 cm or more was seen in 22% of 180 airmen complaining of low back pain. In his review of the Rush-Steiner data, Nichols also concluded that a signif-

609

icant difference in the incidence of back pain and a shortened lower extremity could only be seen when the discrepancy was 11 mm or more. It is these studies that provide some objective information of the effects of lower extremity length discrepancies. Giles and Taylor (178) studied the relationship of lower extremity length inequality and low back pain in 1309 patients and a small control group of 50. The prevalence of a length discrepancy of 1 cm or more was more common in patients suffering from low back pain (18.3% of 1309) than in the normal population (8% of 50 controls). Friberg (170) studied the correlation between lower extremity length discrepancy, low back pain, and chronic unilateral hip discomfort. He used a low-dose radiologic method with the patients in a standard posture with the single radiograph showing the lower spine, pelvis, and hips. The study comprised 1157 subjects: 653 patients with chronic low back pain with or without sciatica, 254 with chronic unilateral hip pain, and 359 symptom-free army conscripts. In the total series the lower extremities were of equal length in only 8% of patients. The left lower extremity was longer than the fight by a ratio of 1.4:1. There was excellent correlation between an increase in the amount of lower extremity length discrepancy and both back pain and hip discomfort. When all patients were assessed, those within lengths from 0.0 to 4.0 mm difference comprised 36% of the population; those 5.0-9.0 mm a further 39%, 10.0-14.0 mm, 17%, and 15.0 mm or more, 8%. Stated another way, those with lower extremity lengths less than 10.0 mm (1.0 cm) comprised 75% of the study, leaving 25% with a discrepancy of 1.0 cm or greater. The incidence of low back pain in the lower extremity length groupings 5.0-9.0, 10.0-14.0, and 15.0 mm or more was 45.3%, 18.4%, and 11.7%, respectively, with the numbers for the symptom-free group at the same length discrepancies much less at 27.9%, 13.4%, and 2.2%. Stated a different way, the ratio of symptomatic to nonsymptomatic back pain patients with limb length differences of 5.0 mm or more was 1.73:1, 10.0 mm or more, 1.93:1, and 15.0 mm or more, 5.32:1. A discrepancy of 1.5 cm or more, therefore, clearly appeared to predispose the individual to a relatively high likelihood of back discomfort; the presence of leg length inequality of 1.5 cm or more was 5.32 times more likely in 653 patients with chronic low back pain than in 359 symptomfree soldiers. Similar findings were found in relation to chronic hip discomfort. In the 254 patients with hip discomfort, the pain was located on the side of the longer extremity in 88.9% with symptoms and in the hip of the shorter extremity in only 11.1% of cases. Sciatica predisposed to the longer side by a ratio of 3.7:1, similar to findings in the small series of Redler (399) in which, in 15 cases of sciatica, it was present on the longer side by a 2:1 ratio. Rossvoll et al. (415) have assessed back pain in young adult patients before and after subtrochanteric shortening osteotomies of the femur performed after skeletal maturity. There were 22 patients followed for an average of 5 years. The mean preoperative length discrepancy was 3.2 cm with

610

CHAPTER 8 ~ Lower Extremity Len~tth Discrepancies

follow-up discrepancy diminished to 0.43 cm. Approximately half of the patients had relatively serious low back pain prior to surgery with the other half having minimal to no low back pain. The mean ages at operation in the two groups were 25.9 and 20.2 years. The degree of low back pain was felt to be significantly reduced after the operation. Other studies with small numbers of patients reflect how well some patients with length discrepancies do. Gibson et al. (176) found that otherwise healthy young adults with an average of 3 cm of limb length discrepancy perceived no functional effect, whereas Gross (202) evaluated 35 marathon runners and found 7 with limb length discrepancies greater than or equal to 1 cm who reported no effects upon their performance. 3. GAIT ASYMMETRY Additional studies have begun to define the nature of gait asymmetry in patients with limb length inequality. Kaufman et al. (264) performed detailed gait studies on 20 subjects to determine the magnitude of discrepancies that resulted in gait abnormalities. A limb length inequality greater than 2 cm (3.7% difference) resulted in gait asymmetry that was greater than that observed in the normal population. This number correlates well with the generally accepted clinical guideline beyond which length discrepancy correction is warranted. Goel et al. (184) performed gait analysis for discrepancies less than 2 cm to determine the maximum moments at the hip, knee, and ankle joints. They concluded that a minor length discrepancy of 1.2 cm did not produce meaningful biomechanical changes and that the body was well able to compensate for minor lower extremity length discrepancies up to 2 cm. Goel et al. studied 10 healthy subjects with equal limb lengths, simulated minor limb length discrepancies using a shoe lift of 1.25 cm, and an additional 10 asymptomatic patients with limb length discrepancies ranging from 1 to 2 cm. Their study "did not find an association between minor limb length discrepancies and predictable changes in lower extremity joint kinetics that might potentially lead to joint abnormalities."

III. LIMB LENGTH DETERMINATION A. Clinical Measurements Clinical examination of a patient with a lower extremity length discrepancy remains the basic form of assessment. When viewed from the back in the standing position, one looks for a compensatory scoliosis, palpates the levels of the iliac crests, and examines for the levels of the buttock and popliteal creases, the presence of a plantigrade foot, and the thigh and calf circumferences. Children with lower extremity length discrepancies use compensatory mechanisms to maintain an uptight alignment. In the standing position with the feet fiat and the knees fully extended, a compensatory

lumbar scoliosis is seen with the curve convex on the shortened side. With forward bend or in the sitting position, the scoliosis disappears as there is no longer need for any compensation. A rotatory thoracic or lumbar component is absent with forward bend in a compensatory scoliosis, whereas in a structural curve it persists. When the child is standing or walking, the discrepancy can be hidden either by flexing the knee on the longer side or by walking with the foot in an equinus position on the shorter side. Assessment of a patient with a lower extremity length discrepancy should check for the range of motion of the ankle as in some of the larger longstanding discrepancies this equinus posturing can become fixed. To measure the extent of discrepancy, rectangular blocks of known height are placed under the foot on the shortened side with the patient standing. The blocks are positioned until the compensatory scoliosis disappears and the iliac crests can be palpated at the same level. This clinical assessment is essential as it accurately denotes the entire extent of any discrepancy, including pelvic, thigh, leg, and foot components. Another characteristic clinical measurement of lower extremity length discrepancy done with a tape measure and the patient lying supine measures the distance from the inferior tip of the anterior superior iliac spine to the inferior tip of the medial malleolus. It is important to check that there is no pelvic obliquity, hip or knee contractures, or femoral-tibial angular deformities in the patients being assessed in this way. If these conditions are present, measurement from the umbilicus to the medial malleolus also can be performed to register what is referred to as the apparent limb length discrepancy. Smith (449) has reported a clinical method for determining lower extremity length discrepancy, which is particularly valuable in those with hip or knee flexion contractures. The method, referred to as the thigh-leg inspection test, is performed by placing the patient in the supine position with the hips and knees flexed 90 ~. Determination of the difference in height of the thighs is made by using a flat protractor on the most superior aspect of the knee joint on the longer side and measuring the difference between the protractor and the shortened thigh at the knee. The shortened leg then is measured as the thighs are held parallel. The difference in length of the longer side heel and the shortened heel is measured by placing the protractor on the plantar surface of the longer heel and measuring the distance between it and the shorter heel. This measurement eliminates flexion contractures at the hips, knees, and ankles as a source of error. It also measures the soft tissue component of the lower extremity as well as the bones. This measurement was found to be more reproducible than tape measurements or block measurements in a comparative study of several practitioners in 96 patients with a mean of 5.2 cm length discrepancy. The technique is yet another way of estimating lower extremity length discrepancies in a quite reasonably accurate fashion clinically and particularly bypasses concerns with hip, knee, or ankle flexion contractures. Morscher and Figner

SECTION IV ~ Causes of Lower Extremity Length Discrepancies (338) have described clinical and radiographic methods of measurement in detail.

B. Segments to Be Considered in Assessing Lower Extremity Length Discrepancies Lower extremity length discrepancies can involve any or all of four segments: pelvic height, femoral length, tibial length, and foot height. The clinical assessment measuring the length discrepancy with the patient standing on blocks takes each of these four segments into consideration, whereas most of the other length determination modalities do not. The clinical measurement from the anterior superior iliac spine to the medial malleolus eliminates consideration of the foot and only partially addresses the pelvic height; the characteristic radiographic determinations of femoral and tibial length obviously do not consider either the pelvic or the foot region, and virtually all of the limb length determinations and surgical corrections relate to the femur and tibia alone without considering the other two segments. Although the femur and tibia are responsible for the vast majority of the limb height, there can be situations in which foot and pelvic abnormalities contribute meaningfully to the length discrepancy. In these situations specific radiographic determination of their height is important.

C. Radiographic and Other Imaging Documentation of Lower Extremity Length Discrepancies Five imaging techniques have been used to document both the length of the respective lower extremity bones and the extent of lower extremity length discrepancies. Many of the screening surveys referred to in Section II documented length discrepancies using standardized standing positions but radiographs of only the lower spine, pelvis, and hips. The relative femoral head positions allowed for length discrepancy measurements but not for absolute femoral-tibialfoot height measurements. The many technical considerations in making accurate radiologic measurements have been detailed particularly well. (161,188, 197, 338) 1. TELEOROENTGENOGRAMS Teleoroentgenograms are limited to use during the first year of life. They refer to a plain X ray of the entire lower extremity centered over the knee. If taken at a 72-in.,height, they are quite accurate in terms of length due to the minimal magnification with the small limb. After this age their accuracy decreases in documenting limb length discrepancies due to angular distortion (Fig. 2A). 2. ORTHORoENTGENoGRAMS This technique was developed by Green and associates (197) in the late 1940s to document accurately lower extremity length discrepancies. Both femurs and tibias are radio-

611

graphed in their entirety. A single long X-ray cassette with a single long X-ray sheet is used. Three X-ray machines are built onto the ceiling of a specifically defined chamber 72 in. from the X-ray film. There is one machine to be centered over the hip, one to be centered over the knee, and one to be centered over the ankle. Three radiographs are taken in rapid succession. Studies have documented that the magnification factor is under 1% with this technique. In addition to providing extremely accurate length determinations, the film allows for radiographic visualization of the entire bone for an indication of structural or angular deformities. This technique is infrequently used today because of concerns about the total amount of radiation exposure with sequential studies (Fig. 2B). 3. SCANOGRAMS This is the technique used most commonly today for lower extremity length discrepancy documentation. It can be accurate if details of performance are rigidly adhered to, although in practice many inaccuracies are seen. Three radiographs are taken similar to the orthoroentgenogram centered over hip, knee, and ankle, but spot films only are taken with the intervening femur and tibia diaphyseal segments spared any radiation. A ruler is placed beside the limb and the radiographic projection of the ends of the femurs and tibias over the ruler allows for a measurement at that level. The ruler should be calibrated in millimeters. A major problem with this technique is patient movement during the repositioning of the single machine. This can lead to inaccuracies that if slight cannot be detected. Very high levels of quality control thus are needed to allow for accurate measurement.

4. COMPUTERIZED TOMOGRAPHY SCANS The CT scan can provide accurate length measurements due to the high resolution calibration of the technique. 5. ULTRASONOGRAPHY Ultrasound can be used as well to document bone length (474). It is quite helpful in the first year or two of life when cartilage elements compose the bulk of the epiphyses.

IV. C A U S E S O F L O W E R E X T R E M I T Y LENGTH DISCREPANCIES A large number of disorders during the growing years can either stimulate or retard the growth of epiphyses unilaterally or asymmetrically such that a lower extremity length discrepancy occurs. Virtually any childhood disorder that affects an epiphysis can lead either to stimulation or retardation of growth, depending on the clinical context, and a possible limb length discrepancy must be considered in relation to overall management. Even disorders that are present throughout the skeleton, such as hereditary multiple exostosis, can affect the two sides unequally. The causes, effects, and extent of lower

612

CHAPTER 8 9 Lower Extremity Len~trh Discrepancies

A A

A'

A"

iP

b T

.

.

..

c

Length of X-ray Shadow

,

.

d'" -~

F I G U R E 2 Radiographic measurements to determine lower extremity lengths. (A) Teleoroentgenogram. A single exposure of the entire lower extremity centered over the knee provides a radiographic image of both the entire femur and the entire tibia and fibula. The longer each bone, the greater the magnification error due to the increased divergence of the rays. (B) Orthoroentgenogram. In performing an orthoroentgenogram, a single long X-ray cassette is used. Three cameras that can be moved along a track are mounted at a standardized 72-in. distance from the cassette holder. Three radiographs are taken in rapid succession with one camera centered over the hip joint, one over the knee joint, and one over the ankle joint. The perpendicular rays intersect the ends of the bones recording the true length. Each long bone is imaged completely, allowing for structural and angular deformity assessments as well. [Reprinted from (197), with permission.]

extremity length discrepancies throughout the spectrum of disorders affecting the growing skeleton are listed in Table I.

V. D E V E L O P M E N T A L P A T T E R N S IN L O W E R E X T R E M I T Y LENGTH DISCREPANCIES The discrepancies that develop in children are susceptible to considerable change with time, as the involved physes have the potential for increasing the discrepancy, maintaining it at a stable level, or correcting it spontaneously. Not all length discrepancies increase continually with time during the growing years. In a review of lower extremity length discrepancies in 803 children who were followed by at least annual orthoroentgenograms for 5 or more years to skeletal maturity or to the time of corrective surgery, it was demonstrated that several patterns of developmental discrepancy can occur (433). These are dependent on the nature of the conditions causing the discrepancies and on the place and time of their occurrence. They do not refer to changes following bone surgery. Table I gives a broad categorization of disorders that can be associated with lower extremity length discrepancies, an indication of whether they cause growth retardation or stimulation, and a range of length discrepancies with which they are associated. Details relating to each specific disorder are described in Section VI.

A. Patient Population The longitudinal data on lower extremity length discrepancies from patients who had been followed in the Growth Study Unit at the Children's Hospital, Boston, over a 40-year period (1940-1980) were studied carefully. The patterns of developmental discrepancy that developed were demonstrated by charting the extent of a discrepancy directly against time as represented by the patient's chronological age (433). The patterns also were related to skeletal age to show their independence from that parameter. Lower extremity length discrepancies were documented by standard techniques using teleoradiographs for patients who were younger than 5 years of age. Orthoroentgenograms, from which femoral and tibial measurements were made, were used for all of the older patients. Skeletal age was determined from posteroanterior radiographs of the left wrist and hand. These radiographs were correlated with the Todd atlas (477) until 1950 and with the Greulich and Pyle atlas (199) thereafter. For inclusion in the review, an individual had to have been followed by radiographic means at the Growth Study Unit for a minimum of 5 years (or from the onset of disease) either to the time of skeletal maturity or to the time of bone surgery. Due to the deep interest of Dr. William T. Green and his staff, virtually all of the patients in the series were assessed annually, and often semiannually, from the time of onset or detection of the disease to maturity. It must be emphasized that these patients were followed prospectively

SECTION V ~ Developmental Patterns in Lower Extremity Length Discrepancies

TABLE I

C a u s e s , Effects, a n d Extent o f Lower Extremity Length D i s c r e p a n c i e s a Stimulation

A 1. Femoral disorders

2. Tibial and fibular disorders

B

5. Infection

C

Retardation

B

C

D

A

Coxa vara

~/

v/

Congenital short femur Developmental dysplasia of the hip

v/

~/

Proximal femoral focal deficiency

D

v/

Subluxated hip Dislocated hip Avascular necrosis Legg-Calve-Perthes disease Slipped capital femoral epiphysis

v/

Status post varus osteotomy

v/

~/

Congenital short tibia (fibular hemimelia) Tibial agenesis (tibial hemimelia)

v/

Pseudoarthrosis of tibia (anterolateral bowing) ___neurofibromatosis Posteromedial tibial bowing 3. Foot disorders 4. Trauma

613

v/

~/

Blount's disease (tibia vara) Clubfoot

~/

v/

Epiphyseal growth plate fracture-separations Diaphyseal fractures Healing with shortening or overgrowth Loss of bone mass Septic arthritis Hip or knee Meningococcal septicemia

v/

,/

Femoral-tibial osteomyelitis (metaphyseal-diaphyseal)

~/

v/

v/

~/

v/

v/

v/

Femoral-tibial osteomyelitis (infantile) Tuberculosis

6. Prolonged unilateral immobilization (9-12+ months) 7. Neuromuscular disorders

8. Vascular disorders

Hip or knee Femoral-tibial shaft Tuberculosis CDH, etc. Poliomyelitis Cerebral palsy-hemiparesis Peripheral nerve injury (unilateral); sciatic nerve paralysis Myelomeningocele Arthrogryposis Diastematomyelia Vascular malformations Klippel-Trenaunay Parkes Weber Proteus

v/

~/

Beckwith-Wiedemann Hemangioma (newer terminology)

~/

v/

Cutis marmorata telangiectatica congenita

~'

v/

Following use of neonatal umbilical or femoral catheters

(continues)

614

CHAPTER 8 9 Lower Extremity Length Discrepancies TABLE I (continued) Retardation

Stimulation

9. Hemihypertrophy, with connective tissue abnormalities

A

B

C

D

A

B

v/

v/

C

Synovial hemangioma (knee)

v/

,i/

Arteriovenous aneurysms (traumatic)

v/

v/

Cerebrovascular malformations

v/

,/'

Lipomatosis

,/

4'

Lymphedema

v/

~/

Lymphangioma Neurofibromatosis (bone structurally normal)

v/ v/

v/ v/

Silver-Russell syndrome

v/

v/

Malignant tumors: Wilm's, adrenal, carcinoma, hepatic carcinoma

v/

v/ v/

v/

4'

4'

v/

v/

v/

4'

v/

V

V

v/

Resection with growth plate damage

v/

v/

Resection with loss of bone mass

v/

v/

v/

Radiation-induced physeal arrest

v/

v/

v/

v/

v/

V

v/

v/

v/

V

v/

,/

v/

v/

v/

v/

v/

10. Hemiatrophy

Otherwise normal appearing limb

11. Inflammatory disorders

Juvenile rheumatoid arthritis

D

Scleroderma 12. Hematologic disorders

Hemophilia Thalassemia

13. Tumors-focal bone lesions

Benign lesions Fibrous dysplasia Unicameral bone cyst Aneurysmal bone cyst Osteoid osteoma

v/

,/

Caffey's disease

,/

~/

Malignant tumors

14. Skeletal dysplasia (asymmetric lengths)

Hereditary multiple exostoses

Bone diseases with high incidence of asymmetric deformities

v/

Ollier's enchondromatosis Stippled epiphyses (chondrodysplasia punctata) Dysplasia epiphysealis hemimelica Melorheostosis Camptomelic dwarfism Congenital banding (Streeter syndrome)

15.

v/

,/

v/

V

v/

v/

~/

V

Osteogenesis imperfecta

v/

v/

v/

Rickets Renal osteodystrophy Vitamin A intoxication of infancy

V

v/

v/

v/

v/

v/

V

~/

Scurvy (lack of vitamin C) 16. External causes

Bums

v/

Frostbite aCode: A = 0-2 cm; B = 2-5 cm; C = 5-15 cm; D = 15 cm+. The A-D groups also provide management guidelines-see Fig. 18. A check (v/) indicates the possible extent of the final discrepancy.

because they had an affection in which lower extremity length discrepancy was known to occur rather than being seen only after a clinically apparent discrepancy had developed. With the exception of the group of patients with a fractured femoral diaphysis, the patients included in this re-

view had to have had a discrepancy of 1.5 cm or more at some time during the period of assessment. The classification does not refer to any change in discrepancy that followed surgical physeal arrest, diaphyseal lengthening, or osteotomy.

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

The disease entities that were studied and the number of patients in each group were as follows. There were 18 patients with proximal femoral focal deficiency, 102 with congenital coxa vara and a congenitally short femur (some with associated anomalies of the leg and foot), 17 with Ollier's disease (enchondromatosis), 21 with physeal destruction, 115 with poliomyelitis, 33 with septic arthritis of the hip, 116 with a fractured femoral diaphysis, 29 with a hemangioma, 17 with neurofibromatosis, 46 with hemiparetic cerebral palsy, 113 with hemiatrophy or hemihypertrophy (anisomelia), 36 with juvenile rheumatoid arthritis, and 140 with Legg-Perthes disease. The distribution of pattern types, the average discrepancy in centimeters, and the range of discrepancies before surgery were assessed for each group.

B. Classification of Developmental Patterns in Lower Extremity Length Discrepancies Type I, upward slope pattern: The lower extremity length discrepancy develops and increases continually with time at the same proportionate rate. Type II, upward slope-deceleration pattern: The lower extremity length discrepancy increases at a constant rate for a variable period of time and then shows a diminishing rate of increase independent of skeletal maturation. Type III, upward slope-plateau pattern: The discrepancy first increases with time but then stabilizes and remains unchanged throughout the remaining period of growth. Type IliA, downward slope-plateau pattern: The discrepancy decreases with time but then stabilizes and remains unchanged throughout the period of growth. Type IIIB, plateau pattern: The discrepancy, detected initially after it has developed, remains unchanged throughout the remaining period of growth. Type IV, upward slope-plateau-upward slope pattern: The discrepancy first increases then stabilizes for a variable but considerable period of time, and then it increases again toward the end of the growth period. Type V, upward slope-plateau-downward slope pattern: The discrepancy increases with time, stabilizes, and then decreases in the absence of surgery. The classification is illustrated in Fig. 3A, and an indication of the various patterns in various disorders is outlined in Fig. 3B.

VI. L O W E R E X T R E M I T Y L E N G T H D I S C R E P A N C I E S IN S P E C I F I C D I S E A S E ENTITIES: PATHoANATOMY, PATHOPHYSIOLOGY, DEVELOPMENTAL PATTERNS, AND RANGES OF DISCREPANCIES In this section we incorporate information from our study reported in the article "Developmental Patterns in Lower

615

Extremity Length Discrepancies," (433) as well as information from the extensive literature on the entire range of disorders that can lead to length differences. The focus is on the pathoanatomy and pathophysiology of the disorders themselves and particularly on the pattern of discrepancy development and the extent of the discrepancies in the specific diseases. In most instances the ranges of length discrepancy values are provided. Some studies refer to percentage shortening in relation to the normal side. Reference to the GreenAnderson tables then can indicate the range of values in absolute terms.

A. Congenital Limb Deficiencies Congenital limb deficiencies are among the most common causes of lower extremity length discrepancies. In the next few subsections we will refer to the most common and most severe types, but in reality they represent part of a spectrum of disorders affecting appendicular development in both upper and lower extremities. Extensive efforts have been made over the past few decades to develop encompassing classifications for these disorders, but they are so variable and the terminology used has been so awkward that there has been no universal agreement on any way of referring to them. As a result, individual terms from differing classifications have come to be used commonly and on occasion different terms are used to refer to the same disorder. 1. FRANTZ AND O'RAHILLY

The classification of Frantz and O'Rahilly (168) is an all encompassing approach that divides congenital skeletal limb deficiencies into terminal, in which no unaffected parts are distal to and in line with the deficient portion, and intercalary, in which the middle portion of a proximodistal series of limb components is deficient but the proximal and distal portions are present. Each of these two main groups then may be either transverse, in which the defect extends transversely across the entire width of the limb, or longitudinal, in which only the preaxial or postaxial portion is absent (hence, the deficiency is longitudinal). Among the terms used in the classification are the following: amelia, absence of the limb; hemimelia, absence of a large part of a limb; phocomelia, a flipperlike limb with a hand or foot attached more or less directly to the trunk; acheiria, absence of a hand; apodia, absence of a foot; adactylia, absence of a digit including the associated metacarpal or metatarsal; and aphalangia, absence of one or more phalanges. Hemimelia may be complete or partial. The term paraxial hemimelia indicates that either the preaxial or the postaxial portion of the distal half of the limb is involved. The anatomical term preaxial refers to the border of a limb on which either the thumb or the big toe is situated and the term postaxial refers to the opposite border. The preaxial paraxial hemimelias therefore are either radial or tibial and the postaxial paraxial hemimelias are ulnar or fibular. The various subtypes of paraxial hemimelia are named

616

CHAPTER 8 9

Lower Extremity Length Discrepancies A J

Type 1 Upward Slope Pattern

J,

g b (/) t5 Age

Type 2 Upward SlopeDeceleration Pattern

Type 3 Upward Slope- Plateau Pattern

I\ '1 Type 3a

Type 3b

D o w n w a r d Slope Plateau Pattern

Plateau Pattern

Type 4 Upward Slope- Plateau Upward Slope Pattern

/ J

f

Type 5 Upward Slope- Plateau Downward Slope Pattern

DISTRIBUTION OF DEVELOPMENTALPATTERNS IN THE VARIOUSDISEASES IN EIGHT HUNDRED AND THREE PATIENTS Condition Proximal femoral focal deficiency Congenitally short femur, including congenital coxa vara (with some associated leg and foot anomalies) Ollier's disease Destroyed epiphyseal growth plates Poliomyelitis Septic arthritis (hip) Fractured femoral shaft Cerebral palsy (hemiparetic) Anisomelia Hemihypertrophy Hemiatrophy Hemangiomas Neurofibromatosis Juvenile rheumatoid arthritis Legg-Perthes disease

No. of Patients

I

II

Pattern Type III

IV

V

18 102

18 65

0 29

0 8

0 0

0 0

17 21 115 33 116 46

17 21 64 14 0 15

0 0 25 4 8 5

0 0 17 12 108" 24t

0 0 9 3 0 0

0 0 0 0 0 2

86 27 29 17 36 140

48 17 9 11 7 21

20 4 8 2 0 8

18 6 10 3 16 52

0 0 1 0 0 10

0 0 1 1 13 49

* Both type-III and type-IIIA discrepancies. I" Many of these discrepancies were detected in the plateau phase (type IIIB). F I G U R E 3 (A) The developmental pattern classification showing types I-V. (B) The distribution of patterns in several of the more frequent length discrepancy categories is shown. [Reprinted from (432), with permission.]

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

617

A TERMINAL

INTERCALARY

EMIMELIA i,',~, ,::t ;: : :~.tlt:] :.~,! i ! i i

i/J I

ii] [/

...:: t;, lelll. i/e . i I ,'~ ~ :

I

RAXlAL FIBULAR

HOCOMELIA / INCOMPLETE ;' : ~'/~ I DISTAL I j l I (Radiusand Ulna absent) i i ~i :I /I

HEMIMELIA

I,i [ /

Fibulaabsent

i~}

(5thor 4th &5th)

/ ,//

! ~ ~t ,ato~,,oe ra~sent

~."

PARAXIAL FIBULR HEMIMELIA Fibula absent All toe rays present

~...." 9 - ,

TRANSVERSE TRANSVERSE

LONGITUDINAL

LONGITUDINAL

Bii

Bi

Biii

9.L:.)

Biv

tEE9

o

,:::.*

o

00000 "''~ 00090 :::!0

CQ.

~;,;.

:!?:0

F I G U R E 4 Classifications of congenital skeletal limb deficiencies are illustrated. (A) The classification of Frantz and O'Rahilly is shown. The terms terminal, intercalary, transverse, and longitudinal are illustrated. (B) The classification of Henkel and Willert is shown in parts i-iv. (Bi) The teratological sequence of dysmelia of the lower extremities is shown. (Bii) Sequential changes of the tibia are shown in a series of cases. The femur is relatively normal throughout. (Biii) Sequential changes of the femur are shown with relative normalcy of the tibia, fibula, and foot regions. (Biv) Sequential changes of the femur are shown along with progressive tibial hemimelia. [Part A reprinted from (168), with permission. Parts Bi-Biv reprinted from (7), with permission.]

after the absent portion; radial hemimelia refers to a deficiency of the radius. The terminology of congenital skeletal limb deficiencies is shown Fig. 4A.

2. DYSMELIA

Henkel and Willert (231) proposed a differing approach to classification of the congenital malformations, which they

618

CHAPTER 8 ~ Lower Extremity Length Discrepancies

felt outlined more accurately the teratological sequence. The term dysmelia is used to refer to limb malformations varying from mild hypoplasia to partial and total aplasia of the tubular bones of the extremities and even to complete nonformation of the extremity. They arrange the abnormalities according to their degree of severity to form a teratological sequence linked by a common morphological pattern, enabling subtle variations to be included and to represent the abnormalities throughout an entire limb. The approach addresses three questions: (1) Which region of the limb and which skeletal elements are affected? (2) In what manner are they affected--by hypoplasia, partial aplasia, or total aplasia? (3) Have the affected skeletal elements also undergone fusion or synostosis? There are five main types of any teratological sequence of dysmelia: (1) distal form of ectromelia (ectromelia refers to involvement of the radius or tibia with its peripheral rays); (2) axial form of ectromelia; (3) proximal form of ectromelia; (4) phocomelia (abnormalities in which no remnants of long bones are seen between the limb girdle and the hand or foot); and (5) amelia (total loss of an extremity). The classification of malformations was derived from a survey of 693 deformed limbs (Figs. 4Bi-4Biv). This approach would seem to offer the best correlation with gene and molecular abnormalities as they are increasingly defined in relation to limb development. 3. INTERNATIONAL TERMINOLOGY FOR THE CLASSIFICATION OF CONGENITAL LIMB DEFICIENCIES

In 1973 the International Society for Prosthetics and Orthotics organized a working group to propose a terminology for limb deficiencies that would be acceptable internationally (267). They utilized both the system of Frantz and O'Rahilly and that of Henkel and Willert along with other terminologies in an effort to reach agreement among a wide number of practitioners. There still is little unanimity of opinion concerning descriptive terms for this wide array of disorders, although with an appropriate clinical and radiologic description rarely is there any doubt as to which entity is being discussed. Even after the adoption of any uniform terminology it would take several years before the studies and literature all conformed to a standard. It remains essential for those involved with these disorders to have a general understanding of the differing classifications used. In the following sections the most common terms will be used. Congenital abnormalities of the femur encompass a spectrum of disorders from those in which the femur is completely absent to those in which it is present, structurally normal, and only somewhat smaller than that on the opposite side. These can be classified into four broad groups including proximal femoral focal deficiency, coxa vara with congenital short femur, congenital short femur with diaphyseal bowing, and anisomelia in which the femur is essentially normally shaped but is smaller than that on the opposite side.

The pathoanatomy and overall management approaches were presented separately in Chapter 5. 4. PROXIMAL FEMORAL FOCAL DEFICIENCY In each of the 18 patients with proximal femoral focal deficiency in our study, severe progressive shortening of the type-I pattern occurred (433). In types A and B proximal femoral focal deficiency as defined by Aitken (11), the proximal part of the femur is intrinsically maldeveloped with no effective capability for normal reconstitution even though the acetabulum and femoral head are present. In types C and D the proximal structures are even more markedly abnormal, with no visible ossified head and the tapered diaphysis displaced proximal to the shallow, often unrecognizable acetabulum (Fig. 5A). Severe growth sequelae in this class of femoral developmental abnormalities are well-known (20). Proximal femoral focal deficiency in our series resulted in an average of 27 cm of shortening, with some lower limbs having as much as a 45-cm discrepancy. The range of femoral shortening averaged 60% (range = 40-80%) compared with the normal side. In patients classified as having type A, B, or C deficiency the shortening averaged 57%, and in type D it averaged 80%. Tibial shortening averaged 7.6% (range = 0-37%) and fibular shortening averaged 28% (range = 0-100%) in all types. This condition caused the most severe discrepancies seen in the series and presents an extremely difficult management problem. Accurate prediction of the final discrepancy is possible from the early years of life in patients with this condition, however, due to the invariable type I pattern. 5. CONGENITAL SHORT FEMUR INCLUDING CONGENITAL COXA VARA

This group in our study was composed of patients with congenital femoral anomalies, including congenital coxa vara, a congenitally short femur with coxa vara, and a congenitally short femur with lateral bowing and sclerosis but without coxa vara. Many of these patients also had associated mild or moderate anomalies of the pelvis, tibia, fibula, and foot. Excluded from this group were the patients with proximal femoral focal deficiency and those with a normally shaped and only mildly shortened femur, who were categorized as having hemiatrophy (anisomelia). The average preoperative limb length discrepancy in this group was 5.92 cm (range = 2.2-15.6 cm). It is important to note that 37 of these patients showed a type II or type III developmental pattern. If a discrepancy reached 6 cm, it generally persisted with a type I pattern. Those patients, however, in whom the discrepancy was less great often had a type II or type III pattern. Ring (408) has noted that patients with a congenital short femur alone--one with lateral bowing, cortical sclerosis, increased hip external rotation, and minimal to absent internal rotation, but without coxa varamwill continue to have an increase in the discrepancy at a regular rate with time (type I pattern). The relatively marked length discrep-

SECTION VI 9 Lower Extremity Length Discrepancies in Specific Disease Entities

619

Bi Class I

Class II

Class III

Class IV

Class

V

Bii Class VI

Class VII

Class VII!

Class IX

C 7-

6-

5-

4-

..j 2-

CHRONOLOGIC AGE SKELETAL

AGE

Orl

2

3

6

8

9

I0 ~

0~

-

2e 3e 4e 6

4

5

8

9

I0 6

FIGURE 5 (A) The classification of proximal femoral focal deficiency of Aitken is shown. Drawings are from Clinical Pediatric Orthopedics by M. O. Tachdjian, copyright 1997 by Appleton and Lange. (Bi, ii) The classification of congenital abnormalitiesof the femur into nine types as defined by Pappas. (C) A type I developmentaldiscrepancypattern in congenital short femur is shown. [Parts Bi and Bii reprintedfrom Ogden, J. (1982). J. Pediatr. Orthop. 2:331-377, 9 LippincottWilliams& Wilkins, with permission.]

ancies in femoral developmental disorders are well-known (20, 46, 277, 283, 517). Two studies on length discrepancies in congenital femoral anomalies have been published in which both proximal femoral focal deficiency and congenital short femur with and without coxa vara have been assessed together. A separate detailed study also primarily based on patients followed longitudinally in the Growth Study unit of Children's Hospital, Boston, was published by Pappas (366), in which the large number of patients assessed allowed a more detailed subclassification into nine types of deformity. Pappas defined the

percent of femoral shortening in each of the nine classes, detailed the femoral and pelvic abnormalities, assessed associated abnormalities of the tibia, fibula, patella, and feet, and defined treatment objectives (Fig. 5B). The large number of patients available for this study demonstrated a continuum of abnormalities. Class I refers to the situation in which the femur is entirely absent and the acetabular region of the pelvis markedly is hypoplastic. In class II, the proximal 75% of the femur is absent. In class III, there is no bony connection between the femoral shaft and head although the femoral head, which has delayed ossification, is present in the acetabulum.

620

CHAPTER 8 ~ Lower Extremity Length Discrepancies

In class IV, the femur is present to approximately one-half its length but the proximal abnormalities show the femoral head in the acetabulum with the head and shaft joined by irregular calcification in a fibrocartilaginous matrix. It is these four disorders that generally are referred to as proximal femoral focal deficiency. In class V, the femur diaphysis and distal end are incompletely ossified and hypoplastic. In class VI, the proximal two-thirds of the femur is perfectly normal and the hypoplasia is in the distal one-third with an irregular distal femoral region and no evident distal epiphysis. Classes V and VI are essentially examples of distal femoral focal deficiency. Class VII is congenital coxa vara with a hypoplastic femur that is shortened and somewhat bowed and also demonstrates lateral femoral condylar deficiency. Class VIII is infrequently seen but involves a proximal femur coxa valga, a hypoplastic femur, and abnormality of the distal femoral condyles with the lateral condyle being somewhat flattened. Most would include congenital short femur in this category, which perhaps most represents class VIII, although it characteristically has anterolateral bowing, which Pappas does not demonstrate. The class IX femur is essentially normal and might be defined by others as having only shortness referred to as hemiatrophy or anisomelia. Pappas also demonstrates the frequently seen underdevelopment of the lateral femoral condyle predisposing one to both a valgus deformity at the knee referable to the femoral deformity and a tendency toward lateral patellar subluxation. The ranges of femoral and tibial discrepancies found in each of the varying categories were listed. In class I the femur was completely absent. In class II the femur was shortened by 70-90% of that on the opposite normal side. The tibia also was shortened. In class III femoral shortening was 45-80% of the opposite side and tibial shortening ranged from 0 to 40%. In class IV femoral shortening was 40-67% of the opposite side and tibial shortening ranged from 0 to 20%. Class V: femoral shortening, 48-85%; tibial shortening, 4-27%. Class VI: femoral shortening, 30-60%. Class VII: femoral shortening, 10-50%; tibial shortening, minimal to 24%. Class VIII: femoral shortening, 10-41%; tibial shortening, 0-36%. Class IX: femoral shortening, 6-20%; tibial shortening, 0-15%. Vlachos and Carlioz (489) studied bone growth in 40 cases of congenital anomalies of the femur. They categorized their patients into five groups, with type I being congenital short femur without coxa vara but with shortening and curvature of the shaft; type II, with congenital short femur and coxa vara; type III, severe coxa vara with a dystrophic or pseudo-arthrotic junction between the proximal femur, which was in coxa vara, and the diaphysis; type IV, coxa vara, coxa vara with severe angular deformity proximally, and discontinuity with the shaft of the femur; and type V, almost complete absence of the proximal femur and no hip joint articulation. Relatively few patients were followed to skeletal maturity, with many being seen only to the ages of 3-10 years, such that definitive pattern progression could not be determined. They felt, however, that all patients regard-

less of diagnostic category increased at a constant rate with time, although this is somewhat distinct from our findings. Assessment of some of their charts also would indicate a type II pattern in some patients. They clearly documented both the percentage shortness and the absolute amount of shortness in centimeters in each group. In the mildest form, type I, the average shortening was 10% of the normal side, ranging between 88 and 97%, with the mean amount of shortening at 13 years of age being approximately 2.8 cm. In type II shortening averaged 30% of the normal side, with a range between 64 and 80% of normal length and the mean amount of shortening around 10 years of age already 9 cm. In type III shortening was in the range of 45% of the normal side, indicating 55% length compared to the opposite side, and associated with a mean discrepancy at age 12 years of 19 cm. In type IV the overall length was only 24-44% that of the normal side, indicating in many cases a 75% shortness that translated into a mean discrepancy of 11 cm, although patients in this group had only been followed to a little more than 2 years of age. In the most severe category, proximal femoral focal deficiency, shortness was 90% of the involved side translating into a length approximately 10% of normal and leading to discrepancies at age 5 years that were already 25 cm. A type I developmental discrepancy pattern in congenital short femur is illustrated in Fig. 5C. 6. CONGENITAL DEVELOPMENTAL ABNORMALITIES OF THE FIBULA: FIBULAR HEMIMELIA Congenital abnormalities of the fibula are the most common of the congenital deficiency syndromes of the lower extremity. Coventry and Johnson (126) noted the fibula to be the most common bone congenitally absent with congenital absence of the tibia, ulna, radius, and femur following in that order of frequency. Farmer and Laurin (159) reviewed congenital absence or severe maldevelopment of the long bones at the Hospital for Sick Children, Toronto, from 1931 to 1957 and noted 32 limbs with congenital absence of the fibula, whereas during that same period complete or incomplete absence of the following long bones also was noted: femur, 16; radius, 13; tibia, 5; and ulna, 2. The fibular abnormalities are referred to as fibular hemimelia or lateral (external) hemimelia. They are always accompanied by tibial shortening and frequently accompanied by same side femoral shortening, and it is this tibial and femoral shortening to which length discrepancy treatment relates. Because the primary bone undergoing treatment is the tibia, the disorder sometimes is referred to as congenital short tibia. In those with a hypoplastic fibula (fibular hemimelia) the major treatment considerations are the limb length discrepancies, although on occasion measures are needed to stabilize the ankle usually by varus osteotomy of the distal tibia and fibula and occasionally by orthotic support (151). Those with complete absence of the fibula present greater management problems because of the more significant length

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

discrepancy and the equinovalgus foot deformity along with a subluxed or dislocated ankle (279). There also can be anterior bowing of the distal one-third of the tibia, absence of one or more rays of the foot on the lateral sidel a tarsal coalition, and a ball-and-socket ankle joint (234). Coventry and Johnson (126) developed a classification with three types. In type I the patients have partial unilateral absence of the fibula with little or no bowing of the tibia. There is little or no deformity of the foot. The extremity always is shortened but the shortening can be quite minimal and usually is handled with an epiphyseal arrest. In type II the fibula is completely or almost completely absent and involvement is unilateral. There is anterior bowing of the tibia, dimpling of the skin, equinovalgus of the foot, and absence or deformity of the lateral rays and tarsal bones. There also is marked shortening of the extremity, and amputation was frequently needed in this group. Coventry and Johnson also defined a type III in which either the type I or type II deformity was associated with other congenital deformities, which were usually either severe deformities of the ipsilateral femur or contralateral deformities of the other leg. At present a slightly different three-part classification is favored by some. Type I is characterized by a slight to moderate shortening of the fibula, proportionately lesser shortening of the tibia, and minimal femoral shortening on some occasions. Catagni et al. (100) report tibial shortening of 3-5 cm at the end of growth with little angular deformity. On occasion, the associated outer fourth or fifth ray of the foot also is abnormal but rarely is this of clinical significance. Type II has major shortening of the fibula with particular underdevelopment or lack of development of the distal one-half to one-third. The lateral malleolus usually is absent and the ankle is unstable with the foot moving into a position of valgus deformation. The tibia is shorter than in type I disorders and tends to a valgus deformation and slight distal recurvatum with posterior bowing and anterior concavity. Type III, the most severe form, is characterized by an absent fibula, showing in addition severe deformation and shortening of the tibia and a deformed foot held in a position of equinus and valgus and often associated with dislocation or severe subluxation of the ankle. Due to the shortness of the extremity, the angulation of the distal tibia, and the deformed foot, much extended orthopedic treatment is needed often including Syme or Boyd amputation for prosthetic fitting. In a large series from Children's Hospital, Boston, presented by Pappas et al. (368), 129 of 291 patients with congenital unilateral shortening of an extremity (44%) showed shortening of the fibula greater than 10%. The extent of fibular shortening in 58% was between 10 and 30%, in 9% it was between 31 and 50%, and in 33% it was more than 50%. Although absolute length discrepancy numbers were not presented, the associated tibial shortening often was in the range of 10% or more with fibular shortening greater than 30%. There was a clear correlation between the fibula shortening and foot deformities. Among the associated limb deformities

621

(.J FIGURE 6 Achtermanand Kalamchi define a classification of fibular hemimelia type Ia (left) with fibular hypoplasia, which is relatively mild, type Ib (middle) with fibular hypoplasia, which is more with a length deficiency distally leading to a tilt of the distal tibia and its epiphysis, and a type II fibular deficiency (right) in which the fibula is completely absent. [Reprinted from (7), with permission.]

was genu valgum and instability, absence of the fourth and fifth rays of the foot, anteromedial shortening and curvature of the tibia, tarsal coalitions involving the talonavicular or talocalcaneal joints, and a domed-shaped talus. Often the fibula was absent in its proximal one-third. Achterman and Kalamchi (7) studied 97 limbs with the diagnosis of congenital deficiency of the fibula. They produced a slightly modified classification, defining type I deformities as those with hypoplasia of the fibula and type II deformities as those with complete absence of the fibula (Fig. 6). They noted that congenital anomalies of the femur were present in 76% of patients with type I deficiency and in 59% with type II. The femoral abnormalities were invariably underdevelopment of the femur, leading to worsening of the limb length discrepancy. Congenital shortening of the femur was present in 46 of the 66 limbs in which femoral abnormality was detected. Approximately 20% of the patients had some bilateral involvement. Measurements were difficult in patients with a proximal femoral focal defect, and leg length inequality was assessed when data were available in 51 cases. In those in which there was complete absence of the fibula (the type III categorization listed earlier), the amount of tibial shortening in the affected limb was 25% of normal with femoral shortening 13% of normal. In those cases in which there was hypoplasia of the fibula, the type I deformity group showed fibular shortening 7% of normal, tibial shortening 6% of normal, and femoral shortening 12% of normal, and in the type II group in which there was major shortening of the fibula, particularly with underdevelopment at the ankle, the fibular shortening was 38% of normal, tibial shortening 17% of normal, and femoral shortening (although on a small number of patients ) 23% of normal. Although detailed growth data were not presented, Achterman and Kalamchi felt that growth of the abnormal limb was proportional to that of the normal limb and that the degree of tibial shortening

622

CHAPTER 8

~

Lower Extremity Length Discrepancies

increased as the fibular deficiency became more marked. Treatment of length discrepancy was by either epiphyseal arrest or tibial lengthening depending on the clinical situation. If percentage shortening numbers are converted to length measurements for a male patient whose height is at the 50th percentile at skeletal maturity, 6% tibial shortening would represent 2.2 cm, 17% shortening 6.3 cm, and 25% shortening 9.3 cm. Lefort et al. (295) reviewed 62 cases of fibular hemimelia, concentrating in particular on the pathoanatomy of the leg and the associated femoral and tibial malformations. They stressed in particular the anterior curvature of the tibia in those cases in which the fibula was either completely absent or absent to a great extent. Absolute values for long bone shortening were not presented although percentage values were. Charts demonstrated a type I pattern of discrepancy development for developmental abnormalities of both the femur and the tibia. Hootnick et al. (237) studied 43 patients with partial or complete absence of the fibula and a congenital short tibia. They also determined that the relative difference in growth between the two limbs remained remarkably constant and thus adhered to the type I pattern of length discrepancy development. The patients studied had a strictly unilateral variant, and all measurements were determined radiographically by scanograms or from films showing both tibias on the same X-ray plate in the youngest children. The serial radiographic measurements of leg length were available in 14 patients coveting an average observation period of 9.3 years. Those with sequential radiographs were in the more severe end of the spectrum with the fibula absent from 11 patients and present but abnormal in 3. The amount of limb shortening was greater as the number of metatarsal bones diminished. There were 36 patients for whom assessments could be made in terms of the number of metatarsal bones and the amount of lower extremity shortening. In 12 patients with 5 metatarsal bones the average shortening was 8.7 cm (range = 3.6-12.7 cm), in 11 patients with 4 metatarsals the average shortening was greater at 9.5 cm (range = 3.8-13.5 c~), in 11 patients with 3 metatarsals the average shortening was 11.8 cm (range = 4.8-16.5 cm), and in 2 patients with only 2 metatarsals the average shortening was 14.6 cm (range = 11.9-17.3 cm). The average age reached in the first three groups was 11 years and in the final group 9.5 years of age. The femur was only minimally affected in these patients. In the 14 followed radiographically there is excellent documentation that the percent inhibition of growth in the affected limb compared to the normal remained unchanged from the earliest documentation to skeletal maturity. Femoral involvement at skeletal maturity was relatively small, ranging from 86 to 96% length compared to the normal side, whereas tibial involvement was somewhat greater, ranging from 73 to 82% length of the normal side. In patients followed for several years, although not quite to skeletal maturity, the same pattern persisted with femoral shortening

in all patients except one being only 92-99% of normal with associated tibial shortening of 61-90% of the normal side. Hootnick et al. felt that, if the predicted shortening was less than 8.7 cm, efforts at limb equalization were warranted, whereas if projected discrepancies were between 8.7 and 15.0 cm, amputation of the modified Syme's type was in order. In those discrepancies projected to be greater than 15.0 cm, retention of the foot and its adaptation to a prosthesis were warranted. The extent of growth discrepancy as well as management considerations was well-assessed by Choi et al. (112). They evaluated 48 extremities in 43 patients with the disorders skewed to the more severe types in their series. There were 7 fibulas of the type IA categorization, 2 type IB, and 39 type liB (complete absence of the fibula or presence of only a distal vestigial fragment according to the classification of Achterman and Kalamchi). Treatment of groups varied between those having amputation and those having lengthening procedures. Choi et al. subclassified their patients according to the amount of inequality projected for the lower limbs. In group I the percentage of shortening was 15% or less with the foot of the shorter extremity at the distal onethird of the contralateral normal limb; group II, between 16 and 25% of shortening with the foot of the shorter extremity at the level of the middle one-third of the contralateral normal limb; and group III, greater than 26% shortening with the foot of the shorter extremity at the level of the proximal one-third of the contralateral normal limb. Choi et al. concluded that lengthening was best suited only for patients in group I who had stable hips, knees, and ankles and a plantigrade foot, whereas patients in groups II and III were best served by ablation of the foot and a prosthetic fitting. Either the Syme or Boyd amputation was used, with the latter increasingly favored. The data provided indicated the extent of shortening in these disorders. The group projected the limb length discrepancy at maturity, which is highly accurate due to the invariably type I pattern in these deformities. In 15 patients in the group I category the mean discrepancy projected to skeletal maturity was 8.85 cm with a range from 5.0 to 12.07 cm. In group II, 20 involved limbs had a mean projected discrepancy of 16.29 cm with a range between 12.5 and 22.5 cm. Deformities were so great, both in terms of extent and bony deformity, in group III that numbers were not provided. Farmer and Laurin (159) recommended early Syme amputation when the length discrepancy was projected to be more than 7.6 cm (3 in.) at maturity, especially when severe foot deformity was present. A similar recommendation was made by Westin et al. (497) in their review of 32 patients with 37 fibular deficiencies. Many of their patients underwent Syme amputation, the two indications of which were a foot deformity so severe that any surgery to make the foot plantigrade and functional was likely to fail and a lower extremity length discrepancy of 7.5 cm or more that would be present at skeletal maturity in the absence of any man-

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

agement. In this group amputation was performed in 29 of 37 cases. Farmer and Laurin considered the results of the Syme amputations to be uniformly good. They noted the growth inhibition to be constant with time. The growth inhibition in those treated without amputation ranged from 7 to 12% in the tibia and from 0 to 14% in the femur. In the amputee group the inhibition was 22-42% in the tibia and 0-22% in the femur. In a listing of nine patients with unilateral Syme amputations, who were followed to skeletal maturity and who had growth data to skeletal maturity, the final femoral and tibial discrepancies ranged between 7.8 and 24.1 cm with a mean of 14.4 cm with the tibial discrepancies themselves ranging from 6.7 to 14.8 cm with a mean of 11.0 cm.

'1

9 Tibia not seen

/

a

9 Hypoplastic lower femoral epiphysis

!I tl

7. CONGENITAL DEVELOPMENTAL ABNORMALITIES OF THE TIBIA: TIBIAL HEMIMELIA

Partial or complete absence of the tibia often is referred to as tibial hemimelia. These disorders are rare and markedly less frequent than the fibular variant. There is marked shortening and bowing of the involved leg, a flexion contracture of the knee, and a rigid varus foot. Four basic patterns have been defined (258). In type IA, the tibia is completely absent and there is a markedly hypoplastic lower femoral epiphysis. In type 1B, the tibia also is completely absent, but there is a normal lower femoral epiphysis. In both instances, the fibula rides high laterally in relation to its normal position had the proximal tibia been present. In type II, the upper proximal tibial epiphysis is present as is a small portion of the metaphysis, but the rest of the tibia distally does not form. In type III, the proximal tibia is absent but the distal one-third is present. In type IV, there is a marked diastasis between the proximal tibia and the fibula and the distal one-third of the tibia is absent. The most distal tibial segment tends to be curved medially (Fig. 7). In the type II deficiency, the knee is well-preserved. In types I, II, and III the foot is in an equinovarus position. In the type IV deformity, the knee tends to be well-formed, but the talus has subluxated proximally between the now separated tibia and fibula. Treatment is directed toward early disarticulation of the knee and prosthetic fitting for type I lesions, tibiofibular fusions and prostheses for types II-IV, fibulocalcaneal fusions for types III and IV, and the frequent need for foot ablation for prosthetic fitting. Schoenecker et al. (425) studied 57 patients with congenital tibial hemimelia from the Shriners Hospital system. There were 33 type IA, 6 type IB, 15 type II, 7 type III, and 10 type IV patients. In the 54 limbs with the type I or II deficiency, there were 22 who had knee disarticulation, 25 a Syme amputation, and 1 a Chopart amputation. The foot was retained in only 6 with these two variants. There were 17 extremities with a type III or IV deficiency, and a Syme amputation was done in 9 and a Chopart in 4. In 4 the foot was retained. Schoenecker et al. report that 56 of 57 patients walked independently. Kalamchi and Dawe (261) simplified the classification, defining three types: in type I there was total absence of the

623

9 Tibia not seen 9 Normal lower femoral epiphysis

9 Distal tibia not seen

II i l 9 Proximal tibia not seen ,,,

9

,,

~

o Oiastasis

FIGURE 7 Classificationof tibial hemimelia(congenitalabsence of the tibia) by Jones et al. is shown. Severalclassificationsare in use, but this one defines the abnormalitiesclearly. [Reprintedfrom (258), with permission.]

tibia; type II, distal absence of the tibia; and type III, distal deficiency with tibiofibular diastasis. The fibula always was present but in type I it was subluxed proximal and lateral to the distal femur, with a similar position also sometimes being shown in the type II variant. In type I the disorder was also invariably associated with a marked flexion contracture of the knee, variable rotation of the leg, and marked inversion and adduction deformities of the foot. The distal femur usually was hypoplastic with marked retardation of the ossification center of the distal epiphysis. The functional status of the quadriceps muscle, the severity of the flexion contracture of the knee, and the position and function of the foot all had to be considered in planning surgical straightening or ablation. Jones et al. stressed that in all three the fibula is relatively normal in form and development, although in types I and II it often is situated proximal to the normal relationship at the knee. In some the tibial segment is greater than it appears at birth because it is present in cartilage manifesting delayed ossification. Careful clinical exam and other forms of imaging are important to clearly define the anatomic structure in the newborn. 8. POSTERoMEDIAL TIBIAL AND FIBULAR BOWING Posteromedial bowing of the tibia, which almost always is associated with bowing of the fibula, must be appreciated as an entity different from those described previously.

624

CHAPTER 8 ~ Lower Extremity Len~lth Discrepancies

FIGURE 8 Radiographsillustrate anteroposterior (A) and lateral (B) views of lower extremityof 20-month-oldfemale with posteromedialbowing. The lateral view shows full correction of posterior deformity. There had also been some correction of medial bowing on AP view. Note cortical thickening of tibia.

There is a considerable tendency for the posteromedial bow to correct during the first several years of growth, although considerable shortening often persists (Fig. 8). The condition is unilateral. The deformity is exclusively in the distal onethird of the leg and is associated at birth with foot deformities of the calcaneovalgus type. The foot and leg deformity responds well to conservative treatment with repeated application of casts or splints. On occasion, osteotomy is resorted to but only after conservative management has reached a plateau. Shortening of the affected leg, however, is progressive, increases with age, and must be followed until skeletal maturity. Sequential studies have demonstrated well the spontaneous correction of the bowing, which in both anteroposterior and lateral projections becomes either perfectly straight or sufficiently straight that osteotomy is rarely needed by the time of skeletal maturity. The most rapid straightening occurs between 6 and 18 months of age. Pappas (367) has indicated that the bowing was reduced by roughly 50% in the first 2 years, but after the age of 3 years the reduction in angulation continued at a much slower rate. Little further correction should be expected after 10 years of age. The posterior bowing almost completely resolves with the medial bowing somewhat less likely to correct fully. The fibular bowing was equal to or slightly greater than the tibial bowing and corrected more slowly, and some posterior bowing persisted in most even at maturity when tibial posterior bowing had fully corrected. The proportionate difference in lengths between the normal and bowed tibiae remains markedly stable throughout childhood, showing a type I discrepancy pattern. In the study of 33 patients by Pappas (367) the female: male incidence was 20:13 (1.5:1) and left:fight involvement

was 19:13 (1.5:1). In those patients whose limb lengths were determined radiographically within the first 2 months of life, the average initial discrepancy was 1.25 cm. Subsequent studies showed a constant increase with time. When all patients had discrepancies calculated to skeletal maturity, thus bypassing valves obscured by epiphyseal arrest surgery, tibial shortening averaged 4.1 cm with a range from 3.3 to 6.9 cm. Femoral lengths were unaffected and foot lengths little affected. Due to the extent of the length discrepancy, some form of limb equalization surgery was done or recommended for each patient, but results were not reported. The abnormality is focused in the entire distal one-half of the tibia and fibula and soft tissues of the leg. The growth discrepancy occurred exclusively at the distal end of the tibia and fibula based on radiologic appearances of proximal and distal tibial and fibular epiphyses. In 4 of the 33 patients, osteotomies were performed at an early age to correct residual bowing; all healed uneventfully. Hofmann and Wenger (235) also noted a marked tendency to spontaneous correction of the posteromedial bowing, with continuing progression, however, of the discrepancy in limb length. In 13 patients studied there was a direct relationship between the degree of initial tibial bowing and the severity of the subsequent discrepancy, which, stated slightly differently, indicates that slightly greater discrepancies in the earlier years of life would lead to greater discrepancies toward skeletal maturity. The mean posterior bowing at diagnosis was 30 ~ (range = 4 - 6 0 ~ and the mean medial bowing was 27 ~ (range = 10-45~ Hofmann and Wenger noted the relatively slow improvement in the posteromedial angulation over a few years compared with the rapid and complete correction of the calcaneovalgus deformity over a few months. The limb length discrepancy was progressive and present in each of the 13 patients. The mean discrepancy was 3.1 cm (range = 1.9-5.4 cm), but none had been followed to skeletal maturity and 10 patients were still between only 1 and 7 years of age. In the oldest 3 patients (10 years 4 months to 15 years 5 months of age) the mean discrepancy was 4.7 cm, a value similar to the 4.1-cm projection of Pappas. There were no femoral length discrepancies. A remarkably similar picture of the effects of a posteromedial angulation was reported by Carlioz and Langlais (95). They reported on 18 cases of congenital posteromedial bowing of the tibia and fibula, all of which also were associated with shortening. Both the valgus (medial bowing) and posterior bowing components corrected over the first few years of life. The valgus or medial bowing ranged between 10 and 56 ~ initially and the posterior bowing or recurvatum ranged between 10 and 65 ~ Although not all of the patients were followed to skeletal maturity, 5 had shown evidence of complete correction of both deformities during growth. The spontaneous correction occurred in 3 - 4 years. The posterior bowing or recurvatum tended to correct more completely than the medial or valgus deformation. Osteotomy was resorted to in 3 patients and in each instance healing was un-

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

eventful. Length differences were invariably seen and ranged between 10 and 20% of the length of the normal tibia. In most the discrepancy increased at a steady rate with time, but on occasion with increased growth the rate of inhibition on the involved side lessened. Five tibial lengthening procedures were performed with the preoperative average discrepancy of 4.42 cm, whereas 3 epiphyseal arrests were performed with a presurgery discrepancy of 3.7 cm. Other patients were still being followed such that additional surgery might well have been needed. The authors projected that untreated discrepancies would have reached between 1.5 and 7 cm at skeletal maturity with the majority being in the 3- to 5-cm range.

B. Skeletal Dysplasias with Asymmetric Involvement In several of the skeletal dysplasias, asymmetric involvement is strongly associated with length discrepancies. The following variants are particularly likely to show such findings. 1. HEREDITARYMULTIPLE EXOSTOSES Femoral-tibial limb length discrepancy measurements in 32 patients from Children's Hospital, Boston, indicated a range from 0.1 to 4.0 cm (437). In 2 patients, femoral and tibial shortening was equal, in 20 femoral shortening was greater than tibial, and in 10 tibial shortening was greater than femoral shortening. On occasion, the shortening was limited to either the femur or the tibia. Of 22 patients who reached skeletal maturity, 11 (50%) had limb length discrepancies in the range for which limb length equalization would normally be recommended. Of these, however, only 5 (23%) actually had the procedure. The limb length discrepancies at the termination of growth in those who did not undergo growth arrest procedures measured 2.1, 2.4, 2.3, 2.6, 3.1, and 3.5 cm. Many of the charts were not clear as to why surgery was not performed, but the following reasons were listed: subsequent planned correction of discrepancies with associated opening wedge osteotomy for deformity on the short side or closing wedge osteotomy on the long side, difficulty of performing epiphyseal arrests in regions in which large exostoses are present, clinical impression of an acceptable situation despite the roentgenographic measurements, and reluctance of patients and their families to undergo yet more procedures. When we take into consideration all 22 patients who had reached skeletal maturity and also the 7 who were close enough that it was possible to say whether they would require an arrest, there were 29 patients, 12 (41.2%) of whom had discrepancies within the recommended range for operative epiphyseal arrest. All 5 patients who were operated on had distal femoral epiphyseal arrests, and 1 had a proximal tibiofibular epiphyseal arrest as well. The limb length discrepancies at the time that the arrest was performed and the eventual limb length

625

discrepancy at the termination of growth were as follows: 2.9 cm corrected to a 1.9-cm discrepancy; 2.8 corrected to 1.2 cm; 2.4 to 1.3 cm; 2.3 to 0.5 cm; and 4.0 to 4.0 cm. There were no overcorrections, and although 4 of the 5 limbs were corrected into an acceptable range, all were short of equalization. The last patient was operated on too late by all criteria. In the other 4, however, the question arose as to whether growth anomalies in hereditary multiple exostoses might make prediction from the normal charts slightly unreliable. We therefore plotted femur-tibia length ratios along the appropriate percentile distribution in all the patients with hereditary multiple exostoses and compared them with the standard charts. The mean value for this length ratio in the patients with hereditary multiple exostoses was 1.27, which was exactly the same as the value from the charts for normal subjects with the same size distribution. Reference to the records of each patient who underwent epiphyseal arrest indicated that there had been considerable difficulty in assessing the skeletal age from the roentgenograms of the wrists. The pattern of limb length discrepancy that can occur in this condition is variable. The discrepancy can remain unchanged for several years, it can increase at slow or moderate rates, which is the usual pattern, or it can, on occasion, decrease spontaneously. The growth study data did not support the belief that there is an increase in longitudinal growth in an affected bone following removal of an exostosis. In addition, no correlation was seen between the degree of shortening in a particular bone and the number or size of the exostoses present. Limb length discrepancies in hereditary multiple exostoses were frequent, and in approximately one-half of the patients they were great enough to warrant epiphyseal arrest. These discrepancies point to the asymmetrical growth pattern in patients with hereditary multiple exostoses. The discrepancies in our series were mild to moderate and were readily managed by appropriately timed epiphyseal arrests. Extremely careful observation is required, however, as the discrepancies can remain stable, increase at varying rates, or even, on occasion, spontaneously decrease. Osteotomies also can alter limb length relationships. Some difficulties were encountered in determining skeletal age accurately due to the associated wrist and knee anomalies, but the GreenAnderson charts were appropriate for predicted corrections in this condition. The limbs were more affected than the spine, both the femur and the tibia were involved, and limb involvement was not invariably rhizomelic.

2. OLLIER'S DISEASE (ENCHONDROMATOSIS) The 17 patients with this intrinsic bone disease demonstrated a type I pattern of discrepancy development (432, 433). As varus or valgus femoral and tibial deformities often were associated with the shortening, corrective osteotomy was performed frequently and length discrepancy data that were unsullied by any bone surgery intervention throughout the growth period were rare. Relentless shortening was

626

CHAPTER 8 ~ Lower Extremity Length Discrepancies

demonstrated, however. One patient with severe involvement who was followed to skeletal maturity, with no surgical intervention, had a type I profile with a 35.7-cm discrepancy and no decline in the rate of increase. In all patients the extent of shortening paralleled the extent of radiographic involvement. The average shortening prior to physeal arrest or diaphyseal lengthening was 9.79 cm. Enchondromatosis was the second most serious condition causing extensive discrepancies, exceeded only by proximal femoral focal deficiency. 3. MAFFUCCI SYNDROME The Maffucci syndrome refers to patients with enchondromatosis, usually but not always unilateral, and hemangiomata. The enchondromas are present primarily in the hands, feet, and tubular long bones. The hemangiomas are either dermal or subcutaneous and adjacent to areas of enchondromatosis in most instances. Thrombosis of the dilated blood vessels with phlebolith formation occurs in almost half of the cases. The hemangiomas usually are absent at birth but appear within the first 4 years of life, with 25% occurring during the first year. Intracranial tumors of cartilaginous origin are seen in approximately 15% of patients. The incidence of chondrosarcomatous change is high, similar to the finding in Ollier's disease, with the reported incidence being as high as 25-50%. The matter is difficult to determine because some authors consider virtually all enchondromatosis tissue after skeletal maturity to be presarcomatous at least. The limb length discrepancy findings are similar to those with Ollier's disease, particularly when unilateral lesions predominate. 4. DYSPLASIA EPIPHYSEALIS HEMIMELICA This rare disorder often is accompanied by a lower extremity length discrepancy particularly if the epiphyseal irregularity is at the distal femur or proximal tibia. It is generally referred to as dysplasia epiphysealis hemimelica, but other suggested terms for the disorder are tarso-epiphyseal aclasis and epiphyseal osteochondroma. Many of the disorders, however, occur at the ankle joint involving either the distal tibia or on occasion the talus. The discrepancies generally tend to be mild to moderate, and more clinical difficulty is encountered with the asymmetric joint surface rather than with the discrepancy itself. Trevor (479) was the first to delineate the disorder formally, describing 8 patients initially. He noted that the initial description of such a disorder was a case described by Mouchet and Bellot (342) in 1926 involving the talus. There was no true shortening in the 8 patients assessed, although specific limb measurements were not taken. Sixteen patients with the disorder were reviewed by Connor et al. (124). Each had only one leg involved but 12 multiple epiphyses were affected. The most common sites were the distal femur, distal tibia, and talus. Treatment of the lesion was generally by local excision and was generally effective around the knee,

although some at the ankle required arthrodesis. The disorder is characterized by asymmetrical overgrowth of one or more epiphyses in a limb or of a tarsal or carpal bone during childhood. In the 16 patients, most of whom were followed to skeletal maturity, inequalities of limb length were apparent in 5, 1 with lengthening and 4 with shortening. Discrepancies generally were of a minor degree and caused few problems. In those with shortness on the involved side, the amounts were 2, 1, 1, and 6 cm. In the one instance in which there was overgrowth on the involved side, it was only 1 cm and there was multifocal involvement of the distal femur, distal tibia, and talus. In the one patient with an extensive 6 cm of shortening, there was major involvement of the lateral half of the right distal femoral epiphysis. Approximately three-fourths of the lesions are concentrated in five regions, which, starting with the most common, involve the talus, distal femoral epiphysis, distal tibial epiphysis, proximal tibial epiphysis, and the tarsal navicular bone. When deformities are present, they tend to involve either genu valgum or genu varum, valgus deformation of the ankle, and equinus deformity of the ankle. Kettelkamp et al. (273) reported 15 new cases and also reviewed the literature; they noted that inequality of limb length was found occasionally and that the affected extremity could be either shorter or longer. The distal fibular epiphysis and the medial cuneiform also were affected in some instances. When the 7 most common sites of involvement were included they accounted for 84% of all lesions. Lower extremity length discrepancies were described in only 3 of their cases, although no specific comments about measurement were made otherwise. In those 3 instances the involvement was marked, with 1 patient having 3.5 in. short and 2 with overgrowth discrepancies of 1 and 2.5 in. Fairbank (157) reported on 14 additional cases. The length of the limb was usually unaffected but on occasion discrepancies did occur. He reported 3 instances of length discrepancy in 14 patients: 0.25 in. of shortening, 1 in. of lengthening, and 0.5 in. of lengthening. If we summarize the length discrepancy descriptions from the papers of Connor et al., Kettelkamp et al., and Fairbank, the number of instances of clinically significant length discrepancy is relatively small with 11 out of 45 involved or about one-fourth (25%) of the patients. Of those with involvement, 6 had shortness on the involved side and 5 were longer on the involved side. The range of discrepancy in those with shortening was from 0.6 to almost 9 cm (0.6, 1, 1, 2, 6, and approximately 9 cm), whereas in those with lengthening the values ranged from 1 to 5.7 cm (1, 1.2, 2.5, 2.5, and 5.7 cm). In summary, therefore, whereas clinically significant lower extremity length discrepancy occurs in only approximately 25% of patients and appears to be equally divided between shortening and lengthening, on occasion it can be marked such that examination through the growing years is essential in regard to length discrepancy as well as angular deformity and range of motion.

SECTION V! 9 Lower Extremity Length Discrepancies in Specific Disease Entities

5. MELORHEOSTOSIS The rare skeletal dysplasia melorheostosis is characterized by linear radiodensity or sclerosis primarily in the metaphyseal and diaphyseal regions and can be associated with length discrepancy due to asymmetric involvement. Daoud et aL (132) reported one case in which the predicted final discrepancy was 2.5 cm of shortness on the involved side. Campbell et al. (86) reported 14 patients with the disorder, which was characterized by the radiographic long bone abnormalities with a primary clinical finding of contractures or limitation of joint motion. The tendency of the disorder is to be either exclusively monomelic or at least concentrated in one or two of the major long bones. Any bone of the body can be involved, however. The clinical findings at birth or in early childhood involve contractures, fibrosis, and abnormal skin with the bone changes occurring over a several-year period afterward. Campbell et al. noted that the affected extremity was usually shorter, although occasionally longer, but that the affected limb usually appeared larger in circumference and often had angular bone deformities. In the epiphyses and carpal and tarsal bones, there is often a spotty or patchy collection of increased radiodensity. An inequality in limb length ranging from 0.5 to 4.5 in. was found in 9 patients; the involved or more severely involved extremity almost always was shorter, except in one instance in which it was longer. Younge et aL (516) performed a study of 14 children with the disorder. The principal presenting clinical features were unilateral soft tissue contractures and inequality of limb length. The initial bony changes involved endosteal thickening or hyperostosis marked by streakiness of the long bones and spotting of the small. As in the Campbell et al. series, there was equal involvement of males and females. Eleven of the 14 were followed to at least 16 years of age. Soft tissue contractures causing severe and rigid joint deformities occurred in all and were the presenting complaint in 11. In 8 patients only one limb was involved, in 3 two limbs, in 2 three limbs, and 1 had all four limbs affected. Contractures were most commonly seen at the hip, knee, ankle (clubfoot), fingers, and iliotibial band. Each of the 14 patients had a lower extremity length discrepancy. In 13 the affected limb was shorter, ranging from 1.2 to 10.0 cm with an average shortening of 4.1 cm. In the one patient whose limb was longer the discrepancy was 2.5 cm. Firm thickening of the skin with tethering of the underlying fascia was seen in 5 patients. The problems of joint deformity, bone deformity, and contracture were marked and poorly responsive to nonoperative and even operative attempts at correction. Indeed amputation was required on four occasions, almost always after failed surgical procedures. Epiphyseal arrest was effective in treating the length discrepancies. More recently, Marshall and Bradish (317) reported successful tibial and fibular lengthening with the callotasis technique. The discrepancy at maturity was 4 cm equally divided between femur and tibia. The regenerated bone in the distraction gap had the radiologic appearance of the original bone.

627

C. Destroyed Physes If destruction and premature fusion of a physis occur, a type I pattern of discrepancy development invariably follows except at the hip (as will be described), with no tendency to compensation by the other physes in the involved bone. Such destruction occurs most commonly today with certain physeal fracture-separations. Significant lower extremity length discrepancies are seen with distal femoral growth plate fracture-separations, especially if there are several years of growth remaining because 70% of femoral growth occurs distally. Successful focal transphyseal bone bridge resection will allow growth to resume in some instances. Other causative factors are physeal ablation during tumor resection and severe osteomyelitis, particularly in the pre-antibiotic era. Wilson and McKeever (507) documented shortening in 18 (21.1%) of 85 infected bones whose physes were damaged from an adjacent focus of osteomyelitis.

D. Abnormal Growth Following Use of Neonatal Umbilical or Femoral Catheters On occasion, prolonged use of neonatal umbilical or femoral catheters for newborn illness has been associated with mild growth stimulation on the involved side. More serious sequelae were reported by McCarthy et al. (321) in four patients in whom significant lower limb length shortening occurred subsequent to neonatal catheter use. This involved either umbilical or femoral arterial catheters or both. The projected length discrepancies were massive, varying from 10 to 20 cm at maturity. Autopsies on other patients with indwelling catheters demonstrated varying degrees of thrombosis of the major associated vessels along with many embolic phenomena, so that the complication of length discrepancy problems surfacing after a prolonged or difficult neonatal hospitalization should be considered.

E. Poliomyelitis In patients with poliomyelitis involving the lower extremity, the type I pattern was seen commonly. There was, however, a distinct tendency for the discrepancy to increase most rapidly in the first 4 or 5 years following infection, with the rate of increase diminishing after that (types II and III) (Fig. 9) (433). This fact has been pointed out previously by Green and Anderson (193), Ratliff (398), and White (500). Assessment of the 115 patients in our series who were followed for 10 years or more, either to skeletal maturity or to the time of surgical physeal arrest, indicated that almost one-third of the patients demonstrated a type II or type III pattern. It has been theorized that improved function due to tendon transfers and bracing is responsible for the lessening rate of discrepancy with time. Detailed studies during the poliomyelitis era described a good but variable correlation between the extent of shortening and the severity of involvement. No correlation

CHAPTER 8 9 Lower Extremity Len~trh Discrepancies

628

7-

6-

5-

4Q ,,J RIGHT

2-

LOWER E X T R E M I T Y

-i I CHRONOLOGIC AGE SKELETAL AGE

i .5

I ,5

I 7'

I 9

I II

Is

2•

304 ~ 54 6 e

7 I~

9~

-

-

-

7

9elO s

4s

-

6e

z

F I G U R E 9 Type I pattern of a developmental discrepancy in a patient with poliomyelitis is shown.

between the age at onset and amount of shortening was found in one major study, although this finding was challenged (Fig. 9). During an era in which polio was common, Ratliff (398) noted that lower extremity shortening was seen in 219 of 225 affected patients. Barr (34) had stated that any child developing poliomyelitis before the age of 10 years with considerable difference in paralysis of the legs was almost certain to develop inequality of leg length. Barr had indicated in an earlier study that the incidence of leg shortening in poliomyelitis was as high as 78% of cases, but more detailed studies revealed an even greater increase in the likelihood of shortening. To a certain extent the shortening was related to the age at onset of the poliomyelitis and to the extent of the musculature weakened but there were not absolute correlations. Barr (34) found that 35% of patients developed a discrepancy of 1.5 in. (3.8 cm) or more, and Green noted that 8% developed a shortening greater than 2 in. (5 cm). Ratliff (398) analyzed several factors relating to limb shortening and poliomyelitis in 225 children in whom paralysis was confined to one leg. Patients were assessed between 5 and 17 years after infection. The study assessed length discrepancy prior to any limb equalization procedures. Measurements were clinical from anterior superior iliac spine to medial malleolus. Shortening had occurred in 219 of 225 children (97%) of whom 190 had some radiographic studies, which led to effective measurement to within 1~6in. The greatest incidence of onset of poliomyelitis was within the first 3 years of life, with relatively few developing the disease after the age of 8 years. The involved leg was always shorter. There were 65 patients followed into adult life, with no surgery having been performed, in which the shortening varied between 0.25 and 5 in. There was only

1 patient with the 5-in. shortening, such that the natural history would best be defined as leading to discrepancies between 0.25 and 3.5 in. because each 0.25-in. gradation between those two numbers had patients involved. The discrepancy was 2 in. or greater in 25 of the 65 patients. The discrepancy was almost always greater in the tibia than in the femur. In 184 patients assessed in this regard, 12 had tibial shortening only, 94 had tibial shortening greater than femoral shortening, 47 had equal tibial and femoral shortening, 30 had tibial shortening less than femoral shortening, and only 1 had femoral shortening only. Ratliff felt, after careful analysis, that it was impossible to predict accurately the distribution of shortening from a study of the distribution of muscles paralyzed. In another subset of patients followed for at least 9 years after disease onset (130 patients), three patterns of progressive shortening were noted. One involved a rapidly progressive discrepancy pattern in which 2.5 in. or more discrepancy occurred within 9 years or less of the onset of the disease, a slowly progressive pattern in which the discrepancy increased at a constant rate amounting to 2-3 in. by adult life, and a nonprogressive variant in which a discrepancy of up to 1.5 in. was present 5 years after onset of the disease and then remained constant until growth ceased (the type III pattern by our criteria). There was no evidence that reduction of shortening ever occurred (absence, therefore, of the type V pattern). Ratliff noted that almost 62% of patients fell into the nonprogressive group III pattern. Once again the classification of paralysis as mild, moderate, or severe could not reflect the pattern of length discrepancy progression. Although there was considerable overlap between groups, in general those with mild paralysis had lesser amounts of shortening than those with severe paralysis, with moderate in between. For example, each of the 5 patients without any shortening had mild involvement. In 102 of the patients assessed 9 years after poliomyelitis, those with mild paralysis had shortening from 0 to 2 in., those with moderate paralysis from 0.5 to 2.25 in., and those with severe paralysis from 1 to 3.5 in. There was no significant difference in the range of shortening that occurred in patients suffering disease onset between 0 and 2 years of age and those between 3 and 7 years of age. The range of shortening also was similar whether the paralysis involved only one muscle or all of the muscles below the knee. No child with paralysis only below the knee, however, showed a discrepancy greater than 1.75 in. On occasion, some transient lengthening of the involved paralyzed leg was noted during the first 2 years after the onset of paralysis, but this was always temporary and no patient was found in the series with lengthening of the paralyzed leg 5 years or more after onset. It is possible that inequality of leg length was present before the onset of the poliomyelitis, which of course would not have been observed. Premature physeal fusion was not a feature of the poliomyelitis disorder.

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

Ring (406) also studied limb shortening and its relationship to paralysis in poliomyelitis. His study did not provide absolute numbers for the amount of shortening but rather assumed (somewhat incorrectly) that it would occur at a constant rate with time. The degree of shortening was basically the same in limbs with muscle power graded as 2, 1, or 0. The amount of shortening progressively increased, however, as the grade of strength diminished from 5 to 4 to 3. Shortening was least in the strongest patients. The important feature, therefore, was not the degree of weakness but whether the muscles could function against gravity; within this group the stronger the muscle the less the shortening. The cause of shortening in poliomyelitis was not felt to be weakness per se but rather the diminished vascularity that accompanied the decreased muscle mass. Absolute correlations, however, were never possible in relation to the age at onset of the disorder or the extent or type of muscle weakness. Gullickson et al. (207) determined that the average percent shortening of unilaterally affected limbs with poliomyelitis did not appear to be different between the 0-5 year age group or the 6-10 year age group in terms of age at onset of the disorder. They also noted no correlation between muscle strength of the leg and shortening of the tibia. There was, however, definite correlation between atrophy of the thigh or leg and shortening of the femur or tibia. They provided no measurements of limb shortening although percentages were provided. In those with age at onset between 0 and 5 years of age, the percentage shortening of the tibia was 2.44% and that of the femur was 2.52% in 47 cases, and in those from 6 to 10 years of age, the percentage shortening of the tibia was 2.37% and that of the femur was 1.96% in 29 cases. Green, in a discussion published in the Journal o f Bone and Joint Surgery after an article published by Stinchfield et al. (464), reviewed the fact that, in 257 cases of poliomyelitis studied in his unit in which one lower extremity was affected and the other was normal, the average maximum growth inhibition occurred from the second to the fifth year after the onset of the disease. There was less inhibition prior to the second year and less inhibition each year after the fifth year following onset. Thus, it was recognized that not all patients with poliomyelitis increased their discrepancy at a constant rate with time. Green and Anderson showed that the final result of epiphyseal arrest treatment for poliomyelitis was within 1.2 cm (0.5 in.) of the predicted amount in 88.5% of 61 cases (193). Barr (34) provided an excellent review of leg length inequality in poliomyelitis. Based on the assessment of 371 cases during an active era of poliomyelitis, with the onset before the age of 16 years, shortening was 0.5 in. (1.3 cm) or less in 41%, over 0.5 in. but less than 1.5 in. (1.3-3.8 cm) in 24%, and 1.5 in. or more (i>3.8 cm) in 35%. Numbers from his clinic indicated that approximately 80% of patients with unilateral poliomyelitis developed shortening on the involved side. Subsequent studies showed the incidence to be

629

higher, with virtually all patients with unilateral poliomyelitis having some degree of shortening. Stinchfield et al. (464) attempted to project how much the growth discrepancy would be based on the muscle power in the affected extremity compared to the muscle power in the normal extremity. A chart was constructed based on data from 64 cases who had reached adulthood such that the final discrepancy was known. The hope for a clear correlation between muscle strength and limb length discrepancy was shortly disproven, however.

F. Hemiparetic Cerebral Palsy Most of these patients have a lower and upper extremity length discrepancy, with the shortening on the more distal parts of the involved side. In our series, type I and type III developmental patterns in the lower extremity predominated (433). Lower extremity shortening in hemiplegic children occurred almost exclusively in the tibia, a correlation also noted by Staheli et al. (456). Growth alterations in the hemiplegic child were studied in 50 children with spastic hemiplegia by Staheli et al. The hemiparetic side was always somewhat shorter than the contralateral normal side. Discrepancies in the affected upper extremities were actually larger than those in the lower extremities. The mean difference in the radius in terms of percent growth inhibition in 25 cases was greatest at 6%, whereas the inhibition in the humerus was approximately 4.2% and in the tibia approximately 3%. There was essentially no difference in the femoral lengths. The limb length inequality was far more significant in terms of functional disability in the lower than in the upper limb, but lower limb discrepancies were not particularly great because femoral shortening was not a factor. A discrepancy of 2 cm or greater occurred in only 2 of 16 patients in the older age group and epiphyseal arrest was rarely used. Of the 16 patients 11 years of age or older, the lower limb discrepancy was 1 cm or more in 10 children and greater than 2 cm in only 2. We have found, however, that lower extremity length discrepancy represents a more important consideration in many hemiparetic patients. Of the 46 patients who were followed in the Growth Study Unit for 5 years or more and who had a discrepancy of more than 1.5 cm, the average discrepancy just prior to physeal arrest or at maturity was 2.0 cm (range = 1.5-3.2 cm). Physician referral strongly influenced our study of this disease entity, unlike other diseases for which the condition itself was the reason for referral. The majority of patients with cerebral palsy in our hospital were not assessed for discrepancies. Femoral-tibial shortening alone does not give a true measurement of the functional discrepancy that may be present in the limb, as subtle dynamic or static hip and knee flexion contractures and an expected, but rarely documented, shortness in the height of the foot may further decrease the functioning length of the hemiparetic limb. If

630

CHAPTER 8 ~ Lower Extremity Length Discrepancies

an Achilles tendon lengthening is done and the lower extremity length discrepancy is not appreciated, there may be a tendency for equinus deformity to recur on a mechanical compensatory basis.

G. Septic Arthritis of the Hip Damage to the femoral capital epiphysis in septic arthritis can produce serious growth discrepancies especially if they occur in infancy (155,335). In our series such discrepancies tended to increase with time, but a type I pattern was seen in only 42% of the patients and most commonly when the infection had occurred relatively late after the age of 7 or 8 years (433). An assessment of pattern development in this group was obscured somewhat more often than in other groups because of the necessity for early and often frequent surgical intervention, although femoral osteotomy per se was done only infrequently in growing children. The patterns in this assessment were based on femoral and tibial lengths. In following such patients, however, it is important to be aware that, if dislocation occurs, the practical consideration in discrepancy relates to the distance between the iliac crest and the floor. This can be documented accurately by orthoradiographs, but a combination of measured blocks under the shortened extremity in association with a standing anteroposterior radiograph of the pelvis also is important, especially if only scanograms have been used to document the discrepancy. Even with complete destruction of the epiphysis, however, femoral shortening did not invariably become worse with time, particularly in the younger patients and those in whom femoral head dislocation did not occur. When the greater trochanter overtakes the involved femoral head in height, the femur resumes a somewhat more regular growth pattern as the greater trochanter and distal femoral physes are normal, thus accounting for the type II and III patterns that were seen. The relatively rare type IV pattern (Fig. 10A) also was seen with septic arthritis (Figs. 10Bi-10Biv). There are instances in which treatment effectively eradicates infection, allowing physeal growth to continue for several years in a seemingly normal fashion. On occasion, however, premature femoral head-neck growth plate closure occurs with no reactivation of infection, leading to a worsening of the discrepancy several years after the infectious insult (Fig. 10). The growth of the proximal end of the femur, particularly the relationship between the capital femoral and the greater trochanteric epiphyses, has been discussed in relation to normal, diseased, and experimental situations. The complexities of this particular growth area must be understood in order to plan the proper time for surgical intervention. Both Betz et al. (45) and Hallel and Salvati (215) found the subsequent length discrepancies after neonatal septic arthritis of the hip to range between 3.0 and 3.5 cm if the femoral head remained located and between 5.5 and 6.0 cm if dislocation occurred.

Hallel and Salvati (215) reported on the end result of 24 cases of infantile septic arthritis of the hip in 21 patients, 3 of whom were involved bilaterally. The disorders occurred in the first 7 months of life. A clear difference in length discrepancy was noted between those hips that had dislocated and those that remained located, with the far more serious length discrepancies occurring in the former group. Fifteen of the 21 patients had reached skeletal maturity and the others ranged between 11 and 14 years of age. Shortening of more than 2.0 cm was noted in 16 cases. In 8 instances it was due to arrested or delayed growth of the physis of the femoral head, and in 8 other cases the epiphyseal damage was worsened by proximal migration of the head in either subluxated or dislocated hips. In those cases in which the head remained located the mean discrepancy was 3.4 cm (range = 2.0-5.0 cm), and in the dislocated and subluxated cases the mean discrepancy was 6.0 cm (range = 3.0-9.0 cm). The growth of the femoral shaft measured from the proximal tip of the greater trochanter to the lateral femoral condyle was not affected by the septic process, and the growth rate of the trochanteric epiphysis remained equal to the opposite side even when the trochanter was placed into the acetabulum in the form of a trochanteric arthroplasty. A long-term multicenter follow-up of the late sequelae of septic arthritis of the hip in infancy and childhood was published by Betz et al. (45). They defined infantile cases as occurring from birth to 3 months of age and the childhood form in those whose onset occurred after age 3 months. All patients had reached skeletal maturity, and the study involved 28 patients with 32 affected hips. The lower extremity lengths were assessed at skeletal maturity in those for whom no epiphyseal arrest had been performed or those who were determined to have the predicted length discrepancy had epiphyseal arrest or lengthening not been performed. In the infantile group, the projected lower extremity length discrepancy was a mean of 3.93 cm with a range from 0 to 7.0 cm. In the childhood group, the mean discrepancy was 4.2 cm with a range from 0 to 8.0 cm. The discrepancy was 2.5 cm or greater in 7 of the 10. Wopperer et al. (512) studied 9 hips in 8 patients at a mean follow-up of 31.5 years after infantile hip sepsis. The authors had defined that group as having the sepsis between birth and 3 months of age. Six of the hips had dislocated and 3 had remained located. The leg lengths were equal in the case with bilateral sepsis and dislocation, leaving 7 patients in whom length discrepancy due to the disorder itself could be determined. In these 7 the discrepancy ranged from 0 (in 2 patients) to a maximum of 6.0 cm. The mean discrepancy was 2.86 cm. Cottalorda et al. (125) reported on the growth sequelae of 72 hips in 60 children with septic arthritis of the hip in early childhood. Treatment was variable. In 28 there was no lower extremity length discrepancy and in 16 the discrepancy was 2.0 cm or less, but in 16 with poor results discrepancies greater than 2.0 cm were present, with an average of 3.5 cm

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

631

A ,32.-

SEPTIC ARTHRITIS HIP

..j

CHRONOLOGIC AGE SKELETAL AGE

I

I

I

I

I

I

I

I

/

I

:3

5

7

9

II

13

15

17

I

2 2u

4 4u

6

8

9

II

3s4 s

5s

7

83

I0411 s II s 1 3 s l 4 s

0 e2

2s

12 13 14 15

F I G U R E 10

(A) Type IV pattern in a patient with septic arthritis of the right hip in infancy is shown. The hip was treated early such that there was continuing growth of the physis for several years followed by premature cessation of growth, which added an additional centimeter to the discrepancy at the time of skeletal maturation. (Bi-Bvi) Radiographs of the proximal femurs and hips at 1, 2, 4, 8, 13, and 15 years of age are shown. Note the excellent structural recovery by 8 years of age, mild coxa vara at age 13 years with early evidence of premature physeal closure, and clear trochanteric overgrowth with a shortened femoral neck at age 15 years due to premature physeal closure years after the initial infectious insult. [Part A reprinted from (432), with permission.]

and a range of 2.0-7.5 cm. The type IV developmental growth pattern is limited almost exclusively to abnormalities of the proximal end of the femur such as occur with septic arthritis of the hip, osteomyelitis of the femoral neck, LeggPerthes disease, and avascular necrosis of the femoral head complicating treatment of congenital or developmental dislocation of the hip. In septic arthritis patients in whom damage was relatively mild, premature fusion of the proximal femoral capital physis was sometimes noted years after the infectious insult. The premature fusion can be detected 2 or 3 years prior to skeletal maturation by the progressive change in the relationship of the level of the greater trochanteric physis to that of the proximal femoral capital physis. Therefore, it is extremely important to continue periodic assessments of these children by monitoring carefully the relationship of the head and neck to the greater trochanter until

skeletal maturity, even if the discrepancy has been in a plateau phase for several years. Although the average increase in the late phase was only approximately 1 cm, this amount could convert a clinically insignificant discrepancy to one of 2.4 cm or more and thus warrants special consideration. Gage and Cary (172) have shown the value of trochanteric epiphysiodesis in patients with severe growth damage to the femoral head-neck physis.

H. Tuberculosis Tuberculosis was frequently associated with lower extremity length discrepancy especially prior to the advent of antibiotics (250). The hip was a common site of infection with the knee involved as well fairly often. The serious nature of the disease overall and the major sequela of joint destruction

632

CHAPTER 8 ~ Lower Extremity Lenyth Discrepancies

limited the study of length discrepancy alone. There was early awareness, however, that tuberculous infection of the bones in childhood could affect growth. Langenbeck (284) described decreases in growth with tuberculous involvement of the joints in the lower extremities. Dollinger (144) described 41 cases of tuberculosis of the knee joint and found that, in the active phase, the diseased leg grew either at the same rate or on occasion more quickly. Because tuberculous infection was relatively indolent there was a chronic synovitis associated with it, and this stimulation of growth was similar to what one sees today in childhood rheumatoid arthritis. Dollinger also noted that retardation of growth occurred only when the disease became quiescent, by which time in that era sufficient damage had occurred to destroy the physis and also further decrease height by associated articular cartilage destruction. An increase in the length of the long bones after tuberculous infection also was noted by Reschke (400) and Pels-Leusden (378). In spite of the fact that tuberculosis of the hip had been a common disorder, very little has been written on the extent of length discrepancies resulting, although one would expect values similar to those with septic arthritis. Although the joint affected most commonly in tuberculosis was the hip and some of the femoral shortening was due to damage at the proximal end of the femur, it was frequently noted that there was premature distal femoral growth plate closure, which sometimes led to marked length discrepancies. Gill (179) demonstrated that, contrary to the previous feeling that it was disuse that led to the length discrepancy, there was often premature central closure of the epiphyseal cartilage plate of the distal femur and occasionally the upper tibia that caused most of the discrepancy. He described 15 cases of childhood tuberculosis of the hip in which the complication occurred. In each case the limb had been immobilized for an extremely long period of time. In each instance the age of the patient at the onset of tuberculosis was less than 7 years. The distal femoral arrests were almost invariably central epiphyseal-metaphyseal bone bridges, which, with continuing growth peripherally, caused the formation of an inverted V shape to the distal femoral epiphysis with a deeper bicondylar notch than on the normal side. In the tibia, central fusion also tended to occur, leading to continuing peripheral overgrowth and a saucer shape for the proximal end. With any degree of eccentricity of the central fusion angular deformity also occurred. In each instance the premature closure of the distal femoral or proximal tibial growth plate occurred on the side of the diseased hip. There was no evidence of tuberculosis of the knee joint to serve as a direct cause of the physeal fusion. Gill noted marked osteopenia of the entire affected femur, which he felt was due to a long period of immobilization, leaving the cartilage plate more susceptible to traumatic damage even if slight. Parke et al. (370) also noted a relatively high frequency of premature epiphyseal fusion at both the distal femur and the proximal tibia in tuberculous disease of the hip. They

also felt that it was due to the associated severe osteopenia of prolonged immobilization, leading to rupture of the physis and transphyseal bone bridge formation. Much of the osteopenia also was due to the generalized suppression of new bone formation by the tuberculous disease. These severe sequelae were described in the pre-antibiotic era. The physeal fusion was primarily central initially. The occurrence was with prolonged disease, which had to have been present for at least 2 years and frequently for much longer. They assessed 91 diseased hips with disease duration varying from 4 months to 22 years. It was not trauma per se but rather severe osteoporosis that was the precursor to premature fusion. They noted 29 cases of premature fusion at the ipsilateral knee in the 91 tuberculous hips. The tibial epiphysis was involved in 26 patients and the femoral in 17. Both bones were affected in 14 cases, the tibia alone in 12, and the femur alone in 3. A central bulging defect was seen most commonly in the proximal tibia, whereas in the femur a more fragmentary type of central physeal lesion was seen. Physeal growth of the normal limb was invariably unaffected. There was virtually no growth plate problem when the disease was less than 2 years in duration. The longer period of time the disease had been present the greater the incidence of fusion, such that each of the 7 patients with a 10-year history of the disorder or longer showed premature fusion. Sissons (447) confirmed the histopathology of extreme osteoporosis in the knee joint region in eight cases of joint tuberculosis from amputation or postmortem specimens, three of which were studied in detail. The osteoporosis appeared quite marked due to removal of the transverse trabeculae, making the longitudinal ones more prominent radiographically and also involving the subcortical bone adjacent to the articular surfaces and the bone in the neighborhood of the epiphyseal plates. Sissons could not directly address the physeal cartilage because his studies were performed in young adults. The central tibial arrest led to a saucer-shaped conformation of the proximal tibial epiphysis and articular surface. Wilson and Thompson also commented on the negative effects on growth with tuberculosis of the hip. They did note, however, that in relatively mild tuberculosis of the knee itself growth stimulation sometimes occurred.

I. Premature Epiphyseal Fusion at the Knee Complicating Prolonged Lower Extremity Immobilization Although premature epiphyseal arrest at the knee complicating prolonged immobilization of the hip for tuberculosis became a well-recognized clinical entity, in reality it was as much the prolonged immobilization as the tuberculosis itself that caused the disorder in the knee region. Subsequent studies showed that prolonged immobilization for many other hip disorders led to the same negative sequelae, including such conditions as septic arthritis, chronic osteomyelitis, and slipped capital femoral epiphysis treated with casting. Simi-

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

lar negative sequelae have been described with prolonged immobilization for congenital dislocation of the hip in works by Kestler (272) and Botting and Scrase (63). These negative sequelae occur with treatments in abduction splints in recumbency and then in hip spicas for 12 months or longer. With prolonged immobilization the disorder appears to be related to severe osteopenia such that even minor trauma can lead to transphyseal injury followed by linkage of the epiphyseal and metaphyseal circulations, leading to bone bridge formation. Ross (414) also pointed out the occurrence of distal femoral and proximal tibial premature growth plate closure in association with prolonged disability of an extremity in which the site of pathology was well away from the epiphyseal areas that subsequently fused. He studied 13 patients, 9 of whom had tuberculosis of the hip, with the others suffering from septic arthritis of the hip, slipped capital femoral epiphysis, polio, and osteomyelitis of the femoral shaft. Both of the epiphyses at the knee can be involved or there can be involvement of either the distal femoral or proximal tibial region alone. When the distal femur was involved there was a high tendency for closure to occur centrally in almost all instances, leading to the inverted V appearance of the physis and increased angulation of the intercondylar notch area. In the proximal tibia, peripheral fusions occurred more frequently than central fusions. The proximal fibula was not affected in any instance. Post-immobilization epiphyseal fusion occurred after extremely long periods of cast or splint protection of the limb, especially in comparison to current treatment protocols. In general, the immobilization was greater than 1.5 years. Examples from the case reports indicate the following periods of immobilization: 2 years and 10 months; 16 months (plaster cast); intermittent immobilization of the lower extremity for 3 years between 3.5 and 7.5 years of age; immobilization in plaster dressings between the ages of 6 and 9.5 years; 4.5 years (plaster cast) of immobilization during the period of rapid growth with weight bearing not allowed until the age of 13.5 years; a long leg brace between the ages of 6 and 8 years; 1 year and 11 months (plaster cast); 7.5 months (plaster cast); and plaster immobilization between the ages of 4 and 7 years. In those patients for whom radiographs of the knee region were available during the course of the disorder, one of the early changes of a developing growth disturbance was thinness of the physis and a transverse zone of dense bone on its metaphyseal aspect, which would indicate a Harris arrest line. The growth retardation scars were numerous and osteoporosis of the regional bone was pronounced. The contour of the epiphyseal cartilage was irregular. Bony bridges were then noted to cross the physis. At the distal end of the femur the point of arrest was commonly posterior to the central portion of the disk leading to a flexion deformity as well as shortening. At the proximal tibia the arrest was often in the postero-medial quadrant, leading to a tibia vara deformity. The tibial tubercle was sometimes the site of premature un-

633

ion, leading to genu recurvatum. Histologic examples of transphyseal bone bridge formation were found. Cartilage from the physis showed fibrillation, disorganization, and shortening of the cartilage cell columns and diminution of endochondral growth. Histologic studies from the fibula showed normal physeal cartilage. The osteoporosis led to inappropriate support of the physis, and with any weight bearing transphyseal injury was prone to occur. The interpretation that weight beating played a major role in causing further damage is supported by the fact that the proximal fibular physis almost invariably remained intact and continued to grow. This physis clearly would have been subject to the same immobilization osteopenia, but because it is a nonweight-beating structure no negative sequelae occurred.

J. Osteomyelitis Both overgrowth and growth retardation can occur in long bones that are the site of osteomyelitis. Growth changes in osteomyelitis vary depending on whether the infection is controlled in the acute phase or whether it persists as a subacute or chronic osteomyelitis (439, 440). In addition, the precise site of the infection dictates the growth sequelae. If it traverses the physis, then premature growth retardation occurs leading to a shortening, whereas if it remains in the diaphysis and metaphysis juxtaposed only to the physis, then the hyperemia leads to growth stimulation with the maintenance of physeal function. Pandey et al. (365) have pointed out the change over several decades in growth phenomena related to osteomyelitis. In the pre-antibiotic era physeal damage with growth retardation was more common, whereas afterward the disorder was better controlled although often not eradicated quickly, such that periphyseal hyperemia persisted and overgrowth predominated. Hentschel (232) pointed out as early as 1908 that osteomyelitis of the proximal tibia in the growing child could interfere with growth. In a series of growing bones affected by osteomyelitis, he was able to note instances of both shortening and lengthening of the affected bone (452). McWhorter also noted a case in which osteomyelitis in a growing child led to 1.5 in. overgrowth (452). Speed (452) pointed out that acute inflammation either arising directly in the epiphyseal area or spreading by continuity from the diaphysis could damage and destroy physeal cells, leading to growth arrest problems. The associated thrombosis served to damage further the blood supply to the growth regions. Depending on the extent of the physeal damage, there would be either complete stoppage of growth or uneven growth resulting from focal physeal destruction. If it became clear that the damaged epiphysis had ceased all growth and was no longer functional, Speed referred to the possibility of excising the contralateral epiphysis to stop its growth and thus limit any further worsening of the discrepancy. In a subsequent work Speed (453) commented on specific instances of overgrowth of a long bone due to stimulation of

634

CHAPTER 8 ~ Lower Extremity Lenyrh Discrepancies

the epiphyseal cartilage by inflammatory processes. He recognized that there was an actual stimulation of growth and that the increased length was not due to a delay in closure of the involved physis. He presented one case with chronic osteomyelitis that began at 6 years of age and persisted throughout the growing period. In adulthood the involved side was 2.25 in. longer with 1.75 in. increased length in the infected tibia. One inch of overgrowth was noted in a chronic instance of osteomyelitis of the tibia in a 5-year-old child. Speed was one of the earliest to note that overgrowth of long bones could follow infections of those bones, particularly when the infection did not directly involve the epiphyseal region. He pointed out that patients should be warned that after infectious disorders of a long bone either shortening or lengthening could occur. Wilson and McKeever (507) studied bone growth disturbances in 85 cases of osteomyelitis, also in the pre-antibiotic era. They also noted that both lengthening and shortening could occur after such infections in a growing bone. The onset of infection in all of their cases was prior to 12 years of age. Shortening of the bones was noted in 18 of 85 infections (21.2%). The focus of osteomyelitis, which caused an interruption of growth, was always located in the metaphysealdiaphyseal region adjacent to the physeal cartilage. Serial radiographs showed the epiphyseal line to be interrupted, narrowed, and eventually prematurely closed in many. The amount of shortening ranged from 1.0 to 4.0 cm although there was one case in which 6.0 cm of shortening occurred in a distal femur, which was complicated by a pathologic fracture. The amount of shortening would be dependent on the extent of the growth arrest and the age of its occurrence. Fifteen of the 18 episodes of shortening occurred in the femur, tibia, and humerus, and the mean amount of shortening in these 3 bones was 2.6 cm. Lengthening of the involved bones, however, occurred to the same extent, with 18 of 85 infected bones (21.2%) showing an increase in length of the involved bone. Infection in all patients with bone lengthening was located in the metaphysis and diaphysis but not necessarily immediately adjacent to the physis. Lengthening occurred in the 18 instances in osteomyelitis of the femur, tibia, and humerus. The range of lengthening was between 1.0 and 3.0 cm except for one instance of a 5.0-cm lengthening, which appeared to be somewhat usual. When that single case was eliminated, the 17 instances of overgrowth had a mean of 1.74 cm per bone. Trueta and Morgan (482) performed a long-term assessment of the late results in the treatment of 100 cases of acute hematogenous osteomyelitis shortly after the introduction of penicillin. The earliest studies of the antibiotic era soon began to show that shortening was becoming much less common than previously observed. They noted 7 instances in their series, and in each it was due to primary epiphyseal damage usually in infants. Increased growth in length of the affected bones, however, was common and 32 cases demonstrated lengthening, although in no case was the increase

more than 2.0 cm. Trueta and Morgan felt that the increase was due to the increased vascularization of the periphyseal area following damage of the nutrient vessels by the infectious process. Overgrowth following osteomyelitis lasted until medullary recanalization had occurred, a type III pattern in our classification, by which time the sequestra would have been resorbed and more normal vascular patterns would have been established (481,482). In chronic recurrent osteomyelitis of childhood, however, overgrowth will persist until the inflammatory focus is totally eradicated. With the increasingly widespread use of chemotherapy, such as sulfathiazole after 1941 and penicillin after 1946, the mortality from osteomyelitis dramatically diminished as did many of the bone deforming complications in those who survived. Patients with infantile osteomyelitis, within the first 3 months of life, continue to have serious bone sequelae. These were well-illustrated by Roberts (409), who documented the long-term disturbed epiphyseal growth at the knee after osteomyelitis of the distal femur or proximal tibia in infancy. The distal left femur was involved in 13 and the proximal tibia in 2. In many instances in the distal femur, the physis and epiphysis were not affected equally across the diameter of the bone such that severe varus or valgus malformation occurred. Multiple osteotomies may become needed along with attention to the length discrepancy. The lower extremity length discrepancies are of great magnitude due to the early stage at damage and due to the fact that the majority of lower extremity growth occurs at the knee region. In the 15 patients followed either to skeletal maturity or into the second decade of life, the maximum discrepancy reached a mean of 8.1 cm with a range from 1.0 to 14.0 cm.

K. Meningococcemia Meningococcal septicemia associated with necrotic purpura, cardioshock, and neurologic signs is a relatively rare infectious disorder that usually requires intensive resuscitation. The patient usually can be stabilized medically, but a serious complication of disseminated intravascular coagulopathy (DIC) can lead to ischemic lesions often with necrosis and on occasion full gangrene of the extremities. Once the initial and intermediate symptoms are stabilized, there is frequently evidence that a growth disturbance of the physes of several lower extremity bones has occurred secondary to emboli of the epiphyseal vessels directly linked to the DIC state. Formal recognition of the late sequelae of meningococcemia on physeal growth has occurred only within the past 20 years, with two early papers documenting the phenomenon. Fernandez et al. (160) reported on three patients developing epiphyseal-metaphyseal abnormalities limited to the lower extremities in multiple joints, whereas Patriquin et al. (374) described growth-related physeal changes in four children in whom the development of such lesions had been clinically unsuspected at the time of disease occurrence. Several years after the septicemic event, premature fusion of part

SECTION V! ~ Lower Extremity Length Discrepancies in Specific Disease Entities

F I G U R E 11 Radiograph of bone bridge formation following meningococcemia of infancy. X ray at 7 years of age shows the central bone bridge clearly (arrow).

of several physes with subsequent shortening and angular deformity was noted. Patriquin et al. commented on the frequency of central physeal fusions resulting in a cone-shaped epiphysis. Robinow et al. (411) also reported partial destruction of the right humeral and right femoral head and physeal regions in a 30-month-old girl 2 years after recovery from meningococcal septicemia and DIC. There also were symmetrical epiphyseal-metaphyseal lesions of the lower femoral and upper and lower tibial physes. The disorder was characterized radiographically by progressive narrowing of the physis, which could be either uniform across the entire extent or focal, and many of the focal deformities tended to be centrally situated (Fig. 11). A report from the University Hospital of Geneva, Switzerland, assessed 46 patients with meningococcemia with an average age of occurrence at 4.5 years (514). Twenty-six of the patients required immediate resuscitation in the intensive care unit and 15 of these subsequently died. Most of the survivors had serious complications involving cutaneous necrosis and 4 instances of gangrene of the upper and lower extremities. Two suffered serious bone growth disturbance due to partial or complete ischemic destruction of the physis with DIC. The negative growth sequelae usually do not become manifest for 1-2 years postinfection. They involve shortening of the affected limb and frequently angulation due to asymmetric involvement. Radiographs will show generalized narrowing of the physis in the more severe instances, although often a transphyseal bone bridge forms with the rest of the physis continuing its growth. O'Sullivan and Fogarty (360) reported two distal tibial physeal arrests as complications of meningococcal septicemia, which occurred initially in a boy 2 years 6 months and in an 18-month-old boy. The first patient returned at 8.5 years of age with a short leg and deformed ankle, whereas a diagnosis of growth abnormality was made 1 year postinfection in the 18-month-old boy. The fibula was unaffected.

635

Barre et al. (37) documented epiphyseal and physeal abnormalities following meningococcal sepsis and disseminated intravascular coagulation, They also reviewed papers from the literature comprising nine patients:, In each instance there was no evidence of bone or joint sepsis during the period of hospitalization and the acute phase of the disorder. The changes became evident relatively late, anywhere from 4 months to 6 years after the septic episode. In these four studies the patients were young, averaging a mean age of 14 months with a range from 2 weeks to 48 months, and had meningococcemia, with or without meningitis, complicated by septic shock and DIC. The skeletal abnormalities were of the epiphyses and physes and involved angular deformities and lower extremity length discrepancies. Involvement of the upper extremities was rare. The late radiographic changes either appeared as destructive of the secondary ossification centers reminiscent of avascular necrosis or involved epiphyseal-metaphyseal defects, by which is meant narrowing of the physis usually in an asymmetric fashion or with premature central physeal closure that produced cupping or peaking of the metaphyseal regions and often transphyseal bone bridge formation. Involvement has been reported at the proximal capital femoral growth plate epiphysis, leading to a coxa vara deformity, the distal femoral epiphysis at which central involvement often leads to the inverted V appearance, the proximal tibial physis at which central involvement again was relatively common, and the distal tibial epiphyses at which asymmetric involvement frequently led to varus or valgus deformation. In those instances in which the lesions were bilateral, the length discrepancy either was not present or was minor. Due to the age of the patients at the time of infection, major discrepancies have been reported particularly when involvement is at the knee and is associated with a bone bridge, which often can extend to involve almost the entire physis.

L. Physeal Damage Following Irradiation for Childhood Tumors Limb length inequality remains a fairly frequent occurrence following radiation therapy in childhood for malignant tumors of the lower extremities and also for those of the kidney, abdomen, and pelvis. Robertson et al. (410) documented lower extremity length discrepancy in 12 of 67 patients treated in childhood by radiation therapy in the previously mentioned regions. The use of high dosage and early age of treatment correlate with the severity of the growth complications. The more common tumors in the group of patients were Wilms' tumor, acute lymphocytic leukemia, non-Hodgkin's lymphoma, Ewing's sarcoma, Hodgkin's lymphoma, and rhabdomyosarcoma. No lower extremity length discrepancies were noted in any child who had symmetric radiation to the abdomen or pelvis. If the primary tumor was in tibia or femur, limb length inequality developed frequently. It also developed where there were asymmetric irradiation fields to

636

CHAPTER 8

9

Lower Extremity Length Discrepancies

the abdomen, kidneys, or pelvis. Limbs were equal when the radiation dose was less than 2400 cGy, whereas length discrepancy was quite common in those for whom the mean level was between 4000 and 5000 cGy. The development of lower extremity length inequality was uniformly related to long bone physeal irradiation of 4500 cGy or greater. Of the 12 children who developed lower extremity length discrepancies as assessed at skeletal maturity, 5 had length discrepancies of 2 cm or less and 7 of the patients developed length discrepancies ranging from 2.5 to 9.0 cm. Katzman et al. (263) pointed to the many skeletal abnormalities in patients undergoing radiation therapy in the childhood years. They assessed material from 19 survivors of 51 patients with Wilms' tumor and 13 survivors of 46 patients with neuroblastoma treated partially or completely with radiation. Epiphyseal damage was common. Although limb length discrepancy and long bone deformity were noted, no detailed analysis of these particular parameters was performed. Lewis et al. (299) studied longer term morbidity in 55 patients with Ewing's sarcoma who survived 2 years or longer. Unfortunately, the dose level required to treat the tumor effectively with a minimal likelihood of recurrence is very close to the dose level that severely damages or destroys physeal cartilage. The length discrepancy also is dependent on the age of treatment and the region affected. In those patients who received less than 5000 rad, 18% or 64% had minimum or moderate morbidity. A dose of 5000 rad was insufficient to ablate reliably a primary tumor, with many at that level showing recurrence. The intermediate dosage level of 5000-6500 rad was most effective in terms of tumor ablation, although at the higher levels the skeletal morbidity was proportionately increased. All patients had some degree of morbidity. Those defined as having minimal problems had less than 4 cm of shortening along with joint flexion deformities and muscle atrophy and fibrosis, causing only minor limitation of activities. In the moderate group shortening was defined as from 4 to 8 cm. In the severe group, gross shortening was present that could not be compensated for by epiphyseal arrest of the opposite extremity. Dawson (135) described the orthopedic problems with skeletal radiation for various lesions in 35 children. Anisomelia or lower extremity length shortening was present in 8 of 9 receiving abdominal, 1 of 2 receiving pelvic, and 6 of 7 receiving lower extremity irradiation. There is an extremely long history of awareness of the negative effects of radiation therapy on epiphyseal function dating back to the work of Perthes (384) in 1903.

M. Fractured Femoral Diaphysis The stimulation of femoral growth after a diaphyseal fracture in children who are 2-11 years old has been well-documented. It is an obligate phenomenon and occurs regardless of whether a fracture has healed with an overlap, end to end, or in a lengthened position or whether it occurred in the proximal, middle, or distal one-third of the femur. The average

3

FRACTURE FEMUR RIGHT

,.J

i I CHRONOLOGIC AGE SKELETAL AGE

i `5

t 5

; 7

i 9

I II

I 13

I 15

38 4 4 6 5 5 6 6 2 7

8 ~ 9~

II

13

15

28 , 5 - - 5

76 88

II 6

13z

146

6 6

F I G U R E 12 Type III in the developmental pattern classification is shown in this example of length discrepancy following a fractured femur and healing associated with overgrowth. [Reprinted from (432), with permission.]

femoral overgrowth from the time of fracture healing in our series was 0.92 cm (range = 0.4-1.8 cm) (431). Ipsilateral tibial overgrowth, averaging 0.3 cm, occurred in 82% of the patients. Seventy-eight percent of the overgrowth had occurred by 18 months after injury. In 85% of the patients, the overgrowth had terminated at an average of 3 years 6 months after fracture. The overgrowth phenomenon manifested itself as the type III slope-plateau pattern in 108 (93%) of the patients, with the limb length discrepancy remaining unchanged throughout the remainder of growth (Fig. 12). If a fracture heals at length or with lengthening, the overgrowth produces an upward slope-plateau pattern. If a fracture heals with shortening, the overgrowth leads to a downward slopeplateau representation. A type II pattern occurred in eight patients whose fractures had healed with excessive angulation. In these, continuing overgrowth presumably occurred due to the prolonged remodeling process. The methods of treatment and the angular deformities were reported earlier by Griffin et al. (201). Martin-Ferrero and Sanchez-Martin (319) studied femoral overgrowth in 71 patients under the age of 14 years. Each had an isolated unilateral femoral shaft fracture. Femoral overgrowth averaged 0.86 cm (range = 0.1-2.1 cm). They felt that the greatest amount of overgrowth occurred in those between 3 and 9 years of age who had had the most severely displaced fractures. Ipsilateral tibial overgrowth occurred in 60% and averaged 0.1 cm. Most of the overgrowth occurred during the first year after fracture, but it continued to a lesser extent during the second year and for as long as the fifth year postinjury in 27%. After this time the growth rate of both femurs was equal in all. These values were remarkably similar to those from the Children's Hospital, Boston, study. All patients had been treated in various forms of traction except for a hip spica cast, which was applied immediately after fracture in 7%. Reynolds (401) specifically studied growth rate changes after fractures of the shaft of the femur and tibia in children using serial radiographic measurements of length accurate to the nearest millimeter. The increased growth rate postinjury was greatest within 3 months of injury and was 38% in excess of normal with both femoral and tibial fractures. The rate then decreased but remained significantly raised for 2 years, returning to normal in the tibia approximately

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

40 months after injury and in the femur between 50 and 60 months after injury. In unilateral femoral fractures, the uninjured tibia in the same limb also underwent an acceleration of growth but to a much lesser degree. An uninjured femur was not particularly affected by fracture of the ipsilateral tibia. Reynolds concluded that the acceleration in growth of the fractured bones reached a maximum between 3 and 6 months after injury, with subsequent acceleration decreasing and the growth rate returning to normal between 3 and 5 years after injury. Stated another way, he felt that significant measurable overgrowth ceased within 2 years after fracture of the femur and within 1.5 years after fracture of the tibia. The observations of greatest value were on 55 children with tibial fractures and 32 with femoral fractures followed for between 2 and 5 years. The average increase in femoral length was 0.7-0.8 cm (range = 0.1-1.7 cm), and every femur exhibited some increase in the growth rate. The average increase in the tibia was 0.3-0.4 cm (this for actual tibial fractures) with a maximum of 1.1 cm. Overgrowth occurred in all except 3. Stephens et al. (463) studied leg length discrepancy after femoral shaft fractures in children and assessed 30 skeletally mature patients. Only isolated closed femoral shaft fractures without other injury to the limbs were assessed. All patients were treated conservatively in skeletal traction. When the fracture occurred between the ages of 7 and 13 years, the limb overgrew by about 1 cm regardless of sex, age, fracture site, or configuration. Stephens et al. recommended that treatment aim for 1 cm of overlap at union to compensate for the postfracture overgrowth phenomenon. Treatment was by either skeletal or skin traction for an average of 6 weeks followed by a spica plaster or cast brace. The average tibia overgrowth was 0.18 cm. After fracture at 7-13 years of age, limb overgrowth averaged 1.1 cm (range = 0-2.4 cm). Because tibial overgrowth accounted for only 20% of the total, the average femoral overgrowth was 0.92 cm. A more detailed review of the Children's Hospital, Boston, study is presented (431) next. Level of Fracture: In the entire group of 116 patients, 63% of the fractures occurred in the middle one-third of the femur, 28% in the proximal one-third, and 9% in the distal one-third. In the 74 patients studied in greater detail, a similar distribution was seen with 66% in the middle one-third, 27% in the proximal one-third, and 7% in the distal onethird. The site of fracture was similar regardless of age or sex. Extent of Femoral Overgrowth: In the 74 patients with initial radiographs within 3 months of fracture, overgrowth of the fractured femur occurred universally. The average femoral overgrowth in all cases from the time of healing onward was 0.92 cm. The extent of overgrowth was not dependent on sex, age at the time of fracture, the position of healing, or the level of fracture. Temporal Aspects of the Overgrowth Phenomenon: Two patterns of overgrowth were seen. In the more common pattern, overgrowth continued after fracture healing for a limited time period and then ceased with no change in dis-

637

crepancy throughout the remainder of skeletal growth. This is referred to as the plateau pattern or plateau phenomenon. Much less frequently, overgrowth continued until skeletal maturity although at a much slower rate after the first 18 months following fracture. In the group of 74 patients with early and continuing documentation of lengths, 92% (67/74) showed temporally limited overgrowth (plateau pattern) whereas 9% (7/74) continued overgrowth with time. In the entire group of 116 patients, 93% (108/116) demonstrated the plateau phenomenon whereas 7% (8/116) persisted in overgrowth. In the 74 completely studied patients, 64% of the documented femoral overgrowth occurred within 9 months of healing (1 year postfracture). By 1 year 6 months postfracture overgrowth was complete in only 12%, by 2 years it was complete in 45%, by 2 years 6 months it was complete in 45%, by 3 years in 77%, by 3 years 6 months in 85%, and by 5 years 9 months in 91%. Premature epiphyseal closure on the fractured side with a late change in discrepancy did not occur. Tibial Overgrowth: The tibia on the side of the fractured femur increased in length from the time of femoral fracture healing, such that at skeletal maturity 82% of patients had slightly longer ipsilateral tibias. In 13% tibial length was equal, and in only 5% was the tibia longer on the contralateral, nonfractured side. The average tibial discrepancy was 0.29 cm longer on the ipsilateral side (range = 0.1-0.5 cm). Epiphyseal Arrest: In the group of 116 patients, 28 underwent epiphyseal arrest. The average preoperative discrepancy was 2.39 cm (range = 1.7-3.4 cm). The discrepancies occurred as a combined result of overgrowth and healing in an anatomical or slightly distracted position. The average discrepancy post-epiphyseal arrest at skeletal maturation was 0.66 cm (range = 0-1.5 cm), with 86% of those operated showing a discrepancy of less than 1.0 cm. Five of the 28 patients requiting epiphyseal arrest had continued to increase their discrepancy with time due to continuing stimulation on the fractured side. The 74 completely documented patients with fractured femurs were studied prospectively solely on the basis of femoral shaft fracture rather than on the basis of other clinical criteria. Overgrowth was a universal phenomenon occurring in each of the 74 patients. This finding is similar to those who studied large numbers of patients by radiologic measurements: Hedberg (224), who demonstrated overgrowth in 86% (38/44); Aitken (10), who documented overgrowth in all fractured femur patients but one (64/65); and Viljanto et al. (487), who documented overgrowth in 50 out of 51 patients over 2 years of age. The average documented femoral overgrowth in our series was 0.92 cm, which compares well with others: Viljanto et al. (487), 1.07 cm; Aitken (10), 1 cm from position on discharge; Hedberg (224), 0.9 cm; MartinFerrero and Sanchez-Martin (319), 0.86 cm; Reynolds (401), 0.7-0.8 cm; and Stephens et al. (463), 0.92 cm. The same average amount of overgrowth occurred regardless of the age at fracture when the patient group was divided into the age brackets 2-4, 5-7, and 8-12 years. Hedberg (224) and

638

CHAPTER 8 9 Lower Extremity Len9th Discrepancies

Staheli (455) noted slightly greater overgrowth in those 4-8 years of age and 2-8 years of age, respectively, but Viljanto et al. (487) found no statistically significant difference in average overgrowth in those less than 3 years old, 3-9 years old, and more than 9 years old. Overgrowth also occurred regardless of whether the fracture had been allowed to heal in a shortened position, at length, or in a lengthened distracted position. This is an important finding regarding the cause of overgrowth and is in agreement with Staheli (455) and Viljanto et al. (487). Overgrowth did not appear to be influenced by whether the fracture was in the proximal, middle, or distal one-third. Because of the large size of this series and the accurate method of assessment using frequent orthoroentgenograms, these data appear to reflect the actual situation more closely than studies that rely on clinical measurements or less accurate radiologic measurements. Truesdell (480), in one of the earliest documentations of the overgrowth phenomenon, noted that overgrowth occurred whether the fracture was in the upper, middle, or lower one-third of the femur. Similar overgrowth regardless of level disagrees somewhat with the opinion of Staheli (455), who felt that proximal fractures demonstrated more overgrowth, but is consistent with the work of Viljanto et al. (487). The overgrowth phenomenon was appreciated well over a century ago by Oilier (354) and received ample documentation early in this century (78, 117, 134). Increased blood supply to the healing bone was felt by Oilier (354), Levander (298), and Bisgard (49) to be the primary cause of the overgrowth. Although there was early disagreement as to the cause of the phenomenon, with some attributing it either to "young bone yielding to pull" as the shortening corrected itself (117) or to a law of compensatory overgrowth (134), most investigators now feel that the overgrowth is a physiologic process (55) associated with the increased vascularity of the involved bone due to healing and remodeling. The increased vascularity extends to the epiphyseal plate regions where the overgrowth stimulus occurs. This now appears to be amply confirmed especially with the demonstration that overgrowth occurs regardless of the position of fracture healing and that it occurs in all patients, thus indicating that it is an obligatory phenomenon rather than one called into play only to compensate for shortening. In addition, it has been demonstrated to occur with humeral (225) and tibial (198) fractures. Kellernova et al. (268) demonstrated increased vascularity to the entire limb following experimental tibial fracture. The tibial overgrowth documented here also provides evidence for a total limb response. Increased length of the ipsilateral tibia averaging 3 mm in 82% of the patients with only 5% showing a longer contralateral tibia is taken as presumptive evidence of overgrowth in association with femoral fracture. Such tibial overgrowth also was documented by Stephens et al. (463) at 0.18 cm and by MartinFerrero and Sanchez-Martin (319) at 0.1 cm. The frequent length assessments and the accuracy of the orthoroentgenographic method have allowed for more de-

tailed study of the temporal aspect of overgrowth than has previously been reported. The impression that most of the overgrowth occurs within the first year of fracture and that it is virtually complete by 18 months (78, 134, 55) is valid, but it is demonstrated that the overgrowth phenomenon can persist for 3 or 4 years and, more importantly, that in from 7 to 9% of patients it continues for the remaining period of skeletal growth. Prolongation of overgrowth beyond 18 months or 2 years has been alluded to by Hedstrom (225) and Viljanto et al. (487) on the basis of remodeling, which can continue for that period of time. These two findings are important in following children with femoral fractures especially if they have been allowed to heal at length or with some distraction. In the eight patients whose overgrowth continued, assessment following fracture demonstrated overgrowth averaging 1.98 cm in contrast to the entire group, which averaged 0.92 cm. The overgrowth 18 months following fracture was only 39%, in comparison to the overall group where 78% of overgrowth had occurred by that time. Overgrowth was continuing 8 years postfracture in these patients. Five of the eight had a discrepancy sufficiently large to require epiphyseal arrest. In four of the eight patients no unusual factor could be identified that might have contributed to the continuing overgrowth, but in four of the patients hyperemic stimuli may well have persisted due to excessive angulation, which prolonged the remodeling phase, and to myositis ossificans, which also is associated with an increased blood supply. During the early weeks of fracture healing there is slight motion at the fracture site, and it is neither feasible nor necessary to perform accurate orthoroentgenographic length measurements. One virtually never has accurate radiographic documentation of the lengths prior to fracture: Barford and Christensen (33) in a clinical study of 431 normal children found 8% with unequal length of the lower limbs, although only 0.7% had a 1 cm or more difference. These limitations in all clinical studies have been discussed in detail (225). Both Hedstrom (225) and Bisgard (49) attempted to assess overgrowth from the time of fracture in experimental animals. It is unknown, however, whether the vascular response begins simultaneously throughout the whole extent of the femur or whether it spreads from the fracture site toward the epiphyses. If the former mechanism occurs, then overgrowth probably would be somewhat greater as it would begin earlier; if the latter, overgrowth may well represent primarily a postconsolidation repair and remodeling phenomenon. It is our feeling that the latter is the case and that radiologic measurements begun at the time of healing reflect the total overgrowth accurately. It has been suggested that overgrowth either will not be noted or will be markedly diminished in those with childhood femoral shaft fractures who are treated with external fixation devices. The matter is of some importance because it will define the mode of reduction. If the increased stability leads to more rapid healing and anatomic reduction favors less need for remodeling, then stabilization at full length

SECTION VI 9 Lower ExtremiW Length Discrepancies in Specific Disease Entities

639

A 7-

6-

HEMIATROPHY (ANISOMELIA)

5-

4-

..J

J

J

S

2-

J B 3

JUVENILE RHEUMATOID ARTHRITIS

..J

LU

",J

I I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

.5

5

7'

9

II

13

I

,3

5

7

9

II

13

15

9

I0 II

I0 II

12 13

CHRONOLOGIC AGE

It2 r

5z 6

7

SKELETAL AGE

Is 2 e

5

6a 8

6

8

9e 0s 2

CHRONOLOGIC AGE 5 z 6 SKELETAL

AGE

7

8

9

43 6 5 7 t) 85 9

I0 12 121~

F I G U R E 13

(A) Developmental pattern type I in patient with hematrophy. (B) Type V discrepancy pattern in juvenile rheumatoid arthritis. The initial stimulation led to a discrepancy with ipsilateral overgrowth, which then reached a plateau, and eventually selfcorrection with premature fusion due to the continuation of the synovitis. [Reprinted from (432), with permission.]

with anatomic reduction would be needed. If overgrowth occurs because of the fracture itself, then anatomic reduction and either internal or external fixation would still lead to overgrowth. Until such a time as detailed study of a large number of cases is complete, external fixation with the 1-cm overlap should still allow healing, and remodeling of the bayonet opposition should favor the overgrowth phenomenon. Viljanto et aL (488) reported on growth responses following operative treatment of femoral shaft fractures in children. During a 10-year period from 1957 to 1966, they treated 35 femoral shaft fractures (18% of their patient population) by operative means and assessed 19 of these patients at skeletal maturity. The average age of the group at the time of fracture was 9.8 years compared with 7.2 years in the group treated conservatively. The age range at fracture in the 19 patients undergoing assessment was from 2.6 to 16 years. Sixteen of the 19 patients demonstrated femoral overgrowth. One patient with a comminuted fracture lost 7 mm of length. The overgrowth phenomenon was demonstrated, however, even in those having surgical correction. The mean longitudinal overgrowth of the fractured femur treated by operation was 9.8 mm, with a wide range from 0 to 30 mm. The corresponding group of 52 patients treated by traction and casting had an overgrowth of 10.7 mm. Slightly less overgrowth was seen in those treated with intramedullary nailing with the overgrowth value registering 7.2 mm, whereas those treated with other means of osteosynthesis had a 13.5-mm overgrowth. In the 19 patients assessed, intramedullary nailing with the Kuntscher nail was done in 7, intramedullary Rush pins in 4, cerclage wiring in 3, screw fixation in 2, plate fixation in 2, and open reduction in 1.

N. Fractured Tibial Diaphysis Tibial overgrowth following tibial fracture has been reported to be most marked in patients who are less than 9 years old. In an isolated tibial fracture, overgrowth rarely is severe enough to require continuing long-term length assessment, but it can be troublesome when there is an ipsilateral femoral fracture.

O. Hemihypertrophy and Hemiatrophy (Anisomelia) This group of patients was discussed together in our study even though two different diagnoses, hemihypertrophy and hemiatrophy, were made (433). As the study reviewed the cases of patients who were assessed over a 40-year period, it was frequently not clear what criteria were used to include a patient under each particular designation, but the diagnosis of hemihypertrophy does not include patients who were noted to have hemangiomas, lymphangiomas, lipomatosis, or neurofibromatosis (who were assessed separately). At present the term hemiatrophy is applied to patients in whom both limbs individually appear to be normal, with the short limb diagnosed as being hemiatrophic. The developmental patterns in both groups were similar, however, and for the purposes of this classification the entity is referred to as anisomelia (Fig. 13A). The term anisomelia, which means a condition of inequality between two paired limbs, is quite obviously nonspecific and is infrequently used today as a diagnostic term. Most of these patients (57%) demonstrated a type I pattern, with the remainder equally divided between types II and III. The average maximum discrepancy in these

640

CHAPTER 8 ~ Lower Extremity Length Discrepancies

113 patients was 3.16 cm (range = 1.5-6.90 cm). Beals (40) mentions 2 patients with hemihypertrophy who experienced spontaneous correction of limb length inequality in early childhood, with the maximum discrepancies radiographically documented at 1.4 and 1.0 cm prior to 4 years of age with no discrepancy at 6 years of age. We did not note this pattem in our patient group in which inclusion required at least a 1.5-cm discrepancy at some time. Pappas and Nehme (369) also reported on lower extremity length discrepancies associated with hypertrophy in a separate study of patients from the Children's Hospital, Boston, Growth Study Unit. Many of these patients would have been included in the developmental pattern paper (433), so it is not surprising that the results were similar. Pappas and Nehme divided their assessments into patients with idiopathic hypertrophy and those with vascular disease, neurofibroma, and lymphangioma. Graphs showing the progression of length discrepancies with age showed a distribution with what we referred to subsequently as type I, type II, and type III patterns. In 35 patients with idiopathic hypertrophy, the lower extremity length discrepancy immediately prior to epiphyseal arrest averaged 3.6 cm. In those with vascular disease the mean discrepancy in 18 patients preoperatively was 3.3 cm. The mean discrepancies in the neurofibromatosis group were somewhat larger, reaching a mean of 5.3 cm preoperatively. Many patients with the neurofibromatosis disorder, however, had either no hypertrophy or clinically insignificant hypertrophy. In 90 patients followed to the time of epiphyseal arrest, the average preoperative discrepancy was 3.4 cm. In each of the four groups mean values indicated that there almost always was both femoral and tibial lengthening but that tibial lengthening was greater than femoral in each of the subgroups, usually by a 2:1 margin. Hemihypertrophy is rarely a simple increase in the lengths of the femur and tibia on the involved side. Some or many mesodermal abnormalities almost always are associated, many in particular with vascular anomalies, and there also is a high incidence of neuroectodermal abnormalities. Bryan et al. (77) studied the orthopedic aspects of congenital hypertrophy, documenting 27 cases of congenital hemihypertrophy in which the entire side of the body was affected or there was segmental hypertrophy in which only one particular extremity was markedly affected compared to the contralateral side. The maximal lower extremity length discrepancies were measured. The ranges of length discrepancy involvement seemed to be comparable in the hemihypertrophy and segmental hypertrophy cases. In 22 patients in whom discrepancy data were listed the range varied from 0.9 to 6.4 cm. Although the large majority of the patients had reached skeletal maturity, a few still had several years of growth remaining. The mean discrepancy in patients 12 years of age or older (13 patients) as measured prior to surgical correction was 3.92 cm. In a brief review of hemihypertrophy, MacEwen and Case (306) reviewed 32 cases, noting a limb length discrepancy in

26. Of these they felt that 65% had, or would develop, a discrepancy great enough to require epiphyseal arrest. In 6 patients who had already been treated with epiphyseal arrest, the mean discrepancy was 3.3 cm. In reality there is a very large subset of patients with hemihypertrophy, primarily associated with a wide range of vascular and other connective tissue abnormalities. An effort is made to delineate these in the next few sections. 1. INITIAL DELINEATION OF THE HEMIHYPERTROPHY SYNDROMEmTRELAT AND MONOD

Initial delineation of the hemihypertrophy syndrome occurred in an 1869 monograph by Trelat and Monod (478). Their work reviewed the clinical findings of the entity in which they considered previous case reports, including one of their own, in great detail. Their work was titled, "On Unilateral Partial or Total Hypertrophy of the Body," and included use of the term hemihypertrophy to describe the entity. Trelat and Monod indicated that Geoffroy SaintHilaire, in his earlier book on developmental anomalies in humans, had commented on asymmetric development, which often involved only a region or small part of one side in relation to the other. They reviewed in detail case reports beginning from 1836, which led to their presentation of "a general history of this defect of conformation." They delineated it specifically as being a hypertrophy of one side and not an atrophy of the other because atrophic conditions appeared almost always to be associated with neuromuscular abnormalities and a weakened state. The initial clinical reports were accompanied by extremely detailed measurements, and Trelat and Monod were able to indicate that the average discrepancy in hemihypertrophy was between 3 and 5 cm at maturation but that the range was great, extending from 1.5 to 19.0 cm. These numbers are still good approximations. The disorder not only involved increased length but also a proportionate increase in soft tissue size in those areas affected. Trelat and Monod also clearly pointed out the large number of vascular abnormalities present on the hypertrophic side. In virtually all instances the limb itself was the site of abnormality, which in the standing position caused an elevation of the pelvis most obvious as an elevation of the iliac crest on the hypertrophic side; it was not the pelvis or trunk itself that was abnormal, but rather these changes of position were due to the increase in lower extremity length. The truncal abnormalities and pelvic obliquity were not fixed deformations but rather were due to positional effects based on limb length discrepancy in the uptight and walking position. In 11 of the 12 cases the skeleton of the trunk was normal. Almost invariable skin changes occurred on the involved side including discoloration and changes in thickness. In many instances the skin was thickened and elevated by numerous swollen venules. In one of the cases with the largest discrepancy, enormous hypertrophy of the lower extremity (19 cm), congenital lipomas, and true elephantiasis were observed. By review of these many case studies, Trelat

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

and Monod came to define hemihypertrophy as a true congenital malformation. They noted that "vascular dilatations were frequent and encountered in varying degrees in the large majority of instances." There were two types involving skin capillaries in some and presenting with birthmarks (nevi) in others involving lesions of the subcutaneous veins, which often were true varices. The cutaneous and vascular abnormalities were always present on the hypertrophic side, and they themselves were unilateral, exactly limited to the hypertrophic regions. The cutaneous nevi were present from birth. These lesions rarely, if ever, crossed the midline, and the spots, veritable nevi of reddish discoloration, had a variable series of configurations but were always limited to the particular region of the hypertrophy. Varicose veins were often present although somewhat less frequently than the cutaneous vascular changes. These too were limited to the regions of hypertrophy, and it was felt that they were specifically connected with the general hypertrophy. Many observers also noted superficial arteries to be more dilated on the hypertrophic side. In summary, the large number of early observations made did not note any particular changes of the deep arterial circulation but rather frequent modifications of the veins and capillaries of the involved side of a general character of "angiectasies." Trelat and Monod felt that the condition was not hereditary, although they considered that the unilateral hypertrophy or hemihypertrophy was congenital. It progressed after birth and during the entire period of development, leading to the increasing limb length discrepancies. The disorder produced hypertrophy of several tissues, but some of the findings such as the varices were variable. 2. ASSOCIATION OF HEMIHYPERTROPHY WITH NEOPLASIA

There is a small, but well-documented incidence of neoplasia in association with hemihypertrophy. In many instances the hemihypertrophy had been documented well before development of the tumors. The neoplasms tend to be visceral, with involvement of liver, adrenals, and kidneys. An incidence of 2.75% hemihypertrophy was documented in a large series of Wilms' patients. Other associated tumors include adrenal carcinoma, carcinoma of the liver, pheochromocytoma, retroperitoneal sarcoma, testicular carcinoma, and cerebellar hemangioblastoma. Thus, it is evident that any patient being followed with hemihypertrophy must be assessed periodically for visceral neoplasms. No specific standard of evaluation has evolved, but renal ultrasound is an easy and relatively effective way of screening for renal Wilms' tumor and other abdominal and retroperitoneal lesions. 3. ASSOCIATION OF HEMIHYPERTROPHY WITH SILVER-RUSsELL SYNDROME

Both Silver (444, 445) and Russell (417) described a syndrome, which subsequently became well-defined and is currently most often referred to as the Silver-Russell syndrome. It was described by Silver as congenital hemihypertrophy,

641

shortness of stature, and elevated urinary gonadotropins. Russell commented on the syndrome as involving intrauterine dwarfism recognizable at birth, with craniofacial dysostosis, disproportionately short arms, and other abnormalities. With further study, manifestations of the syndrome came to involve the following: significant asymmetry; shortness of stature present at birth even though the child was born at term; variations in the pattems of sexual development, which involved elevated urinary gonadotropins, early sexual development, or markedly retarded skeletal age in relation to sexual development; unusually short fifth fingers often with increased curvature; a triangular shape to the face; and occasional cafe au lait areas of the skin. A detailed natural history study of the Silver-Russell syndrome by Tanner et al. (470) showed height at referral (4.6 years mean) averaging 3.6 standard deviations (SD) below the mean, with that level persisting throughout growth. The height at a mean age of 13 years was 3.4 SD below the mean. The predicted adult height in males was 153.5 cm and in girls was 147.0 cm. Tanner et al. felt that the limb asymmetry was a hemihypertrophy of the longer side rather than an atrophy of the shorter. The asymmetry was relatively mild, being less than 1.0 cm, which they felt was a normal variation, in 31 of 36 patients. When hemihypertrophy was present, the discrepancies in children with some growth remaining were 1.0, 1.17, 1.25, 3.33, and 6.12 cm. Specht and Hazelrig (451) reviewed the lower extremity length discrepancies in Silver syndrome in detail including 4 of their own cases and 47 from the world literature. The 51 cases gave a good overview of the length discrepancy involvement. The asymmetry was noted either at birth or during the first year of life in 29 of 40 instances where data were available. Specht and Hazelrig felt that the length difference increased in proportion to the child's skeletal growth. The length discrepancy varied from 0.5 to 6 cm, although the length data included relatively few who had reached the age of skeletal maturity. There was no sex preference for the disorder. In 4 patients who had reached skeletal maturity, each had a significant discrepancy from 4 to 6 cm. There were 7 instances in younger children in whom the discrepancy was already in excess of 3 cm. There were 40 patients for whom the length discrepancy was discussed. There appeared to be a continuing increase in the discrepancy with time. If the numbers provided are studied by age groupings, 13 values of lower extremity length discrepancy were listed from birth to 5 years of age, 15 measurements from >5 to 10 years of age, and 12 measurements from > 10 years of age to skeletal maturity. In the youngest group of 13, the mean discrepancy was 1.75 cm with a range from 0 to 5.0 cm. In the next age group the 15 measurements indicated a mean discrepancy of 2.44 cm, with a range from 0 to 4.0 cm. In the oldest age group, the 12 values listed had a mean discrepancy of 3.44 cm with a range from 1.0 to 6.0 cm. The impression is that of a type I pattern with perhaps some showing type II in the later years of growth.

642

CHAPTER 8 9 Lower Extremity Lenfth Discrepancies

Beals (40) describes a patient with Silver-Russell syndrome with a leg length discrepancy of 3.0 cm at 11 years of age. 4. ASSOCIATION OF HEMIHYPERTROPHY WITH ABNORMALITIES OF THE CEREBRAL VASCULATURE

Fischer et al. (162) described two patients with congenital hemihypertrophy and associated vascular abnormalities of the brain on the side of the hypertrophy and in the posterior fossa. The abnormalities included giant aneurysm, capillary hemangioma, and arteriovenous malformation. Literature review indicated only one previous similar patient, a girl who died at the age of 6.5 years with a vascular malformation of the thalamus. The hemihypertrophy in one patient reached 5.0 cm at 11 years of age, at which time an epiphyseal arrest was performed, whereas in the other patient the maximum length discrepancy reached was 1.6 cm at 2.5 years of age, after which the discrepancy diminished by a few millimeters over the next 15 years and surgical correction was not required. The vascular abnormalities were assessed by CT scans, arteriograms, and examination at open neurosurgical intervention in one case. The extent of the hemihypertrophy did not correlate with the presence or extent of associated cerebrovascular malformations. Other neurological abnormalities reported with hemihypertrophy include mental retardation in as many as 20% of patients, ipsilateral loss of sweating, ipsilateral indifference to pain, neurofibromatosis, metachromatic leukodystrophy, ipsilateral ventricular enlargement, and bilateral and ipsilateral enlargement of the cerebral hemispheres. Only one of the previous reports described abnormalities of the cerebral vasculature. 5. ANGIODYSPLASTIC DISORDERS ASSOCIATED WITH LOWER EXTREMITY LENGTH DISCREPANCIES

Many lower extremity length discrepancies are associated with vascular anomalies of the affected limb. This correlation has been known for some time, but in reality there is a vast array of involved vascular disorders with many patterns of irregularity noted at histopathologic examination. In addition, most patients with these disorders do not undergo surgical exploration such that clinical and some imaging criteria alone are used for diagnosis. Many of these disorders are grouped under the term "hemangiomas" in the orthopedic literature, even though, in a histopathological sense, the lesions associated with major length discrepancies are not true hemangiomas. Two patterns of nomenclature have evolved for this group of disorders with only partial correlation between them. The more specific terminology has been utilized by pathologists, cardiovascular surgeons, and plastic surgeons in their dealings with these disorders, whereas orthopedic surgeons and geneticists tend to use broader syndromal descriptions based on the clinical appearance of the limbs and particularly the nature of overgrowth (hemihypertrophy) or less frequently diminished growth (hemiatrophy)

features. Some have used the general term "angiodysplastic disorders" to refer to this broad array of conditions. The two-part article by Malan and Puglionisis (310, 311) in 1964 remains one of the clearest and most detailed correlations of the clinical symptoms, anatomic findings, and histopathologic descriptions of the wide array of congenital angiodysplasias. Mulliken and associates (343, 344) have attempted clarification by introducing a biological classification of cutaneous vascular anomalies incorporating cellular features, physical findings, and the natural history of the various disorders. The two major categories of cutaneous vascular anomalies are hemangioma, a lesion demonstrating endothelial hyperplasia, and malformation, a lesion with normal endothelial turnover. The use of the term hemangioma should be restricted to a lesion of vascular origin that grows by cellular proliferation. It is the most common tumor of infancy and demonstrates spontaneous regression. Malformations result from errors of vascular morphogenesis and are subdivided into slow-flow and fast-flow lesions. The slowflow lesions encompass capillary malformations, lymphatic malformations, and venous malformations, while fast-flow lesions involve arterial malformations, arteriovenous fistulae, and arteriovenous malformations. Combined vascular malformations are seen frequently involving capillary-lymphatic, capillary-venous, lymphatic-venous, and arteriovenous lesions. By using this classification approach, it became apparent that the vast majority of skeletal changes were associated with vascular malformations. The term hemangioma was restricted to common childhood tumors distinguished by rapid postnatal growth but followed by slow involution. When skeletal abnormalities were assessed in relation to the hemangioma-malformation categorization, it was noted that of 356 hemangiomas only 3 (1%) had bone changes, whereas 224 vascular malformations demonstrated bone changes in 77 (34%) (67). Of the 77 patients with vascular malformations, 27 were in the head and neck region and 50 were in the extremities. In the extremity group, lymphatic malformations were frequently associated with hypertrophy and on occasion with distortion of the shape of the bone. The extremity venous malformations, however, were frequently associated with hypoplasia and demineralization. When the vessel malformations were of the combined type, then both shape distortion and hypertrophy as well as hypoplasia and demineralization were found. High-flow malformations tended to produce hypertrophy and shape distortion of the bones in the terminology used by Mulliken and Glowacki. Parkes Weber syndrome is an extremity arterial malformation with arteriovenous fistulae and skeletal hypertrophy, and KlippelTrenaunay syndrome is a combined lymphatic-venous malformation with cutaneous portwine stain, with or without associated bone hypertrophy or hypoplasia. This group indicated that skeletal alterations commonly were associated with vascular malformations and rarely seen with heman-

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

giomas. For each type of vascular malformation there might be characteristic skeletal changes, although further study would be needed in that regard. Hypoplasia, for example, was characteristic of venous or combined extremity malformations and demineralization was another common finding in venous malformations, whereas intraosseous and lytic changes were characteristically seen with high-flow lesions. Many of the complex vascular abnormalities are associated with a spectrum of disordered neuroectodermal and mesodermal elements, often with skeletal overgrowth. 6. HEMANGIOMAS" OLDER GENERALIZED "ORTHOPEDIC" TERMINOLOGY As used for a diagnostic category in our paper on developmental patterns, "hemangioma" encompassed a wide histopathological variety of vascular anomalies, including capillary hemangioma (portwine stain), cavernous hemangioma, arteriovenous aneurysms and fistulae, congenital varicosities, and mixed lymphangioma-hemangioma lesions. The term "hemangioma" has been used by the orthopedic surgeons over a period of several decades in a genetic sense, that is, to refer to any type of vascular anomaly. Ipsilateral overgrowth occurred in 29 (83%) of 35 patients, whereas in the remainder the limb was shorter on the ipsilateral side (433). Nine (31%) of the 29 patients who showed overgrowth had the type I pattern, the remainder being type II or type III. Involution is a well-known occurrence in some types of hemangiomas and may account for slowing of growth stimulation. Although there were some well-documented instances in our series when partial resection of the soft tissue lesions also diminished the growth stimulation, most patients demonstrated a type II or type III pattern in the absence of any surgery. The average discrepancy prior to bone surgery in this group of patients was 3.09 cm (range = 1.8-5.60 cm). The developmental pattern in this group must be observed carefully in the middle years of the first decade of life, as considerable discrepancy may develop and projections that are based on the expectation of the same rate of growth stimulation can be misleading. In a study of hemangioma of the extremities in 35 cases in which a broad array of histopathologic diagnoses were included, McNeill and Ray (323) noted 16 limbs with equal lengths, 8 with overgrowth on the involved side, and 11 with shortening or atrophy on the involved side. It appears from the work that overgrowth, when present, was far more extensive than shortening because no indication of the amount of shortening, other than its presence, was made. In those with overgrowth, when amounts were listed, the extent ranged from 1 to 8.75 cm. A generalized overview was provided by Maroteaux (316). a. Newer Terminology. If the more specific terminology of Mulliken and Glowacki (344) is used, a different pattern of length discrepancy involvement occurs. Due to the fact that the vast majority of hemangiomas undergo spontaneous

643

regression, axial skeletal overgrowth is almost never seen in conjunction with these lesions. Most hemangiomas first appear in the early neonatal period and 80% grow as a single lesion with the rest as multiple lesions. They are far more common in females with a 3-5:1 female:male ratio. Once established, there is rapid neonatal growth during the first 6 - 8 months of life with a plateau in size being reached at 1 year, following which the lesions grow proportionately with the child with regression then beginning around 5 years and continuing until approximately 10 years of age. As the tumor proliferates in the superficial dermis, the skin becomes raised and develops a vivid crimson color. If it is deeper in the dermis and into the subcutaneous layer, the overlying lends a bluish color to it. Its contained blood cannot be evacuated completely by manual pressure. Hemangiomas can be divided into a proliferating phase, during which they enlarge, and the subsequent involuting phase. Histologically there is endothelial cell proliferation. Active pericytes are often seen as well. As the tumor regresses endothelial cell activity also diminishes. Mast cells make their appearance during the involuting phase. 7. HEMANGIOMAS OF BONE Hemangiomas of bone do occur but are extremely rare. In addition, when present they almost invariably affect either the vertebral bodies or the skull. In those infrequent instances when they affect the long bones, they tend to be innocuous in terms of clinical significance. Cohen and Cashman (115), however, have reported one instance in which hemihypertrophy of a lower extremity was associated with multifocal intraosseus hemangioma. In one patient the described involvement of both the right femur and right tibia led to overgrowth on the right side, but this was due to a combination of the right tibia being 2.9 cm longer at 12 years of age while the femur was 1.1 cm shorter. The hemangioma was present in both epiphyseal and metaphyseal bone, but the intervening growth plate was structurally normal. The authors noted that, in the few previous cases of long bone hemangioma described, no growth changes were reported. 8. VASCULAR MALFORMATIONS Vascular malformations are all present at birth but on occasion may not be obvious. Venous malformations in particular may manifest later during childhood. Males and females are affected equally. Vascular malformations tend to grow proportionately with the child. Each structural category of vascular malformation has a typical cutaneous appearance. Each of the four major categories of vascular malformation has its particular histopathological appearance, but combined forms (CVM, CLM, LVM) can be difficult to distinguish even histologically. The characteristic histopathologic finding is a lining of endothelial cells without high activity, similar to what is seen in normal vessels. Vessel walls particularly with venous and lymphatic anomalies are of variable

644

CHAPTER 8 9 Lower Extremity Length Discrepancies

thickness, and mast cell distribution in vascular malformations is normal. Treatment of the vascular malformations is beyond the scope of this work but has been well-discussed. The malformations are either arterial, venous, capillary, lymphatic, fistulae, or combinations of these. Two characteristic hematologic findings are associated with these disorders. In hemangiomas, platelet trapping, referred to as the Kassabach-Merritt phenomenon, is a rare complication of large hemangiomas and usually occurs in the neonate showing thrombocytopenia, usually less than 10,000 per mm 3. In vascular malformations, a coagulopathy can occur but is usually associated with larger extensive venous anomalies and is a true intravascular coagulative defect. It is the vascular malformations that are associated with the various syndromes leading usually to hemihypertrophy and appendicular skeletal overgrowth. The reasonably well-defined syndromes associated with hemihypertrophy, or infrequently hemiatrophy, are described in the following sections. It is important to recognize that terminology used between different institutions and in different papers, particularly in papers written several decades apart, can vary significantly. Those disorders or syndromes seen frequently with lower extremity length discrepancies include the Klippel-Trenaunay syndrome (KTS), Parkes Weber syndrome, Proteus syndrome, Beckwith-Wiedemann syndrome, congenital arteriovenous fistula, cutis marmorata telangiectatica congenita, and Maffucci disease. Some link the first two syndromes, referring to the Klippel-TrenaunayWeber syndrome. Maffucci disease was discussed earlier in conjunction with Ollier's disease. a. Klippel-Trenaunay Syndrome. This named syndrome derives from an initial description in 1900 by Klippel and Trenaunay (275). It refers to a congenital abnormality consisting of a cutaneous nevus (portwine hemangioma), varicose veins, and bone and soft tissue hypertrophy. It is usually unilateral and affects the lower limb, but occasionally more than one limb is involved. In their initial description, Klippel and Trenaunay stressed that the bones on the hypertrophic side, though larger, maintained a normal anatomic shape and proportion. They called the syndrome and their article, "le noevus variqueaux osteo-hypertrophique." Klippel and Trenaunay noted that there was considerable awareness of the existence of hemihypertrophy, often including skull and facial asymmetry, asymmetric soft tissue development of the extremities, and particularly asymmetric vascular anomalies of the skin and subcutaneous tissues. In describing the syndrome, which subsequently came to bear their names, Klippel and Trenaunay commented on the unique triad of abnormalities involving the nevus, varices, and osteohypertrophic changes. They pointed out that Trelat and Monod (478) as early as 1869 had described a syndrome in which the characteristics were unilateral bony hypertrophy generally involving the lower extremities and frequently accompanied by a vascular dilation, which could be of two types: capillary involving the nevi and subcutaneous venous

involving the varices. In the words of Klippel and Trenaunay, however, "they did not observe the remarkably frequent coexistence of the nevus, the hypertrophy of the skeleton and the venous dilations," feeling that they were secondary occurrences to the primary symptom, which was the hemihypertrophy. Klippel and Trenaunay, on the other hand, insisted "on the simultaneous presence of these three principal signs." They pointed out several examples from the literature of their era of the soft tissue abnormalities associated with hemihypertrophy and referred to previous cases to support their contention that the triad of abnormalities was a unique congenital lesion apparent since birth. They stressed that the nevus and the hypertrophy had existed since birth and that the varices became evident around 8 or 10 years of age. The length discrepancies described in their own cases and in examples of the triad from the literature were extensive; they listed discrepancies in the lower extremities of 4.5, 4, 2, 4, and 9 cm. The triad of disorders present in the same subject were not lesions grouped by accident or coincidence but rather resulted from a single disorder. The hemihypertrophy sometimes involved an entire side, including the face and skull, but was often segmental and occasionally just involved either the hand or the foot and sometimes individual digits. The bone was uniformly increased in size in terms of length, width, and thickness. In spite of this, however, the anatomic shape was normal and there were no angular or other deformations. The hypertrophy was present at birth but progressed and thus worsened with growth. On occasion, the discrepancy increased to 10 cm as noted by Oilier. Due to the limb length discrepancy, secondary deformations of the pelvis and lower spine occurred. The disorder was not particularly rare. Generally it is accepted today that the affected tissues do not contain hemodynamically significant arteriovenous communications, but often there are other soft tissue, lymphatic, and bony abnormalities. It is similar to the Parkes Weber syndrome and many studies link the two therefore describing the Klippel-Trenaunay-Weber syndrome. Baskerville et al. (39) note that the presence of arteriovenous fistulae excludes the diagnosis of Klippel-Trenaunay syndrome and is characteristic of the Parkes Weber syndrome. In their detailed study of 49 KTS patients with 56 abnormal limbs (53 lower extremity and 3 upper extremity) the male:female ratio was 1.3:1 (39). All 49 had visible varicosities, 47 had a nevus, and 47 had limb hypertrophy. In 43 of the patients the abnormality was noted at birth. There was no clinical evidence of an arteriovenous fistula in any of the 49 patients, including 22 who had formal arteriography. In 36 patients the affected limb was longer (greater than 2 cm) and in only 2 was it shorter. The feet were also usually hypertrophic. Varicose veins were visible in all 49 patients. Sixty-eight percent had a large, incompetent vein on the lateral aspect of the limb, which arose on the dorsum of the foot or ankle and extended a variable distance up the leg. Phlebography showed that approximately 50% of the abnormal lateral veins drained

SECTION VI 9 Lower Extremity Length Discrepancies in Specific Disease Entities

into the main stem leg veins, with 33% extending to the buttocks and draining via the gluteal veins into the internal iliac vein. More than one-fourth of the patients had intrapelvic venous abnormalities as well. Other abnormalities included 15% with lymphedema and 22% with cutaneous lymphatic vesicles. Five patients demonstrated gigantism of the toes. A convincing argument is made to separate the Klippel-Trenaunay syndrome in which arteriovenous fistulae are absent and the Parkes Weber syndrome, which is characterized by arteriovenous fistulae (38). It is also evident that appreciable overgrowth of the limb may occur in the absence of arteriovenous fistulae. The difference in lower extremity length discrepancies rarely increased after the age of 12 years, which would imply a type II or type III discrepancy pattern. Absolute numbers for the extent of the discrepancy were not provided, but epiphyseal stapling was performed in only 4 patients (4.5 %). It should be noted, however, that problems with vascularity render length discrepancy management potentially dangerous. Baskerville et al. suggested that KTS was caused by a mesodermal abnormality during fetal development (38). There appears to be no true atresia of the deep veins with abnormalities concentrated in the superficial system. Histologic studies, similar in all patients, showed an increase in the number and diameter of the venules in a cross section of the deeper layers of the dermis and subdermal fat. There was also widespread hypertrophy of the smooth muscle in the walls of the subcutaneous veins due to response to chronically increased flow. There were normal deep veins and normal calf pump function in 60% and 84% of patients, respectively. Both Bourde (64) and Baskerville et al. (38, 39) suggest that KTS is due to persistence of part of the embryological vascular system and that a mesodermal defect acting primarily on angiogenesis could explain the condition. The findings were felt to be consistent with a later regression than normal of the embryonic vascular reticular network in the developing limb bud. This itself would lead to increased capillary and venular blood flow during intrauterine development and to the superficial varicosities. A mesodermal abnormality would also explain the other venous abnormalities with the syndrome involving developmental abnormalities, such as the absence of valves in the deep veins or reduplication of axial veins and a large, often valveless, lateral venous channel. Management was suggested to concentrate on the correction of bony overgrowth, excision of soft tissue hypertrophy, and removal of varicose veins but only in those veins causing pain or discomfort (38, 39, 388). The widespread removal of varicose veins had not proved to be particularly successful. b. Parkes Weber Syndrome. Parkes Weber pointed out the combination of vascularization abnormalities with hemihypertrophy of the limbs. A particular syndrome has come to be associated with his name; it involves the KTS triad of cutaneous nevus (portwine hemangioma), varicose veins, and soft tissue and bone hypertrophy with arteriovenous

645

malformations. In his initial presentation in 1907, he drew attention "to a group of cases in which hypertrophy of one limb, or else hemi-hypertrophy, is found to be associated with tumor-like overgrowth in the corresponding portion of the vascular system" (371). Weber felt that the disorder was congenital in most instances. In differentiating a particular syndrome from previously known developmental abnormalities of the vascular and lymphatic systems, he pointed out that the condition under consideration was distinguished by the associated vascular abnormalities and the actual increase in length of the bones of the affected limb. He pointed out that even in the late 1800s there were many cases of hemihypertrophy reported with some form or other of angiomatous formation. It was, however, in Weber's second communication in 1918 that he linked the congenital limb hypertrophy specifically to dilatation of arterial and venous trunks with a specific arteriovenous communication (372). He did this by specifically indicating that "the communication between the arterial channels and the venous channels may be so free that in it a definite kind of thrill or pulsation, rhythmical with the heart's contractions, is transmitted to the veins as in cases of arterio-venous anastomosis of traumatic origin." Weber referred to the condition as "congenital or developmental phlebarteriectasis" or hemangiectatic hypertrophy of limbs. Both of his papers were accompanied by abundant descriptions of vascular anomalies and limb overgrowth from the late eighteenth and early nineteenth centuries from English, French, and German reports. c. Proteus Syndrome. The Proteus syndrome is similar to the Klippel-Trenaunay and Parkes Weber syndromes but is characterized by more frequent progression of hamartomatous growth. It is truly a syndrome in the sense that there are an extremely large number of associated abnormalities. The disorder was first recognized by Cohen and Hayden (116), who differentiated the various symptoms from other overgrowth syndromes. The name Proteus, however, was suggested a few years later in an article by Wiedemann et al. (504), who described 4 patients documenting partial gigantism of the hands or feet, nevi, hemihypertrophy, subcutaneous tumors, macrocephaly or other skull anomalies, accelerated growth, and possible visceral affections. In a later note, yet further developmental abnormalities were described including, but not limited to, progressive kyphoscoliosis, subcutaneous abdominal lipomas, dilated veins and/or hemangiomas, facial anomalies, possible mental retardation, and occasional seizures (79). One of the characteristics of this disorder is the changing phenotype with time. Many of the patients are normal at birth to clinical assessment and develop the characteristic findings over the first year of life. Once established, the abnormalities tend to be progressive throughout childhood with growth of the hamartomata and generalized hypertrophy increasing. The disorder is stable, however, after puberty. Morbidity is considerably greater than with the KT or PW syndromes. Of 11 patients evaluated by Clark et al. (113), 2 required amputation of the leg, 6 had fingers or toes

646

CHAPTER 8 9 Lower Extremity Length Discrepancies

removed, and 2 women had breast implants and reconstruction. Spinal stenosis and neurological sequelae can develop due to vertebral anomalies or tumor infiltration. There is concern about neoplastic disorders of many types with this syndrome. Because it was described only in 1979 and clearly defined in 1983 careful assessment is needed. An extremely large number of developmental abnormalities can be associated with this syndrome. The characteristic findings involve hemihypertrophy, generalized thickening and soft connective tissue swellings of the skin and subcutaneous tissue, and macrodactyly. The overgrowth may involve the whole body, it may be unilateral involving one limb, or occasionally it is localized even to a digit. The skin and subcutaneous thickening is associated with lipomata, lymphangiomata, or hemangiomata. There is a relatively high proportion of skeletal abnormalities other than the hemihypertrophy, including bony prominences over the skull, angular deformities of the knees, scoliosis or kyphosis along with the dysplastic vertebrae, hip dislocation, and hallux valgus. There is disproportionate involvement of the hands and feet with macrodactyly and often soft tissue hypertrophy particularly over the plantar surfaces of the feet. d. Beckwith-Wiedemann Syndrome. Anothersyndrome occasionally associated with hemihypertrophy leading to lower extremity length discrepancy is the Beckwith-Wiedemann syndrome (41,503). This clinical entity is characterized by macroglossia, omphalocele or other umbilical anomalies such as umbilical hernias, macrosomia with large muscle mass and thick subcutaneous tissues, linear creases in the lobule of the external ear, and large kidneys. Other developmental anomalies are seen in differing patients. Hypoglycemia is present in early infancy in approximately one-half of the cases. e. Congenital Arteriovenous Fistula. Horton (239)drew attention to the overgrowth phenomenon with congenital arteriovenous fistula involving the extremities in an early report. He detailed findings in 23 upper and lower extremity disorders in which actual documentation of the length discrepancy was made. In the group of 23 patients described, the limb circumference was from 2 to 8 cm greater than the corresponding normal side, and in 18 of the cases there was an increase from 0.5 to 7.0 cm in the length of the bones on the abnormal side. With lower extremity involvement, tilting of the pelvis and lateral curvature of the spine invariably were seen. The involved extremity was hypertrophied and showed marked evidence of engorgement, swelling of superficial veins, and skin ulcers in most. There was a marked increase in the pulsations of the arteries and a definite increase in the surface temperature of the extremity involved. Horton used the term arteriovenous fistula to designate any abnormal communication or communications between arteries and veins by means of which arterial blood passes from an artery to a vein without passing through a capillary bed. There were 15 patients with lower extremity length discrepancies described, all of whom had reached

skeletal maturity except for 3 at 4, 5, and 9 years of age. Two of the patients had no increase in length on the involved side, but the others all showed an increase in length. The discrepancies varied from 0.5 to 7.0 cm. In the 13 patients with lower extremity length overgrowth, the mean discrepancy was 3.2 cm with 9 of the 13 showing a discrepancy of 2.5 cm or more. McKibbin and Ray (322) implicated abnormalities of venous return in experimental arteriovenous fistulae with bone overgrowth. The direction of blood flow in the vein distal to the fistula was reversed for a considerable difference. As the venous collateral channels, including those in the bone, developed the periphyseal blood supply was also altered. f. Cutis Marmorata Telangiectatica Congenita. Cutis marmorata telangiectatica congenita is characterized by a persistent vascular mottling of the skin usually involving the limbs and usually in an asymmetric pattern. The disorder was described initially by Van Lohuyzen (484) in 1922, and a detailed review by Gelmetti et al. (175) in 1987 listed approximately 150 cases described in the literature. Spontaneous regression has been observed in the majority of patients in the first few years of life, but many lesions do persist to adulthood. The disorder is referred to as congenital phlebectasia by many. Lower extremity length discrepancies occur on occasion in the disorder. As with many congenital vascular abnormalities, there is a high frequency of multiple associated congenital abnormalities such that it is a syndrome including other neuroectodermal and mesodermal defects. Gelmetti et al. pointed out that abnormalities of the central nervous system, musculoskeletal system, and vascular system were involved. Several instances of hemiatrophy or hemihypertrophy on the involved side have been described. In the detailed review referred to earlier there were 11 cases of shortness or hemiatrophy of the involved side and 7 cases of hemihypertrophy of the involved side, as well as other categorizations involving retardation of growth and asymmetric growth in which specific limb length determinations were not clear. Dutkowsky et al. (146) described a 15-year-old male with 2.9 cm of shortness on the involved side and 2 other patients 2-3 years old with 1-1.6 cm of shortness on the involved side. These studies indicate that some discrepancy in limb length could be present in as high as 25% of cases, that both hemiatrophy and hemihypertrophy on the involved side could be seen, and that evidence of growth retardation perhaps is somewhat greater than that of growth stimulation. 9. GENERAL CLASSIFICATION OF ANGIOMATOUS LESIONS BASED ON THEIR SIZE AND POSITION (GOIDANICH AND CAMPANACCI) Goidanich and Campanacci (189) derived a classification of the congenital and developmental vascular disorders based on their position and size in the extremities. Their approach was based on the dissatisfaction with confusing and variable descriptive terms used for this group of disorders.

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

They felt that the clinical management could best be guided by consideration of the size and position of the vascular anomalies. Their six groups included (1) localized cutaneous and subcutaneous vascular hamartomata; (2) localized deep vascular hamartomata; (3) extensive deep vascular hamartomata; (4) multiple deep vascular hamartomata; (5) diffuse deep vascular hamartomata; and 6) infantile angioectatic osteohyperplasia. Goidanich and Campanacci felt that each group had distinctive clinical and pathological characteristics but agreed that there were cases that showed transitional features. For each group, they documented whether there was hemihypertrophy or hemiatrophy and also assessed bone changes radiographically, varicose veins, skin temperature, skin angiomata, pain, functional impairment, and swelling. The classification was derived from a study of 94 cases with the largest group, deep localized vascular hamartoma, comprising 45 cases and the smallest, infantile angioectatic osteohyperplasia, comprising 7 cases. Goidanich and Campanacci considered all of the angiomata of the soft tissues of the extremities to be hamartomatous, with the lesion being present from birth and growth ceasing after skeletal maturity. In terms of extremity length, examples of both overgrowth and retardation were noted. In patients with cutaneous and subcutaneous vascular hamartoma, limb length was essentially equal in each of 11 cases. In all 7 cases with infantile angioectatic osteohyperplasia overgrowth occurred on the involved side. In the other four groups, however, were examples of both overgrowth and retardation in association with the vascular lesions. In the four groups were 16 patients with overgrowth and 17 with decreased growth. In terms of length discrepancy, therefore, even site and extent of the lesion with the two exceptions noted earlier offer little prognosis as to either the extent of any discrepancy or whether there will be overgrowth or retardation. 10. LOWER EXTREMITY LENGTH DISCREPANCIES IN VASCULAR MALFORMATION SYNDROMES Most studies of Klippel-Trenaunay syndrome in the orthopedic literature link it with Parkes Weber syndrome, referring to the entire entity as Klippel-Trenaunay-Weber (KTW). In addition, relatively few reports exist on lower extremity length discrepancies with the Proteus syndrome because it was described only recently. Peixinho et al. (377) described 8 patients with KTW syndrome treated for lower extremity length discrepancy. There were 7 patients who had growth remaining at the time of treatment, ranging between 10 and 15 years of age, and 1 patient treated at skeletal maturity at 19 years of age. This report concentrates on the more severe variants but does show the extent of the length discrepancies that can develop. The range of disorders treated was between 1.5 and 10.0 cm, with those undergoing epiphyseal arrest between 2.9 and 10.0 cm. In these 6 patients the average discrepancy was 6.35 cm. In the 5 patients with epiphyseal arrest followed to skeletal maturity, the mean discrepancy initially was 6.82 cm and at skeletal ma-

647

turity it had decreased to a mean of 2.1 cm. An additional patient had shortening of the femur at skeletal maturity, diminishing the discrepancy from 4.8 to 0.8 cm. Guidera et al. (205) described 28 patients with limb overgrowth and a diagnosis of either KTW (18 patients) or Proteus (10 patients) syndrome. The results were pooled. Most patients had a huge array of mesodermal abnormalities, indicating the need for extremely careful and detailed total body assessment and careful follow-up with growth. All patients but 1 had extremity involvement. Twenty-seven of 28 had lower extremity involvement, consisting of overall limb hemihypertrophy in 15, localized gigantism in 6, and macrodactyly in 10. Nine patients had lower extremity length discrepancies varying from 1 to 12.8 cm. Two patients exhibited clinodactyly. Other deformities similar to those mentioned earlier were present. No specific analysis of the discrepancies was made. The authors warned along with others of using extreme caution concerning the use of surgical intervention. Soft tissue debulking operations and vascular operations in particular often appeared to be of dubious benefit especially with the danger of heavy bleeding. Guidera et al. felt that the timing of epiphyseal arrests was not predictable in equalizing limb length discrepancies, although that does not appear to be the experience of others. Amputation was frequently required for the more massive limb deformities generally described as "grotesque." Rogalski et al. (413) utilized the term angiodysplastic lesions of the extremities to review 41 patients. They used this terminology rather than the syndromal terms referred to earlier, indicating that "identification of a specific syndrome appeared to depend on the speciality and training of the treating physician and was not predictive of initial symptoms or outcome." Rogalski et al. felt they could not categorize the vascular malformations according to either histologic or angiographic descriptions. They preferred to describe the malformations utilizing the system of Goidanich and Campanacci based on the size and location of the malformations. Eleven patients had limb length discrepancies with hemihypertrophy and 4 of these had epiphyseal arrest. No specific details concerning the extent of the discrepancies were given. The shape and size of the extremities and the associated vascular problems as well as the large number of additional malformations throughout the body made the length discrepancy itself relatively less important than in other individuals. The authors felt that anatomic location and overall size were predictive of symptomatology and thus were the more important factors at our current level of knowledge in relation to patient assessment. Rogalski et al. determined that the majority of the lesions (59%) were subcutaneous, whereas 20% were specifically identified as intramuscular. No bony lesions were identified. All limb length discrepancies were associated with overgrowth of the involved limb. The authors also commented on the limitations of surgical interventions particularly when cosmetic improvement was being sought. Paley and Evans (363) also felt that it was the depth and

648

CHAPTER 8 ~ Lower Extremity Length Discrepancies

extent of the lesions that were the most significant prognostic characteristics. A review of 40 patients with KT syndrome from the Mayo Clinic documented an equal male:female distribution of 1:1 (183). Thirty patients had findings noted immediately at birth, with 3 being diagnosed before the age of 1 year and 4 between 1 and 6 years of age. A strict distinction was made between Parkes Weber syndrome and KTS with patients with arteriovenous fistulae not included in the KTS grouping. The lower extremity was involved in 38 patients (95%) and the upper extremity in 6 patients (15%), with 4 patients (10%) bilateral. The disorder was unilateral in 34 patients (85%), bilateral in 5 (12.5%), and crossed bilateral in 1 case (2.5%). The affected extremity was longer in every case. In 29 patients documented by scanogram the average difference in documented length was 2.39 cm, with the largest at 12 cm. Hemangioma was present in 39 cases (97.5%), whereas lymphangioma was the histologic diagnosis in 1 case (2.5%). The two were found together in 5 additional cases. The typical portwine fiat cutaneous hemangioma was present in 30 cases (75%). Varicosities were marked and significant with incompetent perforators in 25 cases (62.5%). Seven of the 25 had persistent large embryonic veins on the lateral aspect of the thigh. Arteriography was performed in 9 cases with no arteriovenous fistulae diagnosed. All except 2 of the patients were still under 10 years of age in terms of their assessment for length discrepancy. It again was stressed that the KTS diagnosis required three main symptoms: varicosity, hemangioma, and extremity hypertrophy. The most common form of hemangioma was the capillary type or portwine nevus, which had a pink to purplish color and represented diffuse telangiectasias of the superficial vessels of the dermis. P. N e u r o f i b r o m a t o s i s

Neurofibromatosis is associated with length discrepancies in two types of clinical situations (8, 320, 336, 485). 1. BONES APPEAR STRUCTURALLY NORMAL The bones may be structurally intact, but the soft tissues of the affected limb are characterized by care au lait spots of the skin and/or actual subcutaneous neurofibromas. Growth stimulation was documented in our series on the involved side in 17 patients who did not have a tibial pseudarthrosis (433). Prior to physeal arrest, the average discrepancy was 4.40 cm (range = 2.0-8.8 cm). Shortening was also associated with neurofibromatosis in the 6 patients in whom a pseudarthrosis occurred. The type I pattern was common, although type II, type III, and type V patterns were seen also. 2. CONGENITAL PSEUDARTHROSIS OF THE TIBIA There may be a congenital pseudarthrosis of the tibia, which is often seen with neurofibromatosis, although cutaneous abnormalities or neurofibromas of the involved segment may not be present. Because surgical intervention in an

attempt to establish union was so frequent in the patients with pseudarthrosis, the natural length discrepancy patterns in our series were infrequently available for assessment. Some information is available, however, on the maximum lower extremity length discrepancies reached from large series of pseudarthrosis of the tibia described previously. Virtually all of these patients will have had surgical interventions, often multiple times, on the affected tibia in efforts to obtain straightening and union. The discrepancies were at times worsened by angular deformity, the need to resect scarred sclerotic diaphyseal tissue, and intramedullary rod passage through the tibial epiphyses in efforts to enhance stabilization. Van Nes (485) pointed out, as many had previously, that a considerable amount of the shortening in congenital pseudarthrosis of the tibia was due to associated developmental abnormalities of the distal tibial epiphysis, which often led not only to diminished growth but also to premature fusion. The distal tibial physis was often delayed in appearance, indicative of early problems with normal development. Throughout the growing years the physis can often be noted to be misshapen and thin. Van Nes stated that the retardation of growth in the distal tibial epiphysis indicated involvement with the same segmental dysplasia as the distal part of the diaphysis that caused the pseudarthrosis originally. Both the pseudarthrosis and the retardation of growth were to be considered symptoms of the same developmental defect of the distal tibia. The closer the pseudoarthrosis to the epiphyseal region, the greater the epiphyseal involvement and the less the associated growth. Van Nes clearly pointed out that spontaneous physeal fusion could occur as early as 10-12 years of age. He presented case reports on 22 individuals. Length discrepancy measurements were listed in 18 patients, virtually all after 10 years of age and prior to any epiphyseal arrest for the length discrepancy. Though not a natural history study, the data showed the extent of the discrepancies in the more severe variants of the disorder. One patient had no discrepancy and the others ranged from 1.25 to 11.0 cm. The mean discrepancy in the 18 listed patients was 5.4 cm. The discrepancies were due to a combination of factors, including angular deformity, multiple surgical procedures some of which involved resection of bone in efforts to obtain union, diminished function of the distal tibial epiphysis due to its involvement in the dysplastic process, and premature fusion of the distal tibial epiphysis. On occasion, the physis was damaged by intramedullary nails passed through it, although many instances of continued growth following this procedure also were noted. In a long-term study from the Mayo Clinic, Masserman et al. (320) reviewed 52 cases. Of these, 20 (38%) had a known diagnosis or the clinical manifestations of neurofibromatosis. Lower extremity length discrepancies were documented in 32 of the 52 patients. Of these, 5 had no specifc number mentioned but appeared to be insignificant. In all instances except 1 the involved leg was shorter. There were 2 patients whose limbs were equal, 8 in whom the difference

SECTION VI ~ Lower Extremity Length Discrepancies in Specific Disease Entities

was between 0 and 1 in., 5 between 1 and 2 in., 3 between 2 and 3 in., 4 between 3 and 4 in., 1 between 4 and 5 in., and 3 with 6 in. of discrepancy. The 1 patient with tibial overgrowth had a 2.0-cm discrepancy. Morrissy et al. (336) analyzed 40 cases of congenital pseudarthrosis of the tibia in detail. Half of the patients in the series had neurofibromatosis. As in the study by Masserman et al., the diagnosis of associated neurofibromatosis does not affect the overall result in comparison with the group without that diagnosis. The amount of shortening in congenital pseudarthrosis of the tibia correlated extremely well with the eventual result achieved. This serves as a biological reflection of the extent of the bone abnormality, which affects not only the diaphyseal regions but also the entire bone. In those with good results the average amount of shortening was 1.4 cm (range = 0-4.0 cm), with fair results the average amount of shortening was 3.4 cm, with 1 patient having a 6.0-cm shortening and another a deficit at one time of 8.1 cm, and with poor results the average shortening was 5.5 cm (range = 2-8 cm), with all but 1 patient having at least 4.0 cm of shortness. Amputation was required eventually in 14 patients. As the results indicate, a significantly greater amount of shortening existed in patients with a tenuous union or nonunion. Much of the growth discrepancy was secondary to failure of growth in the distal physis, whereas relatively normal growth proximally continued.

Q. Juvenile Rheumatoid Arthritis Lower extremity length discrepancies occur commonly in patients with juvenile rheumatoid arthritis in which joint inflammation is prolonged and asymmetric. There had been recognition for some time that asymmetric bone growth occurred in many instances of moderate to severe juvenile rheumatoid arthritis and that this was characterized by overgrowth on the involved side early in childhood and by a tendency to premature physeal closure, leading to shortening in those affected toward the end of skeletal growth (24, 25, 70, 81, 99, 280, 290, 468). Griffin et al. (200) noted overgrowth as great as 2.6 cm in 1 patient with knee involvement and shortening as great as 3.8 and 5.1 cm in others. A retrospective study at Children's Hospital, Boston, determined the course of limb length discrepancies occurring in patients with monoarticular and pauciarticular juvenile rheumatoid arthritis (446). Data were assessed on 36 patients followed to skeletal maturity, (group I), 15 patients who had not reached skeletal maturity but who had been followed for 4 years or more, (group II), and 49 patients followed for 3 years or less (group III). In 72 of the 100 patients the onset of the disease occurred before they were 5 years old, and 90 patients had involvement of the knee. All patients in whom the disease developed before the age of 9 years had overgrowth of the involved extremity, but that overgrowth never exceeded 3.0 cm. The major discrepancy developed within the first 3 - 4 years and either in-

649

creased very slowly thereafter, remained level, or decreased. Of the 36 patients who were followed to skeletal maturity, in 29 a discrepancy of 1.5 cm or more developed at some time during the period of assessment. Twelve of the 36 patients had diminution of the discrepancy to the extent that epiphyseal arrest was not required. Fifteen eventually had an epiphyseal arrest. Rapid premature closure of the epiphyseal growth plate occurred only in those patients in whom the disease developed after the age of 9 years. This led to immediate shortening of the involved side and on occasion to marked limb length discrepancies of as much as 5.1 and 5.9 cm. Of the 51 patients included in groups I and II, although all had some length discrepancy, in 35 (70%) a length discrepancy of 1.5 cm or more developed during the study period. Twenty-one patients had a discrepancy of between 2.0 and 2.9 cm, and in 3 it was 3.0 cm or more. In the patients with unilateral disease whose disease onset was before the age of 9 years, the involved side was almost invariably the longer one (39 of 40 cases). Those whose disease onset occurred within the first 3 years of life tended to have a discrepancy greater than 1.5 cm (24 of 34 patients) relatively more often than children 3-8 years old, but overgrowth in the younger children never amounted to more than 3.0 cm. When the disease occurred initially after the patient was 9 years old (5 patients), the involved side usually became shorter (with one exception). Regardless of age at onset, the major discrepancy that developed did so within the first 4 years after disease onset. Thereafter, in the group I patients the discrepancy either increased very slowly (6 patients), remained unchanged (14 patients), or decreased spontaneously (12 patients). The other 4 patients, whose discrepancies continued to increase, were patients with late onset of the disease whose epiphyses fused prematurely on the involved side. Continuing involvement of the knee for several years in 1 patient resulted in a continuing increase with time until a prearrest discrepancy of 2.4 cm was reached, but such an occurrence was unusual. In the 12 patients whose discrepancies decreased, there was a gradual inhibition of growth in the involved limb over several years. In 7 of the 12, the discrepancy became clinically insignificant. The changes in discrepancy with time were 2.8 to 0.3, 2.5 to 1.2, 2.3 to 1.3, 2.2 to 0.4, 2.0 to 0.2, 1.8 to 1.4, and 1.5 to 0.9 cm. Evidence for rapid premature epiphyseal growth plate closure was noted in only 4 patients who had discrepancies because of shortness on the involved side. In patients with the monoarticular or pauciarticular form of juvenile rheumatoid arthritis, who have predominant involvement of the major joints of one lower extremity, the length discrepancy occurs as a result of two factors: (1) stimulation of the epiphyseal growth plates, predominantly about the knee joint, during the time the disease is active and for some time afterward and (2) inhibition of the growth potential of the involved extremity.

650

CHAPTER 8 9 Lower Extremity Length Discrepancies

The epiphyseal growth plates at the knee account for 70% of the growth potential of the lower extremity. They are sufficiently close to the synovial capsule to be affected by the hyperemia that occurs during the inflammatory process, but are not adversely affected by the concomitant destructive process. Thus, because the knee is commonly involved in the type of arthritis under discussion and so often is involved unilaterally early in the course of the disease, a great potential exists for growth stimulation of the involved lower limb. However, the severity of the disease often causes a decrease in the use of the involved extremity, either because of the patient's symptoms or because of the treatment. Such decreased use can explain the gradual reduction in epiphyseal stimulation and inhibition of growth. When the patient has activity of the disease in early adolescence, consequent hyperemia about the knee could cause rapid and premature fusion of one or both of the growth plates, with a sudden decrease in growth of the involved lower limb. In this study, the most common and well-defined pattern of development and progression of length discrepancy was ipsilateral lengthening in patients who had onset of the disease before the age of 5 years. The major part of the discrepancy then occurred within the first few years of the disease. Thirty-nine of the 40 patients with unilateral involvement whose disease began before the age of 9 years showed this pattern. We suggest that early in the course of the disease the predominant factor is stimulation of the epiphyses about the involved knee joint. The pattern of development of a limb length discrepancy described most often is followed by lack of a significant continuing increase in the discrepancy. Although occasionally (in two patients in this series) the discrepancy will increase over the ensuing years, in most patients it will decrease or remain unchanged with time. It is important to note the rapid premature closure of the epiphyseal growth plates about the involved joint that occurred at the end of growth in four children in this series. All of them had onset of the disease after the age of 9 years, and all showed shortening of the involved side. Their length discrepancies ranged from 1.9 to 5.9 cm, with the two patients who were 9-10 years old at onset of the disease showing larger discrepancies at skeletal maturity than the other two, who were 11-12 years old at onset. Early epiphyseal growth plate fusion has been recognized as a complication of juvenile rheumatoid arthritis for many years. In patients with monoarticular or pauciarticular juvenile rheumatoid arthritis, variable developmental patterns occurred. Types I, II, III, and V were all seen (Fig. 13B). The knee is the most common area of involvement in juvenile rheumatoid arthritis and involvement there is most likely to result in clinically significant discrepancies. The type I pattern was seen most frequently in patients whose initial attack of rheumatoid arthritis occurred after the age of 9 years. In these patients, a type I pattern with shortening on the involved side developed due to the relatively rapid, prema-

ture physeal fusion of the bones comprising the involved joint. The type II and type III patterns were seen most often in patients whose initial synovitis occurred in the first few years of life and resulted in physeal stimulation and overgrowth. In our series, once the synovitis had resolved, physeal growth altered toward a more normal rate and the discrepancy either persisted unchanged or increased at a much slower rate. The type V pattern resulted from a slowing of physeal stimulation over a few years prior to plate closure (Fig. 13B). Whether the type V pattern was due to decreased use or due to an alteration in the timing mechanism for closure due to disease, or to both, is uncertain, but the phenomenon itself was well-documented. It was not possible to predict which patients would have a type II, III, or V pattern. Similar overgrowth can occur following inflammatory conditions about the knee in childhood such as tuberculosis, septic arthritis, and hemophilia. Indeed, in what appears to represent a description of a type V pattern, Phemister (387) quoted Bergmann as observing "equalization of length years after overgrowth produced by tuberculosis of the knee beginning in early childhood."

R. Thalassemia Premature fusion of the epiphyses has been recognized as a fairly common occurrence in thalassemia of the homozygous or major type. In a series of 79 patients surveyed with the disorder, 14% showed premature physeal fusion almost always most marked in one growth plate (129). All instances, however, were noted in those greater than 10 years of age. When that age group alone was assessed, fully 23% of 48 patients demonstrated the physeal arrest phenomenon. The most common growth plate involved was that of the proximal humerus, with frequent occurrence in the distal femur and examples also in the proximal and distal tibia and fibula. The study by Currarino and Erlandson also noted that the premature fusion was almost always focal and generally was peripheral, such that any shortening present was usually complicated by angular deformity. The growth plate arrest in the proximal humerus was almost always seen medially, leading to varus tilt of the head in relation to the glenoid. The abnormality was not demonstrated in any of the 31 patients under 10 years of age. The incidence in males and females was approximately equal. Very little documentation exists concerning the extent of the discrepancy or the nature of the angular deformity and also little indication that surgical intervention was performed. No definitive cause of the disorder has been described.

S. Hemophilia Lower extremity length discrepancies can occur in hemophilia (412). The mechanism appears to be similar to that in juvenile rheumatoid arthritis in that a recurrent synovitis in a single joint, commonly the knee, leads to growth stimula-

SECTION VI ~ Lower ExtremiW Length Discrepancies in Specific Disease Entities

tion in the early years up to approximately 8 years of age, following which continuing synovitis tends to premature closure of the distal femoral and to a lesser extent proximal tibial physes. It is common for hemophilia to occur in a recurrent fashion in one joint, which is referred to as a target joint, with the three most common affected regions being the ankle, knee, and elbow. Caffey and Schlesinger (84) described overgrowth of the epiphyses themselves in all dimensions in joints with hemophilic arthropathy, but they did not study overall bone length. Kingma (274) described overgrowth of an extremity affected with recurrent hemarthrosis in one joint. Three patients, all less than 10 years of age, suffered repeat knee hemarthrosis and after straightening of flexion contractures were noted to be longer on the involved side. Each was still growing and report of the final discrepancy was not made. The overgrowth in an 11-year-old boy was 2.5 cm, in a 5-year-old boy it was 2 cm, and in a 7-year-old boy it was 2.5 cm. Overgrowth in hemophilia was also described by Harris (220), who reported on a 1-in. overgrowth on the involved side at 11 years of age, and by Heim et al. (227), who reported a 2.2-cm overgrowth on the side of the involved knee at 5 years of age. Length discrepancy is seen less frequently now because of improved medical control limiting hemarthrosis and synovitis. Overgrowth discrepancies were often hidden and minimized by flexion contractures and articular cartilage degeneration of involved joints.

T. Synovial Hemangioma of the Knee Joint Synovial hemangioma, a rare disorder, can occur with the most frequent site of involvement being the knee joint. Moon (333) performed a careful review of the literature in 1973 and documented 137 patients with synovial hemangioma of the knee. There was equal occurrence in males and females, with the initial occurrence of symptoms concentrated in the childhood years from birth to late adolescence. Approximately 75% of patients were symptomatic prior to age 16 years. In those reports that mentioned limb length, only 14 cases were described as having increased limb length on the involved side, with many showing only about 1 cm difference. In 8 cases the limbs were of equal length, and in 4 the involved limb was somewhat shorter. Limb length discrepancies thus were variable, although a slight majority of patients had slight overgrowth.

U. Legg-Calve-Perthes Disease Data for the extent of lower extremity length discrepancy in Legg-Perthes disease were reported in our study of a large group of patients treated with a unilateral abduction brace. The data show how both the disease and the mode of treatment used impact length differentials. In the 147 patients with a lower extremity length discrepancy associated with unilateral Legg-Calve-Perthes disease, the involved side

,~, 2 5

Y

o 2~) >0 z 1.5
651

'

0 CA SA

;

I

I

I

7 46

8 66

9 86

10 90

; 11

i 12

I 13

101~

124

140

I 14

15 157

FIGURE 14 TypeIV pattern in a patient with Legg-Perthes disease who maintainedproximal femoral growth for several years after the initial insult, but then the proximal femoral capital epiphysis prematurely fused, increasing the discrepancy.

was always shorter at some time during the period of assessment (430). The average maximum femoral discrepancy in this group of patients was 1.38 cm, none of whom had femoral or pelvic osteotomy. All five types of discrepancy pattern were seen. Femoral shortening occurs due to cessation of growth during the phase of necrosis of the secondary ossification center, due to subchondral collapse with the coxa plana deformity, and due to disuse in association with therapy, and it has long been recognized as part of the disease entity (185,422, 441). If proximal femoral varus osteotomy is performed, shortening will almost always be increased (30). Innominate osteotomy tends to increase length on the operated side by 1 cm (422). The femoral shortening was frequently associated with shortening of the ipsilateral tibia due to decreased use of the limb with unilateral brace therapy. The average maximum tibial discrepancy was 0.93 cm, and the average maximum combined lower extremity length discrepancy was 2.14 cm. Once bracing was discontinued, the tibial discrepancy decreased. Twenty-one patients demonstrated a type I pattem, 8 a type II, 52 a type III, 10 a type IV, and 49 a type V. When only the femoral lengths were assessed, 14 patients (10%) demonstrated a type IV pattern (Fig. 14). The occurrence of the type IV pattern was analogous to that seen in some patients with septic arthritis of the hip with only mild destruction. Premature fusion of the capital femoral epiphyseal growth plate occurred, with a late alteration of the femoral head-greater trochanter relationship. It is probable that the type IV pattern would have been seen more often, but the performance of distal femoral epiphyseal arrests in these patients made it difficult to document the type IV change. There was a good correlation between the age of the patient at disease onset and the final discrepancy pattern. The average age at onset in the patients who showed the type I pattern was 8.7 years, for type II it was 6.5 years, for type IV it was 5.6 years, and for type V it was 5.3 years. These numbers reflect the better healing that occurs in younger patients with Legg-Perthes disease, who

652

CHAPTER 8 ~ Lower Extremity Length Discrepancies

have a longer time available for the slow repair process. A more detailed presentation of lower extremity length discrepancies in Legg-Perthes disease can be found in Chapter 4.

V. Slipped Capital Femoral Epiphysis Clinically significant lower extremity length discrepancy is relatively infrequent in slipped capital femoral epiphysis. There are several reasons for this, the two most important being the relatively late age of occurrence in skeletal development and the fact that only 30% of femoral and 15% of lower extremity length are due to growth at the proximal femur. Even though treatment induces premature physeal fusion, there is insufficient growth remaining on the contralateral proximal femur to lead to a clinically significant discrepancy. Other limiting factors include the fairly high incidence of bilateral involvement and the slight discrepancy caused by the mild to moderate slippage itself. Concern about discrepancy is increased with unilateral disorders in those 11 years of age or younger particularly with moderate or severe displacement. This profile often is seen in the subgroup of slipped epiphysis associated with other medical disorders.

W. Limb Length Discrepancies Due to External Causes 1. BURNS Lower extremity length discrepancies can occur following severe third-degree burns particularly about the lower part of the leg and ankle joint. Four cases were reported by Frantz and Delgado (167). The growth slowdown was considered to be due primarily to scarring in the periphyseal region, which could serve as a mechanical tether to growth and also limit vascularization of the physeal area. The possibility of direct thermal damage to the epiphyseal plate was also raised. In two instances there was premature closure of the distal tibial physis, leading to a maximum discrepancy of 4.6 cm in one case and 4.7 cm in another as measured prior to contralateral epiphyseal arrest for correction of the discrepancy. In a third patient the discrepancy prior to epiphyseal arrest was 2.0 cm, although premature closure of the distal tibial plate did not develop and it was felt that periphyseal scar formation was limiting the growth. In a final patient, a discrepancy of 1.0 cm was noted with early central sclerosis of the distal tibial physis and the suspicion that growth would worsen with time because several years of growth still remained. Evans and Smith (153) studied several cases of burn patients, many in the childhood years, and described bone and joint changes. In spite of the fact that many children were assessed, no specific comment on physeal changes was made. They defined skeletal alterations as falling into three groups: (1) alterations limited to bone, with osteoporosis and periosteal new bone formation seen; (2) alterations involving pericapsular structures such as pericapsular calcification, osteophytes, and heterotopic para-articular ossification; and

(3) alterations involving the joint itself with progressive articular destruction and joint fusion. Tethering of the physis would clearly occur with alterations involving the pericapsular structures, and in those instances in which the articular cartilage is destroyed adjacent physeal cartilage would also have a high likelihood of destruction. 2. EPIPHYSEAL GROWTH ARREST SECONDARY TO FREEZING

Severe frostbite in childhood has been shown clearly to lead to epiphyseal damage with premature growth arrest following. The large majority of cases described involve the hands with the phalangeal epiphyses most commonly affected. Initial observation of this occurrence was made by Lohr (302), who demonstrated physeal closure in several phalanges beginning several months postexposure. Other instances of phalangeal destruction due to freezing were reported by Bennett and Blount (42), in which there was complete destruction of the epiphyses of the distal phalanges of the second to fifth digits and the middle phalanges of digits 2, 3, and 4. The physes in the less involved contralateral hand remained open. Thelander (475) also reported a case of unilateral frostbite in a 6.5-year-old boy, who demonstrated premature physeal closure at 9 years of age involving each of the epiphyses of the distal and middle phalanges from the second to the fifth digits. An extensive study by Bigelow and Ritchie (48) described 13 patients with frostbite of the hands during the childhood years, all of whom had loss of one or more epiphyses. The thumb was shown to be involved only rarely, and it was invariably the distal and then middle phalanges of the fingers that were affected first. In any instance in which the proximal phalangeal epiphysis was affected those distal to it were also affected. In most instances shortening occurred without angular deformity, but in some instances the growth arrest was partial and subsequent growth led to varus or valgus angulation. It is felt that the epiphyses are damaged both by the direct effect of the freezing itself and then by vascular changes secondary to frostbite including thrombosis.

X. Infantile Cortical Hyperostosis-Caffey's Disease In this disorder there is thickening of the periosteum and radiodensity of unknown etiology, but the bone involvement is selective and can lead to increased length on the involved side (83). In those instances in which the tibia and fibula are involved overgrowth has been described. Jackson and Lyne (249) reported an unusual complication of Caffey's disease, however, which involved shortening on the involved side secondary to direct involvement of the distal tibial and fibular epiphyses. Radiographs showed cortical hyperostosis of the entire length of the tibia and fibula on one side, as well as the left mandible. Synostoses eventually developed between

SECTION VII ~ Projection of Limb Length Discrepancies by the Time Skeletal Maturity Is Reached

the tibia and fibula proximally and distally. In addition, the distal tibial epiphysis was wedge-shaped. The fibula was shortened 1.1 cm relative to the opposite side and the tibia was shortened 0.7 cm. The authors felt that it was not the synostosis that caused the growth retardatiori but primarily the periosteal inflammation and scarfing in the epiphysealmetaphyseal regions that affected the vessels to the epiphyseal plates. In this report the patient was followed only to 4 years of age.

VII. P R O J E C T I O N O F L I M B L E N G T H D I S C R E P A N C I E S BY T H E T I M E S K E L E T A L M A T U R I T Y IS R E A C H E D The projection of limb length discrepancies that would be present at skeletal maturity became of practical importance with the demonstration by Phemister (387) that surgical arrest of a physis on the longer limb with growth remaining allowed limb length equalization as the shorter limb continued to grow. Accurate timing of an epiphyseal arrest procedure became feasible as documentation of the percentage of growth at each long bone epiphysis and the normal range of femoral and tibial bone lengths was obtained.

A. Percentages of Growth at Each End of Major Long Bones Digby (142) determined percentages of growth in human bones by measuring from the nutrient canal to either end taking the nutrient canal position as a standard and unchanging marker. He documented proximal femoral growth as contributing 31.2% of entire length and distal growth 68.9%. Other measurements included proximal tibia 57.1%, distal tibia 42.9%, proximal humerus 80.8%, distal humerus 19.2%, proximal radius 25%, distal radius 75%, proximal ulna 18.6%, and distal ulna 81.4%. Bisgard and Bisgard (50) performed similar studies in goats and showed remarkable similarity of growth patterns. By using the nutrient canal method, their findings for the proximal femur were 32.7%, distal femur 67.3%, proximal tibia 56%, distal tibia 44%, proximal humerus 81.7%, distal humerus 18.3%, proximal radius 25.5%, distal radius 74.5%, proximal ulna 19.6%, and distal ulna 84.4%. The rounding off of numbers widely accepted for clinical use is proximal femur 30% of growth, distal femur 70%, proximal tibia 55%, distal tibia 45%, proximal humerus 80%, distal humerus 20%, proximal radius 25%, distal radius 75%, proximal ulna 20%, and distal ulna 80%. In terms of overall lower extremity growth, general percentages are upper femoral epiphysis 15%, lower femoral epiphysis 35%, upper tibial epiphysis 30%, and lower tibial epiphysis 20%. Two major studies in the American literature have provided data that remain valuable in terms of overall lengths of femurs and tibias throughout the childhood years. These are the reports by Maresh (35) and

653

Anderson, Green, and Messner (23). Dimeglio and Bonnel (143) have assessed growth data to indicate mean amounts of growth remaining in males and females at distal femur and proximal tibia at varying ages. Pritchett (395) indicated that growth rates at each physis were not uniform throughout growth. Over the final 2-3 years, the proportion of growth at distal femur and proximal tibia increased. This appears to be due to earlier closure of proximal femur and distal tibial physes [see also reference (80) and Chapter 7, references (171) and (172)].

B. Systems for Projecting Limb Length Discrepancy at Skeletal Maturation Many systems for predicting limb length discrepancy at the time of skeletal maturity have been presented. Accurate knowledge of the amount of growth remaining in femoral and tibial physes at any particular age became of practical importance after Phemister demonstrated the relative ease and value of arresting physeal growth in the longer bone.

1. HATCHER Hatcher developed the initial approach for determining the best timing for epiphyseal arrest procedures in the distal end of the femur and proximal end of the tibia (87). Although the figures were based on averages at chronological age periods, they provided the first quantitative information on expected growth increments. Hatcher compiled data from detailed studies, which included knowledge of the percentages of growth of the bones of the lower extremities from the proximal and distal epiphyses as follows: femur, 23% proximal and 77% distal; tibia, 56% proximal and 44% distal. Growth percentage contributions of each of the four major lower extremity epiphyses to overall lower extremity length were then determined as proximal femur 12%, distal femur 38%, proximal tibia 29%, and distal tibia 21%. From available growth tables, Hatcher was able to calculate and chart the averages of normal growth to be expected from the femur and tibia for males and females from age 4 to 17 years. Although the length numbers provided were averages, they provided the initial information in terms of final projected lengths of the femur and tibia and the average increment per year noted. Hatcher's growth increment curve was published in the first edition of Campbell's Operative Orthopaedics in 1939 (87). With the three previously mentioned pieces of information one could make a relatively crude but effective estimation of the appropriate timing for any epiphyseal arrest.

2. WHITE White timed all epiphyseal arrests on the basis that the distal end of the femur grew 9.5 mm (0.38 in.) per year and the proximal ends of the tibia and fibula grew 6.4 mm (0.25 in.) per year (500). White and Warner (502) described their technique for epiphyseal arrest in 1938, slightly modifying Phemister's approach. They clearly credited him, however,

654

CHAPTER 8 9 Lower Extremity Length Discrepancies

with the original idea of epiphyseal arrest and confirmed its great value in treating lower extremity length discrepancies. They utilized a square mortising chisel, which, for example at the distal femur, was utilized in the midline medially and laterally. The square chisel was made to straddle the growth plate diagonally so that two of the diagonal points were forced into the epiphyseal line and two into bone on either side. The chisel was then driven into the femoral condyle to a depth of approximately 0.75 in., and the square of bone and interposed physeal cartilage tissue were removed in one piece. The metaphysis and physis were then further damaged by a curette, and the plug of bone and cartilage removed by the chisel was rotated through an angle of 90 ~ and replaced in the femur so that the portion containing the epiphyseal line was directed in the long axis of the femur and solid bone readily bridged across the physis. A light plaster cast was employed to immobilize the leg for 2 weeks, after which physiotherapy was begun with weight bearing allowed in 4 weeks. White and Warner reported successful transphyseal bone bridging bilaterally with no deformities produced in approximately 30 cases followed long enough to allow conclusions to be reached. White and Stubbins (501) reported on additional cases a few years later, again stressing the relative ease and effectiveness of the square chisel approach to epiphyseal arrest. As the operation came to be more widely used they stressed (1) the importance of recording lower extremity length discrepancies more accurately than by tape measurements and (2) the use of a simple method of calculation to project the appropriate timing for the intervention. In a short period of time, the epiphyseal arrest procedure dramatically limited use of the more involved femoral shortening procedure. They felt that the bone block removed by Phemister (387) did not, in many instances, include sufficient depth of bone and that the technique they had developed favored solid bony union because there was more extensive bone surface opposed. Phemister had mentioned chiseling out the epiphyseal plate to a depth of 1 cm anterior and posterior to the block of bone and cartilage removed, but apparently some surgeons were somewhat lax in doing this. With rotation of the square bone plug 90 ~ the contained epiphyseal cartilage was at a fight angle to the persisting physis; on either side of it bone tissue crossed the physis, and with healing metaphyseal bone is continuous with epiphyseal bone. White and colleagues began using standardized radiographic projections to document length discrepancy using a standard tube distance. A formula for calculating growth remaining in the femur and tibia was also established. Regardless of the age and size of the child, a growth arrest operation at the distal femoral epiphysis would retard growth at the rate of 0.38 in. per year, whereas at the proximal end of the tibia and fibula it would retard growth by 0.25 in. per year. They h "adaaccumulated ' a series of 202 separate growth arrests; of these57% were for poliomyelitis and 11.5% for osteomyelitis.

3. WILSON AND THOMPSON Wilson and Thompson (508) also derived a rough guide for timing epiphyseal arrest based on the expected amount of growth from each of the epiphyses. Several years of documented lower extremity lengths were needed for each individual. If a discrepancy had not been increasing with growth, the present discrepancy would equal the expected discrepancy. If the rate of growth in the longer leg had been greater than in the shorter extremity, a discrepancy would continue to increase proportionately. They then calculated the expected discrepancy, related it to the expected growth in the longer extremity, and chose the appropriate physis for ablation. Fourteen patients averaging 12.5 years of age at the time of surgery had been observed from 1 to 4 years postsurgery. Seven of these showed a lessening in discrepancy of 1 to 1.75 in., in 3 the decrease was 0.5 to 1 in., in 2 the difference between the two legs remained the same, whereas in 2 the discrepancy increased slightly. Even in those cases in which no limb length equalization occurred there was at least stabilization of any progressive discrepancy. 4. GILL AND ABBOTT Gill and Abbott (180) improved the accuracy of growth projections greatly by using percentile height tables and determinations of skeletal age. Their method took the individual's relative growth and maturation rate into consideration rather than using average values, as had been done previously, although the limb lengths were based on percent determinations from data for total body height. Gill and Abbott pointed out five major criticisms in relation to the methods used by Phemister, Hatcher, White and colleagues, and Wilson and Thompson. Their work began the era of accumulation and use of more accurate growth data. They pointed out the following: average figures for the length of legs should not be applied because of the wide variations in the final length of individual children; leg lengths determined as proportions of overall height were far less reliable than the radiographic measurements that they were beginning to use; no allowance had been made previously for variations in sexual and skeletal development in children of the same age; use of parents' measurements were not reliable in determining the future growth of the child; and the numbers used by Digby presupposed that the femur and tibia grew at the same relative rate during the entire period of growth. Many studies, even then, showed this latter point to be untrue, with some data showing that the tibia achieved its final growth before the femur. Gill and Abbott's method was based on three major principles: (1) the final stature of a child could be predicted by use of the percentile method; (2) the accuracy of growth projection would be increased if the bone maturation age of the child was considered; and (3) the relative proportions of the length of the femur and the tibia to overall stature were maintained with only small variations throughout the adolescent period. Among the principles of general and limb development that they utilized were the

SECTION VII 9 Projection of Limb Length Discrepancies by the Time Skeletal Maturity Is Reached

relationship of overall patient height to age and percentile, separate tables for males and females, and assessment of skeletal maturation using Todd's Atlas of Skeletal Maturation. Gill and Abbott began to utilize skeletal maturation ages for growth calculations if skeletal development was greater than 6 months advanced or retarded compared to the chronologic age. Length data were then placed in a specific percentile range to allow the final length to be determined relatively early. Femoral and tibial lengths were quantitated radiographically by using either teleoroentgenograms or scanograms. The expected final lengths of each bone were calculated. By subtracting the present length of the normal femur from its expected final length, the future expected growth of the bone was obtained. This was done separately for the femur and the tibia. It was then possible to determine growth from the distal part of the femur, which would be 70% of expected femoral growth. Similar approaches were used for the proximal tibia: the expected proximal tibial growth was 55% of expected. Gill and Abbott calculated expected final lengths of femur and tibia by relating them to projected adult stature because numbers with appropriate percentiles were available for overall stature at that time but not for femoral and tibial lengths. Direct femoral and tibial length measurements were eventually determined by Maresh (315) and by Anderson, Green, and Messner (23). The relationship of the femur and tibia to overall stature, however, was felt to be stable with growth. Based on a femoral radiograph and total body height measurement, Gill and Abbott were able to determine the femoral percentage of stature, which, for example, was frequently 28%. They then determined the patient's final overall stature from the charts and calculated the final femoral length as being 28% of that. The remaining expected growth of the femur and tibia could also be calculated, as could the growth at the proximal and distal ends of each of the major long bones. The accuracy of their predictions was dependent on the exactness with which height could be predicted by the percentile method. The method was felt to be valuable because any given child tended to maintain his or her rank in stature from one age to another. Much data also were presented to indicate that the relative proportions of femoral and tibial length to overall stature were maintained throughout growth with only small age variations. The percentage of growth from each end of the bone could then be calculated on the basis of the proximal femur 30%, distal femur 70%, proximal tibia 55%, and distal tibia 45% numbers. The growth remaining could be further calculated. Gill and Abbott showed clearly the reasonably high accuracy with which they could estimate expected growth of the normal limb. The next problem in terms of the treatment of lower extremity length discrepancies involved estimates of the expected growth of the abnormal limb. In experimental efforts to determine how much growth was occurring in physes that were open but growing at a slower than normal rate, for example, with poliomyelitis, Gill and Abbott inserted small metallic markers in the af-

6S~5

fected bones through a special hypodermic needle, although this method could not be used widely in a clinical setting. 5. GREEN AND ANDERSON

Shortly after Phemister demonstrated the value of epiphyseal arrest in the longer extremity to allow for relatively straightforward correction of lower extremity length discrepancies during the final years of skeletal growth, it became evident that more accurate growth data would greatly enhance the timing and thus the accuracy of the procedure. Green, Anderson, and Messner (22, 23) of Children's Hospital, Boston, played a major role in developing femoral and tibial growth data. They worked in conjunction with Stuart and Reed (465) and associates at the Harvard School of Public Health, who were performing a longitudinal series of child health and development studies from 1930 to 1956. The patients assessed involved 134 healthy children, 67 boys and 67 gifts, who were followed from birth to 18 years of age. Assessments began in 1930 when the first mother was enrolled in the study and continued until 1956 when the last child was discharged from assessment at 18 years of age. Part of the study involved a completely longitudinal series of radiographs of the lower extremities in the 67 boys and 67 girls who had annual lower extremity radiographs as part of the observations in the comprehensive longitudinal study program. Very early in the program the standardized lower extremity radiographs utilized the orthoroentgenographic technique, which ensured a high degree of accuracy with each assessment. Radiographs were performed once yearly. In 1964, Anderson, Messner, and Green (23) published their normal femoral and tibial growth charts. Separate assessments were made for the femur and the tibia in both boys and girls. Values at each year included the mean length of the bone, including the epiphyses, and publication of growth charts in which the mean value was indicated as well as those values 1 and 2 standard deviations above and below the mean. Values were listed from 1 to 18 years of age. These are printed as Figs. 15A and 15B. The socioeconomic status and national origin of the patients in this study were defined by Stuart and Reed (465). Enrollment was limited to white females of predominantly northern European stock, and all but 48 of the 592 parents of the children initially enrolled were born in the United States, with most of the others born in Ireland and brought to the United States early in life. The vast majority of children, therefore, were born in North America'and family origin was from North America, the British Isles or northern European countries. The predominant national origin of the families enrolled was Irish. The patient population was drawn from the clinics at Children's Hospital such that it tended to exclude those of high economic status. Additional growth data generated by the study allowed for charts to be constructed indicating recumbent and erect overall height in boys and gifts. Recumbent height was measured between 1 and 6 years of age and standing height bet.ween 6 and

F I G U R E 15 (A) Femoral and tibial length chart for girls is shown. (B) Femoral and tibial length chart for boys is shown. (C) The growth remaining chart documents distal femoral and proximal tibial growth remaining in relation to skeletal age for boys and girls. (D) Annual increments in length of femur and tibia in males and females are shown below compared to total height above. [Parts A and B reprinted from (23), parts C and D from (22), with permission.]

SECTION VII ~ Projection o f Limb Length Discrepancies by the Time Skeletal Maturity Is Reached

657

C GROWTH REMAINING IN NORMAL DISTAL FEMUR AND PROXIMAL TIBIA FOLLOWING CONSECUTIVE SKELETAL AGE LEVELS MEANS AND STANDARD DEVIATIONS DERIVED FROM LONGITUDINAL SERIES 50 GIRLS AND 50 BOYS KEYs

---

IEln|,.~,.

SKELETAL AGES FROM G R E U L I G H - PYLE ATLAS

7

I

6

!

2

16 SKELETAL AGE IN YEARS

15 SKELETAL AGE IN YEARS

50

GIRLS

=.,,,= STATURE ' - ' = ' FEMUR ---TIBIA

50

BOYS

/4\ .... o

~5

.....

lull ~

4'

i.U 13. I

-rI--.

, .............

\J

!

j .,,,

........

0 n,-

m -'qR,r

" .....

""

,,X,

,

AGI

FIGURE 15 (continued)

18 years of age. Charts incorporating the data outlined the mean values and also those 1 and 2 standard deviations above and below the mean. Sitting height charts were also produced for boys and girls. In a separate publication, Anderson, Green, and Messner derived charts indicating the growth remaining in the normal

distal femur and proximal tibia observed in a longitudinal series following given skeletal ages. The "growth remaining" data were derived from 100 children, 50 girls and 50 boys, measured at least once per year over the 8 years before growth terminated in their lower extremities. Of this number, 51 children were normal (25 girls and 26 boys) and 49 children

658

CHAPTER 8 9 Lower Extremity Lenttth Discrepancies

(25 girls and 24 boys) had poliomyelitis, which affected only one lower extremity with the opposite normal extremity used for data accumulation. The maturity of each child was evaluated from the radiographic appearance of the bones in the hand and wrist with the skeletal ages being read from the Greulich and Pyle atlas. The normal children were part of the longitudinal study from the Harvard School of Public Health, whose radiographic studies were also used to derive the normal femoral and tibial length charts. The normal and poliomyelitis groups were analyzed separately before the data were combined. No statistically significant differences were found between the two at any age either in the patterns of maturation or in the average amounts of growth that occurred after specific skeletal ages. The growth remaining charts indicate values for the distal femur and proximal tibia in boys and girls, with data for the girls beginning at 8 years of age and for the boys at 10 years of age. Values are listed as the mean amounts plus those in the first and second standard deviations above and below the mean (Fig. 15C). The Growth Study Unit at the Children's Hospital, Boston, was founded by Green in 1940, at which time the major cause of lower extremity length discrepancies was poliomyelitis. The unit was initiated with support from the National Foundation for Infantile Paralysis. Details relating to the use of the Green-Anderson method for projecting timing of epiphyseal arrest will be described in the next two sections. In a study of 125 epiphyseal arrest procedures in their unit followed to skeletal maturity, correction to within 1.2 cm (0.5 in.) of prediction occurred in 89% of cases (195). 6. MENELAUS Menelaus developed a simplified approach for projecting eventual length discrepancies based on chronological rather than skeletal age (325). It was assumed that the lower femoral epiphysis provides 0.38 in. and the upper tibial epiphysis 0.25 in. of growth each year. It was further assumed that growth of these epiphyses stops at the chronological age of 16 years in boys and 14 years in girls. Skeletal age is estimated primarily to allow for a calculation of the predicted adult height. Epiphyseal arrest is then resorted to only if this adult height will be acceptable. In addition, a marked difference between skeletal and chronological ages indicates that any calculation for the timing of the epiphyseal arrest is likely to be inaccurate. In a study of 44 patients having had 53 epiphyseal arrests who were followed to skeletal maturity, Menelaus (325) indicated that 52% were within 0.25 in. of calculated discrepancy and an additional 41% were within 0.75 in. Those with more than 0.75 in. error were only 7% of the operative cases. He felt that this approach compared quite favorably with the use of skeletal age as proposed by Green and Anderson. In their 1957 report of patients with poliomyelitis, 89% were within 0.5 in. of calculated discrepancy, whereas in Menelaus' series those with poliomyelitis showed an 85% effectiveness within the same 0.5 in. (195).

In a later study of 94 patients described by Westh and Menelaus (496), 85.1% were within 0.5 in. of that calculated and 94.1% of those had a final discrepancy within 0.75 in. of that calculated. The values are expressed in that fashion because equalization was not specifically sought in many of the patients who had poliomyelitis because the weakened limb was often deliberately left somewhat shorter than the stronger. 7. MOSELEY Moseley, using the Green-Anderson data, developed a straight line graph for leg length discrepancies by converting the normal growth curve for logarithmic methods into a straight line (340, 341). Moseley based his straight line graph method on two concepts not previously used in growth projections. The first was that the growth of the legs could be represented on a graph by straight lines and the second that a nomogram relating leg length to skeletal age could provide a mechanism for taking the child's growth percentile into account in predicting at what lengths the growth of the legs will stop. The first concept simply involved converting the exponential curve of normal growth to a straight line by plotting the data against a logarithmic scale. The Anderson, Messner, and Green data were used to construct the chart. Moseley pointed out several of the important consequences of this way of depicting the growth data: (1) the growth of the short leg was also represented by a straight line, which was positioned below that of the longer leg and tended to have a different slope; (2) the leg length discrepancy is represented by the vertical distance between the two lines; (3) the percentage inhibition of growth of the short leg is represented by the difference in slope of the two growth lines, designating the slope of the normal leg as 100%; (4) the growth of the leg that has undergone surgical lengthening thereafter follows a straight line of the same slope, which is displaced upward on the graph by an amount equal to the lengthening achieved; and (5) the length of a leg that has undergone epiphyseal arrest will follow a straight line of decreased slope in which the decrease in slope exactly equals the percentage contribution that the fused growth plate would otherwise have made to the total growth of the extremity. Because the contributions of the proximal tibial and distal femoral growth plates are 28% and 37%, respectively, of the total growth of the leg, it is possible to predict the amount of inhibition to be introduced by an epiphyseal arrest. The Moseley approach has been widely adopted and details of its use are available in each of the major general textbooks of pediatric orthopedic surgery.

8. HECI-IARDAND CARLIOZ Hechard and Carlioz (223) also developed a growth chart based on the data of Green, Anderson, and associates (Fig. 16). The limb lengths from 15 to 54 cm were listed on a single chart, as were bone ages. Because, in the vast majority of instances, normal growth persisted along its same percentile, projections were readily made. By plotting the length of the

SECTION VII ~ Projection o f Limb Length Discrepancies by the Time Skeletal M a t u r i t y Is Reached

DIAGRAMME DE CROISSANCEDES OS LONGS P. HECHARD d "aprL,s les donnees de GREENet ANDERSON NOM" ne (e) le :

9

Age osseux fille

7

;

.

8

.:

..-

....!

G:~.

.-".../

/ / . 5 C . / / . 5 . ~ . . / : / ...

/

J

J.

./

/f

//:G

~./C.:/.-.

//.,s~:5-~-~..~..,~,..-" '--~

Age osseux g arGon

FIGURE

16

'

9

9

-

320 /

--.,/.~-~,,-':/

/

~,.~:~'_.-.-

L./

,

~.~

~

_

~"~

38o

/

" 7JL~"

~S-'_P-.5<>:%--J--<--"P--'_ ~ - ' % - - ~ ~

> /

..

. . . . . . :.:~~-S.-"..-

//S///5-~

~ . . / .

i-"/
/f:... ,'::

.-

,,,1

/'//'/490

. . . /, / -/ .,r . !../, / .... :3-"//.// > ' / . . ....I';V'_-/ / . / : / , - " / /.!.-':/L/.......r./b,," / . / , / / / / / ~:)-~/." " .-"/".',/./,//:/3-" /

,

, ,

E E

.~

,,o

.2, 4oo

,,. , - ' 3 " /

I I

J

-'~

10

16~

The H e c h a r d - C a r l i o z growth chart depicts the length of

the lower extremity bones from 150 to 540 mm and allows for representation of boys and girls, with the bone age in girls extending from 5 to 14 years and in boys from 6 to 16 years. [Reprinted from Hechand and Carlioz (1978), Rev. Chir. Orthop. 64:81-87, 9 Masson Editeur, with permission.]

abnormal extremity with time, its pattern of change in relation to the normal side could be viewed and projections made. Hechard and Carlioz considered the method for projecting the inequality in length to be easily used as well as providing a document that could follow growth evolution. Before 4 or 5 years of age the rate of growth per year increased from year to year; between 4 and 13 years for girls and 5 and 14 years for boys the rate of growth annually was constant, whereas toward the approach of skeletal maturity the rate of growth diminished in a regular fashion. The majority of children examined for a lower extremity length discrepancy were seen during the period of linear increase in growth. By eliminating the extremes of age at either end, a linear depiction of growth was felt to be accurate. The values for bone ages and femoral-tibial lengths were listed for girls from 5 to 14 years of age and for boys from 6 to 16 years of age. 9. E A S T W O O D AND C O L E Eastwood and Cole (147) described a clinical method for

the graphic recording, analysis, and planning of lower extremity length discrepancies. Their chart lists length discrep-

659

ancy in centimeters along one axis and chronological age in years along the other (Fig. 17). The average maturity lines were marked for girls at 14 years and for boys at 16 years. Superimposed on the graphs are epiphysiodesis reference slopes (slopes 1-3), which converge to the skeletal maturity lines at zero leg length discrepancy. The slopes of these lines are based on the average annual growth of 0.6 cm from the proximal tibial growth plate and 1.0 cm from the distal femoral growth plate after the age of 8 years in girls and 10 years in boys. The graphs depict the estimated mature discrepancy and timing of surgery. The pattern of differential growth of the legs is determined from the graph such that the patterns defined by Shapiro are documented and then used to predict the pattern of further differential growth and eventual leg length discrepancy projected for skeletal maturity. The observed discrepancy line (line 2) is projected to the skeletal maturity line (line 3). The point of intersection (Y) gives the estimated mature discrepancy. The mature discrepancy line (line 4) is drawn horizontally from point Y and may intersect one or more of the epiphysiodesis reference slopes. The slopes are for proximal tibial arrest alone, distal femoral arrest alone, or a combination of the two. Vertical lines are dropped from these points of intersection to give the chronological ages for epiphysiodesis of the appropriate growth plates (X and X1) (Fig. 17). This method incorporates the different patterns of discrepancy into the plotting of the appropriate time for surgery without the need for specific calculation of the growth inhibition rate as is done in the Green-Anderson method.

C. Discussion of Methods The previous subsections 1-9 have shown the evolution of approaches to determining the expected discrepancy at skeletal maturity and the appropriate time for epiphyseal arrest. Additional charts are still being created. Pritchett and Bortel (Clin. Orthop. Rel. Res. 342:132-140, 1997) incorporated information on the late increased proportion of distal femoral and proximal tibial growth into straight line graphs. Paley et al. (J. Bone Joint Surg. 82A:1432-1446, 2000) developed a multiplier method for predicting limb-length discrepancy (in the type I developmental pattern) from 1-2 measurements. Femoral/tibial lengths (from existing databases) at skeletal maturity were divided by femoral/tibial lengths at each age for each percentile to obtain the multiplier. Limb length discrepancy and growth remaining values could be calculated. Femoral/tibial male/female multiplier charts were made from birth to 18/16 years for the mean and 1, 2 standard deviations above/below the mean. Gill and Abbott basically developed the concepts needed for accurate growth determination while Green and Anderson and colleagues provided the data needed to construct the appropriate growth charts for femoral and tibial lengths and for growth remaining. Both Green and Anderson and Menelaus derived relatively simple formulae to aid in determination of the

660

CHAPTER 8 9 Lower Extremity Len~trh Discrepancies H-H+++~!:: 11

I~J I[__j

~,~ E

~o 9

8

7 ~n

--

:

7

!::::: 1:7: :: :

: i :

LL_II: :::l ::: :: I !:

1 :::: [_:::i

::::I:]::

I i

:::!

:

iZ:l_1]:

:

I: 1 !~: ] :I:'-

!I _ 2 : : : :!!

i : : 2! I:: :; :~2 :

I

l___:! 22122I::2

:B 4

i:

:!7::12]~'

I I I 1

_: 2!

2 :22::1 :

!! I1'II:~11I/:III!!:!::[1 iI:I1~I!s ll.![:llll~.II~lllll:li:II

I::: I::. I.:: 1

11::l[ll:il[i;I

l.~i

!:7]lil!!!l]:il L .:l:I:~:I-:l:

!1 ~ : ! ; i : ,i!l [[~ l{l l ! i ! 721][

! ii!!!!!!~!

I lilll;lll}! I :illil:li~l 12:21 ]ili[l

'-

;

T~-'---'0

1

-

1

. . . . . .

2

i

.

:-'--~ 3

:l::I

!~:!:

:1::1!~I[1I ~I:::I!I111 :I:!I!!!111

l!~l:]l::i_il:~i::l:I::il:~I~:l:::::i-: IIIN .!1::1:i1~::::i:I:_I1:![::!:: I1! ~ ll::::jl:: :1:I::[I:I[::1

il:Ii:llLII

I111.X1:::.I1::.

:I:I1:11:II

l.I: I_!

!.~

1:1li

i1.11 91:1!

~

i!-;!::7

l:ll!lll~!:~!i 1~!1!!I\!:::I:::

,D.!!

::

:

_~.22221;2:.:I:72_:I

::~Jl:iSi:!li:i_:l:Sr :: :2

:I_:[21I[1121I, :I :!:!1:!1!,11

2-I -

11';

6

7

1

"

y

~

~

:

_Xl

:

"

9

' [7[

10

11'

"

";

-

;;

I

H\_I::7-"- ~ 11

12

13

14

!7: 1-

_ : : : 7 7 :

[ 2 I72 I72 r ! ~-: 7-: 175 172

: :

1 ~!I : :'':7::1

:

t !

I I

!I I

i I ! ! ] ! I i

I

I I

:

. . . . . .

r

"'-" ~ , ~ . 16

L ?

:::7:

:iI

15

_ 7

: -

-:-'I:LLL

' 1 i

:

I ['"

:: :I 7L:-"%I_'~N.._212:7:77222:1 : ;I -::IF'~"~X" -: --i::2:1:2:2i__~_'~I:: :~'

]7~I '-'L~

:~ :J 7 I :

:::7:::71 ::I7:::11

-",L_I'X:---I-?

- ' s 1

"__--.'-T'...;;:;;, 8

::::! " -

_ :

I'~

I:: ;7:

:: I,:: ::

-1 ~I

_

!

: :

:

1-~. lk._ .I _[ ~: 71:1__!I_[[ CI:2"%_:I: N ~II_ii":[:?

~ ::I~.'T'q-T,T~.."-TL~_.II2 ..... I -:~il ~--%t_ -

2Ii:1:1I:i1ZI~ii12s :I !i II 1171121Ii]1 :1 [i:I!:111'11]i1:!1

7I

-:1 ::1 :1 :1

~

!![!i~!-~;::!--:::!_a~.:!:.i:;!:

1

_

[!~I::

I ; [ ; ~ t - Ufl~ ~ :I:FT_TTI[ i :l:::'l .~a :_:1:::::g:

Illl:]l_iN::l:::_:ll:;:7ll::::l: III1:!1 i N::l:]:::ll?: ' ~ i ---2.-i.-~ .4,,po'4.,--

; ' ~ - - ~ _i:II~llffll"1.d;__l_I." ::12:1]!i11171117"~.12[IJ71 :::I2: - 1::::II:!tblI]!!---'%1_221

--'-" 7--'-- , 5

::::1

221 __Jill:iS I 7 .l.:i:J|iiillkr~L:__2.1::l 121 ..21_2_,~_~ _l:i! _ 72I:.2~:12!1! I::2 :ISSiS~lllihLl:N~_iI::i . L 22 J~'r._?l:[[iL[ ! 7:1:222:1:i11 I777 :lTili!l:lll. lll[~_Ili:: :'_s i 2:I:21::1::[~[I ::_:I:!i;IIilII'II:[:_N/12: ~'g,'~_tl iI2::2:~:2:7:!_~ :_1:2:2:I'::7~I 72 :I 11:111111'!I :?:'~1[11,,1~ ~L[ _xa 7 1 7 7 ; ' - 1 - - - 2 : 1 • :21:i:::l?:/]il ~lflna ~ [ A I H l ~ t S 2 ~:t :k 21.: llna ~ --1:: 721LS?:]I?]]]ili"~.v'vp ~ ~ k L ~ ~ • ~::,il :?k!l:~,u~ ~ _T'T.i:: :21::2_I ::.~1_:2 ' : ~ ~ i . 1-~]1!12~J21 :: ~1 21_:!~.3.~7.1:: 211.; 1 : Z : : l _ i : ~ ~ i : i i . l : k 1 :: P! :!:i[.::.:l::

-

~ 4

ZI

: : r ~

_.7

i

I:

::l:::ill~!ll I Z:lllll!l

~:::I~+:]!l!]]IIl:li::l:::!ll::l::l--

:1

:l::::;llllil~llllll

!11:1:112~1 L.~I:IIII~: ~1221 !;

I::::: ~ 222I 222212;IL1111I!Ll::LSLIi::!:l:]:2:l 1:2~2:I2712]1 ]:2:I2II111IIii11:2::!122!2717::2]I:S? |::IS:I:TS~I:II]IIIIIS]IllltlI:i:-'l::]:[I?:::!I::!

i:i!iil:ii::l

!l:I~:Illlllll!lllll:l:2;l_]7:II::I::I:-:''l:i-::l:

11 f l ]SI0~0

[ ..... 17722:l::-2]~,dl~2212212:S:Ll:S:S:L_i:]:l::2 1 ~ 1 2 : 2 2 ~ . ~ , , ~ _ _ ~ _ ' ~ 1 ~ ' ~ 1 ! : : ? 2 1 2 2 : : : I - 2 : 2 i I

=--:1

+ l'"''-t-t ...... ! ,~..~TTI:...'TY'L-~.?::S?| II " ;--i--'-1 }~'"-*~['~""-~-'-*-f'-"'~*

:: !~::~::[:rr[:l:::t~[-:_"'-71 :: ;I::!:;I:[I;:!::: :! :.:II[2i::I:i[[: 9 :---I:: 1_:ll::l~:l-!'I'I'-~'~.."7T~.I::

Ili:i:[llll~llilllIl:l~:iI

!_:[ 1 :r

!I :-! Z~

:: : : : : :

.....

;--I:IZ~I::

1 l

. . . . . . 1II ~ ~ 1:__I ---rl :lSi_:lI:!;!iltl]/;l!!~::l.::?[l~-.~l'~TT".i : 1 : ! : : : 1 2 2 : ! 1 1 : I I ~ : 1 : 1 ! : 2 I _ _ / _I12_:i12.~...~..~

::

:rr::

' . . . . ~. . . . . . . . . . IIXVVI-TLj_~II::;:~I : : : , ~ : : ~ : ] ~ : [ [ : 2 1 _ : ~ L : Z ~ | : !

~:[r l~:~t

::~I1::1:.1:::!

l:!]l!![l'l

I

::rx :

~

!::! I:[; !_1: I :~ 1. I !: '

:]::l:::::l:i]! l::ll:::::l~!li !:]:1::[:!1:1:1

ll[ll]illll

:m: II.l.:_,_:J,_.:i:il

:I:I !T[IITVIWHTEITI7 7I-7TITrT"~ 21"!!I.llllllll[lI!ll!:r:II::!::I:!_ :I lltlllllIlIlllliI:!:.; ~ N ~ : : -_ 2 : _ i l :~':. . ;. :. . i s 0p0 I~J~L~~H~.~"t-t...:II.::-;I

lllll:li]ll 1!Ili11]!11

!

r:

I_ : : I 7: i ,: i:;

11[!llLtlll !1111;11!11 :!l!l:!!~!l

112 2 1_:2 :I:.: l_:.]liIil:il ! S::11:2:221:::2 17:[I11111!;1 I ::::1::2211_:2: 17::iilqlll:l I:LS::I:::S!IL::: 1:2!111111!]I 1:277-I-722;1::: l::!:~l~lir;|]: [ 2: :I::_21 ::::l:• I:2~21 ]::212222:!]]]111][!]]I12 I 77 21 7:_:1 -~__12:21]12~111I:[ I :: 21 2ZS:l_-~T-.'l~i!:[!l=~:i~;i i .... :I !.12:!1]; 1:722:I

:111

-'[llIlI!7"~-~L"~-t-l:11l J : ~

l I

t,,,H-H,

I:::::I:,:II[.LLLLLil/j_.LLI:;i

:-1:11! ~ ill .... N I!II1111II::I:1!?:I:111 ~ ~ \ 1!l!!1111!::!:!:::I:I11 ~ ~ \ LLLII_LIII~::I:!:::IiiII ~ \ I-FTTq-FIIII:2 I];: !1LII: ~.l[IlllllII:!i!~::!li!11 " ~ l l ~ 1 ; l l l l I l T : l : : : : I I : l l l 1:: ! : :1 7 ] ~ : ] I : : : : I : : : :1:!!!

\

:

!

U o3

~

:4m

, ...................

LL%llll!::il!l]::l [TI:III:]::I171:::I!II! tlilll i:7~ -,~.~I11111!l-l

X

Tibia~\

1--7] !~] |'-"] E : [~..] [::

l]lLlilil:!ili::::l:lil

r .... r~mul

I=] [=1

I:::

I.....

........

] . . . . .

L'2-=] r e m u r / l l B l a %

17

-I 2 ~ |

]

i 18

Chronological age (years) F I G U R E 17 The clinical leg length discrepancy graph for boys derived by Eastwood and Cole is shown. Line 1 shows the skeletal maturity line, which for boys is 16 years of age and for girls is 14 years of age. Line 2 documents the observed discrepancy in chronological years. Line 3 is the projected discrepancy. Line 4 is the mature discrepancy line. It is drawn horizontally from point Y. Slope 1 is the femoral and tibial epiphyseal arrest reference slope, slope 2 is the femoral epiphyseal arrest reference slope, and slope 3 is the tibial epiphyseal arrest reference slope. Y is the estimated mature discrepancy, with X and X' representing the chronological ages for femoral or femoral and tibial epiphyseal arrests, respectively. [Reprinted from (147), with permission.]

appropriate time for epiphyseal arrest. Green and Anderson based their determinations on skeletal age, whereas Menelaus used chronological age for most patients. Green and Anderson were able to derive a method utilizing their growth charts. They incorporated the concept of a growth inhibition formula to aid in the ultimate timing for epiphyseal arrest. Growth inhibition was calculated as a formula: (growth of the long leg - growth of the short leg)/growth of the long leg. This formula enabled them to determine a rate of growth inhibition, which then served to indicate how much the discrepancy would increase over the remaining years of growth. For example, from any time period reference to the normal growth charts would indicate how much growth of the longer or normal leg was expected prior to skeletal maturity. Growth inhibition was then calculated from a time period of sufficient length to provide accurate data in that regard. The assumption was then made that the growth inhibition would be constant throughout the period of growth. It is evident that this is not true for all discrepancies. In practice, however, the Boston Children's Hospital Growth Study Unit frequently took a slowing of growth inhibition into account, although there was no specific formulaic method for this. If the growth inhibition was calculated as 0.4 and the future projected growth of the normal leg was 10 cm, then the

future increase in discrepancy was considered to be 10 • 0.4 or 4 cm. This would be added to the discrepancy at the time that the projection was made to yield the final discrepancy at skeletal maturity. The appropriate timing for epiphyseal arrest would then be determined from the growth remaining charts. Because the correction would be made by the shortened leg, the percentile along which the shortened leg was growing was determined from the normal femoral and tibial length charts. If this was one standard deviation below the mean, then that particular line was referred to on the growth remaining chart and the appropriate skeletal age to make up the specific discrepancy was decided upon. The conceptual changes introduced by the subsequent Moseley, Hechard and Carlioz, and Eastwood and Cole graphs primarily involved depiction of the growth data graphically such that growth inhibition would not have to be specifically calculated but was simply taken into account by plotting the growth of the normal and affected limbs on the chart, which allowed the discrepancy to be read directly. These methods, each of which has its advocates, are indeed simpler than calculation of the growth inhibition itself and are now widely in use. It has been recognized for some time that the length parameter of normal growth can be represented accurately by logarithmic plotting (242, 402). It is incorrect, however, to

SECTION VIII ~ Use of the Developmental Pattern Classification in Projecting Limb Length Discrepancies assume that pathological processes are as readily predictable by logarithmic plotting or any other formula. The straight line graph of Moseley (340, 341) or the growth inhibition method of Green-Anderson (195, 196) might well lead to inaccurate projections, particularly if the assessments are done too early in patients with conditions in which type II, late plateau type III, type IV, or type V patterns may be evolving. The types of pattern that occur in each disease category have been delineated. Knowledge of the developmental pattern classification, the natural and specific history of the condition causing a particular discrepancy, and the pattern type or types that occur in the condition allow the physician to project the ultimate extent of the discrepancy with clinically acceptable accuracy. The frequency with which clinical and radiographic evaluations of the discrepancy should be done must strike a balance. The evaluations should not be terminated too early or done too frequently on the expectation that straight line graph or growth inhibition projections always will suffice, and on the other hand they should not be done unnecessarily often as though the eventual outcome were totally in doubt. The developmental patterns themselves cannot be used to make accurate mathematical projections because growth, particularly during growth spurt periods and immediately prior to skeletal maturity, is not linear with time. The patterns do, however, permit accurate projections of discrepancy to be made using the femoral and tibial length charts and the femoral and tibial growth remaining, charts, which do take the nonlinearity of growth into consideration. The method of Eastwood and Cole (147) incorporates both the discrepancy pattern classification and the Menelaus method and appears attractive in that regard. The length and growth remaining charts were developed from information obtained by making yearly orthoradiographs of 67 boys and 67 girls between the ages of 1 and 18 years; they give the most accurate indication of individual bone lengths currently available. Their value lies in indicating the lengths of the femur and tibia and the growth remaining in those bones in relation to the standard deviation position. Smooth curves of growth are shown, with the individual growth spurt that occurs between the ages of 10 and 14 years blurred by averaged data. When an individual child's growth is plotted, the growth spurt will often change the standard deviation position of the limb lengths. If maturation is relatively early, the limb length will be on a higher percentile; if maturation is late, the limb length will be on a relatively lower percentile. Awareness of this factor is important in determining the amount of growth remaining in a bone and in projecting its final normal length. Growth is generally linear between the ages of 4 and 10 years, and if a child is on the first standard deviation above the mean at the age of 7, it is very likely that at skeletal maturity limb length will also rest along that percentile. Thus, length data obtained before the adolescent growth spurt are of great value in indicating what the child's projected mature level will be. If information is available only from the period between

661

10 and 14 years, however, awareness of the relationship of skeletal age to chronological age is important. If the skeletal age is retarded or advanced by 6 months or more in relation to chronological age, the correct growth percentile can best be determined by plotting the femoral and tibial lengths in relation to skeletal age, not chronological age. The Green-Anderson growth charts are derived from studies of white North American and northern European children during the time frame 1930-1956. Thus, they reflect the growth characteristics and height variations of that group, which would differ slightly from other racial groups and even from similar racial groups at differing time periods under altered socioeconomic climates. Even if the absolute height values between groups are slightly different, however, the pattern and percentile distribution would be unlikely to change at least in any meaningful clinical way. Because the values are read from the appropriate percentile and not simply determined as means or averages, placement of any individual on his or her percentile, even if this were somewhat higher than the percentile placements for other groups of relatively smaller stature, would still lead to the appropriate projections with time. Although it is unlikely that long-term, serial, longitudinal radiographic studies will be repeated, the ability to perform accurate imaging assessments without radiographic means, for example, by use of ultrasonography, should enable newer charts of differing racial groups and in differing socioeconomic settings to be established. The assessment of skeletal age is important in using the Green-Anderson method. Although a wide variation in skeletal age reading can be demonstrated among readers who do it infrequently, the assessments become highly reproducible when done by readers who do many. Although the Greulich and Pyle atlas (199) has certain limitations, it still serves as a clinically reliable guide to the rate of skeletal maturation. Management of the growing patient with a limb length discrepancy can be improved by knowledge of the classification of developmental patterns, the type or types of patterns that can occur with the particular disease process, radiographic documentation of the lengths of the lower extremities, a chart of the relationship between discrepancy and age to outline the developmental pattern that is evolving, the percentile standing of the normal limb and the abnormal limb, and the patient's skeletal age.

VIII. U S E O F T H E D E V E L O P M E N T A L PATTERN CLASSIFICATION IN P R O J E C T I N G L I M B LENGTH DISCREPANCIES A. Type I The type I discrepancy increases at a constant rate with time, as the rate of inhibition or stimulation remains uniform throughout the growth period (433). If one is certain that a

662

CHAPTER 8 9 Lower Extremity Length Discrepancies

type I pattern will evolve, one radiographic assessment of length, especially after the age of 2 years, will suffice for an accurate determination of the final discrepancy, although more determinations are always performed. In the first 2 years of life there can be considerable shifting of length between various percentiles, whereas afterward the distinct tendency is for normal growth to persist along the same percentile. For example, if at the age of 4 years the involved femur in a child with proximal femoral focal deficiency is 63% as long as the normal femur, one can project the final discrepancy by determining the length percentile on which the normal femur lies from the femoral and tibial length charts and noting the femoral length at maturity for that percentile. Sixty-three percent of the value represents the projected final length of the involved femur, and the difference between the two lengths represents the projected femoral length discrepancy. When the type I pattern is due to physeal destruction, the femoral and tibial growth remaining data can be localized accurately to the involved physis, and the values for the distal end of the femur and proximal end of the tibia can be read directly from the chart. If the proximal femoral physis has closed, the projected growth loss is determined on the basis that 30% of the remaining growth of the normal femur would occur at the proximal physis (and 70% at the distal physis). Similarly, if the distal tibial plate has fused, projected growth loss is determined on the basis that 43% of the remaining growth would occur at the distal tibia and 57% proximally. The amount of growth remaining in the entire femur or tibia is determined from the line that corresponds to the standard deviation position of the normal bone on the femoral and tibial length charts. Thirty percent of the difference between the present normal femoral length and the projected final length along the patient's percentile is the growth remaining in the normal proximal femoral physis. B. Type II This can be a difficult pattern to project because the discrepancy shows a decremental rate of increase, which varies from patient to patient and from condition to condition. The information available from the period of constant increase has no predictive value, as the discrepancy values themselves cannot "be aware" that a change in discrepancy pattern is about to occur. This group, therefore, requires especially careful monitoring. For example, at the age of 11 years, a child's femoral discrepancy measures 5.0 cm. Length on the short side is a cumulative 87% of normal. The growth percentile on which the normal femur lies allows one to project its final length. The growth rate in the most recent 6-month period, however, indicates that the short femur has shown 93% growth in relation to the normal side, thus demonstrating a deceleration in the development of discrepancy. The growth remaining in the normal femur is 8.6 cm, as indicated by the femoral and tibial length chart. A projection of the change in discrepancy with time indicates that growth on the

shorter side, based o.n the recent 6-month deceleration, would be at least 93% of 8.6 cm, such that the discrepancy would increase by only 7% of 8.6 cm or 0.6 cm, yielding a final maximum projection of 5.6 cm of discrepancy. If there is more time before surgical intervention, a further 6-month growth assessment might allow for an additional calculation. By this time, projections that allow for a clinically acceptable result (discrepancy of less than 1.0 cm) can be made. C. Type I I I Once a plateau has been reached, the lower extremity length discrepancy will not change throughout the remaining period of growth. The prototypical type III pattern is seen with overgrowth following fracture of a femoral diaphysis. The timing for the corrective physeal arrest is determined by using the femoral and tibial length charts and the femoral and tibial growth remaining charts. The final discrepancy is known once the plateau phenomenon has been documented to have occurred, as neither further stimulation nor inhibition will occur. D. Type IV Type IV discrepancies characteristically are seen after hip diseases in childhood that affect the proximal femoral capital epiphysis, such as septic arthritis of the hip with mild-tomoderate damage, Legg-Perthes disease, and avascular necrosis of the femoral head in association with treatment of congenital or developmental dysplasia of the hip. Premature closure of the proximal femoral capital epiphysis can occur after the discrepancy has remained in a plateau phase for as long as a decade. Radiographic indication of premature fusion of the proximal femoral capital epiphysis is demonstrated by a change in the relationship of the femoral head to the greater trochanter due to relative overgrowth of the latter. The growth discrepancy to be expected from premature fusion, once it has occurred, is obtained by determining the growth remaining in the entire normal femur, multiplying that value by 30% to give the amount of overgrowth expected from the proximal end of a normal femur, and, because growth is not occurring, adding this value to the preexisting discrepancy to give the projected final discrepancy. E. Type V If a discrepancy is beginning to correct itself, the growth charts are used to determine how much growth remains. A determination can then be made as to whether the spontaneous correction will be insufficient, result in equal limb lengths, or result in overcorrection. The type V pattern is seen characteristically with chronic inflammatory disorders not fully responsive to therapy, which stimulate growth under 10 years of age but lead to premature growth cessation toward the end of skeletal growth. The type V pattern is well-documented in juvenile rheumatoid arthritis and ap-

SECTION IX ~ Management o f Lower Extremity Length Discrepancies

pears to occur in many cases of hemophilia and tuberculosis for which therapy is less than fully effective. The developmental pattern classification provides a visual representation of the varying directional changes that can occur with time in lower extremity length discrepancies (Figure 3A). The dependence of the patterns on the causes of the discrepancies and on the time and anatomical locations of their occurrence is stressed. The demonstrated relationships between the pattern type and the particular disease entity (Fig. 3B) should aid in planning the nature and frequency of discrepancy assessments. In those conditions in which several pattern types occur, the classification serves mainly to point out that variability. Some of the factors contributing to the various patterns within each disease entity have been assessed further. The patterns alone do not provide for an accurate projection of a final discrepancy (except in type III) as growth, particularly during the adolescent growth spurt and immediately prior to skeletal maturity, is not linear with time. The patterns do, however, permit accurate projections of discrepancy to be made using the femoral-tibial length and growth remaining charts of Green and colleagues or the Eastwood-Cole chart.

IX. M A N A G M E N T O F L O W E R EXTREMITY LENGTH DISCREPANCIES

A. General Considerations As a general guideline, any discrepancy projected to be less than 2.0 cm at skeletal maturity should not require limb equalization; those discrepancies between 2.0 and 5.0 cm are usually treated with contralateral epiphyseal arrests to shorten the longer side. Discrepancies greater than 5.0 cm warrant consideration for ipsilateral lengthening, those beyond 8 cm often benefit from a combination of ipsilateral lengthening and contralateral shortening, and massive discrepancies in the 15-cm range or beyond might require prostheses with or without partial amputation. The aim of management is to ensure a discrepancy of less than 1.2 cm at skeletal maturity. This goal can be achieved in four basic ways: (1) by epiphyseal growth plate arrest in the longer limb at the appropriate time before skeletal maturity; (2) by metaphyseal or diaphyseal shortening, removing a segment of bone at skeletal maturity; (3) by lengthening the shorter extremity using metaphyseal or diaphyseal osteotomy and gradual distraction, transphyseal distraction, or transiliac osteotomy; and (4) by combinations of shortening and lengthening approaches in particularly difficult cases. For relatively massive discrepancies that leave the foot on the shortened side in the region of the midleg or knee of the longer side, prosthetic fitting and some or all of correction of angular deformity, joint stabilization, limb lengthening, distal limb rotationplasty, or amputation to maximize prosthetic fit may be required (Fig. 18).

663

Several excellent reviews of the lower extremity length discrepancy entity have been published and continue to warrant study for their insights (44, 120, 256, 260, 441).

B. Procedures to Shorten the Longer Limb 1. THERAPEUTIC ARREST OF GROWTH PLATE Therapeutic arrest of the growth plate requires knowledge of the amount of further growth to be expected in each of the growth plates at a particular age and an accurate projection of the expected discrepancy at skeletal maturity. The lengths of the femur and tibia have been documented radiographically and plotted in percentile charts showing standard deviations, and charts of the amount of growth remaining in these bones have been developed. The approximate contributions of each of the major long bone epiphyses to growth have been known for some time. In patients with a discrepancy in the length of the lower extremities, Green and colleagues (22, 193-196) have used an index of the rate of growth inhibition to project the final discrepancy. Moseley has developed a straight line graph for projecting length discrepancies by using logarithmic methods to convert the normal growth curve (340, 341). Both the growth inhibition and the straight line graph methods can lead to inaccurate projections in patients in whom the rate of change varies over time, if the assessments stop too early. The five patterns of discrepancy and their prevalence in each of the major conditions causing discrepancies in length have been delineated. Not all discrepancies increase at a constant rate, but the discrepancy at skeletal maturity can still be projected if the disease and the pattern of development are assessed carefully. The growth of the distal femoral, proximal tibial, and proximal fibular growth plates may be arrested when discrepancies are projected to be less than 5 cm at skeletal maturity. The function of the growth plate can be arrested surgically by inducing premature fusion between the epiphyseal bone of the secondary ossification center and the metaphysis; the length discrepancy is then corrected by continuing growth of the shorter side. Complete growth plate arrest of a normal physis can be performed if the contralateral affected epiphysis is still functioning. The procedure allows the shorter side to catch up in terms of growth. If the affected contralateral epiphysis no longer has any growth, the epiphyseal arrest prevents any discrepancy from worsening. Complete epiphyseal arrest is most commonly used in treatment of lower extremity length discrepancies. The timing of the procedure is crucial to its success but can be determined effectively using any of several prediction systems. The most common areas of performance of elective growth plate epiphysiodesis are at the distal femur and proximal tibia and fibula. In limb segments with two long bones (the leg and forearm), complete arrest of one growth plate often mandates arrest of the adjacent growth plate to prevent worsening of the deformity and to maintain articular alignment.

664

CHAPTER 8 ~ Lower Extremity Length Discrepancies

Management Guidelines for Lower Extremity Length Discrepancies A. 0-2 cm

No treatment; + small lift in shoe

B. 2-5 cm

Epiphyseal arrest Diaphyseal shortening (at skeletal maturity)

C. 5-15 cm

Lengthening

~

Shorten longer side Lengthen shorter side

Lengthening + contralateral -'~ epiphyseal arrest or ~ Combination diaphyseal shortening Approach D. 15 r (short sidef o o t

Prosthesis Possible

distal

amputation/

nearlongside knee) rotationplasty re prosthetic fit F I G U R E 18 Management approaches to discrepancies of increasing magnitude are shown. The management profiles A - D in each of the defined disorders leading to either stimulation or retardation of growth are shown in Table I.

Four technically effective ways of inducing premature epiphyseal arrest have been used. a. Phemister Technique. At open operation, a periphyseal rectangle of bone and cartilage is removed, rotated 180~ and replaced at both the medial and lateral sides of the involved bone end (387) (Fig. 19A). The rectangle of tissue removed involves metaphyseal bone, epiphyseal bone, and the intervening growth plate with two-thirds of the rectangle length being on the metaphyseal side and one-third on the epiphyseal side. The size of the segment removed and then repositioned varies depending on the size of the bone. Phemister defined a 3 cm • 1.5 cm • 1 cm block of tissue with curettage of the physis anterior and posterior to the block of tissue removed to a depth of 1 cm. Once removed it is rotated 180 ~ such that the larger metaphyseal fragment bone completely bridges the remaining epiphyseal growth plate. The medial and lateral transphyseal bone bridges stop growth as soon as bone repair occurs and their tethering effect enhances central physeal fusion. The White modification has been popular with many (Fig. 19B). b. Green-Phemister Technique. Green, in his modification of the Phemister approach, removed a larger and deeper block of bone and cartilage medially and laterally, obliterated the remaining growth plate with drills and a curette, and packed adjacent metaphyseal bone into the physeal defect (195, 196). The rectangle of bone removed was approximately 1.5 in. long (1 in. diaphyseal, 0.5 in. epiphyseal), 1 in. wide, and 0.75 in. deep. At the end of the procedure it was reversed 180~ as in the Phemister approach and replaced

into the defect, and the periosteum was resutured in place. Metaphyseal-epiphyseal bone fusions lead to immediate growth cessation. c. Blount Stapling Technique. B lount used three large metallic staples placed medially and laterally at the anterior, middle, and posterior aspects of the physes to halt growth (56-59). The principle involved is different from in the two techniques described previously; growth cessation is gradual because the physis must continue to grow until the prongs of the staple mechanically prevent further expansion. The original reason for this approach was to allow for subsequent removal of the staples and the resumption of growth if the timing of epiphyseal arrest proved to be too early, such that the discrepancy not only was eliminated but continuing growth from the shorter side was about to reverse the discrepancy. Unfortunately, this rationale was not always realized because after the staples had led to cessation of growth there was often no continuing growth of the physis once the staples were removed. Blount and Clarke (58) made their initial report on the control of bone growth by epiphyseal stapling and clearly laid out the principles of the approach. They pointed out the work of Haas (211,212), who had both proposed and demonstrated retardation of physeal growth by a circumferential wire loop, a discovery of "the principle of temporary arrest of epiphyseal growth." This approach utilized the principle of mechanical diminution of growth and was attractive to B lount because when Haas either removed the wire or the wire broke growth continued. Pressure inhibition of growth

SECTION IX ~ Management of Lower Extremity Length Discrepancies

665

Ci

curette

1

F I G U R E 19 Technical approaches to epiphyseal arrest are illustrated. (A) Drawing from Phemister's original work shows his outline of the reversed bone block technique. (B) The White modification of the Phemister technique is illustrated. At left the medial and lateral distal femoral and proximal tibial blocks to be removed are outlined. Once removed, the 0.5-in. square plugs containing epiphyseal bone, the epiphyseal growth plate cartilage, and metaphyseal bone are rotated 90 ~ and reinserted. Bone tissue now completely covers the physis and the bone bridge formed bilaterally tethers growth and leads to its cessation. The fibular block soon was recognized as being unnecessarily large, and for fibular arrest most now simply curette the physis, which also minimizes the chance of damage to the peroneal nerve. (C) Surgical approaches for the Blount stapling technique are shown in part (Ci). The medial approach is shown at left and the lateral approach at right. (Cii) The correct insertion for the distal femoral and proximal tibial medial staples is shown. Each prong is equidistant from the physis, and the alignment of the staple is parallel to that of the epiphyseal growth plate with the cross bar at right angles to the physeal cartilage and parallel to the bone surface. Three staples were placed medially and three laterally in each bone requiring arrest. [Part A reprinted from (387), Part B from (56), Part Ci from (58), and Part Cii from (56), with permission.]

had also been demonstrated experimentally by Arkin and Katz (26). Haas (211) had utilized a mechanical principle of limiting physeal growth by passing a wire around the epiphyseal plate with one transverse path across the metaphysis and the other across the secondary ossification center with the ends twisted together to provide a continuous loop. Several experiments in the dog were performed, each of which showed a definite loss in length growth of the bone. In some instances

the wire either broke or came loose at which time growth continued, indicating that physeal growth while restrained by the intact wire did not lose its full potential, which could be realized once the restraint was released. A few similar procedures were performed on patients with some definite evidence of growth retardation noted. In the human growth also continued after breakage of the wire. Haas (212) performed additional investigations in an effort to make the technique of clinical value. In a second series of studies

666

CHAPTER 8 9 Lower Extremity Length Discrepancies

staples were used instead of wire loops. The staples applied unilaterally across the physis arrested growth on the side of insertion and also restricted it on the opposite side to a lesser degree. In many instances there was evidence of complete cessation of physeal growth. Enormous forces generated by the growing physis were readily apparent because either a single staple or the wire loop of Haas often broke, allowing growth to continue. Even when two staples were placed there was often separation of the tips or widening of the staples, again indicating the powerful forces of growth not fully controlled by the staple. Blount and Zeier (59) pointed to the work of Strobino and Colonna showing that a force greater than 120 lb was needed to halt proximal tibial growth in a calf. The routine use of three staples on the medial and lateral sides of the physis was then adopted for clinical cases and the procedure was proposed for distal femoral and proximal tibial growth arrests, which during that era generally were for poliomyelitis. The procedure, which appeared technically quite simple, was adopted widely with relatively less consideration for timing because it then was accepted that any imperfect timing could be remedied simply by removing the staples, at which time growth would resume. In a second report, Blount and Zeier (59) reviewed 117 staplings noting few complications. They concluded that staples could be left in place for at least 2 years and still removed with the expectation that growth would be resumed. After the removal of staples there was usually a local growth spurt lasting a few months. Frantz (166) reviewed 10 clinical papers summarizing the first two decades of work within the orthopedic community with this technique. Benefits and drawbacks became more clearly defined. One of the problems, which was basically present in any epiphyseal arrest operation, was that of timing. Green and Anderson reported on both formal epiphyseal arrest and stapling and felt that both procedures were satisfactory with the incidence of complications being relatively insignificant, although they used the stapling for definitive growth cessation. The complications reported, which tended to appear early in any series, included buffed staples, metal reaction, overcorrection, premature physeal closure, peroneal palsy, knee joint laxity, misplaced staples, fractured staples, extrusion of staples, angular deformity, infection, genu valgum, and false aneurysm. It was widely agreed that stapling was not warranted under the age of 8 and preferably 9 years. In an experimental series of studies, Heikel (226) demonstrated that epiphyseodesis of the proximal tibial plate had no effect on subsequent longitudinal growth distally. Siffert (438) performed asymmetric stapling of the distal femoral epiphysis in rabbits and noted production of the varus deformity with gradual histologic thinning of the physis and eventual transphyseal bone arrest. Goff (186) studied 120 biopsies of children at various stages of growth deceleration and arrest following stapling. He observed that the direct compression by staples served to inhibit the proliferation stage of the physis. The thinness of the disk increased

with time. The most sensitive signs of diminished growth were increased degeneration and shortening of the hypertrophic region. Eventually all of the hypertrophic cells disappeared and new bone formation crossed from metaphysis to epiphysis. Bone bridge formation was present invariably after 4 years or 48 months, although the markedly abnormal structure prior to that time would appear to have had little potential for regrowth. Bylander et aL (80) studied growth of the physeal regions following stapling using their highly accurate radiographic stereophotogrammetric analysis. There was a uniform pattern of growth retardation following stapling, which lasted over a period of several months to years. This pattern indeed was evidence of the applicability of the original theory of stapling because it indicated slowing of physeal growth to basal levels rather than complete cessation of growth. Bylander et al. calculated that growth at the distal femur and proximal tibia in human patients continued at a low basal level of about 5-10 Ixm per day, particularly when stapling was performed at younger skeletal ages. Blount (57) later pointed out the need for precise placement of the staples, feeling that many of the imperfect results reported were due to improper timing or less than ideal technique. The upper and lower prongs of the staple were to be equidistant from the physis and the cross member was to be perpendicular to the growth plate and parallel to the surface of the bone into which it was being driven (Figs. 19Ci and 19Cii). The staples were to be angled such that the tips of the prongs pointed toward the central axis of the distal femur or proximal tibia. Stapling was best performed after the patient had reached the skeletal age of 8 years. It was inappropriate to perform the stapling procedure according to a schedule set up for epiphyseal arrest because the stapling procedure was based on a different principle, namely, allowing for correction with growth but not being designed to cause complete cessation of growth. Staplings were thus performed earlier than epiphyseal arrests. It was important not to bury the staples under the periosteum because they would be difficult to find at the time of removal and more likely to cause periosteal new bone formation and growth plate bridging. Blount indicated that "stapling of an epiphysis retards growth 80-90% for the next few years. It causes a temporary growth spurt at the other end of the bone. At the stapled epiphysis elongation is decelerated only 50% during the first six months. Some growth continues until a year or less before the normal time for epiphyseal closure." Many continue to use this technique because it is an effective, accurate, and relatively simple way of causing the cessation of growth if the staples are not removed prior to skeletal maturity. Sengupta and Gupta (429) reported on the value of epiphyseal stapling. They used two staples on each side of the distal femur effectively in the large majority of cases, rarely resorting to three per side. The two staples on each side were equidistant from each other with their tips pointing toward the center of the physis. Seventy-one percent

SECTION IX ~ Management of Lower Extremity Length Discrepancies

of the 503 procedures led to a discrepancy of 0.5-1.0 cm of shortening at the end of growth with only 3% showing more than 2 cm of shortening. The staples should be removed at skeletal maturity. Stapling increasingly is used in situations in which only a medial or lateral growth arrest is desired to allow for the correction of angular deformity without performing an osteotomy. d. Percutaneous Technique. In this procedure, growth plate obliteration is performed through a small incision with physeal visualization by fluoroscopic image intensification (65, 92). Lateral and medial incisions are made immediately over the physis to be ablated, but for some surgeons the approach is only from one side. The soft tissues are dissected down to the physeal region at which time a guide wire and cannulated 4 - 6 mm wide drill bit is inserted. Under radiographic control, drilling is performed across the physeal region anteriorly, at the midline, and posteriorly. This serves both to destroy the cartilage and to allow for communication between epiphyseal and metaphyseal vessels, thus leading to transphyseal bone bridge formation. The postoperative scars are smaller and rehabilitation is quicker than in the previous physeal ablation techniques. The procedure was reported by Bowen and Johnson (65) in 1984 and by Canale, Russell, and Holcomb (92) in 1986. Ogilvie (351) provided an experimental report in 1986 and a clinical assessment (352) in 1990. Each group reports some differences in technique, although the principles of small incision, percutaneous surgery, and growth plate obliteration under fluoroscopic control are common to all. The physis has been damaged with the use of a drill, drill and curette, drill and burr, or osteotome and curette. Ogilvie (351) has demonstrated well the importance of several passes through the physeal plate in a fanlike pattern to assure complete obliteration. Excellent long-term results have been reported both by the original authors and by others subsequently adopting the technique. In the original report of Bowen and Johnson (65), the resuits of 12 percutaneous epiphyseal arrests noted no complications. In their later series, complications were uncommon. There were no fractures, no neural or vascular complications, and no angular deformity. Canale et al. (92), in their initial report on 13 percutaneous epiphysiodesis operations, reported that all growth plates appeared to be fused with no major complications and no clinical evidence of subsequent angular deformity. A later report by Canale and Christian (93) on 22 children with percutaneous epiphyseal arrest noted that arrest was achieved in all with no patient developing angular deformity. Ogilvie and King (352) in 7 epiphyseal arrests reported no failures of fusion, postoperative infections, restricted joint motion, or angular deformities. Horton and Olney (240) reported 42 percutaneous epiphyseal arrest procedures in which all patients achieved physeal arrest radiographically and clinically and no patient developed angular deformity from an incomplete arrest. There were no neurovascular complications or fractures. Timperlake et al.

667

(476), in a detailed study from the DuPont Institute, reported on 53 consecutive percutaneous epiphysiodeses. In their procedure, the medial and lateral two-thirds of the growth plate are ablated, but the central one-third is preserved for stability. They approach the physis from both sides, use a 3-mmwide osteotome driven 1 cm into the growth plate and rotated 180~ to create a hole in the cortex, and then use a 3-mm-wide oval curette that is swept across the growth plate to ablate it. A report by Gabriel et al. (171) recorded the results of percutaneous epiphyseal arrest in 56 physes using a cannulated 10-mm drill bit over a guide pin followed by curettage. There were no severe complications of angular deformity, deep infections, or neurovascular problems. 2. RESULTS OF TIMING DECISIONS FOR EPIPHYSIODESIS

The timing of epiphyseal arrest is still an imperfect science, and studies continue to appear indicating a range of discrepancies at skeletal maturity from fully acceptable to amounts still leaving lower extremity length discrepancy out of the desired therapeutic range. It is not essential in a clinical sense for limb lengths to be equalized because a fully acceptable result by all the criteria is to move the discrepancy to a value under 0.5 in., which is generally reported in papers as ranging between 1.0 and 1.2 cm. Even those patients with a discrepancy moved to within 0.75 in. or less than 2.0 cm would appear to have few long-term problems because of the slight difference. Because many factors go into an accurate projection of the timing for epiphyseal arrest, the better results appear to be produced by those centers at which large numbers of procedures are performed and at which a small group of individuals or a specific unit is responsible for timing the procedures. Relatively few problems are reported with the various surgical techniques used to cause the epiphyseal arrest. Two problems characterize the studies reporting less than perfect results. One is the widespread recognition of the relative inaccuracy of the skeletal maturation tables used as a key indicator of timing in most systems. The Greulich and Pyle atlas using posteroanterior radiographs of the left wrist and hand remains the most widely used indicator. There is still considerable subjectivity, however, with this system and, in addition, the radiographs used are spaced at either 6-month or 1-year intervals, which represents a considerable margin of difference between age gradings. The other problem appears to be referable to the nature of the predominant disorders being studied in any particular series. Very few studies take into account the differing developmental patterns of the lower extremity length discrepancies that have been described. In those series in which type I and type II patterns predominate, there is relatively little problem with utilizing the more straightforward projections of the Green-Anderson, Moseley, or Menelaus methods. Where, however, there are relatively large numbers of type III, IV, and V discrepancies, the failure to recognize these patterns can further worsen the

668

CHAPTER 8 ~

Lower Extremity Length Discrepancies

accuracy of timing. Studies reporting on timing in epiphyseal arrest procedures should be read with these considerations in mind. A study by Blair et al. (54) reviewed retrospectively 67 distal femoral and proximal tibial epiphyseal arrests performed over a 14-year period. Only 22 patients had a final discrepancy of less than 1 cm. Setting aside 10 of 45 failures due to inadequate surgical technique, the remaining 35 failures were secondary to errors in timing. This report utilized the Green-Anderson growth predication tables. Porat et al. (389), on the other hand, reported good results in 90% of their patients, although the series was small involving only 20 children. In 5 children with anisomelia whose expected discrepancy was 4.5 cm, results at maturity show an average discrepancy of 0.7 cm. In 10 girls with lower extremity length discrepancies caused by ischemic necrosis with congenital dislocation of the hip, the average discrepancy at maturity was 0.6 cm with the expected nontreated value 4.0 cm. In 5 children with the discrepancy caused by infection, the average discrepancy was 3.8 cm at the time of epiphyseal arrest, whereas at maturity it had diminished to 0.5 cm. This group utilized the Moseley straight line graph, CT scanograms for length determination and the percutaneous epiphyseal arrest procedure. Lampe et al. (282) performed a prospective study in 30 children who underwent 33 epiphyseal arrest procedures using the Moseley straight line graph method. The mean predicted length discrepancy was 5.2 cm and the mean discrepancy at the end of growth was 1.4 cm with a range from 0 to 4.3 cm. In 9 patients out of the 30 (30%) the final length discrepancy exceeded 1.5 cm. They felt that the altered skeletal maturation was most problematic in those cases in which the projection was inaccurate. Variation in radiographic determination of skeletal age was also a source of error as noted by Cundy et al. (128). Little et al. (301) reviewed 71 epiphyseal arrest procedures in effort to compare the Green-Anderson, Moseley, and Menelaus methods of projection. They felt that each of the different methods did not have a meaningful superiority in projecting the end result and that all had somewhat limited accuracy. They thus advocated the Menelaus method due to its simplicity. The mean preoperative discrepancy in their series was 3.12 cm (range = 1.4-7.4 cm) and the mean discrepancy at follow-up was 1.05 cm (range - - 2 to 4.4 cm). Little et al. determined that 34% of the patients (24) had discrepancies greater than 1.5 cm at skeletal maturity and 27% (19 patients) had discrepancies greater than 2.0 cm. This study was quite detailed in that eight different methods were assessed involving four variations of the Anderson and Green technique and two each of the Menelaus and Moseley techniques. They concluded that it was the inherent variability of the individuals requiring epiphyseal arrest that prevented the methods from predicting the outcome more satisfactorily. Little et al. (301) also felt that it was the inability to project the date of skeletal maturity that played the major role in the imperfect results. The group also felt that

CT documentation of the length discrepancies provided the most accurate determination with the least amount of radiation. In reality there should be relatively little difference between the Green-Anderson and Moseley approaches, and even that of Menelaus, because each of these approaches uses the same source of growth data, this being the GreenAnderson femoral and tibial length charts and growth remaining charts. The Hechard-Carlioz (223) chart is also a differing method of expression of the Green-Anderson data. 3. PARTIAL THERAPEUTIC GROWTH PLATE ARREST There are two treatment situations that can lead to a recommendation for the performance of an asymmetric or partial growth plate arrest. a. To Complete an Already Existing Focal Arrest to Prevent Further Angular Deformity. Partial arrest of an affected epiphysis can be performed to terminate all growth in a growth plate that has suffered a focal or partial arrest that is considered to be too extensive for bone bridge resection. By completing the arrest across the entire width of the growth plate, development or worsening of the angular deformity is prevented. Projection of the amount of growth remaining in the opposite physis should be made. If the patient is near skeletal maturation, no additional measures are needed. If further growth is in the 2- to 5-cm range, contralateral epiphyseal arrest is warranted to prevent an invariable discrepancy from developing. If the growth remaining is greater than 5 cm, the need for ipsilateral lengthening should be discussed. b. To Treat Angular Deformity without the Need f o r Osteotomy. The second reason for the performance of an asymmetric or partial growth plate arrest relates more to treatment of the angular deformity than it does to treatment of shortening. It was recognized early in this century that the creation of an asymmetric growth plate arrest would still allow the remaining physis to continue to function such that it could be used to correct the angular deformity without the need for complete osteotomy. The earliest proponents of widespread use of this treatment were Blount (57-59) and colleagues, who utilized stapling on one side of an epiphysis to correct angular deformity. The technique has its major application at the distal femur or proximal tibia for angular deformities centered at the knee. Blount and colleagues (520) reported on 82 knees treated with epiphyseal stapling over a 20-year period and followed to skeletal maturity. The deformities were allowed to overcorrect before the staples were removed and an effective rebound phenomenon occurred in 22 patients with 35 deformities. In older children the staples were removed when the legs looked straight. Blount felt that exaggerated physiological deformities may correct spontaneously and should not be stapled before the skeletal age of 11 years in girls and 12 years in boys. Deformities secondary to specific disease processes could be corrected earlier although rarely below 8 years of age. B lount et al. concluded that results were satisfactory or improved 87% of the deformities corrected. Two staples were used

SECTION IX

9

Management of Lower Extremity Length Discrepancies

in most patients. There were 64 valgus deformities and 18 varus deformities. The two largest groups of patients were idiopathic-physiologic or secondary to poliomyelitis, with other scattered disorders involving hemihypertrophy, rickets, and occasional skeletal dysplasias. Two features are associated with removal of the staples when growth is still active. The group generally allowed for some overcorrection, counting on the rebound growth phenomenon to occur in which the stapled side of the bone elongated more rapidly than the other for a few months after staple removal. Following that the rate of growth tended to remain equal and then the physis closed 4 - 6 months prematurely on the stapled side. It is evident that a considerable amount of personal experience and judgment go into the timing of such procedures. Bowen e t al. (66) established a chart in an effort to project the appropriate time for the correction of any particular degree of angular deformity (Fig. 20). Rather than relying on stapling with removal of the staple after a certain degree of overcorrection, they performed a bony epiphysiodesis asymmetrically based on a timing chart. Their operation was performed for idiopathic or physiologic genu valgum or varum in the adolescent patient. With an asymmetric stapling, for example, on the medial side of the distal femur, continued growth from the lateral physis would be expected. The growth, however, would not be linear but would represent an arc of a circle with the radius equal to the width of the bone measured at the physis. The arc of continued growth relates to the angle of deformity as the circumference of the circle relates to the total number of degrees in the circle. With the use of a specific formula a chart was constructed to relate the amount of growth remaining to the angular change for varying physeal distances. This information was then combined with the Green-Anderson growth remaining chart, which allowed the angular deformity to be related to the linear growth remaining for the patient's skeletal age. An early study of 13 extremities treated surgically indicated that, as a general rule, following partial tibial epiphyseal arrest 5 ~ of angular correction could be expected for each year of remaining growth and following partial distal femoral epiphyseal arrest 7 ~ of correction could be expected for each year of remaining growth. The mean ages at surgery were quite similar to those recommended by B lount and his group. In the 7 patients having procedures on 13 extremities, the average chronological age was 12 years 8 months with the girls averaging approximately 12.5 years of age and the boys slightly more than 13 years. The average preoperative deformity was 11.6 ~ of femoral-tibial valgus and the average deformity at follow-up was 6.6 ~.

4. METAPHYsEAL SHORTENING OSTEOTOMIES Wagner (492) has pointed out the value of correcting a longer femur or tibia, which also has angular or rotational deformity, by performing a metaphyseal shortening osteotomy. The most common site is the proximal femoral metaphysis followed by the distal femoral metaphysis. He stressed main-

GIRLS

9 ~ ~ -M,,a,, t 10" ~ a n t 2O"

DISTAL FEMUR

Greul,ch-Pykl A l l .

Skeletal A p t From

669

BOYS DISTAL FEMUR

CM 8 7 6 5

CM

FIGURE 20 Asymmetricstapling is performedto allow for the correction of angular deformity without need for osteotomy. A chart developed by Bowen et al. estimates the correct time of asymmetricepiphyseal arrest in relation to the patient's skeletal age and the degree of angular deformity correction sought. [ReprintedfromDePalmaand Cotler(1956),Clin. Orthop. Rel. Res. 8:163-190, 9 LippincottWilliams& Wilkins, with permission.]

tenance of a medial cortical strut in continuity with the lesser trochanter to allow for stabilization and a greater area of bone repair at the osteotomy site. An AO large fragment blade plate is used for stabilization. Shortening osteotomies of the distal femoral metaphysis are carried out in a similar fashion. A strut of medial cortical bone in the metaphyseal region is also left intact. The distal femoral metaphyseal shortening osteotomy is often complicated by relatively slow rehabilitation due to periarticular scarring and knee joint stiffness and due to disruption of the quadriceps mechanism. That site is chosen, therefore, only if the angular bone deformity necessitates correction there. Bianco (47) has reviewed shortening and angular correction procedures in the femur. Metaphyseal shortening osteotomy of the tibia and fibula is recommended only if angular deformity necessitates such a procedure. The tibia should rarely be shortened by more than 4 cm. The advantage of the metaphyseal site for shortening is the rapid bony consolidation. If metallic fixation is used, closure can be difficult and complications can occur in relation to the peroneal nerve with a relatively marked shift in length and alignment. If possible diaphyseal procedures are recommended when tibial shortening appears mandatory.

5. DIAPHYSEAL SHORTENING (RESECTION) The diaphysis can be shortened at skeletal maturity to make the final length of the limbs the same. The exact

670

CHAPTER 8 9 Lower Extremity Length Discrepancies

amount of bone required is removed and the bone is stabilized with a metal plate, screws, or an intramedullary rod. When performed at skeletal maturity there is no concern about overgrowth following surgery and no need to project the final discrepancy. This approach is recommended primarily for discrepancies between 2.5 and 5 cm that were not detected by or persist at maturity. Shortening of the bones of the longer leg to achieve length equality is a much older and simpler procedure than elongation of the shorter leg. Blount and Zeier (59) have credited Rizzoli of Italy with the earliest descriptions of this procedure. White (499) and Wilson and Thompson (508) reported early attempts from the late nineteenth and early twentieth centuries, which generally involved fracturing of the shaft of the femur or open oblique osteotomy, allowing the fragments to overlap the necessary amount, and then treating the limb until healing occurred. The first formal shortening of the tibia and fibula was reported by Brooke (75) in 1927 in which segments of bone, 1 and 2 in. long, respectively, were removed and then applied as grafts to enhance healing. The femur could also be shortened by a stepcut method followed by internal fixation. The first large series of cases was reported by Camera (85), who resected a portion of the femoral shaft and then used the resected bone as an intramedullary graft. He reported on 32 cases with the average period of external fixation being 50 days. Moore (334) reported on 13 femoral shortenings with bone resection with the average amount of shortening obtained 2.5 in. and the average period for complete union being 2.5 months. White (499) reported on 45 cases of femoral shortening with a transverse osteotomy of the shaft, overlapping of the two fragments by the necessary amount, and fixation by long metal pins, which were then incorporated into a hip spica cast. Treatment then followed with casting and eventual bracing. Many patients were under age 14 years; in younger children he shortened 0.5 in. more than necessary to compensate for the overgrowth phenomenon. The quadriceps muscle always maintained its ability to fully extend the knee with no loss of strength detected. The average amount of shortening obtained was 2.5 in. with a range between 2 and 3.12 in. Wilson and Thompson (508) reported 5 femoral shortening operations using the White technique with an average shortening of 2.12 in. obtained. They reviewed the major series of lower extremity shortening procedures. In the works reported previously plus their own there were 98 femoral and 2 tibial shortening procedures for the 100 cases. The only 2 tibial procedures were those reported by Brooke. The average shortening obtained was 2 in. and the complications reported were relatively small, involving separation of fragments necessitating reoperation in 2 cases, infections in 7, delayed union in 2, and no instances of nonunion or angular deformity. In comparison to the results reported during that era for limb lengthening the relative simplicity of the procedure was clearly shown.

The large majority of lower extremity diaphyseal shortening procedures are still performed in the mid-diaphyseal or proximal subtrochanteric region of the femur where negative postoperative sequelae are less marked. There have been reports of the inadvisability of performing shortening of the tibia and fibula due to increased problems with nerve or vessel kinking or muscle control at the foot and ankle region afterward. The muscle compartments are much tighter than those of the thigh and they have more structures passing through them in a smaller area. As a result, the lag effect and the compressive effect of vascular disruption are greater in the leg. Some reports of leg shortening with good results are described next, however. An effective approach is to shorten the femoral diaphysis by resection, following which an intramedullary rod is placed to allow for rigid stabilization and immediate weight bearing (Fig. 2). Our preference is to perform open shortening of the diaphysis to remove the diaphyseal segment of the bone. There are also reports of closed shortening in which the osteotomy site is not opened so as to limit the incidence of infection and perhaps hasten the rate of healing. The operative success rate is improved with the use of an AO universal intramedullary nail locked statically at both proximal and distal sites (269). Excellent results can be achieved if the complications of malrotation and distraction at the osteotomy site are prevented (Fig. 21). As with any femoral fracture or osteotomy in the nonchildhood years, fat embolism can be encountered. Shortening of up to 2 in. or 5 cm can be performed readily in the femur in one stage. There is concern about performing more shortening than this due to the lag effect of the shortening on the quadriceps muscle. In the large majority of studies performed, any quadriceps lag can be overcome usually within 1 year of surgery with intensive physical therapy. It is, however, difficult to guarantee this effect if more than 5 cm of bone is removed. Two complications of the shortening technique that must be guarded against carefully are stabilization of the femoral fragments with inappropriate rotation and subsequent loss of close apposition of the fragments, which not only increases the likelihood of delayed or nonunion but also leads to a failure to gain a complete correction of the discrepancy. Both the rotational and distraction possibilities with intramedullary rodding can be minimized with the application of a small four-holed AO side plate with a unicortical grip or with the use of a universal AO nail with static locking proximal and distal screws. Liedberg and Persson (300) reported on 11 midshaft femoral shortening procedures in which stability was provided with a reamed Kuntscher intramedullary rod and a step osteotomy for rotational control. D'Aubigne and Dubousset (133) described excellent results with a step cut or Z shortening followed by intramedullary stabilization and fixation with screws to control rotation. Femoral shortening can also be done with the correction of proximal and distal angular deformity by removing appropriately shaped trapezoidal wedges.

SECTION IX ~ Management o f Lower Extremity Length Discrepancies

671

F I G U R E 21 Radiographs of a case of femoral shortening stabilized by an intramedullary rod are shown. Five centimeters were removed from the longer femur at skeletal maturity at open operation. (A) An anteroposterior film of the proximal two-thirds of the femur approximately 10 weeks postsurgery shows the healing osteotomy site. There are two transfixation screws proximally and two distally to serve as controls, preventing rotation or lengthening at the osteotomy site. (B) A lateral radiograph at the same time shows the maintained position of the IM rod and healing progressing well. Anteroposterior (C) and lateral (D) radiographs show the fixation device distally. (E) Anteroposterior radiograph of entire femur at healing is shown.

Wagner (492) has commented on some of the important principles underlying diaphyseal femoral and tibial shortening osteotomies. Wherever possible he has recommended the use of an intramedullary nail as distinct from a side plate and cortical screws because the intramedullary nail allows for early weight bearing. With femoral shortening he recommends removing no more than 6 cm of bone to avoid muscular insufficiency, and he supports the use of open osteotomy to remove the bony segment from the diaphysis. Attention to rotation is important. Even with a tibial and fibular shortening osteotomy an intramedullary tibial nail is favored. Reaming of the medullary cavity is essential to ensure a tight fit and minimize the likelihood of rotational abnormalities. The tibia is resected in its narrowest portion at the middle one-third. The stable mechanical fit by an intramedullary rod makes the diaphyseal site for shortening preferable to metaphyseal except for those disorders in which axial correction is needed. a. F e m o r a l S h o r t e n i n g . In a review of 46 limb shortening operations, 37 in the femur and 9 in the tibia, Kenwright and Albinana (270) felt that shortening of as much as 7.5 cm could be done in the femur and 5 cm in the tibia in adults of normal height without any loss of function. Major problems with the technique were technical in nature and involved inadequate stabilization at the osteotomy site. Attention to detail, however, should minimize or eliminate these problems. They concluded that the optimal site for femoral shortening was at the subtrochanteric region using an open approach and stabilization with an intramedullary rod with

ml mlproximal locking. Extreme importance was attached to preventing separation at the osteotomy site postoperatively and in achieving appropriate rotation. After tibial shortening, one problem, cosmetic in nature, was complaint by the patient of a localized increase in the bulk of the leg. For that reason Kenwright and Albinana recommended that tibial shortenings be limited to 4 cm or less. Closed nailing had been shown to be reliable by Winquist (509) and by Blair et al. (54), who also expressed concern about controlling rotation, possible instrumentation breakage, and the mass of bone that persisted and was occasionally troublesome. They also felt that healing in the tibial shortening procedure would be most enhanced with excision at the level of the flare in the lower diaphysis. They used either a standard AO intramedullary nail or an AO plate but felt that, with intramedullary nailing, locking mechanisms should be incorporated. Sasso et al. (423) also demonstrated results after closed femoral shortening. Shortening averaged 4.4 cm with a range from 3 to 5 cm in 18 cases. Complications included 1 episode of fat embolism and 3 cases of loss of fixation. They limited shortening to 5 cm because of concern about quadriceps lag, and in no instance at this amount of resection were any negative sequelae noted with a full knee range of motion and excellent strength regained. One possible but rare complication of closed intramedullary nailing of the femur in association with shortening is avascular necrosis (AVN) of the femoral head. Mileski et al. (326) have reported such a problem in an adolescent female who at age 11 years underwent a closed intramedullary

672

CHAPTER 8 ~

Lower Extremity Lenyth Discrepancies

femoral shortening. The rod subsequently was removed after healing. AVN was not diagnosed after rod placement but was noted 15 months after rod removal. The authors feltthat the vascular insult may have been caused by using a rod entry point that was placed slightly medially, although it was also possible that rod removal had contributed to the vascular insult. This represents the first report of this complication with femoral shortening, although Herzog et al. (233) reported 4 cases of femoral head necrosis following intramedullary Kuntscher nailing of femoral fractures in 26 children and O'Malley et al. (355) reported AVN in a 13-year-old boy with a closed intramedullary nailing of a femoral diaphyseal fracture. b. Tibial-Fibular Shortening. Broughton et al. (76) reported on 12 patients who underwent tibial and fibular shortening by diaphyseal resection to correct limb length discrepancies over a 25-year period. Each of the 12 patients did well with no major complications. A step-cut technique or Z-type shortening was employed for the tibia accompanied by a midfibular resection of the required amount. The tibial fragments were then stabilized by two screws. All operations except 2 were performed either at or just beyond skeletal maturity. The shortening achieved ranged from 2.5 to 5.1 cm. Normal function and appearance were documented in all following uneventful healing.

C. Procedures to Lengthen the Shorter Limb 1. STIMULATION OF EPIPHYSEAL GROWTH PLATE OF THE SHORTER LIMB

The earliest approaches to the treatment of lower extremity length discrepancy beginning in the late nineteenth century and continuing well into the first half of the twentieth century involved efforts to stimulate growth of the shorter side prior to skeletal maturity. There was early recognition of the fact that periosteal irritation frequently led to long bone overgrowth, and the therapeutic method that evolved from this observation was an attempt to stimulate and irritate the periosteum to allow for increased growth of the adjacent physis. Ollier (354) was the first to develop this technique. He performed periosteal elevation in the rabbit tibia causing overgrowth of 2-5 mm within 3 months. Over the subsequent few decades many methods were used to stimulate the periosteum, including cutting the periosteum, circumferential stripping or elevation of the diaphyseal periosteum, and placing objects underneath the periosteum to allow for chronic irritation (these objects included ivory pegs). Several differing techniques were used, however, both experimentally and clinically to enhance physeal growth. a. Sympathectomy In 1930 Harris (219) reported on lumbar sympathectomy performed on the short side of a patient with poliomyelitis to take advantage of the observation that those having sympathectomy for vascular disease frequently developed vascular dilatation and increased warmth of the affected side. It was postulated that this increased vascular

response would enhance physeal growth. In many instances overgrowth was indeed caused by the sympathectomies, but its occurrence was unpredictable and rarely exceeded 1 cm. These techniques are not used today. In a more detatiled presentation Harris and McDonald (220) in 1936 reported on the response to 46 lumbar ganglionectomies. Use of the procedure was based on the observation that in certain pathological conditions in growing children characterized by prolonged hyperemia in the neighborhood of the epiphyseal growth plates there was overgrowth of the involved extremity. They attempted to reproduce the earlier clinical finding by performing lumbar sympathectomy in several kittens, puppies, and lambs, but in no instance were they able to reproduce growth stimulation on the ipsilateral side. They performed 70 lumbar sympathectomies in patients with poliomyelitis and assessed the clinical response. Forty-six were available for review. Harris and McDonald showed many instances in which lumbar sympathectomy enhanced the growth of the extremity. In 21 of 46 patients (46%) the shortness had decreased by amounts varying from 0.12 to 1 in. The average age at time of surgery in this favorable group was 8.5 years. In 8 cases (17%) the amount of shortness present at operation remained unchanged, which still represented a positive response because the progressive shortening had ceased. Beneficial results were demonstrated in 63% of the cases. In 17 cases (37%) the shortness progressed in spite of operation. In many of this latter group the authors felt that effective sympathectomy had not been either obtained or maintained. When they subdivided their assessment to include only those with very clear sympathectomy, 20 of 29 cases showed diminution of shortness, 4 showed no increase in shortness, and in only 5 did the shortness increase. Beneficial effects on growth were thus increased to 82%. Barret al. (35) in 1950 assessed the results of 23 unilateral lumbar gangliomectomy procedures in patients with poliomyelitis. Ipsilateral lumbar ganglionectomy had, in some instances, a stimulating effect upon the growth of the shorter extremity. The most favorable interpretation of their results showed that the patients had an average decrease in discrepancy of 1.5 cm, based upon projections of the expected discrepancy. In the control group of 23 cases, 21 increased and 2 decreased with an average increase of 1.8 cm; in the ganglionectomy group of 23 cases, 13 increased, 9 decreased, and 1 was unchanged with an overall average increase in discrepancy of 0.3 cm. It was calculated that the average decrease in discrepancy with ganglionectomy was 1.5 cm. b. Surgically Induced Arteriovenous Fistula. Observations made in the late 1800s and early 1900s reported overgrowth of childhood limbs in patients who had sustained an arteriovenous (AV) fistula. Horton (239) reported 23 cases of congenital arteriovenous fistula with overgrowth of the involved extremity almost always seen. In midcentury efforts were made to incorporate this observation into clinical practice by surgically inducing arteriovenous fistulas in the mid-thigh region on the short side to treat developing limb

SECTION IX ~ Management of Lower Extremity Length Discrepancies

length discrepancies. James and Musgrove (252) showed the growth stimulation effect of an experimentally created arteriovenous fistula. Janes created the first AV fistula in a child with a short limb due to polio and 10 years later reported his results (251). Mears et al. (324) induced 55 fistulas and studied their results in detail. The fistula was placed between the superficial femoral artery and vein. There were no major cardiopulmonary complications. Thirty-nine patients were available for long-term review. All patients but 3 had a short limb due to polio. In 28 there was either a decrease or at least no increase in the amount of discrepancy. In 11 the limb length inequality continued to increase after the fistula was established. Thus, 72% of the group had a discrepancy that was diminished or stabilized by the fistula. Lengthening of the shorter leg after establishment of the AV fistula was 0-0.5 cm in 11, 0.5-2.5 cm in 13, and 2.5-5 cm in 4. When continued shortening occurred after the AV fistula, the amounts were still much less than would have been anticipated, with 5 patients having 0-0.5 cm of additional shortening and 6 having 0.5-2.5 cm of additional shortening. The growth pattern following establishment of the fistula was unpredictable and the response was quite variable in terms of extent. Petty et al. (386) assessed their results following surgical creation of a femoral arteriovenous fistula to treat limb length discrepancy in 28 patients. Of the fistulae made, 21 of 28 were performed when the patient was felt to be the optimal 8 or 9 years of age. The fistula was created between the femoral artery and vein in the mid-thigh region. Considerable complications occurred secondary to fistula creation although none was limb threatening. Closure of the fistula was eventually performed in all by 16 years of age. Of the 28 patients operated, 17 subsequently had an epiphyseal arrest on the opposite side. The average length discrepancy in this group was 4.6 cm at the time of fistula creation and the average increased to 5.9 cm at the time of epiphyseal arrest. Eleven patients did not have an epiphyseal arrest, and in this group the average discrepancy at fistula creation was 4.1 cm and at fistula closure it had decreased to 2.4 cm. Only 9 of 28 patients (32%) showed a decrease in length discrepancy of more than 1 cm as a result of an AV fistula alone. Twentyone of 28 patients (75%) showed no further increase in discrepancy, however. The authors concluded that "artificially created arterio-venous fistulae can accelerate growth in the lower extremity, but the results vary greatly and are unpredictable." By 1970 both Janes and Sweeting (253) and Petty et al. (386) no longer performed or recommended the procedure for treatment of length inequality. c. Elevation and Stripping o f the Metaphyseal and Diaphyseal Periosteum. Increased growth in length long had been noted after stripping of the periosteum of the metaphysis and diaphysis in several experimental animal procedures. An experimental study of stimulation of longitudinal growth of long bones by periosteal stripping in dogs and monkeys was reported by Sola et al. (450). They assessed

673

not only an initial periosteal stripping but also the effects following a second stripping 1 or 2 months after the first procedure. Once again there was a tendency to show increased growth, although it was not particularly marked nor was it invariably seen. In operations on dogs involving a single stripping of the femur and tibia, the mean increase in length was only 0.16 cm with 63% showing an increase of growth on the operated side. When two stripping procedures were done the mean increase was greater at 0.35 cm, but still only 69% of the animals showed an increase on the operated side. When two procedures were done on monkeys the increase was only 0.17 cm, although 87.5% showed an increase. The stripping procedure was extensive going from growth plate to growth plate in both femur and tibia. A detailed study of the effects of stripping of the periosteum in rabbits was performed by Wu and Miltner (513). Periosteal stripping was performed in variable parts of the fight tibia and fight femur. In all instances there was overgrowth on the operated side. Twenty-two rabbits were used. The fight tibia alone was operated upon in 18 animals and both the right femur and tibia were operated upon in 4. Definite longitudinal overgrowth of the operated bone was observed in all instances except 3. The amount was small and primarily limited to the first 3 months postsurgery, with the overgrowth varying from 0.5 to 6 mm. Chan and Hodgson (109) applied this procedure to 45 patients suffering from poliomyelitis, with the operations performed between 1961 and 1968. The age at surgery ranged from 5 to 13 years and all patients had a short limb at the time of surgery, averaging 3.4 cm (1.1-9.5 cm). The periosteum was completely stripped with an elevator in both the femur and tibia. A definite overgrowth greater than 4 mm was noted in 31 patients (69%), there was no significant increase or decrease in growth over the normal side in 9 (20%), and there was continuing shortness in 5 (11%). In the favorable group the average overgrowth was 1.3 cm during the mean time period of 9 months with a range from 0.6 to 4.4 cm. In many patients the overgrowth effect was noted to persist for as long as 1, 2, and even 4 years postsurgery. The average period of stimulation, however, could not be determined accurately. The authors concluded that it was best to perform the stimulation operation when the child was 8 years of age. No meaningful or long-term complications were seen. Jenkins et al. (257) studied 13 of these patients at a later time from 3 to 5 years postsurgery. They continued to note some degree of stimulation but again a variable response. In 28 femurs assessed 3-5 years following periosteal stripping, there was a mean increase in growth of 0.5 cm in 17, a decrease in growth of 0.81 cm in 8, and no change in 3. In the 26 tibias, 18 had an increase of 0.75 cm, 5 a decrease of 0.5 cm, and 3 were the same. d. Shortwave Diathermy. Doyle and Smart (145) reported that shortwave diathermy enhanced epiphyseal growth in rats. Preliminary experiments indicated that a temperature of 40~ would be effective to induce increased growth without

674

CHAPTER 8 9 Lower Extremity Len~tth Discrepancies

tissue damage. The right lower extremity was maintained at this temperature throughout the treatment. Twenty female rabbits were used in which insulated copper plates, 2 • 8 in., permitted the administration of diathermy to 4 - 6 animals via a shortwave medical diathermy apparatus. Treatment was directed to the epiphysis of the right knee. The animals were treated for 0.5-1 hr each day or on altemating days from the 21st to the 70th day of life. The average total duration of diathermy was 25 hr. Of those animals appropriate for assessment, the treated right hind limb was longer than the untreated left in all instances. The increase in the combined length of the treated tibias and femurs varied from 0.4 to 2.8 mm, averaging 1.4 mm. The advantage of the procedure was that it would increase the temperature in tissues at deeper levels without causing bodily damage. Diathermy acted by producing deep tissue heating and increased blood flow. Shortly thereafter, Granberry and Janes (191) repeated the experiment on dogs without showing beneficial effects. They used microwave diathermy to increase the temperatures 3-5~ in the tibia. In their experiment, one knee each of 7 young dogs was heated by microwave diathermy at 100 W for a total of 100 hr, but they noted no significant growth alteration.

e. Efforts to Stimulate Epiphyseal Growth by Insertion of Multiple Implants and Creation of Bone Damage in the Metaphyseal Regions. Wilson and Thompson (508) noted that the many attempts to stimulate epiphyseal growth had not been successful enough to warrant clinical use, a feeling still prevalent after additional attempts since then. Compere and Adams (122) addressed the effects of trauma to the diaphysis on subsequent longitudinal growth. They performed two series of experiments in rabbits with limited trauma. Three drill holes were placed through both lateral and medial cortices near the upper and lower metaphysis and in the middle one-third of the shaft. This served to disrupt the medullary blood supply. When animals without fracture during the postoperative phase were assessed, there were no differences in the length of either the tibia or femur. This led to the conclusion that trauma sufficient to interrupt the medullary blood supply but not great enough to cause regional hyperemia did not consistently cause growth stimulation. Fractures were then made in the femur and tibia between metal markers to assess the overgrowth phenomenon. Compere and Adams demonstrated, in agreement with Bisgard, that overgrowth of a long bone may occur following fracture even without shortening and that the longitudinal overgrowth occurred entirely from stimulation of the epiphyseal growth cartilage. They then assessed patients in relation to the growth of tibias from which bone grafts had been taken. Growth arrest lines were used for the assessment. Small amounts of increased growth ranging from 0.1 to 0.8 cm occurred following tibial bone gratis. Evidence was clear that the growth stimulus lasted only as long as healing of the defect site was occurring. Compere and Adams concluded that minimal trauma to the shaft or to the metaphysis of the long bone with or without interruption of the medul-

lary blood supply did not produce any definite increase in longitudinal bone growth. Gross trauma such as that caused by a fracture or removal of a large segment of bone for grafting, both of which necessitated extensive bone repair, did reproduce epiphyseal stimulation and increased longitudinal overgrowth. The increased rate of growth continued during the period of healing but not much beyond. The growth stimulation appeared secondary to the hyperemia, which included the epiphyseal region. Wu and Miltner (513) reviewed clinical situations in which overgrowth occurred. Metaphyseal and diaphyseal fractures were known to cause overgrowth in long bones in children. Infection could clearly damage growth if it involved the physeal cartilage, but in instances in which the physeal cartilage persisted the increased hyperemia led to overgrowth. They performed several experiments on rabbits aged 5-8 weeks to assess growth phenomena. Group 1: Insertion of foreign material into a drill hole placed immediately distal to the proximal epiphyseal cartilage of the tibia. The foreign materials included cotton, gauze, paper, wood, brass, and iron shot. There was no difference in the length of the bones operated. Group 2: Indirect interference of circulation of bone. The epiphyseal circulation was left intact but the nutrient arteries and periosteal vessels were damaged extensively. There was no appreciable change in longitudinal growth of the bone after the experiments. Destruction of the nutrient artery and the extra-periosteal blood supply in particular caused no changes in the longitudinal growth of the bone, agreeing with the extensive studies of Oilier and Haas. Group 3: Curettage of bone marrow. The tibial bone marrow was curetted through a metaphyseal drill hole. There was no significant change in the length of the operated bones. Group 4: Stripping of the periosteum. Many variable patterns were used, and in virtually all instances definite longitudinal overgrowth of the operated bone was seen from 0.5 to 6.0 mm. Chapchal and Zaldenrust (111) assessed the effect of various metals, metal alloys, and ivory placed in the metaphyses and also in the epiphyses of the bones comprising the knee joints of several animals. They concluded that some lengthening of the bones was obtained but that the amount was minimal and uncertain. Pease (376) attempted to assess overgrowth using foreign bodies in the metaphyseal regions. His work also involved clinical investigation of the phenomenon. He placed transverse screws across the entire metaphyseal diameter of the distal femur and/or proximal tibia using vitallium, stainless steel, vanadium, and ivory screws. Two screws per region were used. In all cases stimulation of growth followed the operation to a variable degree, and there were no deformities indicative of asymmetric stimulation. The screws were placed parallel to each other and to the adjacent growth plate and extended to or slightly through the opposite cortex. Seven patients were operated with two screws placed in the tibia and femur in most with 1 patient having the procedure only in the femur. The operation was

SECTION IX 9 Management of Lower Extremity Length Discrepancies

repeated occasionally. Seventeen segments were stimulated. In two instances there was no overgrowth, whereas in the others overgrowth stimulation varied from 0.1 to 2.2 cm. The mean length increase for 17 cases with growth stimulation was 0.7 cm. Carpenter and Dalton (97) repeated the clinical work in 30 cases in which epiphyseal stimulation was attempted by the use of intramedullary implants in a distal femur and proximal tibia. The periosteum was elevated and a cortical window was made to the metaphyseal side of the distal femoral and proximal tibial growth plates. The medullary canal was curetted and the cavity then tightly packed with small chips of ivory. Each patient was followed for a minimum of 2 years with radiographic and clinical measurements made at 3-month intervals. Some increase in growth was obtained in 26 of the 30 patients. The gain was 0.12 in. in 10, 0.25 in. in 11, 0.5 in. in 3, 0.75 in. in 1, and 1 in. in 1. In 70% of the cases the maximum gain was only 0.12-0.25 in., and it was concluded that the degree of stimulation was neither great enough or predictable enough to warrant clinical use. Tupman (483) reported a detailed study to stimulate bone growth by inserting beef bone pegs into the epiphyseal and metaphyseal regions in children. The first clinical attempt to stimulate growth in dogs by introducing ivory pegs into the femur and tibia was in 1869 by vonLangenbeck, who claimed 1 cm of overgrowth in 3.5 months. Tupman reviewed 28 patients who had a total of 51 operative procedures. Insertions of material were in the distal femoral and proximal tibial metaphyses. Three tunnels were made with a drill in the metaphysis close to but not involving the physis. Into each of these tunnels in the femur and tibia, a beef bone peg was inserted across the bone diameter. In some, ivory pegs were inserted for comparison. In one part of the series performed at one hospital, no leg subjected to a single operation showed any evidence of acceleration of growth. It was felt, however, that documentation might have been somewhat inadequate. In a second series at another hospital, there were groups in which effective stimulation took place, some in which the operation had no effect on progression, and some in which the operation was fully ineffective. Only limited conclusions could be made. The operation appeared to be more effective in younger children than in those close to puberty. The best that could be said was that the growth was stimulated somewhat in 12 of 28 patients, with the best resuits seen when the operation was performed between 6 and 12 years of age. Once again some stimulation was noted, but it tended to be of a small amount and unpredictable. 2. LENGTHENING OF THE DIAPHYSIS

a. Clinical Approaches from 1900 to the 1960s. Much interest in, and occasional efforts at, lengthening lower extremity long bones was reported in the nineteenth century. No acceptable techniques evolved, but valuable biological and mechanical principles gradually became evident with continued work over the next several decades.

675

Codivilla: Well-documented attempts at long bone lengthening date from accounts by Codivilla of Bologna, Italy, in 1903 and 1905 (114). He began one of his articles with the following, still relevant sentence: "The difficulties to be encountered in lengthening a shortened limb are found in operation to be greater as regards the fleshy parts than as regards the bones." Following osteotomy or fracture of the bone, skin traction had been applied characteristically with limited effect due to pressure necrosis of the skin, pain, and the fact that much of the force applied "did not reach the skeleton." Codivilla went on to describe the evolution of his technique, which eventually involved osteotomy of the bone, application of a hip spica cast, removal of the foot portion of the cast, application of a force directly to the skeleton with transverse placement of a calcaneal wire of 5 - 6 mm diameter, incorporation of the transverse pin and two side bars into the plaster cast, cutting of the cast at the level of the osteotomy, and application of counterbalanced traction to gain length immediately followed by completion of the cast at the desired length until healing. Codivilla reported on 26 patients who gained between 3 and 8 cm. Freiberg (169) supported the validity of the approach, using it following a femoral fracture in a 9-year-old boy to reduce a 2.25-in. shortening to 0.5 in. 5 weeks after injury. A double skeletal transfixator method, after the length had been achieved, then evolved. Ombredanne: Ombredanne (356) began utilizing principles subsequently incorporated in more formal apparatuses at later times. In 1913 he described an oblique osteotomy and lengthening of the femur slowly and gradually with an apparatus fitted to the side of the thigh and working against one pin inserted above the osteotomy site and one below. He achieved up to 4.0 cm of lengthening, but no detailed followup of the procedure was performed. Putti: The next technical advance in limb lengthening was described by Putti (396), also of Bologna, in 1921. He described a need for lengthening of the femur when the discrepancy was greater than 2 in. and questioned early on whether lengthening of such a magnitude was "possible without damaging the muscles, nerves and vessels." His own work and that of Magnusson then showed the possibility of lengthening safely by 2-3 in. Putti defined the need for continuous traction. He developed a unilateral distraction apparatus for lengthening the femur, which used two large transcortical metal pins on either side of the osteotomy held apart by a telescoping tube that contained a strong spring press moved by a screw. The apparatus was designed to be sufficiently strong to overcome resistance, to stabilize the osteotomy site and the alignment, and to provide traction. Gradual traction was applied to separate the bone fragments, but the time taken for lengthening was not reported. Putti (397) elaborated on operative lengthening of the femur in 1934, with an altered technique. He stressed that operative bone lengthening was particularly valuable during the period of "childhood when the reparative power of the

676

CHAPTER 8 ~ Lower Extremity Length Discrepancies

long bones is most vigorous thereby minimizing the danger of non-union." Skeletal traction mandated the use of skeletal countertraction. The Z osteotomy was replaced with an oblique osteotomy. The patient then was put in bed with traction placed on the upper wire. Increasing traction was applied to the lower Kirschner wire until the desired length, usually from 2.5 to 4 in., was attained. This gradual skeletal traction usually required 18-21 days. A hip spica cast was then applied with the transfixion wires included, at which time the traction was discontinued. Results were reported in 11 patients. Putti added that "in the eleven cases I have treated by this method no complications whatever arose other than a single case of temporary 'toe drop' caused by overstretching of the external popliteal nerve presumably due to faulty position of the knee. The paralysis promptly cleared up with rest." The concept of gradual traction was introduced, and Putti indicated that the time required to obtain the desired lengthening was approximately 20 days. The period of immobilization in the corrected position was 4 months in a plaster cast with an additional period of limited support for several months, during which physical therapy continued. No indication was made of the amount of lengthening achieved. Magnuson (308) described femoral lengthening of 2.5-4 in. in 14 patients using double transfixion wires and plaster with lengthening at one sitting by skeletal traction over 20-30 min. The osteotomy was a Z-type to allow for bone contact to enhance healing. The apparatus was left on for 30 days, following which a cast was applied. In 10 cases of femoral shortening, which were operated, the lengthening varied from 3 to 4 in. Abbott: Abbott (1) directed his attention to lengthening the tibia and fibula in patients with poliomyelitis and reported a method of lengthening in 1927 that gained wide acceptance. He used Putti's concepts of skeletal traction and countertraction but developed a lengthening apparatus, which was the true precursor of apparatuses used even now. His method employed a preliminary Z lengthening of the Achilles tendon, oblique osteotomy of the distal one-third of the fibula, insertion of single proximal and distal 0.19-in.-wide traction pins, which passed completely through the limb to enable fixation to a biplanar distraction apparatus to control angulation, circumferential division of the periosteum at the osteotomy site, a Z-type tibial osteotomy, and a delay in lengthening from 7 to 10 days postsurgery. The two distraction devices consisted of telescoping brass tubes and a strong coil spring with lengthening performed by turning of a thumb screw. The limb was immobilized on a Thomas splint to which were also attached two metal stabilizers to prevent anterior angulation. Lengthening was done once daily and ranged from 0.06 to 0.12 in. In his initial report, the maximum amount of length that Abbott felt could be achieved safely was 2 in. (5 cm). The entire time in traction was between 3 and 4 weeks. Overall the apparatus remained in place for 8-10 weeks, at which time the limb was immobi-

lized in a cast. The external apparatus and the wires were removed from 4 to 5 months after surgery when there was sufficient callus to permit the patient to walk with a splint. Abbott indicated that "in every case treated by this method callus filled in this space lying between the fragments in a comparatively short time." The operation was performed in those cases with 1.5 in. or more of shortening. Abbott warned of the extreme care needed in the postoperative phase and suggested "the surgeon who has not the time to give for daily adjustment of the apparatus should leave it entirely alone." Abbott then provided a very detailed account of his first six patients, one of whom was regarded as a failure. In the other five the gains in length varied from 1.75 in. to 1.88 in., with all of these increased lengths secured from 21 to 28 days. Union of the fragments sufficient to allow for weight beating in a splint was present in all cases from 4 to 5 months after surgery, and in all of the completed cases consolidation of the callus had taken place in 6 months. The amount of lengthening was initially kept relatively small because of concerns about negative sequelae with larger lengthenings. The recommendation and determination of the initial paper, however, allowed for increasing lengthening to 2 in. The next year Abbott and Crego (2) reported on a similar procedure for operative lengthening of the femur. Eight cases were reported with the gain in length ranging between 1.5 and 3.5 in. The authors felt that with experience in an average case a gain of 2.5 in. could be secured without producing injury to the blood vessels or nerves. The principles again involved osteotomy of the bone, direct bone traction that was gradual and continuous, and maintenance of alignment and contact of the fragments obtained by transverse wires entirely through the limb above and below the osteotomy and attached to the screw extension pieces. The osteotomy was accompanied by sectioning of deep fascial structures, the iliotibial band, and the biceps tendon to diminish the resistance of the soft parts. The authors emphasized that "by far the most important and difficult part of the entire procedure is the post-operative care of the patient during the lengthening process." The lengthening began approximately 5 - 6 days after operation. Lengthening was performed once daily with average daily gains of 0.12 in. and the entire time of traction extending over a period 4 - 5 weeks. The apparatus was removed and a cast applied in 10-12 weeks, protected weight bearing was allowed at 5 months, and full weight bearing occurred in 7-8 months. Abbott (3, 4) next presented a review of 48 tibial and fibular lengthenings in which he again stressed his three principles: (1) to lengthen a bone, traction and countertraction must be taken directly on that bone; (2) to overcome the elastic resistance of the soft parts, the traction must be slow and continuous; and (3) after osteotomy and the application of traction, complete control of the fragments including their appropriate alignment must be maintained during the lengthening process. The apparatus evolved to the use of two pins above and two below the osteotomy site. The lengthening

SECTION IX ~ Management of Lower Extremity Length Discrepancies

began only when all swelling had disappeared, which usually was at 7-10 days. In the 48 patients the gain in length ranged from 1.5 to 3.25 in. The average time of union to permit weight bearing with a splint was 4-5 months, and consolidation of callus generally occurred in 6-7 months with restoration of the medullary canal in 10-12 months. The complications listed by Abbott were surprisingly infrequent, involving two fractures, one worsening of paralysis of the dorsiflexor of the foot, and one infection, which involved an osteomyelitis of the bone, and occasional overlengthening of the tibia allowed for valgus deformation at the ankle. In a later paper, Abbott (4) reviewed his results in 73 procedures, 48 of which were in the tibia and fibula and 25 in the femur. In the femur the maximum gain was 3.5 in. and the minimum was 1 in. This paper describes the technique of the apparatus and the intraoperative and postoperative approaches in great detail. Abbott noted that the tibia and fibula provided much more dependable results than the femoral lengthenings, which had more complications (5). There were only two incomplete fractures following the tibial lengthenings, but seven fractures occurred through femoral callus during the early weight bearing phase. There were three cases of nerve paralysis. One involved paralysis of the sciatic nerve associated with subluxation of the knee. Infection of the pin sites was commonly seen but only one deep osteomyelitis occurred. Other Reviews of Abbott Technique: Following Abbott's development and subsequent refinement of his technique, limb lengthening enjoyed a surge of popularity. Three major reviews of relatively large series of the Abbott procedure appeared in the 1930s. These series served to define both the benefits and negative aspects of these interventions, although the detailed descriptions of the complications considerably diminished enthusiasm for lengthenings. Haboush and Finkelstein (213) reported on 17 tibial-fibular lengthenings, 16 of which were done for poliomyelitis. They used Abbott's technique although they did modify his apparatus somewhat. Their major problems involved (1) anterior-medial angulation during the bone lengthening process, (2) more tibial elongation than fibular, (3) valgus of the foot, (4) equinus of the foot, (5) osteomyelitis both at the site of the tibial osteotomy and at the site of pin insertions, (6) delayed union, leading in some instances to nonunion. The modifications to their apparatus were designed to address these problems. One of their major conclusions was that the fascia in particular yielded poorly to lengthening, which led to both the angular problems and the severe pain that they described as being "a rather constant feature in this series of cases." One of the major changes in their operative technique was a more thorough division of the fascia in the limb along with the interosseous membrane and complete circumferential division of the periosteum of both tibia and fibula. A more positive review was published in 1935 by Brockway (74), who provided a clinical resume of 46 leg lengthening operations using the Abbott technique and apparatus.

677

He provided average values for the key criteria in his longrange assessment, although ranges of numbers and case by case details were not given. His report is most instructive to note in comparison with current approaches using the distraction osteogenesis principles. The average age of the patient was 14 years. The number of days following surgery before the lengthening process was started averaged 7.5. The average length obtained was 1.9 in. The number of days required to obtain this length was 35, making the average daily increase in length 1.3 mm. The average time before the pins were removed (at which time the long leg cast was applied) was 11.4 weeks. The average time that plaster was worn following the removal of the pins was 13 weeks, and the average time before full weight beating was allowed postsurgery was 9.5 months. Brockway did note complications with the procedure, but his overall impression was positive and he indicated that "on the whole, very gratifying results have been obtained by this operation and it is now a routine procedure." Poor results involved fracture of the tibia, some cases of skin slough, delayed bone healing, and anterior bowing of the tibia. The most critical of the early reviews was published in 1936 by Compere (121), although his paper is somewhat unique by current standards. The large list of complications that he referred to led him to indicate his belief that shortening was by far the more preferable approach to limb equalization. The paper describes five patients in detail, each of whom had a large number of complications listed. There is no indication, however, as to how many patients were operated. The discussion also focuses on the negative aspects of the procedure, which is important, but almost totally neglects any positive indications in either his own work or the work of others. Compere then listed 14 complications, which, in the absence of any indication that he had any good results at all, clearly led to a major dampening of enthusiasm for this intervention particularly in North America. The complications he listed follow: (1) stretch paralysis of the sciatic or the external popliteal nerve; (2) increased weakness of lengthened muscles in old cases of poliomyelitis; (3) fracture of the osteotomy; (4) malunion; (5) delayed union or nonunion; (6) osteomyelitis from wound infection; (7) traumatic arthritis and limitation of motion in the knee; (8) late fracture; (9) pressure or stretch necrosis of the skin in the zone of lengthening; (10) necrosis of bone due to excessive subperiosteal stripping, which also might increase the likelihood of infection; (11) malposition of the foot due to rotation following lengthening; (12) circulatory disturbance with prolonged edema in the lengthened limb; (13) displacement of the head or of the distal end of the fibula when this bone is not lengthened as much as the tibia; and (14) protrusion of the osteotomy fragment of the tibia through the skin. Complications have characterized lengthening procedures from the beginning. Table II lists the large group of possible disorders. Many of these are seemingly inherent with the procedures, but awareness should help to minimize them.

TABLE II Bone

Skin

Muscle

Nerve

Vessel

Joint

Physeal cartilage

Possible Complications in Limb Lengthening Procedures Premature union Incomplete osteotomy Long latency period prior to beginning lengthening Interruption of lengthening due to other causes Malunion Angular deformity/axial deviation Unstable apparatus Imperfect pin placement or initial malalignment postosteotomy Altered muscle pull with increased extent of lengthening Proximal femur, varus Distal femur, valgus Proximal tibia, valgus Distal tibia, varus Tibial lengthening greater than fibular ~ valgus ankle Angular deformation post fixator removal (softened bone at distraction site) Scanty union Delayed union Nonunion Osteomyelitis Lengthening site Pin tract site Osteoporosis Late fracture Shortening Angular deformity Pin-wire pull-out Pin site irritation-infection Skin slough secondary to malaligned fragment pressure Fragment protrusion Fibrosis Contracture Weakness Myopathic Neurogenic Intraoperative phase Pin skewers nerve Distraction phase Excessive stretching; sciatic, peroneal, posterior tibial, radial nerve Sensory: paresthesia, hyperesthesia, anesthesia, transitory, permanent Motor: partial paralysis; paralysis Intraoperative phase Pin skewers vessel Vessel damaged during osteotomy Distraction phase Excess vessel stretching Compartment syndrome Hypertension Excessive arterial stretching Thrombophlebitis Stiffness: ankle equinus, knee extension, hip adduction-flexion contractures Cartilage degeneration Subluxation Knee Hip Dislocation Hip Septic arthritis (pin placed into joint cavity) Increased growth of femur following completion of lengthening a Diminished growth of tibia following completion of lengthening a

aln relation to prelengthening growth rates.

SECTION IX ~ Management of Lower Extremity Length Discrepancies

The major English language orthopedic journals did not publish any subsequent large limb lengthening review for 12 years until the report of Allan (48). He used a tibial distraction apparatus, which incorporated the principles of Abbott and some modifications of Haboush and Finkelstein as well as his own. An oblique osteotomy of the bone was performed to enhance repair after the sliding lengthening. Allan applied a plaster of Paris long leg cast and incorporated it into the distraction apparatus with the cast cut at the osteotomy site. This was designed to minimize angulation. The periosteum was left intact and careful surgery sought to minimize damage to the soft tissues. Distraction began immediately and proceeded at a rate of 0.06 in. per day. Allan reported that "little pain should be experienced." When sufficient callus had formed as indicated on X rays but prior to firm union, the plaster cast and wires were removed, any malalignment was adjusted, and the limb was stabilized in plaster until union was complete. Femoral lengthening produced greater difficulty than tibial. In 47 cases of tibial-fibular lengthening, the average time before weight bearing was 6.5 months and the average lengthening obtained was 2.33 in. In 40 femoral lengthenings, the average time for weight bearing was 5 months and the average lengthening obtained was 1.62 in. The technique was evolving during the course of the series. Allan felt that bone was laid down in parallel lines between the fragments with the osteogenic material strung out across the gap and that most of the repair came from periosteum. He noted that all bones returned to normal radiographic appearance within a year or so of consolidation. Union eventually occurred in every case but there was a marked delay in 12 cases, with union taking from 9 to 16 months. Although Allan recognized that the most resistant structures to stretching were the periosteum, the interosseous membrane, and the deep fascia, he specifically mentioned that these were to be left intact as much as possible. The blood vessel response distal to the lengthening site was benign. He felt that the external popliteal nerve and the compartment part of the great sciatic nerve could be stretched to 2 in. in the thigh without losing function and that they could be stretched to 3 in. with only temporary impairment. Indeed, he felt that certain vascular complications experienced by other surgeons were attributable to sub-periosteal bone exposure and to dividing the periosteum and fascial structures transversely. Anderson: The next technical advance in limb lengthening accompanied by a burst of renewed popularity came from adoption of the technique described by W. V. Anderson (21) of Edinburgh, Scotland, in the mid-1960s. This technique evolved from an operation in 1954, but he did not report the technique in detail in writing until 1967 and even then apologized for including no numerical data. His paper described the evolution of his technique from that of Abbott, which was practiced with some modifications in his institution for approximately 20 years prior to 1954 with "satisfactory results obtained, with increases of length up to 389 inches in the tibia, with no complications of any sever-

679

ity." We quote from Anderson's paper, which eloquently reviews the subsequent reactions to the Abbott procedure after its seemingly excellent early results. Unfortunately, this procedure appeared to be so simple and satisfactory that its rapid and deserving popularity almost brought about its eclipse. It was performed widely by many surgeons who failed to realize the fundamental importance of suiting the operation to the patient; by this failure they were directly responsible for the numerous and terrifying complications which followed this tragic misuse. These ranged from gross sepsis to amputation, following vascular failure. The procedure was condemned loudly and bitterly because it produced too much pain and was technically difficult and destructive by those who had, in fact, made it so themselves. For many years, it was practically given up in the land where it originated. In the years following, the alternative methods of equalization were more fully developed with the perfection of epiphysiodesis and stapling, whereas shortening was much more in favor than lengthening. Eventually the Edinburgh group in which Anderson worked modified the apparatus and the procedure and began distraction on the operating room table. They noted that "contrary to the findings in America, we had none of the complications which ended in the condemnations of the operation. The pain was minimal, even from the day of the operation. There were no vascular or neurologic changes of any importance and no major sepsis occurred. The operation was accepted by operating and nursing staff, and the patient, as a simple routine procedure." One complication that was reported by others and indeed appeared in the Edinburgh series was the slowness of the lengthening of the fibula, which led to valgus deformation of the ankle. This was greatly minimized by a distal tibial-fibular synostosis obtained prior to the lengthening procedure. The Anderson technique as subsequently practiced evolved in a patient who suffered a transverse fracture of the mid-tibia as he was waiting for a more formal Abbott-type lengthening. This opportunity presented itself such that the apparatus was applied for stability and lengthening was performed. There was surprisingly good maintenance of alignment and it was noted that the lengthened site healed readily. The advantage of this approach was that only a very small linear incision was required, there was very little disruption of the periosteum, and the transverse osteotomy was made following positioning of several transcortical drill holes and manual osteoclasis. There was a 0.12-in. lengthening on the operating room table with subsequent lengthenings starting on the third or fourth day of one turn daily equaling 0.06 in. Anderson indicated that "in the average case the length of 2 to 3 inches may be obtained without difficulty" and "up to 1 and 89 inches can be expected before any tendency to foot deformity (tendoAchilles tightness) becomes evident." The patient remained in bed with the limb suspended in a traction frame. When new bone was seen radiographically, external fixation was removed and the limb was immobilized in plaster. Anderson

680

CHAPTER 8 9 Lower Extremity Lenyth Discrepancies

noted empirically that 0.06 in. increase per day remained the most satisfactory rate of lengthening. Anything less may predispose one to early bone union with callus formation overcoming the rate of lengthening, thus preventing the full length desired from being obtained, and a faster rate may cause a delay in union, considerable pain, and possible nerve and vascular complications. Anderson specifically remarked that "pain is so exceptional (in their lengthening unit) that apart from the immediate post-operative pain (in itself minimal) it is regarded as indicative of something abnormal and subsequently of importance." Following new bone formation deemed sufficient to prevent collapse, the apparatus was removed and the limb was placed in a long leg cast followed by eventual transfer to a walking brace. Femoral lengthening was performed in a similar fashion. A major review of the Anderson approach was published by Coleman and Noonan (118) from the Salt Lake City Shriners Unit in 1967 and by Coleman alone later on (119). Thirty-one tibial lengthenings using the Anderson technique with limited surgical exposure were reported. They performed the Anderson technique as he had described it, emphasizing "the advantage of osteotomy of the tibia by a limited surgical exposure, in which the hematoma remains localized, periosteal stripping is avoided, there is relatively little soft tissue damage, and the periosteal tube is preserved." A distal tibial-fibular synostosis was performed initially prior to lengthening to prevent valgus deformity at the ankle, but eventually stabilization was performed using a transfixion screw at the same time as the lengthening procedure. A cast was applied at the same time as the distraction apparatus and osteotomy to help minimize the development of equinus at the ankle. Distraction was at the rate of 0.06 in. per day. In the 31 patients, satisfactory union was obtained in all with an average gain in length of 5.0 cm (range = 2.0-6.0 cm). Nonunion occurred in 4 patients but satisfactory union was subsequently obtained with bone grafting. The authors modified their technique to perform a bone grafting procedure in all patients who at 4 months after surgery "show lack of complete bone bridging." There were no wound infections and only pin tract infection. There was no permanent detectable injury to the neurovascular structures. Manning (314) reported on a large series of femoral and tibial lengthenings using the Anderson apparatus and technique. The patient remained in bed for the lengthening procedures with the distraction apparatus supported in traction. Distraction took place once daily with a gain of 0.06 in. (1.6 mm). Once length had been achieved and early healing was underway, the patient was transferred to a cast and ultimately to a brace. Lengthenings were done with 211 procedures performed, 161 on the tibia and 50 on the femur. The average gain was 3.06 cm, with 122 of the tibias lengthened by 5.0 cm or more and 21 femurs by 5.0 cm or more. Major complications related to fracture after lengthening had been completed and delayed union or nonunion. Fractures occurred in 30 limbs (14%). These all subsequently united relatively quickly but shortening occurred in many, lessening

the advantage of the original lengthening procedure. Bone grafting was resorted to for slow union or nonunion present 6 months after the lengthening had been completed. Such grafts were used in 23 tibial lengthenings (14%) and 2 femoral lengthenings (4%). Each then healed uneventfully. Chacha and Chong (108) reported on overall favorable results with 35 tibial lengthenings (31 Anderson-type) with an average gain of 5.2 cm and a range of 2.8-6.5 cm. Malhis and Bowen (312) reported on 12 tibial lengthenings using the Anderson method. The mean amount of lengthening achieved was 6 cm (range = 3-10 cm) and the mean percent lengthening was 24.5% (range = 13-42%). Bosworth (61) used the Abbott technique but recommended not beginning lengthening until 10 days after the osteotomy. One-Stage Lengthening ProceduresmLe Coeur: Le Coeur (291) of Paris described a one-stage lengthening procedure with immediate stabilization and applied the technique 169 times between 1952 and 1962. The amount of length gained varied between 3.0 and 4.7 mm, and due to the immediate fixation used the amount gained was maintained with certainty. The period of immobilization was similar to what occurred with a fracture. The tibia was the most favorable site for lengthening, but femoral procedures were also performed. The operation was accompanied by multiple transverse muscle and fascial releases, which allowed for lengthening of the soft tissues readily in conjunction with the bone elongation. A lengthy oblique incision of the tibia was made, following which a bipolar traction apparatus was applied to upper and lower regions of the tibia. Once the soft tissue releases had been performed, the oblique osteotomy made, and the traction apparatus positioned, lengthening began with the knee in partial flexion. The knee remained in flexion during the lengthening procedure and during the time in cast following stabilization. Once the desired length had been reached and the surrounding muscle and fascial tissues were released, osteosynthesis was performed with four or more transverse screws. No cast was used. The patient remained in bed with the knee flexed for approximately 45 days to allow healing to occur. The patient then began walking with crutches. Le Coeur indicated that femoral lengthening could also be performed with the same technique, although it was more difficult and laborious to perform. He briefly reviewed his results in 125 cases involving 88 tibial and 37 femoral lengthenings. The amount gained ranged between 3.0 and 4.7 mm on a regular basis. In younger children with open growth plates, there was often an added effect of overgrowth providing an additional 1-2 cm of lengthening. There were no vascular complications. There were 11 neurological complications in the 125 procedures, the large majority of which resolved fully. Two full paralyses of the sciatic nerve occurred, which lasted for a year, two partial sciatic lesions, which cleared in 1 or 2 months, three paralytic lesions, and four partial neuralgias with the tibial lengthenings. Bone repair was uneventful, occurring in most in 45 days with no pseudarthroses created. Fractures did compli-

SECTION IX ~ Management of Lower Extremity Length Discrepancies

cate the procedure and occurred both during the early repair phase and also following repair anywhere from 3 months to 4 years postprocedure. There were 14 fractures reported in the 125 lengthenings (11%), with 7 in the femur and 7 in the tibia. The vast majority of lengthenings were performed for shortening secondary to poliomyelitis. Cauchoix and Morel et al.: A one-stage femoral lengthening was also described by Cauchoix, Morel, and colleagues in 1963 (105), with results from the 100 initial patients reported in 1972 (106) and a total of 180 cases reviewed in 1978 (107). The operation involved a lengthy middiaphyseal Z osteotomy of the femur in the frontal plane with the length of the longitudinal cut being 8 cm in the adult and 6 cm in the growing child. Distraction was performed against two transverse 5-mm Steinmann pins placed in the transverse axis through the proximal and distal femoral fragments. The knee remained flexed during the lengthening procedure and care was taken to keep the bone fragments in contact using a bone holder, which was alternately opened and closed during the lengthening procedure. Considerable releases of the fascia were completed along with decortication of the middle one-third of the femur and formal exposure of the sciatic nerve to check that it was not excessively stretched during the lengthening procedure. Once lengthening had been achieved internal fixation was secured by a posteriorly placed vitallium plate. The bone gaps left by the Z lengthening were then filled with iliac crest autogenous cancellous bone graft. The average gain in 180 cases was 3.7 cm, with 169 of 175 one-stage one-time lengthenings ranging between 2.6 and 4.5 cm. In 5 instances greater lengthening was achieved by performing the procedure twice. The operation was performed with relatively strict limits in terms of the maximum amount of lengthening that could be achieved primarily due to limitations involving stretching of the sciatic nerve. A total of 111 of the 175 cases was concentrated in the lengthening range of 3.1-4.0 cm. The usual limit in the child was 3.5 cm, and the maximum lengthening to be aimed for in the adult was considered 4.5 cm. The extent of lengthening was in the range of 10-15% of the length of the femoral shaft. Only one case of nonunion was observed in the children, with 11 cases in adolescents and adults (15%). In the latter group the average time for bony union was over 6 months. During the evolution of the procedure, the subperiosteal approach to the femur was gradually replaced with the musculo-periosteal decortication approach. Knee motion was either not affected or only minimally affected by onestage lengthening of the amounts described. Of 180 patients, 157 experienced no change in knee motion. There were no vascular complications in the series. Nerve complications were evident, however; there was 1 case of sciatic nerve palsy with a 4-cm lengthening and 2 cases of quadriceps palsy, both with femoral nerve involvement with partial improvement. Additional complications involved infection, early breakage of internal fixation, nonunion, and late fractures of the lengthened femur, which occurred on 8 occasions. The authors stressed that sustained flexion of the

681

knee during the procedure prevented vascular and nervous complications. They felt the procedure was warranted for moderate discrepancies, although strict observance of the maximum lengthening guidelines was essential to minimize complications. D'Aubigne and Dubousset: D'Aubigne and Dubousset (133) described a one-stage lengthening of the femur using a transverse osteotomy with lengthening performed immediately over an intramedullary rod and the lengthening stabilized by insertion of a cortical bone block. They described 16 patients in whom transient peroneal palsy occurred 3 times, delayed union or nonunion 3 times, and knee flexion contractures requiring surgical release twice. Wagner (492) has stressed that one-stage lengthenings should be restricted to the femur and that a maximum of 4-cm increase can be obtained safely. Kawamura: Kawamura and associates (265) began lengthening immediately after osteotomy, feeling that "since operative damage to soft tissues is minimized it is not necessary to delay the start of bone lengthening for a few days." Further lengthening was accomplished intermittently in 3-5 sessions under anesthesia. Their method emphasized leaving the periosteum intact. In their technique the periosteum was elevated circumferentially, following which osteotomy was performed. When the periosteum was elevated along the line of osteotomy in experimental dog models with 10% lengthening, the periosteum tore almost completely. In the second group the periosteum was incised longitudinally and detached for about 5 cm above and below the line of osteotomy. With distraction the periosteum persisted as a tube localizing the fracture hematoma, and rupture did not occur until 20% lengthening. Histologic study of 130 dog procedures was done. After tibial osteotomy, lengthening was performed gradually up to 10% over 2 - 4 weeks. At 3 weeks new bone formation from the inner surface of the periosteum was seen. Kawamura et al. felt that preservation of the periosteal tube was helpful in enhancing early bone repair. The repair response was mediated by both periosteal and nutrient vessels. Kawamura also felt that injury to the nutrient artery should be prevented if at all possible. Oblique osteotomy was used to lessen slightly the length in the gap needing repair. The nutrient artery was intact at this time frame also. In the center of the lengthened area hematoma was seen. At 3 weeks there was excellent vascular supply to the repair callus, and at 5-7 weeks there was vigorous proliferation of arterial supply from nutrient branches, although a slight avascular zone was still seen centrally. By 8 weeks most of the avascular zone had disappeared, indicating that the lengthened area had reunited. At 12 weeks there was a wellreconstituted medullary and cortical blood supply. The study showed that, after a 10% gradual lengthening following cortical osteotomy, the bone union progressed similar to that seen after a fracture. Kawamura also noted the effect of diaphyseal lengthening on the physis in dogs at either end. The effect of tibial lengthening on the longitudinal growth of the bone was

682

CHAPTER 8 ~ Lower Extremity Length Discrepancies

assessed in young dogs between 1 and 3 months of age. Osteotomy was performed and lengthening was carried out to 0, 10, 15, and 20% in 4 - 6 stages over 2 weeks. Subsequent longitudinal bone growth was then studied between 2 and 18 weeks postoperation. In lengthenings of 15 and 20%, the growth plate showed marked narrowing by both radiographic and histologic study. There was marked growth plate deformity by 3 weeks in bones lengthened between 15 and 20% with disturbance of endochondral ossification. Early closure of the epiphyseal growth plate was seen in one-third of the dogs with 10% lengthening and in all dogs lengthened 15 and 20%. In those animals lengthened 15 and 20%, the ultimate gain compared to those lengthened only 10% was negated "since the normal bone growth was lost in the larger mechanical lengthenings." Kawamura et al. concluded that lengthening should be carried out gradually and limited to about 10% of the original bone length. Technique: The distal metaphysis of the fibula was resected sub-periosteally for a distance of 2-3 cm, and when lengthenings were projected to be more than 11%, it was felt to be advisable to fix the distal fibular fragment to the tibia with a screw. Four Steinmann pins were then introduced into the tibia, two above and two below the proposed osteotomy site, and the leg was placed in the distraction apparatus cradle. A small 1-cm incision was made over the mid-tibial region with the periosteum incised longitudinally and elevated from the tibial surface. The periosteum was free circumferentially. The oblique osteotomy was outlined by several drill holes penetrating the cortices with the osteotome subsequently cutting only the cortex in an effort to spare the nutrient intramedullary vessels. The initial lengthening was carried out to a distance not exceeding 3% of the tibial length immediately. Kawamura did not find it necessary to divide the deep fascia, interosseous membrane, or intermuscular septum. Lengthening of the Achilles tendon was performed if it was tight or if tightness was anticipated. Because operative damage to soft tissues was minimal, it was not considered necessary to delay the start of lengthening to allow for soft tissue repair. Further lengthening was achieved by a small amount each day or in 3-6 sessions a few days apart with more lengthening done. It was felt that 3-6 weeks might be required to gain the desired length. When the required gain was achieved and the consolidation of callus had occurred, the limb was placed in a plaster cast incorporating the pins. A similar procedure was performed for femoral lengthening. Over a 17-year period this group performed 252 tibial and 58 femoral lengthenings in children (266). The tibial technique was standardized and 223 of 252 cases were reported as obtaining "highly satisfactory results." The 58 femoral lengthenings were also considered to be satisfactory. The vast majority of patients had good or excellent results with tibial lengthening up to 15%, with similar findings in the femur with lengthenings up to 11%. Their observations in young dog experimental lengthenings done before the age of

closure of the epiphyseal growth plates showed that gains in length were preserved in lengthenings of 10%, but in lengthenings of more than 15% subsequent growth was often markedly decreased. Fifty-seven patients were studied in terms of subsequent longitudinal growth following tibial lengthening. Of the total of 57 patients, 35 showed a decrease in expected growth. Kawamura stressed that complications in this regard could be avoided with lengthenings kept to within 10-15% of overall length. The lengthening was to be carried out gradually in 3-6 stages over 3-6 weeks, and the initial lengthening performed immediately should not be more than 3% of the bone length. Complications: Kawamura listed in detail the possible complications with limb lengthening. (1) Complications due to overstretching. These included angular deformity, arthritis, stiff joints, loss of muscle power, stretch paralysis, and neuralgia. (2) Complications due to interference of blood supply. These included delayed union and nonunion, bone necrosis, and circulatory disturbances distal to the operative site. (3) Complications due to inadequate fixation of fragments. These involved overlying skin slough, fragment protrusion, anterior bone angulation, late fracture, and pin site infection, which could lead to osteomyelitis. (4) Direct operative complications included fracture of the osteotomized shaft, pin pull-out, hypertension and shock. b. Improved Results Due to More Rigid Fixators. Judet Technique: The Judet technique (259), developed in France in 1969, provided much more rigid stabilization, thus enhancing comfort in association with lengthening. It involved a lengthening by gradual distraction in association with an oblique osteotomy and decortication to hasten bone repair, but use of neither graft nor internal metallic stabilization. The external fixator used 5-mm-diameter pins with three or four placed in the proximal fragment and an equal number distally. A unilateral distractor was attached. The procedure was designed for use in the tibia. The distal fibula was stabilized by a transfixion screw to the distal tibia to allow for lengthening of both bones simultaneously and maintenance of the orientation of the ankle mortice. The oblique osteotomy of the tibia was made in the frontal plane and was of the greatest length possible, with the average amount being 10.5 cm. A plaster splint then immobilized the knee in extension and the foot at a fight angle at the termination of the operative phase. The limb was maintained in the cast during the period of the lengthening, which was performed at a rate of 1.5 mm per day. Each day hip, knee, and ankle range of motion exercises were done. When the lengthening was completed, the lower extremity was placed in a long leg cast, which was then replaced at varying times with a lighter plastic brace. At 6 months the distractor and fibular-tibial pins were removed with the tibia continuing to be protected in a brace for an additional 6 months. Pouliquen et al. (391) reviewed 108 tibial lengthenings performed by the Judet method. Of these, the large majority, 79, were due to poliomyelitis. The average lengthening ob-

SECTION IX ~ Management o f Lower Extremity Length Discrepancies

tained was 4.37 cm with a range from 2.6 to 6.0 cm. The average gain was in the range of 16%. Bone union in the most favorable cases was obtained at an average of 4 months and a secondary bone graft was needed in only 5 instances. There were 17 instances of neurological impairment, although only 2 showed persistent anesthesia and 1 permanent paralysis of the extensor hallucis longus muscle. Postlengthening fractures occurred in 9 patients, 5 of which were non-displaced and treated only with a simple cast with 4 requiring additional surgical intervention. Overall, 10 secondary bone grafts were required, 5 for a delay in primary healing and 5 for pseudarthrosis or fractures. The paper also performed a detailed comparison with tibial lengthening studies from 7 other series. Excellent reviews of the more recent limb lengthening techniques have been provided by Caton (101) and Paley (361). Some of the less well-known European precursor techniques have been reviewed by Wiedemann (505). Wagner Technique: In 1978 Wagner (493) reported his technique, which had major improvements from previous approaches and restimulated widespread interest yet again in lengthening procedures. His unilateral fixator was structurally very stable and allowed patients to be ambulatory with crutches immediately after the surgery and throughout the lengthening procedure. This alone was a major advance because virtually all previous lengthenings required prolonged bed rest, at least during the lengthening and early consolidation phases. Diaphyseal lengthening began immediately with approximately 1 cm of distraction performed in the operating room at the time of instrumentation (described by Wagner as lengthening "until stabilization is achieved by soft tissue tension"). Once daily lengthening of 1.6 mm was performed. Once the desired length was reached, a second operative procedure was performed in all patients involving the application of a long side plate, autogenous iliac crest bone grafting, and removal of the external fixator (Fig. 22). The advantage of the second operative approach was that rigid internal stabilization was achieved immediately and rehabilitation was enhanced because range of motion exercises could be performed more easily once transfixing pins were removed. The Wagner technique represented a significant advance over the Anderson and previous methods due to the fact that the patient could remain ambulatory and pain was considerably diminished due to the increased stability of the external fixator. By the time Wagner developed his approach, the etiology of limb length discrepancy had changed with fewer children suffering from poliomyelitis. Of the 58 patients in his initial paper only 10 had poliomyelitis, with the largest group involving congenital short femur and the next largest epiphyseal fracture. As a result of these differing etiologies, Wagner found that preoperative consideration of the entire limb was of increasing importance. Of particular importance prior to lengthening procedures were the correction of joint contractures and surgery to correct acetabular dysplasia and angular

683

FIGURE 22 Illustrations of the Wagner lengthening technique are shown. The end result followinglengthening of a femur is shown. Following attainmentof the desired length, operationis performedat which time a Wagner plate and bone graft are applied. The plate designed by Wagnerfor this procedure has no holes spanning the lengthenedgap as these are prone to breakage if a regular AO plate is used. Althoughadditional surgeries are required with this technique, there is one major advantage to it in that rehabilitation in terms of range of motion of the adjacentjoints is markedly quickened and improved compared with subsequent techniques in which the distraction device is left on until there is full healing. Anteroposterior (A) and lateral (B) projectionsat healing.

bone deformities. He detailed the advantages and disadvantages of operative limb lengthening and the various shortening procedures. Carlioz et al. (94) reported on their first 30 cases of the Wagner lengthening involving 15 femoral lengthenings, 7 tibial lengthenings, and 8 cases associated with the correction of angular deformities or pseudarthrosis. In the straightforward femoral lengthenings the gain ranged between 4.0

684

CHAPTER

8

~

Lower Extremity Length Discrepancies

and 7.0 cm. Among the complications were 3 subsequent fractures. The percentage of lengthening was in the range of 15% with no cases beyond 20%. In the straightforward tibial lengthenings, the range of correction was between 4.0 and 5.8 mm with an average percent increase of 15% with one case being 26%. Early experience of this group was quite favorable. A few years later their group reported on 48 femoral lengthenings using the Wagner technique (Blachier et al.) (53). The mean lengthening obtained was 5.2 cm with the most extensive at 11.2 cm. In terms of percent gained, lengthening values were 15.6% of initial bone length. On 6 occasions the lengthening was greater than 20% with a maximum of 24%. Many complications were noted varying from minor to much more serious in importance. Superficial infection was frequent but easily managed. In 22 of the 48 cases, varus deformation occurred during lengthening but was easily corrected in most at the time of plate application and grafting. In some instances, however, the varus was not corrected even at the time of plate application. Subluxation at the hip and knee was seen on occasion as were contractures without subluxation at these two joints. Five patients had neurological complications; in 4 there was full and rapid recovery of the anterior foot dorsiflexors, but in 1 paralysis of the peroneal and tibialis posterior nerves was complete and recovering only slowly at the time of publication. There were 12 deep infections, which on occasion compromised the long-term result. On occasion fracture or pseudoarthrosis at the site of infection further complicated treatment. Caton et al. (102) also reported early favorable results with the Wagner technique in 33 lengthenings, 20 of which were in the femur and 13 in the tibia. The mean lengthening obtained was 5.35 cm, and although many complications were seen 70% of cases had none. The mean percent lengthening was in the range of 16.2% with a mean lengthening of 4.8 cm for the tibias and 5.7 cm for the femurs. The use of more rigid fixators such as those of Wagner or Judet became popular in the 1970s. Some series are reported in which the fixators were used, although differing techniques from those initially described by the developers were appropriated. A study by Rigault et al. (403) assessed 36 femoral lengthenings in which 21 used the Judet distractor and 15 the Wagner. In most instances, an oblique femoral osteotomy was used, there was an initial 5-10 mm of lengthening at the time of surgery, and progressive lengthening was then performed at a rate of 1-2 mm per day. The bone was then allowed to heal without application of the graft and side plate as recommended by Wagner. The mean lengthening gained in the 36 cases was 5.2 cm, which was an 18.3% increase in length. The series was complicated by a high rate of fracture with 9 of 36 or 25%. Those having the oblique osteotomy had a mean gain of 5.6 cm, whereas those having more transverse osteotomies had a mean gain of 4.7 cm. Bone graft was performed in 6 of the 36 cases because of delayed healing. Rigault et al. (404) also described 48 tibial lengthenings using either the Judet or Wagner apparatus along with a long oblique osteotomy, decortication, and mul-

tiple soft tissue releases to stabilize the knee and minimize equinus deformities at the ankle. In general, the distractor was left in place until radiographic bone consolidation was evident, following which a brace or walking cast was applied to protect the limb during the return to full weight bearing. The mean lengthening was 4.2 cm or 17.5% of preoperative bone length. Fractures were seen in 10%. In the 10 more complicated cases the mean lengthening was 5.5 cm or 16.5%. Many groups reported their results with the Wagner technique. Wagner (493) reported on 58 femoral lengthenings in patients below the age of 17 years in which the average gain in length was 6.8 cm. Bjerkreim and Helium (52) reported an average lengthening of 5.8 cm in the femur, Aldeghiri et al. (14) reported 4.9 cm in the femur and 4.0 cm in the tibia, and Paterson et al. (373) reported 5.8 cm in the femur and 5.2 cm in the tibia. Stephens (462) reported an average femoral lengthening of 5.7 cm in 18 Wagner procedures and an average tibial lengthening of 5.6 cm in 7 Wagner procedures. Osterman and Merikanto (360) reported on a mean increase with 26 tibial lengthenings of 4.1 cm and with 9 femoral lengthenings of 4.9 cm. Mahlis and Bowen (312) reported 27 femoral lengthening with a mean increase of 6 cm (range = 3-9.5 cm) (mean percent increase 17.6%, range = 7-36%) and 11 tibial lengthenings with a mean increase of 6.1 cm (range = 4.5-9 cm) (mean percent increase 20%, range = 10-32%). Ahmadi et al. (9) compared results with 50 Anderson, 40 Rezaian, and 51 Wagner lengthenings all performed for poliomyelitis. The mean gain was 4.8 cm and complications showed 6 refractures (4%) as well as others commonly seen, but in relatively low rates. Results in the three techniques were similar; the Wagner was favored for relative ease of use. Complications with the Wagner Procedure: The study of and literature on the complications of limb lengthening surgery are somewhat unique in relation to the rest of the orthopedic literature, primarily because of the intrinsic nature of the difficulties with this procedure. A characteristic approach is to divide what would generally be considered complications into (1) problems, which are felt to be intrinsic, cannot be avoided, and include such disorders as delayed union and nonunion, pin track infections, and transient restriction and motion of adjacent joints, and (2) complications, which are extrinsic and should be avoided, including infection, nerve damage, fractures, and subluxations. Others use the terms minor and major complications, an approach that is preferred in this chapter. DeBastiani et al. (137) reported a 26% complication rate with the Wagner method. Hood and Riseborough (236) noted 37 complications in 40 procedures, whereas Wagner himself noted a complication rate of 45% in 58 patients. Luke et al. (305) specifically described fractures after the Wagner limb lengthening procedure. In a series of 27 cases, there were 10 fractures following lengthening in 8 patients, 6 of whom were spontaneous and 4 traumatic. The fracture occurred through the lengthened area after plate and screw removal in 8 patients, through a proximal screw hole and

SECTION IX 9 Management of Lower Extremity Length Discrepancies

plate in 1 patient, and through an external fixator pin hole with hardware intact in another. Seven fatigue fractures occurred after plate removal in Wagner's series. Hood and Riseborough (236) reported 4 fractures after 40 Wagner lengthenings; Mosca and Mosely (339) reported 63 Wagner lengthenings with 16 subsequent fractures; and Chandler et al. (110) reported 21 lengthenings with 2 subsequent fractures. The preceding 4 studies thus reported a total of 182 Wagner lengthenings with 29 subsequent fractures (16%), whereas the most recent report of Luke noted a 37% rate (10 of 27). The amounts lengthened in both tibia and femur were not remarkable in those suffering either spontaneous or traumatic fractures with the tibial percent lengthening at approximately 16% and the femoral 20%. Some attempted to minimize the fracture sequelae by a four-stage procedure involving the osteotomy, plating and bone grafting of the lengthened area, and then, instead of removing the hardware at one stage, doing a two-stage procedure with the removal of alternate screws and partial loosening of the plate followed several months later by removal of all remaining hardware. Osterman and Merikanto (359) noted that a major complication was late femoral fracture, which occurred in 6 of the 26 instances although there were no tibial fractures. There was 1 hip dislocation, 1 talar deformation, 1 peroneal nerve entrapment, and 1 infection, which delayed bone union. The authors cautioned about early removal of the stabilizing plates and also indicated the need for bone lengthening to be performed by well-trained teams with great experience. Karger et al. (262) reported primarily on Wagner lengthenings involving 51 femurs and 18 tibias. Complications were more marked when the lengthening exceeded 25% of initial bone length. They were also much higher in the femur than in the tibia. Some modifications to the Wagner procedure were incorporated; whereas 44 patients had the entire sequence described by Wagner himself, 18 had only the stage one procedure but did not have grafting and plating because the authors felt initially that healing would be appropriate. In 51 femurs lengthened by the Wagner technique, the mean lengthening achieved was 7.57 cm, which represented a 25% increase, whereas in 18 tibias, the mean length achieved was 6.07 cm, accounting for a 23% increase in length. The complications were divided into groups referred to as problems, obstacles, minor sequelae, and major sequelae. They were commonly seen in the femoral lengthenings but somewhat less frequently in the tibial. Fractures were common, occurring in 18 of 51 femoral procedures and 3 of 18 tibial. Angulation was noted in 25 in the femoral group, all of which were varus with a mean deformity of 25 ~ Some had flexion deformities as well. There were 10 of 18 angular deformities in group 2, all of which were valgus with a mean of 16~ Pin tract infections and deep infections were noted as were joint restrictions of motion, including 12 transient knee subluxations. Salai et al. (418) described three hip subluxations during femoral lengthening with the Wagner technique. Each of

685

the patients had some predisposing hip and acetabular abnormality, indicating again the importance of assessing the hip prior to and during the course of lengthening and fully correcting malposition surgically prior to the lengthening procedure. c. Distraction Osteogenesis. Although many good and excellent outcomes were obtained with the Anderson and Wagner techniques, many problems and complications occurred. The Anderson technique frequently required supplementary bone grafting, whereas the Wagner approach required at least three surgical interventions. Patient reviews reported numerous additional problems. Work continued with different techniques to improve results. Clinical Techniques: Techniques developed by Ilizarov (243-247) in Russia, Monticelli and Spinelli in Italy (332), DeBastiani, Aldegheri, and associates in Italy (137), and Canadell and associates (88-90) in Spain have allowed for lengthening and bone repair such that bone grafting and plating were infrequently required. These techniques, described collectively under the term distraction osteogenesis or callotasis, are dependent on four principles, the first two of which are truly integral to the effectiveness: (1) delay prior to the initiation of bone lengthening for 7-10 days to allow early repair, or callus formation, to occur, (2) gradual lengthening of 1 mm per day performed at four separate times to minimize damage to newly formed vessels and repair tissue, (3) corticotomy leaving the medullary vasculature intact or at least minimally disturbed, and (4) maintenance of an intact periosteum. There has been somewhat of a polarization between schools performing the distraction osteogenesis technique based on the type of external fixator used. The Ilizarov technique uses circular and hemispheric external fixators stabilized by multiple narrow wires, the Monticelli-Spinelli method uses a single circular fixator to hold epiphyseal wires and two diaphyseal level tings held together by longitudinal rods between which the metaphyseal corticotomy was performed, and the Orthofix and Monotube methods use a unilateral fixator with two upper and two lower 5-mm-wide pins (Fig. 23A). The early results are promising, but even with these techniques problems similar to those reported with earlier methods are being encountered. Healing time still remains prolonged and associated with osteopenia, muscle atrophy, and joint stiffness. Canadell has summarized the extensive experience of his group from Pamplona, Spain. A detailed assessment of 93 lengthenings performed over a 3-year period using the principles of distraction osteogenesis and a unilateral fixator has been reported along with conclusions realized over a 25year period involving more than 800 lengthenings (88-90). The 93 lengthenings involved 27 with unilateral discrepancy due to pathologic conditions and 34 patients having bilateral lengthenings due to symmetrical shortening with skeletal dysplasia disorders. The average lengthening obtained was 8.37 cm with a complication rate of 2.1 per lengthening. The repair index was the same for both femur and tibia, but humeral lengthenings healed in a much quicker fashion. There

686

CHAPTER 8 9

Lower Extremity Len~tth Discrepancies

F I G U R E 23 (A) Examples of the distraction osteogenesis technique are shown. (Ai) Anteroposterior film of tibia and fibula shows distraction gap with early repair at 6 weeks using the Orthofix apparatus. Lengthening was begun 1 week postsurgery. Note that early bone regenerated is more dense adjacent to the upper and lower persisting bone, with the central part of the gap region showing lesser ossification and radiodensity. (Aii) Results of a femoral lengthening are shown using the distraction osteogenesis principle and the Orthofix apparatus. Anteroposterior films of the femur are shown from the time of initial osteotomy and insertion of apparatus to complete healing at 9 months. Note the close apposition of the cortices at the time of the initial procedure. Lengthening was begun at 1 week. At 12 days, there is just the beginning trace of new bone formation lateral to the cortical regions. At 1 and 2 months, new bone clearly is forming in the gap and slightly lateral and medial to it along with increasing length noted. A central gap region is seen, in particular at 2 months, with more bone formed adjacent to the cortices above and below. Progressive new bone formation is seen at

SECTION IX ~ Management o f Lower Extremity Length Discrepancies

FIGURE 23 (continued)

687

688

CHAPTER 8 ~

Lower Extremity Length Discrepancies

F I G U R E 23 (continued) 4, 5, and 6 months. Note the marrow cavity reconstitution at 6 months particularly in comparison with the appearance at 5 months when bone across the gap was relatively uniform. Final cortical and marrow reconstitution at 9 months of age is seen. The patient has continued with normal function and increased bone density with a full range of motion at the hip and knee. (B) Classification of the shape and structure of the distraction callus by Hamanishi et al. [from (217), with permission.] The external and straight variants are ideal. (C) Radiographs of rabbit tibia undergoing distraction. The films are from 7 days (Ci), 13 days (Cii), and 22 days (Ciii) following surgery. (D) Specimen radiographs allow for better demonstration of bone repair at varying times following surgery and distraction. Distraction began at the time of surgery. The specimen radiographs in (Di) were performed 7 days after surgery and initiation of distraction (AP and lateral); (Dii) 14 days (AP and lateral); (Diii) 19 days (AP and lateral); (Div) 32 days (AP and lateral); and (Dv) 44 days postsurgery. Note the excellent cortical reconstitution and marrow reformation in the final lateral radiograph. (E) Specimen photographs obtained after decalcification and hemisectioning the bone repair gap and the adjacent cortices prior to histologic sectioning. Photograph from a rabbit sacrificed at 46 days shows the excellent reconstitution of the cortex as well as marrow continuity. The lengthened region, which totaled 9 mm or 9% of the initial bone length, is indicated by the darker stained more vascular marrow centrally. (F) The lengthened tibia is shown at sacrifice after the fixator had been removed and all surrounding soft tissues had been dissected free. This had been lengthened by 11 mm, an amount that corresponded to 10.7% of its initial length. Lengthening proceeded for 14 days and the fixator was left on for an additional 30 days, with sacrifice 44 days postsurgery.

was only a slight overall difference in repair between metaphyseal and diaphyseal osteotomies and those involving distraction epiphyseolysis. The diaphyseal and metaphyseal lengthenings provided slightly greater increases in length.

The rate of distraction was somewhat slower in the epiphyseal distraction procedures, although the rate of healing was somewhat quicker once length had been obtained. Canadell also noted that repeat lengthenings could be performed read-

SECTION IX ~ Management of Lower Extremity Length Discrepancies

689

range 4-15 cm. More recent results from the European and ily on the same bone generally with an interval of 2 years North American literature have been reported as the Ilizarov between each procedure. In fact, the number of complications technique became widely adopted. Paley has reviewed the was greater during the course of the first lengthening than complications with the technique in detail (362). with the second. Two femurs had been lengthened 5 times Stanitski et al. (460) reported on 62 tibial lengthenings and 2 others had been lengthened 3 times. In his patients, using the Ilizarov technique with the average lengthening of lengthening was resorted to much more readily than is the 7.5 cm (range = 3.5-12 cm) representing the equivalent of practice in North America with discrepancies of 3 cm or a 32% average overall increase. Stanitski et al. (458) also more being candidates for lengthening procedures. In terms reported the results in 36 femoral lengthenings using the of surgical technique, it was difficult to prevent traumatic Ilizarov technique, with the average lengthening being 8.3 cm rupture of the medullary cavity and in only 30% of cases (range 3.5-12 cm) and a lengthening index of 0.74, months were the contents intact. Protection of the periosteum was of treatment/cm of lengthening. considered essential. The ideal rate of distraction appeared Franke et al. (165) reported good results with the Ilizarov to be 1 mm per day at four evenly spaced time periods with technique whether they used distraction epiphysiolysis or 0.25 mm lengthened at each time slot. Lengthening was conmetaphyseal corticotomy. In the distraction epiphysiolysis traindicated after the age of 30 years, and the ideal age to procedure, they lengthened 22 tibias with an average lengthconduct the lengthening was between 8 and 12 years at which ening of 8.25 cm (range = 4-18 cm) and 30 tibias using time the osteogenic capacity was greatest. In those younger the metaphyseal-diaphyseal corticotomy with an average of than 8 years of age, repair was often so rapid that appropriate lengthening could not be achieved. Considering all of length7.9 cm of lengthening (range = 4-15 cm). Their report repenings done the index of maturation was 1.16 months per cm. resents one of the most detailed in terms of assessing the Distraction was begun at different times after percutaneous average time of distraction, the average time to removal of metaphyseal or diaphyseal osteotomy with application of a the apparatus, and the average time to full weight bearing. monolateral fixator dependent on the age of the patient using They also subdivided results into the amount of lengthena rough guideline of 1 day of delay per year of age. Thus, a ing from 4 to 5 cm in one group, 5.5-9.5 cm in another, and child of 8 years of age had lengthening started 8 days after ml 10 cm or greater in the next. initial surgery, and for someone 15 years of age Canadell Monticelli and Spinelli (332) reported 43 cases of metaet al. waited 15 days. The quickest healing rate was most physeal lengthening using the distraction osteogenesis techfavorable in those with short stature conditions, averaging nique with their own fixator, with a mean gain of 7 cm and a 0.8 months per cm, and it was much longer in those with range from 4 to 10 cm. DeBastiani and colleagues (137), length discrepancies in the range of 1.5 months per cm. The using open corticotomy and callus distraction with their greatest lengthenings were also obtained in those with short monolateral fixator (Orthofix), performed limb lengthenings stature with an average of 11.2 cm per segment lengthened. on 100 segments with a lengthening index of 1.2 months for In 8 patients with a chondrodysplasia, Canadell reported exthe femur, 1.4 months for the tibia, and 0.8 months for the tensive lower extremity lengthenings gaining 23.2 cm dihumerus. The average amount of lengthening obtained in the vided between 12.35 cm in the femur and 10.85 cm in the femur in patients with achondroplasia was 7.8 cm (range = tibia. The ideal site for osteotomy was the metaphysis. More 5.5-12 cm), representing a mean 26% increase, whereas in sensitive angiographic studies indicated that in 90% of diathe tibia the average amount of lengthening obtained was physeal osteotomies the medullary vessels were damaged, 7.8 cm (range = 6-10.5 cm), representing a 36% increase. which was one of the reasons the metaphyseal site was faIn patients with limb length discrepancy from congenital and vored because the vascularity was more diffuse and richer acquired disorders, the average amount of lengthening in in that region. On both a clinical and experimental basis, it the femur was 4.7 cm (range = 3-9 cm, 11%), in the tibia was vastly more important to respect the continuity of the 4.7 cm (range = 3-9 cm, 17%), and in the humerus 7.5 cm periosteum than the medullary circulation. The group also (range = 7-8 cm, 34%). Aldegheri et al. (14) reported on strongly supported the importance of dynamization in en270 femoral and tibial lengthenings using the Orthofix callohancing repair. tasis method. Ninety-five patients had limb length inequality Paley (361) summarized the work of Ilizarov and the and 45 had achondroplasia-hypochondroplasia. The average Kurgan school, who worked extensively on distraction length increase was 6.6 cm or 24.6% of initial length. The osteogenesis from 1950 on. He indicated that more than mean healing index was 39 and the complication rate 13.3%. 1000 publications from their institute alone had described With the passage of time following the introduction of the various approaches to the technique, with most published in callotasis technique, long-term studies of relatively large Russian. Major studies were reported by Ilizarov of 237 femnumbers of patients have begun to accumulate and, perhaps oral lengthenings with an average lengthening of 7.4 cm, not surprisingly, show a pattern of findings similar to other range 3-15 cm, and complications listed in 12 patients lengthening techniques. Glorion et al. (181, 182) reviewed (5.6%). Ilizarov also published results of 217 tibial lengthen79 cases of femoral lengthening by the callotasis technique, ings in both children and adults with the mean gain of 7 cm, with 9 patients having the Judet apparatus and 70 the Orthofix.

690

CHAPTER 8 9 Lower Extremity Length Discrepancies

They concluded that the incidence of complications did not seem to be less than that encountered with previous methods of lengthening. The complications were the same in terms of nature and extent whether the callotasis technique had been performed using the Ilizarov or the Orthofix technique. The averaging lengthening achieved was 5.2 cm (range = 3.58.5 cm), which represented a mean increase in femoral length of 17.7%. In the 79 cases, however, there were 87 complications, a rate of 110%, although in general several complications were often encountered during one procedure. Glorion et al. noted that 30% of the lengthenings were performed without any complication and 62% with relatively mild complications such that additional surgery or anesthetic procedures were not needed. The healing index was 39.6 days per cm, which was comparable to other callotasis studies. Dynamization was considered to be an important adjunct to the healing process. Fractures continue to be seen fairly often even after distraction osteogenesis lengthenings. Danziger et al. (131) reported 9 femoral fractures after 18 Ilizarov femoral lengthenings but no tibial fractures with 8 tibial lengthenings. Glorion et al. (182), using primarily the Orthofix technique, in 61 lengthenings had fractures in 6 instances. Their technique involved percutaneous osteotomy, referred to as compactotomy when at the level of the metaphysis or corticotomy when done within diaphyseal bone. The cortex alone was cut with a small 5-mm osteotome with care taken not to enter the medullary cavity in an effort to spare the nutrient artery and medullary circulation. The periosteum remained intact. The remainder of the cortex was then broken either by rotating the osteotome to distract the bone ends or by rotating the fixator pins. Distraction did not begin for 7 days postsurgery and was at a slow rate of 0.25 mm of elongation four times daily. The average time to healing is still extensive, and the lengthening index that has been established is months per centimeter of lengthening, with most studies showing an index around 1 (if 5 cm are lengthened, the time to healing is 5 months). Suzuki et al. (467) studied 26 femoral lengthenings using the Orthofix callotasis technique. Lengthening began 1 week postsurgery with the rate of distraction 0.25 mm every 6 hr. The mean amount of lengthening obtained was 5.0 cm with a range from 2.0 to 7.5 cm. The study particularly assessed dislocation and subluxation of the femoral head during the lengthening procedure. One group of 14 hips with a CE angle of greater than 20 ~ at the start of lengthening showed no deterioration in position with the lengthening, whereas the other group of 12 hips with an angle of 20 ~ or less showed deterioration of femoral head position in 5 of the 12 hips. One developed a complete dislocation and the other 4 subluxed, showing a decrease in the CE angle. Four of the 5 had a history of congenital dislocation of the hip and the other had multiple epiphyseal dysplasia. The authors recommended that, in cases in which the CE angle was 20 ~ or less preoperatively, bone procedures such as innominate osteotomy should precede the femoral lengthening.

Limb lengthenings were performed increasingly with the callotasis technique for those with symmetrical limb shortening due to skeletal dysplasias (17, 394, 457); these will be discussed in greater detail in Chapter 9. Comparison of Techniques within the Same Centers: Many studies from centers in which large numbers of lengthening procedures were performed began to present data comparing differing techniques. An excellent review by Pouliquen et al. (393) compared femoral lengthenings in 82 cases divided between six techniques, five of which were used relatively frequently. These involved the one-stage lengthening (14 cases), Judet technique (20), Wagner (13), a transverse osteotomy and graft technique (11), and callotasis (20). The authors concluded that the callotasis technique was the best because there were no serious complications out of 20 cases. The amounts lengthened, however, were similar with the several techniques with the exception of the one-stage lengthening, which was reserved for relatively smaller discrepancies. In that group the average lengthening was 3.6 cm or 7.8% of bone length. The other four techniques had lengthenings on average ranging from 4.6 to 5.5 cm or 12.7 to 15.5%. In the one-stage lengthening, a Judet distractor was applied followed by performance of an oblique osteotomy, the application of temporary cerclage wires, lengthening by the distraction technique intraoperatively with the knee flexed, and osteosynthesis with a side plate once the desired amount of lengthening had been achieved. The Judet technique involved a unilateral distractor similar to the Wagner but with four heavy pins below and four above the osteotomy site. The lengthening was at a rate of 1.5 mm per day. Once lengthening had been achieved, the distractor was left in place while healing was allowed to occur. Walking began again under protection of a brace, which incorporated both the pelvis and the external fixator, and was generally maintained until 12 months after initial surgery. The complication rate in the callotasis technique was extremely low at 5%, whereas it ranged from 27 to 35% in the four other major approaches. The study also reviewed the literature from the 1970s and 1980s in relation to each of the major approaches then used involving the one-stage lengthening, the Judet lengthening, the transverse plus graft lengthening, Wagner lengthening, Ilizarov lengthening, and callotasis lengthening. The one-stage lengthenings assessed involved 229 cases with the lengthenings achieved ranging between 3.2 and 3.7 cm and a complication rate between 10.5 and 35%. Two series of Judet lengthenings with an oblique osteotomy were reviewed involving 56 cases with a mean lengthening of 5.2 cm and a complication rate between 25 and 41%. There were 11 cases of the transverse osteotomy plus bone graft group also with a 5.5-cm mean lengthening and a 27% complication rate, 120 cases of the Wagner lengthening with a range of 5.2-6.8 cm increase and a complication rate of 12.5-31%, Ilizarov lengthenings in 21 cases with a mean lengthening of 5.06.1 cm and a complication rate of 6-25%, and the most favorable group was the callotasis technique involving 78 cases

SECTION IX ~ Management of Lower Extremity Length Discrepancies

with a mean lengthening of 4.7 cm and only a 6% complication rate. This report remains one of the best detailing the techniques of the various procedures and providing a de, tailed review of results both from the literature and within the 82 cases of varying techniques from one unit. Faber et al. (156) reviewed several limb lengthening procedures divided between the Wagner approach, the MonticelliSpinelli metaphyseal corticotomy and distraction, and the Monticelli-Spinelli distraction physiolysis. Complications with each of the three procedures were reviewed in detail and were frequent. They concluded that the Wagner diaphyseal osteotomy involved more cases of delayed union, late fracture, and axial deviation. The list of complications in 17 femoral lengthenings included 3 with delayed union, 8 axial deviations, 3 late fractures, 2 losses of length, and 1 premature consolidation. As far as the joints were concerned, there was restriction of motion at the hip in 2, knee in 14, and ankle in 2, with 1 hip subluxation, 2 proximal tibial subluxations, and 1 patellar subluxation. There was 1 deep wound infection and 10 pin track infections. One patient had loss of muscle power, and complications leading to discontinuation of the lengthening procedure occurred in 1 with serious restriction of hip motion and in another with hip dislocation. Only 6 tibial procedures were performed with complications involving 1 nonunion, 3 axial deviations, 4 late fractures, and 2 premature consolidations. As far as the joints were concerned, there was restriction of motion at the knee in 1 and at the ankle in 3 and also 1 pin tract infection. A review of 100 lower limb lengthenings from Brazil assessed 25 tibial lengthenings by the Anderson technique, 45 femoral lengthenings by the Wagner technique, and 16 femoral and 14 tibial lengthenings by the Ilizarov technique. Once again, an assessment of the amount of lengthening achieved and the average healing time showed very little difference between the techniques. The complications seen widely in limb lengthening were present in each, although certain types of complication tended to occur with certain types of lengthenings. As with previous studies, good results could be achieved with each and considerations such as comfort and relative ease of the procedure for both the patient and the surgical team would play a primary role in choice. In the Anderson group the average lengthening achieved was 4.2 cm, range = 3-6 cm, and the average healing time was 197 days with a lengthening index of 1.72 months. In the Wagner group the femoral lengthenings had an average of 4.6 cm, range = 1-12.5 cm, with an average healing time of 185 days with a lengthening index of 1.32 months in a subsection having percutaneous osteotomy and a healing time of 166 days with a lengthening index of 1.23 months in a subgroup having corticotomy. In the Ilizarov group the average femoral lengthening was 4.7 cm, range = 1-7.5 cm, with an average healing time of 186 days and a lengthening index of 1.31 months. For the Ilizarov tibial lengthenings the average was 4.5 cm, range = 1-7.5 cm, with a healing time of 184 days and a lengthening index of 1.35 months. With the Anderson method the most common complication was

691

delayed union, with the Wagner technique the most common complications related to bone healing and subsequent fracture, and with the Ilizarov method the most common complication was incomplete corticotomy. Effect of Lengthening on Muscle Strength, Articular Cartilage Structure, and Nerve Function: Three other areas of concern with lower extremity lengthening have been subject to more refined analyses than those available by clinical assessment alone. These involve the effects of lengthening on (1) muscle strength, (2) articular cartilage structure, and (3) nerve function. (1) Muscle strength before and after femoral and tibial lengthening. Maffulli and Fixsen (307) studied quadriceps strength in those with congenital short femur before and at the termination of femoral lengthening. The Orthofix technique was used. Seven patients had an average lengthening of 7.1 cm, a 23.5% lengthening of the congenital short femur. The normal side was stronger initially than the shortened side. The differences in strength, however, between the two sides did not meaningfully change with a difference of 15.7% at the beginning of the procedure and 13.1% at the end of the study. When the relationship of the knee extensor strength to the muscle and bone area of the mid-thigh was calculated, there was no change postoperatively in the normal side but a slight increase in the extensor strength in the operated side. That report is similar to the clinical impression that, once lengthening has been completed effectively and range of joint motion regained, the muscle strength is maintained. Lee et al. (292) studied changes in the gastrocnemius muscle in relation to the percentage of lengthening in the rabbit tibia using the callotasis technique. Lengthenings of 10, 20, and 30% were performed assessing 25 rabbits in each group or 75 overall. The study was based on histopathologic and morphometric assessments of the muscle. Biopsies were obtained from the medial gastrocnemius of both hind legs immediately prior to sacrifice at termination of the lengthening procedure. As compared with the control side, the lengthened side had substantial differences with fiber size variation noted in all three lengthening groups. Significant differences, however, in internalization of muscle nuclei and endomysial fibrosis, which represent more definitive changes, were observed only after 20 and 30% lengthenings. There were no differences in degeneration or regeneration among any of the lengthening groups. The fiber size variation was thought to be due to an increased number of atrophic fibers rather than due to the presence of hypertrophied fibers. As the lengthening percentage increased, the histopathologic scores of each parameter of the lengthened side showed a linearly increasing trend reflecting an increasing severity of the histopathologic changes. Because the rabbits were sacrificed at the termination of lengthening, no information was available as to whether some of these changes might have regressed. Kawamura et al. (265) also noted no significant histochemical or electromyographic changes in those lengthened up to 10% of their initial bone length. Carroll et al.

692

CHAPTER 8 9

Lower Extremity Length Discrepancies

(98), using a relatively rapid distraction model in sheep, noted that tibial lengthening greater than 11% of initial length consistently produced irreversible changes in the gastrocnemius and flexor digitorum profundus muscles, including loss of myofibrils, central migration of nuclei, and irregular shapes and sizes of myofibrils. Other observers have reported that up until approximately 20% lengthening the muscle actually lengthens throughout its entire extent, after which lengthening appears to be localized at the osteotomy site and is more associated with fibrosis, which itself would tend to weaken the muscle. Lee et al. showed, therefore, that lengthenings up to 10% have little histopathologic change in muscle and that progressively greater lengthenings to 20% and then 30% led to more conspicuous changes. They also felt that, once lengthening extended beyond 20%, irreversible changes were more likely to occur in the muscle itself. Carroll et al. felt that the changes in the muscle were primary rather than being secondary to nerve stretch phenomena. They also noted histologic changes in the articular cartilage of the tibiotarsal joint at both gross and histologic levels, with tibial lengthenings including fibrillation, empty lacunae, and matrix degeneration. The tibia had been rendered short initially by a proximal tibial epiphyseal arrest using the Phemister technique, after which a lengthening was performed in 16. Carroll et al. concluded that lengthening of the tibia by more than 11% consistently produced muscle changes in the leg and cartilage damage in the ankle joint. (2) Effects of lengthening on articular cartilage of adjacent joints. Stanitski et al. (459) also documented the effect of femoral lengthening on the articular cartilage. They felt that 30% femoral lengthening causes reproducible knee cartilage injury, which was evident by actual loss of cartilage substance and fibrillation. Application of a modified Ilizarov apparatus to the femur and tibia with coaxial hinges at the knee followed by 30% lengthening resulted in less severe damage than when the femur was lengthened independently, suggesting that there was a protective effect of the femoraltibial apparatus on joint compression. Nakamura et al. (346) studied knee articular cartilage changes in association with limb lengthening by the callotasis technique in 18 rabbits with a distraction rate of 1 mm per day. Distraction began the day after operation. On the fight side the frequency was 0.5-mm increments every 12 hr, whereas on the left it was controlled automatically leading to 120 smaller incremental increases, which averaged 0.0083 mm every 12 min. Histologic changes were much less in the multistep autodistractor technique side than in the side undergoing twice daily 0.5-mm increments. This study was also divided into assessments with length increases of 10, 20, and 30%. The incidence of cartilage degeneration on the 2-step side was 2/5, 5/6, and 6/7 at the 10, 20, and 30% length increases, respectively, whereas on the 120-step side it was much less at 0/5, 1/5, and 1/7 at the corresponding length increases. The numbers 5, 6, and 7 refer to the number of animals in each group.

(3) Nerve changes due to the lengthening procedure have been assessed. It has been recognized for several decades that nerve stretching in association with limb lengthening can lead to sensory and motor nerve deficits. In most patients, however, even with significant lengthening motor and sensory function is maintained. The more sensitive neurological testing is beginning to show that the margin between maintained function and diminished function is narrow. In those patients who develop weakness in association with significant lengthening, there has also been the question as to whether it was myopathic or neuropathic in nature. Young et al. (515) studied six consecutive patients completing tibial lengthening by the Ilizarov method by electrodiagnostic techniques. Nerve conduction studies and electromyography (EMG) were performed. At the termination of lengthening there were no complaints of sensory or motor abnormalities in the group, and all patients were normal to clinical examination. All six subjects demonstrated significant sensory and motor nerve response abnormalities. Electrodiagnostic testing showed abnormalities in six of six deep peroneal nerves and five of the six demonstrated abnormalities in superficial peroneal sensory responses. Two of six demonstrated abnormalities related to the posterior tibial nerve. Five of six patients demonstrated needle EMG abnormalities. Although the study was limited, there was clear evidence for an axonal neuropathy based on the nerve conduction and EMG results. A purely muscle etiology would not be expected to demonstrate sensory nerve abnormalities. The authors performed additional studies in an effort to implicate slightly increased compartment pressure as part or all of the causation of the neuropathic findings. In the study by Young et al., the mean lengthening was 5.6 cm with a range from 4.0 to 7.0 cm; the lengthenings, therefore, were well within the normal range in terms of extent. Galardi et al. (173) assessed peripheral nerve damage during limb lengthening. Electrophysiologic studies on limbs having five bilateral tibial lengthenings showed reduced motor conduction velocity in two tibial and three common peroneal nerves after a mean lengthening of 27%. Makarov et al. (309) reviewed much of the literature concerning neurological problems in relation to limb lengthening and published the data in chart form in terms of both intraoperative and total neurologic complications. A study of 8 reports encompassing 946 cases reported 51 intraoperative nerve injuries (5.4%), whereas total neurologic complications from 12 reports described 215 complications in 1214 patients (17.7%). A study of the Ilizarov technique by Erohin and Makarov (309), originally published in Russian, accounted for 703 of the patients, and results from this large series were parallel to the overall reports with 17% total complications and 5% intraoperative complications. Makarov et al. used intraoperative somatosensory evoked potentials (SSEP) to detect acute peripheral nerve injury during external fixation application. There were 42 Ilizarov surgical procedures of the lower extremities reported in 40 children. Significant deterioration or total loss of SSEP response dur-

SECTION IX ~ Management of Lower Extremity Length Discrepancies

ing surgery occurred in 4. They proposed the use of monitoring to detect early abnormalities and possibly to minimize or eliminate their long-term effects by changes in surgery pattern. Distraction Osteogenesis Research: Unlike previous lengthening methods, the distraction osteogenesis-callostasis technique has a considerable body of animal research data associated with it. The extensive work of Ilizarov and associates has been presented in English by Paley (361) and more recently in translation by Ilizarov (243-247) himself. Ilizarov showed that the proper biomechanical environment was extremely important for bone regeneration, which involved not only the interfragment stability but also the timing of the beginning of lengthening and the rate of lengthening. His early work demonstrated that preservation of the intramedullary circulation, particularly the nutrient artery, was important, but subsequent assessments have shown that even when cut the nutrient artery will repair quickly over a period of a couple of weeks as long as stability is present. Ilizarov also showed that the rate of osteogenesis was closely related to the distraction rate and that the optimal rate was 1 mm per day, with quicker rates of 1.5 and 2 mm per day slowing osteogenesis and a slower rate of 0.5 mm leading to premature consolidation. Repair was improved with four separate lengthenings of 0.25 mm each 6 hr apart distinct from one lengthening per day for the entire distance. The smaller, more frequent lengthenings minimize damage to the repair microvasculature and to the early repair cells and matrices. Ilizarov showed that bone formed during the course of distraction osteogenesis is well-organized and longitudinally oriented in the direction of the distraction forces. New bone forms initially in the medullary canal adjacent to the cut cortices and then passes progressively toward the center of the distraction gap. The central region between either cut cortical zone is referred to as the interzone and is the region in which bone forms latest and distraction occurs longest. The region tends to be filled with immature fibroblastic cells, which transform relatively late into osteoblasts. In most instances bone formation is via the intramembranous route with no cartilage forming in the gap region. The new bone is oriented along the longitudinal microvasculature and quickly forms a lamellar orientation. The interzone region ossifies quickly after distraction has stopped. New bone formation is seen as early as the second or third week after beginning distraction, and the interzone region then usually is seen as a central transverse radiolucency. During the several months of the remodeling phase, the cortex thickens and eventually the medullary cavity is reformed. The original work of DeBastiani and associates is almost totally clinical in nature, although use of the apparatus in experiments distracting the epiphyseal plate has been reported. Specific reports on distraction osteogenesis in the rabbit have been published by White and Kenwright (498), who noted that delayed distraction in the skeletally mature rabbit tibia led to more vigorous osteogenesis compared with immediate distraction. Kojimoto and associates (276) also

693

performed callus distraction in the rabbit and demonstrated excellent bone healing, even following medullary vessel destruction as long as the periosteum was carefully preserved. Experimental studies in the dog have been reported by Aronson and associates (27), quantifying mineralization by CT methods, and by Delloye and associates (140), documenting regenerate bone formation using microradiography and histology. DePablos and Canadell (138, 139) studied lengthening in a sheep model using a unilateral fixator with long-term assessments by radiographic and biomechanical techniques. Aronson et al. (28) studied the histology of distraction osteogenesis using Ilizarov and Wagner external fixators for tibial lengthenings with 8 dogs in each group. The distraction osteogenesis procedure was performed after a 7-day latency period and distraction at 0.25 mm every 6 hr until an osteotomy gap of 2.8 cm (about 15% of initial tibia length) was achieved. Correlative histologic and radiographic studies were then made at varying time periods. Both groups healed equally well. Radiodense columns of bone appeared regularly between days 21 and 28 of the distraction osteogenesis procedure. These took their origin from either bone end with the most central part of the gap persisting as a radiolucent band. Areas of bone repair were aligned in linear fashion parallel to the long axis of the gap and of the entire bone. The next phase of healing encompassed continuous bone tissue traversing the gap from end to end. The radiolucent band corresponded histologically to parallel bundles of collagen intermixed with cells also oriented in the direction of the distraction force. The vascular channels also were noted to be longitudinally oriented and there was no mention of cartilage formation. At higher power magnification there was direct transformation of fibrous matrix into bone strikingly similar to the intramembranous ossification characteristic of the embryonic phase. In a subsequent histologic analysis of the repair gap in tibial distraction osteogenesis by the Ilizarov method in dogs it was shown again that intramembranous ossification proceeded from each cortical end toward the central fibrous interzone. Good correlation between histologic repair and mineralization as shown by CT scanning (27). Delloye et al. (140) studied bone repair during distraction lengthening on the forearms of mature dogs using the Ilizarov system. With distraction both periosteal and medullary callus on either side of the gap gave rise to new bone trabeculae. These were oriented along the direction of distraction and progressively approached one another. The characteristic central transverse region of gap radiolucency was also reproduced. Bone in longitudinal alignment traversed the entire gap region, linking proximal and distal bone fragments 4 weeks after the end of the lengthening period. Most of the new bone formed by intramembranous ossification with some foci of cartilage seen. Specific delineation of the cortex began to be noted at 3 months but was still not fully achieved at 5 months. Procedures were performed on 13 adult female dogs, some unilateral and some bilateral, totaling 20 operative procedures. Slight variations in technique were used to

694

CHAPTER 8 9 Lower Extremity Length Discrepancies

assess their influence on repair. In general, initiation of periosteal and endosteal callus at the bone ends became apparent at 3 weeks, and with progressive distraction bone repair along the long axis of the gap and adjacent to either cut end was noted. Bone regeneration occurred equally at proximal and distal ends with the central radiolucent region seen. Bone bridging of the gap was usually achieved in a month after the end of distraction, but full cortical definition was not noted circumferentially even at 20 weeks. Microangiography revealed that the integrity of the medullary artery could be preserved after careful corticotomy. Callus was vascularized by both periosteal and medullary systems. The microradiographic study showed the first signs of osteogenesis at both ends of the distracted bone segments from intramedullary and sub-periosteal sources at 2-3 weeks. Striated longitudinally oriented callus was noted at 4-6 weeks. No evident difference was observed between lengthening after corticotomy or full transverse osteotomy with respect to the amount of callus. Histologic evidence of bone repair was, as expected, slightly in advance of radiologic manifestations. Medullary and periosteal osteogenesis was particularly active at 2 weeks. Woven bone was synthesized initially and there was abundant vascularity associated with this. With progressive distraction and time the longitudinal orientation of the new bone trabeculae was seen. Bone formed from a membranous ossification sequence with the advancing fronts of osteogenesis approaching each other from either side and fusing approximately 4 weeks after distraction was ended. The healing sequence was the same regardless of whether the bone had been broken by corticotomy or by transverse osteotomy. Some areas of highly cellular cartilage and fiber cartilage were noticed during the first 2 months, but these were invaded soon by vessels followed by endochondral ossification. Bone marrow appeared to be the larger contributor to the amount of interfragmentary callus, but periosteal callus also constantly supplied the peripheral part of the regenerating bone. Similar findings were reported by Lascombes et al. (289) and Saleh (420) in studies of bone biopsies in human lengthenings. Lascombes et al. (289) were able to harvest 11 biopsies of repair bone during bone lengthening following the Ilizarov technique. The mean age of the patients was 13.5 years and the delay after the initial procedure ranged from 23 to 502 days. Bone was noted histologically along a long axis of the gap as early as the third week. There was a distinct linear alignment to the bone trabeculae along the long axis of the bone. New bone formation was of the intramembranous type without evidence of a cartilage stage. Osteoblastic and osteoclastic activities were prominent, and remodeling was continuing even 1 year after initial intervention. Mature lamellar bone was noted, however, by the fourth month postsurgery. Saleh et al. (420) analyzed bone from 8 patients undergoing distraction osteogenesis using the Orthofix technique. The specimens were obtained from 102 days to 4 years postsurgery. The earliest bone synthesized was

woven with high cellular activity. This was soon covered by lamellar bone, which with time developed a characteristic Haversian architecture. The results of each of several experimental reports on the histology of repair in distraction osteogenesis have been similar. Many have been correlated with clinical studies, primarily radiologic but on occasion utilizing biopsy material. Among the characteristic features are the orientation of newly synthesized collagen and then bone trabeculae along the long axis of the repair gap parallel to the distraction forces. In the vast majority of instances, there is direct intramembranous bone formation present initially adjacent to either cut end, with the central region or interzone healing last. There is a contribution from the inner layer of the reconstituting periosteum, which also tends to be along the longitudinal axis and to represent new bone formation. On occasion, cartilage can be seen within the distraction gap and this subsequently turns to bone by the endochondral mechanism. The presence of cartilage, however, is best interpreted as a sign of less than optimal stability and does not represent true endochondral growth, but rather the formation of cartilage on the basis of increased interfragmentary movement and then conversion of that cartilage to bone once better stabilization occurs. Callotasis means stretching of the bony callus. It is evident that not all instances of bone lengthening heal with a uniform distribution of bone surrounding a central interzone region. Hamanishi et al. (217) classified the radiologic pattern of callus formation seen with the Orthofix procedure in 35 limbs (Fig. 23B). One of the continuing problems with bone lengthening procedures is this variable state and pattern of bone formation, even when the surgeons involved appear comfortable that a relatively uniform technique is being used. The categorization defined by Hamanishi et al. involved (1) the external pattern with lateral bulging of the bone, (2) a straight pattern in which the gap filled uniformly, (3) an attenuated pattern in which the diameter of bone formed centrally was less than at either end and the opposite pattern in which, usually with angular deformity, more bone was formed on the concave than on the convex side, (4) the pillar category in which a thin linear bone collection formed centrally, and (5) the agenetic form in which there were only isolated spicules of bone within the gap. The healing index correlated nicely with the pattern, as would be expected. The index (correlating the number of months per 1 cm of lengthening) was 1.1, 1.3, 1.5, 2.1, 3.7, and 4 in the respective types, with an overall mean of 1.7 because most patients were in the external or straight healing pattern category. One of the main principles articulated by Ilizarov (243, 246, 247) and DeBastiani (137) was the need for a delay in beginning the lengthening to allow for early revascularization and early bone formation, which subsequently would enhance the repair process. Aside from the excellent clinical evidence, experimental studies also confirmed the value of delay in distraction. Gil-Albarova et al. (177) used the Or-

SECTION IX ~ Management o f Lower Extremity Length Discrepancies

thofix fixator to compare results in femoral diaphyseal osteotomy on 24 3-month-old lambs, beginning distraction in half on the first postoperative day and delaying until the tenth postoperative day in the other half. The femur was lengthened by 2 cm at a rate of 1 mm per day. Both radiographic and densitometric studies of the lengthened callus at 1, 2, and 3 months showed that delayed distraction, when compared with immediate distraction, improved the quality of the callus with quicker, denser, and more homogeneous bone formation. The value of delay prior to distraction was also shown clinically by Lokietek et al. (303), comparing clinical results with the Wagner technique with immediate intraoperative lengthening of 1 cm and postoperative distraction of 1 mm per day and the Ilizarov technique, which did not begin lengthening until 5-7 days after osteotomy. They concluded that autogenous new bone formation in limb lengthening related primarily to the management protocol and was distinct from the external device used or even the location of the osteotomy. Their best results were reached when the surgical technique involved circumferential decortication, partial corticotomy with osteoclasis or cracking of the posterior element, a postoperative period of fixation without lengthening of 5-7 days, and an eventual distraction rate of 1 mm per day. Cell and Matrix Deposition Patterns in Distraction Osteogenesis-Rabbit Model: Studies have been performed in our laboratory on the histologic responses of tibial bone lengthening procedures using 4- to 5-month-old rabbits (436). The Synthes Mini Lengthening apparatus (Synthes, USA, Inc., Paoli, PA) was found to be placed easily and well-tolerated by the animals. Figs. 23C-23Ciii show radiographs of the mini-lengthening apparatus attached to the rabbit tibia at 7, 13, and 22 days, respectively. The initial work with this apparatus involved the use of skeletally immature 4- to 5-month-old rabbits with distraction begun on the first day postsurgery. Tibial lengthenings were then performed on 20 skeletally immature rabbits using the Synthes apparatus. A complete 360 ~ turn of the spindle knob gives 0.7-mm distraction. This represents the amount of daily lengthening and was performed in two stages, one in the morning and the other in the late afternoon. Subsequent studies assessing variable age and distraction parameters increased the total number of procedures to 60. Animals were sacrificed at varying intervals from 6 to 46 days postsurgery. Nine animals were allowed to heal for additional periods after lengthening was completed. The lengthening achieved in those lengthened to the time of consolidation of the regenerate bone, which was generally 21 days, ranged from 11.8 to 13.9% of total preoperative tibial length. Periodic X rays of the leg were obtained at 1- to 2-week intervals. At sacrifice the entire tibia was removed and specimen X rays were obtained in two planes (Figs. 23Di-23Dv). The specimens were then processed for light microscopic histologic study. Tissue preparation involved removal of the distracted segment and adjacent bone with a saw, decalcification, sec-

695

tioning in sagittal or coronal planes, and slide preparation using the JB4 plastic embedding technique. This allowed excellent visualization of the histologic detail, which was then correlated with the radiographic appearance. Specimen photographs were also taken after decalcification and halving of the specimen (Fig. 23E). The specimen X rays of the entire bone were taken using a standard technique with a magnification factor of less than 0.01. This allowed accurate measurements to be made of the length of the entire bone and of the distraction gap. A lengthened tibia at sacrifice after fixator removal is shown in Fig. 23E Histologic sections were processed from all animals. At 6 days following lengthening, the gap was filled with blood clot and fibrinous tissue with no geometric pattern of organization noted. Mesenchymal cells were accumulating adjacent to each of the bone ends with early intramembranous bone formation seen. The histologic studies from intervening time periods demonstrated the pattern of new bone formation. Animals that had undergone lengthening for 16 and 20 days showed the entire spectrum of repair cells and matrices within the gap. The repair was not uniform across the lengthening gap. Immediately adjacent to the cut cortical bone ends the repair bone was being transformed to a lamellar configuration, although evidence of initial woven bone persisted. Closer toward the center of the distraction gap the bone repair matrix was aligned longitudinally, as were the accompanying blood vessels. Further toward the center of the distraction gap the repair tissue was increasingly more woven in configuration with progressively less lamellar bone deposited. In the center of the distraction gap, mesenchymal cells persisted and in some sections areas of organizing clot persisted. This histologic picture correlated extremely well with the specimen photographs, specimen X rays, and the clinical finding of an inability to further distract the bone after 3 weeks. The model reproduces the clinical and radiographic findings being described in human distraction osteogenesis. Examples of repair sequences shown by histologic processing are illustrated in Figs. 24A-24E d. Lengthening along an Intramedullary Rod. Lengthening along an intramedullary rod has two attractive features. First, the rod helps to maintain alignment and generally eliminates translational and angular deformation. Second, external fixation can either be removed earlier or in some apparatus not be used, which hastens rehabilitation. Intramedullary Rod with External Fixator Distraction: For several decades some surgeons have performed lengthening along an intramedullary rod, which serves to enhance stability and to help maintain alignment at the lengthening site. The earliest formal paper with this approach was by Bertrand (43), who used one or two narrow intramedullary rods during distraction of the femur. Bost and Larsen (61) used an intramedullary rod to assist femoral lengthening in 23 procedures. The intramedullary rod, generally a large Rush rod, was narrow enough to allow distraction to occur and wide enough to allow for stabilization. The oblique, step

696

CHAPTER 8 "

Lower Extremity Length Discrepancies

F I G U R E 24 Histologic study of the gap region correlates well with the radiographic studies showing advancing fronts of repair adjacent to each sectioned end rather than a uniform repair process throughout the entire gap simultaneously. (A) The varying regions from which the histologic studies in this figure were taken. (B) Early new bone formation adjacent to cortex. Below is undifferentiated mesenchymal tissue at the center of the gap. (C) Histologic section from the persisting cortex (PC) and gap interface at 10 days following surgery and initiation of distraction. The initial bone synthesized is woven (W), which is more darkly stained here. Shortly thereafter, better oriented lamellar (L) bone is synthesized on the woven scaffold. Osteoblasts (arrows) line up nicely on the lamellar surfaces. (D) A higher power view in another rabbit in this same general region 2 weeks following surgery now shows a predominance of lamellar (L) bone with only small remnants of woven (W) bone persisting. Note the orderly array of osteoblasts (arrows) on the lamellar trabecular surfaces. (E) A region closer toward the center of the gap at 3 weeks. Note the predominantly longitudinal orientation of the new bone formation. This is related primarily to the associated longitudinal orientation of the vasculature. At the fight, a primarily lamellar trabeculum of bone is seen lined with osteoblasts. At the left, some initially deposited woven bone persists. (F) Tissue from the center of the gap region at 2 weeks. This is the most radiolucent appearing region on the radiographs as it is the newest site of bone formation. The mesenchymal cells have formed a primarily woven bone matrix. Note, however, the longitudinal orientation, the early formation of lamellar (L) bone, the lining up of osteoblasts on these lamellar surfaces, and in particular the marked osteoclastic resorption (arrows) occurring even as synthesis proceeds.

cut, and t r a n s v e r s e o s t e o t o m i e s e a c h w e r e used, b u t ulti-

tion a n d c o u n t e r t r a c t i o n w e r e p r o v i d e d t h r o u g h S t e i n m a n n

m a t e l y t h e y f a v o r e d the t r a n s v e r s e p r o c e d u r e b e c a u s e it w a s

p i n s into the p r o x i m a l a n d distal f e m o r a l f r a g m e n t s f o l l o w e d

simpler, led to b e t t e r c o n t r o l o f a n g u l a t i o n , and s h o w e d no

b y either the a p p l i c a t i o n o f t r a c t i o n u p o n the l i m b s u s p e n d e d

d i f f e r e n c e in h e a l i n g t i m e f r o m the o t h e r t w o patterns. T r a c -

in a T h o m a s splint or the u s e o f a t r a c t i o n - c o u n t e r t r a c t i o n

SECTION IX ~ Management of Lower Extremity Length Discrepancies

apparatus designed by themselves, which had an upper femoral ring and a frame on both the inner and outer aspects of the thigh and leg. Traction was started immediately after the operation and continued at a low rate until the desired length was obtained or no further length'~was obtainable. The patient remained in bed during the lengthening procedure. The amount of lengthening of the femur varied from 0.38 to 4.25 in. In the entire series of 23 lengthening operations, the average gain was 2.18 in. or slightly less than 2.25 in. or 5.5 cm. In 20 of the straightforward patients, the average time required for the lengthening was about 11 weeks. In 13 of the 23 procedures union of the bone occurred in an average of 32 weeks and did not vary between the various types of osteotomy. After 10 of the osteotomies, one or more bone grafting procedures were necessary to obtain bony union. All of the bone grafting procedures were performed as secondary interventions because of delayed union because no primary bone grafts were used at the time of lengthening. Both of the authors, as well as Abbott and Saunders (5) in 1939, specifically noted that lengthening of the callus was occurring. Bost and Larsen indicated that "during its growth the bone callus may be stretched out in length." They pointed out that major discrepancies could be markedly improved by a combination of lengthening on one side and shortening on the other. There was one late disturbance in circulation during the process of lengthening but this resolved with reduction of the traction weight. A palsy of the peroneal nerve occurred in 7 patients but in 5 it was transient. There were 5 patients with a posterior subluxation of the tibia on the femur, but after management no major problems resulted. Fractures of the lengthened femurs occurred in 4 of the 23 cases. The primary aim of the procedure, however, the control of alignment during lengthening, was successfully obtained thus eliminating many of the difficulties associated with lengthening procedures. Transverse osteotomy led to healing as quick and as sound as the more commonly used step cut or oblique osteotomies. Once reasonable healing had occurred, the patients were protected with either a cast or a brace during the transition phase to full unprotected weight bearing. Wasserstein (494) used an unreamed, flexible intramedullary rod in association with distraction through an external fixator of the Ilizarov type, followed by a cortical allograft of the distraction gap using a tubular bone segment once the lengthening had been achieved. Wasserstein transplanted cylindrical allografts into the distraction gap both to decrease the treatment time as well as to increase the stability of the fixation and insure proper alignment of the limb. The technique was used in patients between 5 and 25 years old, although in selected cases those up to 35 years of age were operated. He felt that the technique was best used for discrepancies greater than 6 cm because lengthenings under 6 cm appeared to heal adequately with distraction osteogenesis techniques alone. The technique was used in 300 patients over a 15-year period for both femoral and tibial lengthenings.

697

FIGURE 25 Femoral lengthening around an intramedullary rod is shown. [Reprintedfrom (364), with permission.]

More recently, Paley et al. (364) have performed femoral lengthening with the Ilizarov or Orthofix distractors over an intramedullary nail and compared it with a matched group of patients having lengthening with the standard Ilizarov technique (Fig. 25). The Ilizarov external fixator was used in 11 lengthenings and the Orthofix fixator in 21 with a 10-mm intramedullary femoral rod. The mean amount of lengthening in 32 procedures was 5.8 cm (range = 2-13 cm) and the mean age of the patients was 26 years (range = 1053 years). Results in several categories were compared with the standard Ilizarov femoral lengthening that had been performed in 32 matched patients. The authors noted that lengthening over an intramedullary nail reduced the average time of external fixation by almost one-half. The range of motion of the knee returned to normal an average of 2.2 times faster in the group that had lengthening over the intramedullary nail. There were 6 refractures of the distracted bone in the standard Ilizarov group but none in those protected with the intramedullary nail. Paley et al. concluded that the advantages of lengthening over an intramedullary nail included a decrease in the duration of external fixation, protection against refracture, and earlier rehabilitation.

lntramedullary Telescoping Rod for Lengthening without the Need for External Fixators: The most recent technological innovation in limb lengthening has been the use of an intramedullary rod that contains an internal telescoping mechanism, such that the lengthening can be performed in the complete absence of any external fixators. Distraction is performed following osteotomy and fixation of the upper and lower portions of the intramedullary rod to the bone by

698

CHAPTER 8 ~ Lower Ex~cremity Length Discrepancies

use of a rotation maneuver of the limb, which serves to distract the two segments of the rod at the internal telescoping mechanism using a ratchet effect. Guichet and colleagues (205) in France developed a model for this form of lengthening. A precursor of this technique had been described by Gotz and Schellmann (190) in which a hydraulic distractor was placed in a modified interlocking intramedullary nail to provide for continuous distraction. The system developed by Guichet and collaborators has the two telescoping fragments relating to each other internally such that, with external rotation of the limb, a clicking and locking mechanism allows for elongation with no movement possible in the reverse direction. Fifteen external rotation maneuvers of the lower extremity, which serve to rotate the distal femoral fragment to which the intramedullary rod is attached, allow for a lengthening of 1 mm per day. The device was developed for femoral lengthening. An osteotomy is performed such as would be done for a closed femoral shortening procedure. The principles of distraction osteogenesis are used with lengthening beginning on the eighth postoperative day. The 1 mm per day lengthening is achieved in four separate time periods 6 hr apart, much as is used in the Orthofix and Ilizarov approaches. Once the appropriate length has been achieved, rotation is no longer performed and the rod serves as a regular intramedullary rod until healing occurs. Guichet et al. refer to the mechanical intramedullary system, System for Progressive Intramedullary Lengthening (SAPI), as one destined to be used as an internal dynamic fixator for the progressive lengthening of segments of the lower extremity. The apparatus also has a dynamization capability built in once lengthening has been reached. The rod is fixed to the femur both proximally and distally, allowing for stabilization at the same time that lengthening is occurring. The external rotation movement allowing for the lengthening to occur is one of 20 ~ following which the limb then returns to its normal position. Each rotation corresponds to a lengthening of 0.067 (1/15) mm such that 15 movements correspond to 1 mm of lengthening. Each movement provides an audible clicking sensation. Caton et al. (103) have presented a brief report with the elongating intramedullary nail based on a series of experiments in sheep. They inserted the nail in the femurs of four sheep bilaterally. In their report, each operation applied to the limb allowed 0.1 mm of lengthening with elongation of 1.25 mm per day. In the first group of animals the mean lengthening obtained was 3.2 cm over 24 days with a percentage elongation of 14.2%. Regenerated bone was noted radiographically at 15 days and consolidation took place at 5 months. In a second series the internal device was compared with lengthening using external fixation. In the second series the mean lengthening up to 90 days was 3.9 cm, which was actually somewhat more than with the external fixator. Caton et al. felt that bone regeneration in the intramedullary group was completely satisfactory. Although the technique has not yet achieved wide usage, its advantages are attractive

in the sense that there is no external apparatus and the skin is virtually intact. There are no structures to impede the muscles such that joint motion should be more readily obtained. Stability is maintained and angulation is either prevented or markedly minimized. e. Longitudinal Growth after Diaphyseal Lengthening Done prior to Skeletal Maturity. Clinical Studies: Although an increased rate of femoral growth has been reported after one-stage procedures for lengthening of the femur and a variable rate of growth has been noted after Judet-type procedures for lengthening of the femur and tibia, there have been few detailed radiographically documented studies of growth of bone after lengthening of the diaphysis. Authors of earlier papers that have included the results of lengthening of the diaphysis have commented on somewhatvariable, but generally good growth after the procedure. However, these studies were not directed toward the specific assessment of growth after lengthening, nor did they report data from radiographic measurements; thus, the exact rate of growth after lengthening cannot be determined from them. We noted from the data on growth in patients at Children's Hospital, Boston, that when lengthening was performed on a bone that had several years of growth remaining the lengthened bone continued to grow at a slightly increased rate in some patients, whereas in others growth became more inhibited than it was before the operation (435). In our study, data on growth were assessed from 18 patients who underwent lengthening of the femur or tibia by mid-diaphyseal osteotomy and the gradual distraction techniques of Anderson and Wagner. The goal was to define the growth responses to these lengthening procedures. Femoral Lengthening: In each of seven patients, the rate of postoperative growth of the lengthened femur, in relation to that of the normal bone, was increased compared with the preoperative rate. The average rate of preoperative growth of the short femur was 82% that of the normal side, whereas the average postoperative rate of growth was 90% of normal. In one patient the rate of postoperative growth was 21% greater than the preoperative rate, but in all of the others the increase was from 5-8%. The amount of surgical lengthening of the femur averaged 18% of the preoperative length of the bone, with a range of 6-35%. Tibial Lengthening: In all 11 patients, the rate of growth diminished from the preoperative levels, ranging from a 46% diminution to a 3% diminution. The average preoperative rate of growth of the shortened tibiae was 88% that of the normal side, whereas the average postoperative rate diminished to 64% of the normal side. The amount of lengthening of the tibiae averaged 20% of the preoperative length of the bone, with a range of 14-30%. Long bones have differing growth responses after lengthening of their diaphyses. However, when growth is assessed according to the specific bone that was lengthened (that is, according to whether the femur or tibia was operated on), more uniform patterns of response are seen. A slight increase

SECTION IX ~ Management of Lower Extremity Length Discrepancies

in the rate of growth was noted in each of 7 patients who had lengthening for congenital short femur, and the increased rate was maintained for several years after the procedure. There are insufficient data to determine whether this stimulation of growth is maintained until growth ceases. Suva et al. (466) documented an almost invariable tendency for increased growth in the femur after a one-stage lengthening. They noted stimulation of growth in 33 of 36 patients who had poliomyelitis, 8 of 13 who had a congenital short femur (including some who had proximal femoral focal deficiency), and 4 of 5 in whom the shortening had another etiology. Overgrowth after fracture of the femoral diaphysis has been recognized for several decades. Assessment of 74 patients using serial orthoroentgenograms documented overgrowth as a universal phenomenon in patients who are less than 13 years old, regardless of whether the fracture healed with anatomical reduction, shortening, or distraction. Increased vascularity to the entire bone brought about by the repair process has been thought to stimulate growth at the proximal and distal growth plates. Because the repair response that is engendered by lengthening of the diaphysis is much more extensive and prolonged than that after fracture of the diaphysis, overgrowth is expected. The primary pathological condition that causes the discrepancy and resistance of soft tissue to distraction, which can exert compressive forces on the growth plates to restrain their growth, can minimize the effects of stimulation. However, in the series reported here the patients who had lengthening of the femur all had an increased rate of growth. The decreased rate of growth that was seen in all patients after tibial lengthening in the present series does not appear to occur universally. Variable growth responses to Judet-type lengthening have been documented radiographically by Pouliquen and Etienne (392) and by Pouliquen et al. (390). In the more detailed of the two articles, the authors reviewed the results after 39 lengthening procedures; 6 were performed on the femur and 33 on the tibia although the report did not deal with the femoral and tibial procedures separately. Their report and the present series are not entirely comparable as the criteria for inclusion of patients and documentation were much stricter in the present series. Still, comparisons can be made. In the series of Pouliquen et al., of the patients who had poliomyelitis, the growth was normal after lengthening in 18, slowed in 7, and arrested in 2. Of their patients who had congenital agenesis, growth was normal in 1 and slowed in 5. The findings in the latter group were similar to those in the patients who had lengthening of a congenital short tibia in the present series. Pouliquen et al. noted almost no retardation of growth after lengthening of 5.0 cm or less and a progressively greater slowdown of growth when the lengthening, expressed as a percentage of the preoperative length of the bone, increased beyond 15%. The crucial determinant appears to be the percent of lengthening rather than the absolute amount of the lengthening. Of their patients who had lengthening of 10-15%,

699

only 1 (11%) of 9 had a slowing of growth, whereas 9 (45%) of 20 who had lengthening of 15-20% and 8 (80%) of 10 who had lengthening of 20-25% had a slowdown or cessation of growth. Hope et al. (238) noted no change in growth following 10 femoral and 10 tibial lengthenings using the Wagner technique for congenital shortening of the lower limb. It is unclear why their data differ from the Children's Hospital, Boston, and the Pouliquen group data or the experimental data of Lee et al. reported later. The absolute amounts of lengthening were not given by Hope et al., only growth ratios, and it is known for the tibia that the greater the amount of lengthening the greater the growth slowdown. No patient in the present series who had lengthening of the femur had a slowdown of growth, even though the increases in length were often greater than 15% of the preoperative length. Reports made before 1978 on lengthening of the tibia raised the matter of postoperative growth, but none included rigorous radiographic documentation of this specific phenomenon. The absence of detailed data allows for only general, qualitative impressions concerning growth after lengthening. Moore (334) noted that, of 19 skeletally immature patients with poliomyelitis who had lengthening of the tibia using the Abbott method, the correction was maintained in 15, growth actually increased beyond the normal side in 3, and only 1 showed increased retardation. In another report on Abbot-type lengthening, many patients who were operated on before the age of 12 years showed an increased discrepancy between the lengths of the extremities at skeletal maturity compared with the amount of lengthening achieved. In these reports, it is unclear how much of the final discrepancy was due to the condition itself, postoperative complications, or postoperative retardation Of growth. A review of the results after 31 Anderson procedures for lengthening of the tibia revealed a variable degree of postoperative recurrence of limb length discrepancy in patients who had undergone the procedure between the ages of 8 and 19 years, especially in those who had a congenital short tibia (118). Gross (203) indicated that in some patients growth was stimulated after Anderson lengthening of the tibia, whereas in others the opposite occurred. The majority of patients referred to in these four reports had poliomyelitis. The present series included only 1 patient who had poliomyelitis. It has been proposed that extensive resistance by soft tissues in the leg is responsible for increased inhibition of growth with lengthening of the tibia. The interosseous membrane and the Achilles tendon appear to be more resistant to stretching than the tissues surrounding the femur. The negative effects of increased pressure on epiphyseal growth have been well-outlined. Shortening was noted in this series even after extensive releases of soft tissue, including lengthening of the heel cord. Attempts have been made to document the impression that retardation of the growth plate can be due to increased pressure in the limb generated by the distractive forces during lengthening of the tibia. A pressure

700

CHAPTER 8 ~ Lower Extremity Length Discrepancies

gauge was developed by Pennecot et al. (383) to measure the forces generated during distraction, and the measurements were then correlated with the rate of subsequent growth. It was concluded that for the tibia there was good correlation between retardation of growth, lengthening of greater than 15% of the preoperative length of the bone, and the increased forces that were registered. In the present series, the growth response after lengthening of the tibia was different from that after lengthening of the femur. In each of 11 patients who had lengthening of the tibia, growth was retarded in comparison with the preoperative rate. The patients who had a congenital short tibia and those who had Ollier's disease had marked retardation of growth of the tibia, whereas the 1 patient who had poliomyelitis had only a 3% decrease. Greiff and Bergmann (198) demonstrated overgrowth in the tibia after tibial fracture. As the process of repair after lengthening of the tibia is more extensive than that after fracture, it is likely that the factors that cause stimulation after fracture are present after lengthening. In the patients in the present series, however, it appears that the factors limiting growth were more influential than those stimulating it. Growth responses do not appear to be dependent on the techniques that are employed to lengthen the bones. In this study, femoral growth was stimulated with the use of the Wagner apparatus, and in another it was stimulated with the use of a one-stage lengthening. Growth was maintained in some patients who had poliomyelitis after the use of the Abbott, Anderson, and Judet methods, whereas retardation of tibial growth increased in patients who had a congenital short tibia, Ollier's disease, or another nonparalytic condition using the Anderson, Wagner, and Judet methods. In the human, variable growth responses are demonstrated after lengthening of the diaphysis in bones that have several years of skeletal growth remaining. In our series, the 7 lengthening procedures that were performed for congenital short femur all led to an increased rate of growth. It is anticipated that lengthening can be performed on femurs that have several years of growth remaining with the expectation of continuation of growth at a slightly increased rate. The tibial lengthening procedures that were done for patients who had a congenital short tibia or Ollier's disease all led to retardation of growth that was more extensive than the preoperative retardation. Hadlow and Nicol (214) have incorporated this growth information into a formula used to aid in timing for femoral and tibial lengthenings, incorporating altered growth expectations as well as projections of the preoperative growth rate. Lengthening of the tibia should include a slight overcorrection to compensate for an expected retardation of growth if it is performed on a bone that has several years of growth remaining. Preferably, the tibia should be lengthened at or near skeletal maturity to avoid a loss of correction secondary to retardation of growth. The timing of surgical intervention for discrepancies in the lengths of the lower extremities before skeletal maturity should be improved by considering both the developmental

patterns that have been described previously and the patterns of growth after lengthening that have been described here. Experimental Studies: A detailed experiment assessing longitudinal growth of the rabbit tibia after distraction osteogenesis was reported by Lee et al. (293). They divided 99 5-week-old immature rabbits into five groups according to the percentage of lengthening done with group I at 10%, group II 20%, group III 30%, group IV 40%, and group V a sham operation with osteotomy without lengthening. They clearly demonstrated that tibial lengthening did not cause retardation of growth when the bone was lengthened by 1020%, but in those instances in which it was lengthened by 30-40% growth retardation was evident. These data correlate well with our clinical studies and those of Pouliquen reported earlier. In groups I, II, and V, no statistically significant growth differences were noted between the operated and control tibias. There were significant differences in the growth ratio in groups III and IV with relative growth ratios of left to fight decreased significantly in group III (average = 4.2%) and in group IV (average = 7.0%). Histomorphometric studies were also performed on the physes in each of the groups. These studies correlated well with the gross measurements of length. In groups I, II, and V there were no statistically significant differences, but in groups III and IV there were significant decreases in the total thickness of the operated tibial growth plates, both proximally and distally, compared with controls. There was thinning of both the proliferative and the hypertrophic zones. The overall heights of the growth plate were measured such that the average decrease in the proximal growth plate was 10.4% in group III and 23.9% in group IV. Distally it was 11.9% in group III and 12.4% in group IV. Similar ratios were found with diminution of the thickness of both the proliferative and hypertrophic zones studied separately. f Increased Awareness of the Need for Joint Stabilization and Axial Correction as Well as Limb Length Equalization in Complex Abnormalities. As the profile of lower extremity length discrepancies changed over the past few decades from one in which the major discrepancies in most series were due to the sequelae of poliomyelitis, the deformities became more complex such that shortness of the limb was often combined with subluxation and dislocation of associated joints and axial malalignment. Wagner (493)clearly pointed out the need to stabilize the joints and correct any axial malalignment prior to initiation of any lengthening procedure and also the need to watch carefully for joint and alignment changes during the course of any lengthening procedure. A report by Saleh and Goonatillake (419) strongly reiterated the need for adherence to these principles of treatment particularly with congenital lower extremity length discrepancies. The disorders they discussed fell into the range of femoral, tibial, and fibular congenital abnormalities in 92 patients, with the three most common groups encompassing 77 patients involving proximal femoral focal deficiency, proximal femoral focal deficiency and fibular hemimelia, and hemihypertrophy. Saleh and Goonatillake indicated that

SECTION IX ~ Management of Lower Extremity Length Discrepancies

joint stabilization was mandatory for good function and was an absolute prerequisite prior to beginning limb lengthening. In efforts to prevent or minimize the likelihood of hip subluxation or dislocation, therefore, femoral head-acetabular congruity would have to be established, such that pelvic or shelf osteotomy along with proximal femoral osteotomy would be needed. The greatest challenge is in cases of proximal femoral focal deficiency in which there often is a need for the previously mentioned procedures and on occasion, in the more severe variants, femoral-pelvic fusion or at least definitive placement of the proximal femoral shaft into the acetabulum. Knee instability due to anterior cruciate ligament and/or posterior cruciate ligament deficiency is common in dysplastic limbs. Although no specific treatment is needed for this, the occurrence of subluxation of the knee during the lengthening procedure must be observed for carefully and managed as well as possible during the procedure itself. Areas of concern at the ankle involve Achilles tendon tightness leading to an equinus deformity and varus-valgus instability, with abnormal relationships of the lateral malleolus to the medial malleolus or with changes in the rate of distal tibial versus distal fibular lengthening. In some instances the Achilles tendon lengthening, posterior ankle capsulotomy, and distal tibial-fibular stabilization are performed prior to the lengthening procedure. In other instances orthotic devices are used along with physical therapy to maintain looseness and anatomic integrity at the ankle region, with intervention occurring only with changes that develop. Abnormalities in femoral or tibial alignment of more than a few degrees should be corrected prelengthening with appropriate osteotomies. Soft tissue contractures at hip, knee, and ankle must also be released. The final area of stabilization needed prior to lengthening relates to a pseudoarthrosis either of the proximal region of the femur, such as can be seen in a proximal femoral focal deficiency, or a formal congenital pseudarthrosis of the tibia. In general, however, lengthening of a congenital pseudarthrotic tibia is rarely performed, although on occasion it has been attempted in that part of the tibia well away from the pseudarthrosis in which the bone structure appears normal. g. Humeral Lengthening. Attention has been directed to humeral lengthening for relatively extensive length discrepancy problems (104, 130, 141,385, 426). Because 80% of the growth of the humerus occurs from the proximal growth plate, injury to this region particularly in the early childhood years can lead to major limb length discrepancy. Dick and Tietjen (141) reported on a humeral lengthening following neonatal growth arrest in 1978 using the Wagner technique. The most common causes of humeral lengthening involve negative sequelae following a proximal humeral unicameral bone cyst, neonatal sepsis, humeral trauma, humerus varus, Ollier's disease and a severe skeletal dysplasia. Peterson (385) reviewed 12 cases from the literature in 1989 and added 1 of his own. The operations proceeded quite nicely with surprisingly few complications reported. This has been our observation as well. Humeral lengthenings

701

are considered to be the least problematic of the major limb bone lengthenings. Many of the problems characterizing lower extremity lengthenings are infrequently seen, such as delayed union or nonunion and angular deformity. The Wagner technique is readily applicable to the humerus, as is distraction osteogenesis using the Orthofix apparatus. We have performed 4 of the latter, each with a 5.5-cm gain in length and no problems referable to delayed union or nonunion, angular deformity, joint stiffness, or neurovascular issues. In a review of several reports, the amount of lengthening achieved varied from 20 to 125 mm with a mean of 66 mm (6.6 cm). In 7 of the 12 patients reported by DalMonte et al. (130), the mean lengthening was 5.0 cm and the mean percent lengthening was 25.2%. Cattaneo et al. (104) performed 43 humeral lengthenings with the Ilizarov technique on 29 patients, 14 with achondroplasia, with an average lengthening of 9 cm achieved (range = 5-16 cm). There were 7 fractures in 6 patients following removal of the apparatus, all of which were treated successfully. There were no permanent neurovascular problems. 3. TRANSPHYSEAL LENGTHENING: DISTRACTION EPIPHYSEOLYSIS OR CHONDRODIATASIS

a. Early Experimental Findings. Transphyseal lengthening procedures were conceived initially and performed experimentally by Ring (407). He achieved lengthening of between 11 and 32 mm in 20 puppies between 4 and 6 months of age using an external distraction apparatus, which elongated both distal radius and ulna. He recognized that when sufficient traction was applied there was a physical transphyseal separation, after which continued distraction opened up a space allowing lengthening to occur with subsequent repair with a cylinder of new bone from the periosteum. The physis continued to function, and both it and the metaphysis filled the distraction gap with repair bone centrally. Some animals suffered premature growth plate fusion to limit somewhat the length gain achieved, but some continued with growth postlengthening. Continuous transphyseal traction was applied by Fishbane and Riley (163) across the proximal tibial growth plate in 10 puppies. Histologic examination revealed fracture to have occurred in all cases through the metaphyseal zone of primary trabeculae just distal to the hypertrophic zone of the cartilaginous growth plate. The physis itself appeared undamaged. Growth was felt to continue, but complete follow-up to skeletal maturity revealed early fusion preceding the normal limb by several weeks in 5 puppies. Rapid bony healing was noted in the distraction gap. The authors felt that the technique could be applied to the human as an effective and reasonably safe way of obtaining increases in limb length. Sledge and Noble (448) performed transphyseal lengthening experiments in the distal femur of the rabbit. They also varied the forces across the physis and compared histologic findings in an effort to determine the precise site of lengthening. A Salter-Harris type I transverse fracture occurred in

702

CHAPTER 8 ~ Lower Extremity Length Discrepancies

13 of the 16 animals to which 5 kg or more of force had been applied. The fractures occurred at the lower part of the hypertrophic zone. At lesser levels of force, however, there were markedly fewer or no fractures and lengthening appeared to have occurred in relation to increased thickness of both the proliferative and hypertrophic zones of the physis. Thirteen of 21 animals to which 2 kg were applied and 18 of 19 animals to whom 1 kg was applied showed only a partial microscopic fracture or no fracture. In 17 of 56 animals distracted there was no fracture of any type, and overall there were 34 rabbits in which gross disruption at the plate did not occur. The response of the physis to transphyseal force, therefore, clearly was dependent on the amount of force applied. The combined force in excess of 2 kg almost always caused fracturing. There was, however, a consistent increase in length on the involved side whether or not fractures in the plate were produced. In the absence of fracturing, Sledge and Noble felt that hyperplasia and hypertrophy of the plate were representative of an increased physeal stimulation, which in turn led to the length increase. Jani (254, 255) also noted fracture in the hypertrophic zone and subsequent healing by endochondral ossification in 42 puppies in whom distraction epiphyseolysis had been performed. Further insight into the effect of chondrodiatasis on the physis itself was provided by Elmer et al. (149). They performed a study in the rabbit to assess cell activity in the physeal region and noted that the procedure did not produce any significant changes in the percentage of cells labeled with tritiated thymidine, the intensity of radioactive sulfate labeling the matrix, or the blood supply of the physis. Thus, Elmer et al. felt that lengthening was not the result of stimulating cell division or increased synthetic function of the plate, but rather the result of stretching of the matrix passively. Transphyseal lengthening is referred to as distraction epiphysiolysis when the distraction force is sufficiently great that a physeal fracture-separation occurs, or chondrodiatasis in which lesser distraction forces leave the physis intact with hyperplasia occurring in the proliferating and hypertrophic zones. In the former repair occurs by an intramembranous bone mechanism, whereas in the latter the endochondral mechanism continues. The distraction device spans the growth plate with pins anchored in the secondary ossification center of the epiphysis and in the metaphysis, and the distraction forces lengthen the limb by causing a separation at the plate. DeBastiani (136) slowed the rate of lengthening to 0.25 mm every 12 hr and advanced the concept that, in so doing, a transphyseal fracture did not occur but that a stretching of the hypertrophic zone only was the lengthening phenomenon. The term chondrodiatasis was used to refer to this occurrence. b. Clinical Use The usual sites of limb lengthening in the human have been within the diaphyseal or metaphyseal regions of bone but work by Zavyalov and Plaksin (519), Ilizarov and Soybelman (248), and Monticelli and Spinelli

(329-331) showed that transphyseal lengthening was both clinically feasible and advantageous in some regards. These advantages involved the fact that healing was much quicker than in diaphyseal lengthenings because dense cortical bone did not have to be repaired but rather only metaphyseal bone, much as would occur in a physeal fracture. Early experiments showed that actual physical separation occurred in the hypertrophic zone of the physis, leaving the major growth part of the physis intact, a phenomenon referred to as distraction epiphyseolysis. The transphyseal lengthening then occurs mechanically and subsequent growth continues once the physeal defect within the hypertrophic zone of the metaphysis has been repaired. Although this experimental technique can lead to superb results both clinically and experimentally, currently it is not widely used when there are a few years of growth remaining. There have been reports of premature growth plate cessation following this procedure. In patients, however, who have virtually no growth remaining, the procedure is attractive. It can be performed using either the circular distraction devices of Monticelli and Spinelli and Ilizarov or unilateral lengtheners such as the Orthofix device. Eydelshtyn et al. (154) reported extensively on the distraction epiphyseolysis in a clinical setting in association with meaningful limb lengthening. They noted epiphyseal separation radiologically within 7-10 days and the ability to obtain length increments of 4-7 cm. They suggested that in most instances there was no negative influence on further growth. The cleavage fractures appeared radiologically to have occurred in the metaphysis in 26 of 33 patients, with the remaining 7 occurring at the lowest levels of the growth cartilage but preserving the growth mechanism more proximally. DeBastiani and associates (136) performed chondrodiatasis using the unilateral Orthofix apparatus in 40 segments of patients with limb length discrepancies, gaining a mean of 3.3 cm in length (range = 1.5-7.0 cm), and in 60 segments in 25 achondroplastic patients, gaining a mean of 7.1 cm (range = 3-10.5 cm). In 16 distraction epiphyseolysis procedures reported by Monticelli and Spinelli (331) the tibia was lengthened by an average of 6 cm (range = 3 10 cm). The patients were all between 13 and 16 years of age with little to no growth potential persisting. Aldegheri et al. (15, 17) reported on chondrodiatasis used for elongation of 170 bone segments in 75 children, 41 with limb length discrepancies and 34 with achondroplasia. All were operated on with the growth plate open. The Orthofix apparatus was used to lengthen either the distal femoral or the distal tibial epiphyseal plates. Distraction began the day after operation at a maximum rate of 0.5 mm per day in several stages. There were 92 femoral and 78 tibial lengthenings. The average age was 11.9-12.1 years. In those treated for limb length inequalities the mean lengthening obtained was 3.4 cm (10.9% of average initial length). The mean lengthening of the femur was 3.0 cm and that of the tibia was 3.7 cm. In achondroplastic patients, the mean lengthening

SECTION IX ~ Management of Lower Extremity Length Discrepancies

obtained was 7.4 cm with virtually equal amounts in the femur and tibia. Most of the complications recorded were in the achondroplastic patients, and most of these were in the tibias primarily because of the abnormal shape of the distal tibial epiphyses. The procedure is generally performed at the distal femur or distal tibia just before skeletal maturity. It does not require the use of plates or grafts, because metaphyseal bone heals readily in the gap produced. Franke et al. (165) performed distraction epiphyseolysis using the Ilizarov apparatus in 22 lower limb segments with an average lengthening of 8.25 cm (range = 4-18 cm). Some of the patients had achondroplasia, and it was these patients in whom some of the larger lengthenings occurred. The healing was considerably quicker in the distraction epiphyseolysis group compared with the partial metaphyseal corticotomy group summarized previously. In a group of 9 patients lengthened between 6 and 9.5 cm, the average time to full weight bearing was 9.5 months with the repair index 43.6 days per cm. c. Growth Consequences Related to the Force o f Distraction and Mechanism o f Lengthening. Although initial reports from the European literature indicated that growth can continue following healing, there have been more recent reports of premature growth cessation such that the procedure is best performed within 1 year of expected growth plate closure or in situations in which metaphyseal or diaphyseal lengthening is not possible. Clinical papers reporting early physeal closure after chondrodiatasis are increasingly common. Hamanishi et al. (216) reported femoral chondrodiatasis in 5 patients, but in 4 of the 5 the physis closed shortly after lengthening and loss of gained length or further shortening occurred in each. They used the Orthofix apparatus and the lengthening rate was 0.25 mm every 12 hr. The 5 femurs were lengthened by a mean of 32 mm (range = 2543 mm) after 70 days of distraction. Subsequent growth, however, was markedly diminished in 4 and moderately diminished in the other. Hamanishi et al. felt that the 0.5 mm per day distraction caused physeal separation rather than hyperplasia of the growth plate alone. Bjerkreim (51) evaluated 10 consecutive proximal tibial physeal distraction cases with a mean lengthening of 6.7 cm. In 6 cases lengthened at 1 mm per day under 13 years of age, subsequent growth was only 6 mm compared with the normal side of 32 mm. Growth retardation has been seen in several experimental animals after epiphyseal distraction. Letts and Meadows (297) performed distraction epiphysiolysis of the proximal tibia in 18 rabbits. The average distraction gained was 0.54 cm, which was 6% of the length of the tibia at the time of intervention. Union invariably occurred. In each of 12 younger rabbits operated, once the distracted area united it was shortly followed by premature fusion of the growth plate. In 6 older rabbits operated close to the time of skeletal maturity no negative growth sequelae occurred. Subsequent histology showed no normal epiphyseal growth plates after repair of the gap. In the study by Fjeld and Stein (164), subsequent

703

growth retardation was consistently experienced in all animals with an average reduction in growth ranging from 40 to 70%. The experimental groups involved 10 distal femoral epiphyseal distractions in the goat, 14 in the proximal tibia of the goat, and 18 in the distal femur of the dog. They felt that the reduction in gained length with time after the end of distraction must have been due to a retardation of endochondral growth after the elongation procedure. Jani (254, 255) also reported negative growth sequelae after distraction epiphyseolysis experiments in the rabbit. Alberty (12) performed a vascular study of the growth region following physeal distraction in the rabbit. The microangiography assessments were performed in relation to 45 distal femoral transphyseal lengthenings. Most of these were performed at a rate of 1.0 or 1.5 mm once daily increases. Both marked enlargement of the epiphyseal arteries and defective metaphyseal capillary filling were noted after 3 days of distraction, changes that persisted in specimens distracted as long as 21 days. New capillaries were observed in the hyperplastic physes and in separation gaps at 21 days. Of note, however, was the fact that vascular anastomoses were noted across the physes at 6 weeks of follow-up. The latter phenomenon was noted in those distracted from 9 to 21 days who then were assessed after an interval of 6 weeks between discontinuation of the distraction and microangiography. A premature closure or impaired function of the physis was common in association with the transphyseal vascularity. In spite of the comments of DeBastiani concerning the absence of transphyseal fracture with slower rates of distraction, dePablos et al. (138, 139) found that, in each of three models used, production of a fracture between the metaphysis and the epiphysis always occurred but that the lower the distraction rate employed the greater the viability of the growth cartilage. The optimal rate for distraction was 0.5 mm per day. The Orthofix distractor was used in 45 lambs divided into three groups each, with the rates of distraction being 2, 1, and 0.5 mm per day. Histologic studies showed that at the slowest rate the growth cartilage remained essentially normal, whereas in femurs lengthened at a rate of 1 or 2 mm per day obvious lesions of physeal cartilage were observed, particularly in those studied 45 days following the conclusion of lengthening and at 6 months of age. Spriggins et al. (454) assessed the response of the growth plate to increasing force of distraction. They studied the upper tibia in 24 rabbits close to skeletal maturity with distraction rates of 0.13, 0.26, and 0.53 mm every 24 hr. They noted two distinct patterns of response. In the group in which forces increased to maximum values of 20-22 N and then suddenly decreased subsequent distraction fracture of the growth plate had occurred, whereas in the other group lower forces of 16-18 N produced and continued to the end of the distraction period were associated only with physeal hyperplasia without fracture. These results were consistent with those of Sledge and Noble and DeBasiani et al. that a slowed rate of distraction could allow for lengthening without

704

CHAPTER 8 9 Lower Extremity Length Discrepancies

fracture. It was in the two lower distraction rates that histologic evidence of fracture was not seen. d. Histologic Findings in Physeal Distraction. With the application of physeal distraction, therefore, the mode of lengthening is of two types. If distraction pressures are comparatively low, there is apparent continuity of the physis with hyperplasia of cells in the lower proliferating and hypertrophic zones. In most instances, however, there is a transphyseal fracture with the break occurring in linear fashion within the hypertrophic zone. Peltonen: Repair phenomena were studied by Peltonen (380, 381) in the distal radius of 40 growing sheep. The distraction was at a rate of 0.5-1 mm per day. In this model the first signs of ossification in the distraction gap were seen radiographically about 3 - 4 weeks after distraction began. Collagen fiber bundles were oriented parallel to the long axis of the bone in the gap region and served as the focal point for new bone formation. Abundant osteoid was demonstrated in the calcified sections. Thin trabeculae from both the epiphyseal and metaphyseal sides grew toward the center of the distraction gap. At the periphery there was active bone formation from the inner layer of the periosteum. Clinical and radiologic consolidation of the distraction area occurred within 10 weeks of the surgery. The consolidated bone area was composed of partly woven bone and lamellar trabecular bone, with most of the lamelli organized in the direction of distraction. In those instances in which the physis remained intact there was an apparent stretching of the endochondral growth mechanism due to the distractive force. In many instances, however, there was actual transphyseal rupture through the hypertrophic zone. At 12 weeks postoperation the repair bone resembled normal metaphyseal bone. Once the hematoma of the stretch injury had been removed, the gap was filled with a collagenous matrix and collagen bundles could be seen organized in the direction of the distraction. Shortly thereafter, trabecular bone had been formed from the inner layer of the periosteum and from epiphyseal and metaphyseal sides. Peltonen et al. studied gradual physeal distraction of the distal radius in 2- to 5-month-old sheep (379, 382). The bone formation occurred from the inner layer of the periosteum, from the metaphysis, and from the epiphyseal side toward the center of the distraction area. Prior to bone formation, collagen bundles were organized in the direction of distraction. Bone formation occurred from the epiphyseal and metaphyseal sides. Under polarized light microscopy lamellar trabecular bone was seen forming along the collagen bundles. Woven bone was frequently seen. In the distraction area at 11 weeks in the study by Peltonen both lamellar and woven bone were present within individual trabeculae, but it was clear that the lamellar bone predominated. Alberty and Peltonen (13) showed that all physes in 12 growing rabbits undergoing distal femoral distraction showed widening of the proliferative and hypertrophic zones, as well as a fracture-separation at the hypertrophic

zone or at the junction of the hypertrophic zone and the metaphysis in 11 of 12. Proliferation of the hypertrophic cells with 5-bromo-2-deoxyuridine (Brd Urd) labeling was noted even though normally not seen. The labeling of the physeal regions above, however, was normal in both control and distracted physes, being positive in germinal and proliferating cell zones with no labeling in the hypertrophic zone. DePablos et al: In the histologic study of dePablos et al. (138, 139), local fracture was noted in each instance. The physis tended to be poorly organized in some instances and otherwise normal in some. The lengthened zone was first wholly occupied by a hematoma, which quickly underwent fibrous organization. This then was replaced by a rich granulation tissue composed of numerous fibroblasts and collagen fibers, which aligned themselves parallel to the long axis and the traction axis of the bone. Ossification of the lengthened zone was first observed on the 20th day of lengthening. This was referred to as desmal ossification, although there was some endochondral ossification taking place in the growth cartilage at its lowest regions. The ossification progressively replaced the fibrous granulation tissue. Four months after the lengthening process had begun all of the tissue in the lengthened zone showed complete ossification. Monticelli and Spinelli: Monticelli and Spinelli (329, 330) showed that the gap invariably healed with bone, provided that the surrounding perichondrium and periosteum were not excised. The resumption of endochondral ossification was infrequently seen and was incomplete when present. The new bone consistently formed along the fibrils as well as along the undersurface of the peripheral periosteum. The bone was noted to be well-oriented in most instances. Polarization microscopy showed the fibrils to be parallel to the long axis of the bone and to the axis of the distraction. Monticelli and Spinelli carried out proximal tibial lengthening experiments on 41 sheep between 3 and 5 months of age. The duration of the distraction varied from 1 to 16 weeks. The lengthening rate varied between 0.5 and 1.5 mm per day with most either 0.5 or 1.0 mm/day. Separation of the epiphysis from the metaphysis was felt to occur after 3-5 days, and the force needed to induce fracture was approximately 18 _+ 4 kg. Separation was always noted within the hypertrophic zone. During the first and second weeks, the region between the epiphysis and the metaphysis referred to as the interzone was transparent on radiographs. At the beginning of the third week, when the interzone was 10-20 mm wide, faint irregular shadows became visible suggesting early osteogenesis. New bone formation was seen next to the epiphysis and the metaphysis with the central region initially showing little bone formation. After 1 month of distraction, the interzone was 25-30 mm wide and the new bone formation sites began to link longitudinally. Longitudinal bundles were continuous now between the epiphyseal and the metaphyseal ends. The bone formation from the metaphysis generally was more advanced than that from the epiphysis. After 2 months of distraction the interzone had increased to 42-60 mm and

SECTION IX ~ Management of Lower Extremity Length Discrepancies

the radiographic shadows were more clearly striated in a longitudinal direction. Ossification proceeded upward and downward toward the center of the interzone where bone repair was ultimately slowest. Only when distraction ceased did the central region begin to show uniform radiodensity. After lengthening of 30-40 mm it took about 2 months for the bone cortex to reform. In a few animals premature epiphyseal fusion occurred, which led to the recommendation that transphyseal lengthening be performed only when the subject's bone growth was nearly complete. The repair bone from the sheep was then subjected to a more detailed morphologic study at 1-4 weeks and 2, 4, 6, and 12 months (330). One week. An epiphyseal fracture (epiphyseolysis) occurred in all specimens. The fracture gap was filled with hematoma. There was no calcification or bone at this stage. The fracture line occurred within the growth plate around the level at which matrix calcification was beginning to occur in the hypertrophic zone. Two weeks. The hematoma occupied a wider space but persisted. Most of the chondrocytes in the physis itself were unchanged, although cell columns of the upper cartilage segment were sometimes distorted. The cartilage on the metaphyseal side was penetrated by many capillary vessels and in general looked more irregular than in the normal control. Three weeks. The distance between the upper and lower cartilage cell segments had widened due to the continuing distraction. Hematoma was becoming organized with fibrous material. No bone was yet formed. Four weeks. The gap was now filled with a compact, translucent, grayish tissue and the lower cartilage segment was less evident than previously. There now was a thin shell of bone under the periosteum surrounding the distraction space. This was also seen radiographically. The hematoma had almost been completely resorbed and replaced with connective tissue consisting mainly of elongated cells surrounded by thin fibrils. Many of these were irregularly oriented but the beginning tendency to longitudinal orientation was seen with fibrils being aligned in the same direction of the distraction force. A thin layer of circumferential periosteal bone composed of thin trabeculae of primary bone was seen. Two months. A compact gray tissue zone was evident grossly below the epiphyseal cartilage. Sub-periosteal bone formation was seen along with metaphyseal ossification. No traces of hematoma were recognizable, it having been replaced completely by connective tissue consisting of elongated fibroblasts and thick bundles of collagen fibrils, both of which were preferentially oriented parallel to the tensile force of the distraction. This was particularly evident in the middle portion of the elongated segment. Some of the collagenous material was calcified and there were small spicules of osteoblasts seen. Four and six months. A thin surrounding shell of cortical bone was seen at the distraction area. Ossification was continuing in the sub-periosteal area. New bone was also seen near both the epiphyseal and metaphyseal zones. Long thin bundles of collagen fibrils were

705

present in the elongated area. Most of these bundles were now calcified, forming thin parallel oriented trabeculae whose longitudinal orientation was the same as that of the distraction force. Together with these calcified collagen bundles true bone trabeculae were found, part of which developed from the epiphyseal cartilage and part from the dislocated metaphyseal trabeculae and the periosteum. The tips of these trabeculae were surrounded by osteoblasts. They also appeared osteoblastic by electron microscopy. The persisting physeal cartilage in some instances retained its normal structure, but in others it was irregular. Of note, however, was the fact that no bone trabeculae were present near the epiphyseal cartilage, indicating disruption of the normal endochondral sequence. Twelve months. Ossification had progressed in the elongated segment and the fibrous tissue had been replaced almost completely by bone trabeculae. The sub-periosteal bone had a compact appearance. On the whole, the elongated tibias were almost indistinguishable from the contralateral side. The process of distraction epiphyseolysis was divided structurally into three stages: (1) epiphyseolysis with formation of the hematoma; (2) resorption of the hematoma and the formation of fibrous tissue, which increases in length as long as the distraction force is applied; and (3) ossification of the fibrous tissue and reconstitution of the periosteal bone. As the fibrous tissue develops in the second phase, the cartilage on the metaphyseal side is resorbed and replaced by fibroosseous and osseous tissues, which are well-vascularized. In the third stage there is first calcification and then true ossification of fibrous tissue with the formation of longitudinal trabeculae lying almost parallel to each other. Ossification takes place directly from the periosteum and from the cellular elements of the fibrous tissue undergoing osteoblastic differentiation in connection with the displaced metaphyseal segment. Ossification also takes place from the epiphyseal cartilage but is less regular and less impressive than from the other two sites. On occasion, the endochondral ossification sequence of the physis persisted, but in others it was damaged. Cessation of endochondral ossification was not of clinical significance if the distraction epiphyseolysis was carried out at an age approaching skeletal maturity. Varying reports in the literature could easily be due to the differing types of distraction apparatuses used and the different degrees of stabilization provided. 4. TRANSILIAC LENGTHENING Millis and Hall (327) showed that, in patients with a discrepancy in lower extremity length associated with acetabular dysplasia, primary intrapelvic asymmetry, or a decompensated scoliosis, 2.5 cm of length can be gained through transiliac lengthening by means of a modified innominate osteotomy. This approach both increases lower extremity length and corrects structural pelvic abnormalities associated with the discrepancy. Three possible complications involve delayed union, sciatic nerve palsy, and sacroiliac joint

706

CHAPTER 8 9 Lower Extremity Length Discrepancies

disruption. The procedure is most strongly indicated for primary correction of pelvic deformity in which slight limb lengthening is also desired.

X. DIRECT OPERATION ON EPIPHYSES TO ENHANCE GROWTH POTENTIAL

BY REMOVING FOCAL TRANSPHYSEAL TETHERS A. Bone Bridge Resection Partial destruction of the function of the growth plate is associated with the formation of localized transphyseal bone bridges (287, 313). These develop most commonly after certain growth plate fracture-separations (6, 60, 73, 304, 350, 421,434, 461,495), in severe cases of Blount's disease (infantile tibia vara), and after infection (409). The bone bridges retard growth in a localized part of the physis and predispose one to angular deformity as well as shortening because the remaining physeal tissue continues to function. The possibility of removing focal bone bridges was raised over 100 years ago by Ollier (354). He was able to demonstrate the formation of transphyseal bone bridges in experimental animals. Experimental work in rabbits by others at this time also demonstrated partial bone bridging of the physis (434). There was good clinical awareness in the latter stages of the nineteenth century of bone bridge formation following epiphyseal trauma with its subsequent effect on growth. Ollier himself made efforts to remove bone bridges surgically, but recurrence was common due to the failure to use an appropriate interpositional tissue to prevent recurrent bridge formation. The mechanism of bone bridge formation was described extensively in Chapter 7. The position of the bone bridge defines not only the type of deformity but also the surgical approach to removing the bridge and the type of material interposed to prevent reformation (Fig. 26A). Central bone bridges lead to shortening without angular deformation, whereas peripheral bone bridges lead to angular deformity as well as shortening. Bright (72) classified partial growth arrest lesions into three types: type I, peripheral lesion; type II, central lesion; and type III, combined central-peripheral lesion. The exact position and extent of bone bridges can be shown by tomography, CT scanning, and magnetic resonance imaging (Fig. 26B). Examples of bone bridges are shown in Fig. 26A. Langenskiold (285-287) refocused attention on focal bone bridge formation and developed bridge resection and the implantation of fat for use as a clinical tool. In a series of experiments in his laboratory many types of interpositional materials were used, but fat was both the easiest and the most effective in preventing the reformation of bone bridges and thus maintaining physeal function. Fat is a minimally vascularized tissue and generally persists as fat when interposed in growth plate defects, thus keeping the epiphyseal and me-

taphyseal circulations separate and allowing the remainder of the physis to continue growth. Another interpositional material used was cartilage, which in experimental studies had just as effective a result as the fat (357). In commentary on Langenskiold's first 43 clinical procedures excising local bone bridges and interposing autologous fat grafts, the results in general were good to very good; only 7 showed questionable benefit "mainly because the procedure was carfled out too close to the end of the growth period" (287). The large majority of procedures involved the distal femur, proximal tibia, and distal tibia. The etiology of partial closure in 38 growth plates was fracture in 28, osteomyelitis in 8, and tuberculosis and Blount's disease in 1 each. Following interposition of the fat graft the radiolucent area of the fat transplant usually has a rounded or oval shape, whereas following subsequent growth the radiolucent area becomes elongated. The fate of the fat implants was studied experimentally by making round cavities in the proximal end of the tibial growth plates in pigs and filling them with autologous fat. Studies indicated that the volume of fat tissue implanted in the cavities continuously increased in parallel with the growth in length of the bone. It appeared that the fat was augmented by fat cells in the metaphysis. Langenskiold et al. (288) recalled 3 patients several years after surgery for CT scan assessment of the epiphyseal-metaphyseal region. They concluded that the former resection cavities were filled primarily by fatty tissue and that the portion of implanted fat had grown in size corresponding with the growth in length of the bones in the affected ends. Some strands of fibrous tissue were intermingled with the fat. A layer of dense bone remained interposed at the periphery of the fat graft. Langenskiold concluded that the free fat grafts implanted at the time of resection continued to grow and thus had filled out the elongated cavities. The fat persisted well beyond the period of growth termination, and the cavities were not filled with fluid or bone. Examples of central bone bridge resection are shown in Fig. 26C. Other clinical studies have assessed the treatment of partial physeal growth arrest by bridge resection and fat interposition. Vickers (486) reported on 15 patients with good early results. Williamson and Staheli (506) assessed 29 physeal resections, 22 of which were followed for more than 2 years. They interpreted their results in the longer term group as 11 excellent, 5 good, 2 fair, and 4 poor. Twenty of the 29 bridges were caused by fracture, 3 by tumors, 3 by tibial traction pins, and 1 by infection. Twenty of the bridges were peripheral, 6 were central, and 3 were combined. The results correlated inversely with bridge size. They were uniformly excellent for bridges less than 25% of the physeal volume, bridges between 25 and 50% yielded good to excellent results in 9 of 12 cases, and results were generally poor in bridges greater than 50% with only 1 of 4 yielding a good result. Bright (72) reported briefly on 100 patients followed for more than 2 years with silastic interposition material, with 81% of the patients demonstrating some growth after

F I G U R E 26 (A) Mature transphyseal bone bridges can be seen on plain radiographs. (Ai) A central transphyseal bone bridge (arrow) is shown at left in a 7-year-old girl who suffered a distal tibial growth plate arrest subsequent to meningococcemia of infancy. (Aii) A peripheral bone bridge (arrow) of the proximal media tibia following Blount's disease is seen (right). Varus deformation of the tibia has developed. (B) Magnetic resonance imaging defines the extent of the bone bridge (white arrow; bone bridge is black, persisting physis is white). Image is from A, left. MRI is from the coronal (lateral) plane. (C) A series of radiographs shows the operative approach for removal of a central bone bridge (Ci, Cii), filling of the defect by fat followed by reinsertion of the bone window (Ciii, Civ), and results following growth resumption several months later (Cv).

708

CHAPTER 8 ~

Lower Extremity Len~tth Discrepancies

bridge resection and 70% with good to excellent results. Aufaure et al. (29) studied 18 cases of bone bridge resection in childhood and concluded that there were 9 good results and 9 failures. The best results were obtained in cases in which the bridge was peripheral, because it was approached more easily, and following a traumatic injury in young children. The larger the bone bridge, the greater the likelihood of failure. Extensive bridges particularly those located centrally, therefore, had a poor prognosis and all bone bridges due to osteomyelitis were failures. Most of the bone bridges were resected at the distal femoral and distal tibial growth plates. Resection has proved to be clinically feasible in many instances if one-fourth or less of the growth plate is involved and there is sufficient growth remaining to warrant removing the focal tether. Interposition of fat, cartilage, silastic, or methyl methacrylate can keep the epiphyseal and metaphyseal circulations separate, thus preventing the formation of further bone bridges and allowing the unaffected growth plate cartilage to continue to grow normally. Each of these methods has advocates; the interposition of fat is the easiest and most commonly used approach clinically. Examples of a bone bridge resection procedure are shown in Fig. 26C. The extent of the bone bridge must be determined prior to making a decision to resect the bridge. Plain biplanar radiographs are inadequate. Tomography has been shown to provide a good percentage estimate of physeal area replaced by the bone bridge, but CT or MR scanning with threedimensional reconstruction is used currently (96).

B. Varieties of Interpositional Materials Several interpositional substances have been used following resection of a transphyseal bone bridge to keep the epiphyseal and metaphyseal circulations and bone apart so as to allow physeal growth to continue. Cady et al. (82) performed fat implants into 14 New Zealand white rabbit physes in an effort to determine whether physeal regeneration occurred. They found no instance in which the physis regenerated transversely across the gap in either the fat-implanted or the control femurs. The fat remained viable and was gradually replaced with fibrous tissue. Three inert materials have also been used as interpositional substances with good effects described. These have included silastic (71, 72), methyl methacrylate (313), and bone wax. Among biological substances cartilage itself was shown to be excellent in preventing transphyseal revascularization. Lennox et al. (296) performed a comparative experiment in 5- to 6-week-old New Zealand white rabbits in which two adjacent 4-mm-diameter defects were drilled in the distal lateral femoral epiphysis. In the control group, valgus angulation had a mean of 43 ~ with 2.4 cm of shortening compared to the opposite side. The group with fat interposition showed a diminution of the valgus angle, although it was still 28 ~ with shortening of 1.9 cm. Better results were achieved with the interposition of femoral head cartilage using punch bi-

opsy plugs of fresh bovine cartilage in one group and frozen bovine cartilage in another. In both the fresh and frozen groups, distal femoral growth continued with the difference being only 0.6 cm from the opposite side. The fresh and frozen cartilage also led to the least extent of valgus angulation, showing 11 ~ Eulert (152), also working on the distal femur of the developing rabbit, eliminated or greatly minimized partial premature distal physeal closure by transplantation of iliac crest growth cartilage. The best results were obtained when the growth cartilage was transplanted alone or with a thin layer of bone. The results were poor when the layer of bone transplanted with the cartilage was too thick. Eulert concluded that iliac crest cartilage could be used as a graft following bone bridge resection. When a region had been resected and no interpositional tissue placed, the mean valgus at 24 weeks postsurgery was 60 ~ When the cartilage graft was inserted with a thick lamella of bone attached, angular deformity was less but still approximately 45 ~. The best results were achieved when the cartilage graft had only an extremely thin rim of bone attached, which led to angular deformity of only 10~ As an extension of this work, early efforts at chondrocyte implantation have been presented, but the results at present are insufficient to be recommended for clinical use. Although some evidence of cartilage survival is seen, no columnation reproductive of a true physis has been identified. (These will be reviewed later.) Bright (71) showed the value of silicone rubber implants in dogs in which they were effective in preventing bone bridge reformation in a distal femoral epiphyseal growth plate model. Wirth et al. (510) even tried implanting periosteum to determine whether it would modulate into a cartilage phase; this did not occur and direct bone formation resulted. Studies have also been performed attempting to define the use of fat grafts in relatively large central defects. Osterman (358) removed approximately 65% of the central part of the plate of the distal femur in a 3-week-old rabbit and interposed autologous fat. This work attempted to show that even large defects involving more than half of the epiphyseal plate can be successfully treated. He subsequently removed the lateral third of the distal femoral growth plate in rabbits and replaced the cartilage with free fat tissue transplants which minimized angular deformity and growth loss compared with controls where the defect was left empty (358).

C. Treatment of Premature Physeal Closure by Means of Growth Plate Transplantation 1. FREE AUTOGENOUS ILIAC CREST PHYSEAL GRAFTSmFOCAL DEFECTS Efforts have been made in our laboratory to reestablish growth by the transplantation of a free partial growth plate after resection (353) (Fig. 27). This procedure is designed to keep the epiphyseal and metaphyseal circulations apart; the partial growth plate also actively contributes to growth rather than serving as a passive spacer, as does fat, silastic, or

F I G U R E 27 A series of illustrations shows the use of focal iliac crest physeal transplants in a rabbit model. (A) Photograph demonstrating the defect in the lateral aspect of the distal femoral physis. (B) Photomicrograph showing bone bridge formation following creation of a physeal defect. The lateral femoral physis had been removed, but no graft inserted, 4 weeks before the animal was sacrificed. The persisting physis (P) is on the left. There is no evidence of an attempt by the physeal cartilage to grow laterally to reform cartilage tissue at the site of the defect. Bone fills the defect site (right) and an extensive bone bridge unites epiphyseal bone (EB) with metaphyseal bone (MB). (C) Photomicrograph illustrating the iliac apophysis and the metaphyseal bone. Fibrocartilaginous layer (FC);

F I G U R E 27 (continued) cartilage of the apophysis (C): cytologically specialized region of the cartilage referred to as the physis (P); and metaphyseal bone (M). The arrow represents the line of demarcation between the green-staining fibrocartilaginous tissue and the red-staining cartilaginous tissue, as demonstrated on a Safranin O-fast green preparation. (D) Photomicrograph of the iliac apophysis after its separation from the metaphysis. Separation has occurred through the zone of hypertrophic cells, with a few spicules of metaphyseal bone persisting on the transplant specimen. (E) Photomicrograph showing an iliac physeal graft in position in the defect immediately after transplantation. The cartilage graft is below and the femoral metaphysis is above. A narrow space is present between the graft and the femoral metaphysis, which ensures a firm fit. (F) Photograph showing an iliac physeal graft in position in the lateral

F I G U R E 27 (continued) femoral physeal defect. (G-N) This series of photomicrographs demonstrates the histological appearance of the physeal graft in relation to the persisting physis and associated epiphyseal and metaphyseal bone following transplantation. Except (G), the photomicrographs are positioned with the epiphyseal areas at the top and the metaphyses below. (G) Low-power photomicrograph illustrating the distal femoral physeal graft at the right 3 months after insertion. Normal growth has occurred and no distal femoral valgus deformity is present. The graft is approximately two times as thick as the persisting nonoperated physeal tissue. Cartilage union has occurred at the graft-physis junction, and there is no continuity between the epiphyseal and metaphyseal bone and no angular deformity. (I-I) Photomicrograph of the medial aspect of the nonoperated distal femoral physis, demonstrating the normal architecture. (I) Photomicrograph from the same specimen as in (I-I), showing physeal and metaphyseal tissue from the lateral physeal transplant. Cartilage and physeal tissue are at the top and metaphyseal tissue at the bottom. Note the excellent orientation of the proliferating and hypertrophic cell layers of the transplanted physis and the smooth junction with the metaphyseal tissues. (J) Lowpower photomicrograph made 3 months after physeal transplantation. The transplanted physis (TP) is at the left and the persisting physis (PP) is at the fight. There is firm coaptation between the two physes. Persisting epiphyseal bone is at the top and metaphyseal bone at the bottom. (K) Higher power photomicrograph from the same specimen illustrated in (J) demonstrating the junction of the epiphyseal bone above and the transplanted cartilage below. Vascular invasion of the transplanted cartilage has not occurred. (L) Photomicrograph illustrating the physeal graft (fight) and the persisting physis (left). Cartilage continuity between the two has been established. At the upper fight, note the bone that has invaded the cartilage part of the transplant as distinct from the physeal part. The metaphyseal bone adjacent to the persisting physis merges imperceptibly with the metaphyseal bone adjacent to the graft. (M) Higher power photomicrograph showing, from top to bottom, epiphyseal bone, bone in the cartilaginous part of the graft, cartilage and physeal cartilage from the graft, and metaphyseal bone produced by the graft. (N) Higher power photomicrograph showing the junction between the persisting physeal cartilage (PP) and the physeal transplant (TP). There is persistence of the proliferating and hypertrophic cell zones in the graft tissue and continuity between metaphyseal bone from both segments. At the upper left, epiphyseal bone remains separate from metaphyseal bone. (O) Photograph demonstrating a poor result 4 months after transplantation (Oi). When the animal was sacrificed, the graft was found to have displaced from the defect site, allowing a bone bridge to form with significant shortening and valgus deformity. The nonoperated control femur is on the left. (Oii) Radiographs of the two bones (image reversed). (P) Photograph illustrating an excellent result 4 months after growth plate transplantation (Pi). The control femur is on the left. (Pii) Radiographs of the two bones illustrated in (Pi) demonstrating an excellent result following transplantation. (The radiograph has been reversed and the transplanted bone is on the left.) [Reprinted from (353), with permission.]

712

CHAPTER 8 9 Lower Extremity Len~trh Discrepancies

methyl methacrylate. This approach takes advantage of the fact that the physeal cartilage itself has no direct blood supply but, rather, a dual supply from epiphyseal vessels on one side and metaphyseal vessels on the other. If the defect in the growth plate is at the periphery, its removal leaves the epiphyseal cartilage and bone, the metaphyseal bone, and, therefore, the dual blood supply intact. In rabbits, oriented cartilage from the iliac crest, transferred after the outer fibrocartilaginous portion has been carefully removed, has been shown to survive and function because the transplanted tissue is nourished by the intact blood supply of the host bone. Focal lesions of the physis continue to cause considerable morbidity in terms of angular deformity and shortening. Osteotomy can correct deformity, although it must be repeated frequently if the growth problem has occurred several years prior to maturity. A variety of procedures have evolved to treat the length discrepancies. In an effort to restore normal growth, well-localized focal bone bridges have been resected and a biological or prosthetic material has been inserted to prevent reformation. Free physeal and epiphyseal transplantation is attractive, but previous experimental and clinical attempts have not been successful consistently. Vascularized epiphyseal transplantation has been used with promising results in experimental animals. The cartilaginous physis has a dual epiphyseal and metaphyseal blood supply. From the epiphyseal vessels, nutrients are carried by diffusion through the cartilaginous extracellular matrix to chondrocytes; the metaphyseal vasculature serves as a source of osteoprogenitor cells that lay down bone on the calcified cartilage matrix to complete the endochondral sequence. Basic requirements for the survival and normal function of free physeal transplants are that this dual blood supply be preserved at the graft site and that the free physeal graft not be covered by tissues that impede the diffusion of nutrients from blood vessels into the extracellular matrix of the physis. We have reported our studies on the transplantation of free autogenous iliac crest physeal grafts into defects in the lateral aspect of the distal femoral physis in rabbits. The graft site in the lateral aspect of the distal end of the femur was carefully fashioned to expose the epiphyseal and metaphyseal bone and its vessels. Fibrocartilaginous and perichondrial tissues were removed from the surfaces of the free physeal graft to facilitate the diffusion of nutrients into the physeal cartilage. The physeal grafts were readily incorporated into the graft site, the morphology of the physis was retained, and the physeal transplant prevented bone bridge formation, growth arrest, and valgus deformity. Focal physeal bone bridges develop most commonly following certain fracture-separations or as medial bridging of the proximal tibial physis in association with severe Blount's disease. In such situations, the epiphyseal and metaphyseal blood supply to the bone adjacent to the physis is generally normal. The results of this study indicate that free physeal transplants might be used to prevent or decrease the bone bridge formation, growth arrest, or angular deformity that can occur in these conditions.

A procedure was first developed for creating a standard focal defect in the lateral aspect of the distal femoral physis of 3- to 4-month-old rabbits, which consistently led to the formation of a bone bridge between the distal femoral epiphysis and metaphysis. By using a dissecting microscope, the defect was created by removing the outer one-half of the lateral half of the physis. The gross appearance of the defect in a midcoronal section of the distal end of the femur is shown in Figure 27A. In creating the defect, care was taken to remove all of the cartilage from the surfaces of the epiphyseal and metaphyseal bone to leave the bone intact and, thus, to preserve the host's dual epiphyseal-metaphyseal blood supply to the physis. In the early stage following creation of the defect, undifferentiated mesenchymal cells migrated into and filled the defect, but did not differentiate into cartilage. Woven bone formed, which then transformed to the lamellar conformation. By 4 weeks after surgery, a bone bridge had formed between the epiphysis and the metaphysis (Fig. 27B). There was no tendency for the remaining physeal cartilage to grow laterally into the defect. At the periphery of the defect a fibrocartilaginous mass was sometimes identified, but this never developed into a growth plate. Seven of the 8 defects that were created resulted in bone bridge formation, growth arrest, and marked valgus deformity. In the 1 rabbit, operated on early, in which normal growth occurred, the dissecting microscope had not been used, and we assumed that the physis had not been removed completely. A procedure was then developed to remove and transplant a free autogenous iliac crest physeal graft into the focal defect created in the lateral part of the femoral physis. The graft was taken from the posterior part of the iliac crest apophysis. The overlying muscle was freed from the crest, and two parallel incisions approximately 2 cm apart were made through the fibrocartilage and physis to the underlying bone. These incisions were carried through the periosteum onto the inner and outer iliac bone surfaces and then were joined by transverse incisions. By elevating the periosteal sleeve, the iliac crest apophysis was gently separated from the metaphysis at the junction of the hypertrophic chondrocytes and the metaphysis and was pulled free of the underlying bone. On occasion, a few spicules of metaphyseal bone came free with the cartilage fragment, but specific attempts to include bone were not made. Histological studies confirmed the separation between the hypertrophic chondrocytes and the metaphysis. The graft was trimmed under the dissecting microscope to remove the periosteal-perichondrial sleeve and much of the overlying fibrocartilage. The graft was then placed in the femoral defect with the correction orientation, and the femoral periosteal sleeve was sutured to the intact femoral periosteum with 4-0 Dermalon to hold the graft in place. The animals were not immobilized postoperatively. The iliac apophysis is composed of a physis, epiphyseal cartilage, and a fibrocartilaginous layer (Fig. 27C). The apophysis was separated gently from the metaphysis of the iliac crest at the junction between the lowermost hypertrophic

SECTION X 9 Direct Operation on Epiphyses to Enhance G r o w t h Potential

chondrocytes and the metaphysis (Fig. 27D). Most of the fibrocartilaginous layer and all of the perichondrium and periosteum were removed by using the dissecting microscope. The physeal graft was then placed in the femoral defect and the overlying femoral periosteal sleeve was sutured into place (Figs. 27E and 27F). In the transplant specimens assessed as early as 1 week postoperatively, the physeal graft appeared to be intact and viable. The morphology and organization of the physis were retained. Union of the graft by cartilage at the graft-host junction was seen as early as 2 weeks. Viability of the growth plate was evident histologically by the production of metaphyseal bone, the maintenance of physeal height and cytological organization, and the absence of vascular invasion or bone formation in the germinal, proliferating, and columnar cell layers. The physis maintained its bright red stain with Safranin O. The residual overlying fibrocartilaginous tissue of the graft underwent vascular invasion and ossification with time, but the growth plate itself appeared immune. A detailed description of the histological characteristics of the transplanted physis is presented in Figs. 27G-27P. In the definitive study of the capacity of the physeal transplant to prevent bone bridge formation, shortening, and valgus deformity, 33 rabbits received transplants. Twenty-seven of the rabbits were sacrificed 21 days or more after surgery. The capacity of the graft to prevent growth arrest or valgus deformity was then assessed by its gross and radiographic appearance and by measurements of the distances between the femoral head and the medial and lateral condyles of the distal end of the femur. The results were classed as excellent, good, fair, or poor, as already defined. Photographs of the gross appearance and radiographs of the femurs demonstrate an obvious, striking difference between an excellent (Figure 27P) and a poor (Figure 270) result. Following transplantation, 16 results were rated as excellent, 3 were good, 4 were fair, and 4 were poor. In the definitive group, therefore, a good or excellent result was seen in 70% of the animals. The type of growth plate defect created in this study led to bone bridge formation and valgus deformity in experiments in which no cartilage was interposed. Similar defects created by other investigators have also led repeatedly to bone bridge formation and angular deformity. There have been reports of defects that involved only a narrow, peripheral portion of the plate in which a bone bridge formed but the bone subsequently yielded to growth, with angular deformity not occurring as the intact physis overcame the small bridge. There also have been reports of small central bone bridges forming but not restricting growth. Thus, there appears to be a relationship between the size of a physeal defect and bone bridge formation and the effect of the bridge on subsequent growth. Because in this study normal growth did not continue in the sham experiments after creation of the growth plate defect alone, the large number of good and excellent results following physeal transplantation is attributable to insertion

713

of the graft. The graft may have functioned simply as a passive barrier to vascularizafion, keeping the epiphyseal compartment separate from the metaphyseal compartment, without actually contributing actively to growth. If so, the absence of deformity might have been due to the persistence of growth of the remaining growth plate alone. Separation of the two circulations is the prime function of any interpositional material, be it biological or prosthetic. Intimate communication between the epiphyseal circulation with its associated osteoprogenitor cells and the metaphyseal circulation with its osteoprogenitor cells allows a bone bridge to form. Examples of passive barriers include heterogeneous deep-frozen hyaline cartilage, fat, silastic, methyl methacrylate, and bone wax. Physeal cartilage is avascular and also possesses an anti-angiogenesis factor, which inhibits vessel ingrowth. The histological studies of the transplants reported here indicate that the morphology, viability, and normal function of the transplanted physis were retained. The proliferating chondrocytes remained organized in orderly rows and contributed to longitudinal growth. The persisting epiphyseal blood supply adjacent to the defect provides nourishment for the resting, germinal, and proliferating cell layers of the transplanted iliac physis; as long as the graft is sufficiently thin, handled gently, and well-positioned, it appears that it will survive. Cartilage union at the persisting physis-graft junction was demonstrated histologically in the several specimens from animals that were sacrificed at early time periods. Endochondral bone formation occurred beneath the physeal transplant in a fashion almost indistinguishable from that of the persisting physis. The importance of a narrow graft must be stressed. As the cells of the growth plate are supplied by diffusion, removal of most of the overlying fibrous and fibrocartilaginous tissue from the iliac crest cartilage graft allowed vascular diffusion from the epiphyseal side to supply the transplant almost immediately. With time, the fibrocartilage remaining on the graft underwent vascular invasion, followed shortly by endochondral bone formation such as occurs in iliac apophyseal ossification prior to skeletal maturation. This bone soon merged with that of the secondary ossification center of the epiphysis. The physeal portion of the graft, however, remained intact without suffering vascular invasion until the time that the entire distal femoral physis reached maturity. A bed of well-vascularized epiphyseal and metaphyseal bone is essential because the nutrition of the free graft is derived exclusively from its surrounding tissue. Lalanandham et al. (281) transplanted scapular cartilage in the rabbit into the distal ulna following curettage of the ulnar physis in animals 5-6 weeks old. They then assessed the viability and biochemical function of the transplanted chondrocytes and their histologic appearance at varying time periods. They concluded that the cartilage transplanted in an avascular fashion could remain viable, synthesize proteoglycan, and also be associated with active growth (although less than normal). Histologic sections at 2 weeks showed the

714

CHAPTER 8 ~ Lower Extremity Length Discrepancies

transplanted block of cartilage to be intact, and by 8 weeks the transplanted cartilage appeared to be participating in endochondral ossification on the metaphyseal side. Histochemical studies from the graft cartilage were consistent at 2, 7, and 14 days and showed lactic dehydrogenase to be strongly present in both transplanted and unoperated control cartilage. Radiosulfate administered intravenously was clearly present throughout the transplanted tissue also at 2, 7, and 14 days. This incorporation was indicative of chondrocyte synthesis of proteoglycan. Thymidine incorporation was seen but was diminished relative to controls. This experimental approach is for well-localized, focal lesions, in this instance involving approximately one-fourth of the physis in the coronal plane and one-half of that segment in the sagittal plane. The shape of the iliac physis must also be considered in relation to the shape of the defective physis, as few long bone physes are as level as diagrammatic and radiographic projections often imply. It appears that full physeal and certainly epiphyseal grafts can be most successful only if they are transplanted with an associated vascular supply. The possibility of using vascularized iliac crest apophyseal transplants is being investigated. In focal physeal lesions, however, when the host epiphyseal and metaphyseal bone is present and well-vascularized, free autogenous iliac crest might be effective clinically as demonstrated in our study. Although not all of the transplants were successful, the dramatic differences demonstrated between the excellent, good, and even fair results in comparison with the poor and sham results clearly indicate that free physeal transplants in the appropriate environment can function well. The graft must be freed from much of the overlying fibrocartilage and surrounding periosteum and perichondrium and must be fitted gently but firmly into narrow defects so that coaptation is intimate. In addition, the host periosteum at the region of the groove of Ranvier should be sutured over the graft to provide both mechanical and physiological support. Because the majority of lesions due to trauma and Blount's disease leave the epiphyseal and metaphyseal bone intact and well-vascularized, free iliac crest physis transplants may be useful in the treatment of focal physeal arrest in patients with such conditions. 2. VASCULARIZEDAUTOGENOUS EPIPHYSEAL ILIAC CREST GRAFTS Because studies of nonvascularized transplants over several decades stressed poor or imperfect results due to failure of rapid revascularization, the use of vascularized transplants held great attraction, and once vascularized bone transplants became feasible investigation was extended to this area. Nettelblad et al. (347) transplanted successfully the proximal one-third of the fibula including the entire epiphysis, adjacent metaphysis, and diaphysis in 22 puppies, showing the feasibility of the technique. In the experimental groups the fibula switch was performed, selecting one fibula as a vascularized graft and the other as a nonvascularized graft. Con-

tinuous growth was observed in the vascularized epiphyseal transplants and in the controls with no statistical difference noted, whereas the nonvascularized transplants exhibited considerably less or no growth. Varying techniques subsequently confirmed the continued viability of the vascularized epiphyseal transplants in contrast to the nonvascularized procedures. Teot and associates (471-473) have reported on their investigations concerning vascularized partial iliac crest growth cartilage transplants. An initial study performed in 50 childhood cadavers and 25 immature dogs assessed more precisely the vascularization of the lower end of the femur and the iliac crest (473). They were able to identify regions of cartilage from the epiphyses that could be transplanted with anastomosis of appropriate pedicles. In a second brief presentation, the value of iliac crest cartilage as a graft source again was reviewed (471). Teot et al. utilized iliac crest vascularized transfers in an experimental sense in 38 puppies replacing approximately 80% of the distal femoral growth plate region. Initial indications were of growth in the range of 85% of that on the opposite normal side. An early report was also made of 1 case in a 2-year-old patient with a total epiphyseal arrest of the distal femur, who had a vascularized iliac crest transplantation after bone bridge resection. The patient was doing well 16 months later but no further follow-up was reported. The second proposed utilization was in reestablishing acetabular growth for the dysplastic hip. Allieu also provided a brief report on the feasibility of iliac crest cartilage and bone pedicle transplants into the acetabular and proximal femur regions (19). The vascular pedicle had its origin from the deep circumflex iliac artery. A more formal presentation of iliac crest pedicle graft transplantation was reported from 48 immature dogs in relation to repairing distal femoral growth plate abnormalities (472). In one group the graft was pedicled on the superficial circumflex iliac vessels and reimplanted in situ. In group 2 the pedicle graft was transferred to the groin area as an island graft. These two control groups demonstrated conservation of growth activity when the graft was pedicled on its epiphyseal vessels. In group 3 the graft was transferred to the distal epiphyseal area of the femur after resection of the portion of the growth plate (approximately two-thirds) located inside the perichondrial ring of Ranvier, with conservation of 80% of the outer cylinder. Microsurgical revascularization was achieved by using the saphenous vessels. In group 4 the latter technique was used without revascularization. Results were far more favorable in terms of growth restoration in the pedicled than in the nonvascularized transplant. In the vascularized distal transplant the possibility remained that some regeneration of cartilage was from the surrounding epiphyseal plate, which had been left intact. Teot et al. concluded that the pedicle graft "appears to act as a catalyst in the formation of a new growth plate, preventing the formation of bony bridges between the epiphysis and metaphysis." Vascularized iliac crest transplantation has been used exper-

SECTION X ~ Direct Operation on Epiphyses to Enhance Growth Potential imentally to augment acetabular growth in dogs with severely deformed hips and damaged femoral growth as a result of epiphyseal lesions. 3. PHYSEAL RECONSTRUCTION USING TISSUE FROM FETAL AND EARLYPOSTNATALEPIPHYSES Zaleske and colleagues have worked extensively on attempting physeal reconstruction with extremely young fetalneonatal tissue in an effort to take advantage of its greater growth potential (36, 127, 424, 511). A series of experiments have been done involving full and partial physeal reconstruction in mice, with most work involving 4-day-old postnatal distal femoral tissue. Whereas growth in length over brief periods of time has tended to be limited to the 25% range, autoradiographic studies using tritiated thymidine show the persistence of cell proliferation after avascular transplantation. Isolated physeal regions appear to maintain their kinetic activity at least in short-term implantations. In the 4-day-old distal mouse femur, which serves both as the source of tissue and as the area into which tissue is implanted, the epiphysis is completely cartilaginous and avascular. The work, summarized by Barr and Zaleske (36), showed that as a group the transplanted physeal blocks resulted in femurs of significantly shorter overall length. Metabolic and kinetic analyses, however, showed both tritiated thymidine and radioactive sulfate incorporation, indicating that cell viability continued even though normal function was not reconstituted. They concluded that blocks of cartilage containing important cell populations can be transplanted in a nonvascularized fashion with at least partial maintenance of viability. The bulk of this transplantation investigation was done with highly inbred strains allowing for syngeneic transplantation. The work was expanded by using complete epiphyseal replacement via knee transplantation in the murine model but utilizing tissue of different developmental times in neonatal mice in an effort to determine whether a specific stage of epiphyseal chondrogenesis led to improved results. Studies were done in 4-day-old postnatal mice, but the distal femoral and proximal tibial chondroepiphyses were transplanted in their entirety from a 4-day-old postnatal mouse, a 1-day-old postnatal mouse, and a 17-day-old fetal mouse. Histologic, metabolic, and kinetic analyses were performed similar to those that had been used previously. The animals were followed for a period of 2 months postsurgery. There was clear variability in morphology and growth from the transplanted syngeneic knee in all experimental groups, but the important observation was felt to be the presence of an unequivocal joint with distal femoral and proximal tibial secondary centers of ossification in adjacent physeal regions, implying some continuing function of the chondroepiphyseal transplant. No differences could be noted, however, between the three groups of slightly differing ages. Becausee the chondroepiphyses were transplanted prior to their vascularization, the fact that they survived and subsequently were vascularized to form secondary ossification centers demonstrated the ability to utilize

715

developmentally immature tissue, which could progress to increased development in its new position.

D. Comparative Studies in Experimental Animal Models of Differing Focal Physeal Implants Several groups have used animal models to compare the effects of interpositional tissues. Martiana et al. (318) used many interpositional tissues in physeal regions of the distal left femur of a 3-month-old rabbit, including muscle, fat, physeal allograft, and iliac apophyseal autograft. A standard defect was created in the lateral distal physis of the left femur in all rabbits. A control group had no interpositional material. At 12 weeks following surgery assessments involved limb length discrepancy and angular deformity. Muscle, fat, and iliac apophyseal autograft had less severe limb length discrepancies and angular deformities than did the control group and the physeal allograft group. In terms of limb length discrepancy and angular deformity, the best results were seen with the iliac crest autograft. The second best tissue in each regard was the interposed muscle, with fat third best. Lee et al. (294) used a model excising the medial half of the proximal tibial epiphyseal growth plate in the rabbit to create a partial growth arrest and then excising the bone bridge and inserting either iliac crest physis, fat, or silastic, into the gap. Their study showed the iliac crest physeal transfer to be superior to silastic, which in turn was superior to fat. The tibias that received free fat as interpositional material developed severe varus angulation and failed to grow in length. The bony bridge redeveloped in all cases in the transferred adipose tissue, contrary to the findings of Langenskiold and his group. In the experimental group with iliac physeal grafts, varus angulation was much less and the amount of longitudinal growth much greater than those of the other groups with fat and silastic. In many of the animals, "the transferred physis remained viable and therefore could have contributed to growth and also could have prevented reformation of the bony bridge through its spacer effect." E. Transplantation of Entire Physes

and Epiphyses 1. EARLYTRANSPLANTATIONEXPERIMENTS1899-1914 Epiphyseal transplantation has intrigued investigators for over 100 years. The results of the initial experimental transplantations of entire growth plates in dogs were almost all poor in allografts, but there were some favorable reports of reimplantation and autograft viability and growth. Haas (208, 209) reviewed in great detail the studies on growth plate transplantation that had been reported in the German literature between 1899 and 1914. Although an optimistic report had been presented concerning the effectiveness of allograft (homoplastic) transplantation of the upper end of

716

CHAPTER 8 ~ Lower Extremity Len9th Discrepancies

the radius, including the epiphyseal cartilage allowing for growth in rabbits, he felt that the interpretation of that work by the authors was inaccurate and remained unconvinced that actual growth had occurred because very little growth occurred normally in the proximal radius. In assessing all of the reports, virtually no growth was noted in any of the allograft transplantations, although on occasion there was some growth in the autotransplantations and often considerable growth in reimplantations. The term reimplantation refers to removing a segment of bone or cartilage and then replacing it in its same site. In none of these approaches were any vascular repairs performed. Among the earliest detailed epiphyseal transplantation studies were those of Helferich (228) and Enderlen (150), who provided separate reports in 1899. The investigators, who worked together, reimplanted the epiphysis in rabbits using the lower epiphyseal cartilage of the ulna with an adjoining piece of epiphysis and diaphysis. Helferich (228) reported on the macroscopic findings that the epiphyseal cartilage, under favorable conditions, need not lose its property of producing length growth. A lessening of this ability was noted, but generally there was not a complete loss of growth. Enderlen (150) reported on the microscopic findings and indicated that some of the reimplanted epiphyseal cartilage remained viable to a large extent, particularly those parts adjacent to the perichondrium. Von Tappeiner (490, 491) performed reimplantations and allograft transplantations on dogs 6 and 12 weeks of age using the distal half of the metatarsal. In the reimplantations he found no disturbance in length growth even after 6 months, whereas in the allograft transplantations growth was markedly diminished. Obata (349) performed reimplantations, autografts, and allografts of the entire metatarsal-phalangeal joint with either a partial or an entire metatarsal and phalanx. He felt that shortening occurred in all cases but that it was most marked in the allograft transplantations. An extensive study was performed by Heller (229, 230), who carried out reimplantations of the distal epiphysis of the radius and ulna as well as homotransplantations (allografts). The epiphysis did not keep its normal length growth in any of the 45 experiments, in which the epiphyseal cartilage was transplanted in the form of half joints. The best results occurred in the reimplantation group, whereas the worst, by which is meant the greatest degree of shortening, occurred in the allograft between nonrelated animals. In the allograft transplantation group there was complete cessation of epiphyseal function. Heller concluded that in autografts there was active regeneration of the epiphyseal cartilage from the perichondrium but that bone growth did not remain at a normal amount. The most favorable conditions for epiphyseal cartilage transplantation would be in the form of a thin sheet of physeal cartilage without adhering bone particles, so that the cartilage could come directly into contact with nourishment from the host. Heller reported almost normal growth in instances in which autograft transplantation had been performed transferring physeal cartilage only. Minoura (328)

transplanted metatarsal-phalangeal joints of 2-month-old rabbits into soft tissue. Varying models were used, but he too concluded that autografts were much superior to allografts and that even with autografts there was not regular growth of the epiphyseal cartilage, even though some of the tissue survived and in no case was there lengthening of the transplanted joint. Axhausen (31) had also transplanted the lower one-fourth of the femur from a growing rat into the subcutaneous tissue of another rat. The physeal cartilage was noted to degenerate such that at 20 days only the peripheral parts remained alive. With further time the entire physis was replaced by fibrous tissue. Similar findings were noted when the same experiment was repeated in the rabbit. Von Tappeiner (490, 491) performed 3 reimplantations and 8 allografts in dogs. In the reimplantations there were practically no changes in the epiphyseal cartilage line, even after 6 months as assessed microscopically. Even in the allograft group some histologic evidence of continuing function was noted. The findings of Von Tappeiner were not confirmed by Haas. Obata (349) described a progressive degeneration of the epiphyseal cartilage line in the reimplantation group until the 50th day, at which time there was some repair of function. He also ascribed almost normal functional properties to this regenerated epiphyseal cartilage. By 70 days, however, degeneration again had occurred, leading Haas to interpret the fact that under some favorable circumstances some of the physes continued to function. Otherwise the changes described by Obata were similar to those by Haas. In Obata's allograft work the epiphyseal cartilage underwent progressive degeneration with practically no regeneration. Heller noted some favorable results in reimplantations. There was some regeneration of the epiphyseal cartilage line and some new growth, but overall the bone growth was retarded. In allograft transplants the results were invariably poor. Haas felt that use of the distal ends of the radius and ulna made the interpretation of one bone transplantation more difficult than in the simpler and more straightforward metacarpal or metatarsal model. The findings of Minoura agreed closely with those of Haas. In none of the experiments was there any increase in the length of the bone after transplantation. 2. HAAS Haas reported on his 75 experimental procedures performed on dogs in relation to the effectiveness of epiphyseal transplantation in maintaining growth (208). The majority of the animals were from 1.5 to 4 months of age at the time of surgery. The metacarpals and metatarsals were selected for transplantation because they had only one epiphysis, making it easy to determine whether growth from the transplanted region had occurred. The bones were also quite stable after repositioning because of the nonoperated adjacent bones. Review of the literature indicated the failure of allograft transplantation to be effective in virtually all studies, and thus Haas concentrated on reimplantation and autograft transplantation. Following reimplantation of an entire metacarpal or metatarsal, which had been removed from its posi-

SECTION X 9 Direct Operation on Epiphyses to Enhance G r o w t h Potential

tion with the articular surface and periosteum intact and then immediately replaced, there was complete cessation of growth of the epiphysis. Autograft transplantation of the entire metacarpal or metatarsal, in which an entire bone was transplanted from one foot to another foot of the same animal, also showed no evidence of growth. No effective growth was seen with split metacarpal reimplantation or autograft transplantation. In the group most likely to succeed, namely, those having reimplantation of the epiphyseal cartilage, the cartilage was transplanted with a piece of adjoining epiphysis and diaphysis, and even here there was no effective growth. A similar approach with autograft transplantation from one foot to the other showed no definite evidence of growth, with the conclusion that after autograft transplantation of the epiphyseal cartilage its function for linear growth was entirely lost or present to only a minimal degree. There was also uniform failure of growth with reimplantation of varying lengths of the epiphyseal end of the metacarpal and metatarsal bones and uniform failure of growth after autotransplantation of varying lengths. Haas' conclusions were quite straightforward; the epiphyseal cartilage lost its power to function after transplantation of either the reimplantation or autograft transplantation type and with the physis transplanted either alone or with an accompanying piece of epiphysis and diaphysis. When Haas used the term epiphyseal cartilage "alone," he still obtained a piece with small adjacent regions of epiphyseal and metaphyseal bone. He concluded that the epiphyseal cartilage was very vulnerable and that its viability was directly dependent upon its blood supply. In no instance was any vascular repair performed with his transplants, and the physeal cartilage rarely had direct access to the persisting vascularity of the host bone because it was always transplanted with a thin rim of bone on the epiphyseal and metaphyseal sides. Haas himself addressed this question in an abstract of the discussion printed immediately following the article. He noted that "in all the excisions of the epiphyseal cartilage a piece of adjoining epiphyseal and diaphyseal bone was removed so as not to injure that particular region." Heller had commented on this matter and felt that in his work he had succeeded in transplanting the epiphyseal cartilage in the form of very thin sheets of cartilage, which subsequently resulted in almost normal growth. Haas felt that Heller's conclusions were incorrect in the sense that he could not truly effectively transplant just cartilage alone, leaving open the possibility that it was the persisting cartilage not removed for transplantation that enabled growth to occur. In a subsequent article Haas reported on the macroscopic, microscopic, and in some cases radiologic details from 58 of the 75 procedures (209). The microscopic description involved not only the epiphyseal growth plate cartilage but also the articular cartilage, the marrow and bone of the secondary ossification center, the marrow and bone of the metaphysis, and cortical bone. With epiphyseal cartilage reimplantation, the first evidence of degeneration of the epiphyseal cartilage was seen at 23 days and appeared as a

717

cleavage line extending through the cartilage columns dividing the physis into a proximal two-thirds and distal one-third. With time progressive degeneration of the epiphyseal cartilage line occurred, and by 85 days there was almost complete disappearance of the physeal cartilage. With autotransplantation the epiphyseal cartilage also showed early degeneration, and at 23 days a line of cleavage was noted through what appeared to be the hypertrophic zone. By 135 days only a slight remnant of physeal cartilage persisted. Following reimplantation with varying lengths of the epiphyseal and of the metacarpal and metatarsal bones, degeneration of the cells at the junction of the proximal two-thirds and distal one-third of the physis was noted followed by fissuring and eventual degeneration of the entire cartilage. With autotransplantation, the epiphyseal cartilage line underwent progressive and complete degeneration. The epiphyseal cartilage line also underwent progressive and complete degeneration with either reimplantation or autograft transplantation of the entire metacarpal or metatarsal bone. Haas reconfirmed his observations that the epiphyseal cartilage ceased to function after either reimplantation or autograft transplantation in each of several approaches. The longitudinal growth ceased in each case. The cartilage degenerated and frequently there were transverse and vertical fissures followed by disappearance of the cells, fibrous substitution, and eventually bone transformation. The only evidence for regeneration was near the periphery beneath the perichondrium, although the new cartilage possessed none of the length performing functions of the normal physeal pattern. Haas concluded that the epiphyseal cartilage was the least transplantable of any of the components of bone due to damage to the vascular supply to the epiphyseal region. Haas later performed another series of transplantation procedures because others continued to describe some effective results with epiphyseal transplantation (210). He concluded again, however, following 20 additional procedures, that "the epiphyseal cartilage plate loses its power of causing length growth after transplantation." Haas defined the potential clinical value of epiphyseal transplantation, but he concluded after extensive and repeated experimentation that longitudinal growth ceased after reimplantation, autograft, or allograft transplantation of the epiphyseal cartilage, whether by itself or with a neighboring piece of epiphyseal or diaphyseal bone. Entire widths of physis in dogs were transplanted, and most of the many variations used were unsuccessful, as Haas recognized, due to failure to provide nutrition to the resting, germinal, and proliferating chondrocytes of the physis. Even transplants of what Haas referred to as the epiphyseal cartilage line, however, included a thin rim of adjoining epiphyseal and metaphyseal bone. 3. PHYSIOLOGIC CONCERNS IN PHYSEAL TRANSPLANTATION Subsequent studies on physeal transplants continued to use entire ulnar, radial, metacarpal, and metatarsal physes, most of which had some epiphyseal bone attached both to

718

CHAPTER 8 ~

Lower Extremity Length Discrepancies

protect the physis from mechanical damage during removal and to ensure its complete transplantation. The results were variable. It is probable that even thin layers of bone attached to physeal transplants on the epiphyseal side inhibit rapid diffusion of nutrients to the physeal cartilage, result in chondrocyte death, and lead to growth plate resorption and replacement by bone. In support of this contention is the report by Heller (229) of good results in physeal transplants when a thin layer of cartilage was transplanted without a sliver of epiphyseal bone attached to the graft, although Haas has pointed out that growth might have been due in such situations to parts of physeal cartilage left behind that continued to function. Similarly good results have been reported with focal cartilage transplants performed using iliac crest cartilage without overlying epiphyseal bone. Additional problems with previously reported experimental methods can be considered. Some grafts appear to have led to poor results after transplantation with the surrounding perichondrium intact. This tissue layer would inhibit nutrition by diffusion from the host tissues, which is essential in free transplants. In addition, most of the studies on free epiphyseal transplantation have involved total physes, which undoubtedly represent too extensive an amount of tissue for consistently good results, especially if epiphyseal bone, perichondrium, and periosteum are left attached and a firm fit in the defect is not achieved. The previously mentioned technical factors, pertaining ultimately to physeal nutrition, have been discussed and appear to be essential considerations for successful free physeal transplants.

Ring then performed comparative experiments involving autograft transplantation within the same animal from right to left and vice versa and allograft transplantations from separate rabbits. Detailed and accurate length determinations of normal rabbit ulnar growth and that of the transplanted animals then were made over a several-week period well beyond skeletal maturation. Histological and radiographic studies were made. Ring concluded that only 5 of the autograft transposition group could be assessed as completely successful out of 18 procedures. The ulnar growth had to exceed 75% of that predicted and stringent radiographic criteria had to prevail to give the successful grading. In the allograft group 26 procedures were performed. By using the criteria for successful transplantation developed for the other two groups, 21 of the 24 were clear failures, and whereas 3 looked to have acceptable radiographic appearances, in each of these the epiphyseal cartilage was narrow and in none of the 3 was the growth in the experimental limb within 75% of the control. Ring thus concluded that allograft transplantation was unsuccessful in all of the animals studied, a conclusion invariably reached by investigators throughout the century. In the autograft transposition group normal growth occurred in only 5 of 18 procedures. Some of the failures were due to technical difficulties in the sense that it was difficult to obtain a snug fit of the transplanted physis into its new host bone. Even when this was present, however, growth failure occurred because of the inability to restore full revascularization in an appropriate period of time.

ETAL. Harris et al. (220) achieved a 50% success rate in autogenous transplants of immature rabbit whole distal ulnar physes by leaving the host perichondrium intact and transplanting only physeal cartilage without a sliver of epiphyseal bone, thus substantiating Heller's results at least partially, as well as the importance of allowing nutrition from the host to support the graft. An extremely careful technique was utilized to obtain the graft. Following sub-periosteal dissection of the distal ulnar metaphysis and adjacent epiphysis for about one-third of its circumference, the epiphysis was broken from the shaft with ease by manual pressure. Because the physis was quite straight there was a constant line of separation through the hypertrophied cells of the zone of provisional calcification, leaving the bulk of the epiphyseal plate attached to the epiphysis. Harris et al. then inserted a small scalpel blade as close to the bone plate of the epiphysis as possible, excising only the cartilaginous physis. Because the germinal layer was relatively thick during early development it was easier to obtain an adequate transplant in the very young animals. They performed reimplantation in 65 rabbits, autogenous transplantation (fight to left) in 65 rabbits, and allograft transplantation using animals paired as closely as possible by weight in 65 animals. A stainless steel wire was then placed through the ulnar shaft approximately 0.5 cm proximal to the plate to serve as a radiographic guide 5. HARRIS

4. RING Interest in physeal transplantation has continued due to the high potential value of such a technique, but daunting difficulties persist. Ring performed reimplantation, autograft, and allograft transplantations of the entire distal ulnar epiphysis in 3- to 6-week-old rabbits (405). He had calculated that approximately 85% of the growth of the ulna occurred distally. He also recognized that, due to the tight bond between the distal radius and the distal ulna, there could be some lengthening of the distal ulnar region after transplantation, which need not represent physeal growth but undoubtedly was due to a mechanical pulling apart of the distal ulna by the continuing growth of the radius. After a period of time, however, failure of ulnar growth would be reflected by tethering of the adjacent radius and its growth in a curved pattern. The ulnar physeal cartilage was removed following knife cuts made transversely through the bony epiphysis and metaphysis, such that the isolated cartilage was transplanted "with its thin attached slivers of bone," following which it was gently freed from the radius and removed together with its surrounding perichondrium. In a series of reimplantation experiments of 9 animals followed for 5 weeks or more, 6 showed little or no shortening and 3 were failures. The reimplantation procedure represents the most favorable in terms of prognosis.

SECTION X ~ Direct Operation on Epiphyses to Enhance Growth Potential for length measurements. Sacrifice ranged from 1 to 84 days with nine time periods in the first 2 weeks and studies then at 21, 28, 56, and 84 days. Studies were histologic. In the allograft transplantations 7 of 13 long-term (56 and 84 days) animals were successful. Normal growth was virtually never noted in comparison to controls regardless of the type of procedure. In the reimplantation group, 11 of 25 available long-term specimens showed normal histological appearance of the plate and 80% or more of anticipated growth. In the allograft transplantations there was a remarkably uniform success rate up to 28 days. All physes subsequently degenerated, however, and were destroyed between the 56th and 84th days. Harris et al. thus concluded that satisfactory survival with up to 80% of normal growth for the 12 weeks of the experiment was achieved in half of the autografts, with homografts fairing quite poorly particularly after the first 4 weeks. Technical features for a good result involved the necessity of including the germinal cells of the physis within the transplant, an adequate supply of tissue fluid (i.e., vascularization) reaching the transplant from the epiphyseal side, and a snug fit of the transplant in the recipient bed. Their interpretation, which appears to be highly accurate, suggested that the transplants had to survive an avascular stage, obtaining nutrition from the surrounding tissues before healing and stabilization occurred such that true longitudinal growth could continue. Care had to be taken to ensure that the recipient bed particularly on the epiphyseal side had been curetted to bleeding bone because it is the vascularity of the host that provides nutrition to the transplanted or re-implanted physeal graft tissue.

6. SILFVERSKIOLD;FARINE Silfverskiold (443) performed ulnar epiphyseal cartilage transfers in the rabbit and found that allografi transplantation produced a complete cessation of growth, whereas autograft exchange produced cessation of growth in 6 of 11 animals and considerable retardation in the rest. Farine (158) also performed distal ulnar transplantation procedures in the young rabbit. Sacrifice in most was from the 1st to the 90th day after surgery with occasional animals followed to 180 days. Operations were performed in 16 rabbits at 5 weeks of age. The transplant included the entire distal ulnar physis with two thin epiphyseal and metaphyseal lamellae of bone. Autograft transposition from fight to left was done in 16, and in a second group an additional 16 had reimplantation procedures performed. The results in the two groups were similar. The subsequent study was histologic and microangiographic. A major part of the cartilage transplant survived in each of the rabbits, and on one occasion there was total survival. Any imperfect result was due to the production of a hematoma, which limited passage of host vessels to the graft and thus limited revascularization. Measurements in terms of physeal function were not made, however. Experimental work in immature rabbits by Zaleske et al. (518) has shown that reimplantation of vascularized whole

719

knee joints, including the epiphyses, can survive and continue growth. At present, however, such procedures have no clinical applicability, primarily because the epiphyses that can be sacrificed for transplantation do not fit the anatomical and mechanical needs of the area to which they would be transferred.

F. Implantation of Chondrocyte Suspensions The ability to grow chondrocytes, which synthesize a cartilage matrix in vitro, has led to the hope that chondrocyte suspensions grown in tissue culture subsequently might be placed within focal physeal defects to reconstitute a functioning physeal region. A few experimental reports of this possible approach have appeared. Hansen et al. (218) reported on the growth of chondrocytes into cartilage disks after culturing isolated epiphyseal chondrocytes from fetal lambs. The best results are achieved in experimental animals using fetal cartilage as the source of chondrocytes. Whereas there is some evidence that a cartilage tissue subsequently forms in the physeal defect and proliferates, there is no evidence that the cytologically specialized physis has been reconstituted. Regardless, this avenue of approach is promising as expertise is improving in cell culture and in providing a mechanical substrate on which the chondrocytes can proliferate.

G. Treatment of Premature Physeal Closure by Means of Physeal Distraction Transphyseal distraction has been used in both experimental and clinical situations to disrupt focal transphyseal bridges. DePablos et al. (32, 138, 139) used the procedure in 30 lambs at 1.5 months old after partial epiphyseal arrest in the distal femur had been induced. Physeal distraction was then performed followed by no subsequent intervention in some animals and the interposition of fat in others. They demonstrated the ability to disrupt the focal physeal bridge by the distraction technique but recommended that fat be interposed because the bridge reformed in all instances in which distraction alone had been used. DePablos et al. recommended physeal distraction to pull apart the bone bridge, and if clinically meaningful amounts of growth were still left, there was value in performing fat interposition to prevent bridge reformation. The need to resect the bridge, however, was bypassed. Connolly et al. (123) applied transphyseal traction to correct acquired growth deformities in the immature dog. Bone bridges were created across the medial distal femoral physis to produce a varus deformity. They were able to bring about correction by transphyseal lengthening both after removal of the bone bridge directly and by not intervening on the bone bridge but stretching and breaking it with the traction procedure. Many of the animals suffered premature growth plate closure. They showed, however, that mechanical epiphyseolysis after a bone bridge had formed offered

720

CHAPTER 8 ~

Lower Extremity Length Discrepancies

the possibility of treating large areas of epiphyseal arrest in order to restore length and correct angulation. Correction of alignment was also performed in two patients with shortening and angular deformity due to enchondromas but without bone bridges. Correction and lengthening were obtained, but premature fusion limited effectiveness. Kershaw and Kenwright (271) were able to pull apart bone bridges by transphyseal distraction, but rabbits sacrificed 3-6 weeks after the distraction showed complete physeal closure, suggesting that distraction epiphyseolysis along with bone bridge disruption would have a high potential for producing premature physeal fusion. Bollini et al. (60) used the Ilizarov device to treat a centrally located bone bridge in the lower tibia of a 10-year-old gift caused by an epiphyseal fracture-separation. The epiphyseolysis occurred on the 4th postoperative day following lengthening or distraction of 0.25 mm per day. The bone bridge, which remained attached to the metaphysis, was surgically removed following distraction and was prevented from recurring following interposition of methyl methacrylate. Follow-up at 2 years showed no recurrence and normal growth. Canadell and DePablos (91) presented four clinical examples of breaking bone bridges by physeal distraction, thus eliminating the need for the complex and relatively inaccurate open resection of such transphyseal tissue. They used an angulated monolateral fixator, which served with distraction both to pull apart the bone bridge and then to allow for angular correction with time. The procedures were performed three times in the distal femur and once in the distal tibia. The physeal bridge broke in each instance a few days after distraction began. In those patients close to growth maturation, no effort was made to interpose tissue. If patients were treated at a younger age, then the possibility of a secondary procedure to insert an interpositional tissue would be strongly considered. Canadell and DePablos made one important technical point. It was important to fracture the bone bridge first with lengthening forces across and parallel to the angulated physis. The distraction would have to be symmetrical, and only after the bridge had broken would angular correction be performed. If angular corrective forces were applied from the beginning, they would subject the healthy part of the physis to compression pressures, which could lead to its permanent damage. Physeal distraction was also applied by Aldegheri et al. (16) to pull apart bone bridges and then allow for angular correction, utilizing the principle of epiphyseal distraction referred to as hemichondrodiatasis. The Orthofix articulated dynamic axial fixator was used to provide for the asymmetric pressures. Two cancellous screws were placed in the epiphysis and two cortical screws in the diaphysis. Earlier work by DeBastiani was quoted to show that it was possible to elongate just the lateral portion of the distal growth plate of the femur, with histologic assessment showing an increase in thickness only in the lateral part of the physis due to cellular hyperplasia and hypertrophy. The deviation was achieved without fracture of the

growth plate. Aldegheri et al. felt that the best results were achieved in posttraumatic deformities when the bone bridge occupied less than 20-30% of the epiphyseal plate. They recommended performance of the procedure with little growth remaining because the possibility of postprocedure growth arrest still existed.

References 1. Abbott LC (1927) The operative lengthening of the tibia and fibula. J Bone Joint Surg 9:128-152. 2. Abbott LC, Crego CH (1928) Operative lengthening of the femur. South Med J 10:823-832. 3. Abbott LC (1931) The operative lengthening of the tibia and the fibula. West J Surg Obst Gynec 39:513-519. 4. Abbott LC (1932) Lengthening of the lower extremities. Cal West Med 36:6-13. 5. Abbott LC, Saunders JB, DeC M (1939) The operative lengthening of the tibia and fibula: A preliminary report on the further development of the principles and technique. Ann Surg 110:961-991. 6. Abbott LC, Gill GG (1942) Valgus deformity of the knee resulting from injury to the lower femoral epiphysis. J Bone Joint Surg 24:97-113. 7. Achterman C, Kalamchi A (1979) Congenital deficiency of the fibula. J Bone Joint Surg 61B:133-137. 8. Adkins JC, Ravitch NM (1979) Neurofibromatosismvon Recklinghausen's Disease. In: Pediatric Surgery, 3rd ed, eds MM Ravitch, CD Benson, E Aberdeen, JG Randolph. pp. 1497-1502, Chicago: Year Book Medical. 9. Ahmadi B, Akbarnia BA, Ghobadi F, Ganjavian MS, Nasseri D (1978) Experience with 141 tibial lengthenings in poliomyelitis and comparison of 3 different methods. Clin Orthop Rel Res 145:150-153. 10. Aitken AP (1940) Overgrowth of the femoral shaft following fracture in childhood. Am J Surg 49:147-148. 11. Aitken GT (1969) Proximal femoral focal deficiency. In: Limb Development and Deformity: Problems of Evaluation and Rehabilitation, ed CA Swinyard. pp. 456-476, Springfield, IL: Charles C. Thomas. 12. Alberty A (1993) Effects of physeal distraction on the vascular supply of the growth area: A microangiographical study in rabbits. J Pediatr Orthop 13:373-377. 13. Alberty A, Peltonen J (1993) Proliferation of the hypertrophic chondrocytes of the growth plate after physeal distraction. Clin Orthop Rel Res 297:7-11. 14. Aldegheri R, Renzi-Brivio L, Agostini S (1989) The callotasis method of limb lengthening. Clin Orthop Rel Res 241: 137-145. 15. Aldegheri R, Trivella G, Lavini F (1989) Epiphyseal distraction. Clin Orthop Rel Res 241:117-127. 16. Aldegheri R, Trivella G, Lavini F (1989) Epiphyseal distraction: Hemichondrodiatasis. Clin Orthop Rel Res 241:128-136. 17. Aldegheri R, Trivella G, Renzi-Brivio L, Tessari G, Agostini S, Lavini F (1988) Lengthening of the lower limbs in achondroplastic patients: A comparative study of four techniques. J Bone Joint Surg 70B:69-73. 18. Allan FG (1948) Bone lengthening. J Bone Joint Surg 30B: 490-505.

References 19. Allieu Y (1982) Indications generales, perspectives d'avenir et conclusion. Rev Chir Orthop Supp II 68:43-44. 20. Amstutz HC, Wilson PD, Jr (1962) Dysgenesis of the proximal femur (coxa vara) and its surgical management. J Bone Joint Surg 44A: 1-24. 21. Anderson WV (1967) Lengthening of the lower limb. Its place in the problem of limb length discrepancy. In: Modem Trends in Orthopedics, volume 5, ed WD Graham. pp. 1-22, London: Butterworths. 22. Anderson M, Green WT, Messner MB (1963) Growth and predictions of growth in the lower extremities. J Bone Joint Surg 45A: 1-14. 23. Anderson M, Green WT, Messner MB (1964) Distribution of lengths of the normal femur and tibia in children from 1 to 18 years of age. J Bone Joint Surg 46A:1197-1202. 24. Ansell BM, Bywaters EGL (1956) Growth in Still's disease. Ann Rheum Dis 15:295-319. 25. Ansell BM, Bywaters EGL (1963) Rheumatoid arthritis (Still's disease). Pediat Clin North Am 10:921-939. 26. Arkin AM, Katz JF (1956) The effects of pressure on epiphyseal growth. The mechanism of plasticity of growing bone. J Bone Joint Surg 38A:1056-1076. 27. Aronson J, Good B, Stewart C, Harrison B, Harp J (1990) Preliminary studies of mineralization during distraction osteogenesis. Clin Orthop Rel Res 250:43-49. 28. Aronson J, Harrison BH, Stewart CL, Harp JH (1989) The histology of distraction osteogenesis using different external fixators. Clin Orthop Rel Res 241:106-116. 29. Aufaure P, Filipe G, Carlioz H (1986) La desepiphysiodese chez l'enfant. Rev Chir Orthop 72:557-565. 30. Axer A, Schiller MG, Segal D, Rzetelny V, Gershuni-Gordon DH (1973) Subtrochanteric osteotomy in the treatment of Legg-Calve-Perthes syndrome. Acta Orthop Scand 44:31-54. 31. Axhausen G (1912) Ueber den histologischen vorgang bei der transplantation von gelenkenden, in besondere ueber die transplantationsfaehigkeit von gelenkknorpel und epiphysenknorpel. Arch Klin Chir 99:1. 32. Azcarate J, dePablos J, Canadell J (1992) Treatment of premature partial physeal closure by means of physeal distraction: An experimental study. J Pediatr Orthop 1:39-44. 33. Barfod B, Christensen J (1958-1959) Fractures of the femoral shaft in children with special reference to subsequent overgrowth. Acta Chir Scand 116:235-250. 34. Barr JS (1948) Growth and inequality of leg length in poliomyelitis. New Engl J Med 238:737-743. 35. Barr JS, Stinchfield AJ, Reidy JA (1950) Sympathetic ganglionectomy and limb length in poliomyelitis. J Bone Joint Surg 32A:793-802. 36. Barr SJ, Zaleske DJ (1992) Physeal reconstruction with blocks of cartilage of varying developmental time. J Pediatr Orthop 12:766-773. 37. Barre PS, Thompson GH, Morrison SC (1985) Late skeletal deformities following meningococcal sepsis and disseminated intravascular coagulation--A case report. J Pediatr Surg 5:584-588. 38. Baskerville PA, Ackroyd JS, Browse NL (1985) The etiology of the Klippel-Trenaunay syndrome. Ann Surg 202: 624-627. 39. Baskerville PA, Ackroyd JS, Thomas ML, Browse NL (1985) The Klippel-Trenaunay syndrome: Clinical radiological and

40. 41.

42. 43. 44.

45.

46. 47. 48. 49. 50. 51. 52. 53.

54.

55. 56.

57. 58. 59. 60.

61.

62. 63.

721

hemodynamic features and management. Br J Surg 72: 232-236. Beals RK (1982) Hemihypertrophy and hemihypotrophy. Clin Orthop Rel Res 166:199-203. Beckwith JB (1969) Macroglossia, omphalocele, adrenal cytomegaly, gigantism, and hyperplastic visceromegaly. Birth Defects 5:188. Bennett RB, Blount WP (1935) Destruction of epiphyses by freezing. J Am Med Assoc 105:661-662. Bertrand P (1951) Technique d'allongement du femur dans les grands raccourcissements. Rev Chir Orthop 37:530-533. Bertrand P, Trillat A (1948) Le traitement des inegalites de longueur des membres inferieurs pendant la croissance. Rev Orthop 34:264-311. Betz RR, Cooperman DR, Wopperer JM, Sutherland RD, White JJ, Schaaf HW, Aschliman MR, Choi IH, Bowen JR, Gillespie R (1990) Late sequelae of septic arthritis of the hip in infancy and childhood. J Pediatr Orthop 10:365-372. Bevan-Thomas WH, Miller EA (1967) A review of proximal focal femoral deficiencies. J Bone Joint Surg 49A:1376-1388. Bianco AJ (1978) Femoral shortening. Clin Orthop Rel Res 136:49-65. Bigelow DR, Ritchie GW (1963) The effects of frostbite in childhood. J Bone Joint Surg 45B:122-131. Bisgard JD (1936) Longitudinal overgrowth of long bones with special reference to fractures. Surg Gyn Obstet 62:823-835. Bisgard JD, Bisgard ME (1935) Longitudinal growth of long bones. Arch Surg 31:568-578. Bjerkreim I (1989) Limb lengthening by physeal distraction. Acta Orthop Scand 2:140-142. Bjerkreim I, Helium C (1983) Femur lengthening using the Wagner technique. Acta Orthop Scand 54:263-266. Blachier D, Trevoux L, Carlioz H (1986) Allongements progressifs du femur selon la technique de Wagner. Rev Chir Orthop 72:495-499. Blair VP, Walker SJ, Sheridan JJ, Schoenecker PL (1982) Epiphysiodesis: A problem of timing. J Pediatr Orthop 2: 281-284. Blount W (1955) Fractures in Children. Baltimore: Williams and Wilkins. Blount WP (1960) Unequal leg length. In: Instructional Course Lectures, The American Academy of Orthopedic Surgeons, volume 17. pp. 218-245, St. Louis: CV Mosby. Blount WP (1971) A mature look at epiphyseal stapling. Clin Orthop Rel Res 77:158-163. Blount WP, Clarke GR (1949) Control of bone growth by epiphyseal stapling. J Bone Joint Surg 31A:464-478. Blount WP, Zeier F (1952) Control of bone length. J Am Med Assoc 148:451-457. Bollini G, Tallet JM, Jacquemier M, Bouyala JM (1990) New procedure to remove a centrally located bone bar. J Pediatr Orthop 10:662-666. Bost FC, Larsen LJ (1956) Experiences with lengthening of the femur over an intramedullary rod. J Bone Joint Surg 38A: 567-584. Bosworth DM (1938) Skeletal distraction of the tibia. Surg Gyn Obstet 66:912-924. Botting TDJ, Scrase WH (1965) Premature epiphysial fusion at the knee complicating prolonged immobilization for congenital dislocation of the hip. J Bone Joint Surg 47B:280-282.

722

CHAPTER 8 9

Lower Extremity Length Discrepancies

64. Bourde CH (1974) Classification des syndromes le KlippelTrenaunay et de Parkes-Weber d'apres les donnees angiographiques. Ann Radiol 17(2):153-160. 65. Bowen JR, Johnson WJ (1984) Percutaneous epiphysiodesis. Clin Orthop Rel Res 190:170-173. 66. Bowen JR, Leahey JL, Zhang ZH, MacEwen GD (1985) Partial epiphysiodesis at the knee to correct angular deformity. Clin Orthop Rel Res 198:184-190. 67. Boyd JB, Mulliken JB, Kaban LB, Upton J, Murray JE (1984) Skeletal changes associated with vascular malformations. Plast Reconst Surg 74:789-795. 68. Brashear HR (1959) Epiphyseal fractures. A microscopic study of the healing process in rats. J Bone Joint Surg 41A: 1055-1064. 69. Breathnach AS (1965) The lower extremity: Pelvis and femur. In: Frazer's Anatomy of the Human Skeleton, 6th ed. pp. 103-131, London: J and A Churchill. 70. Brewer EJ (1970) Juvenile Rheumatoid Arthritis. Philadelphia: WB Saunders. 71. Bright RW (1974) Operative correction of partial epiphyseal plate closure by osseous-bridge resection and silicone rubber implant. J Bone Joint Surg 56A:655-664. 72. Bright RW (1982) Partial growth arrest: Identification, classification, and results of treatment. Orthop Trans 6:65-66. 73. Bright RW, Burstein AH, Elmore SM (1974) Epiphysealplate cartilage. A biomechanical and histological analysis of failure modes. J Bone Joint Surg 56A:688-703. 74. Brockway A (1935) Clinical resume of 46 leg-lengthening operations. J Bone Joint Surg 4:969-977. 75. Brooke JA (1927) Shortening of bones of the leg to correct inequality of length. Surg Gynec Obstet 44:703-706. 76. Broughton NS, Olney BW, Menelaus MB (1989) Tibial shortening for leg length discrepancy. J Bone Joint Surg 71B: 242-245. 77. Bryan RS, Lipscomb PR, Chatterton CC (1958) Orthopedic aspects of congenital hypertrophy. Am J Surg 96: 654-659. 78. Burdick CG, Sifts IE (1923) Fracture of the femur in children. Ann Surg 77:736-753. 79. Burgio GR, Wiedemann HR (1984) Further and new details on the Proteus syndrome. Eur J Pediatr 143:71-73. 80. Bylander B, Hansson LI, Selvik G (1983) Pattern of growth retardation after Blount stapling: A roentgen stereophotogrammetric analysis. J Pediatr Orthop 3:63-72. 81. Bywaters EGL, Ansell BM (1965) Monoarticular arthritis in children. Ann Rheumat Dis 24:116-122. 82. Cady RB, Spadaro JA, Fitzgerald JA, Pinkes J, Albanese SA (1992) The effects of fat interposition for central physeal defects: A histologic study in rabbits. Clin Orthop Rel Res 282: 305-309. 83. Caffey J (1946) Infantile cortical hyperostoses. J Pediatr 29: 541-549. 84. Caffey J, Schlesinger ER (1940) Certain effects of hemophilia on the growing skeleton: Some roentgenographic observations on overgrowth and dysgenesis of the epiphyses associated with chronic hemarthrosis. J Pediatr 16: 549-565. 85. Camera U (1933) 32 casi di accorciamento dell'arto inferiore sano a scopo ortopedicoi, indicazioni, tecnica, risultati. Chir Org Movim 17:569.

86. Campbell CJ, Papademetriou T, Bonfiglio M (1968) Melorheostosis. J Bone Joint Surg 50A: 1281-1304. 87. Campbell WC (1939) Operative Orthopedics, pp. 952-955, St. Louis: CV Mosby. 88. Canadell J (1993) Bone lengthening: Experimental results. J Pediatr Orthop 2B:8-10. 89. Canadell J (1994) Etude prospective et resultats experimentaux des allongements des membres. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp 88-92, Paris: Masson. 90. Canadell J, Aquerelta D, Forriol F (1993) Prospective study of bone lengthening. J Pediatr Orthop 2B:1-7. 91. Canadell J, DePablos J (1985) Breaking bony bridges by physeal distraction. A new approach. Internat Orthop 9: 223-229. 92. Canale ST, Russell TA, Holcomb RL (1986) Percutaneous epiphysiodesis: Experimental study and preliminary clinic resuits. J Pediatr Orthop 6:150-156. 93. Canale ST, Christian CA (1990) Techniques for epiphysiodesis about the knee. Clin Orthop Rel Res 255:81-85. 94. Carlioz H, Pichon F, Barthelemy A, Lebard JP, Filipe G (1980) Allongements progressifs selon la technique de H. Wagner. Resultat et etude critique des 30 premiers cas. Rev Chir Orthop 66:473-483. 95. Carlioz H, Langlais J (1986) Les courbures congenitales de jambe a concavite anterieure 18 observations chez l'enfant. Rev Chir Orthop 72:259-266. 96. Carlson WO, Wenger DR (1984) A mapping method to prepare for surgical excision of a partial physeal arrest. J Pediatr Orthop 4:232-238. 97. Carpenter EB, Dalton JB (1956) A critial evaluation of a method of epiphyseal stimulation. J Bone Joint Surg 38A: 1089-1095. 98. Carroll NC, Grant CG, Hudson R, Gilbert J, Mubarak SJ, Warren R (1980) Experimental observations on the effects of leg lengthening by the Wagner method. Clin Orthop Rel Res 160:250-257. 99. Cassidy JT, Brody GL, Martel W (1967) Monoarticular juvenile rheumatoid arthritis. J Pediatr 70:867-875. 100. Catagni MA, Cattaneo R, DeRosa V (1994) Le traitement de l'hemimelia externe avec la methode d'Ilizarov. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp 177-181, Paris: Masson. 101. Caton J (1994) Les techniques d' allongement. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp. 76-88, Paris: Masson. 102. Caton J, Dumont P, Berard J, Michel CR (1985) Etude des resultats a moyen terme d'une serie de 33 allongement des membres inferieurs selon la technique de H. Wagner. Rev Chir Orthop Suppl II 71:44-48. 103. Caton J, Panisset JC, Rubini J, Fau D, Rachidi I, Michel F, le Genac CR (1993) Progressive limb lengthening with an elongating intramedullary nail. J Bone Joint Surg Supp II 75B: 156. 104. Cattaneo R, Villa A, Catagni MA, Bell D (1990) Lengthening of the humerus using the Ilizarov technique. Clin Orthop Rel Res 250:117-124. 105. Cauchoix J, Rey JC, Heripret G, Cotrel Y, Morel G (1963) L'allongemont du femur dans le traitement des inegalites de longueur des membres inferieurs. Rev Chir Orthop 49: 193-203.

References 106. Cauchoix J, Morel G, Rey JC, Cotrel Y, Ghosez JP (1972) L'allongement extemporane du femur. Rev Chir Orthop 58: 753-774. 107. Cauchoix J, Morel G (1978) One stage femoral lengthening. Clin Orthop Rel Res 136:66-73. 108. Chacha PB, Chong KC (1977) Experience with tibial lengthening in Singapore. Clin Orthop Rel Res 125:100-106. 109. Chan KP, Hodgson AR (1970) Physiologic leg lengthening: A preliminary report. Clin Orthop Rel Res 68:55-62. 110. Chandler D, King JD, Bernstein SM, Marrero G, Koh J, Hambrecht H (1988) Results of 21 Wagner limb lengthenings in 20 patients. Clin Orthop Rel Res 230:214-222. 111. Chapchal G, Zeldenrust J (1948) Experimental research for promoting longitudinal growth of the lower extremities by irritation of the growth region of femur and tibia. Acta Orthop Scand 17:371-396. 112. Choi IH, Kumar SK, Bowen JR (1990) Amputation or limblengthening for partial or total absence of the fibula. J Bone Joint Surg 72A:1391-1399. 113. Clark RD, Donnai D, Rogers J, Cooper J, Baraitser M (1987) Proteus syndrome: An expanded phenotype. Am J Med Genet 27:99-117. 114. Codivilla A (1905) On the means of lengthening, in the lower limbs, the muscles and tissues which are shortened through deformity. Am J Orthop Surg 2:353-369. 115. Cohen J, Cashman WF (1975) Hemihypertrophy of lower extremity associated with multifocal intraosseous hemangioma. Clin Orthop Rel Res 109:155-165. 116. Cohen M, Hayden PW (1979) A newly recognized hemartomatous syndrome. Birth Defects 15:291-296. 117. Cole WH (1922) Results of treatment of fractured femurs in children. With special reference to Bryant's overhead traction. Arch Surg 5:702-716. 118. Coleman SS, Noonan TD (1967) Anderson's method of tibial-lengthening by percutaneous osteotomy and gradual distraction. Experience with thirty-one cases. J Bone Joint Surg 49A:263-279. 119. Coleman SS (1978) Tibial lengthening. Clin Orthop Rel Res 136:92-104. 120. Coleman SS (1978) Lower limb length discrepancy. In: Pediattic Orthopaedics, eds WW Lovell, RB Winter. pp. 850-880, Philadelphia, JB Lippincott. 121. Compere EL (1936) Indications for and against the leglengthening operation: Use of the tibial bone graft as a factor in preventing delayed union, non-union, or late fracture. J Bone Joint Surg 3:692-705. 122. Compere EL, Adams CO (1937) Studies of longitudinal growth of long bones 1. The influence of trauma and the diaphysis. J Bone Joint Surg 4:922-936. 123. Connolly JF, Huurman WW, Lippiello L, Pankaj R (1986) Epiphyseal traction to correct acquired growth deformities. Clin Orthop Rel Res 202:258-268. 124. Connor JM, Horan FT, Beighton P (1983) Dysplasia epiphysialis hemimelica: A clinical and genetic study. J Bone Joint Surg 65B:350-354. 125. Cottalorda J, Bollini G, Jouve JL, Tallet JM, Labriet C, Bouyala JM (1992) Les sequelles des osteoarthrites de hanche en periode de croissance. Rev Chir Orthop 78:544-551. 126. Coventry MB, Johnson EW (1952) Congenital absence of the fibula. J Bone Joint Surg 34A:941-955.

723

127. Cundy PJ, Jofe M, Zaleske DJ, Ehrlich MG, Mankin HJ (1991) Physeal reconstruction using tissue donated from early postnatal limbs in a murine model. J Orthop Res 9:360-366. 128. Cundy P, Paterson D, Morris L, Foster B (1988) Skeletal age estimation in leg length discrepancy. J Pediatr Orthop 8: 513-515. 129. Currarino G, Erlandson ME (1964) Premature fusion of epiphyses in Cooley's amenia. Radiology 83:656-664. 130. DalMonte A, Andrisano A, Manfrini M, Zucchi M (1985) Humeral lengthening in hypoplasia of the upper limb. J Pediatr Orthop 5:202-207. 131. Danziger MB, Kumar A, DeWeese J (1995) Fractures after femoral lengthening using the Ilizarov method. J Pediatr Orthop 15:220-223. 132. Daoud A, Ganansia P, Clement JL, Descamps L, Maestro M (1994) Etiologies des inegalities de longueur des membres infurieures chez l'enfant. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp. 23-30, Paris: Masson. 133. D'Aubigne M, Dubousset J (1971) Surgical correction of large discrepancies in the lower extremities of children and adults. J Bone Joint Surg 53A:411-430. 134. David VC (1924) Shortening and compensatory overgrowth following fractures of the femur in children. Arch Surg 9: 438-449. 135. Dawson WB (1968) Growth impairment following radiotherapy in childhood. Clin Radiol 19:241-256. 136. DeBastiani G, Aldegheri R, Brivio LR, Trivella G (1986) Chondrodiatasis-controlled symmetrical distraction of the epiphyseal plate: Limb lengthening in children. J Bone Joint Surg 68B:550-556. 137. DeBastiani G, Aldegheri R, Renzi-Brivio L, Trivella G (1987) Limb lengthening by callus distraction (callotasis). J Pediatr Orthop 7:129-134. 138. DePablos J, Canadell J (1990) Experimental physeal distraction in immature sheep. Clin Orthop Rel Res 250:73-80. 139. DePablos J, Villas C, Canadell J (1986) Bone lengthening by physial distraction. Internat Orthop 10:163-170. 140. Delloyle C, Delefortrie G, Coutelier L, Vincent A (1990) Bone regenerate formation in cortical bone during distraction lengthening. Clin Orthop Rel Res 250:34-42. 141. Dick HM, Tietjen R (1978) Humeral lengthening for septic neonatal growth arrest. J Bone Joint Surg 60B:1138-1139. 142. Digby KH (1916) The measurement of diaphysial growth in proximal and distal directions. J Anat Physiol 50:187-188. 143. Dimeglio A, Bonnel F (1994) Croissance des membres inferieures. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp. 6-12, Paris: Masson. 144. Dollinger J (1888) Ueber zuruckbleiben im wachsthum der kranken extremitat bei tuberculoser kniegelenksentzundung. Centralbl Chir 149:897. 145. Doyle JR, Smart BW (1963) Stimulation of bone growth by short-wave diathermy. J Bone Joint Surg 45A: 15-24. 146. Dutkowsky JP, Kasser JR, Kaplan LC (1993) Leg length discrepancy associated with vivid cutis marmorata. J Pediatr Orthop 13:456-458. 147. Eastwood BM, Cole WG (1995) A graphic method for timing the correction of leg-length discrepancy. J Bone Joint Surg 77B:743-747.

724

CHAPTER 8 ~ Lower Extremity Length Discrepancies

148. Edinger A, Biedermann F (1957) Kurzes bien--schiefes Becken. Fortschr Rontg 86:754-762. 149. Elmer EB, Ehrlich MG, Zaleske DJ, Polsky C, Mankin HJ (1992) Chondrodiatasis in rabbits: A study of the effect of transphyseal bone lengthening on cell division, synthetic function, and microcirculation in the growth plate. J Pediatr Orthop 12:181-190. 150. Enderlen (1899) Zur reimplantation des resecirten intermediaerknorpels beim kaninchen. Deut Z Chir 51:574-598. 151. Epps CH, Schneider PL (1989) Treatment of hemimelias of the lower extremity. J Bone Joint Surg 7A:273-277. 152. Eulert J (1979) Transplantation du cartilage de la crete iliaque apres desepiphysiodese. Rev Chir Orthop 65:65-75. 153. Evans EB, Smith JR (1959) Bone and joint changes following bums. J Bone Joint Surg 41A:785-799. 154. Eydelshteyn BM, Udalova NI, Bochkarev GF (1973) Dynamics of reparative regeneration after lengthening by the method of distraction epiphyseolysis. Acta Chir Plast 15: 149-154. 155. Eyre-Brook AL (1960) Septic arthritis of the hip and osteomyelitis of the upper end of the femur in infants. J Bone Joint Surg 42B(1): 11-20. 156. Faber FWM, Keessen W, Roermund PMV (1991) Complications of leg lengthening 46 procedures in 28 patients. Acta Orthop Scand 4:327-332. 157. Fairbank TJ (1956) Dysplasia epiphysialis hemimelica. J Bone Joint Surg 38B:237-257. 158. Farine I (1966) Etude experimentale de la transplantation du cartilage conjugal. Rev Chir Orthop 52:669-680. 159. Farmer AW, Laurin CA (1960) Congenital absence of the fibula. J Bone Joint Surg 42A:1-12. 160. Fernandez F, Pueyo I, Jimenez JR, Vigil E, Guzman A (1981) Epiphysiometaphyseal changes in children after severe meningococcic sepsis. AJR 136:1236-1238. 161. Ferran JL, Couture A, Veyrac C, Baud C, Dimeglio A (1994) Place de la radiologie dans l'exploration des asymetries de longueur des membres inferieures. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp. 31-44, Paris: Masson. 162. Fischer EG, Strand RD, Shapiro F (1984) Congenital hemihypertrophy and abnormalities of the cerebral vasculature. J Neurosurg 61:163-168. 163. Fishbane BM, Riley LH (1978) Continuous transphyseal traction: Experimental observations. Clin Orthop Rel Res 136: 120-124. 164. Fjeld TO, Steen H (1990) Growth retardation after experimental limb lengthening by epiphyseal distraction. J Pediatr Orthop 10:463-466. 165. Franke J, Hein G, Simon M, Hauch St (1990) Comparison of distraction epiphyseolysis and partial metaphyseal corticotomy in leg lengthening. Internat Orthop 14:405-413. 166. Frantz CH (1971) Epiphyseal stapling: A comprehensive review. Clin Orthop Rel Res 77:149-157. 167. Frantz CH, Delgado S (1966) Limb-length discrepancy after third-degree burns about the foot and ankle. J Bone Joint Surg 48A:443-450. 168. Frantz CH, O'Rahilly R (1961) Congenital skeletal limb deficiencies. J Bone Joint Surg 43A:1202-1224. 169. Freiberg AH (1912) Codivilla's method of lengthening the lower extremity. Surg Gyn Obstet 14:614-617.

170. Friberg O (1983) Clinical symptoms and biomechanics of lumbar spine and hip joint in leg length inequality. Spine 8:643-651. 171. Gabriel KR, Crawford AH, Roy DR, True MS, Sauntry S (1994) Percutaneous epiphyseodesis. J Pediatr Orthop 14: 358-362. 172. Gage JR, Cary JM (1980) The effects of trochanteric epiphyseodesis on growth of the proximal end of the femur following necrosis of the capital femoral epiphysis. J Bone Joint Surg 62A:785-794. 173. Galardi G, Comi G, Lozza L, Marchettini P, Novarina M, R Facchini, Paronzini A (1990) Peripheral nerve damage during limb lengthening. J Bone Joint Surg 72B:121-124. 174. Garson JG (1879) Inequality in length of the lower limbs. J Anat Physiol 13:502-507. 175. Gelmetti C, Schianchi R, Ermacora E (1987) Cutis marmorata telangiectatica congenita. Ann Dermatol Venereol 114: 1517-1528. 176. Gibson PH, Papainnau T, Kenwright T (1983) The influence on the spine of leg-length discrepancy after femoral fracture. J Bone Joint Surg 65:584-588. 177. Gil-Albarova J, dePablos J, Franzeb M, Canadell J (1992) Delayed distraction in bone lengthening. Improved healing in lambs. Acta Orthop Scand 6:604-606. 178. Giles LGF, Taylor JR (1981) Low-back pain associated with leg length inequality. Spine 6:510-521. 179. Gill GG (1944) The cause of discrepancy in length of the limbs following tuberculosis of the hip in children: Arrest of growth from premature central closure of the epiphyseal cartilages about the knee. J Bone Joint Surg 26:272-281. 180. Gill GG, Abbott LC (1942) Practical method of predicting the growth of the femur and tibia in the child. Arch Surg 45: 286-315. 181. Glorion C, Pouliquen JC, Langlais J, Ceolin JL, Kassis B (1995) Allongement de femur par callotasis. Etude d'une serie de 79 cas chez l'enfant et l'adolescent. Rev Chir Orthop 81:147-156. 182. Glorion C, Pouliquen JC, Langlais J, Ceolin JL, Kassis B (1996) Femoral lengthening using the callotasis method: Study of the complications in a series of 70 cases in children and adolescents. J Pediatr Orthop 16:161-167. 183. Gloviczki P, Hollier LH, Telander RL, Kaufman B, Bianco AJ, Stickler GB (1983) Surgical implications of KlippelTrenaunay syndrome. Ann Surg 197:353-362. 184. Goel A, Loudon J, Nazare A, Rondinelli R, Hassanein K (1997) Joint moments in minor limb length discrepancy: A pilot study. Am J Orthop, Dec:852-856. 185. Goff CW (1954) Legg-Calve-Perthes Syndrome and Related Osteochondroses of Youth. Springfield, IL: Charles C. Thomas. 186. Goff CW (1967) Histologic arrangements from biopsies of epiphyseal plates of children before and after stapling. Am J Orthop May 87-89. 187. Gofton JP (1971) Studies in osteoarthritis of the hip: Part IV. Biomechanics and clinical considerations. Can Med Assoc J 104:1007-1011. 188. Gofton JP, Trueman GE (1971) Studies in osteoarthritis of the hip: Part II. Osteoarthritis of the hip and leg-length disparity. Can Med Assoc J 104:791-799. 189. Goidanich IF, Campanacci M (1962) Vascular hamartomata and infantile angioectatic osteohyperplasia of the extremities. J Bone Joint Surg 44A:815-842.

References 190. Gotz J, Schellmann WB (1975) Kontinuierliche verlangerung des femur bei intramedullarer stabilisierung. Arch Orthop Unfall-chir 82:305-310. 191. Granberry WM, Janes JM (1963) The lack of effect of microwave diathermy on rate of growth of bone of the growing dog. J Bone Joint Surg 45A:773-777. 192. Grant JCB (1972) An Atlas of Anatomy, 6th ed. Baltimore: Williams and Wilkins. 193. Green WT, Anderson M (1947) Experiences with epiphyseal arrest in correcting discrepancies in length of the lower extremities in infantile paralysis. A method of predicting the effect. J Bone Joint Surg 29:659-675. 194. Green WT, Anderson M (1955) The problem of unequal leg length. Pediat Clin North Am 2:1137-1155. 195. Green WT, Anderson M (1957) Epiphyseal arrest for the correction of discrepancies in length of the lower extremities. J Bone Joint Surg 39A:853-872. 196. Green WT, Anderson M (1960) Skeletal age and the control of bone growth. In: Instructional Course Lectures, The American Academy of Orthopaedic Surgerons, volume 17. pp. 199217, St. Louis: CV Mosby. 197. Green WT, Wyatt GM, Anderson M (1946) Orthoroentgenography as a method of measuring the bones of the lower extremities. J Bone Joint Surg 28:60-65. 198. Greiff J, Bergmann F (1980) Growth disturbance following fracture of the tibia in children. Acta Orthop Scand 51: 315-320. 199. Greulich WW, Pyle SI (1959) Radiographic Atlas of Skeletal Development of the Hand and Wrist, 2nd ed. Stanford, CA: Stanford University Press. 200. Griffin PP, Tachdjian MO, Green WT (1963) Pauciarticular arthritis in children. J Am Med Assoc 184:23-28. 201. Griffin PP, Anderson M, Green WT (1972) Fractures of the shaft of the femur in children. Treatment and results. Orthop Clin North Am 3:213-224. 202. Gross RH (1983) Leg length discrepancy in marathon runners. Am J Sports Med 11:121-124. 203. Gross RH (1971) An evaluation of tibial lengthening procedures. J Bone Joint Surg 53A:693-700. 204. Guarniero R, Montenegro NB, Guarnieri MV (1993) Comparative study of Ilizarov, Wagner, and Anderson methods for limb lengthening (100 lengthenings in 98 patients). J Pediatr Orthop 2B:28-34. 205. Guichet JM, Grammont P, Casar RS, Alexander H, Frankel VH (1994) Proprietes mecaniques du cloud' allongement progressif. In: Les inegalites de langueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon, pp. 136-138, Paris: Masson. 206. Guidera KJ, Brinker MR, Kousseff BG, Helal AA, Pugh LI, Ganey TM, Ogden JA (1993) Overgrowth management in Klippel-Trenaunay-Weber and Proteus syndromes. J Pediatr Orthop 13:459-466. 207. Gullickson G, Olson M, Kottke FJ (1950) The effect of paralysis of one lower-extremity on bone growth. Arch Phys Med 31:392-400. 208. Haas SL (1915) The experimental transplantation of the epiphysis. J Am Med Assoc 65:1965-1971. 209. Haas SL (1916) The transplantation of the articular end of bone including the epiphyseal cartilage line. Surg Gyn Obstet 23:301-332.

725

210. Haas SL (1931) Further observation on the transplantation of the epiphyseal cartilage plate. Surg Gyn Obstet 52: 958-963. 211. Haas SL (1945) Retardation of bone growth by a wire loop. J Bone Joint Surg 27:25-33. 212. Haas SL (1948) Mechanical retardation of bone growth. J Bone Joint Surg 30A:506-512. 213. Haboush EJ, Finkelstein H (1932) Leg lengthening with new stabilizing apparatus. J Bone Joint Surg 14:807-821. 214. Hadlow AT, Nicol RO (1990) A formula for diaphyseal limb lengthening. J Bone Joint Surg 72B: 136. 215. Hallel T, Salvati EA (1978) Septic arthritis of the hip in infancy: End result study. Clin Orthop Rel Res 132:115-128. 216. Hamanishi C, Tanaka S, Tamura K (1992) Early physeal closure after femoral chondrodiatasis: Loss of length gain in 5 cases. Acta Orthop Scand 2:146-149. 217. Hamanishi C, Yasuwaki Y, Kikuchi H, Tanaka S, Tamura K (1992) Classification of the callus in limb lengthening. Radiographic study in 35 limbs. Acta Orthop Scand 4:430-433 218. Hansen AL, Foster BK, Gibson GJ, Binns GF, Wiebkin OW, Hopwood JJ (1990) Growth-plate chondrocyte cultures for reimplantation into growth-plate defects in sheep. Clin Orthop Rel Res 256:286-298. 219. Harris RI (1930) The effect of lumbar sympathectomy on the growth of legs shortened from anterior poliomyelitis. J Bone Joint Surg 12:859-866. 220. Harris RI, McDonald JL (1936) The effect of the lumbar sympathectomy upon the growth of legs paralyzed by anterior poliomyelitis. J Bone Joint Surg 1:35-45. 221. Harris WR, Martin R, Tile M (1965) Transplantation of epiphyseal plates. J Bone Joint Surg 47A:897-914. 222. Hasse C, Dehner (1893) Unsere truppen in korperlicher beziehung. Arch Anat Entwick 249-256. 223. Hechard P, Carlioz H (1978) Methode pratique de prevision des inegalities de longeur des membres inferieurs. Rev Chir Orthop 64:81-87. 224. Hedberg E (1944) Femoral fractures in children. Acta Chir Scand 90:568-588. 225. Hedstrom O (1969) Growth stimulation of long bones after fracture or similar trauma. A clinical and experimental study. Acta Orthop Scand, Supp 122. 226. Heikel HVA (1961) Has epiphysiodesis in one end of a long bone a growth stimulating effect on the other end (An experimental study)? Acta Orthop Scand 31:18-23. 227. Heim M, Horoszowski H, Martinowitz U (1985) Leg-length inequality in hemophilia. Clin Pediatr 24:600-602. 228. Helferich (1899) Versuche ueber die transplantation des intermediaerknorpels wachsender roehrenknochen. Deut Z Chir 51:564-573. 229. Heller E (1914) Experimentelle untersuchungen uber die transplantation des intermediarknorpels in form der halbseitigen gelenktransplantation. Arch Klin Chir 104:843-932. 230. Heller E (1918) Versuche uber die Transplantation der Knorpelfuge. Arch Klin Chir 109:1-62. 231. Henkel L, Willert HG (1969) Dysmelia. J Bone Joint Surg 51B:399-414. 232. Hentschel (1908) Z Orthop Chir 22:1647. 233. Herzog B, Affolter P, Joni L (1976) Late results following marrow nailing of femur fractures in childhood. Z Kinderchir 19:74-80.

726

CHAPTER 8 9

Lower Extremity Length Discrepancies

234. Hiroshima K, Kurata Y, Nakamura M, Ono K (1984) Balland-socket ankle joint: Anatomical and kinematic analysis of the hindfoot. J Pediatr Orthop 4:564-568. 235. Hofmann A, Wenger DR (1981) Posteromedial bowing of the tibia. J Bone Joint Surg 63A:384-388. 236. Hood RW, Riseborough EJ (1981) Lengthening of the lower extremity by the Wagner method. J Bone Joint Surg 63A: 1122-1131. 237. Hootnick D, Boyd NA, Fixsen JA, Lloyd-Roberts GC (1977) The natural history and management of congenital short tibia with dysplasia or absence of the fibula. J Bone Joint Surg 59B:267-271. 238. Hope PG, Crawfurd EJP, Catterall A (1994) Bone growth following lengthening for congenital shortening of the lower limb. J Pediatr Orthop 14:339-342. 239. Horton BT (1932) Hemihypertrophy of extremities associated with congenital arteriovenous fistula. J Am Med Assoc 98: 373-377. 240. Horton GA, Olney BW (1996) Epiphysiodesis of the lower extremity: Results of the percutaneous technique. J Pediatr Orthop 16:180-182. 241. Hult L (1954) Cervical, dorsal and lumbar spinal syndromes. Acta Orthop Scand, Supp 17:1-102. 242. Huxley JS (1932) Problems of Relative Growth. London: Methuen. 243. Ilizarov GA (1988) The principles of the Ilizarov method. Bull Hosp Joint Dis Orthop Inst 48:1-35. 244. Ilizarov GA (1989) The tension-stress effect on the genesis and growth of tissues. Part I: The influence of stability of fixation and soft-tissue preservation. Clin Orthop Rel Res 238:249-281. 245. Ilizarov GA (1989) The tension-stress effect on the genesis and growth of tissues. Part II: The influence of the rate and frequency of distraction. Clin Orthop Rel Res 239:263-285. 246. Ilizarov GA (1990) Clinical application of the tension-stress effect for limb lengthening. Clin Orthop Rel Res 250:8-26. 247. Ilizarov GA (1992) Transosseous Osteosynthesis. Theoretical and Clinical Aspects of the Regeneration and Growth of Tissue. Berlin, Heidelberg: Springer Verlag. 248. Ilizarov GA, Soybelman LM (1969) Some clinical and experimental data on the bloodless lengthening of lower limbs. Exp Khir Anes 4:27-32. 249. Jackson DR, Lyne ED (1979) Infantile cortical hyperostosis. J Bone Joint Surg 61A:770-772. 250. Jaffe HL (1972) Metabolic, Degenerative and Inflammatory Diseases of Bones and Joints. p. 806, Philadelphia: Lea and Febiger. 251. Janes JM, Jennings WK, Jr (1961) Effect of induced arteriovenous fistula of leg length: 10-year observations. Proc Staff Meet Mayo Clin 36:1-11. 252. Janes JM, Musgrove JE (1950) Effect of arteriovenous fistula on growth of bone. An experimental study. Surg Clin N Am 30:1191-1200. 253. Janes JM, Sweeting RC (1971) Experiences with discrepancy in length of lower extremities. Mayo Clin Proc 46:489-491. 254. Jani L (1973) Tierexperimentelle studie uber tibiaverlangerung durch distraktionsepiphyseoloyse. Z Orthop 111:627-630. 255. Jani L (1975) Die distrakionsepiphyseolyse tierexperimentelle studies zum problem der beinverlangerung teil 1. Z Orthop 113:189-208.

256. Jansen K (1957) Inhibition and stimulation of growth. Acta Orthop Scand 26:308-316. 257. Jenkins DHR, Cheng EHF, Hodgson AR (1975) Stimulation of bone growth by periosteal stripping: A clinical study. J Bone Joint Surg 57B:482-484. 258. Jones D, Barnes J, Lloyd-Roberts GC (1978) Congenital aplasia and dysplasia of the tibia with intact fibula. J Bone Joint Surg 60B:31-39. 259. Judet J, Judet R, Rigault P, Plumerault J (1969) Traitement des inegalites des membres inferieurs. Allongement du tibia. Mem Acad Chir 95:532-536. 260. Kaelin AJ (1994) Les pieges, les techniques. In: Les inegalites de longueur des membres, eds A Dimeglio, J Caton, C Herisson, L Simon. pp 157-164, Paris, Masson. 261. Kalamchi A, Dawe RV (1985) Congenital deformity of the tibia. J Bone Joint Surg 67B:581-584. 262. Karger C, Guille JT, Bowen JR (1993) Lengthening of congenital lower limb deficiencies. Clin Orthop Rel Res 291: 236-245. 263. Katzman H, Waugh T, Berdon W (1969) Skeletal changes following irradiation of childhood tumors. J Bone Joint Surg 51A:825-842. 264. Kaufman KR, Miller LS, Sutherland DH (1996) Gait asymmetry in patients with limb-length inequality. J Pediatr Orthop 16:144-150. 265. Kawamura B, Hosono S, Takahashi T, Yano T, Kobayashi Y, Shibata N, Shinoda Y (1968) Limb lengthening by means of subcutaneous osteotomy: Experimental and clinical studies. J Bone Joint Surg 50A:851-963. 266. Kawamura B, Hosono S, Takahashi T (1981) The principles and technique of limb lengthening. Internat Orthop 5:69-83. 267. Kay HW, Day HJB, Henkel HL, Kruger LM, Lamb DW, Marquardt E, Mitchell R, Swanson AF, Willert HG (1975) The proposed international terminology for the classification of congenital limb deficiencies. Dev Med Child Neurol Supp 34, 17:1-12. 268. Kellerova E, Delius W, Olerud S, Strom G (1970) Changes in the muscle and skin blood flow following lower leg fracture in man. Acta Orthop Scand 41:249-260. 269. Kempf I, Grosse A, Abalo C (1986) Locked intramedullary nailing: Its application to femoral and tibial axial, rotational, lengthening, and shortening osteotomies. Clin Orthop Rel Res 212:165-173. 270. Kenwright J, Albinana J (1991) Problems encountered in leg shortening. J Bone Joint Surg 73B:671-675. 271. Kershaw CJ, Kenwright J (1993) Epiphyseal distraction for bony bridges: A biomechanical and morphologic study. J Pediatr Orthop 13:46-50. 272. Kestler OC (1947) Unclassified premature cessation of epiphyseal growth about the knee joint. J Bone Joint Surg 29: 788-797. 273. Kettelkamp DB, Campbell CJ, Bonfiglio M (1966) Dysplasia epiphysealis hemimelia: A report of fifteen cases and a review of the literature. J Bone Joint Surg 48A:746-765. 274. Kingma MJ (1965) Overgrowth in hemophilia. Clin Orthop Rel Res 39:199-204. 275. Klippel M, Trenaunay P (1900) Du noevus variqueux osteohypertrophique. Arch Gen Med 185:641-672. 276. Kojimoto H, Yasui N, Goto T, Matsuda S, Shimomura Y (1988) Bone lengthening in rabbits by callus distraction: The

References

277.

278.

279. 280. 281.

282.

283.

284.

285.

286.

287. 288.

289.

290. 291.

292.

293.

294.

295. 296.

role of periosteum and endosteum. J Bone Joint Surg 70B: 543-549. Koman LA, Meyer LC, Warren FH (1982) Proximal femoral focal deficiency: Natural history and treatment. Clin Orthop Rel Res 162:135-143. Krakovits G (1967) Uber die auswirkung einer beinverkurzung auf die statik und dynamik des huftgelenkes. Z Orthop 102:418-423. Kruger LM (1980) Recent advances in surgery of lower limb deficiencies. Clin Orthop Rel Res 148:97-105. Kuhns JG, Swaim LT (1932) Disturbances of growth in chronic arthritis in children. Am J Dis Child 43:1118-1133. Lalanandham T, Ehrlich MG, Zaleske DJ, Deeney VF, Mankin HJ (1990) Viability and metabolism of cartilage transplanted to physeal regions. J Pediatr Orthop 10:450-458. Lampe HIH, Swierstra BA, Diepstraten AFM (1992) Timing of physiodesis in limb length inequality: The straight line graph applied in 30 patients. Acta Orthop Scand 63: 672-674. Lange DR, Schoenecker PL, Baker CL (1978) Proximal femoral focal deficiency: Treatment and classification in fortytwo cases. Clin Orthop Rel Res 135:15-25. Langenbeck B (1869) Ueber krankhaftes langenwachsthum der rohrenknochen und seine verwerthung fur die chirurgische praxis. Berl Klin Wchnschr 6:265-270. Langenskiold A (1967) The possibilities of eliminating premature partial closure of an epiphyseal plate caused by trauma or disease. Acta Orthop Scand 38:267-279. Langenskiold A (1975) An operation for partial closure of an epiphyseal plate in children, and its experimental bases. J Bone Joint Surg 57:325-330. Langenskiold A (1981) Surgical treatment of partial closure of the growth plate. J Pediatr Orthop 1:3-11. Langenskiold A, Videman T, Nevalainen T (1986) The fate of fat transplants in operations for partial closure of the growth plate. J Bone Joint Surg 68B:234-238. Lascombes P, Membre H, Prevot J, Barrat E (1991) Histomorphometrie du regenerat osseux dans les allongements des membres selon la technique d'Ilizarov. Rev Chir Orthop 77: 141-150. Laurin CA, Favreau JC (1963) Rheumatoid disease in children. Can Med Ass J 89:288-301. Le Coeur P (1963) Egalisation des membres inferieurs par allongement avec fixation immediate. Rev Chir Orthop 49: 217-227. Lee DY, Choi IH, Chung CY, Chung PH, Chi JG, Suh YL (1993) Effect of tibial lengthening on the gastrocnemius muscle. Acta Orthop Scand 64:688-692. Lee DY, Chung CY, Choi IH (1993) Longitudinal growth of the rabbit tibia after callotasis. J Bone Joint Surg 75B: 898-903. Lee EH, Gao GX, Bose K (1993) Management of partial growth arrest: Physis, fat or silastic? J Pediatr Orthop 13: 368-372. Lefort J, Carlioz H, Pere C (1976) Aplasies du perone et malformations associees. Rev Chir Orthop 62:621-634. Lennox DW, Goldner RD, Sussman MD (1983) Cartilage as an interposition material to prevent transphyseal bone bridge formation: An experimental model. J Pediatr Orthop 3:207-210.

727

297. Letts RM, Meadows L (1977) Epiphysiolysis as a method of limb lengthening. Clin Orthop Rel Res 133:230-237. 298. Levander G (1929) Uber die Behandlung von Bruchen des Oberschenkelschaftes. Acta Chir Scand Supp 12:1-237. 299. Lewis RJ, Marcove RC, Rosen G (1977) Ewing's sarcoma-Functional effects of radiation therapy. J Bone Joint Surg 59A:325-331. 300. Liedberg E, Persson B (1978) Technical aspects of midshaft femoral shortening with Kuntscher nailing. Clin Orthop Rel Res 136:62-65. 301. Little DG, Nigo L, Aiona MD (1996) Deficiencies of current methods for the timing of epiphysiodesis. J Pediatr Orthop 16:173-179. 302. Lohr W (1930) Die verschiedenheit der auswirkung gleichartiger bekannter schaden auf den knochen jugendlicher und erwachsener, gezeigt an epiphysenstorungen nach erfrierungen und bei der hamophilie. Zentral f Chir 57:898-909. 303. Lokietek W, Legaye J, Lokietek JC (1991) Contributing factors for osteogenesis in children's limb lengthening. J Pediatr Orthop 11:452-458. 304. Lombardo SJ, Harvey JP (1977) Fractures of the distal femoral epiphyses. Factors influencing prognosis: A review of 34 cases. J Bone Joint Surg 59A:742-751. 305. Luke DL, Schoenecker PL, Blair VP, III, Capelli AM (1992) Fractures after Wagner limb lengthening. J Pediatr Orthop 12: 20-24. 306. MacEwen GD, Case JL (1967) Congenital hemihypertrophy: A review of 32 cases. Clin Orthop Rel Res 50:147-150. 307. Maffulli N, Fixsen JA (1995) Muscular strength after callotasis limb lengthening. J Pediatr Orthop 15:212-216. 308. Magnuson PB (1913) Lengthening shortened bones of the leg by operation. Surg Gyn Obstet 17:63-71. 309. Makarov MR, Delgado MR, Birch JG, Samchukov ML (1996) Intraoperative SSEP monitoring during external fixation procedures in the lower extremities. J Pediatr Orthop 16:155-160. 310. Malan E, Puglionisi A (1964) Congenital angiodysplasias of the extremities. J Cardiovasc Surg 5:87-130. 311. Malan E, Puglionisi A (1965) Congenital angiodysplasias of the extremities. J Cardiovasc Surg 6:255-345. 312. Malhis TM, Bowen JR (1982) Tibial and femoral lengthening: A report of 54 cases. J Pediatr Orthop 2:487-491. 313. Mallet J (1975) Les epiphysiodeses partielles traumatiques de 1'extremite inferieure du tibia chez l'enfant. Rev Chir Orthop 61:5-16. 314. Manning C (1978) Leg lengthening. Clin Orthop Rel Res 136:105-110. 315. Maresh MM (1955) Linear growth of long bones of extremities from infancy through adolescence. Am J Dis Child 89: 725-742. 316. Maroteaux P (1988) Les asymetries corporelles et les hypertrophies partielles. Ann Dermatol Venereol 115:221-228. 317. Marshall JH, Bradish CF (1993) Callotasis in melorheostosis: A case report. J Bone Joint Surg 75B:155. 318. Martiana K, Low CK, Tan SK, Pang MWY (1996) Comparison of various interpositional materials in the prevention of transphyseal bone bridge formation. Clin Orthop Rel Res 325:218-224. 319. Martin-Ferrero MA, Sanchez-Martin MM (1986) Prediction of overgrowth in femoral shaft fractures in children. Internat Orthop 10:89-93.

728

CHAPTER 8 ~

Lower Extremity Len~lth Discrepancies

320. Masserman RL, Peterson HA, Bianco AJ (1974) Congenital pseudoarthrosis of the tibia: A review of the literature and 52 cases from the Mayo Clinic. Clin Orthop Rel Res 99:140-145. 321. McCarthy RE (1985) Leg length inequality associated with catheter use in neonates. J Pediatr Orthop 5:742. 322. McKibbin B, Ray RD (1967) Experimental study of peripheral circulation and bone growth: The pattern of venous return in experimental arteriovenous fistulae, Part II. Clin Orthop Rel Res 53:175-183. 323. McNeill TW, Ray RD (1974) Hemangioma of the extremities: Review of 35 cases. Clin Orthop Rel Res 101:154-166. 324. Mears TW, Vesely DG, Kennedy H, III (1963) Effect of surgically induced arteriovenous fistula on leg length inequality. Clin Orthop Rel Res 30:152-161. 325. Menelaus MB (1966) Correction of leg length discrepancy by epiphyseal arrest. J Bone Joint Surg 48B:336-339. 326. Mileski RA, Garvin KL, Huurman WW (1995) Avascular necrosis of the femoral head after closed intramedullary shortening in an adolescent. J Pediatr Orthop 15:24-26. 327. Millis MB, Hall JE (1979) Transiliac lengthening of the lower extremity. A modified innominate osteotomy for the treatment of postural imbalance. J Bone Joint Surg 61A:I 182-1194. 328. Minoura M (1914) Studien ueber gelenktransplantation (inkl. Intermediaerknorpel) in weichteile. Frank Z Path 15:397. 329. Monticelli G, Spinelli R (1981) Distraction epiphysiolysis as a method of limb lengthening: I. Experimental study. Clin Orthop Rel Res 154:254-261. 330. Monticelli G, Spinelli R, Bonucci E (1981) Distraction epiphysiolysis as a method of limb lengthening: II. Morphologic investigations. Clin Orthop Rel Res 154:262-273. 331. Monticelli G, Spinelli R (1981) Distraction epiphysiolysis as a method of limb lengthening: III. Clinical applications. Clin Orthop Rel Res 154:274-285. 332. Monticelli G, Spinelli R (1983) Leg lengthening by closed metaphyseal corticotomy. Ital J Orthop Traum 9:139-150. 333. Moon NF (1973) Synovial hemangioma of the knee joint: A review of previously reported cases and inclusion of two new cases. Clin Orthop Rel Res 90:183-190. 334. Moore BH (1941) A critical appraisal of the leg lengthening operation. Am J Surg 52:415-423. 335. Morgan JD, Somerville EW (1960) Normal and abnormal growth at the upper end of the femur. J Bone Joint Surg 42-B: 264-272. 336. Morrissy RT, Riseborough EJ, Hall JE (1981) Congenital pseudarthrosis of the tibia. J Bone Joint Surg 63B:367-375. 337. Morscher E (1977) Etiology and pathophysiology of leg length discrepancies. In: Leg Length Discrepancy/The Injured Knee, ed DS Hungerford. pp. 9-19, New York: Springer Verlag. 338. Morscher E, Figner G (1977) Measurement of leg length. In: Leg Length Discrepancy/The Injured Knee, ed DS Hungerford. pp. 21-27, New York: Springer Verlag. 339. Mosca V, Moseley CF (1986) Complications of Wagner leg lengthening and their avoidance. Orthop Trans 10:462. 340. Moseley CF (1977) A straight line graph for leg length discrepancies. J Bone Joint Surg 59A: 174-179. 341. Moseley CF (1978) A straight line graph for leg length discrepancies. Clin Orthop Rel Res 136:33-40. 342. Mouchet A, Belot J (1926) La tarsomegalie. J Radiol Electrol 10:289-293.

343. Mulliken JB (1993) Cutaneous vascular anomalies. Sem Vasc Surg 6:204-218. 344. Mulliken JB, Glowacki J (1982) Hemangiomas and vascular malformations in infants and children: A classification based on endothelial characteristics. Plast Reconstr Surg 69:412-422. 345. Munter AH (1936) A study of the lengths of the long bones of the arms and legs in man, with special reference to AngloSaxon skeletons. Biometrika 28:258-294. 346. Nakamura E, Mizuta H, Sei A, Takagi K (1993) Knee articular cartilage injury in leg lengthening. Acta Orthop Scand 64: 437-440. 347. Nettelblad H, Randolph MA, Weiland AJ (1984) Free microvascular epiphyseal plate transplantation. J Bone Joint Surg 66A: 1421-1430. 348. Nichols PJR (1960) Short-leg syndrome. Brit Med J i: 1863-1865. 349. Obata K (1914) Ueber transplantation von gelenken bei junger tieren, mit besonderer berucksichtigung des verhaltens des intermediarknorpels. Beitr Path Anat u Allgem Path 49:1. 350. Ogden J (1982) Skeletal growth mechanism injury patterns. J Pediatr Orthop 2:371-377. 351. Ogilvie JW (1986) Epiphysiodesis: Evaluation of a new technique. J Pediatr Orthop 6:147-149. 352. Ogilvie JW, King K (1990) Epiphysiodesis: Two-year clinical results using a new technique. J Pediatr Orthop 10:809-811. 353. Olin A, Creasman C, Shapiro F (1984) Free physeal transplantation in the rabbit. J Bone Joint Surg 66A:7-20. 354. Ollier L (1867) Traite experimental et clinique de la regeneration des os et de la production artificielle du tissue osseux. Paris: Victor Masson et Fils. 355. O'Malley DE, Mazur JM, Cummings RJ (1995) Femoral head avascular necrosis associated with intramedullary nailing in an adolescent. J Pediatr Orthop 15:21-23. 356. Ombredanne L (1913) Allongement d' un femur sur un membre trop court. Bull Mem Soc Chir Paris 39:1177. 357. Osterman K (1972) Operative elimination of partial premature epiphyseal closure: An experimental study. Acta Orthop Scand Supp 147:1-79. 358. Osterman K (1981) Longitudinal growth of a tubular bone after a large surgical defect in the lateral part of the epiphyseal plate. Acta Orthop Scand 52:439. 359. Osterman K, Merikanto J (1991) Diaphyseal bone lengthening in children using Wagner device: Long-term results. J Pediatr Orthop 11:449-451. 360. O'Sullivan ME, Fogarty EE (1990) Distal tibial physeal arrest: A complication of meningococcal septicemia. J Pediatr Orthop 10:549-550. 361. Paley D (1988) Current techniques of limb lengthening. J Pediatr Orthop 8:73-92. 362. Paley D (1990) Problems, obstacles, and complications of limb lengthening by the Ilizarov technique. Clin Orthop Rel Res 250:81-103. 363. Paley D, Evans DC (1986) Angiomatous involvement of an extremity. Clin Orthop Rel Res 206:215-218. 364. Paley D, Herzenberg JE, Paremain G, Bhave A (1997) Femoral lengthening over an intramedullary nail. J Bone Joint Surg 79A: 1464-1480. 365. Pandey S, Sinha SK, Prasad MB, Singh PD, Pandey AK (1995) Limb length disparity in experimentally produced pyogenic osteomyelitis. J Orthop Surg 3:11-16.

References 366. Pappas AM (1983) Congenital abnormalities of the femur and related lower extremity malformations: Classification and treatment. J Pediatr Orthop 3:45-60. 367. Pappas AM (1984) Congenital posteromedial bowing of the tibia and fibula. J Pediatr Orthop 4:525-531. 368. Pappas AM, Hanawalt BJ, Anderson M (1972) Congenital defects of the fibula. Orthop Clin N Am 3:187-199. 369. Pappas AM, Nehme AE (1979) Leg length discrepancy associated with hypertrophy. Clin Orthop Rel Res 144:198-211. 370. Parke W, Colvin GS, Almond AHG (1949) Premature epiphysial fusion at the knee joint in tuberculous disease of the hip. J Bone Joint Surg 31B: 63-73. 371. Parkes Weber F (1907) Angioma formation in connection with hypertrophy of limbs and hemihypertrophy. Br J Derm 19:231-235. 372. Parkes Weber F (1918) Hemangiectatic hypertrophy of limbsncongenital phlebarteriectasis and so-called congenital varicose veins. Br J Child Dis 15:13-17. 373. Paterson JMH, Waller CS, Catterall A (1989) Lower limb lengthening by a modified Wagner technique. J Pediatr Orthop 9:129-133. 374. Patriquin HB, Trias A, Jecquier S, Marton D (1981) Late sequelae of infantile meningococcemia in growing bones of children. Radiology 141:77-82. 375. Pauwels F (1973) Atlas zur Biomechanik der gesunden und kranken hufte. Berlin: Springer. 376. Pease CN (1952) Local stimulations of growth of long bones. A preliminary report. J Bone Joint Surg 34A: 1-23. 377. Peixinho M, Arakaki T, Toledo CS (1982) Correction of leg inequality in the Klippel-Trenaunay-Weber syndrome. Internat Orthop 6:45-47. 378. Pels Leusden F (1899) Ueber die bei tuberculose des kniegelenkes zu beobachtenden wachsthumsveranderungen am femur. Deut Z Chir 51:257-280. 379. Peltonen J (1989) Bone formation and remodeling after symmetric and asymmetric physeal distraction. J Pediatr Orthop 9:191-196. 380. Peltonen J, Kahri A, Karaharju E, Alitalo I (1988) Regeneration after physeal distraction of the radius in sheep. Acta Orthop Scand 59(6):675-680. 381. Peltonen J, Kahri A, Lindberg LA, Heikkila PS, Karaharju EO, Aalto KA (1992) Bone formation after distraction osteotomy of the radius in sheep. Acta Orthop Scand 6:599-603 382. Peltonen JI, Karaharju EO, Alitalo I (1984) Experimental epiphysial distraction producing and correcting angular deformities. J Bone Joint Surg 66B:598-602. 383. Pennecot GF, Herman S, Pouliquen JC (1983) Retentissement de l'allongement progressif sur le cartilege de croissance. Interet de la mesure du couple. Rev Chir Orthop 69:623-627. 384. Perthes G (1903) Uber den einfluss der rontgenstralen auf epitheliale gewebe, inbessondere auf das carcinom. Arch Klin Chir 79:955-1000. 385. Peterson HA (1989) Surgical lengthening of the humerus: Case report and review. J Pediatr Orthop 9:596-601. 386. Petty W, Winter RB, Felder D (1974) Arteriovenous fistula for treatment of discrepancy in leg length. J Bone Joint Surg 56A:581-586. 387. Phemister DB (1933) Operative arrestment of longitudinal growth of bones in the treatment of deformities. J Bone Joint Surg 15:1-15.

729

388. Phillips GN, Gordon DH, Martin EC, Hailer JO, Casarella W (1978) The Klippel-Trenaunay syndrome: Clinical and radiological aspects. Radiology 128:429-434. 389. Porat S, Peyser A, Robin GC (1991) Equalization of lower limbs by epiphysiodesis: Results of treatment. J Pediatr Orthop 11:442-448. 390. Pouliquen JC, Beneux J, Mener G, Pennecot GF (1979) Etude de la croissance du membre inferieur apres allongement segmentaire chez l'enfant. Ann Orthop Ouest 11:95-98. 391. Pouliquen JC, Beneux J, Vemeret C, Hardy J, Mener G (1984) Allongement de tibia selon la methode de Judet: Apropos de 108 cases chex l'enfant. Rev Chir Orthop 70:29-39. 392. Pouliquen JC, Etienne W (1978) La reprise de croissance apres allongement progressif du membre inferieur chez l'enfant. Chir Pediat 19:179-183. 393. Pouliquen JC, Gorodischer S, Vemeret C, Richard L (1989) Allongement de femur chez l'enfant et l'adolescent. Etude comparative d'une serie de 82 cas. Rev Chir Orthop 75:239-251. 394. Price CT (1989) Limb lengthening for achondroplasia: Early experience. J Pediatr Orthop 9:512-515. 395. Pritchett JW (1992) Longitudinal growth and growth-plate activity in the lower extremity. Clin Orthop Rel Res 275:274-279. 396. Putti V (1921) The operative lengthening of the femur. J Am Med Assoc 934-935. 397. Putti V (1934) Operative lengthening of the femur. Surg Gyn Obstet 58:318-321. 398. Ratliff AHC (1959) The short leg in poliomyelitis. J Bone Joint Surg 41B:56-69. 399. Redler I (1952) Clinical significance of minor inequalities in leg length. New Orl Med Surg J 104:308-312. 400. Reschke K (1922) Verlangerung der rohrenknochen bei arthritis deformans jugendlicher. Deut Z Chir 168:136-139. 401. Reynolds DA (1981) Growth changes in fractured long-bones: A study of 126 children. J Bone Joint Surg 63B:83-88. 402. Richards OW, Kavanagh AJ (1945) The analysis of growing form. In: Essays on Growth and Form, eds WEL Clark, PW Medawar. p. 188-230, Oxford: Clarendon Press. 403. Rigault P, Roucquey P, Padovani JP, Raux Ph, Finidori G (1980) L' allongement progressive du femur cehz 1'enfant. A propos de 36 cas. Rev Chir Orthop 66:13-22. 404. Rigault P, Dolz G, Padovani JP, Touzet Ph, Mallet JF, Finidori G (1981) L' allongement progressif du tibia chez l'enfant: A propos de 48 cas. Rev Chir Orthop 67:461-472. 405. Ring PA (1955) Transplantation of epiphysial cartilage. J Bone Joint Surg 37B:642-657. 406. Ring PA (1957) Shortening and paralysis in poliomyelitis. Lancet ii:980-983. 407. Ring PA (1958) Experimental bone lengthening by epiphysial distraction. Brit J Surg 46:169-173. 408. Ring PA (1959) Congenital short femur. Simple femoral hypoplasia. J Bone Joint Surg 41B:73-79. 409. Roberts PH (1970) Disturbed epiphysial growth at the knee after osteomyelitis in infancy. J Bone Joint Surg 52B: 692-703. 410. Robertson WW, Jr, Butler MS, D'Angio GJ, Rate WR (1991) Leg length discrepancy following irradiation for childhood tumors. J Pediatr Orthop 11:284-287. 411. Robinow M, Johnson GF, Nanagas MT, Mesghali H (1983) Skeletal lesions following meningococcemia and disseminated intravascular coagulation. Am J Dis Child 137:279-281.

730

CHAPTER 8 ~ Lower Extremity Len~tth Discrepancies

412. Rodriguez-Merchan EC (1996) Effects of hemophilia on articulations of children and adults. Clin Orthop Rel Res 328: 7-13. 413. Rogalski R, Hensinger R, Loder R (1993) Vascular abnormalities of the extremities: Clinical findings and management. J Pediatr Orthop 13:9-14. 414. Ross D (1948) Disturbance of longitudinal growth associated with prolonged disability of the lower extremity. J Bone Joint Surg 30A: 103-115. 415. Rossvoll I, Junk S, Terjesen T (1992) The effect on low back pain of shortening osteotomy for leg length inequality. Internat Orthop 16:388-391. 416. Rush WA, Steiner HA (1946) A study of lower extremity length inequality. Am J Roent 56:616-623. 417. Russell A (1954) A syndrome of "intra-uterine" dwarfism recognizable at birth with cranio-facial dysostosis, disproportionately short arms, and other anomalies (5 examples). Proc Roy Soc Med 47:1040-1044. 418. Salai M, Chechick A, Ganel A, Blankstein A, Horoszowski H (1985) Subluxation of the hip joint during femoral lengthening. J Pediatr Orthop 5:642-644. 419. Saleh M, Goonatillake HD (1995) Management of congenital leg length inequality: Value of early axis correction. J Pediatr Orthop 4:150-158. 420. Saleh MB, Stubbs DA, Street RJ, Lang DM, Harris SC (1993) Histologic analysis of human lengthened bones. J Pediatr Orthop 2B: 16-21. 421. Salter RB, Harris WP (1963) Injuries involving the epiphyseal plate. J Bone Joint Surg 45A:587-622. 422. Salter RB (1973) Legg-Perthes, Part V. Treatment by innominate osteotomy. In: Instructional Course Lectures, The American Academy of Orthopedic Surgeons, volume 22. p. 309, St. Louis: CV Mosby Company. 423. Sasso RC, Urquhart BA, Cain TE (1993) Closed femoral shortening. J Pediatr Orthop 13:51-56. 424. Savarese JJ, Brinken BW, Zaleske DJ (1995) Epiphyseal replacement in a murine model. J Pediatr Orthop 15:682-690. 425. Schoenecker PL, Capelli AM, Millar EA, Sheen MR, Haher T, Aiona MD, Meyer LC (1989) Congenital longitudinal deficiency of the tibia. J Bone Joint Surg 71A:278-287. 426. Schopler SA, Lawrence JF, Johnson MK (1986) Lengthening of the humerus for upper extremity limb length discrepancy. J Pediatr Orthop 6:477-480. 427. Schultz AH (1937) Proportions, variability, and asymmetries of the long bones of the limbs and the clavicles in man and apes. Human Biol 9:281-328. 428. Schwerz F (1912) Die Alamannen in der Schweiz. Z Morph Anthrop 14:609-700. 429. Sengupta A, Gupta P (1993) Epiphyseal stapling for leg equalisation in developing countries. Internat Orthop 17:27-42. 430. Shapiro F (1982) Legg-Calve-Perthes disease. A study of lower extremity length discrepancies and skeletal maturation. Acta Orthop Scand 53:437-444. 431. Shapiro F (1981) Fractures of the femoral shaft in children. The overgrowth phenomenon. Acta Orthop Scand 52:649-655. 432. Shapiro F (1982) Ollier's disease. An assessment of angular deformity, shortening, and pathological fracture in twentyone patients. J Bone Joint Surg 64-A:95-103. 433. Shapiro F (1982) Developmental patterns in lower extremity length discrepancies. J Bone Joint Surg 64A:639-651.

434. Shapiro F (1982) Epiphyseal growth plate fracture-separations. A pathophysiologic approach. Orthopedics 5:720-736. 435. Shapiro F (1987) Longitudinal growth of the femur and tibia after diaphyseal lengthening. J Bone Joint Surg 69A:684-690. 436. Shapiro F, Rand F, Upton J, Barone L (1992) Histologic patterns of bone formation in rabbit distraction osteogenesis. Orthop Trans 16:561-562. 437. Shapiro F, Simon S, Glimcher MJ (1979) Hereditary multiple exostosis. Anthropometric, roentgenographic, and clinical aspects. J Bone Joint Surg 61A:815-824. 438. Siffert RS (1956) The effect of staples and longitudinal wires on epiphyseal growth. J Bone Joint Surg 38A:1077-1088. 439. Siffert RS (1957) The effect of juxta-epiphyseal pyogenic infection on epiphyseal growth. Clin Orthop Rel Res 10: 131-139. 440. Siffert RS (1966) The growth plate and its affections. J Bone Joint Surg 48A:546-563. 441. Siffert RS (1987) Lower limb-length discrepancy. J Bone Joint Surg 69A:1100-1106. 442. Siffert RS, Feldman DJ (1980) The growing hip. The dynamic development of the normal adult hip and the deformed hip of Legg-Calve-Perthes' disease. Acta Orthop Belg 46: 443-476. 443. Silfverskiold N (1934) Uber Langenwachstum der Knochen und transplantation von epiphysenscheiben. Acta Chir Scand 75:77-104. 444. Silver HK (1964) Asymmetry, short stature, and variations in sexual development. Am J Dis Child 107:495-515. 445. Silver HK, Kiyasu W, George J, Deamer WC (1953) Syndrome of congenital hemihypertrophy, shortness of stature, and elevated urinary gonadotropins. Pediatrics 12:368-375. 446. Simon S, Whiffen J, Shapiro F (1981) Leg-length discrepancies in monoarticular and pauciarticular juvenile rheumatoid arthritis. J Bone Joint Surg 63A:209-215. 447. Sissons HA (1952) Osteoporosis and epiphysial arrest in joint tuberculosis: An account of the histological changes in involved tissues. J Bone Joint Surg 34B:275-290. 448. Sledge CB, Noble J (1978) Experimental limb lengthening by epiphyseal distraction. Clin Orthop Rel Res 136:111-119. 449. Smith CF (1996) Instantaneous leg length discrepancy determination by "thigh-leg" technique. Orthopedics 19:955-956. 450. Sola CK, Silberman FS, Cabrini RL (1963) Stimulation of the longitudinal growth of long bones by periosteal stripping: An experimental study on dogs and monkeys. J Bone Joint Surg 45A:1679-1684. 451. Specht EE, Hazelrig PE (1973) Orthopaedic considerations of Silver's syndrome. J Bone Joint Surg 55A:1502-1510. 452. Speed K (1922) Growth problems following osteomyelitis of adolescent long bones. Surg Gyn Obstet 34:469-476. 453. Speed K (1923) Longitudinal overgrowth of long bones. Surg Gyn Obstet 36:787-794. 454. Spriggins AJ, Bader DL, Cunningham JL, Kenwright J (1989) Distraction physiolysis in the rabbit. Acta Orthop Scand 60:154-158. 455. Staheli LT (1967) Femoral and tibial growth following femoral shaft fracture in childhood. Clin Orthop Rel Res 55: 159-163. 456. Staheli LT, Duncan WR, Schaefer E (1968) Growth alterations in the hemiplegic child. A study of femoral anteversion, neck-shaft angle, hip rotation, CE angle, limb length, and

References

457.

458.

459.

460.

461.

462. 463.

464.

465. 466.

467.

468. 469. 470.

471. 472.

473.

474.

475. 476.

477.

circumference in 50 hemiplegic children. Clin Orthop Rel Res 60:205-212. Stanitski DF (1997) Limb lengthening in the skeletal dysplasias and short stature conditions: State of the art in 1997. Growth Genet Horm 13:17-22. Stanitski DF, Bullard M, Armstrong P, Stanitski CL (1995) Results of femoral lengthening using the Ilizarov technique. J Pediatr Orthop 15:224-231. Stanitksi DE Rossman K, Torosian M (1996) The effect of femoral lengthening on knee articular cartilage: The role of apparatus extension across the joint. J Pediatr Orthop 16:151-154. Stanitski DF, Shahcheraghi H, Nicker DA, Armstrong PF (1996) Results of tibial lengthening with the Ilizarov technique. J Pediatr Orthop 16:168-172. Stephens DC, Louis DS (1974) Traumatic separation of the distal femoral epiphyseal cartilage plate. J Bone Joint Surg 56A:1383-1390. Stephens DC (1983) Femoral and tibial lengthening. J Pediatr Orthop 3:424-430. Stephens MM, Hsu LCS, Leong JCY (1989) Leg length discrepancy after femoral shaft fractures in children. J Bone Joint Surg 71B:615-618. Stinchfield AJ, Reidy JA, Barr JS (1949) Prediction of unequal growth of the lower extremities in anterior poliomyelitis. J Bone Joint Surg 31A:478-486. Stuart HC, Reed RB (1959) Longitudinal studies of child health and development: Series II. Pediatrics 24:875-974. Suva D, Naccarato LH, Heripret G, Cauchoix J, Morel G (1984) Une inconnue apres allongement chirurgical du femur chez l'enfant. L' allongement parasite spontane post-operatoire. Rev Chir Orthop 70:631-636. Suzuiki S, Kasahara Y, Seto Y, Futami T, Furukawa K, Nishino Y (1994) Dislocation and subluxation during femoral lengthening. J Pediatr Orthop 14:343-346. Tachdjian MO (1972) Pediatric Orthopedics. Philadelphia: WB Saunders. Taillard W, Morscher E (1965) Beinlangenunterschiede. Basel: Karger. Tanner JM, Lejarraga H, Cameron N (1975) The natural history of the Silver-Russell syndrome: A longitudinal study of thirty-nine cases. Pediat Res 9:611-623. Teot L (1982) Les transferts osseux libres vascularises avec cartilage de croissance. Rev Chir Orthop Supp II: 68:40-42. Teot L, Bosse JP, Gilbert A, Tremblay GR (1983) Pedicle graft epiphysis transplantation. Clin Orthop Rel Res 180: 206-218. Teot L, Gilbert A, Katz D, Pous JG, Carlioz H, Bonnel F (1982) Vascularisation epiphysaire pendant la croissance. Rev Chir Orthop 68:357-364. Terjeson T, Benum P, Rossvoll I, Svenningsen S, Floystad Isern AE, Nordbo T (1991 ) Leg-length discrepancy measured by ultrasonography. Acta Orthop Scand 62:121-124. Thelander HE (1950) Epiphyseal destruction by frostbite. J Pediatr 36:105-106. Timperlake RW, Bowen JR, Guille JT, Ho Choi I (1991) Prospective evaluation of fifty-three consecutive percutaneous epiphysiodeses of the distal femur and proximal tibia and fibula. J Pediatr Orthop 11:350-357. Todd TW (1937) Atlas of Skeletal Maturation. Part 1: The Hand and Wrist. St. Louis, CV Mosby.

731

478. Trelat V, Monod A (1869) De l'hypertrophie unilaterale partielle ou totale du corps. Arch Gen Med 6th series, 13:536558 and 676-705. 479. Trevor D (1950) Tarso-epiphysial aclasis: A congenital error of epiphysial development. J Bone Joint Surg 32B'204-213. 480. Truesdell ED (1921) Inequality of the lower extremities following fracture of the shaft of the femur in children. Ann Surg 74:498-500. 481. Trueta J (1953) Acute haematogenous osteomyelitis. Its pathology and treatment. Bull Hosp Joint Dis 14:5-23. 482. Trueta J, Morgan JD (1954) Late results in the treatment of one hundred cases of acute haematogenous osteomyelitis. Brit J Surg 41:449-457. 483. Tupman GS (1960) Treatment of inequality of the lower limbs. The results of operations for stimulation of growth. J Bone Joint Surg 42B:489-501. 484. Van Lohuizen Ch (1922) Uber eine seltene angeborene hautanomalie (Cutis marmorata telangiectatica congenita). Acta Derm Venereol 3:202-211. 485. Van Nes CP (1966) Congenital pseudarthrosis of the leg. J Bone Joint Surg 48A: 1467-1483. 486. Vickers DW (1980) Premature incomplete fusion of the growth plate: Causes and treatment by resection (physiolysis) in fifteen cases. Aust N Z J Surg 50:393-401. 487. Viljanto J, Kiviluoto H, Paananen M (1975) Remodeling after femoral shaft fracture in children. Acta Chir Scand 141: 360-365. 488. Viljanto J, Linna MI, Kiviluoto H, Paananen M (1975) Indications and results of operative treatment of femoral shaft fractures in children. Acta Chir Scand 141:366-369. 489. Vlachos D, Carlioz H (1973) Les malformations du femur. Leur evolution spontanee. Rev Chir Orthop 59:626-640. 490. Von Tappeiner Fr H (1913) Studien zur frage der transplantationsfahigkeit des epiphysenknorpels und des gelenkknorpels. Z Ges Exp Med 1:491. 491. Von Tappeiner Fr H (1916) Neue experimente zur frage der homoplastichen transplantationsfahigkert des epiphysen-und des gelenkknorpels. Arch Klin Chir 107:479-508. 492. Wagner H (1977) Surgical lengthening or shortening of femur and tibia. Technique and indications. In: Leg Length Discrepancy/The Injured Knee, ed DS Hungerford, pp. 71-94, New York: Springer Verlag. 493. Wagner H (1978) Operative lengthening of the femur. Clin Orthop Rel Res 136:125-142. 494. Wasserstein I (1990) Twenty-five years' experience with lengthening of shortened lower extremities using cylindrical allografts. Clin Orthop Rel Res 250:150-153. 495. Weber BG, Brunner C, Freuler F (1980) Treatment of Fractures in Children and Adolescents. New York: Springer Verlag. 496. Westh RN, Menelaus MB (1981) A simple calculation for the timing of epiphyseal arrest: A further report. J Bone Joint Surg 63B:117-119. 497. Westin GW, Sakai DN, Wood WL (1976) Congenital longitudinal deficiency of the fibula. J Bone Joint Surg 58A: 492-496. 498. White SH, Kenwright J (1990) The timing of distraction of an osteotomy. J Bone Joint Surg 72B:356-361. 499. White JW (1935) Femoral shortening for equalization of leg length. J Bone Joint Surg 17:597-604.

732

CHAPTER 8 9

Lower Extremity Length Discrepancies

500. White JW (1949) Leg-length discrepancies. In: Instructional Course Lectures, The American Academy of Orthopedic Surgeons, volume 6. pp. 201-211, Ann Arbor, MI: JW Edwards. 501. White JW, Stubbins SG (1944) Growth arrest for equalizing leg lengths. J Am Med Assoc 126:1146-1148. 502. White JW, Warner WP (1938) Experiences with metaphyseal growth arrests. South Med J 31:411-413. 503. Wiedemann HR (1964) Complexe malformatif familial avec hernie ombilicale et macroglossiemun syndrome nouveau? J Genet Hum 13:223. 504. Wiedemann HR, Burgio GR, Aldenhoff P, Kunze J, Kaufmann HJ, Schirg E (1983) The Proteus syndrome. Eur J Pediatr 140:5-12. 505. Wiedemann M (1996) Callus distraction: A new method? A historical review of limb lengthening. Clin Orthop Rel Res 327:291-304. 506. Williamson RV, Staheli LT (1990) Partial physeal growth arrest: Treatment by bridge resection and fat interposition. J Pediatr Orthop 10:769-776. 507. Wilson JC, McKeever FM (1936) Bone growth disturbance following hematogenous acute osteomyelitis. J Am Med Assoc 107:1188-1193. 508. Wilson PD, Thompson TC (1939) A clinical consideration of the methods of equalizing leg length. Ann Surg 110:992-1015. 509. Winquist RA (1986) Closed intramedullary osteotomies of the femur. Clin Orthop Rel Res 212:155-164. 510. Wirth T, Byers S, Byard RW, Hopwood JJ, Foster BK (1994) The implantation of cartilaginous and periosteal tissue into growth plate defects. Internat Orthop 18:220-228.

511. Wolohan MJ, Zaleske DJ (1991) Hemiepiphyseal reconstruction using tissue donated from fetal limbs in a murine model. J Orthop Res 9:180-185. 512. Wopperer JM, White JJ, Gillespie R, Obletz BE (1988) Longterm follow-up of infantile hip sepsis. J Pediatr Orthop 8: 322-325. 513. Wu YK, Miltner LJ (1937) A procedure for stimulation of longitudinal growth of bone. An experimental study. J Bone Joint Surg 4:909-921. 514. Wyssa B, Le Coultre C, Kaelin A (1992) Orthopaedic and surgical complications of meningococcemia. J Pediatr Orthop 1B:73-77. 515. Young NL, Davis RJ, Bell DF, Redmond DM (1993) Electromyographic and nerve conduction changes after tibial lengthening by the Ilizarov method. J Pediatr Orthop 13:473-477. 516. Younge D, Drummond D, Herring J, Cruess RL (1979) Melorheostosis in children. J Bone Joint Surg 61B:415-418. 517. Zadek I (1935) Congenital coxa vara. Arch Surg 30:62-102. 518. Zaleske DJ, Ehrlich MG, Piliero C, May JW, Mankin HJ (1982) Growth-plate behavior in whole joint replantation in the rabbit. J Bone Joint Surg 64A:249-258. 519. Zavyalov PV, Plaksin JT (1968) Distraction epiphyseolysis in prolongation of the lower limb in children. Khirugiia 44: 121-137. 520. Zuege RC, Kempken TG, Blount WP (1979) Epiphyseal stapling for angular deformity at the knee. J Bone Joint Surg 61A:320-329.

CHAPTER

9

Skeletal Dysplasias IX. Histopathologic Changes in Specific Chondrodysplasias X. Orthopedic Deformities in Skeletal Dysplasias--Regional Abnormalities and Their Relation to Clinically Significant Deformity XI. Limb Lengthening XII. Review of Specific Skeletal Dysplasias: Pathobiology, Clinical and Radiographic Characteristics, and Orthopedic Management XIII. Anesthetic Implications in the Skeletal Dysplasias

Terminology Classification Approaches III. Prevalence of Skeletal Dysplasias IV. Diagnosis of Skeletal Dysplasias V. Chromosome Abnormality Sites in Skeletal Dysplasias VI. Genetic and Molecular Abnormalities in Skeletal Dysplasias VII. Lethal Perinatal Skeletal Dysplasias VIII. Microstructural-Morphologic Abnormalities of the Epiphyses and Metaphyses in Skeletal Dysplasias I.

II.

seal dysplasias), have come to be defined by biochemical abnormalities. Virtually all of the skeletal dysplasias are due to mutations in genes involved in skeletal development. Although the need to catalogue clinical and radiographic variations among these disorders is evident, deeper biologic understanding is being provided by elucidation of underlying gene and molecular defects in relation to collagen, glycoprotein, proteoglycan, fibroblast growth factor receptor, and other molecular components of epiphyseal cartilage, growth plate cartilage, and bone. Histology, histochemistry, and transmission electron microscopy frequently reveal structural abnormalities of physeal and epiphyseal cartilage and bone, although morphologic changes alone rarely are pathognomonic of specific diseases. The term skeletal dysplasia does not refer to all short stature syndromes but only to those in which the shape and structure of the bones are abnormal. One of the earliest distinctions made in relation to short stature syndromes was by Parrot in 1878, who used the term "achondroplasia" to distinguish disproportionately from proportionately short stature (230). It soon came to be recognized that the former syndromes were due to primary inherited skeletal abnormalities, whereas in the latter the bones were short secondary to systemic causes often acquired such as endocrine, metabolic, renal, hepatic, cardiac, or other anomalies. By the late 1920s distinctions within the disproportionate group delineated short limb dysplasias, with limb shortening greater than truncal-axial shortening, from short trunk dysplasias, with truncal-axial shortening greater than limb shortening. The prototypic short limb disproportionate dwarfism is achondroplasia and short trunk disproportionate dwarfism is Morquio's disease (mucopolysaccharidosis IV). Further

I. T E R M I N O L O G Y The skeletal dysplasias are developmental disorders of the bones and encompass a large number of conditions, most of which are genetic in origin. They can lead to such features as disproportionately short stature, angular deformities of the long bones, joint surface irregularity, joint contractures or instability, lower extremity length discrepancies, increases or decreases of bone density, and axial deformities, including cervical vertebral abnormalities, thoracolumbar scoliosis and kyphosis, spinal stenosis, and cranial bone malformations such as craniosynostosis and foramen magnum stenosis. Many terms have been used to refer to this large group of heterogeneous disorders, including chondrodysplasia, because the large majority are due to primary defects of cartilage, and osteochondrodysplasia, referring to abnormalities of both cartilage and bone, but the more general term of skeletal dysplasia is used increasingly.

II. C L A S S I F I C A T I O N A P P R O A C H E S Classification criteria for the skeletal dysplasias can be defined into three broad groups: clinical, radiographic, and molecular. Some conditions are so characteristic, such as achondroplasia or diastrophic dysplasia, that they were defined initially and accurately by clinical appearance. Some, such as the multiple epiphyseal dysplasias, spondyloepiphyseal dysplasias, or metaphyseal dysplasias, were defined, named, and categorized most clearly by characteristic radiographic findings, and others, such as the mucopolysaccharidoses (which, according to radiologic criteria, are spondyloepiphy733

734

CHAPTER 9 ~ Skeletal Dysplasias

subdivision has followed but these basic outlines remain valuable. Many of the skeletal dysplasia syndromes were initially described and referred to as dwarfisms, dysostoses, or dystrophies, but the modifying term dysplasia is used increasingly. Over 150 developmental disorders of the skeletal system have been described (17, 262) (Table I; see also Tables III and VIA in Chapter 1). The large number of disorders, the variable clinical and radiological appearance within individual disorders, the rarity of most of the defined disorders, and the fact that many patients with a skeletal dysplasia still cannot be classified definitively even with detailed and conscientious efforts have led to considerable confusion. The rapidly accumulating findings of genetic and molecular abnormalities in some mesenchymal tissue syndromes are instructive in this regard. In osteogenesis imperfecta, the wide spectrum of clinical involvement becomes understandable because virtually no two patients have the same molecular abnormality of collagen with close to 200 different mutations found (54, 156). Similar situations are being found with some of the other skeletal dysplasias. The subdivision of multiple epiphyseal dysplasia into 10-15 subtypes or spondyloepiphyseal dysplasia into even more subtypes based on clinical and radiographic variability almost certainly is reflective of varying molecular abnormalities, most as yet undefined, and of varying degrees of penetrance and expressivity (8). Relatively broad groupings at a clinical-radiographic level must suffice for now with further subdivision best based on the elucidation of genetic-molecular abnormalities. Multiple systems of classification have been described. Although all are instructive, it is essential to recognize that no universal, accurate system exists. Many of the earliest attempts at classification were hindered by the fact that many disorders currently appreciated were not even known due to their rarity and in some instances variable geographic locations. The medical discipline of those describing the disorders also plays a role in defining the format within which the patients are described. Early clinical observers defined proportionate and disproportionate syndromes, clinical geneticists elucidated autosomal dominant or recessive and sex-linked disorders, radiologists tend to favor minute descriptions often noting epiphyseal, metaphyseal, or diaphyseal localization, orthopedic surgeons distinguish between syndromes with and without spinal involvement, neurologic sequelae, angular limb deformity, and articular surface irregularity, and molecular biologists define genetic and molecular abnormalities. Though these findings interdigitate to a significant extent there is not a complete fit between groupings. On the other hand, a relatively small number of entities characterize the large majority of patients with skeletal dysplasias such that good awareness of findings in any patient is achievable. Rubin reviewed the several classification systems that had evolved for the bone dysplasias by the 1970s (271). Murk Jansen was the first to focus on description of the skeletal abnormality based on the region of the bone that was most

deformed, and his basic classification defined epiphyseal dysostosis, metaphyseal dysostosis, and diaphyseal dysostosis (138, 271). This remains a valuable way to assess radiographs initially. Brailsford wrote on developmental bone disorders, recognizing that the dysplasias are variable in degree and severity with many "transitional" forms that made classification difficult (271). He divided the disorders into three groups: (1) defective ossification of the whole skeleton, which was either osteopenic involving osteogenesis imperfecta or radiodense involving osteopetrosis; (2) abnormal metaphyseal growth in which he placed the achondroplasia disorder and chondro-osteodystrophy; and (3) proliferation of mesoblastic cells in which he defined four disorders, only one of which, melorheostosis, would be clearly recognized by most today. Fairbank in his Atlas of General Affections of the Skeleton reviewed several syndromes (74). Although seven categories of general affections of the skeleton were listed, the major value of the book rests in his description of specific entities rather than attempts to force them into specific groupings. Even in the earliest of classifications difficulties are apparent due to the mixing of clinically recognizable syndromes such as achondroplasia with those defined not by the clinical appearance of the individual but rather by the specific radiologic abnormalities. Further difficulties developed by mixing the bases of the classifications even further by defining variables such as tissue of origin, cartilage or bone, often in the absence of any convincing histopathological correlation. With each year more detailed radiographic and clinical information accumulated, and continuing studies over the past 30 years have added and continue to add additional syndromes and classification systems. The uncovering of genetic and molecular abnormalities is a welcome addition to the field, although even delineation of these abnormalities into molecular families will not totally clarify the disorders because considerable areas of overlap between seemingly disparate clinical disorders are already being found. Rubin provided an interesting approach to assessing the skeletal dysplasias based on the dynamics of bone development and the specific regions that were deformed (271). These observations made initially by Jansen and confirmed by many other studies clearly show a regional or localizing tendency for abnormal structure in many dysplasias, such that they can be defined adequately although not completely by these localizations as seen radiographically. Rubin's dynamic classification of bone dysplasias described (1) epiphyseal dysplasias (by which he meant dysplasias of the articular cartilage and secondary ossification centers), (2) physeal dysplasias, (3) metaphyseal dysplasias, and (4) diaphyseal dysplasias. The dynamic nature of his classification is accompanied by detailed reviews of bone morphogenesis and development involving synthesis and resorptive aspects. He stressed the modeling of the various regions with growth, characterizing epiphyseal enlargement as a hemispherization process, physeal function as involving growth, metaphyseal development as a funnelization process, and diaphyseal

SECTION II ~ Classification Approaches TABLE I

Classification o f O s t e o c h o n d r o d y s p l a s i a s a'b

1. Achondroplasia group Thanatophoric dysplasia, type I AD Thanatophoric dysplasia, type II AD Achondroplasia AD Hypochondroplasia AD Other FGFR3 disorders 2. Spondylodysplastic and other perinatally lethal groups Lethal platyspondylic skeletal dysplasias (San Diego, Torrance, Luton) Achondrogenesis type IA 3. Metatropic dysplasia group Fibrochondrogenesis AR Schneckenbecken dysplasia AR Metatropic dysplasia (various types) AD 4. Short rib dysplasia (SRP) (with/without polydactyly) group SRP type I Saldino-Noonan AR SRP type II Majewski AR SRP type III Verma-Naumoff AR SRP type IV Beemer-Langer AR Asphyxiating thoracic dysplasia (Jeune) AR Chondroectodermal dysplasia (Ellis-van AR Creveld dysplasia) 5. Atelosteogenesis-omodysplasia group Atelosteogenesis type 1 (includes Sp "Boomerang dysplasia") Omodysplasia I (Maroteaux) AD Omodysplasia II (Borochowitz) AR Otopalatodigital syndrome type II XLR Atelosteogenesis type III de la Chapelle dysplasia 6. Diastrophic dysplasia group Diastrophic dysplasia AR Achondrogenesis IB AR Atelosteogenesis type II AR 7. Dyssegmental dysplasia group Dyssegmental dysplasia (SilvermanAR Handmaker type) Dyssegmental dysplasia (RollandAR Desbuquois type) 8. Type II collagenopathies Achondrogenesis II (Langer-Saldino) AD Hypochondrogenesis AD Kniest dysplasia AD Spondyloepiphyseal dysplasia congenita AD (SED) Spondyloepimetaphyseal dysplasia (SEMD) Strudwick type AD SED with brachydactyly AD Mild SED with premature-onset arthrosis AD Stickler dysplasia (heterogeneous, some not AD linked to COL2A1) 9. Type XI collagenopathies Stickler dysplasia (heterogeneous) Otospondylomegaepiphyseal dysplasia (OSMED)

73S

AD

10. Other spondyloepi(meta)physeal [SE(M)D] dysplasias X-linked spondyloepiphyseal dysplasia tarda XLD Other late-onset spondyloepi(meta)physeal dysplasias (i.e., Namaqualand d., Irapa D.) Progressive pseudo-rheumatoid dysplasia AR Dyggve-Melchior-Clausen dysplasia AR Wolcott-Rallison dysplasia AR Immuno-osseous dysplasia Schimke AR Opsismodysplasia AR Chondrodystrophic myotonia (Schwartz AR Jampel), type I, type II Spondyloepiphyseal dysplasia with joint laxity AR Sponastrime dysplasia AR SEMD short limb, abnormal calcification AR 11. Multiple epiphyseal dysplasias and pseudo-achondroplasia Pseudo-achondroplasia Multiple epiphyseal dysplasia (MED) AD (Fairbanks and Ribbing types) Other MEDs 12. Chondrodysplasia punctata (stippled epiphyses group) Rhizomelic type AR Zellweger syndrome (several types) AR Conradi-Htinermann type XLD X-linked recessive type XLR Brachytelephalangic type XLR Tibial-metacarpal type AD Vitamin K-dependent coagulations defect AR Other acquired and genetic disorders including Warfarin embryopathy 13. Metaphyseal dysplasias Jansen type AD Schmid type AD McKusick type (cartilage-hair hypoplasia) AR Metaphyseal anadysplasia XLR? Metaphyseal dysplasia with pancreatic insufficiency and cyclic neutropenia (Schwachman Diamond) Adenosine deaminase deficiency AR Metaphyseal chondrodysplasia Spahr type AR Acroscyphodysplasia (various types) AR 14. Spondylometaphyseal dysplasias (SMD) Spondylometaphyseal dysplasia (Kozlowski AD type) Spondylometaphyseal dysplasia (Sutcliffe AD type) SMD with severe genu valgum (includes AD Schmidt and Algerian types) SMD Sedaghatian type AR Mild SMD (various types) 15. Brachyolmia spondylodysplasias Hobaek (includes Toledo type) AR Maroteaux type AR Autosomal dominant type AD 16. Mesomelic dysplasias Dyschondrosteosis (Leri-Weill) AD Langer type (homozygous dyschondrosteosis) AR (continues)

736

CHAPTER 9 ~ Skeletal Dysplasias TABLE I (continued)

17.

18.

19.

20.

21.

Nievergelt type AD Kozlowski-Reardon type AR Reinhardt-Pfeiffer type AD Werner type AD Robinow type, dominant AD Robinow type, recessive AR Mesomelic dysplasia with synostes AD Acromelic and acromesomelic dysplasias Acromicric dysplasia AD Geleophysic dysplasia AR Weill-Marchesani dysplasia AR Cranioectodermal dysplasia AR Trichorhinophalangeal dysplasia type I AD Trichorhinophalangeal dysplasia type II AD (Langer-Giedion) Trichorhinophalangeal dysplasia type III AD Grebe dysplasia AR Hunter-Thompson dysplasia AR Brachdactyly types A1-A4, B, C, D, and E AD Pseudo-hypoparathyroidism (Albright hereditary osteodystrophy) Acrodyostosis Saldino-Mainzer dysplasia AR Brachydactyly-hypertension dysplasia AD Craniofacial conodysplasia AD Angel shaped phalangoepiphyseal dysplasia AD (ASPED) Acromesomelic dysplasia AR Dysplasias with prominent membranous bone involvement Cleidocranial dysplasia AD Osteodysplasty, Melnick-Needles XLD Precocious osteodysplasty (terHaar dysplasia) AR Yunis-Varon dysplasia AR Bent bone dysplasia group Campomelic dysplasia AD Kyphomelic dysplasia ?AR Sttive-Wiedemann dysplasia AR Multiple dislocations with dysplasias Larsen syndrome AD Larsen-like syndromes (La Reunion Island) AR Desbuquois dysplasia AR Pseudo-diastrophic dysplasia AR Dysostosis multiplex group Mucopolysaccharidosis IH AR Mucopolysaccharidosis IS AR Mucopolysaccharidosis II XLR Mucopolysaccharidosis IliA AR IIIB AR IIIC AR IIID AR Mucopolysaccharidosis IVA AR Mucopolysaccharidosis IVB AR Mucopolysaccharidosis VI AR Mucopolysaccharidosis VII AR Fucosidosis AR oL-Mannosidosis AR

Aspartylglucoasminuria gM 1 gangliosidosis, several forms Sialidosis, several forms Sialic storage disease Galactosialidosis, several forms Multiple sulfatase deficiency Mucolipidosis II Mucolipidosis III 22. Osteodysplastic slender bone group Type 1 osteodysplastic dysplasia Type 2 osteodysplastic dysplasia Microcephalic osteodysplastic dysplasia 23. Dysplasias with decreased bone density Osteogenesis imperfecta I (without opalescent teeth) Osteogenesis imperfecta I (with opalescent teeth) Osteogenesis imperfecta II Osteogenesis imperfecta III Osteogenesis imperfecta IV (without opalescent teeth) Osteogenesis imperfecta IV (with opalescent teeth) Cole-Carpenter dysplasia Bruck syndrome Singleton-Merten syndrome Osteopenia with radiolucent lesions of mandible Osteoporosis-pseudo-glioma dysplasia Geroderma osteodysplasticum Hyper IGE syndrome with osteopenia Idiopathic juvenile osteoporosis 24. Dysplasias with defective mineralization Hypophosphatasia-perinatal lethal and infantile forms Hypophosphatasia adult form Hypophosphatemic tickets Neonatal hyperparathyroidism Transient neonatal hyperparathyroidism with hypocalciuric hypercalcemia

AR AR AR AR AR AR AR AR AR AR AR AD AD AD-AR AD-AR AD AD

AR AR AD AR AR AR Sp AR AD XLD AR AD

25. Increased bone density without modification of bone shape Osteopetrosis a. Precocious type AR b. Delayed type AD c. Intermediate type AR d. With renal tubular acidosis AR Axial osteosclerosis a. Osteomesopycnosia AD b. With Bamboo hair (Netherton AR syndrome) Pycnodysostosis AR Osteoclerosis-Stanescu type AD Osteopathia striata a. Isolated Sp b. With cranial sclerosis AD Sponastrime dysplasia AR

(continues)

737

SECTION II ~ Classification Approaches TABLE I (continued)

26.

27.

28.

29.

30.

Melorheostosis Sp Osteopoikilosis AD Mixed sclerosing bone dysplasia Sp Increased bone density with diaphyseal involvement Diaphyseal dysplasia, Camurati-Engelmann AD Craniodiaphyseal dysplasia ?AR Lenz-Majewski dysplasia Sp Endosteal hyperostosis AR a. van Buchem type AD b. Worth disease AR c. Sclerosteosis AR d. With cerebellar hypoplasia AD-AR Kenny Caffey dysplasia AR Osteoectasia with hyperphophatasia (juvenile Pagets) Diaphyseal dysplasia with anemia AR Diaphyseal medullary stenosis with bone AD malignancy (Hardcastle) Increased bone density with metaphyseal involvement Pyle dysplasia AR Craniometaphyseal dysplasia AR a. Severe type AD b. Mild type XLR Frontometaphyseal dysplasia AR-XLR Dysosteosclerosis AD-AR Oculodento-osseous dysplasia AD Trichodento-osseous dysplasia Neonatal severe osteosclerotic dysplasias Blomstrand dysplasia AR Raine dysplasia Prenatal-onset Caffey disease Lethal chondrodysplasias with fragmented bones Greenberg dysplasia AR Dappled diaphyseal dysplasia AR Astley-Kendall dysplasia AR Disorganized development of cartilaginous and fibrous components of the skeleton Dysplasia epiphysealis hemimelica Sp Multiple cartilaginous exostoses (three types) AD .

.

.

.

Enchondromatosis (Oilier) Enchondromatosis with hemangiomata (Maffucci) Spondyloendochondromatosis Spondyloendochondromatosis with basal ganglia calcification Dysspondyloenchondromatosis Metachondromatosis Osteoglophonic dysplasia Genochondromatosis Carpotarsal osteochondromatosis Fibrous dysplasia (Jaffe-Campanucci) Fibrous dysplasia (McCune-Albright) Fibrodysplasia ossificans progressiva Cherubism Cherubism with gingival fibromatosis 31. Osteolyses Multicentric predominantly carpal and tarsal a. Multicentric carpal-tarsal osteolysis with and without nephropathy b. Shinohara carpal-tarsal osteolysis Multicentric predominantly carpal, tarsal, and interphalangeal a. Francois syndrome b. Winchester syndrome c. Torg syndrome Whyte Hemingway carpal-tarsal phalangeal osteolyses Predominantly distal phalanges a. Hadju-Cheney syndrome b. Giacci familial neurogenic acro-osteolysis c. Mandibulo acral syndrome Predominantly involving diaphyses and metaphyses a. Familial expansile osteolysis b. Juvenile hyaline fibromatosis 32. Patella dysplasias Nail patella dysplasia Scyphopatellar dysplasia

Sp Sp AR AR

AD AD AD AD Sp AD AD AR

AD

AR AR AR AD

AD AR AR

AD AR AD AD

.

aModified from International Nomenclature and Classification of the Osteochondrodysplasias (1997) Am J Med Genet 79:376-382. bAR, autosomal recessive; AD, autosomal dominant; XLD, X-linked dominant; XLR, X-linked recessive; Sp, sporadic.

development as a cylindrization process. Abnormal bone modeling was a process characterized by (1) amplification, in which "the bone showing the greatest growth potential will show the greatest change, since it will magnify the same defect to a greater degree," (2) polarity, which defines the maximal direction of longitudinal growth because tubular bones grow in a differential pattern, one end predominating over the other, and (3) changes, which occur over time. The development of changes with time was delineated clinically by congenita and tarda variants. His work is valuable because of efforts to understand the complex patterning of bone formation and to explain developmental bone disease by localizing change to specific areas. Its main fault is the

rigid categorization of every disorder into a framework, which is accurate only for some. A difficulty in radiologic categorization now becoming apparent with the increase in reported syndromes is the fact that many disorders, including those defined by regional terms, affect adjacent areas radiographically and, if not apparent radiographically, must affect them nevertheless. Disorders defined as metaphyseal dysplasias must affect the physes if they are short stature syndromes. Efforts to encompass epiphyseal and metaphyseal involvement utilize terms such as spondyloeipmetaphyseal dysplasias, although the disorder pseudo-achondroplasia also involves each of the three areas. It appears that the limits of radiographic nosology

738

CHAPTER 9 9

Skeletal Dysplasias

have been reached and could benefit from simplification, although observations of extent and position of involvement in any individual case are essential for the patient involved. Bailey approached classification of short stature conditions by simplifying and concentrating on clinical anatomic features (8). Group 1 was disproportionate, absolutely short patients, group 2 was a spectrum group with a wide range of variability in stature and body proportions, and group 3 included proportionate short stature patients, the vast majority of which represent primary nonosseous problems with the bones shortened secondarily due to systemic growth failure. The current all-inclusive classification system is the International Classification of Osteochondrodysplasias defined initially by an international working group on bone dysplasias in 1972 and 1992 (17) and revised most recently in 1997 and published in 1998 (Table I) (262). The osteochondrodysplasias were divided into three groups in the 1992 classification: group A, defects of the tubular (and flat) bones and/or axial skeleton, 24 sections; group B, disorganized development of cartilaginous and fibrous components of the skeleton; and group C, idiopathic osteolyses. In the most recent version, the overall alignment of groups has been retained but the A, B, and C subdivisions have been eliminated. The classification represents an attempt to be all-inclusive, defining the disorders by using clinical and radiographic frames of reference. Over 150 disorders, not including some subtypes, are listed. This classification seeks to define disorders in which the skeletal system is involved either primarily or at least extensively. Though admirably complete it is obviously unwieldy clinically. Virtually all classifications attempted are either too brief to be definitive or so extensive that they cannot be used as clinical tools. Furthermore, any classification using solitary criteria, by which is meant clinical or radiographic or even molecular-genetic findings, invariably seems to lump diverse entities together or place similar entities in differing groups. It also is important to recognize that bone, joint, and limb malformations are present in an even larger number of developmental abnormality syndromes (95). A widely referred to compendium listing essentially all dysmorphic syndromes, Smith's Recognizable Patterns of Human Malformation, notes over 450 syndromes with a large number of disorders even beyond the intemational listing with minimal to severe skeletal involvement, such as facial bone abnormalities, hand and foot abnormalities, limb length alterations, vertebral structural defects with or without kyphoscoliosis, and joint region malformations such as hip dislocation, clubfoot, and elbow contractures (140). Elucidation of the underlying chromosomal, genetic, and molecular abnormalities is leading to more fundamental definitions of skeletal dysplasia and multiorgan dysmorphic disorders and to an evolving change in classification. Our approach considers several general criteria to be of importance in assessing a patient with a skeletal dysplasia (Table II).

lII. P R E V A L E N C E O F SKELETAL DYSPLASIAS In spite of the fact that several dysplastic syndromes have been identified, relatively few of the disorders still account for the majority of cases. These will differ in various areas throughout the world, but data from Europe, North America, and South Africa are roughly comparable. A study by Wynne-Davies and Gormley assessed individuals with skeletal dysplasias in Scotland, England, and Wales over a 30-year period (371). They were able to document 4383 patients with a skeletal dysplasia. The most common disorder, similar to a South African survey described later, was osteogenesis imperfecta involving 859 patients (19.6%). The next groups in frequency were those with a multiple epiphyseal dysplasia, 588 patients (13.4%); hereditary multiple extososis (diaphyseal aclasis), 452 (10.3%); spondyloepiphyseal dysplasia tarda, 406 (9.3%); achondroplasia, 226 (5.2%); pseudo-achondroplasia, 226 (5.2%); spondyloepiphyseal dysplasia congenita, 181 (4.1%); dyschondrosteosis, 181 (4.1%); and metaphyseal and spondylochondrodysplasias 181 (4.1%). These nine disorders represented 75.3% of the diagnosed skeletal dysplasias. In South Africa, a national skeletal dysplasia registry documented 1270 chondrodysplasias and indicated the number of patients seen in each disorder (16). The most common group, encompassing almost 27% of all disorders, was osteogenesis imperfecta with most of these type I or type III and a relatively small number of type II patients. The next most common groups involved those with achondroplasia, 8.6% of all patients (109 with achondroplasia, 21 with hypochondroplasia, and 15 with thanatophoric dysplasia), cleidocranial dysplasia 4.3%, sclerosteosis 6.4%, SED congenita 3.7%, pseudo-achondroplasia 3.6%, and multiple epiphyseal dysplasia 3.3%. In Finland, the most common skeletal dysplasia is diastrophic dysplasia (342).

IV. D I A G N O S I S O F SKELETAL DYSPLASIAS

A. Overview There are both lethal and nonlethal types of skeletal dysplasias. Many of the disorders now can be diagnosed prenatally by ultrasonography. All of the lethal disorders reach clinical awareness in the perinatal period. Some nonlethal conditions can be diagnosed clinically and radiographically at birth, whereas others do not manifest themselves for several months or years (295). Conditions that can be diagnosed at birth include achondroplasia, diastrophic dysplasia, Kniest syndrome, spondyloepiphyseal dysplasia congenita, cleidocranial dysplasia, chondrodysplasia punctata, metatropic

SECTION IV ~ Diagnosis of Skeletal Dysplasias TABLE II

Basic Considerations in Assessment o f a Skeletal Dysplasia Patient

1. Short stature

2. Disproportion a. Disproportionate short stature re extremity-trunk relationship

b. Disproportionate involvement within an extremity

3. Lethal-nonlethal

4. Classification criteria a. Clinical criteria b. Radiographic criteria

c. Genetic-molecular criteria d. Histopathologic criteria 5. Orthopedic considerations a. Spinal deformity

b. Neurological sequelae of spinal deformity

c. Upper-lower extremity deformity

d. Short stature

739

May not be fully established by reference to growth charts until 1 year of age Absolute: below 3rd percentile; <5 ft male and <4 ft 10 in. female at maturity Relative: low normal range but
7'40

CHAPTER 9 ~ Skeletal Dysplasias

dysplasia, mesomelic dysplasia, acromesomelic dysplasia, the various acrodysplasia syndromes, osteogenesis imperfecta types II and (usually) III, and osteopetrosis (malignant recessive variant) (7). Those that tend to become identifiable with time are the metaphyseal dysplasias, spondylometaphyseal dysplasia, pseudo-achondroplasia, multiple epiphyseal dysplasia, Morquio's disease, spondyloepiphyseal dysplasia tarda, hereditary multiple exostoses, hypochondroplasia, osteogenesis imperfecta type I, and benign autosomal dominant osteopetrosis.

B. Prenatal Assessment Over the past two decades, ultrasonography has allowed increasingly accurate diagnosis of skeletal dysplasias in utero (159). Linear array real time (rapid B scan) ultrasonography has established normal ranges for the length of fetal limbs (Figs. 1A and 1B), and several of the lethal and even nonlethal dysplasias now can be diagnosed in the second trimester. Reports of the ability to diagnose skeletal dysplasia by ultrasound began appearing in the late 1970s, and 25 years of experience has shown the value of this technique. The diagnosis is based primarily on measurement of long bone length from metaphysis to metaphysis and comparison of the lengths obtained to normal values. Other abnormalities detected involve decreased bone brightness, angular deformity of the shaft of the long bones, head shape, and fractures of the long bones. Measurement of the diaphyseal-metaphyseal lengths of the fetal long bones was established from 13 weeks of gestational age for the humerus, radius, and ulna and from 14 weeks of gestational age for the femur, tibia and fibula by Jeanty and Romero (267). Normal standards for the lengths of these bones were also established by Elejalde and Elejalde (70). Percentiles from the 3rd to the 97th were calculated for each of the long bones in 102 pregnancies with values listed from 10 to 40 weeks of gestation. The long bones were measured from metaphyseal bone proximally to metaphyseal bone distally with the maximum distances recorded. Larger studies using more sensitive equipment have recently been published documenting femoral length in cross-sectional studies. Chitty et al. (49) measured femoral lengthening in 649 fetuses of 12-42 weeks gestation, and Lessoway et al. (172) measured 790 femurs from 12 to 42 weeks. Both groups published centile charts and tables. There were no significant differences between male and female fetuses. Several studies have documented the accuracy of ultrasonographic determinations. The lethal perinatal type II osteogenesis imperfecta has been diagnosed sonographically between 16 and 19 weeks of pregnancy. Kurtz and Wapner reported 16 pregnancies in 15 women at high genetic risk for having fetuses with skeletal dysplasias in whom ultrasound examinations were performed during the second trimester (158). Eleven of 16 fetuses were normal and 5 were abnormal. At birth all 11 ultrasonographically determined normal fetuses were normal, and at birth or elective termination of

pregnancy the other 5 were abnormal. Hobbins et al. reported on ultrasound investigations of 66 women at risk for fetal skeletal dysplasia using femoral and humeral measurements (113). A positive diagnosis of dysplasia was made in 5 of 7 second trimester fetuses, whereas in 2 others subsequently born with dysplasia the overlap with normal did not allow for definitive diagnosis. All other fetuses were normal at birth following normal ultrasound evaluation. Filly et al. indicated that prenatal diagnosis of short limb dwarfs could be accomplished readily by the ultrasound technique (79), and Kurtz et al. pointed out the range of femoral measurements by ultrasound in achondroplasia (157). Gordienko detected 30 of 39 prenatally diagnosed skeletal dysplasias (77%) in the first and second trimesters including 46% before the 24th week (96). In those instances in which a conclusive postnatal diagnosis was made, the ultrasonographic accuracy in terms of specific syndromes was reasonably high at 73%. The other fetuses had a skeletal dysplasia although definitive verification was not present. There were 3 instances of OI type IIA with gestational diagnosis at 20, 20, and 22 weeks; OI type IIB, 3 cases correctly diagnosed at 19, 19, and 24 weeks; OI type IIC, 2 cases at 23 and 25 weeks; achondroplasia with correct diagnoses at 26 and 34 weeks; hypochondrogenesis diagnosed at 23 and 24 weeks; NoonanSaldino short polydactyly syndrome diagnosed at 19 and 24 weeks; thanatophoric dysplasia at 26 weeks; metatropic dysplasia at 11, 17, and 25 weeks; achondrogenesis type I at 10 and 20 weeks and type II at 23 weeks; polycampomelic syndrome at 32 weeks; diastrophic dysplasia at 22 weeks; chondroectodermal dysplasia at 24 weeks; and asphyxiating thoracic dysplasia at 22 weeks.

C. Clinical Examination At birth concern about a skeletal dysplasia is based on the presence of dysmorphic features, disproportion of limbtrunk relationships, orthopedic deformities, and abnormalities of bone length and shape. Patients who are being assessed in the neonatal period and indeed in the first year of life may not demonstrate short stature based on clinical height measurements and reference to growth charts, although the diagnosis of a skeletal dysplasia syndrome may be evident by clinical features and radiographic abnormalities. After 1 year of age the growth differential between normal and skeletally dysplastic children is sufficiently great that the presence of short stature can be diagnosed by using the growth charts. When a child is suspected of having a skeletal dysplasia syndrome, clinical examination helps to determine whether the relationship of the limbs to the trunk is proportionate or disproportionate. As a general rule, the bone structure of persons with short stature due to skeletal dysplasia is disproportionate, whereas those with short stature due to systemic medical causes such as endocrine, renal, hepatic, or cardiac disorders is proportionate. In patients whose skeletal proportions are normal, the outstretched fingers reach to the mid-

SECTION IV ~ Diagnosis of Skeletal Dysplasias

741

A Length of Humerus (mm)

Length of

Length of

Ulna (mm)

Radius (mm)

Percentile

Percentile

Percentile

Week

5th

50th

95th

5th

50th

95th

5th

50th

12

~

9

~

~

7

~

~

7

95th

13

6

11

16

5

10

15

6

10

14

9

14

19

8

13

18

8

13

14 17

15 16

12 15

17 20

22 25

11 13

16 18

21 23

11 13

15 18

20 22

17 18 19

18 20 23

22 25 28

27 30 33

16 19 21

21 24 26

26 29 31

14 15 20

20 22 24

26 29 29 32

20

25

30

35

24

29

34

22

27

21

28

33

38

26

31

36

24

29

33

22 23 24

30 33 35

35 38 40

40 42 45

28 31 33

33 36 38

38 41 43

27 26 26

31 32 34

34 39 42

25

37

42

47

35

40

45

31

36

41

26 27 28 29

39 41 43 45

44 46 48 50

49 51 53 55

37 39 41 43

42 44 46 48

47 49 51 53

32 33 33 36

37 39 40 42

43 45 48 47

30

47

51

56

44

49

54

36

43

49

31

48

53

58

46

51

56

38

44

50

32 33 34

50 51 53

55 56 58

60 61 63

48 49 51

53 54 56

58 59 61

37 41 40

45 46 47

53 51 53

35

54

59

64

52

57

62

41

48

54

36

56

61

65

53

58

63

39

48

57

37

57

62

67

55

60

65

45

49

53

38 39 40

59 60 61

63 65 66

68 70 71

56 57 58

61 62 63

66 67 68

45 45 46

49 50 50

54 54 55

FIGURE 1 Intrauterinegrowth in length has been documented by ultrasonography from 12 to 40 weeks. Measurements (in mm) include only the bony shaft (metaphysis--diaphysis--metaphysis) with the proximal and distal epiphyses not included. Values include 5th, 50th, and 95th percentile measurements. (A) Humerus, ulna, radius; (B) tibia, fibula, femur. [Reprinted from Hobbins and Benacerrat (1989), "Diagnosis and Therapy of Fetal Anomalies," pp. 174 and 175, Academic Press, with permission.]

thigh, a finding seen even in the neonatal period. In those whose skeletons are disproportionate, the fingers reach either to the trochanteric region, as a result of short limb dwarfism such as achondroplasia, or to the knee region, as a result of short spine dwarfism such as Morquio's spondyloepiphyseal dysplasia (mucopolysaccharidosis IV). In a short limb dwarfism the extremities are more affected than the spine and in a short trunk dwarfism the reverse is true. A second type of disproportion in many skeletal dysplasias is seen within each

extremity with an abnormality in the normal length relationships of the proximal, middle, and distal segment bones. Micromelia refers to shortening of length of the entire extremity but this shortening is usually disproportionate; if the proximal (humeral-femoral) segment is most affected the shortening is referred to as rhizomelic (eg, achondroplasia), if the middle segment (radius-ulna-tibia-fibula) is most affected the shortening is referred to as mesomelic (eg, dyschondrosteosis), and if the distal segment (carpus, metacarpals,

742

CHAPTER 9 9 Skeletal

Week

5th

Dysplasias

Length of Tibia (mm)

Length of

Length of

Fibula (mm)

Femur (mm)

Percentile

Percentile

Percentile

50th

95th ~ ~ 17 20 22 25 27 30

5th

50th

95th

5th

50th

95th

6 9 13 13 15 19

6 9 12 15 18 21

~ ~ 19 21 23 28

4 6 9 12 15 18

8 11 14 17 20 23

13 16 18 21 24 27

23 26

31 33

21 24

25 28

30 33

12 13 14 15 16 17 18 19

~ ~ 7 9 12 15 17 20

7 10 12 15 17 20 22 25

20 21 22 23 24 25 26 27 28 29

22 25 27 30 32 34 37 39 41 43

27 30 32 35 37 40 42 44 46 48

33 35 38 40 42 45 47 49 51 53

21 24 27 28 29 34 36 37 38 41

28 31 33 35 37 40 42 44 45 47

36 37 39 42 45 45 47 50 53 54

26 29 32 35 37 40 42 45 47 50

31 34 36 39 42 44 47 49 52 54

36 38 41 44 46 49 51 54 56 59

45 47 48 50 52 53 55 56 58 59 61

50 52 54 55 57 58 60 61 63 64 66

55 57 59 60 62 64 65 67 68 69 71

43 42 42 46 46 51 54 54 56 56 59

49 51 52 54 55 57 58 59 61 62 63

56 59 63 62 65 62 63 65 65 67 67

52 54 56 58 60 62 64 65 67 68 70

56 59 61 63 65 67 68 70 71 73 74

61 63 65 67 69 71 73 74 76 77 79

30 31 32 33 34 35 36 37 38 39 40 FIGURE 1 (contmue~

tarsus, metatarsals, phalanges) is most affected the shortening is referred to as acromelic (eg, Apert syndrome). Although diagnoses of a skeletal dysplasia can be made frequently in the newborn period, there actually may be no short stature at that time, but by 1 year of age the growth differential between the skeletally dysplastic child and a normal child becomes apparent and those with a skeletal dysplasia are either below the third percentile or in the lowest reaches of the normal range. After the first year of life, however, the child with a skeletal dysplasia will grow along the same established percentile. Some of the most severe dysplasias have premature growth plate closure around 10-

12 years of age, such as occurs in diastrophic dysplasia. In those children who are short because of a medical disorder, the limbs and trunk are symmetrical or proportionate but there almost always is a continuing fall-off in height percentile with time. Growth charts have been established for some of the more common dysplasias, including achondroplasia (125) and diastrophic dysplasia, spondyloepiphyseal dysplasia congenita, and pseudo-achondroplasia (123). Scott has summarized the adult height ranges in several of the more common skeletal dysplasias (277). Three characteristic clinical findings are seen in many types of skeletal dysplasia. The first is a depressed nasal

SECTION IV ~ Diagnosis of Skeletal Dysplasias bridge, because this area is formed initially in cartilage, whereas the rest of the face and skull (excluding the skull base) develop from intramembranous bone which is not affected in chondrodysplasias. The second involves a lack of full extension of the elbows. This is usually a mild, nonprogressive deformity of 10-15 ~ with little to no clinical significance. The third is a lumbar lordosis, which can become rigid in the second decade but is relatively nonprogressive.

D. Radiographic Examinations Radiographs play a major role in diagnosing skeletal dysplasia syndromes, in following their development through time, and in assessing the need for surgical intervention to correct structural abnormalities (Table II). A complete skeletal survey is warranted in the initial assessment. Films that are particularly helpful are anteroposterior and lateral skull, the cervical spine with anteroposterior, open mouth, and lateral views in flexion, neutral, and extension, posteroanterior views of the wrist and hand, anteroposterior and lateral projections of the elbows, hips, and knees, anteroposterior and lateral views of the thoracolumbar vertebrae, and an anteroposterior film of the lumbar and sacral regions. The range of radiographic abnormalities is summarized in Table II and reviewed in detail for specific disorders in later sections. Structural abnormalities of the skull and axial skeleton characterize many skeletal dysplasias, and radiographs of these regions are an integral part of assessment. Radiographic abnormalities of the long bones in the skeletal dysplasias can involve shortening, irregularities of the secondary ossification centers, shaping abnormalities centered primarily in the epiphyses, metaphyses, or diaphyses, irregularity of the articular cartilage surface, angular deformity, and altered radiodensity. Most of the skeletal dysplasias will show radiographic evidence of abnormalities in the secondary ossification centers. This can involve fragmentation of the secondary center, as is seen in multiple epiphyseal dysplasia particularly in the proximal femoral capital epiphysis, or a marked delay in the appearance of the secondary center, as occurs with spondyloepiphyseal dysplasia congenita in the proximal femoral capital epiphysis in which it is absent as late as 7 or 8 years of age and in the knee region in which the centers of the distal femur and proximal tibia are not present at birth (in normal individuals they are). Metaphyseal abnormalities are characterized by a broadening of the region due to abnormalities in resorptive remodeling and irregularities at the physeal-metaphyseal junction with cartilage persistence in the metaphysis, features seen with the metaphyseal or spondylometaphyseal dysplasias. There also can be prominent exostoses at the periphery of the metaphyseal area, for example, in hereditary multiple exostoses. In rare diaphyseal dysplasias, which are not short stature syndromes, the epiphyseal and metaphyseal regions are structurally normal but there is widening of the diaphysis due to

7'43

increased periosteal new bone formation; this is seen in the progressive diaphyseal dysplasia syndrome of CamuratiEnglemann. The entire bone can be radiodense, as seen in osteopetrosis, or radiolucent, as seen in the inherited osteoporosis syndromes, the most common of which is osteogenesis imperfecta. If the patient has a short stature syndrome, the epiphyseal region by definition must be abnormal, although the structural irregularity shown on the plain radiograph may be present only in the metaphysis. The underlying biologic features of the abnormalities must be considered carefully in assessing the nature of a skeletal dysplasia because descriptive terms based on plain radiographs can be misleading. For example, groups of the skeletal dysplasias are referred to as metaphyseal dysplasias or metaphyseal dysostoses, although clinical examination clearly indicates the presence of short stature and radiographs often show some physeal widening. The primary abnormality, therefore, must reside in the physis of the epiphysis even though the main structural or radiographic manifestation is metaphyseal. Radiographic assessment of patients with skeletal dysplasia has two major values. The first is related to syndromal diagnosis and the second to orthopedic management. The current literature is weighted heavily to the diagnostic categorization side with numerous books and articles by radiologists and medical geneticists using the structural abnormalities detected on plain films to categorize patients and syndromes. There is a relative paucity of studies in the orthopedic literature that relate specific structural deformities to the degree of clinical symptoms on the one hand and to the need for and results of operative intervention on the other. There has also been relatively little reference to the changing patterns of radiographic abnormalities throughout the period of skeletal development in each of the various disorders. For example, long bone shape as seen on radiographs of a child with diastrophic dysplasia of 1 or 2 years of age may well be normal, even though by age 10 years the shortened stature, deformed epiphyses, and joint space narrowing will be well-established. In other disorders, such as metatropic dysplasia, the change in radiographic profile with time has been well-documented. The clinical and radiographic features have been welldocumented in such texts as Disproportionate Short Stature by J. A. Bailey (8), Heritable Disorders of Connective Tissue by V. A. McKusick (198), Inherited Disorders of the Skeleton by P. Beighton (15), Les Maladies Osseuses De l'Enfant by Marotaux (191), and Bone DysplasiasmAn Atlas of Constitutional Disorders of Skeletal Development by J. W. Spranger, L. O. Langer, and H. R. Wiedemann (313). This chapter will not concentrate on detailed repetition of these radiographic descriptions.

E. Laboratory Studies Some basic blood and urine studies are warranted in a skeletal dysplasia work-up. The patients are almost always

744

CHAPTER 9 ~ Skeletal Dysplasias

systemically well. Phosphorus levels will be altered in hypophosphatemic rickets (low phosphorus), which can mimic an osteochondrodystrophy. The alkaline phosphatase is elevated in osteogenesis imperfecta and markedly diminished in hypophosphatasia. Osteopetrosis can show anemia, granulocytopenia, and a diminished platelet count. Urinalysis is helpful in hypophosphatemic rickets (increased urinary phosphate), hypophosphatasia (decreased phosphoethanolamine), and the mucopolysaccharidoses (increased loss of heparan, dermatan, and keratan sulfate). Screening for mucopolysaccharides (MPS) in urine is the simplest biochemical test for MPS (177). If positive, then leukocytes and cultured skin fibroblasts are assayed for specific enzyme activities. Skin biopsy for fibroblast culture and collagen typing is used increasingly for specific diagnosis in osteogenesis imperfecta. The more recent detection of molecular abnormalities in several of the skeletal dysplasia syndromes, as outlined in Section VI, is still done in most instances by specific research laboratories, but it is expected that many of these analyses will be available via commercial diagnostic laboratories shortly.

V. C H R O M O S O M E A B N O R M A L I T Y S I T E S IN S K E L E T A L D Y S P L A S I A S Chromosome abnormality sites as well as the gene involved have been mapped in a large number of the skeletal dysplasias (84, 199). These are listed in Table III.

VI. G E N E T I C A N D M O L E C U L A R A B N O R M A L I T I E S IN SKELETAL DYSPLASIAS The vast majority of skeletal dysplasias are hereditary disorders caused by genetic abnormalities. The subsequent structural anatomic abnormality thus is dependent on the specific molecular abnormality and the role that molecule plays in the development or maintenance of the skeleton. There are several ways of assessing the accumulating information on these genetic and molecular abnormalities. These include (1) the mutation families into which the abnormalities are clustered, (2) the type of molecular function defects, and (3) the phase of the developmental pattern of skeletogenesis in which the abnormality has its negative effect. The specific genetic abnormalities are listed in Tables III and IV in this chapter and also in Table VIA of Chapter 1 in relation to the skeletal gene database. Abnormalities in processing the glycosaminoglycans were defined in the 1950s as the underlying causes of the mucopolysaccharidoses. The fact that abnormalities relating to these molecules would lead to skeletal deformity was not surprising because proteoglycans are some of the most prominent and extensive components of the cartilage matrix.

Processing abnormalities of each of the four major proteoglycans have been defined, and six definitive syndromes are associated with these. In the most commonly occurring syndrome, Morquio's disease (MPS IV), the radiographic findings are those of a spondyloepiphyseal dysplasia, but the identification of biochemical abnormalities allowed for classification under that terminology rather than under the radiographic pattern abnormality. There was a relative lack of additional molecular findings in other skeletal dysplasias over the subsequent decades in spite of extensive analyses performed, but identification of genetic and molecular abnormalities has increased markedly recently. With the increasing molecular definition of several types of collagen and recognition of the concentration of certain types in cartilage tissue, many skeletal dysplasias have now been defined as collagen abnormalities. Among the first to be defined was the lethal skeletal dysplasia achondrogenesis, which was shown to be associated with deficient production of type II collagen with massive subsequent implications for cartilage model and skeletal formation. In hypochondrogenesis, another lethal chondrodysplasia, cultured chondrocytes at autopsy showed a type II collagen gene (COL2A1) mutation with a G ~ A transition, causing a glycine substitution at amino acid 574 of the Pro(II) collagen triple-helical domain (124). Other studies have indicated abnormal synthesis of type II collagen in the spondyloepiphyseal dysplasia congenita syndromes, but problems remain in the interpretation of the information (2, 168). Although the hip is almost invariably abnormal in SEDC having a coxa vara deformity and the spine also has abnormalities, in the anterior portion of the cartilage model of the vertebral bodies there is generally a normal appearance of other major joints such as the knee and ankle. Because an abnormality in type II collagen COL2A1 should affect that molecule throughout the body, it remains unclear how the currently defined abnormality can translate on the one hand into major structural abnormalities of the vertebrae and proximal femur while leaving the distal femur, proximal tibia, and other regions structurally unaffected. Several reviews of molecular defects have been presented (118-120, 187, 214, 215,261,368).

A. Mutation Families Several syndromes have now had identification in relation to the chromosomal abnormality, the gene locus, and in many instances the specific mutation. All chromosomes except 13, Y, and the mitochondrial chromosome have been defined as the site of gene mutation responsible for at least one of the skeletal dysplasias. Mapping studies of the skeletal dysplasias have uncovered considerable heterogeneity within individual disorders, for example, in hereditary multiple exostoses (with three types defined), chondrodysplasia punctata, multiple epiphyseal dysplasia, and the several variants of osteogenesis imperfecta. Another finding is that clinically separate

SECTION VI 9 Genetic and Molecular Abnormalities in Skeletal Dysplasias

TAB LE llI Gene symbol

Chromosomal location

AOM2

lp21

EDM2 EDS6

lp32 lp36.3-P36.2

HOPS1

lp36.1-p34

PYCD SJS

lq21 1p36.1-p34

EDS3

2q31-q32.3

EDSA4

2q31-q32.3

FN1

2q34 2q31 2q24-q31

745

Genetic Skeletal Disorders w i t h C h r o m o s o m a l Location and Gene Defect a

Disorder

Chromosome 1 Arthro-ophthalmopathy, progressive, with deafness, Stickler syndrome Multiple epiphyseal dysplasia, Fairbank type Ehlers-Danlos, type VI, ocular scoliotic form (PLOD deficiency) Hypophosphatasia, infantile

Gene defect

Collagen type XI (COLIIA) Collagen type IX, o~2 Lysyl hydroxylase, protocollagen (PLOD) (putative) Alkaline phosphatase, liver, bone, kidney (ALPL)

PSD Y2 SHFM4

A CAA CDLSL CDMJ

3p23-p22 3q26.31-q27.3 3p22-p21.1

HYPOCA

3q 13.3-q21 3q 13.3-q21 3q 13.3-q21

LRS1

3p21.1-p14.1

MFS2 MMDL

3p25-p24.2 3q22.3-q23

MPS4B

3p21.33

NSHPT

3q 13.3-q21

ACH

4p16.3

CDPR

4p16-p14 4p16.3 4p16.3 4p16

FHH1 HPT2

CRS5B CRS7C CRS8

Pycnodysostosis Chondrodystrophic myotonia, Schwartz-Jampel syndrome Chromosome 2 Ehlers-Danlos syndrome, type III, with joint hypermobility, minor hyperextensibility and softness of the skin Ehlers-Danlos syndrome, type IV, arterial, dominant (COL3A1 defect) Ehlers-Danlos syndrome, type X Polysyndactyly, type 2 Split hand and foot deformity, type 2, including syndactyly, brachysyndactyly, camptodactyly (?same gene as PSDY2) Chromosome 3 Rhizomelic chondrodysplasia punctata Cornelia de Lange-like syndrome with dup (3q) Metaphyseal chondrodysplasia, Jansen type Hypercalcemia hypocalciuric, familial, benign Hypoparathyroidism AD Mild hypocalcemia with hyperphosphatemia and normal PTH levels, autosomal dominant Larsen syndrome of multiple dislocations of large joints Marfan syndrome 2 Mesomelic dwarfism, with hypoplastic ulna, fibula and mandible, Langer type Mucopolysaccharidosis, type IVB; Morquio disease B Severe neonatal hyperparathyroidism Chromosome 4 Achondroplasia, hypochondroplasia (FGFR3 defect) Chondrodysplasia punctata, rhizomelic 1 Crouzon syndrome with acanthosis nigricans Pfeiffer syndrome Craniosynostosis, syndromatic 8, Adelaide type

Cathepsin K, lysosomal (CTSK)

Collagen type III, c~l (COL3A1)

Collagen type III, c~l (COL3A1) ?Fibronectin 1 (FBN1) Homeobox HOXD13

?Acetyl-CoA acyltransferase (ACAA) Parathyroid hormone, parathyroid hormonerelated peptide receptor (PTHR) Parathyroid calcium sensing receptor (PCAR1) Parathyroid calcium sensing receptor (PCAR1) Parathyroid calcium sensing receptor (PCAR1)

Galactosidase, p 1 (GLB1) Parathyroid calcium sensing receptor (PCAR1)

Fibroblast growth factor receptor 3 (FGFR3)

Fibroblast growth factor receptor 3 (FGFR3) Fibroblast growth factor receptor 3 (FGFR3)

(continues)

746

CHAPTER 9 9 Skeletal Dysplasias TABLE III (continued)

Gene symbol

Chromosomal location

EVC

4p16

GNPTA

4p21-q23

MPS1

4p16.3

TNTP

4p16.3

ACG1B ATSG2 CCA

5q31-q34 5q31-q34 5q23-q31

CRS4 DTD MPS6

5q34-q35 5q31-q34 5pll-q13

TCOF1

5q32-q33.1

CCD2 CDMS OCDD

6p21 6q21-q22.3 6p22-p21.3

OSMED

6p22-p21.3

ACS3

7p21.2

CRS1 CRS2 EDS7A2 GCPS MFSV MPS7 OI1B

7p22-p21 7p15-p13 7q21.3-q22.1 7p13 7q21.3-q22.1 7ql 1.21-ql 1 7q21.3-q22.1

OI2B

7q21.3-q22.1

OI3B

7q21.3-q22.1

OI4B

7q21.3-q22.1

OPM

7q21.3-q22.1

Disorder

Ellis-van Creveld syndrome, chondroectodermal dysplasia Mucolipidosis, type II, I cell disease; mucolipidosis type III, pseudo-Hurler polydystrophy Mucopolysaccharidosis, type I, Hurler, Scheie disease (IDUA deficiency) Thanatophoric dysplasia, types 1 and 2 Chromosome 5 Achondrogenesis IB, Fraccaro type Atelosteogenesis Contractural arachnodactyly, congenital, Marfanlike disorder Craniosynostosis, familial, primary, Boston type Diastrophic dysplasia Mucopolysaccharidosis, type VI, MaroteauxLamy disease (ARSB deficiency) Treacher Collins-Franceschetti syndrome Chromosome 6 Cleidocranial dysplasia 2 Chondrodysplasia, metaphyseal, Schmid type Osteochondrodysplasia with mild spondyloepiphyseal dysplasia, osteoarthrosis, sensorineural deafness, dominant Sticklerlike without eye involvement Otospondylomegaepiphyseal dysplasia (chondrodystrophy with sensorineural deafness, Nance-Insley syndrome) (COL11A2 defect) Chromosome 7 Acrocephalosyndactyly III, Saethre-Chotzen syndrome with craniosynostosis Craniosynostosis, syndromatic 1 Craniosynostosis, syndromatic 2 Ehlers-Danlos syndrome, type VIIA2 Greig cephalopolysyndactyly syndrome Marfan syndrome variant Mucopolysaccharidosis, type VII; Sly disease Osteogenesis imperfecta, types I and IA with dentinogenesis Osteogenesis imperfecta, type II, congenital, lethal Osteogenesis imperfecta, type III, severe, nonlethal, dominant and recessive Osteogenesis imperfecta, type IV, moderately severe Osteoporosis, postmenopausal (some cases)

Gene defect

N-Acetylglucosamine (GNPTA) phosphotransferase Iduronidase, Ot-L (IDUA) Fibroblast growth factor receptor 3 (FGFR3)

Sulfate transporter (DTDST) Sulfate transporter (DTDST) Fibrillin 2 (FBN2) Homeodomain encoding gene (MSX2) Sulfate transporter (DTDST) Aryl-sulfatase B (ARSB)

Collagen type X, ctl (COLIOA1) Collagen type XI, chain (COL11A2)

Collagen type X1 (COL11A2)

Helix loop, helix transcription factor (TWIST)

Collagen type I, oL2 (COLIA2) Zinc finger protein (GL13) Collagen type I, oL2(COLIA2) Glucuronidase, 13(GUSB) Collagen type I, ct2 (COL1A2) Collagen type I, oL2(COL1A2) Collagen type I, oL2(COL1A2) Collagen type I, oL2 (COLIA2) Collagen type I, ct2 (COL1A2) (continues)

SECTION VI ~ Genetic and Molecular Abnormalities in Skeletal Dysplasias

747

TABLE IIl (continued) Gene symbol

Chromosomal location

SHFM1

7q36 7q36 7q21.3-q22.1

TPT

7q36

CA2

8q22

CCD1 CPDD1

8q22 8q

CRS7A

8pll.2-pll.1

EXT1

8q24.1 8q24.11-q24.13

PPD2 PSDY1

Disorder

Preaxial polydactyly type 2 Polysyndactyly, complex, bilateral Split and foot deformity, type 1, including some ectrodactyly, ectodermal dysplasia, cleft palate, EEC syndrome Triphalangeal thumb, opposable (?and nonopposable) type (same as PSDY 1)

SGM1

8q22-q22.3

TRP1

8q23.3-q24.13

Chromosome 8 Osteopetrosis with renal tubular acidosis and cerebral calcifications Cleidocranial dysplasia 1 Chondrocalcinosis 1 (calcium pyrophosphate deposition disease) with early-onset osteoarthritis Craniosynostosis syndromatic 7, with characteristic broad thumbs and big toes, Pfeiffer syndrome Exostoses, multiple, congenital 1 Langer-Giedion syndrome, trichorhinophalangeal syndrome, may be associated with bilateral agenesis of tibia and first metatarsus (contiguous gene syndrome) Segmentation gene 1 (vertebral fusions with laryngeal abnormalities), Klippel-Feil syndrome Trichorhinophalangeal syndrome I

NPS1

9p13 9q34.1

Chromosome 9 Cartilage-hair hypoplasia Nail patella syndrome 1

ACS1

10q26

CRS5A

10q26

CRS6

10q26

CRS7B

10q26

CRS9

10q26

BWS

11p15.5

EXT2

llp12-pll.12

LGCR

CHH

Gene defect

Chromosome 10 Acrocephalosyndactyle type 1, Apert syndrome with craniosynostosis and syndactyly (FGFR2 defect) Craniosynostosis, syndromatic 5, Crouzon syndrome with craniofacial dysostosis Craniosynostosis, syndromatic 6, JacksonWeiss syndrome with foot anomalies and great phenotypic variability Craniosynostosis, syndromatic 7, Pfeiffer syndrome, with characteristic broad thumbs and big toes (FGFR2 defect) Craniosynostosis, syndromatic 9, BeareStevenson syndrome

Carbonic anhydrase II, cytosolic, ubiquitous

Fibroblast growth factor receptor 1 (FGFR1)

EXT1, tumor suppressor gene

Contiguous gene with deletion of TRP1 and EXT1

Fibroblast growth factor receptor 2 (FGFR2)

Fibroblast growth factor receptor 2 (FGFR2) Fibroblast growth factor receptor 2 (FGFR2)

Fibroblast growth factor receptor 2 (FGFR2)

Fibroblast growth factor receptor 2 (FGFR2)

Chomosome 11 Beckwith-Wiedemann, exomphalosmacroglossia-gigantism syndrome Exostoses, multiple, congenital, 2, also associated with, craniosynostosis, mental retardation in a contiguous gene syndrome with 1lp deletions (continues)

748

CHAPTER 9 9 Skeletal Dysplasias

TABLE III (continued) Gene symbol

Chromosomal location

ACG2 AOM1

12q12-q13 12q12-q13

ATDJ

12p 12.2-p 11.21

HOS

12q23-q24.1

KND MPS3D

12q12-q13 12q14

NSED

12q12-q13

OAP PDDR SEDC SEMD

12q12-q13 12q14 12q12-q13 12q12-q13

VDR

12q13

Disorder

Chromosome 12 Achondrogenesis II, Langer-Saldino type Arthro-ophthalmopathy, progressive, with deafness, Stickler syndrome Asphyxiating thoracic dystrophy of the newborn, Jeune syndrome Holt-Oram syndrome excluding heart-hand syndrome, type III Kneist dysplasia Mucopolysaccharidosis, type III, Sanfilippo disease, type D Spondyloepiphyseal dysplasia, mild, Namaqualand type; Osteoarthrosis, precocious Pseudo-vitamin D dependency tickets, type I Spondyloepiphyseal dysplasia, congenital Spondyloepimetaphyseal dysplasia congenita, Strudwick type Rickets, vitamin D-dependent, type II, with alopecia

Gene defect

Collagen type II, oL1 (COL2A1) Collagen type II, oL1 (COL2A1)

T-box transcription factor (TBX5) Collagen type II, oL1 (COL2A1) N-Acetylglucosamine-6-sulfatesulfatase (GNS)

Collagen type II, c~l (COL2A1) Collagen type II, otl (COL2A1) 25(OH)D 3-1o~hydroxylase Collagen type II, o~1 (COL2A1) Collagen type II, o~1 (COL2A1) Vitamin D receptor

Chromosome 13 Chromosome 14

CRS3 MFS1 PHP1B

15q15 15q21.1 15ql 1-q13

SGS

15q21.1

MPS4A

16q24.3

AP1J

17q21-q22

CMPD1

17q24.3-q25.1

EDS7A1 MPS3A

17q21.31-q22.15 17q25./3

MPS3B

17qll-q21

OI1A

17q21.31-q22.15

OI2A

17q2!.31-q22.15

Chromosome 15 Craniosynostosis, syndromatic 3 Marfan syndrome 1 Albright hereditary osteodystrophy 2, pseudohypoparathyroidism type IB Shphintzen-Goldberg syndrome marfanoid craniosynostosis Chromosome 16 Mucopolysaccharidosis, type IVA; Morquio disease A Chromosome 17 Symphalangism, proximal (ankylosis of the proximal interphalangeal joints, often associated with conductive deafness) Campomelic dysplasia, 1 including mild campomelic Ehlers-Danlos syndrome, type VIIA 1 Mucopolysaccharidosis IliA, Sanfilippo disease type A Mucopolysaccharidosis, type IIIB, Sanfilippo disease type B Osteogenesis imperfecta, types I and IA with dentinogenesis imperfecta Osteogenesis imperfecta, type II, congenital, lethal

Fibrillin (FBN1)

Fibrillin (FBN1)

Glactose-6-sulfatase (GALNS)

DNA binding protein SOX9 Collagen type I, oL1 (COLIA1) Sulfamidase (SPHM) Acetylglucosaminidase, oLM (NAG) Collagen type I, otl (COLIA1) Collagen type I, al (COLIA1) (continues)

SECTION VI 9 Genetic and Molecular Abnormalities in Skeletal Dysplasias

749

TABLE III (continued) Gene symbol

Chromosomal location

OI3A

17q21.31-q22.15

OI4A

17q21.31-q22.15 17q25

SRS

Disorder

Gene defect

Osteogenesis imperfecta, type III, severe, nonlethal, dominant and recessive Osteogenesis imperfecta, type IV Silver-Russell syndrome

Collagen type I, etl (COLIA1) Collagen type I, ctl (COLIA1)

Chromosome 18

EDM1

19p13.1

EXT3 PSA CH

19p 19p 13 19p13.1

AMCD

20q I 1.2

MCAS

20q 13.2

PHPIA

20q 13.2

CDPX1

HSH HYP1

Xp22.32 Xq27-q28 Xpter-p22.2 Xp22.2 Xp22.2-p22.1

HYP2

Xp 11

MPS2

Xq27.3-q28 Xq26-q28 Xp22.3-p21.3 Xq26 Xp 11.22

FHH2

CDPX2 CFND

OPD1 SEDL SHFM2 XLRH

Chromosome 19 Multiple epiphyseal dysplasia, Fairbank type, allelic to PSACH Exostoses, multiple, congenital, 3 Hypercalcemia hypocalciuric, familial, benign Pseudo-achondroplasia Chromosome 20 Acromesomelic chondrodysplasia, HunterThompson type McCune-Albright polyostotic fibrous dysplasia with caf6 au lait spots and endocrine dysfunction Albright hereditary osteodystrophy 1, pseudohypoparathyroidism type I, A Chromosome X Chondrodysplasia punctata 1, recessive type Chondrodysplasia punctata 2, dominant type Craniofrontonasal dysplasia Hypomagnesemia with secondary hypocalcemia Hypophosphatemia with vitamin D-resistant rickets, excluding the gyromouse homologous type Hypophosphatemia with vitamin D-resistant rickets and hypercalcuria, gyromouse homologue, with inner ear abnormalities (allelism with HYP1, not definitely excluded) Mucopolysaccharidosis, type II, Hunter disease Otopalatodigital syndrome, type 1 Spondyloepiphyseal dysplasia tarda Split hand and foot deformity X-linked recessive hypophosphatemic tickets, likely including Dent syndrome

Cartilage oligomeric matrix protein (COMP)

Cartilage oligomeric matrix protein (COMP)

Cartilage-derived morphogenetic protein (CDMP1)

Somatic mutation (GNAS1)

Guanine nucleotide binding protein, OL stimulating activity, polypeptide 1 (GNAS1)

Arylsulfatase E (ARSE)

Phosphate regulating gene (PEX)

Iduronate 2-sulfatase (IDS)

Chloride voltage-gated channel 5 (CLCN5)

aDerived from Frezal J, Le Merrer M, Chauvet ML (1997) Pediatr Radio127:366-387, and Human Genetic Disorders Chart 1997, Journal of NIH Research.

disorders are the result of mutations in the same gene. These disorders are now described as belonging to mutation groups or mutation families relating to allelic mutations of the same gene (Table IV). Spranger introduced the concept of bone dysplasia families in which disorders of differing severity

would be grouped on the basis of common genetic and molecular abnormalities with different specific mutations within the gene (118, 307). At present these are concentrated in 11 families, although further discoveries of abnormality are being presented quite

750

CHAPTER 9 9 Skeletal Dysplasias TABLE IV More Common Skeletal Dysplasia Mutation Groups

1. Fibroblast growth factor receptor (FGFR) mutation group a. FGFR 3

Thanatophoric dysplasia Achondroplasia Hypochondroplasia b. FGFR 2

Crouzon syndrome Apert syndrome Jackson-Weiss syndrome Pfeiffer syndrome c. FGFR 1

Pfeiffer syndrome 2. Collagen mutation group a. Type I collagen COLIA1 COLIA2

Osteogenesis imperfecta b. Type II collagen COL2A1

Achondrogenesis type II Hypochondrogenesis SED congenita Kniest dysplasia Stickler dysplasia Strudwick dysplasia SED tarda c. Type.IX collagen COL9A2

Multiple epiphyseal dysplasia type (EDM2) COL9A3

Multiple epiphyseal dysplasia d. Type X collagen COLI OA1

Schmid metaphyseal dysplasia e. Type XI collagen COL11A2

3. Diastrophic dysplasia sulfate transporter mutation group DTDST

Achondrogenesis type IB Atelosteogenesis type 2 Diastrophic dysplasia 4. Arylsulfatase gene mutation family ARSE

Chondrodysplasia punctata, X-linked recessive 5. Lysosomal enzyme gene mutation family Mucopolysaccharidoses Mucolipidoses 6. Cartilage oligomeric matrix protein mutation group COMP

Multiple epiphyseal dysplasia, type 1 (EDM1) Pseudo-achondroplasia 7. Parathyroid hormone-parathyroid hormone-related peptide receptor mutation group PTHrPR

Jansen metaphyseal dysplasia 8. SOX mutation family SOX 9

Campomelic dysplasia 9. Short stature homeobox-containing mutation family SHOX

Leri-Weil dyschondrosteosis Langer mesomelic dysplasia 10. Alteration of synthesis of cell surface heparan sulfate glycosaminoglycan EXT1 EXT2

Hereditary multiple exostoses 11. Mutations involving transcription factor CBFA1 CBFA1

Cleidocranial dysostosis

Stickler-like, mild SED

rapidly. The abnormalities identified are presented in this section, whereas their functional roles are outlined in Sections VI.B and VI.C. 1. FIBROBLAST GROWTH FACTOR RECEPTOR (FGFR) MUTATION GROUP The fibroblast growth factors and their receptors (FGFR) help to regulate bone growth, and there are many interactions between FGFs and FGFRs in skeletal development and growth that are both temporally and spatially variable. Mutations in FGFRs 1, 2, and 3 have been shown to occur in many skeletal dysplasias (51,210, 354). a. F G F R 3 . The fibroblast growth factor receptor 3 mutation group includes three separate disorders including the lethal thanatophoric dysplasia (330), the most common skeletal dysplasia, achondroplasia (21,269, 290), and the mild variant hypochondroplasia.

b. F G F R 2 . The FGFR2 group has shown mutations with Crouzon syndrome (craniofacial dysostosis), Apert syndrome, Jackson-Weiss syndrome, and Pfeiffer syndrome (51,210). e. F G F R 1 . The FGFR1 group is also seen in some cases of Pfeiffer syndrome.

2. COLLAGEN MUTATION FAMILY The main structural collagen in bone, tendon, ligament, and dermis is type I collagen. There are four main collagens present in cartilage, types II, IX, X, and XI, and each type has been shown to have mutations associated with specific skeletal dysplasias. a. Type I Collagen. Mutations in type I collagen, COL1A1 and COL1A2, are associated with osteogenesis imperfecta. This disorder is reviewed extensively in Sections VII.D and XII.CC in this chapter.

SECTION V! ~ Genetic and Molecular Abnormalities in Skeletal Dysplasias b. Type H Collagen (COL2AI). Mutations of the type II collagen gene are associated with many skeletal dysplasias especially because type II collagen is a primary matrix protein of physeal and epiphyseal cartilage (116, 124, 126, 187). All mutations have involved only one of the two COL2A1 alleles and map to the triple-helical domain. Most involve the substitution of glycine residues critical to assembly of the triple helix. This group encompasses achondrogenesis type II (71), hypochondrogenesis (124, 213), SED congenita (2, 216, 337), SED tarda, Kniest dysplasia (31, 78, 309, 357, 369), Stickler dysplasia, and spondyloepimetaphyseal (Strudwick) dysplasia (142, 336). c. Type IX Collagen. COL9A2: A large family with multiple epiphyseal dysplasia always affecting the knee joints and often other more peripheral upper and lower extremity joints but never affecting the shoulder or hip joints has been found with mutations in the COL9A2 gene (346). This is currently referred to as MED type 2 (or EDM 2). COL9A3: Paassilta et aL (23) and Bonnemann et al. (36) described families with MED also with invariable knee involvement with a mutation in the COL9A3 gene. There was also a mild proximal myopathy detected in the cases studied by Bonnemann et al. (36). d. Type X Collagen (COLIOAI). Schmid metaphyseal dysplasia has been associated with a mutation causing a 13-base-pair deletion in one type X collagen allele (20, 352). The mutations map to the region of the gene that encodes the C-propeptide. The subsequent disruption of helix formation prevents half of the type X collagen chains from entering into molecule formation, diminishing the number of molecules by half. e. Type XI Collagen (COLIlA2). A Stickler-like syndrome with mild spondyloepiphyseal dysplasia, osteoarthritis, and sensorineural hearing loss but no eye involvement has been linked with mutations of type XI collagen, COLllA2 (174, 347).

3. DIASTROPHIC DYSPLASIA SULFATE TRANSPORTER (DTDST) MUTATION FAMILY These disorders are related to the diastrophic dysplasia sulfate transporter gene, which was found to be the causative genetic abnormality in diastrophic dysplasia (102-104) but is also the gene mutation site for achondrogenesis type 1B (Fraccaro type) (43) and atelosteogenesis type II (323-325). Several different mutations have been found in each of the three disorders. Older reports of lethal diastrophic dysplasia are now defined as atelosteogenesis II. 4. CARTILAGE OLIGoMERIC MATRIX PROTEIN (COMP) MUTATION FAMILY These disorders result from mutations from the gene for cartilage oligomeric matrix protein, which now includes pseudo-achondroplasia and some forms of multiple epiphyseal dysplasia (10, 37, 38, 52, 65, 108). MED type 1 (EDM1)

751

is characterized by relatively severe joint involvement in particular hips, knees, and ankles. 5. PARATHYROIDHORMONE RELATED PEPTIDE RECEPTOR (PTHRPR) MUTATION FAMILY Mutations of the gene encoding a G-protein transmembrane spanning receptor for parathyroid hormone-parathyroid hormone-related peptide have been found in Jansen metaphyseal dysplasia (276). 6. SOX9 MUTATION FAMILY SOX9 is a member of a family of transcription factor genes structurally related to the sex determining region Y genes. SOX9 transcripts have been seen in growth plate chondrocytes although their function is not known. Mutations of SOX9 have been reported in several patients with campomelic dysplasia, a usually lethal skeletal dysplasia with bony malformation and sex reversal. 7. SHORT STATURE HOMEOBOX CONTAINING (SHOX) MUTATION FAMILY Mutations (deletions) have been identified in the SHOX gene in patients with Leri-Weil dyschondrosteosis (19, 289) and Langer mesomelic dwarfism. 8. ARYLSULFATASE GENE (ARSE) MUTATION FAMILY The arylsulfatase gene ARSE has been found with mutations in some patients with X-linked recessive chondrodysplasia punctata. 9. LYSOSOMAL ENZYME MUTATION FAMILY Abnormalities of specific lysosomal enzymes lead to the mucopolysaccharidoses and mucolipidoses (115, 221 ). 10. CELL SURFACE HEPARAN SULFATE PROTEOGLYCAN MUTATION GROUP Three genes have been associated with hereditary multiple exostoses, referred to as EXT1 on 8q24.1, EXT2 on 11 p 11-13, and EXT3 on 19p. EXT1 and EXT2 have been found to be proteins with glycosyltransferase activities required for the synthesis of the glycosaminoglycan heparan sulfate. These two members of the EXT family of tumor suppressors are involved in heparan sulfate biosynthesis. Heparan sulfate proteoglycans are present both on cell surfaces and in the extracellular matrix. 11. MUTATIONS INVOLVING TRANSCRIPTION FACTOR CBFA1 Mutations in CBFA1, a transcription factor essential for osteoblast differentiation, have been found to be associated with cleidocranial dysplasia (212) (Table III and Chapter 1 Table VIA). The collagen II and FGFR3 chondrodysplasias are caused by dominant mutations, but the DTDST chondrodysplasias

752

CHAPTER

9 9 Skeletal Dysplasias

are recessive. Some disorders previously grouped together on a radiologic basis as metaphyseal dysplasias are now separated on the basis of differing genetic abnormalities. It is becoming evident that, whereas molecular classification brings us closer to a clarification of the causation of disorders in some instances, it further complicates the matter of classification. Previously, attempts were made to classify disorders based on radiographic structural similarities, which seemed to make good sense. As the metaphyseal chondrodysplasia group shows, disorders with similar radiographic appearances can be caused by different chromosomal, genetic, and molecular abnormalities, whereas disorders from the same gene frequently cause markedly varying dysplasias, for example, thanatophoric dysplasia, which is lethal, and hypochondroplasia, which is extremely mild and in many instances undetectable other than by slightly short stature.

B. Molecular Function Defects Another approach to interpreting genetic and molecular information is to attempt to classify the disorders based on general categories defining the functions of the abnormal genes and molecules. Rimoin has categorized these defects into (1) defects in structural proteins of cartilage, (2) inborn errors of cartilage metabolism, (3) local regulators of cartilage growth, and (4) systemic defects influencing cartilage development (261). He also has suggested additional categories for genes that have been identified with skeletal dysplasias but whose molecular abnormalities are unknown and other disorders in which a gene location has been mapped but no further information is available. Rimoin's categorization is utilized and expanded upon here. 1. DEFECTS IN STRUCTURAL PROTEINS OF CARTILAGE These disorders involve abnormalities of types II, IX, X, and XI collagen and also of the cartilage oligomeric matrix protein (COMP). The disorders of the type II collagen gene (COL2A1), referred to as type II collagenopathies, are responsible for achondrogenesis type II, hypochondrogenesis, and some cases of spondyloepiphyseal dysplasia, spondyloepimetaphyseal dysplasia, Stickler dysplasia, and Kniest dysplasia. Type II collagen is the major collagen of cartilage and consists of three identical polypeptide oL1 (2) chains encoded by the COL2A1 gene located on chromosome 12. Mutations in the COL2A1 gene give rise to a spectrum of clinical phenotypes varying from lethal to mild arthropathy only. Because type II collagen is found in cartilage, the vitreous humor of the eye, and the intervertebral disks, there is a relatively high association of ocular problems such as myopia and retinal detachment and vertebral abnormalities in these dysplasias. These disorders include achondrogenesis II-hypochondrogenesis. The most severe form of achondrogenesis has a striking micromelia, a hydropic appearance,

and death in utero or in the early hours of life. Hypochondrogenesis is characterized by a very short stature and hydrops at birth with a flat face, widely spaced eyes, and small rib cage, also leading to death within hours or weeks. The skeleton is less severely affected. The vertebral bodies are ossified and the tubular bones are longer. In both types there are single amino acid substitutions for a glycine residue, all clustered near the carboxy-terminal end. Hypochondrogenesis, however, is a type II collagenopathy caused by heterozygous substitution mutations of glycine interrupting the obligatory Gly-X-Y repeat of the triple-helical domain of oL1 (2) chain. The mutations result in cartilage matrix with a reduced amount of type II collagen with increased levels of posttranslational modifications. The marked diminution of type II collagen leads to predominantly type I and III collagens in the cartilage matrix and minor amounts of type XI. Even the small amounts of type II collagen produced were posttranslationally overmodified and remained intracellular. Bones were formed, however, suggesting that cartilage formation and bone development can take place in the absence of type II collagen. Over 50 mutations have been found at the type II collagen gene locus (COL2A1). If one considers that the collagen 2A1 molecule involves the triple-helical region, the aminopropeptide end, and the amino telopeptide as well as the carboxy telopeptide and carboxy-propeptide end, all mutations found so far have been in the triple-helical domain. The abnormalities vary from lethal perinatal types involving achondrogenesis type II and hypochondrogenesis, to those of intermediate severity involving SED congenita and Kniest dysplasia, to the mildest forms involving SED tarda with early osteoarthritis. The SED variants have shown a variety of single-nucleotide substitutions throughout the molecule as well as deletions and insertions. Kniest dysplasia has exon skipping mutations clustered around the amino-terminal end of the molecule. Stickler dysplasia also is present. Most of the disorders other than Stickler are felt to disrupt the formation of the type II collagen triple helix, leading to the abnormally deformed molecules that either are degraded prematurely or are incorporated in an imperfect fashion into the extracellular fibrillar network. Rimoin et al. found a direct correlation in type II collagenopathy skeletal dysplasias between the ratio of type I to type II collagen in cartilage and clinical severity. In the most severe lethal achondrogenesis II disorder only type I collagen is present in cartilage. Normally type I collagen is not found in cartilage. In the slightly less severe hypochondrogenesis there is both type I and posttranslationally overmodified type II collagens, whereas in SED type I is not present and type II is both normal and overmodified. Type IX collagen abnormalities (COL9A2 and COL9A3) have been found in some families with the multiple epiphyseal dysplasia type. Both linkage and mutational changes have been noted. Multiple epiphyseal dysplasia cases have

SECTION VI ~ Genetic and Molecular Abnormalities in Skeletal Dysplasias

also been defined in relation to abnormalities of the cartilage oligomeric matrix protein (COMP), although these so far are phenotypically similar to the type IX variants. Heterozygous mutations in type X collagen (COLIOA1) have been defined in Schmid-type metaphyseal dysplasia in the region of the gene, which holds the carboxyl-terminal chain domain and presumably leads to a reduced amount of type X collagen in the matrix. Mutations in type XI collagen (COL11A2) have been described in a family with Stickler syndrome. Mutations of the COL11A2 gene encoding the o~II chain of type XI collagen have been described in families with Stickler dysplasia-like phenotypes presenting with mild spondyloepiphyseal dysplasia, precocious osteoarthritis, and sensorineural hearing loss, although they lack the eye abnormalities typical of the syndrome. Other types of the Stickler syndrome have resuited from type II collagen defects. It is in the collagen type II disorder that the families have severe myopia and vitreoretinal degeneration. The other structural cartilage protein defect identified so far is in the cartilage oligomeric matrix protein (COMP) present in the short arm of chromosome 19. COMP is a member of the thrombospondin family of proteins found in the extracellular matrix of cartilage. Abnormalities of this gene have been detected in pseudo-achondroplasia (PSACH) as well as in multiple epiphyseal dysplasia (Fairbanks type). Some types of MED and pseudo-achondroplasias thus are allelic and appear to share common pathogenetic features. Pseudo-achondroplasia is a genetically distinct autosomal dominant condition in which heterozygous mutations in the cartilage oligomeric matrix protein (COMP) gene have been identified. Several cases of PSACH have been studied in which COMP mutations have been found. Several differing mutations have been found within the COMP region. 2. INBORN ERRORS OF CARTILAGE METABOLISM

There are three disorders involving the diastrophic dysplasia sulfate transporter (DTDST) group found in the long arm of chromosome 5q. Abnormalities of the gene lead to reduced sulfate transport, which has a major effect on sulfation of chondroitin sulfate containing proteoglycans and on type IX collagen, which has a proteoglycan component. Mutations in the DTST gene have been described in diastrophic dysplasia, in a more severe disorder achondrogenesis 1B, and in a third disorder, atelosteogenesis type II. The DTDST mutations identified in each of these conditions demonstrate a series of allelic phenotypes of variable severity. The other two defects in this category involve enzymes of sulfate metabolism and degradation, respectively. The first group includes mutations of genes encoding for enzymes of sulfate metabolism such as arylsulfatase (ARSE), which are associated with chondrodysplasia punctata and lysosomal enzymes, abnormalities of which lead to the mucopolysaccharidoses and mucolipidoses. The mucopolysac-

753

charidoses (MPS) are a heterogeneous group of disorders that result from inherited deficiencies of enzymes involved in the degradation of proteoglycans. They are classifed into seven types based on clinical features and specific enzyme defects. Genetic abnormalities of specific enzymes lead to an intralysosomal accumulation of glycosaminoglycans followed by excess excretion of these substances in the urine. 3. LOCAL REGULATORS OF CARTILAGE GROWTH In early 1994, linking studies placed the achondroplasia gene on the short arm of human chromosome 4. The particular region of the chromosome affected was known to be the position where the gene for fibroblast growth factor receptor 3 (FGFR3) was present. This protein spans the cell membrane and consists of three extracellular immunoglobulinlight domains, a lipophilic transmembrane domain, and intracellular torroisin domains. FGFR3 is expressed in cartilage and brain, and studies in the mouse showed that it was able to mediate the effect of fibroblast growth factor on chondrocytes. The role of FGFR3 is particularly strong in the growth plate, although it appears to involve negative regulation of intrinsic growth rates with the mutation causing the receptor to become constitutively active. Two groups reported analysis of the FGFR3 gene in achondroplastic individuals, with both groups finding an FGFR3 mutation in the DNA from affected persons. No such mutations were found in DNA from unaffected persons. Unlike some other skeletal disorders, both groups found that mutations were at exactly the same nucleotide of the transmembrane domain of the FGFR3 gene. Shiang et al. found that 15 of 16 achondroplasia mutations had a guanine-to-adenine transition at nucleotide 1138 and the 16th had a guanine-to-cytosine transition at the same nucleotide (290). Both mutations resulted in the substitution of arginine for glycine at amino acid 380 of the protein. Rousseau et al. found that all 23 achondroplasia mutations in their patients resulted in the same substitution at the same amino acid of the transmembrane domain at FGFR3 protein (269). Mutations at the same site (codon 380) have now been found in several hundred patients, with the normal glycine residue replaced by an arginine residue, Gly380Arg (21). The high proportion of identical mutations, which is 100% for the amino acid change, helps to explain the consistency of the phenotype in achondroplasia. Prenatal diagnosis of achondroplasia, homozygous achondroplasia, and the unaffected state is now possible. Hypochondroplasia, a mild autosomal dominant skeletal disorder with clinical similarities to achondroplasia, is also linked to the FGFR3 locus on chromosome 4. FGFR3 mutations have been identified and cluster to amino acid residue 540 (N540K) where asparagine is replaced by lysine (22, 23). Unlike achondroplasia, however, hypochondroplasia shows heterogeneity because not all patients have the FGFR3 mutation (193). Rousseau et al. (270) detected the mutation in 21 of 29 cases, whereas Prinster et al. (148) found it in only

754

CHAPTER 9 ~ Skeletal Dysplasias

9 of 18 cases. Both studies found the mutation in those with relative macrocephaly, and Prinster et al. felt that the mutation was found in the more definitive cases with unchangednarrowed lumbar interpedicular distances and fibulas longer than the tibia. The fibroblast growth factor receptor 3 (FGFR3) family of disorders involves achondroplasia, hypochondroplasia, and thanatophoric dysplasia (35, 53). In the FGFR3 gene abnormality family almost complete homogeneity leading to achondroplasia has been demonstrated. Initially, the two separate papers by Shiang and Rousseau indicated that 37 of 39 achondroplastic alleles carried the same mutation, a G-to-A transition at nucleotide 1138, and that the remaining two alleles had a G-to-C transversion at the same nuclear sites, thus leading to substitution of glycine by an arginine residue at position 380 (Gly380Arg) of the protein. Further genetic homogeneity was demonstrated by Bellus et al., who found the Gly380Arg mutation in 153 of 154 achondroplastic alleles (21). Further studies have indicated that 192 of 194 achondroplastic alleles carded the Gly380Arg mutation, of which over 95% are due to the G-to-A mutation. In thanatophoric dysplasia (TD) heterozygous mutations were found to cluster mainly to four locations in the FGFR3 gene. Those with thanatophoric dysplasia type I with straight femurs and the severe cloverleaf skull deformity all have the same mutation with a Lys650Glu substitution in the intracellular tyrosine kinase domain of the receptor. In thanatophoric dysplasia type I with curved femurs with or without the cloverleaf skull, most of the mutations were found in the extracellular domain sharing a substitution of cysteine for another amino acid. Nerlich et al. defined two cases similar to thanatophoric dysplasia in which molecular analysis disclosed distinct mutations of the FGFR3 gene, which were identical to those previously found in other cases of TD, but there were slightly differing clinical manifestations indicating that factors other than the mutation itself appear to modulate the phenotypic expression of the disease (220). The achondroplasia mutations map to the transmembrane domain of the mature protein, hypochondroplasia mutations to the proximal tyrosine kinase domain, thanatophoric dysplasia I to the immunoglobulin domains, and TD II to the distal tyrosine kinase domain.

somal enzymes. Other rare disorders involve adenosine deaminase deficiency (ADA), which can lead to metaphyseal dysplastic changes. There are abnormal genes that have been identified in some disorders, although the actual protein that is affected has not yet been revealed. Hereditary multiple exostoses is an autosomal dominant disorder that has now been shown to be heterogeneous, with different chromosomal locations at chromosomes 8, 11, and 19. On chromosome 8, the trichorhinophalangeal syndrome type II is contiguous to the multiple extososes gene. This work in relation to genetic and molecular abnormalities is beginning to indicate that the number of genes that harbor chondrodysplasia mutations may not be as large as previously suspected. Well over half of all humans with chondrodysplasias so far have been defined to mutations at one of two loci: COL2A1 or FGFR3. Chondrodysplasia families are now being recognized at the molecular level, with the two largest families of COL2A1 representing spondyloepiphyseal dysplasias and FGFR3 representing achondroplasia families. Mutations of COL2A1 tend to be dispersed throughout the gene. Mutations in genes whose products interact functionally with type II collagen, such as COL11A2, cause similar phenotypes. SED phenotypes have wide variability resulting from dysfunction of collagen fibrils containing type II, IX, and XI collagens. On the other hand, mutations of FGFR3 are different showing clustering to a few codons. These result in abnormal signaling disorders, and relatively subtle differences in the dose signals generated by the receptors lead to different degrees of deformity depending on the specific mutation. Mutations in several genes in mice have been noted to be associated with abnormal bone and cartilage development, and similar abnormalities almost certainly underlie several of the human osteochondrodysplasias as the early discoveries reported earlier show. Mutations have been noted in oncogenes c-src and c-fos, homeodomain genes involved in pattern formation such as pax1 and pax3, zinc finger genes, and growth factor genes such as TGF~L

4. SYSTEMICDEFECTS INFLUENCING CARTILAGE DEVELOPMENT Genes that switch undifferentiated mesenchymal cells to specific tissue lines are just beginning to be defined. The Ylinked test is the determining factor; S 1-y starts male gonadal development and also chondrocyte differentiation. Mutations in the S 1-y-related gene SOX9 occur in campomelic dysplasia, which has severe bowing and angulation of long bones, small scapulae, deformed pelvis and spine, and missing fibs. Other factors in cartilage development involve peroxisomal defects such as are found in the rhizomelic form of chondrodysplasia punctata with a deficiency of the peroxi-

Another approach to understanding how specific mutations lead to skeletal dysplasias is to assess skeletal development from the embryonic period on, showing how specific molecules in the developmental cascade appear and how abnormalities follow if they are mutated. Mundlos and Olsen have described heritable disorders of the skeleton using this framework (214, 215). Their approach includes disorders associated with abnormal mesenchymal condensation or differentiation. These early disorders of regional patterning lead to early inappropriate model formation following which skeletal development would of necessity be abnormal. The approach also includes disorders associated with abnormal

C. Phase of the Developmental Cycle in Which Abnormality Has Its Negative Effect

SECTION VII ~ Lethal Perinatal Skeletal Dysplasias

proliferation or maturation of cartilage, by which abnormality of the endochondral ossification sequence was established; abnormalities of intramembranous bone formation; disorders associated with defects in collagenous extracellular matrix components; defects associated with noncollagenous matrix components; abnormalities of sulfate transport and metabolism; and disorders associated with abnormal matrix homeostasis. Each of the mutation families in Section I and the specific functional defects in Section II also can be slotted into the framework outlined in this latter approach. This approach is helpful, however, in relation to the earliest phases of embryonic bone formation in which mesenchymal condensation and differentiation lead to pattern formation with involvement of the homeobox transcription factor family or H O X genes, the paired box PAX genes affecting segmentation and neurogenesis, and the bone morphogenetic proteins (BMPs), which are signaling molecules influencing the size and shape of skeletal elements. Surprisingly, perhaps, relatively few abnormalities of the H O X genes have been discovered in human abnormalities, even though they represent extremely numerous and intricate series of patterning molecules. One such disorder, however, has been the syndrome referred to as synpolydactyly in which, in humans, expansions of a polyalanine stretch within the aminoterminal non-DNA-binding region of H O X D - 1 3 have been shown to cause synpolydactyly characterized by insertion of an extra digit between digits 3 and 4 and variable syndactyly. In the group defined as disorders associated with abnormal proliferation or maturation of cartilage, it is the FGFR group that has been implicated. These involve abnormalities of growth plate function in the achondroplasiahypochondroplasia-thanatophoric dysplasia disorders along with signaling disorders related to mutations in the parathyroid hormone-related peptide (PTHrP) now known to be associated with the Jansen-type metaphyseal dysplasia. Analogous to the FGFR3 abnormality in causing physeal problems are those of FGFR1 and -2, which have a major negative effect on cranial sutures that form by the intramembranous bone mechanism. Mutations here cause varying degrees of craniosynostosis or premature fusion of the cranial sutures along with abnormalities of the fingers and toes and on occasion tarsal and metatarsal bones. Mutations in these acrocephalosyndactyly syndromes are associated with the major role of fibroblast growth factor genes in limb development, many of which are expressed in the apical ectodermal ridge in which proximodistal growth in the developing limb bud is controlled. Following the synthesis of models of the developing bone by the processes of mesenchymal condensation and then cartilage or intramembranous bone model formation, growth and development then are dependent on the presence of normal amounts and types of matrix proteins and their appropriate processing. Abnormalities, therefore, of collagen and noncollagenous proteins such as COMP and abnormal sulfate processing can lead to major deformations as outlined

755

previously with the DTDST family of mutations. FGFR abnormalities also cause premature cranial suture fusion, leading to several craniosynostosis syndromes. The craniosynostosis in Crouzon syndrome is due to mutations in FGFR2. Apert, Jackson-Weiss, and Pfeiffer syndromes have limb defects as well as craniosynostosis and also show mutations in FGFR2 and (Pfeiffer) FGFR1. The latter three sometimes are referred to as acrocephalosyndactyly syndromes. All have craniosynostosis plus limb abnormalities ranging from complete syndactyly (bone and/or skin fusion of the digits) of hands and feet in Apert syndrome, to broad big toes with tarsal-metatarsal fusion in Jackson-Weiss syndrome, to broadening and deviation of thumbs and big toes without syndactyly in Pfeiffer syndrome. The other local cartilage regulator disorder affects the parathyroid hormone-related peptide receptor (PTHrPR). A mutation in this receptor leads to the Jansen-type metaphyseal chondrodysplasia. The mutation was a histidine-toarginine substitution at position 223 in the receptor protein. The functional abnormality relates to alterations in signaling and signal transduction between growth plate chondrocytes.

VII. L E T H A L P E R I N A T A L SKELETAL DYSPLASIAS

A. Diagnostic Profile Some skeletal dysplasias are lethal in the perinatal period, primarily because they have caused severe underdevelopment of the skeletal system in general and the ribs and thorax in particular. The disorders in this group are disproportionate short limb disorders. As detailed studies of stillborn and early postnatal deaths increase, the frequency of lethal neonatal skeletal dysplasias is becoming known. They appear to occur in 1-2 per 10,000 newborns. The most common among these include achondrogenesis, thanatophoric dwarfism, homozygous achondroplasia, short rib syndromes with or without polydactyly, chondrodysplasia punctata (rhizomelic recessive type), campomelic dysplasia, osteogenesis imperfecta type II, and most types of asphyxiating thoracic dystrophy (Jeune's syndrome). Less common conditions are fibrochondrogenesis, hypochondrogenesis, and atelosteogenesis. A detailed list of lethal disorders is presented in Table V. The disorders have been reviewed in detail by Spranger and Maroteaux (311). Orthopedic management is rarely indicated due to the extremely short life span. Each of three large studies of lethal skeletal dysplasias has noted similar findings concerning the incidence of the specific disorders. The studies by Tretter et al. (29 patients) (339), Rasmussen et al. (35 patients) (254), and Sharony et aL (134 patients) (288) each find thanatophoric dysplasia to be the most common (48%, 34%, and 24%, respectively). The next most common disorder is osteogenesis imperfecta type II (21%, 26%, and 26%), followed by achondrogenesis (7%,

756

CHAPTER 9 ~ Skeletal Dysplasias TABLE V Classifications of the Lethal O s t e o c h o n d r o d y s p l a s i a s (LOC)a

1. Hypophosphastasia and morphologically similar disorders Hypophosphastasia Lethal metaphyseal dysplasia 2. LOC with spotted calcifications Lethal Conradi-Htinermann disease Rhizomelic chondrodysplasia punctata Greenberg dysplasia 3. Achondrogenesis and related disorders Achondrogenesis IA (Houston-Harris) Achondrogenesis IB (Fraccaro) Achondrogenesis II Hypochondrogenesis 4. Thanatophoric dysplasia and related disorders Thanatophoric dysplasia type I Thanatophoric dysplasia type II Homozygous achondroplasia Glasgow variant 5. Platyspondylic LOC San Diego type Luton type Torrance type Shiraz type Opsismodysplasia 6. Lethal metatropic dysplasia and related disorders Lethal metatropic dysplasia Fibrochondrogenesis Schneckenbecken dysplasia 7. Kniest-like LOC Dyssegmental dysplasia, Silverman type Dyssegmental dysplasia, Rolland-Desbuquois type Lethal Kniest disease Blomstrand dysplasia 8. Short rib polydactyly syndromes Asphyxiating thoracic dysplasia Saldino-Noonan type Verma-Naumoff type Yang type Le Marec type Majewski type Beemer type 9. LOC with prominent diaphyseal abnormalities Campomelic syndrome Boomerang dysplasia Atelosteogenesis De la Chapelle dysplasia McAlister dysplasia Pseudo-diastrophic dysplasia 10. Osteogenesis imperfecta Osteogenesis imperfecta type II 11. LOC with gracile bones aModified from Spranger J, Maroteaux P (1990) Adv Hum Genet 19:1-103.

3%, and 13%), short rib syndromes (10%, 6%, and 8%), campomelic dysplasia (7%, 11%, and 10%), and atelosteogenesis (3.5%, 0, and 2%). Approximately two-thirds of all lethal skeletal dysplasias are thanatophoric dysplasia, osteogenesis imperfecta type II, and achondrogenesis. Other rarer lethal skeletal dysplasias included anywhere from 0 to 20% of involved cases. In the Maryland study, 20 out of 27 pregnancies were terminated with an average age at detection of 21.6 weeks, whereas 7 pregnancies proceeded to birth at an average at detection of 29.2 weeks. Almost 60% of cases, therefore, were diagnosed prior to 22 weeks of gestation corresponding to the typical timing of mid-second trimester screening sonography. The major diagnostic finding by ultrasound was short limbs in 88%, indicating the necessity of that measurement as part of routine sonography. The authors felt that almost 80-90% of lethal skeletal dysplasias were currently being detected in pregnancies receiving prenatal care, which was indicative not only of the frequency of sonography but also its increasing accuracy. In spite of this, however, although a lethal skeletal dysplasia syndrome could be diagnosed with great accuracy, a specific antenatal diagnosis was made in only 48%. This was almost identical to the 45% and 50% accuracy rates from the other two major studies. In a large retrospective study by Lachman, the most common disorders diagnosed in a large series of 226 fetuses and stillborns included osteogenesis imperfecta type II (18%), thanatophoric dysplasia (14%), campomelic dysplasia (6%), and achondrogenesis (5%). Brief summaries of ultrasound findings in 22 skeletal dysplasias were given (159).

B. Thanatophoric Dysplasia This is the most common lethal skeletal dysplasia. It is the most severe variant of the FGFR3 molecular abnormality group, which also includes the nonlethal conditions achondroplasia and hypochondroplasia. It is characterized by hypotonia, feeble fetal activity, and polyhydramnios. Affected individuals are usually stillborn or die in the neonatal period, but a patient surviving for 169 days has been reported. The clinical appearance is characteristic of a skeletal dysplasia in general and achondroplasia in particular, with short limbs, a prominent forehead, a depressed nasal bridge, and narrow thorax (15). Two types have been recognized. In TDI, the femurs are relatively short and curved and there may or may not be a cloverleaf skull deformity (Fig. 2Ai); in TDII, the femurs are straight and relatively longer but the cloverleaf skull is common (Fig. 2Aii). Other variants are described. The characteristic and pathognomonic radiographic finding in TDI is in the femurs, which are short and remarkably curved, leading to the older "telephone receiver" terminology regarding configuration (Fig. 2Ai). There are also characteristic H-shaped vertebrae on anteroposterior radiographs, narrowed lumbar interpedicular distances caudally, a narrow

SECTION VII ~ Lethal Perinatal Skeletal Dysplasias

F I G U R E 2 Radiographic examples of several cases of lethal skeletal dysplasia. (A) Thanatophoric dysplasia. (Ai) In TD type I there are short, curved femurs with the characteristic " C " or "telephone receiver" shape. The acetabulae are horizontal and there is a small sciatic notch. The latter two radiographic findings also are characteristic of achondroplasia. (Aii) In TD type II the femurs are slightly longer and less curved but pelvic findings are the same as in type I. (B) The lethal perinatal variant of osteogenesis imperfecta (Sillence type II, OI congenita A) is characterized by short, broad, and crumpled femurs along with deformed tibias with 45-90 ~ of diaphyseal bending. Multiply fractured and broadened ribs and a wormian bone skull deformation are characteristic. (C) LangerSaldino achondrogenesis is characterized by a disproportionately large head, a shortened trunk, and even more shortened upper and lower extremities. The head can make up as much as 40-50% of entire body length. There is marked underossification of the axial skeleton particularly of the lower thoracic, lumbar, and sacral regions. The cervical and upper thoracic pedicles almost always are present, but there is variable and diminishing mineralization of the vertebral bodies. Long bone formation is markedly abnormal with the bones shortened and often lacking any of the characteristic shapes. [Reprinted from Eyre et al. (1986)', Am. J. Hum. Genet. 39:52-67, 9 University of Chicago, with permission.] (D) Specimen radiograph of chondrodysplasia punctata distal femur shows characteristic punctate radiodensities in the epiphysis. The calcification is present throughout the epiphyseal region and not simply in that area that ultimately would undergo secondary ossification center formation. (E) Radiograph of a newborn with campomelic dysplasia shows the characteristic diaphyseal bending of both femurs (Ei) and tibias (Eii).

757

758

CHAPTER 9 ~ Skeletal Dysplasias

FIGURE 2 (continued)

sacrosciatic notch, and horizontal acetabulum. Ultrasonic diagnosis in utero has been made.

C. Homozygous Achondroplasia Because achondroplasia is by far the most common of the disproportionate short stature skeletal dysplasias, there have been numerous marriages between two affected individuals. Any child born to two affected parents has a 1 in 4 chance

of inheriting two faulty genes, causing an even more severe dwarfism called homozygous achondroplasia. Those affected look clinically like achondroplasts, but radiographic changes are more marked and respiratory death usually occurs within weeks after birth although some have survived beyond early infancy to 29-37 months. The acetabulae are horizontal as in achondroplasia, but the limbs are proportionally even more shortened (315). Many lethal cases, but not all, are associated with severe stenosis of the foramen mag-

SECTION VIi ~ Lethal Perinatal Skeletal Dysplasias num with ventriculomegaly and cervical myopathy. Surgical decompression has been performed but remains problematic with questions arising about its safety and effectiveness and whether this anatomic site is the sole cause of lethality (208).

D. Osteogenesis Imperfecta Lethal perinatal Sillence type II OI, OI congenita A. These severe variants of osteogenesis imperfecta (OI) are almost invariably lethal, although the affected individual may be stillborn, die in the first few hours after birth, or live for several weeks to months and occasionally even into the second year (280, 296). The disorder is considered in detail in the later section on osteogenesis imperfecta (Section XII.CC). A radiographic example is shown in Fig. 2B.

759

In both types, infants are often stillborn or premature or die soon after birth and demonstrate pulmonary hypoplasia and diminutive thoracic volume. (These cases are distinct from the nonlethal Grebe-Quelce-Salgado disorder termed achondrogenesis II in some older reviews; hence, the variable nomenclature.) In utero diagnosis by ultrasound has been reported. The disorder is due to nonexpression of type II collagen.

F. Hypochondrogenesis This lethal dysplasia is similar to but structurally a little milder than achondrogenesis II (Langer-Saldino). There is extremely short stature, hydrops at birth, an oval, fiat face with widely spaced eyes, often cleft palate, and relatively longer tubular bones (compared with achondrogenesis II) (344).

E. Achondrogenesis Achondrogenesis is a lethal newborn dwarfing syndrome in which the limbs are extremely short and the head is disproportionately large. It is currently classified into two main subtypes, achondrogenesis type I (Parenti-Fraccaro variant) and achondrogenesis type II (Langer-Saldino variant) (71) (Fig. 2C) (344). Type I cases exhibit defective ossification of endochondral and membrane bones and are subdivided further into groups IA and lB. Type IA is characterized by multiple rib fractures, proximal femurs with metaphyseal spikes, a partially membranous bone cranial vault, pedicles ossified in the cervical and upper thoracic regions only, and no vertebral body ossification. Type IB has no rib fractures and distal femoral metaphyseal irregularities, an ossified cranial vault, and ossified pedicles throughout the entire spine. In type II, membrane bone formation appears to be normal. There is normal cranial ossification but severe underossification of the axial skeleton. The pedicles are ossified but vertebral body ossification is variable. The ribs are short but without fracture. Radiographs of achondrogenesis II show clavicles and bones of the cranial vault to be normal, whereas the ribs are slightly short and without fractures. The maxilla and other facial bones are diminutive; the mandible is more nearly normal but is still slender and pointed at the symphysis. The tubular bones are extremely short, appearing somewhat cupped at their ends, but the intrinsic trabecular pattern is normal. The extremities are extremely short; the hands and feet are very small with short bones and very little ossification of the middle and distal phalanges. The bones of the spine consist of tiny ossification centers, paired in the thoracic and lumbar area and somewhat larger in the upper cervical region. The vertebral canal is narrow with the spinal cord fitted snugly within it. No ossification centers are visible in the sacrum or sternum. The iliac bones are small and crescent-shaped with a convex border superiorly and a concave border inferiorly. It is difficult to identify acetabula on radiographs as there is no ossification of pubis or ischial bones; however, the hips are not dislocated.

G. Atelosteogenesis This lethal neonatal dysplasia shows a micromelic dwarfism, bowed lower extremities, clubfoot, dislocated elbows, and often a cleft palate. The vertebral bodies are poorly ossified. The disorder is now recognized as a severe variant of the diastrophic dysplasia sulfate transporter (DDST) mutation family (293). Further characterization shows a large degree of variability, leading to categorization as atelosteogenesis type I (AOI), which includes the previously described boomerang dysplasia, atelosteogenesis type II (AOII), which includes the previously described de le Chapelle dysplasia, and AOIII. Target areas of marked abnormality are the humerus and femur, pelvis, fibula, calcaneus, vertebral bodies, and hand and foot phalanges. The humeral bone can be absent, oval-shaped, round, distally hypoplastic and tapered, or distally bifurcated. The femur can be boomerang-shaped, oval, rectangular, or distally tapered. The fibula is often nonossifled and the tibia is bowed. The calcaneus is often nonossifled. The lumbar vertebral bodies show coronal clefting in their mid-region on lateral radiographs. The proximal and middle phalanges can be nonossified, whereas the distal phalanges are ossified. The ilia are misshapen, whereas their lower one-third plus pubis and ischium often are nonossified.

H. Chondrodysplasia Punctata, Rhizomelic Form This represents a form of stippled epiphysis, which in its characteristic pattern is almost invariably lethal, but it must be distinguished from benign forms of stippled epiphysis included in the Conradi-Htinermann diagnostic category and from stippling of epiphyses accompanying other skeletal disorders such as some variants of multiple epiphyseal dysplasia (247). The characteristic radiographic finding is punctate radiodensities of the epiphyseal cartilage occasionally present in joints but usually involving several including the larger ones. The calcifications are not limited to the central

760

CHAPTER 9 ~ Skeletal Dysplasias

epiphyses but include the entire epiphyseal and often periepiphyseal regions (Fig. 2D). Death usually occurs within the first year of life. Features of lethal and benign variants were outlined by Jeune et al. in 1953 (139). The two general variants cannot necessarily be distinguished at birth so careful follow-up is needed. In a study of 40 cases from the literature, all lethal cases had death before 10 months of age with all beyond that age being long-term survivors. This is an autosomal recessive rhizomelic form characterized by marked symmetrical rhizomelia (proximal limb shortening) with the most prominent areas of involvement being the humerus and femur. Femoral (thigh) shortening is greater than humeral and usually is asymmetric with one side more involved. Lenticular cataracts and dermal abnormalities are present in approximately two-thirds of cases. There is a tendency to short digits with limited joint mobility. Mental deficiency is common. Death from respiratory complications such as tracheal stenosis usually occurs in the neonatal period and almost all die within the first 10 months of life, but some survive for 1-2 years. Spinal involvement is mild but vertebral body stippling can occur.

I. Campomelic Dysplasia The vast majority of patients with campomelic dysplasia are in the lethal category, dying in the neonatal period because of respiratory deficiency. Those affected rarely live beyond 1 year of age, and most are stillborn or die in the first few weeks of life. A few will survive but with feeding problems, failure to thrive, and serious central nervous system deficiency. The disorder is characterized by bending, primarily of the lower limbs, with intrauterine bony curvature of the diaphyseal regions of the femurs, tibias, and fibulas particularly prominent (Fig. 2E). There is delayed osseous maturation in the fetal period. The head is large, and there tends to be a large brain with hydrocephalus and gross cellular disorganization of the cerebral cortex and midbrain. Clubfoot is common. There is no ossification of the distal femoral and proximal tibial epiphyses in the newborn. There is marked fibular and scapular hypoplasia, a tendency to calcaneal-valgus foot deformities, numerical rib defects, ossification defects of the vertebral bodies, flat face with ocular hypertelorism, micrognathia, and cleft palate. The most severe bowing is of the tibia usually present at the junction of the middle and distal one-thirds associated with a cutaneous dimple. Femoral bowing tends to be in the midshaft and often is from 70 ~ to 90 ~. Laterally dislocated hips in association with hypoplastic ilia also are seen. Respiratory difficulties are worsened by hypoplasia of the cartilage of the tracheal tings. Roth et al. described in detail a case of lethal campomelic dysplasia (268). Multiple studies of the bent long bones led to the conclusion that the disorder was caused by an abnormality in the formation of the cartilage model of the long bones rather than a bony fracture, trauma, or pressure phenomena causing mechanical bending. The bone tis-

sue itself appeared to be normal and was synthesized in the appropriate patterns and positions in relation to the marked angular diaphyseal deformity. The epiphyseal cartilage was entirely within normal limits. In relation to the endochondral sequences at the epiphyseal growth plates, each of the stages of cellular maturation was seen. The epiphyses had cartilage canals but there was a clear delay in formation of the ossification centers of the distal femur and proximal tibia, both of which should be present at birth. The anterior and lateral portions of the ring of Ranvier were within normal limits, but posteriorly in the area of the concavity the ring was thicker than normal and the bone was reduplicated. Three clinical types are recognized, involving the longlimbed type, the short-limbed craniosynostotic type, and the short-limbed normocephalic type. It is those with the longlimbed variant that survive. In those few who survive, the bowing is less marked and tends to correct with time (60, 255). They are of markedly shortened stature and have considerable deformities of the cervical and thoracolumbar spine, generally necessitating spine fusion early in the first decade (60). The upper extremities are affected only minimally by the bowing. The hands, however, have many shaping abnormalities. The ossific nuclei of the distal femur, proximal tibia, and usually talus are not generally present at birth in this dysplasia.

J. Hypophosphatasia Hypophosphatasia is a rare but lethal skeletal dysplasia. It is autosomal recessive and characterized by small stature and poorly mineralized bone, particularly poorly and almost nonmineralized cranial bone. The long bones are extremely thin, hypoplastic, and fragile with a marked osteopenia. The ribs are short in association with a small thoracic cage. Death occurs secondary to respiratory insufficiency in the neonatal period, but many are stillborn often with markedly premature birth. The disorder is characterized by severe deficiency of bony tissue, markedly lower levels of serum alkaline phosphatase, and excessive urinary excretion of phosphoethanolamine. Ultrasonography, prenatal radiographs, and demonstration of markedly lowered alkaline phosphatase levels in cultured amniotic fluid cells have aided intrauterine diagnosis. The major differential diagnosis involves distinction from osteogenesis imperfecta type II.

K. Short Rib Syndromes The short rib syndromes, also referred to as "short ribpolydactyly syndromes," are potentially lethal particularly where thoracic constriction is prominent. Respiratory insufficiency and pulmonary hypoplasia contribute to death soon after birth in almost all cases. Three types are defined.

1. TYPEI: SALDINO-NOONAN The entity shows micromelia, postaxial polydactyly, brachydactyly, thoracic narrowing, and abnormalities of the

SECTION VIII 9 Microstructural-Morphologic Abnormalities of the Epiphyses and Metaphyses

cardiovascular system and genitalia. Radiographically there are short horizontal ribs, small iliac bones, metaphyseal spurs on the tubular bones, and deficiency in ossification of the distal extremities.

2. TYPEII: MAJEWSKI This variant has changes similar to those of the SaldinoNoonan form, but there is superadded facial clefting and nose deformities. The pelvis is radiographically normal and the tibia is more involved, being markedly shortened and ovoid.

3. TYPE IIl: NAUMOFF, LETHAL THORACICDYSPLASIA Radiographic features are similar to those of types I and II but thoracic constriction is severe.

L. Asphyxiating Thoracic Dystrophy (Jeune) On occasion patients with this disorder live, but most die particularly those with marked pulmonary hypoplasia and pancreatic and hepatic fibrosis. There are short ribs, horizontal acetabular roofs, and early appearance of the proximal femoral capital epiphysis secondary ossification centers. In those who survive, there is progressive improvement of lung function and often close to normal stature. Table VI lists the most prominent feature or features of these disorders, reference to which may be helpful in directing initial thoughts concerning diagnosis.

V I I I . MICROSTRUCTURAL-

MORPHOLOGIC ABNORMALITIES OF THE EPIPHYSES AND METAPHYSES IN SKELETAL DYSPLASIAS

A. General Considerations Light microscopic and ultrastructural studies of epiphyseal cartilage, growth plate cartilage, and bone in the skeletal dysplasias show many structural abnormalities, although these rarely are pathognomonic by themselves for any particular dysplasia. With the increasing discovery of molecular abnormalities, however, and with refinements of structural techniques such as immunocytochemistry and in situ hybridization at both the light and electron microscopic levels, it is expected that correlation between molecular abnormality and structural defect will play an increasingly large role in assessing the pathogenesis of deformity. One of the most common radiographic abnormalities in a large number of chondrodystrophies is delayed appearance of the secondary ossification center (Figs. 3Ai and 3Aii). We propose a regional and hierarchical approach to structural descriptions of abnormal epiphyses and metaphyses in the skeletal dysplasias (283). Although there have been a few attempts at coordinating descriptions of pathological find-

761

ings in the skeletal dysplasias, rarely have they been adhered to on a uniform basis. A systemic approach to tissue examination was provided by Yang et al. (373). There is a relatively limited number of ways in which structural abnormalities of an affected physis, entire epiphysis, and adjacent metaphysis can be manifested. The purpose of this approach is not to list in detail the structural abnormalities described in the various skeletal dysplasias but rather to outline the broad categories of gross and microscopic structural abnormalities that occur in the various dysplasias so that organized assessments can begin to link molecular abnormalities with eventual structural consequences. The categorization is outlined in Table VII.

B. Structure of Developing Epiphyses and Adjacent Metaphyses Including the Periphyseal Tissues of the Groove of Ranvier The epiphysis refers to all cartilage at the developing end of a long bone. This region is formed initially completely in cartilage and subsequently subdivides during development into three histologically distinct regions: (1) the cartilage immediately adjacent to the joint, referred to as articular cartilage; (2) the cartilage adjacent to the metaphysis, referred to variously as the growth plate, the epiphyseal growth plate, or the physis, is the functionally and cytologically specialized region in which the bulk of longitudinal growth occurs and encompasses the area from the reserve zone of cells to the end of the hypertrophic cell layer; and (3) the cartilage between the articular cartilage and the growth plate cartilage, referred to as the epiphyseal cartilage. Eventually this is transformed entirely into bone and marrow following the appearance and enlargement of what is variously referred to as the secondary ossification center, the bony nucleus, the ossific nucleus, or the bony epiphysis. The epiphysis is sometimes referred to as the chondroepiphysis, but use of this term should be restricted to the time prior to formation of the secondary ossification center. In all long and flat bones the two mechanisms of bone formation, endochondral and intramembranous, are present. The cells and tissues of these two relate intimately, specifically, and invariably to one another at the periphery of the growth plate in a region referred to as the perichondrial ossification groove of Ranvier. The tissues comprising the intramembranous ossification mechanism circumferentially ensheathe and support the physeal cartilage at this area. A circumferential groove is present whose deepest part is opposite the epiphyseal cartilage-physeal cartilage junction. It contains three tissue components: (1) an outer fibrous layer, which is continuous with the outer fibrous layer of the periosteum and inserts beyond the physeal region into the epiphyseal cartilage; (2) a zone of densely packed cells, which is a continuation of the inner cambial layer of the periosteum and is present into the depths of the groove as far as the resting zone of physeal cartilage (this collection of dense cells synthesizes osteoid and intramembranous bone directly); and

762

CHAPTER 9 9 Skeletal Dysplasias TABLE Vl

Radiographic Characteristics of Lethal Skeletal Dysplasias a

Thanatophofic dysplasia Osteogenesis imperfecta, lethal perinatal (Sillence II, OIC-A) Achondrogenesis

Campomelic dysplasia Hypophosphatasia Short fib syndromes Chondrodysplasia punctata (rhizomelic type) Atelosteogenesis

Short, c-shaped, "telephone-receiver" femurs; horizontal acetabulae; narrow sacrosciatic notch; underformed "H-shaped" vertebrae Short, broad crumpled femurs; womuan skull bones; thickened fibs; short tibia with 45-90 ~ diaphyseal angulation; extremely thin to absent cortices Disproportionately large head (approximately 40-50% of total body length) with extremely short limbs; severe underossification of axial skeleton, especially lower thoracic, lumbar, and sacral areas; pelvis, limb bones extremely short with little markedly abnormal shaping Bent diaphyseal regions of femurs and tibias Severe undermineralization of all bones with virtually no mineralization of calvafial bone; metaphyses are widened with marked radiolucency and cupping Thoracic narrowing with short horizontal fibs; polydactyly; micromelia; cardiovascular and multisystem abnormalities Punctate calcifications in epiphyseal and periepiphyseal regions; rhizomelic shortening Clubfoot, dislocated elbows, bowed lower extremities, abnormally shaped (oval, curved, bifurcated) and shortened humeri and femurs, coronal clefting of lumbar vertebral bodies; lower pelvis, calcanei often not ossified.

aFurther subcategorization of most of these syndromes is then possible.

(3) a collection of loosely packed cells between the outermost reaches of the zone of dense cells and the fibrous tissue layer that adds chondrocytes to the periphery of the epiphysis just beyond the physis itself. The intramembranous bone synthesized by the tissues within the groove region sometimes is referred to as the bony bark of Lacroix. Frequently it is discontinuous with the cortical bone of the diaphysis and metaphysis in those areas in which the metaphyseal cutback zone is extensive. The perichondrial ossification groove of Ranvier and its fibrous, chondroprogenitor, and osteoprogenitor cells are an integral part of the epiphyseal region.

C. Histopathologic Classification of Skeletal Dysplasias 1. SHAPING ABNORMALITIES OF THE EPIPHYSES The entire epiphysis including the articular surfaces can be abnormally shaped. The model of the epiphysis can be irregular from the embryonic period on, but in other disorders the articular surface and epiphyseal mass can be normally shaped at birth but become progressively deformed with time. An example of abnormal epiphyseal shaping dating from the embryonic period is epiphyseal dysplasia hemimelica or Trevor's disease, in which part of an epiphysis is markedly and focally enlarged (Fig. 3B). The disorder, likened by some to an epiphyseal osteochondroma, is seen most commonly at the distal tibia, distal femur, and distal radius. An example of a dysplasia in which the entire epiphyseal shape appears to be normal at birth and for the next few years prior to collapsing the articular surface is diastrophic dysplasia. The cumulative effects of abnormal cartilage synthesis, increased cartilage fibrosis, and continued weight bearing on

an imperfect structure lead to collapse of the articular surface and a lack of adjacent surface articular cartilage congruity (Fig. 3C). Many dysplasias have subtle articular surface shape irregularities, which are not of clinical significance in the childhood years but lead to adult-onset osteoarthritis, such as in multiple epiphyseal dysplasia. 2. SHAPING ABNORMALITIES OF THE PERIPHERY OF THE EPIPHYSEAL-METAPHYSEAL JUNCTION

The tissues surrounding the growth plate and upper metaphysis are well-organized into the perichondrial ossification groove of Ranvier. In hereditary multiple exostosis, differentiation abnormalities in this tissue complex lead to irregular sites of cartilage and bone synthesis causing widening of both the physeal and metaphyseal regions by the formation of cartilage outgrowths where normal periosteum should occur and by the absence of osteoclastic resorption (Fig. 3D). In other disorders, deposition of bone in the perichondrial ossification groove is greater than it should be, leading to a larger and more rigid perichondrial bony ring. This occurs in achondroplasia and also was shown in unnamed dysplasias (181) and appears to result from a dissociation of the two normal patterns of growth, because the intramembranous bone formation system that synthesizes the bony ring is functioning normally, whereas the adjacent endochondral bone formation system is less effective than normal. 3. SHAPING ABNORMALITIES OF THE METAPHYSES

The metaphyses are abnormally shaped in an entire subset of skeletal dysplasias referred to as the metaphyseal dysplasias or dysostoses. Because patients with this disorder are of short stature, there is physeal malfunction by definition but

SECTION Vlll ~ Microstructural-Morphologic Abnormalities of the Epiphyses and Metaphyses

F I G U R E 3 Radiographic and histopathologic abnormalities of the epiphyses and metaphyses that can occur in various skeletal dysplasia disorders are reviewed in this figure. (A) Delayed appearance of the secondary ossification center is characteristic of many of the skeletal dysplasias. It is particularly frequent in the proximal femur in spondyloepiphyseal dysplasia congenita in which it often is not seen as late as the end of the first decade. (Ai) Anteroposterior radiograph of hip in 7-year-old with SEDC showing no secondary center of femoral head. (Aii) Arthogram of hip shows dye (white) outlining round cartilaginous femoral head and cartilage of

763

F I G U R E 3 (continued) acetabulum. (B) A localized shaping abnormality of the epiphysis dating from the embryonic period is shown in this example of epiphyseal dysplasia hemimelica of the distal femur. Note the markedly enlarged medial femoral condyle. The disorder has been likened to an osteochondroma of the epiphysis. (C) Shaping abnormalities of the articular surface and subchondral bone region of the epiphysis are seen in many skeletal dysplasias. In virtually all of these disorders, the articular cartilage shape is normal in the newborn period, but imperfect structure due to abnormal cartilage synthesis and increased cartilage fibrosis leads to progressive collapse of the articular surface with weight bearing and the lack of adjacent surface articular cartilage congruity. This example of severe malformation is from a patient with diastrophic dysplasia. (D) The classic example of shaping abnormalities of the periphery of the epiphyseal-metaphyseal junction is that occurring in hereditary multiple exostosis. This radiograph shows characteristic widening of the metaphysis throughout the entire diameter of the distal femur plus prominent localized tissue outgrowths referred to as exostoses (arrows). (E) Shaping abnormalities of the metaphyses are seen in osteopetrosis due to abnormal resorption of bone. This disorder, which is mediated through imperfect osteoclast function, leaves a widened metaphysis due to failure of the normal resorptiveremodeling processes. (F) There can be changes in proportional contributions of varying regions of the growth plate in certain skeletal dysplasias. An example from a naturally occurring mutant cn/cn mouse is a marked diminution in the height of the hypertrophic cell region. In a normal mouse physis (at fight) each of the appropriate zones is represented, whereas in the mutant (at left) the germinal and proliferating zones are seen but there is an abrupt diminution of height in the hypertrophic zone adjacent to metaphyseal bone. [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.] (G) Epiphyseal-metaphyseal tissue from achondrogenesis shows epiphyseal cartilage above and abnormal metaphyseal bone below, but the intervening region shows significant failure of development of physeal structure. (H) Matrix abnormalities at the light microscopic level can involve abnormally high

F I G U R E 3 (continued) deposition of (Hi) fibrinous or fibrotic material or (Hii) preferential proteoglycan staining in the immediate pericellular matrix. When present the fibrotic deposits can be generalized or, as seen in this case of diastrophic dysplasia cartilage, present in localized foci. (I) Ultrastructural abnormalities of chondrocytes are characteristic of many skeletal dysplasias. The most common abnormality is a dilated rough endoplasmic reticulum, interpreted as a sign of imperfect protein synthesis and secretion. In spondyloepiphyseal dysplasia congenita, the dilated RER often appears to fill the entire cell segment being sectioned. In this disorder it is filled with a slightly radiodense material with a grainy appearance but otherwise without distinguishing features. [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.] (J) Matrix abnormalities at the ultrastructural level almost always involve increased matrix fibrosis. This tends to occur initially in the immediate pericellular area, as shown in this example from a patient with diastrophic dysplasia. At higher power there is characteristic appearance of markedly widened and irregularly adherent fibrous bundles with characteristic cross-banding of collagen. [Parts Hii and J reprinted from Shapiro (1992), Calcif. Tiss. Int. 51:324-331, copyright notice of Springer Verlag, with permission.]

766

CHAPTER 9 9

Skeletal Dysplasias

TABLE VII Outline of Structural Morphologic Abnormalities of t h e Epiphyses in Skeletal Dysplasias: A Regional-Hierarchical Approach a

A. Shaping abnormalities 1. Of epiphyses Embryonic-fetal (primary) Juvenile to adult (secondary) 2. Of the periphery of the epiphyseal-metaphyseal junction 3. Of the metaphyses B. Physeal abnormalities 4. Alterations in growth plate thickness Thin Thick 5. Change in proportional contribution of varying regions of the growth plate 6. Disorganization of the cytoarchitecture of growth plate 7. Premature closure of the growth plate C. Epiphyseal and physeal abnormalities 8. Change in cellularity of epiphyseal and physeal cartilage Hypocellular Hypercellular 9. Altered vascularity (cartilage canals) of the cartilaginous epiphyses Hypervascular Hypovascular 10. Altered appearance of the secondary ossification center Timing Uniformity of appearance 11. Stippling of the epiphyses D. Abnormal cell and matrix appearances 12. Abnormal cell appearance at light microscopic level 13. Matrix abnormalities at the light microscopic level 14. Ultrastructural cell abnormalities 15. Ultrastructural matrix abnormalities 16. Abnormalities of the epiphyseal-metaphyseal junction aShapiro F (1998) Structural abnormalities of the epiphyses. In: Skeletal Growth and Development: Clinical Issues and Basic Science Advances. Rosemont, IL: American Academy of Orthopedic Surgeons.

the structural manifestation of the problem comes at the metaphyseal level. Widening of the metaphyses is also seen in osteopetrosis in which the normal resorptive, funnelization process at the lower parts of the perichondrial groove complex is prevented by absent or nonfunctioning osteoclasts (Fig. 3E). 4. VARIATIONS IN GROWTH PLATE THICKNESS

The growth plate can be the same thickness as normal, thicker than normal, or thinner than normal. In many skeletal dysplasias, there is no change in the thickness, as determined by qualitative assessments and on occasion by quantitative assessment. In those abnormalities associated with a defect

of mineralization, the growth plate is thicker than normal with persistence of cartilage in the hypertrophic zone and adjacent metaphysis in which mineralization normally occurs. Thickened growth plates are the hallmark of tickets disorders, regardless of whether these are nutritional in nature or due to genetic disorders such as hypophosphatemic tickets. The physis can also be thicker in metaphyseal dysplasias, although they tend to be so in an irregular fashion with isolated areas of radiolucency in metaphyseal bone adjacent to the physis. The majority of skeletal dysplasias that show changes in physeal thickness have thinner growth plates. This is particularly true in the lethal chondrodystrophies in which cartilage tissue itself is markedly poor in formation and in diastrophic dysplasia toward the end of the first decade when premature growth plate closure further worsens prognosis. 5. CHANGE IN PROPORTIONAL CONTRIBUTION OF VARYING REGIONS OF THE GROWTH PLATE

In addition to whether the growth plate is of normal thickness, thinner, or thicker, it is important to assess the relative contributions of the various cytologically specialized areas of the plate to its thickness. In certain disorders, there is an apparent normal progression from resting and germinal zones through the proliferating zone, but then a marked diminution in the height of the hypertrophic cell region. This is seen in particular in the cn/cn mouse, which is a short stature, naturally occurring mutant (Fig. 3F) (334). In tickets disorders, on the other hand, the hypertrophic zone is markedly increased in height due to the failure of mineralization of the cartilage cores, which normally occurs as part of the endochondral developmental sequence. The unmineralized cartilage fails to trigger the normal resorptive response and persists, increasing the thickness of the physis. 6. DISORGANIZATION OF THE CYTOARCHITECTURE OF THE GROWTH PLATE

One of the most common findings in skeletal dysplasias is a disorganization in the cellular organization of the growth plate. It is important to recognize that columnation in the human is rarely as well-organized as that demonstrated in animal studies from the rabbit, mouse, and rat. In some instances, therefore, cytologic human specimens reported as abnormal really vary little from samples in nonaffected individuals. Disorganization, once present, however, generally tends to be throughout the growth plate beginning in the resting and germinal zones and involving the proliferating as well as the hypertrophic zones. This is not surprising because growth plate cytologic structure represents a coordinated series of phenomena that, if thrown off early in the sequence, demonstrates little ability to rearrange itself into a normal pattern. The disorganization is characterized by a tendency that in the very mild disorders is an irregularity in columnar formation and with major involvement simply leads to the presence of cartilage with essentially no organi-

SECTION VIII ~ Microstructural-Morphologic Abnormalities o f the Epiphyses and Metaphyses

zation at all (Fig. 3G). When the finely organized hierarchical system of the endochondral sequence is disorganized, major alterations in the chain of events can follow. These generally lead to such associated findings as premature vascular invasion of the physeal cartilage with aberrant bone deposition, vascular invasion of the physis associated with fibrosis, fibrosis with very little vascular invasion, and a disorganized metaphyseal region. 7. CHANGE IN CELLULARITYOF EPIPHYSEAL AND PHYSEAL CARTILAGE Some skeletal dysplasias are characterized by increased cellularity of the epiphyseal, including physeal, regions, and some are characterized by a hypocellular cartilage. Achondrogenesis is an example of hypercellular epiphyseal cartilage due to the fact that matrix synthesis is scanty because the defect in the disorder is virtually complete nonexpression of type II collagen synthesis. In most other dysplasias with a chondrocyte cell number abnormality, the hypocellular features predominate. These findings occur when chondrocyte cell proliferation and differentiation are diminished or when the cells die prematurely and are replaced by fibrous tissue. 8. ALTERED VASCULARITY(CARTILAGE CANALS) OF THE CARTILAGINOUSEPIPHYSES The epiphyses contain vessels present in cartilage canals. In the human these are seen initially in the distal femur as early as the third month in utero. The vessels are widely considered to actively invade the epiphysis from the perichondrium. They provide nutrition to the chondrocytes and are present several months before formation of the secondary ossification center. Little information is available on cartilage canals in abnormal epiphyseal development. In achondrogenesis, however, the vascularity is increased in association with the high fibrous content of the epiphyses. It would be interesting to knowwhether delayed appearance of the secondary ossification centers, which characterizes some skeletal dysplasias such as spondyloepiphyseal dysplasia congenita, is or is not accompanied by alterations in the cartilage canals. 9. ALTERED APPEARANCE OF THE SECONDARY

OSSIFICATION CENTER The time of appearance of the secondary ossification center in the skeletal dysplasias can be normal, premature, or delayed. Premature appearance of secondary ossification centers is rare but is seen sometimes with osteogenesis imperfecta, presumably in relation to microtrauma and the increased vascularity of the epiphyseal-metaphyseal region that ensues due to intrauterine or early postnatal fractures and in the short rib-polydactyly syndrome. A more common occurrence is a delay in appearance of the secondary center, which is sometimes the earliest warning of some underlying skeletal dysplasia. The delayed appearance of the secondary ossification center in spondyloepiphyseal dysplasia congenita is particularly noteworthy in the proximal femurs and is

767

often associated with a coxa vara deformity. In those instances in which it is delayed, the most common pattern seen, there is also a frequent occurrence of multiple discrete regions of calcification and bone formation when it does finally form rather than a uniform appearing centrally placed and gradually enlarged radiodensity, as is normally the case. Fragmentation of the secondary center once it appears is also known to occur and is often seen in the femoral head in multiple epiphyseal dysplasia. The fragmentation can lead to confusion with Legg-Perthes disease. 10. STIPPLING OF THE EPIPHYSES Radiodense calcification can be seen throughout the epiphyseal region in a set of disorders referred to as stippled epiphyses (Fig. 2D). The calcification is not necessarily concentrated in what would normally be the position of appearance of the secondary ossification center but can occur throughout the epiphyseal and physeal cartilage. There are several causes of stippled epiphyses and considerable variations in prognosis so that accurate syndromal diagnosis is extremely important (247). Some are present in the lethal chondrodysplasia punctata characterized by marked rhizomelic limb shortening. In the benign form, referred to as Conradi-Htinermann chondrodysplasia punctata, patients survive and these stippled regions are eventually incorporated into bone as it forms and do not lead to any negative longterm sequelae. In addition, stippled epiphyses can occur as an epiphenomenon with few or no negative sequelae in several skeletal disorders, including multiple epiphyseal dysplasia.

11. ABNORMAL CELL APPEARANCE AT THE LIGHT MICROSCOPIC LEVEL Abnormal chondrocyte appearance at the light microscopic level can involve such findings as increased inclusion materials, an increased number of cells, which is generally reflective of a decreased amount of matrix synthesis, and a decreased number of cells, which is generally reflective of the occurrence of premature cell death. In association with the latter is the tendency for the chondrocytes to be smaller and stain more densely with subsequent premature death and replacement by fibrosis. In chondrocyte areas in which the cell has died and been resorbed, the persistent empty lacuna is frequently referred to as a ghost cell. On occasion, circumferential tings of fibrous tissue are deposited in the empty lacuna and pericellular matrix and the degenerated cell is referred to as a target cell. A characteristic cell inclusion in epiphyseal chondrocytes is fat, as demonstrated on tissue sections prepared by plastic embedding and stained by toluidine blue for light microscopy and on tissue examined by transmission electron microscopy. The fat is intracytoplasmic, always assumes a circular globular shape, and stains a green-blue or turquoise with the toluidine blue. Other intracytoplasmic structures stain a light blue and generally assume a less discrete shape. It is important not to interpret epiphyseal chondrocyte fat globules as pathological,

768

CHAPTER 9 ~ Skeletal Dysplasias

although we have not seen them in physeal chondrocytes. Other intracytoplasmic inclusions in either epiphyseal or physeal chondrocytes may be pathological indicators. Histochemical analyses were performed previously by investigators in the field, but reproducibility of results was difficult. Newer immunocytochemical techniques combined with greater knowledge of normal and abnormal molecular components of cartilage should improve light microscopic identification. 12. MATRIX ABNORMALITIES AT THE LIGHT

MICROSCOPIC LEVEL The uniform appearing and uniform staining hyaline cartilage matrix can show abnormalities, which involve (1) replacement by fibrotic material or (2) preferential proteoglycan staining in the immediate pericellular matrix (Figs. 3Hi and 3Hii). Pathologic fibrotic deposits can be present in abnormal localized loci or can be generalized throughout. Fibrotic collections stain a deeper red with hematoxylin and eosin and a deeper blue with toluidine blue and green, instead of a uniform red, with Safranin O-fast green. The presence of pericellular staining for proteoglycans with less prominent staining in interterritorial areas reflects efforts at matrix synthesis, which are continuing but are insufficient to maintain normal cartilage formation throughout the entire mass of articular, epiphyseal, and physeal cartilages. 13. ULTRASTRUCTURALCELL ABNORMALITIESOF THE EPIPHYSIS AND PHYSIS

The disordered chondrocyte as shown by light microscopy will frequently manifest more specific structural abnormalities at the higher power ultrastructural level (Fig. 3I). These can be characterized by the abnormal appearance of organelles or inclusion in the chondrocyte cytoplasm or by premature cell necrosis. Many early reports on ultrastructural abnormalities in the chondrodysplasias commented on and demonstrated increased intracellular fat; it is now recognized, however, that intracytoplasmic fat is a normal finding in many epiphyseal chondrocytes, which is important in interpretation of the findings. Fat globules within the endoplasmic reticulum, however, are abnormal. The most classic occurrence within the chondrocyte cytoplasm at the ultrastructural level in the skeletal dysplasias is the increase in dilated cisternae of rough endoplasmic reticulum. Indeed, in some cells, one can see only one single dilated cisterna filled with moderately electron dense material in a thin section profile. This is interpreted as an indication of the fact that the cell is synthesizing protein, but, presumably because of its abnormal structure, the normal secretion mechanism fails and the protein persists within the cell. Findings such as this are particularly characteristic of spondyloepiphyseal dysplasia congenita. They are, however, not pathognomic for it and are seen in many other disorders. A laminated appearance of alternating electron dense and electron lucent lines is seen within the dilated rough endoplasmic reticulum in pseudo-

achondroplasia and in some variants of multiple epiphyseal dysplasia. 14. MATRIX ABNORMALITIES AT THE

ULTRASTRUCTURALLEVEL The ultrastructural appearance of the normal cartilage matrix is well-known. In many skeletal dysplasias, particularly diastrophic dysplasia, there are increased loci of fibrous tissue within the normally cartilaginous matrix (Fig. 3J). These fibrils characteristically are cross-banded but are quite irregular and markedly thicker both in terms of diameter and in their appearance on longitudinal sectioning. They represent the aggregation of normally dispersed and much thinner type II collagen molecules due to a deficiency of type IX collagen, which normally coats the type II fibril and prevents its aggregation with adjacent fibrils. The initial foci of increased fibrosis are usually in the immediate pericellular area. 15. ABNORMALITIES OF THE EPIPHYSEALMETAPHYSEAL JUNCTION

In the normal endochondral sequence, the lowest regions of the cartilage of the epiphyseal growth plate in the hypertrophic zone become mineralized. Vascular invasion from the metaphysis then occurs, with vessels accompanied by undifferentiated mesenchymal cells passing into the lowermost two or three hypertrophic cell lacunae and coincidentally laying down new bone on the calcified cores of cartilage. The metaphyseal region thus is characterized by the presence of these columns of tissue composed of the central mineralized cartilage cores and the surrounding deposits of the new bone. At the outermost reaches of the metaphysis immediately adjacent to the hypertrophic layer, the columns are composed almost exclusively of calcified cartilage but as one passes deeper into the metaphysis, there is relatively less cartilage and relatively more new bone. The region is also characterized by active resorption both of the calcified cartilage cores and of the recently deposited bone with replacement by trabeculae composed exclusively of bone. The resorption is mediated by the multinucleated osteoclasts. Because there is considerable disorganization of the endochondral sequence in many of the skeletal dysplasias, this disorganization is reflected not only in the growth plate itself, but particularly in the adjacent metaphysis because the new bone is normally deposited on the calcified core of cartilage. If the cartilage is either not deposited, not appropriately mineralized, or deposited in an irregular arrangement, then new bone synthesis in the metaphyseal region of necessity must be disordered itself because the organized linear scaffold effect of the mineralized cartilage is imperfect. The problem can also lie in a failure of appropriate resorption of bone and cartilage by osteoclasts. In some skeletal dysplasias the physeal-metaphyseal interface is irregular rather than being relatively smooth because mineralization, vascularization, and bone formation proceed at different times and levels in response to the disordered phy-

SECTION IX ~ Histopathologic Changes in Specific Chondrodysplasias

seal tissues. In osteopetrosis the metaphyseal regions are both widened and radiographically dense due to failure of normal resorption. The widening is due to persistence of physeal synthesized bone because the osteoclasts at the lower region of the groove of Ranvier are unable to function normally at the so-called "cutback" zone of remodeling.

D. Interpretation of the Pathogenesis of Skeletal Dysplasias in Relation to the Structural Approach Understanding of the skeletal dysplasias was based initially on clinical appearances. The next era of definition was based on radiographic appearances with many disorders defined by the region of prominent radiographic abnormality such as multiple epiphyseal dysplasia or metaphyseal dysostosis. With the recognition of proteoglycan abnormalities underlying many variants of what radiographically are spondyloepiphyseal dysplasias, the focus of investigation has shifted to biochemical and more recently to molecular and genetic abnormalities. Many reports on histologic and ultrastructural aspects of various skeletal dysplasias have appeared, but the findings have rarely if ever been specific to a particular entity. Over the past several years investigation almost exclusively has centered on molecular and genetic abnormalities underlying these disorders. We propose, however, that there is great value in assessing how genetic, molecular, and biochemical abnormalities translate via micro to macro epiphyseal, physeal, and metaphyseal structural changes into the gross alterations seen clinically and by radiographic projection. Rubin in his classic test attempted, with reasonable success, to relate the pathogenesis of the various dysplasias to the regions of the epiphyseal-metaphyseal ends of the bone as defined radiographically (271). With the increasing sophistication of microstructural assessments the cell and matrix functioning in both normal and abnormal epiphyses is becoming better known. The array of structural techniques includes light microscopy, polarizing light microscopy to assess fiber orientation, histochemistry to assess matrix composition, autoradiography, immunocytochemistry, and in situ hybridization at both light and transmission electron microscopy levels; and scanning and transmission electron microscopy. The approach listed does not mean to imply that the findings are isolated phenomena. They represent clear structural abnormalities but almost always three or four or more of the findings are present in any one disorder. Development, be it normal or abnormal, is a series of events in which early developments define and direct those occurrences that follow them. The eventual structural defects will reflect not only the molecular abnormality causing them but also (1) the time of function of the molecule in the developmental cascade, (2) the extent of the molecular abnormality, and (3) the presence or absence of other molecules that could partially or totally replace the functions of the disordered molecules. It

769

now appears meaningful to re-incorporate microstructural morphologic studies into the assessment of skeletal dysplasia disorders. Morphogenesis is an ordered hierarchical series of events in which an abnormality can set up an altered cascade predisposing to even greater deviations from the norm. The regional hierarchical approach to assessing tissue structure in the skeletal dysplasias can bridge the gap between definition of genetic-molecular-biochemical events and the ultimate clinical and radiographic deformities that define the phenotype in the various disorders. In the following sections, we will review many of the reported histopathological findings in the skeletal dysplasias.

IX. H I S T O P A T H O L O G I C C H A N G E S IN SPECIFIC CHONDRODYSPLASIAS Histopathologic changes in the skeletal dysplasias have been reviewed in several articles (89, 223, 263, 265, 294, 311) as well as in individual case reports.

A. Lethal Chondrodysplasias Many of the abnormal findings have been reviewed by Gilbert et al. (89).

1. ACHONDROGENESISI (PARENTI-FRACARO) All zones of the growth plate are densely hypercellular. The normal chondrocyte progression is lost. Calcification of cartilage and trabecular bone formation are disorderly with irregular capillary penetration and little column formation. The cartilage canals within epiphyseal cartilage are markedly increased. There is diminished width of the longitudinal septae between the hypertrophic chondrocytes. The physeal growth zone is extremely retarded developmentally and disorganized. Resting chondrocytes are slightly enlarged and frequently contain characteristic large PAS positive cytoplasmic inclusions, which are spherical to oval in shape. By EM, the vesicles correspond to markedly dilated cisterns of rough endoplasmic reticulum (344).

2. ACHONDROGENESIS II (LANGER-SALDINO) There is a deficiency of cartilage matrix, markedly enlarged cell lacunae, and chondrocytes with abundant clear cytoplasm (71). The entire cartilage region is hypercellular due to a scanty matrix and the physeal growth zone is markedly retarded and disorganized. No growth plate columnation is seen. The epiphyseal cartilage is lobulated and mushroomed with increased vascularity. There is irregular vascular invasion of the hypercellular cartilage (Figs. 4 A 4D). The primary bone trabecular formation is quite abnormal and decreased to the extent that trabeculae are oriented horizontally. There is overgrowth of peripheral membranous bone leading to cupping of the epiphyseal cartilage. Long bones and ribs have grossly thinner physeal growth zones

770

CHAPTER 9 9 Skeletal Dysplasias

F I G U R E 4 Histopathologic findings in Langer-Saldino achondrogenesis are shown. (A) Photomicrograph shows hypercellular cartilage with fibrovascular invasion. The cartilage canals are increased in achondrogenesis in association with a high fibrous content of the epiphyses. The absence of normal type II collagen and associated proteoglycan leads to a relative abundance of type I collagen, which might encourage vascular invasion. [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.] (B) At higher power the cartilage canal centrally is evident. The numerous cells are surrounded by only a scanty matrix. (C) In many areas, regions of chondrocytes with matrix and fibrovascular tissues have a whorled conformation. [Reprinted from Eyre et al. (1986), Am. J. Hum. Genet. 39:52-67, 9 University of Chicago, with permission.] (D) The markedly abnormal and almost absent physis is illustrated here. The epiphyseal cartilage above appears to merge with the metaphyseal bone below with no intermediation of normal appearing physeal tissue. The metaphyseal bone is synthesized around regions of cartilage but appears to occur more in association with the fibrovascular invasion of the physeal region rather than due to any sequence of cartilage changes. [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.]

SECTION IX ~ Histopathologic Changes in Specific Chondrodysplasias

than normal and are fragile and brittle with the cartilage portion separating readily from the underlying metaphysis. Microscopy of the proximal humerus confirmed marked disorganization of the growth zone, with cupping of subchondral bone and lipping at the periphery. The normal parallel configuration of the hypertrophic and proliferative columns was absent. Short transitional columns were seen occasionally but were of haphazard orientation, often interspersed with primitive mineralizing cartilage. Primary osteoid seams were present in some places adjacent to poorly formed and shortened transitional columns. The resting cartilage in all sites was markedly abnormal, being abundantly cellular and composed of large and ballooned chondrocytes with little intercellular matrix. It was arranged in large nodular masses interspersed with abnormally abundant blood vessels. Membrane bone and bone marrow appeared to be normal. a. Electron Microscopy of Cartilage. In cross sections of cartilaginous rib, two zones were apparent, an outer opaque, more fibrous zone and an inner translucent, gelatinous zone. Sections of humeral head cartilage showed similar cartilaginous whorls and interspersed fibrous septa. Electron microscopy of the chondrocytes in the gelatinous region showed abnormal accumulations of amorphous material in dilated lakes of rough endoplasmic reticulum, whereas the extracellular matrix showed a random feltwork of well-spaced, uniformly thin (15-20 mm diameter) fibrils typical of young, growing hyaline cartilages. Perichondral regions, however, showed the parallel arrays of densely packed, much thicker fibrils that are more typical of tissues rich in type I collagen. Although the appearance of the collagen in the gelatinous central regions of cartilage gave the impression of young type II collagen, the biochemical results clearly showed that type II was absent and the bulk of the collagen was type I. b. Cartilage Chemistry. At all sites, hyaline cartilage had a watery, gelatinous texture and translucent appearance in contrast to the stiff, opaque properties of control neonatal cartilages. The collagen content of the gelatinous cartilage core from ribs and scapula was only half that of control human neonatal rib cartilage. The hexosamine content was in the high range expected for hyaline cartilage, and the high ratio of galactosamine to glycocyamine was typical for proteoglycans of hyaline cartilage and similar to that of control human neonatal rib cartilage. The hydroxylysine content of the cartilage relative to hydroxyproline was significantly lower than that of neonatal control cartilage, consistent with type I collagen rather than type II collagen. The concentration of mature, hydroxypyridinium cross-links in the total cartilage collagen was only 15% that of a control neonate. Because type II collagen is the richest known source of these stable cross-linking residues, this also was consistent with the other biochemical findings. Electrophoresis of the total pepsin-solubilized collagen from gelatinous scapula cartilage showed a pattern typical of type I collagen rather than the type II collagen that dominates in normal neonatal human cartilage and in all mammalian

771

hyaline cartilages. The more fibrous peripheral regions of the scapula and rib cartilages showed a higher proportion of what appeared to be a 1(V) and oL2(V) chains, but again with type I collagen chains as the main molecular species. This tissue region by light microscopy appears to be a fibrous ingrowth from the perichondrium populated more by flattened fibroblastic cells than by rounded chondrocytes. Electrophoretic analyses of CNBr digests of the core cartilage also showed a peptide pattern typical of type I collagen. No type II collagen peptides were recovered in detectable amounts from any region of cartilage, including the gelatinous cores of rib, scapula, and humeral head. The resting cartilage had little extracellular matrix surrounding the large, dilated chondrocytes. The low collagen content was consistent with this observation and appeared to result from a failed expression of type II collagen by the chondrocytes. Type I collagen, which was the main type of collagen present in all cartilage samples including the gelatinous rib core, appears, in part, to replace type II collagen in the sparse matrix. Collagen fibrils in the central, gelatinous region of the rib had the characteristic appearance of those in young cartilage matrix with thin, uniform diameters and a well-spaced random organization, despite type I collagen predominating instead of type II. This finding has important implications for the mechanism of regulation of fibril diameters and fibril organization because it suggests that type I collagen produced in the presence of otherwise cartilage-specific macromolecules assumes a polymeric form and architecture typical of cartilage. It implies that the genetic type of fibrillar collagen in itself has little or no control over fibril diameter and organization. It seems likely, therefore, that components of the cartilage extracellular matrix other than the principal collagen type, such as type IX, are primarily responsible for regulating collagen fibril size and architecture. The findings imply that achondrogenesis is based on a genetic defect that prevents type II collagen expression. 3. HYPOCHONDROGENESIS

Histopathologic changes are similar but somewhat milder than those in achondrogenesis II. Matrix deficiency and large ballooned chondrocytes are seen, but there is rudimentary growth plate formation with hypertrophy of cells and some sense of columnation. True endochondral bone formation rarely occurs (344). 4. THANATOPHORIC DYSPLASIA Thanatophoric dysplasia is the most common lethal skeletal dysplasia. There is relatively normal resting cartilage of the physis with subsequent generalized disruption of the physeal endochondral sequence, with no orderly progression of cell change, no attempts at column formation, and an irregular hypertrophic zone. There are some areas of hypertrophic cellular alignment. Vascular invasion of the growth plate occurs at irregular intervals, leading to an irregular

772

CHAPTER 9 ~ Skeletal Dysplasias

array of short, blunt, and broad spicules of calcified cartilage and bone and often with metaphyseal trabeculae aligned horizontally.

gions, although there is some hypertrophy of cells that are arrayed irregularly immediately adjacent to the metaphysis. Premature epiphyseal ossification occurs.

5. HOMOZYGOUS ACHONDROPLASIA There is considerable similarity both clinically, radiographically, and histopathologically between thanatophoric dysplasia and homozygous achondroplasia. The reasons for this are now much more clear because both represent abnormalities of the FGFR3 molecule. Stanescu et al. noted, upon histologic examination of a child who died at 2 months of age, that the growth zone of the upper tibial cartilage was very narrow with the chondrocytes irregularly arranged without column formation (315). In most areas chondrocyte hypertrophy was not seen, although occasional areas of clustered hypertrophic cells were noted. The primary cartilagebone trabeculae were infrequent, short, and irregularly arranged with some of them actually being horizontal to the long axis. In some areas the growth cartilage was replaced by islands of fibrous or fibrocartilaginous tissue. The periphyseal groove of Ranvier, similar to what is seen in achondroplasia, was more prominent than normal and prolonged toward the center of the growth zone by a band of young fibrous tissue bordered by a narrow bone rim. In contrast to the abnormal and scanty physeal cartilage, the epiphyseal cartilage had an apparently normal cell density and distribution with normal cartilage vascular canals. Electron microscopy showed that the chondrocytes both in the physis and in the epiphyseal cartilage were active with a well-developed rough endoplasmic reticulum, Golgi systems, and many mitochondria. Degenerating cells were rare. The matrix of the epiphyseal cartilage appeared to be relatively normal, whereas that of the physis had many thickened collagen fibers irregularly arranged. The pericellular matrix of chondrocytes from the physis showed accumulation of dense, relatively thick, cross-banded collagen fibers with diminished proteoglycan granules, a finding similar to that seen in many skeletal dysplasias.

7. CHONDRODYSPLASIAPUNCTATA This disorder, more commonly known as stippled epiphysis, encompasses several variants, not all of which are lethal. The punctata appearance is an epiphenomenon rather than a primary defect. The lethal disease variant is rhizomelic chondroplasia punctata, but usually nonlethal variants include Conradi-Htinermann chondrodysplasia punctata and sex-linked chondrodysplasia punctata. Stippled epiphyses are also seen with cases of multiple epiphyseal dysplasia and some of the mucopolysaccharidoses. a. Rhizomelic Chondrodysplasia P u n c t a t a . Most of these patients die in infancy. There is increased vascularization of the reserve zone of cartilage with associated fibrosis, dysplastic calcification, and microcystic degeneration. There are decreased numbers of proliferative chondrocytes and endochondral ossification is abnormal. There is decreased vascular invasion at the lower regions and tongues of cartilage extend into the metaphysis. Jeune et al. reviewed the histopathology from 11 autopsy descriptions in the literature with markedly similar findings. The physeal proliferating and hypertrophic columns were short and irregular and sometimes were grouped only in randomly nonoriented masses. The line of ossification was fragmented and irregular with the endochondral ossification retarded and disorganized, whereas periosteal ossification was normal. The physeal cartilage was characterized by fibrinous, mucoid necrosis with fibrovascular invasion. Calcifications were seen throughout the cartilage from articular to physeal regions. These were often present in relation to vessels but did not represent centers of bone formation. Studies from varying stages indicated initial calcification in cartilage alone followed by calcific deposits among reactive giant cells and eventually deposits in relation to fibrovascular tissue. The latter finding was considered to be compatible with eventual resorption of the calcific deposit if the patient survived.

6. SHORT RIB-POLYDACTYLY SYNDROMES SRPD1 (Saldino-Noonan). There is disorganized proliferative and hypertrophic cartilage of the physis with decreased numbers of chondrocytes. The zone of chondroosseous transformation is irregular. Columnation is disorganized and the formation of trabecular bone is abnormal with short, stumpy trabeculae. Often there are central areas of fibrosis in the physis and premature ossification centers of the proximal humerus and proximal femur. By electron microscopy, the physeal matrix is abnormal with fibril disorganization, abnormal fibrils with tapered ends, and abnormal amounts of mature collagen. b. SRPD2 (Majewski). The physeal growth zone is markedly retarded and disorganized. There is only minimal endochondral bone formation. There is no specific delineation between resting, proliferating, and hypertrophic zone rea.

8. CAMPOMELIC DYSPLASIA

a. Long-Limbed Campomelic Dysplasia. Growth plate disorganization has been relatively mild with a normal progression of cells. Any resting chondrocytes appear vacuolated.

b. Short.Limbed Campomelic Dysplasia (Normocephalic Type). c. Short.Limbed Campometic Dysplasia with Craniosynostosis. There have been histopathologic descriptions in the latter two types. 9. FIBROCHoNDRoGENESIS Histopathology of growth regions shows no columnar arrangement of either proliferating or hypertrophic cells. Metaphyseal bone is immediately adjacent to virtually undifferentiated cartilage. There is some calcified cartilage sur-

SECTION IX ~ Histopathologic Changes in Specific Chondrodysplasias

rounded by bone in the metaphyseal region but certainly no orderly array. Resting cartilage is characterized by chondrocytes in a lacy woven fibrillar network. There are numerous fibrous connective tissue septa forming a lattice appearance with small chondrocytes within surrounded by very little matrix. The chondrocytes are often elongated and flattened. There is no regular growth plate but some evidence of cell hypertrophy toward the metaphyseal areas. There is some provisional calcification of the matrix but irregular vascular invasion. 10. ASPHYXIATING THORACIC DYSPLASIA In type I the physis shows a patchy distribution of endochondral ossification with only an irregular cartilage-bone junction. In type 2, a physeal region can be identified but it is shortened and irregular, in particular showing little column formation. Histologic changes are variable. They include reduction in the number of proliferating hypertrophic chondrocytes, irregular vascular invasion, cartilage islands in the metaphysis, and transverse bony plates extending across the metaphysis. By ultrastructure an extensive number of lipid inclusions in chondrocytes have been reported. 11. HYPOPHOSPHATASIA There is marked underformation of bone throughout the skeleton in this lethal condition. Physeal areas, however, show normal resting cartilage, although the chondrocytes are somewhat disorganized. The underlying problem relates to failure of calcification and new bone formation. Although much material in particular from the lethal chondrodysplasias comes from autopsy long bone specimens, samples from living individuals have generally been obtained from iliac crest or costochondral rib junction biopsies. Although the latter clearly show features of the endochondral sequence, they are atypical in the sense that they do not fully represent the structural appearance of long bone epiphyses. Many histologic studies done prior to the era of more rigorous classification also may be inappropriately titled based on our current knowledge. This pertains particularly to studies performed prior to 1950 when the terminology was not well-defined and use of such diagnoses as achondroplasia for many skeletal dysplasias was common.

B. Histopathologic Changes in Nonlethal Chondrodysplasias Histologic studies of normal human iliac crest cartilage serve as an important baseline for the study of pathologic tissue. A report by Ponseti et al. is very helpful in this regard (244). 1. ACHONDROPLASIA Despite the marked short stature, the growth plate is structurally regular with a well-organized endochondral sequence. Column formation appears to be normal as does cellularity. Periosteal bone formation at the perichondral

773

groove of Ranvier is more extensive than normal. Ultrastructurally, the cartilage contains relatively normal chondrocytes and matrix (197). In a study by Rimoin et al., biopsies of rib and iliac crest from 10 individuals with typical achondroplasia ranging in age from 10 months to 35 years have shown that the endochondral sequence is quite regular and well-organized with normal cell columnation and an unremarkable appearance of the matrix (264). He stressed that the ribs were short even though the rib physeal cartilage was structurally unremarkable. One common finding, however, is the relatively more extensive periosteal ossification deep into the perichondrial groove of Ranvier. This led to a relatively cup-shaped appearance at the ends of some long bones and the comparatively short (endochondral effect) and widened (normal intramembranous bone effect) appearance of the long or tubular bones. This was also noted by Keith as early as 1920 (144). It is the endochondral growth sequence that is affected in achondroplasia, whereas the intramembranous bone sequence is not affected. Ponseti also found the iliac crest growth plate to be close to normal in histologic sections (243). When he assessed fibular growth plates, however, some structural abnormalities of the physis were noted. Sections of the proximal fibular epiphyseal regions were present from 7 patients. The plates were V-shaped toward the metaphyseal region, a finding also noted radiographically in larger epiphyses especially the distal femur. The periosteal intramembranous bone was normal overall, and Ponseti also noted that the periosteal bone of the periphyseal groove of Ranvier was abundant and wider and deeper than normal around the adjacent physis. There was more bone formation by the physis of the fibula at its periphery than centrally, and the central regions of the physis had relatively little and markedly thinned cartilage. The physis was somewhat irregular, with the proliferating cells tending to be in clusters rather than in columns and to be separated by wide septae of cartilage matrix, which was more fibrous than normal. The hypertrophic zone was more narrow and irregular than normal, and the normal maturation of the hypertrophic cell was disturbed. The vascular invasion of the growth plate at the hypertrophic zone was very uneven, such that the vascular front "appeared to face an insurmountable barrier of fibrotic cartilage and bone." The central half of the fibular plate was thin and wavy, although that specimen was from a girl 13 years 4 months of age. The secondary ossification centers formed normally. Photomicrographs showed that the structure of the achondroplastic physis, though not perfectly normal, deviates relatively minimally from normal. 2. HYPOCHoNDROPLASIA There is a normal growth plate with regular, well-organized endochondral ossification sequences. Even by ultrastructural analysis the growth plate matrix and cells are normal. Overgrowth of circumferential periosteal bone in the periphyseal groove of Ranvier is not a feature of this disorder.

774

CHAPTER 9 9 Skeletal Dysplasias

3. DIASTROPHIC DYSPLASIA

Light and electron microscopic studies of diastrophic dysplasia iliac crest growth cartilage reveal extensive cell and matrix abnormalities at each time period (Figs. 3J and 5A5D) (66, 281,294, 317). Light microscopy shows atypical chondrocytes with extreme variation in size and shape, premature cytoplasmic degeneration, and target ghost cell formation. Prominent, densely staining fibrotic loci are present throughout the cartilage. Ultrastructure reveals some structurally intact chondrocytes with a single, large fat inclusion, slightly dilated rough endoplasmic reticulum, and abundant glycogen. As early as 1 year of age, cystic degeneration of chondrocyte cytoplasm is evident with indistinct organelles seen. The cartilage matrix demonstrates a general increase in fibrous tissue as well as in the fibrotic foci. The collagen in these loci is remarkably abnormal. It is composed of short, extremely broad fibrils ranging from 150 to 950 nm in width, which are separated at their terminal ends but fused to each other centrally in a random fashion. On cross section, there are very few round fibrils but rather a marked irregularity in shape, giving the appearance of having fibrils randomly added to others to form enlarged, nonuniform fibril aggregates. On longitudinal sectioning, regular cross-banding across the entire fibril width is seen, but fibril splitting and aggregation are highly irregular. Though no specific molecular abnormalities of collagen have been identified, the disordered self-assembly process points to either a modification on one of the collagen molecules, probably type IX favoring the abnormal fibril aggregation, or a defective noncollagenous matrix molecule, which secondarily interferes with normal cartilage synthesis and allows for deposition of a broad, cross-banded collagen in what should be strictly a cartilage domain. 4. METATROPIC DYSPLASIA

The resting cartilage matrix appears to be normal, but the resting chondrocytes are more vacuolated and contain metachromatic inclusions. Cell progression appears to be normal, but vascular invasion is very irregular and loci of ossified cartilage extend into the metaphysis. A thick band of calcified cartilage bridging the chondro-osseous junction has been described. Ultrastructural study shows cytoplasmic vacuolation, some of which is membrane-bound and the rest non-membrane-bound containing the appearance of glycogenlike material. 5. SPONDYLOEPIPHYSEALDYSPLASIA

In SED congenita there is regular chondrocyte proliferation and maturation in most cases, although the physeal columns are short and the septae wide. On occasion, there is irregular vascular invasion from the metaphysis. By ultrastructure, chondrocytes in all zones contain cytoplasmic inclusions of widely dilated rough endoplasmic reticulum containing a homogeneous, mildly electron dense, finely granular material (Fig. 3I) (364).

In SED tarda the epiphyses remain unossified until late in childhood. Cartilage tissue shows normal chondrocyte maturation, although there is clustering of proliferative chondrocytes. 6. KNIEST DYSPLASIA

Resting cartilage is markedly abnormal showing large chondrocytes in a loosely woven matrix with numerous empty spaces, which led to the term "Swiss cheese cartilage syndrome" (91). In young children, the growth plate is hypercellular with large chondrocytes and little intervening matrix. Vascular penetration is irregular, leading to the formation of broad, short, irregular spicules of calcified cartilage and bone. With increasing age, there is some attempt at column formation but the cartilage is still hypercellular. The physeal growth zone is disorganized with little evidence of proliferating and hypertrophic cell columnation. There is an abnormally short proliferative zone. The chondrocytes show cytoplasmic inclusions, which by ultrastructural studies show dilated cistemae of rough reticulum. In the hypertrophic zone, irregular columns of several enlarged cells are seen, and often there are cytoplasmic inclusions in the hypertrophic cells. Throughout the growth plates, scattered foci of fibrillated matrix with mucoid degeneration are found as well as larger acellular areas. 7. METAPHYSEAL DYSPLASIA

Histopathologic changes are similar in variants such as Jansen (the most severe), Schmid (the mildest) (353), and McKusick (cartilage-hair disease variant). The cartilage stains unevenly, the matrix has a fibrillar appearance, and the chondrocytes are larger than normal. There are clusters of proliferating and hypertrophic cells surrounded by dense staining collagen rather than linear columns. The septae are wider than normal and composed of dense fibrous material. Metaphyseal vascular invasion is irregular, and hypertrophic cartilage cells often persist in the metaphysis. The tonguelike extensions of cartilage into the metaphysis result in the radiolucent linear metaphyseal streaking often seen. Chondrocytes at all levels contain rough dilated endoplasmic reticulum as seen by electron microscopy, although this observation has not been made universally (57, 196). 8. SPONDYLOMETAPHYSEALDYSPLASIA

The growth cartilage contains short, irregular columns and often fibrous septae, which are quite wide. There is a reduced proliferative zone and irregular formation of cartilage columns. The hypertrophic zone is present, and vascular invasion is regular in some areas but uneven in others. Electron microscopy shows large inclusion bodies bound by a smooth membrane within the chondrocytes. 9. MULTIPLE EPIPHYsEAL DYSPLASIA

End-stage osteoarthritis characterizes many cases of multiple epiphyseal dysplasia. Gross and histologic findings of

F I G U R E 5 Histopathology from cases of diastrophic dysplasia are shown. (A) Photomicrograph from iliac crest cartilage shows increased tissue fibrosis in association with premature chondrocyte degeneration. The dark staining regions of fibrosis (F) are totally abnormal for this part of the iliac crest growth plate physis. The tissue is relatively hypocellular. Chondrocytes have died and formed either ghost cells (G) or target cells (T). The ghost cells are pyknotic and the chondrocyte lacunae begin to fill with fibrous tissue. The presence of a target cell refers to the pericellular circumferential rings of increased fibrosis. (B) Pericellular fibrosis is seen in this electron micrograph. The chondrocyte shows a normal intracellular fat (f) collection as well as slightly increased glycogen. (C) Electron micrograph shows a pyknotic cell with increased fibrosis within the chondrocyte lacunae. This would appear as a ghost cell at the light microscopic level. (I)) Examples of the fibrotic tissue within cartilage matrix at increasingly higher levels of ultrastructural resolution. Random collagen deposits are seen. The widest collagen fibril (arrow) in this photomicrograph is 375 nm wide. The more normal appearing, adjacent cartilage matrix shows an increased concentration of fibrillar collagenous material. (1)i) Cross sections of the large collagen fibril aggregates highlight the extremely wide diameters and irregular outlines (closed arrows). Some accumulations have distinctly

F I G U R E 5 (continued) abnormal flat borders. The size variability is extensive and several fibrils (closed arrows) give the appearance of having been formed by a random "tacking on" of fibrils at the periphery. The normal appearing collagen components of cartilage matrix (open arrow), which are 20-30 nm wide, are in marked distinction to the large fibril accumulations (x47,250). (Dii, Diii) Collagen fibrils from one of the fibrotic foci. Note the continuity of the cross-banding across the widest diameter of the accumulation (arrow). The fibril is 890 nm wide at this point. As the splitting begins, each individual segment retains its cross-banded appearance (x38,259). [Parts A - C , Dii, and Diii reprinted from Shapiro (1992), Calcif. Tiss. Int. 51:324-331, copyright notice of Springer Verlag, with permission.]

SECTION IX 9 Histopathologic Changes in Specific Chondrodysplasias

F I G U R E 6 Electron micrograph showing parallel linear lamellar collections of alternating electron dense and electron lucent material in dilated rough endoplasmic reticulum from a patient with multiple epiphyseal dysplasia due to a mutation in the ct3 chain of type IX collagen. Fat globules (f) in chondrocyte cytoplasm are a normal finding. [Reprinted from (36).]

osteoarthritis have been demonstrated on specimens removed at the time of total joint arthroplasty, usually for hip and knee problems. There are few if any histopathologic samples of articular cartilage from the early stages of multiple epiphyseal dysplasia. Iliac crest biopsies and a tibial physeal biopsy report, however, show cartilage changes characterized by inclusions in chondrocytes and increased fibrosis in the septae of the hypertrophic zone. A case of MED showed rough endoplasmic reticulum of epiphyseal chondrocytes filled with linear lamellar arrays of alternating electron dense and electron lucent material, presumably abnormally processed type IX collagen (Figs. 6 and 7). Many cases show normal histology at the iliac crest and costochondral junction, although this is not surprising because often only some regions are affected. 10.

CHONDRODYSPLASIA PUNCTATA,

CONRADI-HUNERMANN VARIANT

The proliferative cartilage of the growth plate can show areas of degeneration and cyst formation, often with excessive fibrous tissue and dysplastic calcification. Endochondral bone formation is normal in some growth plates and abnormal in others. Studies of the Conradi-Htinermann form of this disorder show normal resting cartilage but areas of calcification throughout the cartilage matrix, particularly around individual lacunae. There are extensive areas of fibrous tissue replacement of cartilage within the growth plate and bony spicules arising directly from the fibrous connective tissue rather than from a classic endochondral sequence. Isolated areas of endochondral ossification are also seen, however. In those who survive, the calcific punctate lesions begin to resorb toward the end of the first year of life and have usually disappeared by 4-5 years of age. Some appear to undergo actual resorption, but others are incorporated into secondary center bone.

777

F I G U R E 7 Electron micrograph from same iliac crest biopsy as shown in Fig. 6 also shows RER but inclusions now appear as electron dense dots. Linear lamellae in one projection and dots in another lead to the interpretation of a series of rod-like structures in the RER. The circular inclusions of homogeneous material in the RER are fat which are pathologic in this position. [Reprinted from (36).]

11. M U C O P O L Y S A C C H A R I D O S E S

MPSI (Hurler) (297), MPSIII (SanFilippo), MPSIV (Morquio), and MPSVII (Sly) (241) all show similar histopathologic changes. The disorder is associated with large chondrocytes that are filled with several independent lysosomal vacuoles seen by electron microscopy (195). Light microscopy reveals large areas of connective tissue that focally disrupt the growth plate, although the rest of the growth plate undergoes fairly normal endochondral ossification. Where fibrous tissue replaces growth cartilage, fibrous ossification follows at the metaphyseal end of the lesion. 12. MUCOLIPIDOSIS In mucolipidosis II, marked abnormalities of the growth plate with loss of normal cartilage architecture and the absence of endochondral ossification were noted (234). In a child who died at 1 day of age, there was histologic evidence of abnormally long columnar and hypertrophic cell layers of the growth plate, with failure of the normal pattern of vascular invasion and no evidence of metaphyseal bone formation. The cartilaginous septae were not calcified, and invasion of connective tissue with vascularization from the metaphysis into the cartilaginous cells was irregular. 13. PSEUDO-ACHONDROPLASIA

This disorder shows the most characteristic and specific ultrastructural finding in the skeletal dysplasias. Chondrocytes from all layers of iliac crest and proximal fibular physeal cartilage show curvilamellar bodies within the dilated rough endoplasmic reticulum, consisting of alternating electron dense and electron lucent layers arranged in a curvilinear or whorled array. This finding has been noted in several patients in several reports (58, 194, 294, 317).

778

CHAPTER 9 ~ Skeletal Dysplasias

TABLE VIII Abnormalities o f the Lower Extremities-Approaches to Their T r e a t m e n t a Hip

Congenital hip dislocation

Acetabular dysplasia Congenital coxa vara Flexion contracture

Thigh

Abduction or adduction contracture Lack of sphericity of femoral head Anterolateral bowing

Knee

Bowleg (genu varum) or knock-knee (genu valgum) Flexion contracture

Leg Ankle

Foot

Patellar dislocation (lateral) Misshapen epiphyses; lax ligaments Exostosis Anterior or anterolateral bowing of tibia Valgus tilt of ankle with pronated (fiat) foot due to relative shortness of distal fibula; varus tilt of ankle with inverted foot due to relative overgrowth of distal fibula Clubfoot

Toe deformities (hallux valgus, hammer toes)

Closed treatment is often ineffective; may need open reduction, with or without pelvic and femoral osteotomy Pelvic osteotomy Valgus osteotomy of proximal femur Soft tissue release; proximal femoral extension osteotomy Soft tissue release No current therapy Improvement is usually spontaneous; may require femoral osteotomy Osteotomy of proximal tibia-fibula or asymmetric stapling of proximal tibial growth plate Soft tissue release at posterior knee; distal femoral extension osteotomy Soft tissue realignment repair Bracing; surgery produces uncertain results Surgical excision if symptomatic Improvement is usually spontaneous; may require osteotomy Distal tibial fibular osteotomy near skeletal maturity; asymmetric distal tibial and/or fibular growth plate stapling Serial casting beginning at birth; frequently requires soft tissue posteromedial release; calcaneal and tarsal osteotomies or arthrodesis Serial casting; sometimes requires metatarsal osteotomies, surgical correction if symptomatic

aDerived from Shapiro F (1987) Epiphyseal disorders. N Engl J Med 317:1702-1710.

X. O R T H O P E D I C D E F O R M I T I E S I N SKELETAL DYSPLASIAS---REGIONAL ABNORMALITIES AND THEIR RELATION TO CLINICALLY SIGNIFICANT DEFORMITY

A. Overview The skeletal dysplasias are characterized by structural abnormalities of developing bones. These are important in a clinical sense in two broad areas. One concerns diagnosis because the large majority of dysplasias are currently defined by the radiographic appearance of the skeleton. The other concerns bone deformities, some of which are incompatible with life, some of which are amenable to improvement with surgical intervention, and some of which have no clinical significance. There are three general approaches a practitioner can take in assessing a child with a skeletal dysplasia. One is to review the entire spectrum of clinical and radiographic findings in relation to the described syndromes as outlined in

several large compendia, make a diagnosis as to the name of the disorder in question, and then assess the patient, searching for the skeletal abnormalities described in the literature for that particular disorder. A second approach is to focus on specific structural abnormalities such as hypoplastic clavicles, short neck, and coxa vara, and refer to lists available that give differential diagnoses of disorders based on specific anomalies. A valuable book for this approach is Gamuts in Bone, Joint and Spine Radiology by Reeder (257). The third is simply to use a patterned approach to each patient, assessing each region of the body in a generalized to specific fashion, defining structural abnormalities, and deciding on therapeutic approaches with syndromal diagnosis then following. The regional abnormalities characteristic of skeletal dysplasias are summarized in Table II and the lower extremity problems in Table VIII. Over the past several years, the number of distinct entities described has increased greatly as have the various subdivisions, which make yet finer distinctions. Although the majority of dysplasias throughout the world are encompassed within a relatively few diagnostic

SECTION X 9 Orthopedic Deformities in Skeletal Dysplasias

entities, it is necessary to realize that even today many patients fit no described criteria, or do so only with considerable leeway used in definition. For these reasons, assessment region by region in relation to management concerns for any particular case has much to recommend it. The orthopedic concerns in a skeletal dysplasia patient can be divided into four major categories: spinal deformity, neurologic complications of spinal deformity, upper and lower extremity deformities, and limb shortening. The orthopedic consequences of the skeletal dysplasia diagnosis are reviewed in Table II and in more detail later. Reviews of diagnostic and orthopedic management of skeletal dysplasias have been published with varying degrees of detail (11, 12, 154). Spinal deformities have been reviewed well by Tolo (338). A significant number of the spinal deformities in the skeletal dysplasias have neurologic complications, most of which develop slowly over a few years. These involve spinal cord compression varying from myelopathy, paraparesis, and quadriparesis with atlantoaxial instability, cervical kyphosis, or thoracolumbar kyphosis to C 1-C2 abnormalities with ligamentous laxity, cauda equina, or nerve root compression syndromes with spinal stenosis in the lumbar region. If cervical abnormalities are ruled out, spinal cord compression tends to be associated with the severe angular kyphosis usually at the thoracolumbar junction, presenting with a spastic paraparesis.

B. Cervical Spine Abnormalities Developmental abnormalities of the cervical vertebrae occur in many skeletal dysplasias and, if unrecognized, can lead to compression of the spinal cord with gradual paraparesis or quadriparesis or even to acute quadriparesis and sudden death with minimal trauma (160). The skeletal dysplasias with cervical spine problems include about 35 of the 150 well-described disorders. A developing cervical myelopathy is suspected in a child with easy fatigability, hyperreflexia, or weakness. Pain is not a prominent symptom in the childhood skeletal dysplasia patient with cervical instability; in one series of 6 patients, all of whom came to cervical fusion surgery, no patient complained of cervical pain, and in another study (326), none of 11 patients with odontoid dysplasia and atlantoaxial instability in Morquio's disease had significant neck pain (175). All patients with a skeletal dysplasia should have cervical radiographs, especially prior to any anesthesia in which intubation is to occur. Even when vertebral stabilization surgery is performed in the first decade, subsequent recovery of any cervical myelopathy can be limited so that early diagnosis and presymptomatic fusion are desirable. An assessment of upper cervical abnormalities in 171 patients with skeletal dysplasia has been presented (Table IX) (370). Lachman has briefly reviewed the primary cervical spine abnormality seen in each of 35 syndromes

779

(160). The specific cervical spine abnormalities in the skeletal dysplasias follow. 1. MALDEVELOPMENT OF THE ODONTOID PROCESS WITH ATLANTOAXIAL INSTABILITY Maldevelopment of the odontoid process (the dens) of C2 and laxity of the ligaments result in subluxations of cervical vertebra 1 on cervical vertebra 2 (C1 on C2) in many dysplasias (Fig. 8Ai) (155). The odontoid process may be absent, markedly hypoplastic, or not united to the main body of C2. It is important to assess cervical stability because vertebral anomalies alone are not always associated with instability if adjacent ligamentous structures compensate. Diagnosis is confirmed by some or all of lateral cervical radiographs in flexion and extension, open-mouth views to show the atlantoaxial-occipital complex, C1-C2 tomograms, CT scans with myelography, or magnetic resonance imaging studies. The lateral radiographs show odontoid abnormality, but any instability is best shown with the active flexion-extension views with forward shifting of the axis (C1) on C2 with flexion. The atlanto-odontoid interval should not exceed 4-5 mm in the child and 2-5 mm in the adult. All patients with Morquio's disease (MPSIV) have dysplasia of the odontoid process with atlantoaxial instability, 11 of 11 and 15 of 15 in two series studied. There is also a very high incidence of upper cervical bony anomalies in the other mucopolysaccharidoses [75% (15 of 20) of patients with Hurler, Hunter, Scheie, Sanfilipo, and MaroteauxLamy disorders]. A patient (15 years of age) with quadriparesis in MPSVII has been reported. The condition is also extremely common in spondyloepiphyseal dysplasia congenita, occurring in one series in 20 of 24 patients (83%), and pseudo-achondroplasia (13 of 23 or 57%). It is also seen in Kniest dysplasia, chondrodysplasia punctata (C-H type), metatropic dysplasia, spondyloepiphyseal dysplasia tarda, Larsen's syndrome, Dyggve-Melchior-Claussen dysplasia, and spondylometaphyseal dysplasia. All cases of odontoid hypoplasia or absence have a higher than normal risk for cervical spinal cord compression. Neurologic symptoms of cervical compression (myelopathy) include decreased physical endurance, progressive weakness, and pyramidal tract signs, including spasticity, clonus, hyperreflexia, and extensor plantar reflexes. Sensation tends to be maintained for a considerable period. Neurologic findings are often greater on one side than the other because the cervical cord often is displaced to one side. In those children with the triad of odontoid hypoplasia, instability of C 1 on C2, and spinal cord compression, the onset of symptoms often becomes apparent between 5 and 10 years of age. Gait problems are often attributed to the hip and knee deformities of the dysplasia and the cervical component often is delayed in diagnosis. Diagnosis is best performed with plain radiographs, including anteroposterior, open-mouth, and lateral flexion, neutral, and extension views, MR imaging for soft tissue and spinal cord

780

CHAPTER

9

~

Skeletal Dysplasias TABLE IX

No. of patients

Morquio's disease Other mucopolysaccharide disease: Hurler's Hunter' s Scheie's Sanfilippo Maroteaux-Lamy's Spondyloepiphyseal dysplasia: Congenita Tarda Pseudo-achondroplasia Achondroplasia and hypochondroplasia Multiple epiphyseal dysplasia Chondrodysplasia punctata Metatropic dysplasia and Kniest disease Diastrophic dysplasia Spondylometaphyseal dysplasia Total

Cervical Anomalies in t h e Skeletal Dysplasias a

Age range (years)

No. with odontoid hypoplasia

No. with absence

No. with other defects base of skull to C3

No. with known neurological complications

Total no. (%) with bony defects

All 15 (100)

15

1-15

7

8

0

5

6 5 3 3 3

1-13 3-14 6-14 1-11 3-6

4 3 1 3 3

0 0 0 0 0

4 0

0 0

1

0

0 0

0 0

15 of 20 (75)

24 18 23 23

Neonate to 54 3-62 Neonate to 53 2 months to 65

15 3 8 1

1 0 0 0

4 1 5

1 1 0

1

0

20 of 24 (83) 4of18 13 of 23 (27) 2 of 23

18 14 7

4 to adult Neonate to 11 Neonate to 15

6 2 1

0 0 1

0 9 2

0 0 0

6of18 11 of 14 4of7

6 3

Neonate to 48 7-16

1 1

0 0

3 0

1 0

4of6 1 of 3

59

10

30

8

80 (47)

171 aArch Dis Child 64:283-288, 1989.

visualization, and CT scanning for bony anatomy. Tomograms were used previously to outline the poorly formed dens (Fig. 8Aii). Treatment involves an occiput-to-C2 posterior fusion or a C 1-to-C2 fusion if the posterior arch of C 1 is intact (Fig. 8Aiii). Early diagnosis and fusion represent the ideal treatment because recovery of lost function postsurgery is often incomplete if it had reached moderate to severe levels. The skeletal dysplasias with common and significant C 1C2 instability requiring careful early observation are the type II collagenopathies SED congenita, Kniest dysplasia, and spondyloepimetaphyseal dysplasia (Strudwick), pseudoachondroplasia, metatropic dysplasia, chondrodysplasia punctata (Conradi-Hiinermann), diastrophic dysplasia, and Dyggve-Melchior-Clausen dysplasia. 2. MIDCERVICAL KYPHOSIS

Midcervical kyphosis due to underdevelopment of the vertebral bodies of C3, C4, or C5 occurs frequently in diastrophic dysplasia (Fig. 8B). Treatment is by posterior cervical fusion if there is worsening deformity and cord compression, although in the absence of neurological symptoms expectant observation can be used in the first few years

of life because spontaneous correction of the kyphosis with associated stabilization in many has been described up to 5 years of age. If the angulation is worsening, however, posterior fusion regardless of age is warranted because cases of progressive kyphosis, quadriparesis, and death have been described. Cervical kyphosis is also seen in type II collagenopathies of the SED and SEMD types and can be severe in lethal campomelia, in those few campomelic patients who survive, and in de la Chapelle, atelosteogenesis III, chondrodysplasia punctata (rhizomelic), and hypochondrogenesis disorders. 3. CERVICAL SPINA BIFIDA OCCULTA Cervical spina bifida occulta is an almost invariable occurrence in diastrophic dysplasia. It was noted in each of 16 patients in two published series. The spina bifida is usually narrow in the 2- to 4-mm range, but defects as wide as 10 mm have been described. Generally it is present in the mid- and lower cervical vertebrae but can extend to the upper thoracic area. It does not appear to cause deformity by itself but must be carefully assessed at the time of posterior surgical procedures.

SECTION X ~ O r t h o p e d i c D e f o r m i t i e s in S k e l e t a l D y s p l a s i a s

781

F I G U R E 8 Characteristic cervical spine abnormalities in the skeletal dysplasias are shown. (A) Lateral radiograph shows C 1-C2 anterior subluxation in a 7-year-old female with SED who presented with quadriparesis. (Ai) Lateral cervical view in forward flexion shows C1 anterior subluxation on C2 (arrow) (Aii), tomogram shows underdeveloped dens of C2 (arrow), and (Aiii) lateral view shows stabilized alignment following occiput to C2 fusion. (B) Midcervical kyphosis is shown on this lateral tomogram radiograph of a patient with diastrophic dysplasia at 18 months of age. Note the underdevelopment of the anterior parts of the vertebral bodies of C3, C4, and C5.

4. CERVICAL SPINAL STENOSIS

Cervical spinal stenosis of variable degrees usually accompanies foramen magnum stenosis in achondroplasia but rarely is severe enough to require surgical decompression.

5. CORD-ROOT COMPRESSION BY EXOSTOSES Intraspinal protrusions of exostoses in hereditary multiple exostoses can produce cord and root compression at each of the cervical, thoracic, and lumbar levels (183).

782

CHAPTER 9 ~ Skeletal Dysplasias

FIGURE 9 A severe scoliosis that can occur in the skeletaldysplasias. They tend to occur earlier, progress more rapidly, develop rigidity, and cause spinal cord compression to a greater extent than in idiopathic varieties. Child with diastrophic dysplasiawith 53~curve at 1.5 years of age.

C. Thoracolumbar Spine Abnormalities" Scoliosis, Kyphosis, and Kyphoscoliosis There is a high incidence of thoracolumbar scoliosis, kyphosis, or kyphoscoliosis in many dysplasias, particularly including spondyloepiphyseal dysplasia congenita, spondylometaphyseal dysplasia, spondyloepimetaphyseal dysplasia, diastrophic dysplasia, Kniest syndrome, Morquio's disease, metatropic dwarfism, and pseudo-achondroplasia (Fig. 9) (111). These deformities in the skeletal dysplasias have a marked tendency to occur much earlier, progress more quickly, and be more rigid than those in idiopathic conditions. Compression of the spinal cord and spastic paraparesis can occur, primarily with kyphosis or kyphoscoliosis, whereas such neurological sequelae are extremely rare in patients with even advanced idiopathic scoliosis or scoliosis as the main deformity in a skeletal dysplasia. Kyphosis of a moderate to severe form can occur as an exclusive deformity in achondroplasia or as the major component of a kyphoscoliosis with spondyloepiphyseal dysplasia congenita, MPSIV (Morquio's disease), diastrophic dysplasia, pseudo-achondroplasia, and metatropic dysplasia. An isolated thoracolumbar kyphosis is particularly common in achondroplasia. Winter comments that "all newborns with achondroplasia have at ~ i l d thoracolumbar kyphosis,

and this may be associated with anterior wedging of the apical vertebra." The kyphosis usually resolves to a normal configuration on its own as the patient becomes ambulatory, but in about 30% of patients the kyphosis persists and in one-third of these it can progress to a major fixed thoracolumbar kyphosis. Two types are seen: one a long, moderate kyphosis extended over several segments and the other a sharply angulated, localized kyphosis centered at one or two wedged vertebrae and referred to as a gibbus deformity. The latter group is particularly likely to develop neurologic symptoms, but even a relatively gentle kyphosis can lead to symptoms in the achondroplastic individual because of the associated spinal stenosis with the narrowed interpedicular distance. The neurologic complications in achondroplasia are more likely to occur and are far more serious in those with thoracolumbar kyphosis, particularly the short gibbustype deformity. Scoliosis can also occur in achondroplasia but rarely reaches a degree of deformity requiring full-time brace use or spinal fusion procedures. Symptoms referable to the kyphosis with increased discomfort and neurologic dysfunction are generally not seen until the third or fourth decade. Although those with scoliosis in achondroplasia rarely proceed to a significant deformity, kyphosis is different and the progressive nature of the deformity in those in whom it is present is such that aggressive early treatment even in the immature age group is recommended. Bracing is used to hyperextend the deformed region. It remains important in the child with achondroplasia, and indeed with any skeletal dysplasia, to associate any increase in symptoms not just with the deformity itself but also with the possibility of a herniated disk because the narrowed canal leads to cord and root pressure much earlier than is the case when a normal spacious canal is present. Anterior release and posterior spinal fusion or posterior fusion alone is needed depending on the nature of the symptoms and deformity. Hensinger points out that kyphosis may occur in nearly every type of dwarfism and that, even though all vertebrae are dysplastic, the deformity tends to be concentrated at the thoracolumbar junction (110). Kyphosis also is seen fairly frequently in spondyloepiphyseal dysplasia, a disorder characterized by a shortened trunk due to decreased height of the vertebral bodies. SED congenita has the more severe thoracic kyphosis and exaggerated lumbar lordosis. Progressive kyphosis can also occur in pseudo-achondroplasia. Individuals with this disorder have the same basic body appearance as those with achondroplasia, except that the skull and facial features are normal and the hand findings are different. Radiologically the spinal abnormalities are similar to a spondyloepiphyseal dysplasia and there is no decrease in the interpedicular distance of the lumbar spine. It can show exaggerated dorsal kyphosis and lumbar lordosis, and neurogenie degeneration distal to the kyphosis site has been shown. In the mucopolysaccharidoses, spinal deformity, which on occasion is an isolated kyphosis, can be seen with Mor-

SECTION X ~ Orthopedic Deformities in Skeletal Dysplasias

quio's (MPSIV), Hurler's (MPS1H), and Maroteaux-Lamy (MPSVI). One of the earlier features of Hurler's is a prominent gibbus at the dorsolumbar junction with the vertebral body on the lateral radiograph at the apex of the kyphosis wedged with an anterior beak. In Morquio's disease there is universal flattening of the vertebrae (platyspondyly) and kyphosis with anterior beaking and wedging of the vertebrae most marked at the dorsal-lumbar junction. In the MPSVI syndrome deformities are mild with not as severe flattening of the vertebrae and only a mild gibbus at the thoracolumbar junction. In Hurler's syndrome the child's overall condition generally deteriorates over the first decade and treatment would be symptomatic or at most with bracing. In MPSIV and -VI, however, survival is prolonged with good mental status and aggressive brace treatment at the onset of the deformity is used. Operative stabilization is occasionally required with posterior fusion predominating. A thoracolumbar kyphoscoliosis is particularly common in diastrophic dysplasia. It is invariably progressive and often rapidly so even in the first decade. Deformity also occurs with metatropic dysplasia. In those patients with campomelic dysplasia who survive, severe spinal problems are invariably seen. These can involve severe cervical kyphosis and cervical instability, but all will develop a severe progressive thoracic kyphoscoliosis. Late ossification of the midthoracic pedicles is a clear and early diagnostic finding, and vertebral body hypoplasia is the cause of the deformity. In one review of 8 patients (collected from several centers), all had scoliosis averaging 63 ~ at 6 years, and 6 of the 8 had severe kyphosis averaging 126 ~ The cervical kyphosis in 3 was 16~ 32 ~ and 150 ~ Congenital vertebral anomalies were present in all. The hypoplastic vertebral bodies appeared to be the major deforming regions. Late ossifying thoracic pedicles appear to be unique to this disorder. Scoliosis requiting treatment occurred in a high percentage of patients with diastrophic dysplasia, spondyloepiphyseal dysplasia, metatrophic dysplasia, and chondrodysplasia punctata in one review. Bracing is attempted for both kyphotic and scoliotic deformities, but it often is ineffective in the skeletal dysplasias due to the small trunk size, the rigidity of the curves, and the marked amount of trunk deformation. Increasing improvements in orthotic design and fabrication, however, allow for more accurate and comfortable fitting and should soon indicate somewhat improved results. Spinal fusion is resorted to relatively early if the deformity worsens with bracing. In patients between 3 and 10 years of age, worsening scoliosis not amenable to bracing can be treated surgically. A distraction rod, inserted without spinal fusion, allows growth to continue for several years and minimizes the scoliosis; spinal fusion is then performed close to skeletal maturity. The use of anterior and posterior approaches improves correction in this group in which rigidity is more prominent and at earlier ages than in idiopathic groups. In addition, instrumentation

783

systems designed for smaller, younger patients combined with use of somatosensory evoked potentials for intraoperative spinal cord monitoring further increase the value and safety of surgical intervention. Imaging improvements by CT scanning, MR imaging, and three-dimensional reconstructions provide for more specific understanding of the deformity.

D. Lumbar Spinal Stenosis and Lumbar Lordosis In achondroplasia and some other dysplasias, the interpedicular distance of the lumbar vertebrae progressively narrows from L1 to S 1, as seen on anteroposterior radiographs (Fig. 10A) (42, 96 180). The distance in the lower three vertebrae in particular becomes markedly narrowed. This finding is present even in the newborn period and allows for the diagnosis of some skeletal dysplasias at that time. Thoracolumbar kyphosis that worsens instead of resolving and that is associated with spinal stenosis can lead to compression of the spinal cord in the first 10 years of life. The pedicles are also shortened as seen on lateral radiographic projections, further narrowing the spinal canal (Fig. 10B). If herniation of an intervertebral disk occurs in the lower lumbar region, the spinal stenosis limits space in the canal and puts pressure on the cauda equina and the nerve roots, resulting in claudication and flaccid paralysis. These neurological sequelae are far more extensive than those in patients with disk herniation into an otherwise normally sized spinal canal. Progressive neurologic symptoms at any age can necessitate anterior or posterior spinal decompression and spinal fusion. A prominent lumbar lordosis is seen in achondroplasia (Fig. 10B) and in many other skeletal dysplasias, including pseudo-achondroplasia and diastrophic dysplasia. Although antilordotic bracing has been used, there is no published indication of its effectiveness. The lordosis is not of sufficient concern to warrant primary surgery. In some disorders, such as diastrophic dysplasia, much of the lumbar lordosis is secondary or compensatory to marked hip flexion contractures and will improve upon surgical correction of the hip position. Bethem et al. reviewed 80 patients from their institution with disproportionately short stature, with virtually all showing some aspect of a spinal disorder (28). Their study continues to serve as an indicator of the relative frequency of specific deformities in the skeletal dysplasias. Kyphosis was the most common clinically significant axial deformity in the skeletal dysplasias, being present in 57 patients (71%). Scoliosis considerably greater than 30 ~ was present in 21 patients (26%). Atlantoaxial instability was documented in 10 patients (13%) with a total absence of the odontoid process in 5 patients (6%). Neural symptoms occurred in 13 patients (16%), with 10 having paraparesis and 3 quadriparesis. A summary of the spinal disorders in the more common skeletal dysplasia syndromes is shown in Table X.

784

CHAPTER 9 ~

Skeletal Dysplasias

F I G U R E 10 Abnormalities of the lumbar vertebrae are characteristic of achondroplasia and some other skeletal dysplasias, such as thanatophoric and diastrophic dysplasias. (A) Narrowing of the interpedicular distance of the lumbar vertebrae is characteristic of achondroplasia. This is evident radiographically even in newborns. In the normal, there is progressive widening of the distance from L 1 to S 1. In achondroplasia, there is narrowing at each successive level particularly from L3 down to S1. This results in a spinal stenosis. (B) Lateral radiograph of achondroplasia also shows the shortened pedicles contributing to the spinal stenosis. Concavities in the posterior margins of the vertebral bodies are characteristic. A prominent lumbar lordosis is seen. Shortening of the pedicles limits growth posteriorly, whereas anteriorly the vertebral body growth is relatively less affected.

TABLE X

Spinal Disorders Associated with Different T y p e s o f D w a r f i s m Scoliosis

Kyphosis

Atlantoaxial instability

Decreased interpedicular disease

Cervical kyphosis

Spina bifida

Achondroplasia

No

Yes

Spondyloepiphyseal dysplasia

Yes

Yes

No

Yes

No

No

Yes

No

No

No

Mucopolysaccharidosis

No

Diastrophic dwarfism

Yes

Yes

Yes

No

No

No

Yes

No

Yes

Yes

Yes

Metatropic dwarfism

Yes

Yes

Kniest syndrome

No

Yes

Yes

No

No

No

No

No

No

Metaphyseal dyostosis

No

Yes

No

Yes

No

No

No

Chondrody strophia calcificans congenita

Yes

No

No

No

No

No

SECTION X ~ Orthopedic Deformities in Skeletal Dysplasias E. A b n o r m a l i t i e s of the Skull

The skull is often relatively large (macrocephaly) in the skeletal dysplasias because it is preformed in intramembranous bone, which is unaffected in the chondrodystrophies. On occasion, as in achondroplasia, this enlargement is associated with hydrocephalus, although this is usually of the benign, communicating variety without negative sequelae. The nasal region of the face, which is preformed in cartilage, is smaller than normal leading to the depressed nasal bridge appearance, and the base of the skull surrounding the foramen magnum is also of cartilage origin and smaller than normal. Marked delay in closure of the anterior fontanelle occurs in cleidocranial dysplasia, osteogenesis imperfecta, and some other disorders (Figs. 11Ai and 11Aii). Wormian bone formation characterizes types II and III osteogenesis imperfecta but is also seen in cleidocranial dysostosis, hypophosphatasia, and pycnodysostosis (Fig. liB). The softened osteopenic bone in OI leads to a flattened posterior skull region, and there is also prominent frontal bossing. In the infantile malignant form of osteopetrosis, the skull bone is also thickened and dense but the major negative sequelae result from the narrowed optic and auditory canals, which fail to enlarge (remodel) with growth and whose stenosis contributes to pressure on the optic and auditory nerves with progressive blindness and deafness. There is a group of disorders characterized by craniosynostosis and peripheral limb malformations, including Apert syndrome with symmetric syndactyly of hands and feet, Pfeiffer syndrome with broad thumbs and great toes, Jackson-Weiss syndrome with tarsal-metatarsal coalitions, and Crouzon syndrome with midface deficiency (51, 140). [The broader range of craniofacial dysostoses is beyond the scope of our discussions of the osteochondrodysplasias.]

F. A b n o r m a l i t i e s of the Clavicles

One specific dysplasia, cleidocranial dysplasia, is characterized by either absent or markedly shortened clavicles, such that the patient can touch the shoulders to each other in front of the chest (Figs. 12A and 12B).

G. A b n o r m a l i t i e s of the Extremities 1. MICROMELIC, RHIZOMELIC, MESOMELIC, AND ACROMELIC SHORTENING Shortening of specific segments of limbs is characteristic of many disproportionate skeletal dysplasias with short stature, often to the point of being the feature defining the disorder. Micromelic refers to shortening of all parts of a limb; rhizomelic to proportionally greater proximal shortening (humerus and femur), mesomelic to midlimb shortening most involved in the forearm (radius, ulna) and leg (tibia, fibula); and acromelic to most significant distal segment

785

shortening of the hands and feet. In most skeletal dysplasias the shortening is symmetric. 2. LOWER EXTREMITY LENGTH DISCREPANCIES In some of the skeletal dysplasias limb shortening is asymmetric. The discrepancy itself rarely if ever is significant in the upper extremities, but differences in lower extremity lengths can reach clinical significance. The most significant lower extremity length discrepancy in skeletal dysplasias occurs in Ollier's enchondromatosis and in the related Maffucci syndrome. Because the involvement in these disorders is predominantly unilateral by definition, length discrepancies can be considerable. Lesser but still clinically significant length discrepancies also occur in hereditary multiple exostosis in which many of the differences require epiphyseal arrest on the longer side for appropriate clinical management. Significant discrepancies in other dysplasias are infrequent but can occur, usually in association with asymmetric angular deformities in which shortening can be a component not only of the original malformation but also of the treatment used to correct the deformity. 3. ANGULAR BONE DEFORMITY DUE TO ASYMMETRIC PHYSEAL INVOLVEMENT Asymmetric physeal involvement can lead not only to shortening but also to angular deformation due to unequal growth across the entire physis. The angulation, which is centered in the upper metaphysis, can involve any of varus, valgus, flexion, extension, or rotational malpositions. 4. LIMB DEFORMITY DUE TO UNEQUAL GROWTH RATES OF PAIRED LONG BONES IN FOREARM AND LEG In many skeletal dysplasias there is unequal involvement of radius and ulna in the forearm and of tibia and fibula in the leg. Disparate growth rates can lead to deformation particularly by the end of the long bone with resultant angular deformation and joint displacement (Figs. 13A-13C). As a general rule the ulna is affected more seriously in skeletal dysplasias than the radius. The longer radius compensates for its relative overgrowth by one or all of several mechanisms, including increased metaphyseal-diaphyseal curvature with the concavity toward the shortened ulna (Fig. 13A), gradually developing subluxation and even complete dislocation of the radial head proximally at the elbow, and ulnar deformation and angulation at the distal radial physealmetaphyseal region due to medial physeal tethering by the shortened ulna and its associated ligamentous structures (Fig. 13A). In the leg the fibula can be either more affected or less affected than the adjacent tibia. In those situations in which the fibula is relatively longer, it tends to ride higher at its relationship to the proximal tibia, sometimes coming to lie at the level of the knee joint, whereas relative overgrowth distally leads to a tendency to varus deformation at the ankle

786

CHAPTER

9

~

Skeletal Dysplasias

F I G U R E 11 Developmentalabnormalities of the skull are common in many of the skeletal dysplasias. (A) Marked delay in closure of the anterior fontanelle occurs in cleidocranial dysplasia, osteogenesis imperfecta, and some other disorders. (Ai) Anteroposterior and (Aii) lateral skull radiographs in a patient with cleidocranial dysplasia are shown. (B) Radiograph shows wormian bone formation, a disorder of intramembranous skull bone seen in osteogenesis imperfecta of the more severe congenita or Sillence type II. It has no particular clinical significance but is helpful diagnostically.

(Fig. 13B). If the fibula is relatively shorter than the tibia, which most c o m m o n l y occurs in hereditary multiple exostosis, there are few problems proximally but distally the an-

kle first submits to a valgus deformation, and tethering by the shortened fibula and adjacent ligaments on the distal lateral tibial epiphysis further augments both the valgus defor-

SECTION X ~ Orthopedic Deformities in Skeletal Dysplasias

787

F I G U R E 12 Abnormalities of the clavicle characterize some skeletal dysplasia disorders, the most common being cleidocranial dysplasia. The abnormalities include shortened clavicles (in particular laterally), aplastic clavicles with isolated remnants, and areas of discontinuity like with a pseudarthrosis. (A) The fight clavicle is absent and the left is deficient laterally; (B) there is a fight pseudarthrosis, whereas on the left there is only a small central fragment.

mation of the ankle and obliquity of the distal tibial physis and adjacent articular surface (Fig. 13C). 5. J O I N T C O N T R A C T U R E S AND INSTABILITY

On occasion angular deformation is centered at the joints themselves rather than in the adjacent physeal, metaphyseal, or even diaphyseal regions of the developing bones. The contractures are almost always to the flexed position and a particular characteristic of diastrophic dysplasia (Fig. 14A). Joint instability characterized many of the more severe dysplasias. This results from a combination of ligamentous laxity and relative underdevelopment of parts of the entire epiphyses allowing for the more spacious joint cavity and also predisposing one to malalignment (Fig. 14B). Surgery is performed on the lower extremities far more frequently than on the upper, because problems in the lower extremities result in considerable short-term and long-term morbidity and are also more amenable to correction. The operative procedures for the various deformities in the skeletal dysplasias are not specific for the various diseases but simply utilize operative approaches designed for the deformity that occurred regardless of its cause. A large number of deformities of the upper and lower extremities occur in the various skeletal dysplasias. The presence of a deformity alone does not necessitate surgery. Rather, the decision to intervene should be based on the current degree of discomfort or limitation of function. Correction of deformity because of concern about problems that might develop in the future is not necessarily warranted in

the skeletal dysplasia patient. The tendency to operate on deformities based on similar disorders that occur posttrauma in otherwise normal patients has not infrequently led to unimpressive results. The importance of intermediate and longterm studies of the natural history of skeletal dysplasia deformity and postsurgical results is leading to some data accumulation, but much remains to be done.

H. Abnormalities of the Hip Region 1. G E N E R A L CONSIDERATIONS

A large number of patients with skeletal dysplasias, especially the multiple epiphyseal and spondyloepiphyseal types, develop premature osteoarthritis, which necessitates total joint arthroplasty in early to middle adult life. Surgery is performed increasingly in an attempt to forestall expected problems due to deformity, especially in the hip region. The preoperative study of hip geometry has been improved by the use of arthrography, CT scans, MR imaging, and three-dimensional computerized reconstructions. Many types of pelvic and proximal femoral osteotomies have been devised, as have bone fixation devices specifically for children. It is important to distinguish a skeletal dysplasia from bilateral Legg-Perthes disease because both can present with clinical and radiographic abnormalities of the proximal femoral capital epiphyses. The main disorders to be differentiated from Legg-Perthes disease are multiple epiphyseal dysplasia, pseudo-achondroplasia, and spondyloepiphyseal

788

CHAPTER 9 ~

Skeletal Dysplasias

F I G U R E 13 Disparate growth rates between radius and ulna in the forearm and between tibia and fibula in the leg are characteristic of many of the skeletal dysplasias and usually lead to limb deformation. (A) Radiograph of a forearm in hereditary multiple exostosis shows a markedly shortened ulna. The radius is curved primarily due to tethering of its distal medial physis. (B) Relative overgrowth of the distal fibula compared to the distal tibia can lead to varus malformation at the ankle. This radiograph is from a patient with achondroplasia. (C) Anteroposterior radiograph of the ankle in a patient with severe hereditary multiple exostosis shows the shortened fibula in relation to the relatively longer tibia. At the ankle this causes the lateral malleolus to be positioned more proximally than normal, leading to valgus deformation. The lateral physeal region of the distal tibia suffers some limitation of growth due to tethering associated with the shortened fibula.

dysplasia c o n g e n i t a and tarda (62). O f these it is the multiple e p i p h y s e a l dysplasia diagnosis that can be the m o s t difficult to make. T h e m o s t relevant features in differential diagnosis are the size, shape, texture, and density of the capital femoral epiphyses, the a p p e a r a n c e of the m e t a p h y s e s , and the shape and texture of the a c e t a b u l u m . In bilateral L e g g - P e r t h e s disease, present in about 15% of patients, a m o s t striking distinguishing feature of the capital f e m o r a l secondary ossification centers is the a s y m m e t r i c a l i n v o l v e m e n t . T h e hips in L e g g Perthes n e v e r a p p e a r to be equally affected by the disease

F I G U R E 14 Joint contractures and joint instability can be seen in certain types of skeletal dysplasia. Contractures are particularly common in diastrophic dysplasia and they almost always are of a flexion type. Instability tends to occur with underdevelopment of epiphyseal regions of an asymmetric fashion with persistence of relatively normal capsular and ligamentous tissues. (A) Flexion contracture of the knee characteristic of diastrophic dysplasia is shown here on a lateral radiographic projection. The film was taken with the knee in maximum extension. [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.] (B) Joint instability can represent a combination of bone malformation from asymmetric physeal growth and underdevelopment of parts of epiphyses with normal joint contours either persisting or in association with ligamentous laxity. Note lateral patellar dislocations.

process simultaneously, the f r a g m e n t a t i o n is characterized by irregular areas of variable density, there are often cystic c h a n g e s in the m e t a p h y s e s , and the a c e t a b u l u m is a l m o s t always n o r m a l particularly in the early phases. T h e diagnosis

SECTION X ~ Orthopedic Deformities in Skeletal Dysplasias

always becomes evident with time because Legg-Perthes improves with healing of the secondary ossification center to a uniform appearance, whereas in the skeletal dysplasias the irregular radiographic appearance persists longer. In multiple epiphyseal dysplasia, on the other hand, the capital femoral epiphyses initially are almost always of homogeneous texture and symmetrically affected. In more than one-half of the patients they are late in appearing. Initially they are of the normal round shape but subsequently most become flat and appear to arise from multiple centers, although the fragmentation tends to be characterized by uniformly dense bone. There are no cystic changes in the metaphysis, and the acetabulum is often shallow but sometimes normal. At a later stage there is also delayed formation of the secondary centers of the greater trochanter in MED. In SED tarda the appearance is almost indistinguishable from that of MED, although the femoral capital epiphysis has a greater tendency to a low, flattened crescentlike shape. These patients tend to present clinically for diagnosis around 10 years of age so that earlier findings are often not defined. In SED congenita, the disorganization of the hip is usually marked with gross irregularity of the bony pelvis, poorly formed acetabulae, delayed appearance of the secondary ossification center, and delayed closure of the triradiate cartilage. Coxa vara is frequently seen. In general there should be no difficulty in differentiation of SED congenita from other disorders. Pseudo-achondroplasia also presents generally with severe abnormalities in the pelvis and hip, the triradiate cartilage is strikingly wider than normal, and the acetabulum is poorly formed. The capital femoral epiphysis appears late and is small and irregular with marked delay in maturation. The metaphyses usually show medial beaking. The major differential diagnosis for Legg-Perthes is with MED, with each of the other deformities showing characteristics that should not be confused. In summary, the characteristics of bilateral Legg-Perthes disease are its asymmetry of involvement and its changing nature over a 1- to 3-year period, whereas with the inherited dysplasias there is no evidence of a relatively rapid worsening phase with subsequent improvement. In the generalized skeletal dysplasias the capital femoral epiphyses invariably show symmetrical, homogeneous involvement, whereas maturation may be severely retarded and structurally abnormal. The following specific hip abnormalities should be assessed. 2. CONGENITAL HIP DYSPLASIA Congenital hip dislocation does occur in the skeletal dysplasias although there has been no specific increased incidence defined in any of the various syndromes. In some instances the dislocation is of teratologic appearance, but in others the hips are dislocatable and thus relocatable at birth and treatment is the same as that used in idiopathic developmental dysplasia of the hip. Closed or conservative treatment, however, is sometimes ineffective and there is an increased need for open reduction with or without pelvic and femoral osteotomies.

789

3. ACETABULAR DYSPLASIA Acetabular dysplasia is quite common in many of the skeletal dysplasias (Fig. 15A). In MED, virtually all patients have hip involvement, even in those mild forms in which all other joints are normal or close to normal and skeletal stature is also in the low-normal range. There is an understandable tendency to consider treating the acetabular dysplasia usually in the middle of the first decade with an innominate osteotomy. No data are available to assess the overall effectiveness of this intervention. One of the problems in a few cases we have seen is that usually there is lack of femoral head sphericity in association with the acetabular dysplasia due to a primary modeling defect of the epiphyseal region, such that premature osteoarthritis would be expected even in the presence of a normal shape to the acetabulum. The innominate procedure can be recommended in the sense that it would at least minimize hip deformity, and should total hip intervention be needed, the pelvic dimension would be appropriate without the need for acetabular augmentation. 4. COXA VARA Congenital or developmental coxa vara characterizes many of the skeletal dysplasias (Fig. 15B). It is particularly common in spondyloepiphyseal dysplasia congenita. Two types of coxa vara are seen. In some, there is the characteristic nonunited segment of bone at the inferior medial surface of the neck, and in others the growth plate between the head and neck region is present but only minimally to nonfunctional such that trochanteric overgrowth soon occurs. In the former type there is frequently a delay in the appearance of the proximal femoral ossification center, and indeed this may not be present within the first decade of life. The shape and position of the femoral head must be assessed by means other than plain radiographs such as arthrography or magnetic resonance imaging. Treatment involves a proximal femoral valgus osteotomy. Delayed appearance of proximal femoral capital ossification centers is marked in spondyloepiphyseal dysplasia and Kniest dysplasia. 5. HIP CONTRACTURES

Hip contractures can characterize some patients with underlying bone and cartilage hip joint abnormalities in the skeletal dysplasias. Flexion contractures, which are particularly common in diastrophic dysplasia, can be treated with soft tissue releases, although a proximal femoral extension osteotomy may be needed. Abduction or adduction contractures are also seen and are best treated with soft tissue releases. 6. LACK OF SPHERICITY OF FEMORAL HEAD Unfortunately this is a common associated finding in the skeletal dysplasias because the epiphyseal model is misshapen leading to a lack of congruity of adjacent articular surfaces and early osteoarthritic symptoms (Fig. 15Ci15Ciii). In some disorders, such as diastrophic dysplasia,

790

CHAPTER 9 ~

Skeletal Dysplasias

F I G U R E 15 Abnormalities of the hip are common in many of the skeletal dysplasias. (A) Acetabular dysplasia develops in many of the skeletal dysplasias, usually in relation to underdevelopment or subluxation of the femoral head. This patient had a spondyloepiphyseal dysplasia. Note the severe acetabular dysplasia with the coxa vara deformity on the right. (B) Coxa vara is commonly associated with many of the skeletal dysplasias. Bilateral coxa vara in an unidentified dysplasia is illustrated. (C) Development of a misshapen femoral head is common in skeletal dysplasias such as multiple epiphyseal dysplasia. The femoral heads are spherical in the neonatal period. This sequence of anteroposterior hip radiographs shows improvement of the acetabular dysplasia following bilateral innominate osteotomies at 8 years of age but femoral head deformation to age 14 years.

stiffness and joint pain can render walking difficult to impossible as early as the start of the second decade of life. Joint stiffness also characterizes Kniest dysplasias. Unfortunately, there is no current therapy for the misshapen epiphysis other than perhaps a varus or valgus osteotomy if there is one segment of the head that appears to relate well to the acetabulum as determined by arthrography and serial X rays in varying positions. Misshapen epiphyses develop particularly as early as the second decade in multiple epiphyseal dysplasia and pseudo-achondroplasia.

Stelling has summarized hip disorders in the skeletal dysplasias (Table XI) (320).

I. Knee Abnormalities 1. B O W I N G OR K N O C K - K N E E D E F O R M I T I E S (GENU VARUM-GENU VALGUM)

Either bowing (genu varum) or knock-knee (genu valgum) deformities can be seen in the skeletal dysplasias (Fig. 16A). Bowing is more common and is often seen in

Mild to moderate

Mild to moderate

Mild to moderate

Enlarged, irregular severe

Early ossification

Early osification

Hurler's

Hunter's

Polydystrophy dwarf (Lamy-Marateaux)

Morquio

Ellis-van Creveld

Thoracic pelvic phalangeal dystrophy

No

Metaphyseal chondrodysplasia (metaphyseal dysostosis) (Jansen type)

No

No

No

Osteopetrosis

Craniometaphyseal dysplasia

No

No

No

No

Multiple exostoses

Osteogenesis imperfecta

Pyknodysostosis

Cleidocranial dysplasia (cleidocranial dysostosis)

(Pyles)

No

No

Hypophostasia

No

No

No

Protrusion

No

No

No

No

No

Dyschondrosteosis

Rare

No

No

No

No

No

Oblique superior margin in infant

No

Gross enlargement severe

Mild, usually good shelf

Mild, usually good shelf

Mild, usually good shelf

No

No

Dysplastic

No

Mildly irregular

Mildly irregular

Protrusion large and irregular

Acetabular abnormalities

Enchondromatosis

No

Osteodysplasty (Melnick-Needles)

No

No

Cartilage-hair hypoplasia

(Schmid type)

No

Mesomelic dwarf Lamy-Marateaux

No

No

Hypochondroplasia

Mesomelic dwarf Langer

Severe

Epiphyseal osteochondroma

Dysplasia epiphysealis hemimelica (of hip)

No

Stippling early, irregular late

Chondrodystrophia calcificans congenita

Diastrophic dysplasia

Mild to severe

Multiple epiphyseal dysplasia

Achondroplasia

Moderate to severe

Spondyloepiphyseal dysplasia

Proximal femoral capital epiphyseal abnormalities

50%

+ -

+ -

No

No

+ -

+ -

No

+ -

No

Yes

+ -

No

No

No, shortneck

No

No

No

Rare

Rare

Rare

No

No, short neck

+ -

No

+-

+-

Frequent

Rare

No

No

No

No

No

No

No

No

Yes

No

No

No

No

No

No

No

+ -

Yes

Yes

Yes

No

No

No

No

+-

+-

+ -

Coxa valga

No

No

+ -

No

No

No

No

No

No

No

No

+ -

No

No

No

No

No

Yes

Rare

Rare

Rare

No

No

Frequent

No

+-

+-

+ -

Dislocation

+ -

No

No

No

No

No

No

No

No

No

No

Yes

No

No

No

No

No

Yes

Yes

Yes

Yes

No

No

Severe

+ -

Yes

Yes

+ -

Contractures

Hip Disorders in Skeletal Dysplasia

Coxa vara

TABLE Xl

No

No

No

No

No

No

No

No

No

No

No

No

No

No

No

Yes

Yes

No

No

No

No

No

No

No

No

No

No

No

Early

No

No

No

No

No

No

No

No

No

No

No

No

No

No

No

No

No

Yes

Yes

Yes

Yes

No

No

No

No

No +-

No

No

No

No

No

Rare

No

No

No

No

No

No

No

Yes

Yes

Yes

Yes

No

No

Severe

Yes

+ Yes

Yes

Yes

Severe

Yes

Yes

Yes

Late

Head acetabular incongruity

AD

AR

AD

AD

AR

AD

AR AR sev

AD

No genetic basis

Uncertain

AD

AD

AR

AD

AR

AR

AR

AR

AR

X-linked recessive

AR

AD

AD most mutation

AR

AR, also AD type

AD

X-linked tarda AD cong

Genetics

7'92

CHAPTER 9 ~ Skeletal Dysplasias

F I G U R E 16 Knee deformities are common in the skeletal dysplasias. These can involve varus, valgus, flexion, extension, and rotational deformities. (A) Anteroposterior knee radiograph demonstrates bowing. It is important to determine whether the bowing has its origin from ligamentous laxity, from contours of the joint, from the proximal medial tibial physeal region, or more rarely from the distal femur. (B) Lateral knee radiograph of a patient with pseudo-achondroplasia shows a multiplanar deformity. (C) Lateral subluxation or complete dislocation of the patella is seen in many skeletal dysplasias. This can be due to ligamentous laxity but more commonly is associated with asymmetric underdevelopment of the lateral condylar region of the epiphysis, leading to a poorly formed patellar groove. (D) Possible sites of knee deformation in skeletal dysplasia syndromes are illustrated.

achondroplasia, metaphyseal dysplasia (Schmid type), spondylometaphyseal dysplasia, and dyschondrosteosis. Genu valgum is common in Morquio's disease and hereditary multiple exostoses. Surgical treatment for bowing has involved excision of the head of the fibula, although there are no definitive results published for this procedure. The more common approach is a proximal tibial and fibular osteotomy to

correct the deformity. On occasion asymmetric stapling of a proximal tibial growth plate can be performed. Multiplanar and often asymmetric deformity characterizes some skeletal dysplasias such as pseudo-achondroplasia (Fig. 16B). Osteotomy might thus be required to correct, for example, a varus-flexion-internal rotation deformity centered at the proximal tibial metaphysis.

SECTION X 9 Orthopedic Deformities in Skeletal Dysplasias

793

D

Sites of Possible Deformation at Knee in Skeletal Dysplasias 3. Flexion; (Reverse = 4. Extension)

1. Genu Valgum; (Reverse = 2. genu varum)

Lat View

Asymmetric Physeal Involvement 9 Posterior Femur ~ Posterior Tibia

/

AP View

/ /

Posterior Capsular Contraction

Asymmetric Distal Femoral 'Growth Plate Involvement

/'/

Posterior Muscle-Tendon Contraction

Underdevelopment . of l a t e r a l epiphysis (hypoplastic condyle)

Capsular Laxity

abnormalities

Overdevelopment of medial epiphysis, including dysplasia epiphysealis hemimelica

'

" Early Osteoarthritic change (joint space narrowing)

\

5. Lateral Patellar SubluxationDislocation

(

APVie7 ii/ "

\

Fibular Underdevelopment (Overdevelopment = ? varus)

"%1

Underdevelopment of antero/lateral epiphysis. Shallow, hypoplastic patellar groove

,,\ Asymmetric Proximal Tibial Growth Plate Involvement

6. _+Rotational Deformity FIGURE 16

(continued)

2. PATELLARDISLOCATION(LATERAL) Lateral patellar dislocation is relatively frequent in the skeletal dysplasias primarily because of shaping abnormalities of the distal femoral condyles (Fig. 16C). It is fairly common in diastrophic dysplasia, multiple epiphyseal dysplasia, and spondyloepiphyseal dysplasia. Treatment is conservative or surgical as dictated by circumstances. 3. EXOSTOSES Exostoses are extremely common in the knee region in hereditary multiple exostoses due to the fact that most of the growth in the lower extremity occurs in this area. Management is by excision as warranted.

4. INSTABILITY Instability of the knee joint is seen, usually in the more severely deforming dysplasias involving the epiphyses, such as diastrophic dysplasia and multiple epiphyseal dysplasia. It is not primarily ligamentous laxity that is causative, al-

though that is one aspect, but rather asymmetric malformation of the epiphyses affecting one condyle, which can be smaller, more angled, or flatter than normal leading to imperfect articulation. Great care is required in the skeletal dysplasias in determining the site of angular deformity. Figure 16D indicates the possible sites of deformation. 5. CONTRACTURES Flexion contractures are common particularly in diastrophic dysplasia and respond best to distal femoral extension osteotomies rather than to posterior soft tissue releases.

J. Ankle Abnormalities The most common deformity of the ankle in the skeletal dysplasias is a valgus tilt with a pronated fiat foot deformity due to relative shortness of the distal fibula. Treatment is generally by a distal tibial-fibular osteotomy near skeletal maturity. On occasion, hemiepiphyseal stapling of the distal

794

CHAPTER 9 ~ Skeletal Dysplasias

medial tibia plays a role. Varus tilt of the ankle occurs less frequently, usually with relative overgrowth of the distal fibula. Distal prolongation of the fibula is an important diagnostic feature in hypochondroplasia and is also seen in some metaphyseal dysplasias.

langes and contractures of the interphalangeal joints rarely warrant intervention; the former lead to relatively little disability, and the latter cannot currently be repaired effectively.

XI. L I M B L E N G T H E N I N G

K. Foot Abnormalities Foot deformities are common in many skeletal dysplasia syndromes, particularly with diastrophic dysplasia and spondyloepiphyseal dysplasia. The deformities are usually quite rigid being connected with cartilage-bone model defects rather than ligamentous or perinatal positioning irregularities. Metatarsus adductus is frequently seen, with management varying from observation, to serial casting if rigid, and sometimes progressing to multiple metatarsal osteotomies. Clubfoot is treated in the usual fashion beginning with serial casting but frequently, as in diastrophic dysplasia, requiting soft tissue posteromedial releases and often additional calcaneal and midtarsal osteotomies or triple arthrodesis toward skeletal maturity. Toe deformities such as hallux valgus and hammer toes are also seen in these disorders. Management is by surgical correction if symptomatic. Many of the dysmorphic syndromes are accompanied by structural tarsal, metatarsal, and phalangeal anomalies, relatively few of which need surgical correction. Abnormalities include various tarsal coalitions, metatarsal widening with or without fusions, and syndactyly. Abnormalities of the lower extremities are summarized in Tables II, VIII, and XI.

L. Abnormalities of the Upper Extremities Patients with a skeletal dysplasia often lack full extension of the elbow by 10 or 15~ a limitation that is generally nonprogressive and of no clinical importance but that is helpful as a clinical diagnostic sign. The humerus is the site of significant shortening, called rhizomelic shortening, in many skeletal dysplasias including achondroplasia. Significant angular deformities of the humerus are rare, however. The radial head can be subluxed or dislocated and limits pronation and supination in some skeletal dysplasias in which ulnar growth is more limited than radial growth. The radius compensates by developing a curvature of the shaft, but if this is insufficient the head will displace. The most common disorder in which this occurs is hereditary multiple exostosis. Closed or open reduction during the growing years is usually not indicated. Excision of the radial head in adulthood is occasionally required to alleviate discomfort. At the wrist, a shortened, malformed, and dorsally dislocated distal ulna associated with dorsilateral bowing of the radius (Madelung's deformity) can decrease the range of motion and cause ulnar deviation. These deformities are common in hereditary multiple exostosis and dyschondrosteosis, and their management is discussed in the next section. Short metacarpals and pha-

Limb lengthening remains one of the potentially valuable interventions for short stature dysplasia patients but has gained far from universal patient and surgeon acceptance. With each round of technical improvements, however, better results are achieved. Both the circular Ilizarov and monolateral Wagner-Orthofix type lengthenings have been performed (20, 237). Those disorders most amenable to lengthening have normal articular cartilage and normal joints, such as achondroplasia and the metaphyseal dysplasias. Whereas lengthenings can be done in those with articular surface irregularity, contractures, and instability, the amount of lengthening achievable is usually less and the complication rate high, especially in relation to joint problems such as stiffness and subluxation. In addition, bilateral procedures are obviously required. In the lower extremities this usually means bilateral femoral and tibial lengthenings. In those patients and surgeons desiring relatively massive increases in length, two and even three separate lengthenings have been done. Although large numbers of patients have been operated, there are relatively few reports of results. Some of the best functional results occur with humeral lengthenings because upper extremity length increases of relatively shorter amounts provide great help in personal care and independent living capabilities. The procedures used are outlined in the chapter on lower extremity length discrepancies (Chapter 8).

XII. R E V I E W O F S P E C I F I C S K E L E T A L DYSPLASIAS: PATHOBIOLOGY, CLINICAL AND RADIOGRAPHIC CHARACTERISTICS, AND ORTHOPEDIC MANAGEMENT Average heights reached in adulthood are summarized in Table XII (123, 125, 277).

A. Achondroplasia 1. OVERVIEW Achondroplasia is the most common of the skeletal dysplasias. It has been recognized as a specific disorder since Parrot introduced the term in 1878, described living patients with the disorder, and even theorized that the disorder was a "dystrophie du cartilage pirimordial," a fundamental or developmental dystrophy of cartilage (230). It is autosomal dominant with about 85% of cases representing a new mutation. It shows full penetrance and full expressivity. Short stature is an invariable occurrence. The mean adult height in males was 130 cm (range = 118-142 cm) and in 214 fe-

SECTION Xil ~ Review of Specific Skeletal Dysplasias TABLE XII

795

Adult Heights in a Selected Listing o f Skeletal Dysplasias b

Disorder

Achondroplasia Hypochondroplasia Diastrophic dwarfism Pseudo-achondroplasia Metaphyseal chondrodysplasias McKusick type Schmid type Chondrodysplasia punctata Condradi-Htinermann type Chondroectodermal dysplasia Grebe chondrodysplasia Acromesomelic dysplasia Multiple epiphyseal dysplasia Fairbanks type Pycnodysotosis

Spondyloepiphyseal dysplasia tarda Spondyloepiphyseal dysplasia congenita Dyggve-Melchior-Clausen dysplasia Kniest dysplasia

Inheritance a

Adult height

Short Limb Dwarfism F: 112-138 cm ~ AD M: 118-142 cm (42-56 in.) 132-147 cm (52-59 in.) AD 86-122 cm (34-48 in.) AR 80-130 cm (32-51 in.) AD, AR AR AD

105-145 cm (41-57 in.)

AD AR AR AR

130-160 cm (51-63 in.)

106-153 cm (42-60 in.) 99-104 cm (39-41 in.) 97-123 cm (38-48 in.)

AD, AR AR

137-155 cm (54-61 in.) 130-150 cm (51-59 in.)

Mean

125 cm 130 cm 118 cm 118 cm

130-160 cm (51-63 in.)

Short Trunk Dwarfism XLR, AD, AR AD, AR AR AD

132-156 cm (52-61 in.) 84-132 cm (33-52 in.) 130-135 cm (51-53 in.) 104-145 cm (41-57 in.)

115 cm

Storage Disorders Mucopolysaccharidoses Hunter's syndrome (MPSII) Morquio's syndrome (MPSIV)

XLR AR

120-150 cm (47-59 in.) 80-120 cm (32-47 in.)

aAD, autosoml dominant; AR, autosomal recessive; XLR, X-linked recessive; XLD, X-linked dominant. bDerived from Scott CI, Jr (1988) Clinical Symposia, Ciba-Geigy 40:1-32; Horton WA, et al. (1982) Am J Dis Child 1363:316-319; and Horton WA, et al. (1978) J Pediatr 9:435-438.

males it was 125 cm (range = 112-138 cm). The mean weight is 50 kg. Achondroplasia is a clear example of a short-limbed disproportionate dwarfism. The outstretched upper extremities, more affected than the trunk or axial skeleton, reach only to the hip region. The disorder is diagnosable at birth based on clinical and plain radiographic findings. Achondroplasia was the first genetic disorder noted in which an increased incidence of mutation was associated with advanced paternal age (235,333). The sporadic mutations were studied in 40 cases, and in all 40 the achondroplasia mutation (Gl138 FGFR3) occurred on the paternal chromosome (362). This observation suggests that the mutations occur preferentially during spermatogenesis by DNA replication copy errors in the male germ line and not during oogenesis. The risk of recurrence in sporadic cases is less than 1%. The mean paternal age was 35.9 years (n = 40) in this study and 34.4 + 7.2 years (n = 348) in another.

2. CLINICAL AND RADIOGRAPHIC CHARACTERISTICS Marie described the clinical features of the fully developed adolescent and adult achondroplastic individual in 1900 (184). These characteristics include dwarfism, relatively increased head volume, and marked shortening of the upper and lower extremities with conservation of normal length of the trunk (first form of disproportion). The proximal portions of the extremity were more shortened than the distal portions, leading to a rhizomelia with the humerus more involved than the radius and ulna and the femur more involved than the tibia and fibula (second form of disproportion). Other characteristics include normal intelligence, macrocephaly (which in some instances was relative only because the head remained the same size as that of a normal person, whereas in others the enlargement was so great that it raised the question of hydrocephalus, although anatomic studies were not available to make this distinction), a skull

796

CHAPTER 9 ~ Skeletal Dysplasias

that showed both frontal and parietal bossing, a face that was relatively normal with the exception of the nose, which was shortened and flattened at its base, marked lumbar lordosis, small hands with a specific conformation involving virtually equal length of the second through fourth digits and the characteristic (trident) appearance with a separation between the third and fourth fingers with the proximal phalanges in normal position but separation of the middle and distal phalanges beginning at the proximal interphalangeal joint, bowing of the lower extremities centered at the knees, no major angular deformity of the upper extremities, relatively large epiphyseal regions with short straight diaphyses, and no specific deformations of the stemum, ribs, or clavicle (151, 190). Subsequent studies have confirmed the observations of Marie, who was the first to describe the "trident" appearance of the hand and fingers (Fig. 17A). The trunk height is normal at all ages, as determined by sitting height measurements. This was documented clearly by Nehme et al. in a study of 18 achondroplasia patients (219). The sitting height in both males and females was within normal limits with the average just below the normal average at all ages. The extremity shortening is rhizomelic with arms (humeri) shorter than forearms and thighs (femurs) shorter than legs. The hands have a characteristic "trident" appearance with widened spaces between fingers, equal length of 2nd, 3rd, and 4th fingers, and thickened digits. Many are hypotonic in the early months, and this combined with a relatively large head and shortened extremities leads to mild motor delay. There are many typical radiographic skeletal abnormalities, but relatively few of the deformities are of sufficient severity to warrant operative intervention. The long bones have shortened diaphyses, broad and flared metaphyses, and relatively enlarged and slightly irregularly shaped epiphyses, often (e.g., distal femur) with more sharply curved physes (Fig. 17B). The distal femoral physis has an inverted "V" shape into the metaphysis. Its secondary center tends to form close to the physis within the V-shaped notch. The femoral necks are short but in valgus position, and the trochanters are prominent. The fibula and clavicle are less affected than the other long bones. The fibulae tend to be longer than normal particularly distally at the ankle. The 3rd ray of the hand is more shortened than the others with the 2nd, 3rd, and 4th rays being almost the same length. Hip and pelvic radiographs from the newborn period on are characterized by a horizontal acetabulum and small sciatic, notch but hip malposition is extremely rare (Fig. 17C). The acetabular index approaches 0 ~ The major clinical deformity of the lower extremities involves a genu varum, which is seen in most patients. In relatively few, however, is it sufficiently severe to warrant operative intervention. Because the children have developed with the varus deformity from the early weeks of life, the tendency for the knees to be symptomatic in terms of discomfort and later osteoarthritis is relatively small. If intervention is warranted for the bowing, two operative ap-

proaches have been used. The most common is a proximal tibial and fibular valgus osteotomy. Once performed the tendency to recurrence is extremely low. On occasion the fibular head has been removed using the rationale that relative overgrowth of the fibula proximally predisposes one to the varus deformity by limiting growth in the proximal tibia. No published series is available to show results of this intervention. Our own approach is to favor the proximal tibial and fibular valgus osteotomy, although relatively few have been needed in our hospital's achondroplasia population. The mild deformities of the upper extremities rarely warrant intervention in the vast majority of patients. The lack of full elbow extension is rarely of clinical concern. 3. MORBIDITY AND MORTALITY IN ACHONDROPLASIA

The high incidence of achondroplasia has allowed for detailed studies from several centers concerning associated medical and neurological problems. There is a considerable morbidity and occasional mortality with this disorder such that these patients must be followed closely from birth (374). Possible problems include stenosis of the foramen magnum and cervical spinal canal causing cervicomedullary compression with risk of death, central apnea, and neurological dysfunction including paraparesis, respiratory difficulty as a result of central apnea, a small rib cage and upper airway obstruction, gross motor delay associated with macrocephaly, hypotonia, and joint laxity, frequent otitis media, which may be accompanied by hearing loss and delay of speech, genu varum, hydrocephalus, and later onset neurological complaints due to lumbar spinal stenosis. A detailed review of 701 persons with achondroplasia by Hecht et al. showed an increased mortality ratio for those with the disorder with the rate more than double that of the general population (106). The increase in mortality was particularly apparent in early childhood. Sudden death accounted for excess deaths in those less than 4 years of age, with brain stem compression identified as the cause in half of these. The central nervous system and respiratory causes were not significantly increased but did account for half of the deaths in those from 5 to 24 years of age. In the older age groups severe disability resulting from marked spinal canal stenosis was noted. The problematic axial bony abnormalities associated with achondroplasia are stenosis of the foramen magnum and thoracolumbar spinal canal stenosis. The greatest threat to life occurred in the first 5 years of age and particularly during the first year with the problem being sudden death. Most of these were felt to be due to compression of the brain stem and/or upper cervical spinal cord due to stenosis of the foramen magnum. A more recent detailed study by Hunter et al. of 193 patients from several centers reviewed complications of the disorder (128). Otitis media had been reported in 90% of all children by the age of 2 years with 80% having undergone the insertion of ear tubes by 10 years of age. Middle ear infection had contributed to speech delay in 20% of children

F I G U R E 17 Radiographic abnormalities in achondroplasia are shown. (A) Trident appearance of the hand in achondroplasia. There is a relatively shortened hand with equal lengths of the second through fourth digits and a characteristic separation between the third and fourth fingers with the proximal phalanges in normal position, but separation of the middle and distal phalanges beginning at the proximal interphalangeal joint. (B) The long bones are characterized by shortened diaphyses, broad and flared metaphyses, and relatively enlarged and slightly irregular epiphyses, with the distal femur showing a more sharply curved and inverted "V" physis. (Ci, Cii) Hip and pelvic radiographs of a patient with achondroplasia show the horizontal acetabulae, flared metaphyses of the femoral neck, the short sharp sciatic notch, and the (Ciii) narrowed interpedicular distance of the lower lumbar and sacral vertebrae. (Di-Diii) Thoracolumbar kyphosis is quite common in achondroplasia in the first and second years of life. Lateral radiograph shows the spontaneous correction of the kyphosis over a 4-year period. No treatment was used.

798

CHAPTER 9 9 Skeletal Dysplasias

and to conductive hearing loss in 38% of adults. Dental malocclusion due to midface hypoplasia led to associated orthodontic problems. More serious respiratory problems can contribute to major illness in some achondroplastic patients. A study by Tasker et al. separated infants presenting to the respiratory unit with achondroplasia into three problem groups (329). The mildest were in group 1 with obstructive sleep apnea, which responded well to nasal steroid therapy supplemented when necessary by removal of the adenoids and tonsils. Those in group 2 had sleep apnea plus hydrocephalus resulting from jugular foramen stenosis, whereas the most severe group 3 children actually developed cor pulmonale in which 3 of 5 died. The final group of problems are neurologic, ranging from sudden death or progressive paraparesis to milder lower extremity problems relating to weakness and fatigability (99, 105, 106, 260). Smallness of the foramen magnum as well as cervical canal spinal stenosis can lead to central apnea, hydrocephalus, and cervical spinal cord pressure. 4. MACROCEPHALY--HYDROCEPHALUS IN ACHONDROPLASIA

Although the head appears large in achondroplasia, most often this macrocephaly or megalencephaly is not indicative of a serious underlying problem. Hydrocephalus involving head enlargement with mild dilatation of the ventricles (ventriculomegaly) is a consistent finding in achondroplasia but is rarely symptomatic. It appears to be secondary to raised intracranial venous pressure due to back pressure which is hemodynamically significant from stenosis of the jugular foramen. The head growth curves are usually above the 97th percentile but parallel the normal slope for age. Synostosis of the skull seen in achondroplasia narrows and distorts the foramen magnum, which causes the hydrocephalus-like condition. Most patients with achondroplasia, if studied, will show this phenomenon. There is, however, a great difference in neurosurgical centers as to whether this requires treatment by shunting. Although there are many reports of surgery, the issue does not appear to have been settled definitively. Many clinics seeing these patients only observe. Increasingly it is felt that hydrocephalus does not occur to a degree that requires shunting in most cases of achondroplasia. When hydrocephalus occurs it is almost always communicating and appears to relate to venous outflow at the narrowed jugular foramen (319). Shunting is done only in symptomatic cases in which increased intracranial pressure, increasing head size, and neurologic changes have been documented. 5. CRANIOCERVICAL JUNCTION ABNORMALITIES IN ACHONDROPLASIA

Stenosis of the foramen magnum is an invariable finding in achondroplasia. It is well-documented including by CT studies of large numbers of patients from birth to skeletal maturity and beyond, which establish means and standard deviations in achondroplasia as well as comparisons with the

normal population. Hecht et al. documented 157 normal individuals in relation to the size and shape of the foramen magnum, including both transverse and sagittal measurements, and compared them with measurements, from 154 achondroplastic individuals (107). The achondroplastic foramen magnum is small at birth, and during the first year it has a severely impaired rate of growth especially in the transverse dimension. This diminished growth results not only from abnormal endochondral bone growth but also from premature fusion of the synchondroses. The mean size of the adult achondroplastic foramen magnum is that of a nonachondroplastic individual at birth in the transverse dimension and in a 24-month-old child in the sagittal dimension. This study, however, did not show any detectable difference in statistical significance in achondroplastic patients who were clinically normal from those who were experiencing cervicomedullary compression syndromes, although there was a tendency for those with narrower dimensions to have greater clinical problems. It would appear that many if not all cases of achondroplasia suffering from sudden death, usually in the first few years of life, represent examples of cervicomedullary compression due to tightness of the foramen magnum and adjacent upper cervical canal. Symptomatic patients have been helped by surgical decompression of the skull base, although these procedures themselves have a considerable morbidity and as yet there are no definitive standards indicating which patients should undergo surgery. The surgery involves posterior fossa decompression and laminectomy of the posterior body of C 1. In a study of 6 patients by Ryken and Menezes, symptoms warranting operative intervention included occipitocervical pain, ataxia, incontinence, apnea, and respiratory arrest (272). Detailed studies involved plain radiographs with flexion and extension views, CT scanning, and MR imaging. Typical findings included marked foramen magnum stenosis, ventrolateral cervicomedullary junction compression secondary to central and paramesial basilar invagination, and dorsal cervicomedullary junction compression secondary to ligamentous hypertrophy and invagination of the posterior arch of the atlas. It is evident that surgical decompression is not warranted based on the presence of foramen magnum stenosis alone because essentially all patients with achondroplasia have that finding and only 5-10% develop symptoms. Careful exam is essential for symptoms, which include quadriparesis, paraparesis, dysphasia, poor head control that fails to improve with time, hypotonia, and delayed motor developmental milestones. Respiratory difficulties can include apnea spells and cyanosis. Other problems include occipitocervical pain, myelopathy with ataxia, incontinence, spasticity, and hyperreflexia. Because the craniocervical junctional stenosis is a potentially lethal problem in a subset of young infants with achondroplasia prospective studies have been helpful. Pauli et al. evaluated and followed an unselected consecutive series of infants with achondroplasia to assess the development of problems at this area (233). Of 53 prospectively studied in-

SECTION Xll ~ Review of Specific Skeletal Dysplasias fants, 5 were judged to have sufficient craniocervical junction compression to require cervical decompression, a rate of 9%. All children showed subsequent marked improvement of neurological function. With findings of high cervicomyelopathy, the most frequently seen clinical abnormalities included lower limb hyperreflexia or clonus on clinical exam, central hypopnea demonstrated by polysomnography, and foramen magnum measurements below the achondroplastic mean. Hecht et al. showed that the excess risk of death in infants with achondroplasia may approach 7.5% largely because of cervical cord compression (106). In a larger group studied there were 10 (13.3%) infants judged to be in need of suboccipital decompressive surgery including 5 (9.4%) of those studied prospectively. Surgery in the prospective group was at a mean age of 10.7 months (range = 8.3-12.8 months), whereas in the overall group the age was 24.0 months (range = 7.6-52 months). In every instance, at surgery marked abnormalities of the upper cervical cord were demonstrated involving indentation, deformation, marked compression, and ribboning of the cord. In two instances fibrous constriction bands were also identified. Pauli et al. also stressed the importance of finding multiple clinical and radiographic irregularities prior to surgery rather than intervening based on radiographic abnormalities alone. Infants who underwent decompressive surgery had multiple features of cervical myelopathy including hypotonia, weakness, asymmetric motor signs, and hyperreflexia. The three most important variables were considered to be hyperreflexia or clonus, foramen magnum measurements below the mean for achondroplastic patients, and central hypopnea. The mortality rate in achondroplasia is slowly becoming known although there is considerable variability in separate studies. Rimoin indicated that mortality was less in his group than in those of other centers even though his group was less aggressive in performing occipital-cervical compression (260). Much, of course, is dependent on the understanding of the natural history of achondroplasia. The mortality rate estimates in achondroplasia have ranged from 2.7 to 7.5%, with almost all of these deaths occurring in the first few years of life and considered to be due to cranial-cervical stenosis. Rimoin states, however, that even young patients with neurologic abnormalities such as hypotonia, clonus, and hyperreflexia will eventually attain normal motor development and become normal neurologically if left alone. Because the average age of the 5 patients operated on by Pauli was 10.7 months the implication is clear that many of these would have gone on to normal development if left unoperated. Rimoin felt that the two MRI features most indicative of the need for surgery are "a lack of CSF flow both anteriorly and posteriorly and intracord lesions as indicated by abnormal cord signal on T-2 weighted imaging." He felt that those patients with decreased space around the cord at the cervicomedullary junction, occipital bone impingement on the posterior cord, and even lack of cerebrospinal fluid (CSF) flow posteriorly developed normally with no neurological

799

sequelae if left alone. Although the posterior decompressive surgery can be done safely, Rimoin indicated that there were a considerable number of postoperative problems in historical series that could not be overlooked. Although surgery was needed on occasion, his feeling was that only progressive worsening of neurological symptoms to a "severe" stage especially those who were very floppy with significant truncal and limb hypotonia, significant apnea, or evidence on MRI of a total lack of CSF flow anteriorly and posteriorly, and/or intracord lesions on T-2 wave imaging warranted the intervention. 6. DEVELOPMENTALMILESTONES IN ACHONDROPLASIA These are becoming increasingly well-documented as groups study achondroplasia patients specifically. The achondroplastic patient must be followed very carefully in terms of medical and neurological development in the early months and years of life. The Committee on Genetics of the American Academy of Pediatrics has published supervision guidelines for children with achondroplasia (55). They indicate that unexpected infant death occurs in less than 3% of those affected and usually only in the most severe cases. Although motor development is somewhat slow, it should improve progressively. Many achondroplastic infants do not sit without support until 9-12 months of age nor ambulate until a period of time between 1.5 and 2 years of age. Rimoin indicates, however, that almost all achondroplastic children will gain normal strength and muscle tone and catch up on motor skills by 2-3 years of age. 7. SPINAL ABNORMALITIESIN ACHONDROPLASIA a. Cervicothoracic Spinal Stenosis. Although stenosis of the foramen magnum and lower lumbar spinal canal are recognized to occur in all achondroplastic patients, there is considerable evidence that cervical and thoracic stenosis is also present. Clinical compression sequelae are infrequent at these levels compared to the others, but they do occur. There are reports of surgical decompression in achondroplasia at multiple levels for neurologic cord-root compression. Uematsu et al. reported on complete craniospinal posterior bony decompression (brain stem to cauda equina) in 7 patients (341). Adult age group lumbosacral decompression-laminectomies are well-known. An extensive thoracolumbar decompression has been reported based on the clinical areas of symptoms and signs in a patient as young as 7 months. In that case, cord compression was directly due to a thickened layer of fibrous tissue such that laminectomy with resection of the constricting soft tissue was needed. b. Narrowed Interpedicular Distance, L1-S1. Interpedicular narrowing in achondroplasia from L 1 to S 1 represents a spinal stenosis that can predispose one to relatively rapidly developing neurological symptoms in adult life should the patient have a herniated or bulging disk or an arthritic spine. The progressively narrowed interpedicular distance can be

800

CHAPTER 9 9 Skeletal Dysplasias

seen on the anteroposterior spine film even in the newborn. In the normal spine the interpedicular distance widens with each progressively lower vertebra from L1 to S 1, whereas in achondroplasia the distance becomes progressively narrower particularly from L3 to S 1 (Fig. 10A). On the lateral lumbar film in achondroplasia the posterior border of the vertebral bodies is concave and the pedicles are short as well (Fig. 10B). Caffey has estimated that the achondroplastic pedicles can be as short as only one-half the normal length (42). This lumbar spinal stenosis is infrequently a problem in the childhood or adolescent age group but tends to lead to symptoms when the individuals are in their early to mid-adult years. The sacrum articulates with the fifth lumbar vertebra lower than in the normal spine in relation to the adjacent iliac bones. Associated with the spinal stenosis is a lumbar lordosis present also from the early years of life. The lordosis is structural to a great extent. The lumbosacral angle is increased in the newborn and the plane of the sacrum becomes nearly horizontal after walking. This is not progressive. Efforts have been made to treat this with an antilordotic spinal orthosis. Although the brace has been used on occasion in many centers, no report of its effectiveness has appeared. The brace would have to be worn several hours per day for several years to have any chance for a definitive effect. Lutter and Langer studied neurological symptoms in achondroplastic dwarfs severe enough to require surgical treatment (179). Problems were due to distal neurological deficits with 27 cases of the cauda equina syndrome, impending paraplegia, or actual paraplegia. The average age of onset of symptoms in 14 patients was 38 years. All were adults ranging from 23 to 68 years of age. The space for the neural elements already limited by the spinal stenosis with shortened pedicles and a narrowed interpedicular distance was further reduced by a combination of herniation of the intervertebral disk, degenerative arthritis with the development of osteophytes, and anterior wedging of one or more vertebral bodies concentrated from T 11 to L2. The small size of the spinal canal, however, was the principal cause of the neurologic problems. In general, there was a progressive slow onset of the disorder beginning with discomfort and later altered sensations and motor and sensory deficits. Intermittent claudication with pain in the back and limbs was noted frequently. Only rarely was a formal nerve compression syndrome or acute paraparesis seen. Kyphosis was commonly associated with spinal cord and nerve root damage due to increased mechanical compression. Lutter et al. have documented the diminution in the size of the spinal canal (180). Postmortem studies documented a 39% decrease in the transverse area of the first lumbar level and a 27% decrease at the fifth compared to normal controls. c. Thoracolumbar Kyphosis. Another concern in the developing spine relates to the thoracolumbar kyphosis, which is seen frequently in individuals in the first and second years of life (Fig. 17D). There is a marked tendency for this kyphosis to improve spontaneously as the individuals begin

standing on their own and walking (Figs. 17Di-17Diii). Radiographs may show a slight decrease in the height of the anterior vertebral bodies at the kyphotic site on lateral views, but this also tends to self-correct and we have rarely found bracing necessary. In those relatively few patients in whom the thoracolumbar kyphosis persists, neurological complications tend to occur earlier and be more marked due to the associated spinal stenosis. Lutter et al. studied 25 patients with achondroplasia who had reached a minimum of 5 years of age, and in 22 of these standing lateral radiographs showed 15 having a thoracolumbar kyphosis associated with one or more hypoplastic or wedged vertebral bodies (180). The kyphosis came first with subsequent maldevelopment of the bodies secondary. On occasion the kyphosis was marked, with 5 patients showing deformity from 100 to 132 ~ Rare instances of posterior laminectomy and soft tissue decompression in childhood have been described. d. Scoliotic Deformity. Spinal alignment was carefully studied in 94 achondroplastic patients 15 years of age or older with 53 from a series of Langer and 41 from a series of Bailey (6). Scoliosis was centered at the thoracolumbar junction and into the lumbar region. It was defined as mild, less than 20 ~, in 25 patients and moderate, 20-45 ~, in 7. There were no severe curves. Kyphoscoliosis was not as common, with 4 mild deformities, 1 moderate deformity, and 6 with moderate and localized scoliosis in whom, however, the kyphosis was moderate to severe in degree. There were 22 patients with kyphosis, 8 of which were mild, 8 moderate, and 6 severe. The greatest amount of wedging was at the T12 or L1 vertebra with some adjacent vertebrae also slightly wedged anteriorly. On occasion the kyphosis was quite localized with a severe gibbus involving 1-3 vertebrae. 8. OTHER ORTHOPEDIC ASPECTS IN ACHONDROPLASIA One of the most detailed studies of the orthopedic aspects of achondroplasia was performed by Bailey, who reviewed 63 patients studied clinically and 87 studied radiographically (6). There were a large number of achondroplastic patients in whom lower extremity pain and weakness were due to an excessive number of discogenic problems in relation to lumbar spinal stenosis. Bailey felt that serious neurological difficulties would be seen in slightly more than 10% of patients almost exclusively due to the spinal canal stenosis. Bilateral hip flexion contractures of mild to moderate extent were noted in most patients; the extent documented in children from 0 to 12 years of age was an average of 12.5 ~ 13-21 years of age 23.5 ~, and adults over 21 years of age 33 ~. Patellar subluxation or dislocation was not a problem, but the majority of patients had a genu varum deformity. Surgery was rarely required for these mild deviations from the norm. There were no instances of coxa vara, whereas coxa valga was quite constant being present in almost all patients with an average measurement of 162 ~. The femoral head was well-located in the acetabulum, however, as the neck was

SECTION Xll ~ Review o f Specific Skeletal Dysplasias

short. There was no tendency to arthritic changes in the joints in adulthood, although arthrosis was present in cervical and lumbar vertebra. Although hip contractures were documented, they rarely appeared to be of clinical significance to the patient and surgical intervention was not reported. The skeletal features of the spine predisposing one to premature neurological problems related to the occurrence of a herniated nucleus pulposus, exaggerated lumbar lordosis, narrowed spinal canal, articular process arthritis with osteophyte formation, a recessed vertebral body, slight malalignment of the vertebral column, and possibly constriction of the formen magnum. The most common reason for orthopedic treatment in achondroplasia related to the bowleg or genu varum deformity. Braces are of no benefit and even if attempted are poorly tolerated. The failure to document adult-onset arthritis even with the bowleg deformity mandates caution in treating the varus in childhood unless it is symptomatic. Kopits reviewed 158 patients with achondroplasia with no cases of odontoid hypoplasia or congenital hip dysplasia and only 1 case of scoliosis of 40 ~ (154). Osteoarthritis of the hip or knee was seen only occasionally in spite of the considerable bowleg deformity in several.

B. Hypochondroplasia Hypochondroplasia is a distinct clinical disorder with similarities to achondroplasia, although it is much milder and more variable. It is autosomal dominant with a mild rhizomelic pattern. At a molecular level it is part of the F G F R 3 gene abnormality family, which includes thanatophoric dysplasia and achondroplasia (22, 23). It is differentiated from achondroplasia by a lack of craniofacial involvement, with a normal appearing nose, relatively greater height, and milder features in the pelvis and spine (90, 93, 306). Radiographs of skull, hands, and feet are normal. The hands may be short but there is no trident configuration. There is a lower than normal articulation of the sacrum on the iliac bones. Hydrocephalus and spinal stenosis syndromes in childhood or the adult years have not been features. In the spine changes are limited to the lumbar vertebrae. The interpedicular distance does not widen from L1 to S 1, tending to remain the same without the distal narrowing seen in achondroplasia. The vertebral canal is also narrowed in the lateral view with a concave shape to the posterior borders of the lumbar vertebrae because the pedicles are somewhat shorter than normal. Mild lumbar lordosis is present. The fibula is relatively long, leading to altered relations at both knee and ankle. The distal fibula tends to be longer than normal with some varus tilt to the ankle, and the proximal fibula is also longer, in some reaching to the midtibial epiphyseal region. Bowing at the knee is infrequent. The elbows lack full extension and the fingertips do not reach the mid-thigh region. The ulna tends to be somewhat shortened distally although the styloid process is longer than normal. The femoral neck is shortened but

801

there is no coxa vara. The pelvis is virtually normal, the acetabulae are not horizontal, and there is only slight reduction in the size of the sciatic notch. Bowing of the radius and short, broad long bones especially in the metaphyseal regions are seen. Muscle insertion regions of the deltoid at the proximal humerus are prominent. Deformities requiring orthopedic management and neurological sequelae of the mild structural spinal abnormalities are virtually never seen. Adult height ranges from 127 to 152 cm.

C. Multiple Epiphyseal Dysplasia In classic multiple epiphyseal dysplasia the changes in the long bones are limited to the epiphyses, and any spinal involvement is absent or sufficiently mild that it does not affect the structure or height of the vertebral bodies in any clinically apparent way. There are many phenotypic variants of this disorder to the point that virtually no two families have the same pattern of involvement, and variability can occur even between different members of the same family. For this reason, subcategorizations into multiple "types" by radiographic criteria seem fruitless. Most are autosomal dominant. On occasion, joint degeneration is great, whereas stature is minimally affected with patient height normal or low-normal only. Many of these are now found to represent abnormalities of type IX collagen (36, 229, 346). Molecular definition should both explain and perhaps render unnecessary the radiographic-clinical categorizations with multiple subclasses. Because virtually all families differ from one another, many differing molecular changes seem certain. Two of the primary early descriptions of MED led to clinical terminology defining the Ribbing or Fairbank types. In the Ribbing MED, abnormality is almost entirely confined to the proximal femoral epiphyses with short stature and early onset of hip osteoarthritis (259). In the Fairbank MED, involvement is more widespread involving hips, knees, ankles, hands, and other areas (73, 74, 316). Due to the extreme variability of involvement, however, many other patterns of epiphyseal affection are seen. Bailey divided multiple epiphyseal dysplasia into congenita and tarda types with six MEDC variants and eight MEDT variants (with many subtypes) (8). Van Mourik et al. have described a family with an absence of hip or shoulder involvement despite widespread involvement at the knees, elbows, ankles, and hands (346). Other patterns include only lower extremity involvement or only hip involvement. The main determinants of the MED diagnosis involve epiphyseal irregularities on plain radiographs. Suggestive factors for early diagnosis include stippled epiphyses in the newborn (although this is infrequent) and abnormalities of development of the secondary ossification centers, including delayed appearance, fragmented structure, eccentric positioning, subchondral flattening, and smaller size than expected for age. Early clinical presentation, often preceding plain radiographic changes, involves transient episodes of

802

CHAPTER 9 9

Skeletal Dysplasias

FIGURE 18 Multipleepiphyseal dysplasia is characterized by epiphyseal bone irregularity and flattening in the subchondral region leading to premature osteoarthritis. (A) A hip arthrogram in a patient with MED is shown. The femoral head ossification center is irregular in shape, but the cartilage model is still spherical. Examples are shown of progressive involvement of the knee in (Bi-Biv). (Bi) The secondary ossification center of the distal femur is irregular in shape in an 8-year-old child. (Bii-Biv) Progressively worse osteoarthritis in adult members of the same family who also had the type IX mutation is shown. (C) Irregular ankle joint development in a 10-year-old male family member is shown. [Parts Bi-C reprinted from (36).]

joint discomfort, swelling, and decreased range of motion. Hips (Figs. 15Ci-15Ciii and 18A), knees (Figs. 18Bi18Biv), and ankles (Fig. 18C) are the most c o m m o n clinical sites of problems with elbows, hands, and shoulders not frequently involved. Radiographic features in the childhood hip are bilateral, symmetrical, and similar to the fragmentation stage of Perthes disease (62). A major point in differentiation

from Perthes is the fact that Perthes (1) shows either unilateral involvement (80%) or asymmetrical temporal involvement if bilateral with the two sides never equally affected at any one time, (2) has metaphyseal cystic involvement, and (3) generally has normal acetabulae at least early on. The proximal femoral head secondary centers in M E D are always symmetrically affected without cystic changes in the me-

SECTION Xll ~ Review of Specific Skeletal Dysplasias taphyses and with dysplastic acetabulae seen often early in the first decade. The changes in Perthes rapidly worsen and then improve, whereas in MED the changes progress slowly. A skeletal survey can remove many concerns about diagnosis because other major joints will usually show some subtle abnormalities in MED but are fully normal in Legg-Perthes. The symptoms of pain and decreased range of motion are rarely seen in multiple epiphyseal dysplasia at this age, whereas they usually represent the presenting phenomena in Legg-Perthes disease. Knee changes are characterized by less distinct shaping of the secondary ossification centers with relative flattening of the distal femoral condylar contours, a shallow patellar groove on the anterior epiphyseal surface, and a tendency to lateral patellar subluxation or dislocation. It is likely that many patients diagnosed with primary osteoarthritis of the hip or knee represent unrecognized cases of multiple epiphyseal dysplasia. Involvement can also include shoulder, elbow, and wrist joints, irregularly shaped tarsal and carpal bones; and shortening of the fingers due to metacarpal and phalangeal involvement. There is much to be said for returning clinically to a more simple diagnosis of multiple epiphyseal dysplasia describing the presence and extent of any short stature and listing involved joints. Short stature is variable from severe to mild to absent; the height of many is within the normal range. The skull and face are normal. Spinal changes are absent or extremely mild, and radiographic changes should be limited to the epiphyses. The hands are often somewhat short with broad fingers. The major orthopedic problem in multiple epiphyseal dysplasia, other than the short stature, is the high degree of localization at the hip and knee leading to premature osteoarthritis and the need for total joint arthroplasty in mid-adult life. The disorder leads to abnormalities in the shape of the femoral head, distal femur, and proximal tibia leading to imperfect joint contours, collapse of the subchondral bone, and associated irregularities of the articular cartilage surface. In the growing child the disorder may often present initially with a mild acetabular dysplasia. The abnormal shape of the femoral head tends to become apparent late in the first or early in the second decade (Figs. 15Cii and 15Ciii). The radiographic appearance of the hips in the first decade can be almost indistinguishable at first glance from a Legg-Perthes disorder as noted earlier with fragmentation of the developing secondary ossification center of the femoral head. Ingram has indicated that early diagnosis for MED can be enhanced by analyzing wrist radiographs for carpal and metacarpal measurements and anteroposterior films of the distal femur (132). These films allow for measurement of various indices such that 80% of those with MED can be distinguished from normal with good confidence between approximately 5 and 13 years of age. Efforts have been made to forestall hip osteoarthritis when the disorder is accompanied by an acetabular dysplasia by performing a Salter-type innominate osteotomy during the first decade, but no results are reported. Such studies

803

indeed will be very difficult to do because the variability of joint involvement in patients with multiple epiphyseal dysplasia is great even within the same family. Degenerative cartilage change with varus deformation can also be seen at the knee, and valgus osteotomy is generally the recommended approach. On occasion ankle deformities also warrant distal tibial and fibular supramalleolar corrective osteotomies. There are no specific spinal abnormalities with multiple epiphyseal dysplasia, although as with many of the skeletal dysplasias instances of C1-C2 subluxation and thoracolumbar scoliosis can be seen. Lumbar lordosis is not characteristic in multiple epiphyseal dysplasia.

D. Dysplasia Epiphysealis Hemimelica This disorder is characterized by localized overgrowth of an epiphysis of a long bone or of a carpal or tarsal bone, which subsequently leads to joint malalignment (Fig. 3B). It is not hereditary. The disorder is more common in boys than in girls by a 3:1 margin, and lower extremity bones are much more frequently involved than upper extremity bones. In most instances one-half or less of an epiphyseal region is involved (hemimelica), and the medial aspect of the bone is more often involved than the lateral. The most common areas of involvement are the distal tibial and distal femoral epiphyses in the long bones and the talus. Other sites described include the proximal femoral capital epiphysis, acetabulum, patella, proximal tibia, calcaneus, tarsal navicular and first cuneiform, carpal bones, distal ulna (39), proximal and distal radius, proximal humerus, distal humerus (rarely), glenoid of the scapula, and occasional instances at the metacarpal and phalangeal bones (56, 224, 252). A 1994 article noted that approximately 135 cases had been reported in the world literature. In approximately two-thirds of cases multiple lesions are present in an involved limb. A rare instance of involvement of almost all of the epiphyses of the left side of the body, including complete involvement of the epiphyses affected, was reported by Cruz-Conde et al., representing the most marked manifestation of the disorder reported (63). The overgrowth initially affects part of the peripheral cartilage mass of the epiphysis, although diagnosis is generally brought to light by associated angular deformity and radiographic appearance of accessory and eccentric centers of ossification within the cartilage mass. The disorder has been likened to an osteochondroma affecting the epiphysis rather than the metaphysis. Histologic sections show a relatively normal appearing cartilage tissue, which undergoes endochondral ossification in association with mineralization of the cartilage matrix, invasion of vascularity, deposition of bone on cartilage cores, and formation of a marrow. The clinical problems with the disorder relate to the asymmetric cartilage surface of the joints and angular deformity. The initial report is generally attributed to Mouchet and Belot in tarsal bone in 1926 (209). The term tarsoepiphyseal aclasis was used by Trevor, although the generally accepted term is dysplasia epiphysealis hemimelica. (340)

804

CHAPTER 9 ~ Skeletal Dysplasias

With time the bone from the overgrowth regions merges with that synthesized by the normal ossification processes of the secondary center of the rest of the epiphysis. Surgical treatment is indicated when the lesion causes discomfort, altered function, or deformity. Excision even if incomplete can lead to good results. Osteotomy of the adjacent bone might be needed. Asymptomatic lesions are observed because there is no known risk of malignant transformation. Magnetic resonance imaging is now used to outline the extent of the cartilage mass, to more accurately note the developing secondary ossification regions, and to define the region of differentiation of normal epiphyseal cartilage and the affected hemimelic overgrowth. Keret et al. defined a definite cleavage plane between the epiphysis and the abnormal cartilage by MR imaging (146). MR also can define the articular cartilage shape of the affected joint (also shown by arthrogram) and can assess the physis itself, which if contiguous with the hemimelica lesion might provide a reason for the continuing abnormal growth following osteotomy.

E. Metatropic Dysplasia Metatropic dysplasia is a rare disorder with a specific developmental pattern: newborn appearance similar to achondroplasia (short limb dwarfism) with subsequent reversal to a Morquiolike (short trunk dwarfism) appearance in the first few years of life. The vertebrae show platyspondyly at birth but no deformity. There is delay in ossification of the vertebral bodies. With time a severe kyphoscoliosis develops. Atlantoaxial instability is common. The limbs are short and havea characteristic widened metaphyseal appearance (dumbbell shape) with oval cortical outlining. "Knobby" knees reflect this. The acetabulae are initially horizontal with a lateral supra-acetabular notch. A prominent coccygeal skin fold often is seen.

F. Kniest Dysplasia Kniest dysplasia is evident at birth with disproportionately short stature, short limbs, and a prominent, short, barrelshaped chest (15, 313). A depressed nasal bridge and cleft palate are common. The joints tend to be enlarged with limited mobility and the limbs are short and bowed. Hips often are malformed with short necks, coxa vara, malformed heads, and acetabular dysplasia. There is a high incidence of hearing loss. Eye problems are common and severe with myopia, retinal detachment, glaucoma, and eventual blindness in many. Lumbar kyphoscoliosis tends to develop in the middle of the first decade and usually will progress rapidly to a rigid deformity. The kyphotic element is usually more prominent than the scoliosis. It can be high or midthoracic and tends to be localized to a few vertebrae. The incidence of osteoarthritis, even in the second decade, is fairly common with this disorder and thus more marked than in most skeletal dysplasias with the exception of diastrophic dysplasia. It is present in virtually all joints including the hands.

G. Chondrodysplasia Punctata The benign Conradi-Htinermann type of chondrodysplasia punctata is autosomal dominant in comparison to the more severe and usually lethal recessive rhizomelic variant. The appearance of the stippled lesions in the various subtypes is the same radiographically in the first few months of life so that prognosis is guarded depending on the general condition. Even early resorption of some of the punctate areas is not definitive evidence that systemic recovery will occur. The benign autosomal dominant variant refers to those patients with stippled epiphyses but without rhizomelia who survive with only mild to moderate growth deficiency, which is usually unilateral in the limbs but may affect vertebral growth in an asymmetric fashion. The excess areas of calcification are gradually incorporated into the bone or resorbed in the first few years of life, and the radiographic appearance becomes less evident and disappears with development. Orthopedic problems are fairly frequent involving kyphoscoliosis with structural irregularities of the vertebral bodies and asymmetric limb shortening with coxa vara or angular deformity. The patients should be examined for cataracts, which can occur as high as 17% of the time, and for ichthyosiform skin patterns. These two findings are much more common in the lethal rhizomelic form. An X-linked dominant form described by Happle affects females and is accompanied by widespread atrophic and pigmented skin lesions. Distinction must be made between the three specific forms of chondrodysplasia punctata and stippling as an epiphenomenon in such disorders as multiple epiphyseal dysplasia, spondyloepiphyseal dysplasia, hypoparathyroidism, trisomy 18 and 21, fetal warfarin syndrome (83), and maternal alcohol (first trimester) ingestion (247). The more severe the cartilage abnormality, the earlier in fetal life the occurrence of the stippling calcification and the more severe the eventual outcome. Jeune et al. felt that involvement concentrated in the distal femurs and also proximal humeri was a relatively poor prognostic sign, whereas those patients with primarily hand and foot involvement would do better (139). These findings also correlate with poorer prognosis with rhizometric (humeral-femoral) shortening (Fig. 2D).

H. Spondyloepimetaphyseal Dysplasia This disorder has gradually come to be defined as a separate entity. It is characterized by marked shortness of stature, gross joint laxity, and progressive kyphoscoliosis (17). The pelvis is characterized by large iliac wings with small hypoplastic iliac bodies, poorly formed acetabula, and smaller to absent capital femoral secondary bone centers of the epiphyses. The hips often proceed to dislocation, and the secondary ossification centers remain small, irregular, and flattened. The femoral necks are short with the coxa valga deformity, and skeletal age is delayed. Beighton and Kozlowski, among the early definers of this disorder, indicate that "epi-

SECTION XII ~ Review of Specific Skeletal Dysplasias metaphyseal changes are obvious, but slight. The metaphyses are widened with an abnormal trabecular pattern" (17). The kyphoscoliosis is generally absent or minimal in the newborn, but deformity increases rapidly during the first decade of childhood. The spinal deformity has led in severe cases to paraplegia, progressive respiratory failure, cor pulmonale, and even death in the first two decades (367). Early attention to the spinal deformity including surgical stabilization appears to be essential.

I. Diastrophic Dysplasia Diastrophic dysplasia is diagnosable at birth (7). It is an autosomal recessive disorder that was recognized as a separate, distinct skeletal dysplasia and called diastrophic (Greek = tortuous, twisted) in 1960 by Lamy and Maroteaux in a study of 3 of their own patients and 11 previously described in the literature (161). The clinical picture includes hip, knee, and elbow flexion contractures, rigid equinovarus or metatarsus adductus foot deformities, "cauliflower" ear lobes swollen by fluid due to perichondritis of the external ear cartilage, stiff fingers, and a short, abducted "hitchhiker's" thumb (114, 163, 322, 331,349). The first metacarpal is short and positioned more proximally than normal and the thumbs are hypermobile. The ears are normal at birth but develop the swelling from 1 to 6 weeks of age. If not drained, the fluid eventually resorbs but the pinnae remain thickened and deformed and may develop calcification. The joints are rigid but hypermobility is seen in some. The limbs are shortened in a mesomelic fashion causing a disproportionately short stature. Kyphoscoliosis is common in the first decade, and cleft palate can occur. Diastrophic dysplasia is one of the most difficult of the dysplasias to manage from an orthopedic viewpoint for several reasons. The shortness of stature is great, with most patients not destined to reach even the 4-ft level. This is due to the severity of epiphyseal involvement, worsening of the shortness by contractures of hip, knee, and ankle that accompany the disorder, premature cessation of physeal function around the age of 10 or 11 years, and an often progressive kyphoscoliosis. Contractures are present at virtually every joint in the upper and lower extremities. These tend to worsen with time and to be relatively rigid such that soft tissue surgery is often ineffective, necessitating osteotomy. A characteristic finding in many, particularly at the knee and elbow, is associated subluxation due to cartilaginous epiphyseal deformity. Even when corrected the tendency is for the contractures to recur in relentless fashion. Angular deformity of the long bones is not common, with most limb deformity due to contractures and subluxations. Early and relatively extensive surgery is needed to obtain correction. Attention to postsurgery detail is helpful in efforts to minimize recurrence, but if growth remains some recurrence is common. The characteristic contractures involve flexion deformities of the hip, knee flexion contractures, and

805

a rigid talipes equinovarus (clubfoot) or metatarsus adductus deformity. The most characteristic contractures of the upper extremity involve elbow flexion contractures and stiffness and rigidity of the interphalangeal and metatarsal-phalangeal joints in extension in digits 2-5. The hip tends be located at birth, but stance and gait are worsened by the progressive flexion contracture and secondary acetabular dysplasia develops in some. Hip position is amenable to some improvement with anterior soft tissue release, but full correction is generally achieved only with a proximal femoral extension osteotomy. Recovery of hip range of motion is often slow and can be incomplete following hip osteotomy so that intervention must be done with caution. The knee flexion contractures rarely respond completely to soft tissue releases alone and can be dealt with by extension osteotomies of the distal femur (Fig. 19A). It remains essential not to damage further the physeal regions with these procedures. The extension osteotomy is not ideal because the remodeling potential of the distal femoral physis is limited due to its severe involvement in the dysplastic process, although the anterior surface concavity postosteotomy is readily remodeled because this represents a periosteal intramembranous bone function. There is little capability, however, for the distal femoral growth plate to aid in any subsequent remodeling by differential growth of the dorsal in relation to the volar parts. Clubfoot treatment can begin with serial splinting and casting, but surgery is needed almost invariably. Extensive posteromedial release has been effective in obtaining and maintaining correction. Many of these patients also have a metatarsus adductus. This does not tend to be of great clinical significance but is difficult to correct because of the shortness of the metatarsals. If necessary, however, multiple metatarsal osteotomy can be used. Spinal abnormalities are seen in approximately 80% of patients. The spectrum of abnormalities includes a cervical spina bifida, midcervical kyphosis (Fig. 8B) and a relatively high incidence of progressive thoracolumbar scoliosis (Fig. 9), thoracolumbar kyphosis, and increased lumbar lordosis (112). The scoliosis is not seen at birth but develops early in the first 2 years of life, with a tendency to rapid increase and severe rigidity. Most have a decreasing interpedicular distance in the lumbar spine, similar to that seen in achondroplasia. Quadriparesis and paraparesis can occur particularly with rapidly worsening kyphoses (27). Long-term studies of large numbers of patients from Finland have clarified the extent and nature of deformity in diastrophic dysplasia. There is a very high incidence in that country where diastrophic dysplasia is the most common skeletal dysplasia. In a study of 101 patients, 63 of whom were older than 21 years of age, spinal abnormality was assessed (246). Onethird (33%) of the patients had a cervical kyphosis usually associated with hypoplasia of the vertebral bodies of C3, C4, and C5. In most cases the kyphosis was progressive, but in 3 it resolved before 5 years of age. Spontaneous correction, if

F I G U R E 19 Orthopedic management frequently is needed in diastrophic dysplasia but is difficult due to the inexorable tendency to worsening of deformity with growth. (A) Extension osteotomy of the distal femurs often is needed to correct knee flexion deformities. Soft tissue releases can be attempted but rarely are fully effective. Preoperative film showing maximum extension is illustrated in (Ai) and postoperative correction is shown in (Aii). [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.] Note periosteal remodeling along the anterior surface of the femur. (B) Radiographs show examples of clubfoot deformity (Bi) and metatarsus adductus (Bii). (Ci-Ciii) Radiographs illustrate hip appearance in same girl with diastrophic dysplasia at 2 years, 10 years, and 18 years of age. Part (Ci) illustrates absence of proximal fern ossification center; (Cii) shows a thin secondary center and a large medial joint space (cartilaginous head); while (Ciii) shows premature degenerative hip arthritis and a titled pelvis bilaterally.

SECTION Xll ~ Review of Specific Skeletal Dysplasias it was to occur, did so before 5 years of age. Spina bifida occulta from C3-C4 to the upper thoracic spine was seen in two-thirds of patients with a greater incidence in females than in males (73% to 59%). The dens was slightly abnormal in most, sometimes being very large. Two died with severe cervical kyphosis in relation to manipulation with anesthesia. The total frequency of scoliosis was 37% with more women affected (49% to 22%). Only 13% had curves greater than 50. Those destined to fall into the severe group had deformity beginning from 2 - 4 years of age. Curves were of variable patterns including thoracic, lumbar, thoracolumbar, and double. Bracing was ineffective and surgical intervention was becoming more frequent. A narrowing of the interpedicular distance from L1 to L5 was generally seen although subsequent neurological symptoms were not as frequent as in achondroplasia. Two patients were operated because of lumbar root compression syndromes. In 102 patients the feet were almost always deformed, but not all were classical clubfoot deformities (273) (Fig. 19B). Five gradations of deformity were found. The most common pattern was a tarsal valgus with metatarsus adductus (43%), followed by equinovarus (29%), equinus (8%), metatarsus adductus (13%), and normal (7%). Only 37% could be described as having clubfoot, and two-thirds of the feet were plantigrade. Soft tissue surgery alone for the clubfoot deformity almost always was followed by recurrence to the extent of 80% after tenotomy, heel cord lengthening, or posteromedial release. Much of the deformity in tarsal valgus with metatarsus adductus was attributed to ligamentous laxity. Hip development was assessed in 50 patients (342). Abnormality worsened with age in all, some showing moderate and some severe changes (Figs. 19Ci-19Ciii). The hips were normal at birth with 2 fetal autopsy dissections and 1 newborn MRI showing joint congruity and no apparent deformities. Serial studies showed joint deformity to develop in a progressive fashion in all with changes seen by 1-2 years of age. The major deformity was hip flexion contracture seen in 93% with a mean value of 23 ~ Rotation and abduction decreased with time. Acetabular dysplasia, a flattened enlarged femoral head often with medial inferior growth prominence, a shortened femoral neck with trochanteric overgrowth, and variable joint incongruity developed. There was delay in ossification of the femoral capital epiphysis in almost all, and the secondary center was not seen until 12 years of age in one-fourth of the patients. Osteoarthritis in early adult life was common (Fig. 19Ciii). In three-fourths of the patients the femoral head remained spherical, but in the others congruous flattening developed in adolescence.

J. Spondyloepiphyseal Dysplasia 1. SED CONGENITA There is a high incidence of orthopedic deformity with SED congenita, including C1-C2 abnormalities (Figs. 8Ai-

807

8Aiii), thoracolumbar kyphosis, thoracolumbar scoliosis, and coxa vara (Figs. 3Ai, 15A, and 15B) (310, 312-314). Clinical appearance is characterized by short stature, a short neck, barrel chest, kyphoscoliotic deformity, lumbar lordosis, either varus or valgus knee deformities, and a tendency to walk with slight hip and knee flexion. The autosomal dominant disorder is a disproportionate short trunk dysplasia with spinal and proximal epiphyseal abnormalities evident at birth. Characteristically the spine and epiphyses of the proximal joints are involved with the peripheral bones less involved or uninvolved. The hands and feet are normal in length and shape. The odontoid hypoplasia is often accompanied by a cervical myelopathy. In spondyloepiphyseal dysplasia, 7 of 12 patients had increased thoracic kyphosis ranging from 30 to 130 ~ usually accompanied by increased lumbar lordosis. Scoliosis was present in 9 of the 12 and required treatment in 5 whose curves were larger than 35 ~ There is marked delay of secondary center ossification formation particularly at the proximal femoral capital epiphysis (Figs. 3Ai and 3Aii). There is retarded ossification of the vertebral bodies particularly in the lower thoracic region with oval, pear-shaped vertebral bodies on lateral projections. Coxa vara is almost invariable; the medial ossific nonunion of the neck is seen early (Figs. 20Ai and 20Aii). The distal femoral and proximal tibial secondary centers tend not to be ossified at birth, and the primary centers of the calcaneus and talus are also not ossified. Many other secondary centers show variable amounts of delay in onset of ossification. Ophthalmological problems develop in many including myopia and retinal detachment. The disorder is distinguished from Morquio's mucopolysaccharidosis by the fact that it is diagnosable radiographically at birth and there is no increased excretion of urine mucopolysaccharides including keratosulfates. Morquio's is not usually evident radiographically until 1 year of age. Stabilization by occipital-C2 fusion is warranted (Fig. 8Aiii). In adulthood there is short stature, odontoid hypoplasia and nonunion, thoracic kyphosis, lumbar lordosis, occasional severe scoliosis, flattened vertebral bodies, and narrowed disk spaces. Dysplastic acetabulae and coxa vara are seen. The long tubular bones are relatively short. 2. SED TARDA Spondyloepiphyseal dysplasia tarda is similar to the congenita form but tends to be nonrecognizable at birth, with clinical onset at 5-10 years of age. It is less severe. There are examples of both autosomal dominant and recessive transmission (275). The disorder clinically is a disproportionate short trunk dwarfism. Currently there are no detectable biochemical abnormalities of blood or urine. There tends to be flattening of the vertebrae (platyspondyly) with a midthoracic kyphosis due to anterior wedging and a shortened trunk and neck. Scoliosis is mild. The secondary ossification centers of the hip and knee regions are present but tend to develop with irregular shaping. The major problems involve

808

CHAPTER 9 9

Skeletal Dysplasias

F I G U R E 20 The clinical function of the hip can be improved by proximal femoral valgus osteotomy. (Ai) Anteroposterior hip radiograph in 7-year-old with SEDC shows bilateral coxa vara. Hip on a patient's left shows triangular medial neck fragment. Soft tissue outline of head pointing medially and inferiorly (arrow) can be seen. (Aii) Proximal femoral valgus osteotomy improves head-necktrochanter position.

premature osteoarthritis, especially of hips and knees, and cervical arthritis in early to mid-adult life (203).

K. Dyggve-Melchior-Claussen Dysplasia and Smith-McCort Dysplasia Another subgroup with vertebral and epiphyseal abnormalities is the Dyggve-Melchior-Claussen and Smith-McCort dysplasia entities composed of patients with short stature, a short neck, barrel chest, lumbar lordosis, and radiographs showing platyspondyly (flattened, shortened vertebrae), metaphyseal-epiphyseal irregularities, dysplastic hips, and a pathognomonic lacelike configuration of the iliac crest cartilage-bone junction during the growth years (Figs. 21A21C). The classic DMC syndrome is accompanied by mental retardation, whereas the Smith-McCort dysplasia has the same phenotype but the patients are mentally normal.

L. Pseudo-achondroplasia This disorder was defined by Maroteaux and Lamy to refer to patients with the body habitus of achondroplasia but normal facial features (189). It is not a single disorder because there are both autosomal dominant and recessive types with considerable variability in both. In a large review, WynneDavies et al. concluded that they could not differentiate the two types by clinical-radiographic criteria (372). Spinal and limb abnormalities are more marked radiographically than in achondroplasia, leading to the awkward term pseudo-achrondroplastic spondyloepiphyseal dysplasia, usually shortened to pseudo-achondroplasia (PSACH). Patients are described as appearing clinically normal at birth with dysplastic findings becoming clinically apparent between 1 and 3 years of age. It evolves to a short limb dwarfism with the spine relatively less involved than the limbs. Hips can be dysplastic with some coxa vara, but the acetabulae tend to be flat or horizontal. The knees in particular are the site of

major deformity worsened by ligamentous laxity, which can involve varus, valgus, recurvatum, or rotational components and is often multiplanar. Hip, knee, and ankle subluxation each can occur. There is considerable epiphyseal and metaphyseal irregularity seen on radiographs. Premature osteoarthritis results, with most patients showing OA of the knee often as early as 15 years of age. Platyspondyly with odontoid hypoplasia can be seen. Lateral spine radiographs show central, anterior tonguelike projections from the vertebral bodies. Scoliosis can occur, especially if there is pelvic obliquity, but it is readily manageable by bracing or posterior spinal fusion and is not as rigid or severe as in some other dysplasias. In some patients the spine is normal. The most valuable diagnostic radiograph is that of the pelvis in which hip and pelvic maturation is markedly delayed. The triradiate cartilage remains widened and slowly developing as late as 12 years of age, the proximal capital femoral epiphysis appears late and remains small and often fragmented for most of the first decade, the femoral neck is short and beaked, and the pubic and ischial bones do not fully ossify until 810 years of age. In distinction to achondroplasia, there is no caudal narrowing of the interpedicular spaces, no craniofacial abnormalities, and considerably greater vertebral body, epiphyseal-metaphyseal, and joint space deformation.

M. Mucopolysaccharidoses The mucopolysaccharidoses (MPS) represent the first group of skeletal dysplasia disorders for which biochemical abnormalities were detected. They are a group of lysosomal storage disorders caused by a deficiency of enzymes responsible for degrading glycosaminoglycans (formerly referred to as mucopolysaccharides). The enzyme deficiencies tend to block the normal catabolism of dermatan sulfate, heparan sulfate, or keratan sulfate, either individually or in combination, with chondroitin sulfate sometimes also involved. The glycosaminoglycan molecules accumulate in the lysosomes,

SECTION XII ~ Review of Specific Skeletal Dysplasias

809

FIGURE 21 Radiographic characteristics of Dyggve-Melchior-Claussen dysplasia or Smith-McCort dysplasia are shown. (Ai, Aii) Platyspondyly is apparent in both AP and lateral projections involving cervical and thoracolumbarvertebrae. (B) Anteroposterior radiograph of the hips shows platyspondyly, fragmentation of the bone of the femoral heads, and the characteristic lacelike appearance of the iliac crest bone (arrows). (C) Higher power view of the lacey configurationof the iliac crest apophysis is shown.

resulting in cell, tissue, and organ dysfunction. The glycosaminoglycan fragments generated by alternative pathways are then excreted in the urine. There are 10 known enzyme deficiencies giving rise to 7 distinct MPS disorders. The MPS disorders share many clinical features although to variable degrees. The course is chronic and progressive at variable rates, with particular involvement of hearing, vision, cardiovascular function, musculoskeletal function, and men-

tal capability. Severe mental retardation is characteristic of MPSIH, the severe form of MPSII, and all subtypes of MPSIII. Normal intellect is usually retained in the other disorders. The skeletal disorders are particularly marked in MPSIV (Morquio syndrome) with many specific to that disorder. The large majority of MPS disorders are transmitted in an autosomal recessive manner except for MPSII, which is X-linked. Simple enzyme assays are available for the

810

CHAPTER 9 ~ Skeletal Dysplasias

diagnosis of MPS using fibroblasts, leukocytes, or serum. Prenatal diagnosis following amniocentesis or chorionic villus biopsy is possible for all MPS disorders. The conditions are not yet curable, but newer treatments have involved bone marrow transplantation and gene transfer into hematopoietic cells. They are characterized by increased urinary secretion of glycosaminoglycans. Several different types have been defined, but it is the MPSIV Morquio disorder that requires the most frequent and extensive orthopedic management. From a radiographic perspective, Morquio's disease is a spondyloepiphyseal dysplasia but the biochemical abnormality has led to a specific grouping. Table XIII summarizes the MPS disorders in terms of clinical findings, enzyme deficiency, the glycosaminoglycan affected, inheritance, and gene locus (115, 221, 361). 1. MPSIH (HURLERSYNDROME) In this short stature disorder there is a deceleration of growth between 6 and 18 months of age with failure of advancement by 2-5 years. Death usually occurs in childhood, and patients rarely live beyond 10 years of age. There is a prominent coarse facial appearance with flared nostrils, a low nasal bridge, and a tendency toward hypertelorism. The disorder was described in the early literature as gargoylism. Blindness occurs as the corneas become hazy and there is retinal pigmentation. Other abnormalities involve an enlarged tongue, hepatosplenomegaly, inguinal hernias, joint stiffness, a tendency to hip dislocation, and deafness. The bone radiographs are characterized by diaphyseal widening of the shortened and misshapened bones. Frequently there is a short neck, kyphosis, and a thoracolumbar gibbus secondary to anterior vertebral wedging. The disorder is due to the absence of lysosomal-L-iduronidase in all tissues, leading to the accumulation of mucopolysaccharide in mesenchymal and other tissues including neuronal tissues. There is increased secretion of dermatan sulfate and heparan sulfate in the urine and absence of L-iduronidase in cultured fibroblasts. There is relatively little need for orthopedic intervention due to progressive systemic deterioration. 2. MPSIS (SCHEIESYNDROME) The findings in relation to urine secretion and absence of L-iduronidase in cultured fibroblasts are the same as for the MPSIH syndrome, and differentiation is made on the bases of the clinical course, which is far more benign, and the phenotype in which the intelligence is normal, the facial features are much less coarse, and the hepatosplenomegaly is not as marked. There are few orthopedic implications for this disorder. 3. MPSII (HUNTERSYNDROME) The disorder is somewhat milder than the MPSIH syndrome. It is a sex-linked recessive disorder with only males affected. Onset occurs between 2 and 4 years of age and leads to some growth deficiency. In the juvenile type, mental and neurological deterioration at 2-5 years of age leads to hyperactive behavior and spasticity, whereas in those devel-

oping symptoms somewhat later the mental deficiency can be mild to absent. The coarsening of facial features, stiff and contracted joints particularly of the hand, and hepatosplenomegaly are also seen. There is urinary excretion of dermatan and heparan sulfates. The major clinical differences in contrast to Hurler syndrome involve the presence of clear corneas, a less severe gibbus, no females affected, and a more gradual onset of features. Deafness can occur as early as 2-3 years of age. Cardiac complications can lead to death prior to 20 years of age, and in general the more severe juvenile patients frequently die between 4 and 14 years of age. 4. MPSIII (SANFILIPPOSYNDROME) This disorder is characterized by onset in early childhood with slowing of growth occurring after 2-3 years of age. Some appear normal until 4 or 5 years of age. There is slowing of mental development followed by rather marked deterioration including difficulties with gait, speech, and behavioral control. The usual result is severe mental deficiency, although the syndrome may be compatible with long survival. Significant skeletal abnormalities do not occur. Urinary excretion studies reveal increased heparan sulfate. There are four subtypes--A, B, C, and Dmwhich have the same urine findings and clinical appearance but different biochemical abnormalities. 5. MPSIV (MORQUIOSYNDROME) This disorder was described independently by Morquio (207) and Brailsford in 1929 and defined biochemically as a mucopolysaccharidosis in 1963. It represents a clear example of a disproportionate short trunk dysplasia. It is autosomal recessive. Nonosseous problems include corneal opacities due to tissue accumulation of mucopolysaccharides and aortic damage. Morquio pointed out the short trunk with fingers reaching to the knees, the prominent sternum, short neck, severe genu valgum, flat feet, hyperextensible thumb, and, by radiographs, the epiphyseal abnormality, spacious acetabulae, and generalized platyspondyly. Clinical onset becomes evident between 1 and 3 years of age with severe early limitation of growth and eventual premature cessation of growth. Orthopedic deformities in the MPS syndromes are most extensive with this disorder and are of the spondyloepiphyseal type with diagnostic differentiation depending on biochemical testing (202). A waddling gait develops along with a knock-knee deformity and flat feet around 3 years of age. The genu valgum becomes very marked if untreated by 8-10 years of age. There is marked involvement of the spine with flattening or platyspondyly of each vertebra, with some subsequent ovoid appearances with central anterior bone projections of the vertebral body. The neck is short and there is anterior protrusion of the sternum. The incidence of kyphoscoliosis or scoliosis is high, but major deformity occurs infrequently. Odontoid hypoplasia with or without transverse atlantal ligament laxity is always present, and virtually all patients will develop signs of a cervical myelopathy of variable severity. Symptoms of cervical myelopathy tend to

Scheie

Hurler- Scheie

Hunter (severe)

Hunter (mild)

Sanfilippo A

Sanfilippo B Sanfilippo C

Sanfilippo D

Morquio A

Morquio B

No longer used

Maroteaux-Lamy

Sly

MPSIS

MPS IH/S

MPSII (severe)

MPSII (mild)

MPSIIIA

MPSIIIB MPSIIIC

MPSIIID

MPSIVA

MPSIVB

MPSV

MPSVI

MPSVII

Hepatosplenomegaly; wide spectrum of severity, including lethal neonatal form, infantile form with mental retardation and juvenile form with mild mental retardation

Variable short stature, kyphoscoliosis, corneal clouding, normal intelligence; survival to teens in severe form; milder forms exist

Distinctive skeletal abnormalities with extreme to mildly short stature, corneal clouding, odontoid hypoplasia; milder forms exist Spectrum of severity as in IVA

Phenotype similar to IIIA

Phenotype similar to IliA

Profound mental deterioration, hyperactivity, relatively mild somatic manifestations Phenotype similar to IliA

Normal intelligence, short stature, survival to 20s to 60s

Diagnosis before age 4 years, organomegaly, no corneal clouding or kyphoscoliosis, mental retardation, death before 15 years

Phenotype intermediate between IH and IS, mild to severe skeletal involvement, corneal clouding, early death

Late corneal clouding, stiff joints, normal intelligence and life span

Diagnosis before 2 years of age, early corneal clouding, kyphoscoliosis, organomegaly, heart disease, mental retardation, death in childhood

Clinical manifestations

[3-Glucuronidase

N-Acetylgalactosamine 4-sulfatase (aryl sulfatase B)

[3-Galactosidase

Galactose 6-sulfatase

N-Acetylglucosamine 6-sulfatase

c~-N-Acetylglucosaminidase Acetyl-CoA: o~-glucosaminide acetyltransferase

Heparan N-sulfatase

Iduronate sulfatase

Iduronate sulfatase

ct-L-Iduronidase

ct-L-Iduronidase

Ct-L-Iduronidase

Enzyme deficiency

Classification of Mucopolysaccharidoses ~

aDerived from Hopwood JJ, Morris PC (1990) Mol Biol Med 7:381-404; Neufeld E, Muenzes J (1995); and Whitley CB (1993).

Hurler

Eponym

MPSIH

MPS type

TABLE XIII

Dermatan sulfate, heparan sulfate, chondroitin 4-, 6sulfates

Dermatan sulfate

Keratan sulfate

Keratan sulfate, chondroitin 6-sulfate

Heparan sulfate

Heparan sulfate Heparan sulfate

Heparan sulfate

Dermatan sulfate, heparan sulfate

Dermatan sulfate, heparan sulfate

Dermatan sulfate, heparan sulfate Dermatan sulfate, heparan sulfate

Dermatan sulfate, heparan sulfate

Glycosaminoglycan affected

Autosomal recessive

Autosomal recessive

Autosomal recessive

Autosomal recessive

Autosomal recessive

Autosomal recessive Autosomal recessive

Autosomal recessive

X-linked recessive

X-linked recessive

Autosomal recessive

Autosomal recessive

Autosomal recessive

Inheritance

7ql 1.2-22

5q11-13

3p21-cen

12q14

Xq27.3

4p16.3

Gene locus

812

CHAPTER 9 9 Skeletal Dysplasias

begin around 5 years of age. The hip and pelvis are abnormal with a flattening of the femoral head and shortened femoral neck. There is an invariable knock-knee deformity, short stubby hands, and joint laxity mostly evident at the wrist, ankles, and small joints. Much of the knock-knee deformity is attributable to ligamentous laxity. There is an outward flaring of the rib cage and a prominent sternum. C 1-C2 and C2-C3 subluxations are common. There is an excessive excretion of keratan sulfate in urine. There are two variants, types A and B, with type IVA due to a deficiency of Nacetylgalactosamine-4-sulfatase and type IVB due to deficiency of [3-galactosidase. On occasion the enzyme defects can be diagnosed from cultured skin fibroblasts. Maximum adult height is the range of 100 cm. Orthopedic treatment involves atlantoaxial fusion or occipitocervical fusion. Genu valgum often requires soft tissue releases and realignment with femoral or tibial osteotomy. Long leg bracing is usually required postsurgery for additional stabilization. [The MPS V syndrome was formerly known as Scheie's disease, but this has now been reclassified as MPSIS.] 6. MPSVI (MAROTEAUX-LAMYSYNDROME) This variant is distinct from the Hurler syndrome because mental deterioration does not occur during early childhood. Epiphyseal irregularity is seen particularly involving the femur, and the vertebrae are flattened with anterior wedging particularly of T12 and L1. Odontoid hypoplasia, lumbar kyphosis, and genu valgum are seen. Increased urinary excretion involves predominantly dermatan sulfate. 7. MPSVII (SLYSYNDROME) This is a rare variant characterized by a deficiency of [3glucuronidase. There is moderate mental deficiency. This disorder can be recognized in the neonatal period and is associated with corneal clouding as early as 7 months of age. Growth deficiency begins immediately postnatally, and structural abnormalities involve odontoid hyplasia with atlantoaxial instability (242), shortening and anterior inferior beaking of lower thoracic and lumbar vertebrae, a prominent sternum, acetabular dysplasia, and narrowed sciatic notches. The high frequency of atlantoaxial structural abnormality and instability in the mucopolysaccharidoses was documented with 67 patients showing the phenomenon. In 5 there was complete absence of the odontoid process. There were 14 of 18 patients with a thoracolumbar kyphosis, which measured from 14 to 53 ~ and invariably was associated with one or more hypoplastic anterior vertebral bodies at the TL junction. A scoliosis of more than 10~ was seen in 8 patients, with the largest curve at 38 ~.

N. Metaphyseal Dysplasia A group of skeletal dysplasias characterized by radiographic abnormalities most marked in the metaphyses is referred to

as metaphyseal dysplasias, dysostoses, or chondrodysplasias. This group, which is relatively rare, represents an example where "fine-tuning" of descriptive entities has led to more confusion than clarity. If short stature is present, the physeal regions of the epiphyses must be involved because growth occurs here, not in the metaphyses. The term metaphyseal dysplasia thus describes the most evident radiographic sequelae rather than the region of primary involvement. The three main entities seen are Schmid type, Jansen type, and McKusick type (or cartilage-hair hypoplasia) (15). Other metaphyseal disorders described are associated with immunologic or endocrine abnormalities and have radiologic similarities with rickets, especially in relation to metaphyseal irregularities at the physeal-metaphyseal junction. The metaphyseal dysplasias tend to have bowing of the knees and mildly short stature and become evident after 2-3 years of age. Coxa vara is described in many of the entities, although this is clearly an abnormality of femoral neck physeal growth.

1. SCHMIDMETAPHYSEAL DYSPLASIA This is the most common entity and has been recognized on numerous occasions. The transmission is autosomal dominant, and the disorder is characterized by genu varum, often coxa vara, and short stature. Expanded, irregular metaphyses are noted at the hips and knees with the rest of the skeleton virtually normal. Manifestations develop after 2 years of age as bowing persists and short stature becomes more evident. In this "purest" variant of the metaphyseal dysplasias, skull, spine, epiphyses, and carpal and tarsal bones are normal radiologically: There are persisting bowlegs with internal tibial torsion and coxa vara, and metaphyseal irregularities predominate. 2. JANSEN METAPHYSEAL DYSPLASIA The Jansen type of metaphyseal dysplasia is rare but eventually leads to severe dwarfism, often with restricted joint movements and clubfoot. Radiographic abnormalities are centered toward the middle part of the first decade with abnormal architecture of the metaphyses. Asymptomatic hyperglycemia is associated. 3. McKusICK (CARTILAGE-HAIR) METAPHYSEAL DYSPLASIA This variant of metaphyseal dysplasia is prevalent particularly among the Amish in Pennsylvania and Finland. Short stature is moderate and the phenotype is characterized by thin, sparse hair on the body and head. The finger joints are loose and the fingers have widened and shortened nails. Distal prolongation of the fibula can produce inversion deformity at the ankle. Metaphyses are radiographically irregular with cystic changes across the entire width of the bone.

O. Spondylometaphyseal Dysplasia Another variant of descriptive localization in the metaphyseal dysplasias is the spondylometaphyseal group. Many variants

SECTION Xll ~ Review of Specific Skeletal Dysplasias

have been described, but some characteristic features include metaphyseal irregularities, bowlegs, coxa vara, a horizontal acetabular roof (similar to achondroplasia), and platyspondyly with kyphoscoliosis. The epiphyses are normal. Syndromal overlap with spondyloepiphyseal dysplasia and spondylometaphyseal dysplasia is high because rarely is there only epiphyseal or metaphyseal involvement in the latter two. Joint laxity with spondyloepimetaphyseal dysplasia can lead to severe clinical disability including death in the first decade.

P. Cleidocranial Dysostosis (Dysplasia) Cleidocranial dysostosis (dysplasia) is one of the more commonly seen dysplasias with a characteristic series of changes. The clavicles are maldeveloped in association with mildly short stature, delayed closure and widening of the anterior fontanelle and cranial sutures, wormian bones of the skull, delayed mineralization of the pubic bone, late tooth eruption with malformed teeth, frontal facial bossing, hypertelorism, and occasional coxa vara (Figs. 22A and 22B). The hip abnormalities are accompanied by widened triradiate cartilages during the first decade (Fig. 22B). The clavicular abnormalities are always seen but are variable and may be asymmetric (Fig. 12). They include complete absence, but there may be shortened medial segments present. The shoulders can generally touch anteriorly.

Q. Hereditary Arthro-ophthalmopathy (Stickler Syndrome) This autosomal dominant syndrome is characterized by swollen joints (enlargement of large joints such as wrists, knees, and ankles) leading to arthritis in mid-adult life, myopia and choroidoretinal abnormalities, and deafness (24). The hip is usually involved with coxa valga, widening of the femoral neck, and acetabular protrusion. Collagen abnormalities appear to underlie the syndrome.

R. Dyschondrosteosis The initial description and naming of dyschondrosteosis were in 1929 by Leri and Weill (171). It is transmitted as an autosomal dominant disorder with females more severely affected than males. Dyschondrosteosis is a skeletal dysplasia characterized by short stature, Madelung's deformity at the wrist, and mesomelia involving shortening of the middle segments of the upper and lower extremities (Fig. 23) .(164). The most shortened segments involve the radius, ulna, tibia, fibula, and sometimes the adjacent carpal and tarsal bones. Dyschondrosteosis is a specific skeletal dysplasia that, although mild, generally is characterized by the presence of Madelung's deformity of the distal forearm and wrist. Indeed, when a Madelung's deformity is seen, the patient should be carefully assessed because the large majority will have the deformity in association with either hereditary mul-

813

tiple extososis, Ollier's disease, or dyschondrosteosis. The deformity defined by Madelung in 1878 involved marked hypoplasia of the distal ulna along with its dorsal subluxation or dislocation and an ulnar drift of the wrist and hand (182). There may be slight lateral diaphyseal bowing of the radius associated with the shortened ulnar component. If the ulna is particularly shortened, it causes a tethering effect medially on the distal radius, and there is obliquity of the distal radial articular surface causing it to face in an ulnar and palmar direction and associated obliquity of the distal radial secondary ossification center and physis. In severe cases, premature fusion of the medial half of the distal radial physis accentuates the deformity in adolescence. It is extremely difficult if not impossible to diagnose the disorder at birth, but it is usually evident by the middle to end of the first decade. Shortness of stature is an invariable occurrence, although it is relatively mild. Adult height in dyschondrosteosis is mildly reduced and ranges from 137 to 163 cm. Most affected patients are within the normal range for height but at the lower margin. In a report by Beals and Lovrien, documentation of shortening and mesomelia was such that radial length compared to humeral length was 65% (normally 75%), whereas the tibial length decreased to 70% (normally 82%) in relation to femoral length (13). Similar findings were documented by Langer. There are relatively few reports of requirements for surgical intervention for the forearm, wrist, and hand deformities, although with increasing sophistication of these procedures consideration of their use is increasing. Beals and Lovrien indicated, however, that patients with dyschondrosteosis exhibited little functional impairment or cosmetic deformity, and they felt that therapy was usually unnecessary. On occasion, patients are seen with a genu varum severe enough to warrant surgical correction.

S. Other Mesomelic Dysplasias Many disorders are characterized by the mesomelic pattern of forearm (radius-ulna) and leg (tibia-fibula) shortening. Langer mesomelic syndrome has a disproportionate short limb dwarfism with mandibular hypoplasia, upper and lower extremity shortening with particularly marked shortening of the forearms and legs, and greater ulnar shortening compared to the radius and fibular compared to the tibia. Nievergelt mesomelic syndrome presents with forearms and legs about one-third the lengths of arms and thighs. There is a characteristic rhomboidal shape of the tibia and a triangular fibula.

T. Acromesomelic Dysplasia Dysplasias occur in which progressively more extensive middle and distal limb shortening occurs. Acromesomelic dysplasias refer to the presence of both distal (acromelia) hand and foot shortness and shortness of the forearm and leg

814

CHAPTER 9

~

Skeletal Dysplasias

F I G U R E 22 Radiographic characteristics of cleidocranial dysplasia are shown. (Ai, Aii) Cranial X rays show delayed closure and widening of the anterior fontanelle and cranial sutures and wormian bones. (B) Pelvic abnormalities are characterized by coxa vara, widened triradiate cartilages during the first decade, and delayed ossification of the pubic and ischial bones. Eventual ossification of the latter is complete. (Bi-Biii) Note progressive pubic bone ossification from ages 8 to 14 to 23 years.

SECTION Xli ~ Review of Specific Skeletal Dysplasias

815

(mesomelia). These are rare occurrences. Two specific variants are Grebe dysplasia and Hunter-Thompson dysplasia.

1. GREBE CHONDRODYSPLASIA A marked variant of an acromesomelic syndrome is Grebe chondrodysplasia, with profound upper and lower extremity shortening concentrated in the forearms, legs, and digits. In this disorder there is a normal axial skeleton, a relatively normal humerus and femur, but markedly shortened and deformed forearms (radius-ulna) and legs (tibiafibula), and even more severe abnormalities of the hands and feet, with fused carpal and tarsal bones, several metacarpal and metatarsal bones absent, proximal and middle phalanges of the fingers and toes invariably absent, and distal phalanges present. The digits are reduced to mere appendages and there is often postaxial polydactyly (61).

2. HUNTER-THOMPSONACROMESOMELICDYSPLASIA This variant is milder than the Grebe disorder, with mildly shortened forearms and legs, fibulae represented by distal triangular remnants, metacarpals extremely short, several finger phalanges absent, variable hypoplasia of the metatarsals and single phalanges of the toes (129).

U. Acromelic Syndromes (Acrodysplasias) There are literally dozens of syndromes with digital anomalies leading to shortness and malformation present either exclusively or predominantly in the hands and feet. These most distal abnormalities are included in limb shortening disorders referred to as acromelic or as the acrodysplasias. The short hands and feet are usually associated with shortened stubby digits. Although many of the disorders are isolated to the hands and feet many are also associated with craniofacial abnormalities, ectodermal abnormalities including the nails, teeth, gums, and hair, and other organ abnormalities including in particular structural cardiac problems. It is perhaps not surprising that these regions are subject to considerable developmental malformation since there are numerous bones present within a relatively restricted region. With the recent deciphering of gene and molecular abnormalities of skeletal development, many of these disorders are being defined in relation in particular to abnormalities of patterning genes active in early limb bud formation. Even in the relatively pure syndromes limited to hands and feet there are almost always subtle abnormalities detected elsewhere. The classic forms of acrodysplasia however involve shortening of the tubular bones of the hands and feet, cone shaped diaphyses and virtually no changes elsewhere. The most prominent shortening is seen in the metacarpals and metatarsals. Amongst the earliest descriptions of this particular disorder is one by Brailsford who referred to the disorder as peripheral dysostosis.

1. ELLIS-VAN CREVELDSYNDROME The Ellis-Van Creveld (EVC) syndrome is characterized by acromelic limb shortening, polydactyly primarily of the

FIGURE 23 Anteroposteriorradiograph of the forearm in a patient with dyschondrosteosisshows the characteristic Madelung's deformity at the wrist with the shortenedand dorsallydisplacedulna and radial bowing. hands, and ectodermal dysplasia involving the nails, teeth, and gums. Congenital heart disease is present in approximately 50% of patients, most commonly as a single atrium with a cleft mitral valve. The face is characterized by hypertelorism. Also seen is sparse, absent, or silky hair, variable tooth abnormalities involving either early or late presentation but always pegged and hypoplastic appearances, and gums adherent to the lips. The nails of the hands and feet are always deformed. In the upper extremity there is an extra digit always on the ulnar side and an occasional extra digit on the radial side, leading to six or seven digits overall. The clavicles are abnormal, the radial heads usually are dislocated, and a knock-knee deformity almost always develops in the first decade. The genu valgum that results is due to a defect in the lateral portion of the proximal tibia. The EllisVan Creveld disorder is now more commonly referred to as chondroectodermal dysplasia. The abnormality is detected in the newborn. In the trichorhinophalangeal (TRP) syndrome the peripheral dysostosis is associated with short, scant hair, mildly short stature, and a large, pear-shaped nose. The major orthopedic disorder associated with TRP is a Perthes-like hip involvement that often heals with marked malformation (59, 67). 2. CRANIOFACIALDYSOSTOSES Many of the digital abnormalities are associated primarily with major craniofacial developmental abnormalities. Many

816

CHAPTER 9 9 Skeletal Dysplasias

of these have been found to be associated with molecular abnormalities of FGFR2 (51). Craniosynostosis syndromes are increasingly linked to abnormal gene loci (134). Some of the more frequently seen disorders follow (140). a. Apert Syndrome. Among the most common disorders is the Apert syndrome (acrocephalosyndactyly). The main clinical features are a high forehead, flat occiput, midfacial hypoplasia, and fusion of the second, third, and fourth fingers and toes. b. Pfeiffer Syndrome. Pfeiffer reported a family with mild acrocephaly in association with cutaneous syndactyly of the second and third fingers and marked deviation of broadened thumbs and great toes. c. Crouzon Syndrome (Craniofacial Dysostosis). The Crouzon syndrome is associated with craniostenosis with midfacial hypoplasia, hypertelorism, nasal beaking, prominent eyes, dental malocculsion, and deafness. Other of the craniofacial disorders can also be associated with digital abnormalities. 3. POLYDACTYLY SYNDROMES Preaxial polydactyly involves a unilateral extra thumb or big toe and can be seen as isolated autosomal dominant disorders. Postaxial polydactyly is more common with six digits present and can be seen either as an isolated syndrome or more commonly in association with multiple other disorders. 4. SYNDACTYLY Syndactyly refers to bone or soft tissue union of two or more digits. A multiplicity of forms and variants has been described. These can occur as either isolated phenomena or more frequently as components of well-established genetic disorders such as the acrocephalosyndactylys.

5. DIGITAL DYSPLASIA SYNDROMES In the final group of disorders many digital abnormalities are associated with malformations of other organs of other parts of the musculoskeletal system. Among the most common of these is the Holt-Dram syndrome in which unilateral dysplasia of the thumb is associated with malformations of the heart. In some the thumb is totally absent, and limb defects particularly on the radial side can be extensive. There can also be extra or abnormal bones in the carpus. 6. POLAND SYNDROME In this disorder there is the association of unilateral cutaneous synbrachydactyly with ipsilateral absence of the pectoralis minor muscle and the sternal portion of the pectoralis major. Rib defects together with the absence of part or all of the hand is an occasional component. Dextrocardia also can be associated. Brachydactyly refers to shortening of a single or multiple digits due to bony malformation. Other deformities involving increased or decreased length of individual or groups of digits as well as rectangular deformations have been described.

V. Larsen's Syndrome This syndrome is characterized by multiple congenital joint dislocations and a characteristic flattened facial appearance with depressed nasal bridge, widely spaced eyes, and a prominent forehead (166). Though not all joints are affected in each patient, several regions of abnormality must be present for the diagnosis to be made, and the knee with rare exceptions is always involved. Knee involvement ranges from recurvatum to subluxation to complete dislocation (165). As well as knee involvement, the multiple joint dislocations include the hips (dislocatable, subluxed, dislocated), elbows (dislocation, proximal radioulnar synostosis), hands, and wrists (radiocarpal dislocation, digital dislocationssubluxations, polydactyly). Equinovarus foot deformities are common along with metatarsus adductus and flat feet. Spinal problems are extremely prominent, with C1-C2 instability, cervical spinal bifida, and on occasion a midcervical kyphosis due to underdevelopment of the anterior parts of the central vertebral bodies (201). There is a tendency to decreased interpedicular distance in the thoracic region and mild to moderate thoracolumbar scoliosis. There is an extra ossification center in the calcaneus seen in the first decade only. The thumb is widened and flattened. The disorder is more characteristic of a hyperlaxity syndrome than a primary skeletal dysplasia.

W. Ollier's Disease 1. TERMINOLOGY Ollier's disease is a rare, nonhereditary skeletal condition characterized by persisting cartilage masses in the metaphyses and diaphyses, sub-periosteal deposition of cartilage, and asymmetrical involvement of the limbs, with one side being either exclusively or predominantly involved (228). The affected bones often are shortened and deformed (Figs. 24A-24D). In childhood they are subject to pathological fractures and in adult life to malignant degeneration. This bone disease, defined initially by Ollier (227, 228), is included by some in the condition referred to as multiple or general endochondromatosis (232, 238), but topographically it is separate from the condition of bilateral multiple enchondromas, in which involvement of the short tubular bones of the hands and feet is greater.

2. DISEASE PROFILE A study from our institution assessed 21 patients with radiographically documented Ollier's disease seen over four decades in relation to the frequency, extent, and progression of angular deformities, pathological fractures, and limb length discrepancies and to their orthopedic treatment (279). There were 11 boys and 10 girls. Seventeen patients had exclusively unilateral involvement and 4 had predominantly unilateral involvement. There was equal distribution between the sexes and between the left and fight sides. As in

F I G U R E 24 Radiographic findings in Ollier's disease are shown. (A) A characteristic mild focus of cartilage is seen in the proximal femur in the area of the lesser trochanter and proximal medial diaphysis. (B) Slightly more extensive area of involvement affects the greater trochanter and proximal diaphysis. The cartilaginous tissue is concentrated laterally. Punctate foci of calcification are characteristic. (C) Shortening, often of several centimeters, occurs when there is involvement of the distal femur and proximal tibia. Radiolucencies in the metaphyseal region are characteristics. If the involvement is uniform across the diameter of the bone, in both anteroposterior and lateral projections shortening is present but is not complicated by angular deformity, as in this case. (D) Angular deformity occurs when uniform involvement across the diameter is not seen. There is a considerable varus deformation of the distal femur and a mild valgus of the proximal tibia and fibula, leading to marked obliquity of the knee joint. [Reprinted from (279), with permission.]

818

CHAPTER 9

~

Skeletal Dysplasias

previous reviews, the femur and tibia were the bones involved most frequently. Pelvic and fibular involvement did not pose any clinical problems in these patients, and lesions of the hand and foot posed relatively few problems. The lesions in the long bones were located in the metaphyses, with diaphyseal positioning in addition in the patients with more severe involvement. Although the shortening and deformities were progressive, the positions and radiographic features of the lesions did not change, except as described. In some epiphyses after the middle of the first decade of life enchondromas became evident but significant epiphyseal deformity did not occur. As the metaphyseal lesions matured, punctate calcifications were seen in them quite often. The youngest patient showing this characteristic radiographic finding was 6 years old; such changes sometimes were not seen until the age of 13 years. Fairbank described a severe case in which calcification was seen at the age of 2 years 8 months. Huvos has indicated that the occurrence of dense calcification in cartilage lesions is a feature of benign processes, whereas the disappearance of calcified masses may be a sign of malignant change. The radiographs in the present series showed increasing and more normal bone density after skeletal maturation, as in other studies. Normal trabeculation did not occur in any site in which there was a lesion. Anthropometric measurements were available for 18 patients. All measurements of height were within the normal range. Two children were 1.5 standard deviations (SD) above the mean, 2 were 1 SD and 2 others were 0.5 SD above the mean, 8 patients had height measurements at the mean, and 3 were 1 SD and 1 was 2 SD below the mean. The femur and tibia were the bones involved most frequently. The pelvis, in particular the iliac crest, and the fibula also were affected commonly. The metaphyses were the main sites of involvement, although diaphyseal involvement was seen as well in the more severe cases. Epiphyseal irregularities were seen in the patients with the most extensive involvement, usually becoming apparent radiographically after the age of 5 years. In the proximal end of the femur the initial focus of abnormality was usually the lesser trochanter, with involvement seen toward the greater trochanter and inferiorly in more severe cases. The femoral neck was relatively spared. In all patients, the region of involvement did not change with time.

3. CLINICALSEQUELAE The important clinical problems were progressive shortening of the involved extremity, angular deformity, and pathological fracture. Malignant degeneration to chondrosarcoma was not seen, but this should not be considered a definitive observation because the patients were not followed into adult life. Malignant change in Ollier's disease generally occurs in the adult years. a. Angular Deformities. The location and degree of angular deformity were closely related to the position and extent of enchondromas. When a metaphyseal area was in-

volved uniformly across the width of the bone, there was no deformity other than shortening. Epiphyseal plate growth tended to be slow adjacent to metaphyseal enchondromas. Deformity invariably occurred when metaphyseal involvement was not uniform across the entire width of the bone. Angulation was associated in all instances with nonuniform enchondromatous metaphyseal involvement. The concavity of the angular deformity always was adjacent to the more extensive endochondromatous involvement. Angular deformity was common, most often in the distal part of the femur, and was most severe there. Repeated osteotomy often was required to achieve good alignment at skeletal maturity. Although instances of proximal and distal tibial angulation also occurred, a single osteotomy performed toward the latter part of the growth period generally was sufficient in those patients. Growth is slower adjacent to the enchondromas, either because epiphyseal cartilage is abnormal or because there is a tethering effect on the epiphyseal cartilage tissue by an abnormally thick periosteal sleeve in reaction to the lesions. Any nonuniform involvement with enchondromas indicates the direction the deformity will take, and the more extensive the involvement the more likely it is that there will be recurrent deformity after osteotomy. Spontaneous correction of deformity, either partial or complete, was not seen. b. Involvement of the Lower Extremity. Of the 20 patients with involvement of the lower extremity, each of the 6 with proximal femoral lesions had coxa valga, but that deformity always was mild and virtually nonprogressive. There were no instances of subluxation, dislocation, or dysplasia of the hip, and corrective osteotomy never was required. Coxa vara was not seen. Fifteen of the 19 patients with distal femoral involvement had an angular deformity, 8 showing varus and 7 valgus angulation. Two patients also had anterior angulation because of prominent posterior enchondromas. One patient had posterior bowing because of an anterior enchondroma. Of the 8 patients with varus deformity, 5 required an osteotomy. There were 11 osteotomies in all, ranging from 1 to 4 per patient. Of the 7 patients with valgus deformity, 5 required osteotomy, for a total of 13 osteotomies. One patient had 6 osteotomies. The need for repeated osteotomies was manifest primarily in the distal end of the femur, especially with valgus deformity. Eight patients had angular deformities in the tibia, 6 proximally and 2 distally. Proximally, only tibia valga was seen. Of the 6 patients, 4 had an osteotomy. A single procedure was sufficient to provide and maintain correction in each. Distally, 2 patients had a varus deformity, and 3 osteotomies were required. 4. DISTAL FEMORAL AND TIBIAL DEFORMITIES

Of the 8 patients with a varus deformity of the distal end of the femur, 5 had osteotomies. The first patient of the 5 had 35 ~ of varus angulation and underwent 4 medially based opening wedge osteotomies at the ages of 9, 11, 12, and 15 years. In retrospect, all of the wedges were insufficient, and the final varus angle was 20 ~ The second patient had an

SECTION XII 9 Review o f Specific Skeletal Dysplasias

osteotomy at the age of 4 years for a 20 ~ deformity. By the age of 8 years, 20 ~ of deformity had recurred, but it was balanced by 20 ~ of proximal tibia valga. Both deformities then remained virtually unchanged well into adult life. The fourth patient had 25 ~ of varus angulation; 2 osteotomies were done at the ages of 8 and 12 years, and a straight femur was evident at skeletal maturity. The fifth patient had a 20 ~ varus deformity corrected by osteotomy at the age of 17 years. Of the 3 patients who did not have surgical intervention, one was 7 years old with a 25 ~ deformity, one was 15 years old with a 10 ~ deformity balanced by proximal tibia valga, and one was an adult with a 20 ~ deformity in a severely shortened limb that required a prosthesis. Five of the 7 patients with a distal femoral valgus deformity had osteotomies. One had a 25 ~ deformity. At the age of 6 years it was corrected to 10~ with an opening wedge procedure. At the age of 13 years the deformity was 15 ~ and full correction was then attained with another osteotomy. The second patient had a 35 ~ deformity, for which 6 osteotomies were done at the ages of 4, 7, 9, 11, 13, and 14 years. At maturity the femur was straight. The third patient had 2 distal femoral opening wedge procedures, one at 9 years for a 25 ~ deformity and one at 13 years for a 30 ~ deformity. The fourth patient (not tabulated) had a distal femoral osteotomy at the age of 4 years for a 30 ~ angulation but was lost to follow-up. The fifth patient had 2 distal femoral opening wedge procedures, one at 9 years for a 20 ~ deformity and one at 15 years for a 30 ~ deformity. In the 2 patients with valgus deformity who did not have an osteotomy, one (not tabulated) was an 8-year-old child with a 15 ~ angulation and the other was 13 years old with only 8 ~ of deformity. Eight patients had an angular deformity of the tibia. Of the 6 with proximal tibia valga, 4 had a laterally based opening wedge osteotomy: at 14 years old (an 18 ~ deformity), at 8 years old (for a 25 ~ deformity), at 12 years old (for a 20 ~ deformity), and at 14 years old (for a 15 ~ deformity). The angulation had developed relatively slowly in each patient. The 2 patients with proximal tibia valga who did not have an osteotomy also had a varus deformity of the distal end of the femur, so that alignment of the two long bones was nearly normal although the articular surfaces at the knee in both patients were slightly oblique. The 2 patients with varus deformity of the distal end of the tibia each had a closing wedge osteotomy. In one this was performed at the age of 6 years, correcting a 15 ~ tilt to neutral, but by the age of 15 years the deformity had recurred (to 25 ~ and the osteotomy was repeated. A closing wedge osteotomy also was required on the opposite ankle of this patient, although there was no radiographic evidence of endochondromatosis on that side. The second patient (not tabulated) still had a 20 ~ varus tilt at skeletal maturity, despite early osteotomy. The angle had not changed greatly over a decade of observation. The involved side was extremely short and a prosthesis was needed. All femoral and tibial osteotomies healed normally, whether the osteotomy was

819

above or through enchondromatous lesions. Acetabular or other pelvic deformities never were of clinical significance. Angular deformity associated with enchondromas was not seen in the fibula; any fibular deformity appeared to be secondary to tibial involvement. Although the proper timing for an osteotomy depends on many variables, the procedure generally is indicated when the weight bearing alignment of the extremity is altered significantly. As a rough guide, an angular deformity of 25 ~ or more, not balanced by a reverse deformity, is an indication for osteotomy. If possible, procedures should be delayed so as to limit the number required. There were three causes of recurring angulation after osteotomy: (1) The persistence of the disease at the epiphyseal growth plate and metaphysis. The osteotomy is directed at the deformity and not at its primary cause. (2) Inadequate correction of the deformity at the time of operation. Careful review of the immediate postoperative radiographs in some patients in this series showed that undercorrection had been either unrecognized or accepted. (3) Loss of correction postoperatively. If possible, the osteotomy should be made on the diaphyseal side of the enchondromatous area, so that the bone wedge will rest on normal corticocancellous bone. However, an intralesional osteotomy often is required because the lesion is so extensive and because maximum correction is achieved when the osteotomy is performed at the apex of the deformity. Frequently neither the bone wedge nor the metallic fixation had good inherent stability and collapse did occur. The patients must be monitored carefully by radiographs, and the osteotomy should be manipulated into a good position if position is lost. Although it might seem that overcorrection of the deformities could decrease the likelihood of the need for reparation, overcorrection rarely is done because few patients or families readily accept the substitution of one deformity for another. In addition, it is difficult to know how much overcorrection is desirable. 5. INVOLVEMENT OF THE UPPER EXTREMITY

There were two patients with involvement of the upper extremity. In one with bilateral upper extremity involvement, the left side was involved much more extensively. The lower extremities were normal radiographically. In one, the proximal end of the humerus was bowed markedly (50 ~ of valgus angulation) and an osteotomy to correct the deformity was considered, although it was not done. In the other, a distal ulnar resection osteotomy was done at the age of 13 years to correct radial deviation due to relative ulnar lengthening.

6. LOWER EXTREMITY LENGTH DISCREPANCY The extremity with the enchondromas always was shorter than the opposite one, and the extent of the shortening paralleled the radiographic involvement. There was no correlation between degree of angulation and extent of shortening. If the entire width of the distal femoral metaphysis was involved with enchondromatous tissue, the shortening was

820

CHAPTER 9 9 Skeletal Dysplasias

severe but angulation was absent. The maximum discrepancy reached prior to any surgical treatment for correction of discrepancy ranged from 4.3 to 35.7 cm (average = 9.8 cm). There was a progressive increase in the discrepancy in all patients throughout the growth period. The treatment for shortening of a limb is a challenge. The extent of enchondromatous involvement correlated well with the amount of shortening. The rate of inhibition of growth appeared to be constant in all patients. These observations allow for accurate estimates of the ultimate discrepancy. After the first few years of life, the need for osteotomy and perhaps repeat osteotomy will have become apparent. Although the epiphyseal arrests in this series did not always equalize the limb lengths, retrospective assessment indicates that growth remaining charts are useful, but even when the actual amount of correction needed from an arrest is known precisely, it is the deformed limb that must make up the deficit and its growth rate often is extremely slow. The affected bones often are so short that their length is at or below the second standard deviation below the mean. Therefore, the amount of correction to be gained must be extrapolated from the charts, and the arrests may have to be performed at an early age. If the projected discrepancy is 5 cm or less, well-timed epiphyseal arrest alone can produce excellent resuits, but when larger discrepancies exist the loss of height and the uncertainty of correction invite the use of other techniques. Diaphyseal lengthening of some of the involved long bones is a feasible alternative in some patients even though those with the greatest amount of shortening, who would benefit most from bone lengthening, often cannot undergo the procedure because the bulbous enchondromatous rnetaphyseal diaphyseal involvement does not provide a stable site against which the distraction device can pull. In patients in whom the projected discrepancy is 5 cm or more, however, diaphyseal lengthening for partial or full correction warrants strong consideration. Unsatisfactory results may be seen when, particularly in the last few years of the growth period, combinations of osteotomy, epiphyseal arrest, and diaphyseal lengthening are used. Imperfect results from one procedure, particularly diaphyseal lengthening, may require that an epiphyseal arrest be done but may obscure the appropriate timing for that procedure. If done too late, an arrest will not be fully effective. The bone lengthening operation is technically difficult at any age, but a bone studded with enchondromas and perhaps deformed distally or proximally or predisposed to recurrent deformation compounds the problem. A lengthening procedure may have to be coordinated with an epiphyseal arrest on the opposite side. a. EpiphysealArrest. Twenty standard epiphyseal arrests were performed for limb length discrepancy. When patients were referred when they were too old to obtain full correction, only limitation of discrepancies already present was attainable. Three one-sided stapling (medial or lateral) procedures were performed in an effort to correct angular defor-

mity, but each was ineffective. Retrospective reviews of the results achieved with epiphyseal arrest indicate that the corrections gained were what would have been expected based on the growth charts. b. Diaphyseal Lengthening. Five patients had six diaphyseal lengthening procedures, five tibial and one femoral. Lengthening was performed when the actual or projected discrepancy was considerable (6 cm or more). Five lengthenings were done in mildly to moderately involved bones and one was done in a normal tibia, originally the same length as the one on the opposite side. The femoral lengthening (Wagner technique) achieved a 9.0-cm increase. Two tibial lengthenings were done by the Wagner method and the other three by the Anderson method. The five tibial lengthenings achieved an average gain of 4.8 cm (range = 4.0-5.1 cm). Fourteen additional operative procedures associated specifically with the lengthening were required. With the Anderson technique, each patient had a tibiofibular synostosis several months before lengthening, and the pins were removed when healing was well-advanced. With the Wagner technique, secondary plating and bone grafting always was done and the plates were removed when healing was complete. In one patient an Achilles tendon lengthening was needed, and in another the plate fractured and had to be replaced. There were no instances of failure of consolidation of the bone after lengthening. Data on limb length discrepancies are presented in Table liB. The cases of five patients who did not have surgical treatment were not tabulated. One had so short an affected limb (35.7 cm at maturity) that he required a prosthesis. Three were young (4-8 years old) and were under observation, with discrepancies measuring 3.4, 1.3, and 1.0 cm. The fifth patient died of neuroblastoma at the age of 11 years. His discrepancy was 4.0 cm. 7. PATHOLOGICAL FRACTURES Seven of the 21 patients (33%) had an average of 2 pathological fractures each, with a range from 1 to 4. The femur was the site of fracture on 10 occasions and the tibia on 4 occasions. All fractures healed with conservative treatment in the usual amount of time. The fractures occurred throughout the period of childhood. Operative intervention should not be used simply because a pathological fracture is seen because the cartilage provides good stability, displacement is almost always slight, and healing occurs readily.

8. MALIGNANT TRANSFORMATION A major concern in any patient with Ollier's disease is malignant transformation of cartilage tissue at one site to a chondrosarcoma. The frequency of this occurrence is difficult to determine but is considered to be quite high in most studies. Certain guidelines are available. Malignant transformation essentially never occurs until after skeletal growth has ceased, and very few cases are reported under 20 years of age. The most common sites of occurrence, as for all other chondrosarcomas, are the proximal regions involving par-

SECTION XII ~ Review of Specific Skeletal Dysplasias

ticularly the proximal humerus, pelvis, and proximal femur. Distal pathology alone is not a particularly valuable method of diagnosis because the light microscopic appearance of an enchondroma in Ollier's disease even during the growing years is extremely abnormal, with some fearing that essentially all lesions represent at least a presarcomatous appearance. The more convincing findings of malignant transformation are three-fold involving an increase in the size of a lesion after skeletal maturity, discomfort at the site of the lesion when none was previously present, and the radiographic appearance of increased lysis of bone and disappearance of any calcification previously seen. Treatment involves the current protocols for chondrosarcoma. There is also an increased incidence of intracranial tumors, some of which are localized cartilage growths, but gliomas have also been described (48).

X. Maffucci Syndrome The Maffucci syndrome is characterized by multiple enchondromas and soft tissue hemangiomas, both cutaneous and of internal organs. The enchondromas are usually noted in childhood or adolescence and are structurally similar to those in Ollier's disease. It is not a genetic disease. Lower extremity length discrepancy occurs in many. A major additional concern is malignant degeneration with a predilection for intracranial neoplasms in about 15%. These can be chondromas or chondrosarcomas, but other neoplasms such as glioma, pituitary adenoma, or chordoma (217) are seen.

I(. Hereditary Multiple Exostoses 1. OVERVIEW Hereditary multiple exostoses is one of the most common of the skeletal dysplasias. It is an autosomal dominant disorder characterized by widened metaphyses and multiple exostoses or osteochondromas throughout the skeleton, specifically positioned at the periphery of the growth plates and metaphyses at which the endochondral sequence of bone formation and the intramembranous sequence of bone formation are in close relationship (26, 29, 68, 69, 88, 137, 170, 226, 239, 287, 305). The disorder was referred to as diaphysial aclasia in the older British literature (72). It has long been recognized not to affect bones formed exclusively by the intramembranous mechanism such as the cranium and face or by the endochondral mechanism such as the carpal and tarsal bones. Its clinical and pathological features were well-described in the late nineteenth and early twentieth centuries; as early as 1915, Ehrenfried of Children's Hospital, Boston, was able to review over 300 articles presenting about 600 cases (68, 69). The disorder rarely becomes clinically evident before 2 years of age. Virtually all patients require orthopedic surgical intervention during the growing or young adult years, mostly to remove troublesome exostoses or to correct

821

characteristic deformities that accompany the condition. Although exostoses are seen at the metaphyses of almost all of the long bones, the severe and clinically significant deformities involve the forearm and the leg, regions characterized by two bones in close longitudinal relationship. The elbow, wrist, knee, and ankle are particularly vulnerable to deformity. In the majority of cases, the ulna and fibula are disproportionately short in relation to the radius and tibia. The longer of the two bones is concave toward the interosseous membrane and toward the shorter bone, so that it seems reasonable to ascribe a tethering effect of the interosseous membrane and ligaments in association with a disparate growth rate as causative factors in deformation. The epiphyseal changes at the distal ends of the tibia and radius are analogous, with the distal articular surfaces tilted toward the shorter bone and the epiphyseal shape secondarily altered. 2. HISTOPATHOLOGY The exostoses are invariably broad-based in association with a widened metaphysis. When the exostoses are present as elongated projections, the tips invariably are directed toward the diaphysis and away from the epiphysis (Fig. 25A). The exostoses and the widened metaphyses enlarge with growth coming via the endochondral mechanism from the cartilage cap. The cartilage coveting is fibrocartilaginous at its surface but forms a physis immediately adjacent to the bone it is producing (Fig. 25B). The endochondral bone and marrow of the exostosis then merge with those of the adjacent metaphysis without any cortical intervention. The cartilage cap stops growing at the same time that the adjacent growth plate fuses. At that time, the cartilage cap thins and involutes, the exostosis is now covered by a thin rim of cortical bone, and any persisting cartilage is converted to a fibrous tissue coveting. The close relationship of the exostoses to the physeal growth mechanism is further shown by the fact that the largest and most numerous exostoses are found at the most rapidly growing bone regions: the distal femur, proximal tibia, proximal humerus, and distal radius and ulna. Any increase in the size of an exostosis occurring after skeletal maturity raises a high level of concern about malignant transformation. 3. PATHOGENESIS Many theories regarding the pathogenesis of the osteochondromas have been presented but they have been difficult to verify (64, 131, 137). There are few human specimens of sufficient size that include the exostosis and the periphery of the physis from a growing child, and little experimental work has been done. The osteochondroma lesions (1) invariably occur in bones that develop by endochondral ossification and intramembranous bone formation together, (2) are initiated at the periphery of the growth plate region where the intramembranous sequence relates most closely to the endochondral sequence, and (3) cease growing at the same time the growth plate fuses. They seem, therefore, to be

822

CHAPTER 9 ~

Skeletal Dysplasias

F I G U R E 25 Radiographic examples of hereditary multiple exostoses are shown. (A) The characteristic abnormality in hereditary multiple exostoses comprises not only widening of the metaphysis but also its association with randomly placed exostoses (osteochondromas) whose tips point away from the epiphysis. (B) Photograph of a large osteochondroma removed from a patient with hereditary multiple exostoses. At surgery there are many undulations noted at the periphery of the metaphysis and adjacent to the largest of the exostoses. In the growing child, these are covered by thick cartilaginous caps and generally are more prominent to gross inspection at surgery than was apparent radiographically because the plain radiograph shows only the underlying bone. [Reprinted from (283), with permission of the American Academy of Orthopaedic Surgeons.] (C) Theories of pathogenesis of the hereditary multiple exostoses lesions are illustrated. (D) Proximal fibular exostosis (arrow) which led to a gradual but complete peroneal nerve compression. This resolved slowly, but fully after surgical debulking of the osteochondromatous region and freeing of the peroneal nerve. (E) The occurrence of exostoses throughout the skeleton in a study of 32 patients from Children's Hospital, Boston is shown. (Fi) Exostosis of the medial border of the scapula (arrow) is shown on radiograph. (Fii) These are best demonstrated by CT scanning. (Gi) Characteristic lower extremity deformities in hereditary multiple exostoses include proximal tibia valga. Correction is by proximal tibial-fibular varus osteotomy as seen in (Gii) and (Giii). Osteotomy site must be below the expanded metaphyses to prevent peroneal nerve damage. (H) Determination of the degree of tibia valga in hereditary multiple exostoses is shown. (Ii) Degrees of obliquity in hereditary multiple exostoses at the ankle are shown (type 0 to type III) along with radiographs of three cases (Iii-Iiv). (J) Altered relationships of the distal fibula to the distal tibia at the ankle in hereditary multiple exostoses are shown. Normally the distal end of the fibula is distal to the distal end of the tibia, a position that helps to stabilize the ankle. (Ki-Kv) Characteristic deformities of the upper extremity in hereditary multiple exostoses involve primarily the forearm. The ulna is relatively more shortened than the radius

SECTION XII ~ Review o f Specific Skeletal Dysplasias

F I G U R E 25 (continued) and its distal end has a Madelung-like deformation. Multiple radiographs show the variation in deformity that occurs. (Li-Liv) Correction of angular deformity of the distal forearm was performed with distal radial closing wedge and derotation osteotomy. Regional exostoses also were removed. This improved both the clinical appearance and function of the forearm and wrist. (M) A pseudo-aneurysm resulting from pressure from a distal posteromedial femoral exostosis is shown. Most of these vascular lesions develop in midadolescence and early adulthood when involution of the cartilaginous cap of the osteochondroma leaves the underlying bone spicule more prominent. (Mi) Lateral radiograph shows the prominent bone spur at the posteromedial aspect of the distal femur in a 17-year-old boy who had discomfort and gradually increased swelling of the posterior thigh over the previous 6 weeks. (Mii, Miii) Anteroposterior and lateral arteriograms show the femoral artery and a circular collection of dye at the tip of the exostosis at the site of the pseudo-aneurysm. Surgery removed the exostosis followed by segmental excision of the pseudo-aneurysm and arterial repair with a venous graft. (Miv) A postoperative radiograph shows the operative site. [Parts E, H, and Iiv reprinted from (287), with permission.]

823

824

CHAPTER 9 ~

Skeletal Dysplasias

Type 0 Normal

) Type I Mild Abnormafity

Type II Moderate Abnormality

Type III Severe Abnormality

FIGURE

25

(continued)

SECTION XII ~ Review o f Specific Skeletal Dysplasias

825

FIGURE 25 (continued)

closely related to the growth plate apparatus and are considered as perversions of its normal state. Five main theories of pathogenesis (64, 137) (Fig. 25C) have been mentioned repeatedly: (1) Virchow proposed that some growth plate cartilage becomes split off but continues to grow in its abnormal position (348). Growth of the exostosis, therefore, occurs perpendicular to the long axis of the bone. (2) Mueller considered that a periosteal metaplasia in the periphyseal-perimetaphyseal area produced the characteristic lesions such that cells normally defined as bone producing were acting as chondrocytes (211). (3) Keith alluded to inadequacies of the periosteal bone sheath, which he felt allowed for greater than normal outward expansion of endochondral bone (144). (4) Jansen noted that the lesion was most marked by a broadened metaphysis, which indicated a failure of tubulation (resorption) (137). (5) Langenskiold felt

that perichondrial bone of the groove of Ranvier, normally derived in his opinion from tissue within the epiphysis passing outward and forming an osteogenic layer, failed to form because that tissue pathologically retained its chondrogenic properties and was abnormally deposited as a cartilaginous and not osseous mass (162). The failure of deposition of normal groove of Ranvier tissues leads to failure of the normal processes of periphyseal-perimetaphyseal bone formation, tubulation, and remodeling. It appears likely that elements of all five theories could indeed theoretically contribute to the pathogenesis, although the pathoanatomy is consistent with Mueller's theory as the primary cause with those of Keith, Jansen, and Langenskirld accounting secondarily for failure of metaphyseal remodeling. Proximal fibular exostoses (Fig. 25D) demonstrate structural abnormalities well in a reasonably small focus.

826

CHAPTER 9 ~

F I G U R E 25

Skeletal Dysplasias

(continued)

SECTION XI! ~ Review of Specific Skeletal Dysplasias

827

FIGURE 25 (continued)

Separate reports by Leveuf (173) and Ogden (225) have studied pathological tissue from osteochondromas along with adjacent physeal tissue obtained from growing children by biopsy. The photomicrographs show structural differences in physeal and exostosis lesional cartilage even when those are juxtaposed. The perichondrial groove tissue is absent from its normal position adjacent to the physis and metaphysis. It remains unclear whether the osteochondroma cartilage is present due to uncontrolled lateral growth of peripheral physeal tissue or whether groove tissue has differentiated fully to cartilage rather than to bone as it normally does, but the latter interpretation seems more likely. The osteochondroma cartilage cap tissue undergoes the endochondral sequence, merging lesional bone tissue with endochondral bone from the regular physis and thus causing the widening of the metaphysis. The abnormally structured groove region prevents both normal physeal control by the periosteal tissue and

resorption by osteoclasts. Ogden has shown that the groove tissue is normal in those parts of the perimeter of the physis that are not involved in the osteochondromas, and Leveuf has shown that the normal groove tissue is reconstituted well below the physis where metaphyseal widening is no longer present and appropriate tubulation has been reestablished. Leveuf obtained two large biopsies from the distal femur and proximal tibia of a 7-year-old child with hereditary multiple exostoses (173). The biopsy removed a tissue segment in continuity from the periphery of the physis and the widened metaphysis down to the region of the diaphysis where normal tubulation was reestablished. The tissue segment thus included the exostosis plus the normal adjacent physeal and metaphyseal tissue. The histologic and radiographic study noted that there was a considerable distance between the growth cartilage and the origin of the diaphysis because the widened metaphysis was the site not only of the exostosis

828

CHAPTER 9 ~

Skeletal Dysplasias

FIGURE 25 (continued)

but also of the failure of remodeling. Leveuf noted that two types of exostosis could develop: the wide, flattened, sessile variant, which tended not to elongate with bone growth, and

the narrow pedicle type of exostosis, which did increase its length with bone growth. Between the physis and the beginning tubulation of the diaphysis the wall of the metaphysis

SECTION XII ~ Review o f Specific Skeletal Dysplasias

was malformed and composed of cartilaginous plaques, which gave rise to endochondral bone on their inner surface. It was not until the ossification groove of Ranvier was reestablished distal to the widened metaphysis that normal remodeling occurred and the groove tissue synthesized cortical bone. The most characteristic finding between the growth plate cartilage, which was normal, and the diaphyseal cortex, which also was normal but a considerable distance from the physis, was the presence of the metaphyseal wall filled with cartilaginous inclusions, which were thin and flat. The cartilaginous wall was covered by a perichondrium continuous without transition from the perichondrium of the epiphysis above and the periosteum of the diaphyseal cortex below. The metaphyseal wall cartilage was immediately adjacent to the growth plate cartilage but was separated from it by a segment of connective tissue formed of fibers seemingly detached from the perichondrium and oriented in a transverse dimension. Below the cartilaginous metaphyseal wall was contiguous with the origin of the diaphyseal cortex, and at their point of uniqn a normal appearing perichondrial ossification groove as described by Ranvier was seen. He felt that the cartilage in the metaphyseal wall certainly was the developing point of the sessile exostoses. On the inner surface of this cartilage growth was occurring by the endochondral mechanism, which merged with the adjacent metaphyseal bone. Leveuf's careful description of a well-oriented biopsy from a child in the active growth phase with hereditary multiple exostoses clearly shows the absence of the normal perichondrial ossification groove and its replacement by cartilaginous tissue with endochondral growth function. The groove is repositioned at a greater distance from the physis than normal, and where it is present metaphyseal widening abruptly ceases, tubulation or funnelization begins, and normal cortical bone is formed. Leveuf's study indicates that although the physeal cartilage, which is normal, is immediately adjacent to the cartilage of the exostosis the two are separated by a thin patch or region of connective tissue. [Leveuf used this histopathological example to support the bone growth theory of Policard that growth in length of a long bone was triggered primarily by the periosteal bone particularly at either end by the ossification groove bone tissue and that the physeal tissue merely elongated secondarily. This view has never been widely accepted and indeed appears to be inaccurate, but this does not negate his highly accurate description of the periphyseal and perimetaphyseal tissues in hereditary multiple exostoses and the interpretation of the pathogenesis of those specific lesions.] Ogden was able to study a proximal fibula surgically removed from a 4-year-old girl with hereditary multiple exostoses in the course of treatment of a peroneal nerve palsy (225). The crucial juxtaposed regions composed of the normal physis and the cartilage of the osteochondroma were studied. A normal growth plate was found under the lateral 75% of the chondroepiphysitis of the proximal fibula, at which point an abnormal physis of the osteochondroma ap-

829

peared although it was still adjacent to the cartilage of the epiphysis. In different sections "this osteochondroma was directly juxtaposed to the normal growth plate, separated from the normal growth plate by a fibrous (never bony) zone or separated from the normal growth plate by an area of abnormal cartilage proliferating into the epiphyseal cartilage." The abnormal osteochondroma or exostosis cartilage "continued under the medial 25% of the epiphysis and then turned 90 degrees to course distally along the metaphysis." The osteochondroma was always separated from the epiphysis by an intervening fibrous or fibrocartilaginous zone. The cartilage of the exostosis had a considerable increase in the amount of intercellular substance throughout the various zones, and the cells tended not to be arrayed in normal columnar fashion. Despite this, endochondral ossification proceeded in a relatively normal fashion from the cartilage cap. Milgram also demonstrated that the osteochondroma was derived from aberrant cartilaginous growth plate tissue, which appears to proliferate independently and separates from the growth plate near its edge. The cartilage remains sub-periosteal and may continue to proliferate as a young osteochondroma perpendicular to the growth plate from which it derived (204). He illustrates the distal radial growth plate from a 12-year-old boy with HME. Other examples show the lesional osteochondroma cartilage to be clearly differentiated from a normal physeal cartilage, although it is immediately juxtaposed to it. The pertinent histology of the growth plate and its surrounding tissues must be considered. Long and flat bones that develop via an endochondral mechanism also have an associated periosteal membrane bone component. Of importance is the region in which the two developmental sequences are closely associated. The periosteal membrane bone sequence completely ensheathes the cartilage growth plate, being present in the deepest part of Ranvier's perichondrial groove and attaching, via its outer fibrous layer, to the epiphyseal cartilage beyond the plate. Ranvier's perichondrial groove is depressed into the periphery of the growth plate cartilage. At its deepest part is a dense accumulation of cells, which are continuous with the inner osteogenic layer of the periosteum. Alkaline phosphatase stain outlines this area exactly showing the periosteal membrane bone sequence to be in advance of the adjacent endochondral sequence. Safranin O stain also highlights the differences, with the red cartilage juxtaposed to the green osteoid of the periosteal sequence. Fibers continuous with the outer fibrous layer of the periosteum anchor the periosteum into the epiphyseal cartilage. The dense accumulation of cells is responsible for producing cortical membrane bone. Lacroix calls this the perichondrial bony ring of the ossification groove, noting that it is not continuous with the diaphyseal cortical bone. It is resorbed during the metaphyseal funnelization process such that the fibrous layer of the periosteum comes to lie against metaphyseal bone. Transverse sections show how the periosteal bony ring and osteogenic layer diminish, whereas coverage of the metaphysis by the fibrous

830

CHAPTER 9 ~ Skeletal Dysplasias

periosteum persists. The association of osteoclasts with the resorptive processes is seen. Clearly there is a mechanical and biological controlling function for the periosteum and its bony ring in relation to the physis and adjacent metaphysis. In an effort to assess the control features of the perichondrial sleeve, a series of experiments was performed in our laboratory by Ellis and Shapiro (284). The osteochondromas to be described were caused experimentally solely from defects created by entirely removing relatively large periosteal flaps or windows in 10-day- to 6-week-old rabbits. Most of the defects were made at the lateral aspect of the distal femoral growth plate and metaphysis. A transverse cut was made in the periosteum parallel to and 1 cm from the growth plate on its metaphyseal side. Two incisions, about 0.5 cm apart, passed from the transverse incision to the plate. The periosteum was then lifted and easily peeled toward the plate. The growth plate cartilage was uncovered, and the fibrous attachment into the epiphysis beyond the plate was found to stop further periosteal displacement. The periosteal flap was then removed by sharp dissection here, creating a large, windowlike defect. Histological sections were made on the removed periosteal flap. The Safranin O stain of the removed flap shows the green layer of fibrous connective tissue, the green periosteal membrane bone, and flecks of metaphyseal bone-central red cartilage core surrounded by green staining bone. The tissue remaining, therefore, is clearly devoid of its periosteal sheath and of the bony ring. On some occasions, careful dissection still showed periosteal tissue to be present, which probably accounts for the failure to induce osteochondromas on all occasions. The times of sacrifice varied from 3 to 8 weeks, and tissue created in response to the defects was assessed by gross inspection and histological sections. Some were radiographed. In human hereditary multiple exostoses, all cases show metaphyseal widening due to a failure of funnelization as well as localized broadly based osteochondromas standing out from the bone. In our experiment 41 of 62 defects (66%) resulted in osteochondromas. A normal nonoperated bone demonstrates the metaphyseal flaring and is compared with an operated femur. The lesion is prominent. Anterior and posterior views of other lesions are seen. Both sides have lesions in this case. The animal lesions and radiographs are similar to the human cases. The histological appearance of the induced lesions is crucial to their interpretation. The human exostoses are covered with hyaline cartilage, their metaphyseal bone is continuous with the primary metaphyseal bone (not positioned against cortical bone), and the bone develops by endochondral ossification. Histological sections of the induced lesions show three specific responses: (1) at the defect site the metaphysis has not been remodeled but has remained wide and blunt. The endochondral bone has persisted and the non-remodeled area is composed of metaphyseal bone showing central cartilage cores. Funnelization has occurred normally, however, around the rest of the circumference of the distal femur where the periosteum is in-

tact. (2) Many of the osteochondral lesions possess cartilage "caps," which are composed of hyaline cartilage. This cartilage is producing bone by evolving into a well-defined endochondral sequence. (3) The true cortex is absent between the metaphysis and the osteochondral lesion. The cancellous metaphyseal bone from the growth plate is in continuity with the new endochondral bone of the lesion. Comparable experimental findings in the rabbit were described by Hwang and Park (131). Experiments like this attempt to demonstrate the pathogenesis of a developmental abnormality using a surgically created defect in a young otherwise normal animal. The response to what is a traumatic lesion understandably is intense and florid. Nevertheless, the responses unleashed result in the formation of osteochondromas histologically similar to those seen in humans. The postulate that the osteochondromas of hereditary multiple exostoses occur secondary to a physical or functional defect in the periosteal tissue, which normally ensheathes the entire growth plate, appears to have good experimental as well as theoretical backing. The sequence of events is such that (1) the defect allows the endochondral sequence to "escape" because the physical controlling effect of the periosteal sheath including the bony ring and its cellular resorptive effect are both lacking; (2) further endochondral growth of the osteochondroma lesion occurs secondary to the adjacent periosteal reaction, which in response to the defect has undergone a metaplastic change into cartilage; and (3) continuity between the normal cancellous metaphyseal bone and the cancellous bone of the lesion occurs because cortical periosteal bone formation at the defect site is prevented due to the local absence of the periosteum. The induction of osteochondromas by a simple interference defect experimental method serves both to define some of the normal biological functions of the periosteum and to account for the pathogenesis of a well-known entity. A group of experiments was performed by D'Ambrosia and Ferguson attempting to form osteochondromas experimentally by epiphyseal cartilage transplantation (64). Characteristic osteochondromas such as those seen in hereditary multiple exostoses were produced most extensively in the metaphyseal regions and specifically in that portion of the metaphysis closest to the epiphyseal growth plate. Transplants did not produce any growth when they were placed farther than 1.5 cm from the epiphyseal growth plate. Transplants from the periphery of the epiphyseal growth plate, central cores of metaphyseal and epiphyseal tissue, and even some articular cartilage transplants were able to produce exostoses with the cartilage cap, the chondral ossification, and communication with the adjacent metaphyseal bone. The experiments were performed in New Zealand white rabbits at the region of the proximal medial tibia. The transplant was made through a surgically created defect in the periosteum and immediately adjacent to a small hole in the cortical bone made with a drill approximately the same size as the transplant specimen. No direct communication was made with the intramedullary bone. Control on the opposite leg

SECTION Xil ~ Review of Specific Skeletal Dysplasias included the periosteal defect and cortical hole without transplant. In each instance, the farther down the tibia the transplant was placed, the less the chance of producing growth. Polarity of the physeal tissue was crucial to osteochondroma-like growth, and true osteochondromas with typical modeling were produced only when the physeal orientation was maintained such that the hypertrophic zone regions of the physis were adjacent to the metaphysis. When the tissue was reversed only a cartilage island was formed. The authors felt that their transplantation experiments supported Zirchow's theory on the formation of osteochondromas. Rotation less than 90 ~ from a normal plane was essential, however (64). 4. GENERAL STATURE The anthropometric data indicate that persons with hereditary multiple exostoses are of shortened stature but that they almost all fall within the normal range, with actual dwarfism being rare. The short stature is mildly disproportionate, with limb involvement greater than spinal involvement. The outstretched fingers reach to just below the trochanters in moderate and severe cases, rather than to the mid-thigh region as in the normal. There is no uniform inconsistency between skeletal age and chronological age during development. The distribution of general stature in the 30 patients documented in our series is shown (287). In only 2 of the 30 patients was the general stature less than the second standard deviation below the mean, and one of these patients was skeletally immature at the time of final assessment with a skeletal age 2 years less than chronological age. Growth data are represented on charts that depict the range of sizes seen throughout a particular population at various ages. The average of all individuals measured at one particular age is referred to as the mean. For charting and statistical purposes, about 68% of all measurements are considered to fall between 1 standard deviation above and 1 below the mean, and about 97% fall between 2 standard deviations above and 2 below the mean. Thus, only an individual whose stature is below the second standard deviation below the mean could be considered as abnormally small or dwarfed, because stature within the range of 2 standard deviations is arbitrarily considered to be normal. Seven of the patients were of average height, 17 were below the mean, and 6 were above the mean. Thus, the majority of patients were below average in height, but almost all remained within the normal range and only 1 was outside the normal distribution with a final height at 2.5 standard deviations below the mean. Ehrenfried noted that "a shortness of stature is practically constant. The patients are always of medium height or below, and sometimes they a r e . . , d w a r f e d . . . " "The lack of growth is in the legs, and not the trunk . . . . The upper limbs are also short (68)." Separate growth charts that depict sitting height and overall height are available. In 26 of 30 patients, the sitting height was greater than the overall height in terms of the standard deviation as related to the mean. In 3 patients both values were along the same standard deviation, and in one overall

831

stature was relatively greater than sitting height. The sitting height averaged 1.0 standard deviation above overall height. These findings also indicate a mildly disproportionately short stature, with the limbs involved to a greater extent than the spine. 5. CLINICAL PROBLEMS IN HEREDITARY MULTIPLE EXOSTOSES There are six groups of clinical problems that may be seen in patients with hereditary multiple exostoses. These include (1) the exostoses, which can be uncomfortable or unsightly or both; (2) angular deformities of the limbs; (3) length discrepancies of lower and upper limbs; (4) pressure phenomena caused by the exostoses; (5) fracture of the exostoses; and (6) malignant degeneration of the exostoses. Although this skeletal dysplasia does not shorten life expectancy, except in rare patients with malignant degeneration, it does cause considerable difficulty. In the 32 patients in our study, approximately 2.7 surgical procedures per patient had been performed before the age of 20 years. This number is considered to be on the low side, because some of the children with large exostoses and considerable deformities were still being followed at the time of assessment and others had moved. It was the exostoses themselves that led to the greatest number of operations, especially at the knee, wrist, and shoulder. The exostoses of the hand and foot rarely presented sufficient clinical symptoms to warrant excision. a. Occurrence o f Exostoses. The exostoses are present bilaterally in virtually all long tubular bones and frequently in the pelvis, scapula, fibs, and short tubular bones of the hand and foot (Fig. 25E). They occur at the periphery of the epiphyseal growth plate and metaphysis, which is the region of intimate relationship between the intramembranous and endochondral developmental sequences. The epiphyses and round bones of the carpus and tarsus, which develop exclusively by endochondral ossification, are not involved at least primarily. In the upper limb, all lesions were metaphyseal (287). Exostoses appeared almost invariably at the proximal end of the humerus (49/50 or 98%), at the distal end of the radius (36/45 or 80%), and at the distal end of the ulna (35/ 41 or 85%). Lesions of the distal end of the humerus were quite rare, appearing in only 2 (5%) of 39 humeri, and lesions of the proximal end of the radius (14/37 or 38%) and proximal end of the ulna (14/38 or 37%) were relatively infrequent. In the lower limb, all lesions were also metaphyseal (279). They occurred almost invariably at the proximal end of the femur (56/62 or 90%), at the distal end of the femur and proximal end of the tibia (in both cases in 61/62 or 98%), at the proximal end of the fibula (60/62 or 97%), at the distal end of the tibia (57/62 or 92%), and at the distal end of the fibula (53/62 or 85%). Roentgenograms of other regions showed exostoses in the scapula, ribs, pelvis, vertebral spinous processes, metacarpals, and phalanges. The pelvis, scapula (Fig. 25F), and ribs are commonly affected. Large exostoses frequently emanate from the iliac crest. CT scanning is advisable prepregnancy

832

CHAPTER 9 ~ Skeletal Dysplasias

in affected females to assess further any pelvic canal lesions present by plain radiographs. Lesions on the anterior scapula commonly at the medial or vertebral border often are symptomatic, impeding smooth passage over the adjacent fibs, and should be removed.

b. Deformities of the Lower Limb. Lower Extremity Studies: All roentgenograms were examined for the presence or absence of deformities found in hereditary multiple exostoses: coxa valga, femora vara, tibia valga (Fig. 25G), obliquity of the articular surface and subchondral bone of the distal tibial epiphysis, and obliquity of the superior articular surface of the talus. Each of these deformities was assessed along with the lengths of the tibia and fibula, the relationship of the distal end of the tibia (medial malleolus) to the distal end of the fibula (lateral malleolus), and limb length discrepancies. Although the Green-Anderson table of normal values was used as a standard for limb lengths (femur and tibia), no normal values have been reported for the lengths of the tibia and fibula, including the malleoli, or for the various angles that we wished to measure to assess deformity. For this reason, lower extremity orthoroentgenograms from 50 patients with fractured femurs who were being followed in the Growth Study Clinic for limb length discrepancies were used to obtain normal values for the lengths of the tibia and fibula and the associated bone relationships and angles. The degree of tibia valga was assessed as the angle formed between the epiphyseal growth plate and the long axis of the medial cortical surface of the tibia (Fig. 25H). The obliquity of the subchondral surface of the distal tibial epiphysis was noted not to involve the entire surface but usually to begin in its midportion. Where obliquity was noted the medial half of the surface was level, whereas the lateral half was found to slant upward and outward. The degree of obliquity was graded as types O, I, II, and III (Fig. 251). When the entire surface was perfectly level, as in the normal state, the classification was type O. Type I indicated that the obliquity from the central portion of the articular surface of the distal tibial epiphysis angled upward and laterally but the lateral margin of the subchondral bone remained well within the epiphyseal side of the growth plate. Type I appeared as a minimal deformity but was seen commonly enough to be considered a mild variation of normal. This group was a transitional group, with some ankles so graded being very close to type O and others approaching the type II pattern. For purposes of grading, however, because the deformity was quite small it was included with type O and was defined as normal. Type II deformities were those in which the distal subchondral bone sloped into the lateral margin of the epiphyseal growth plate. Type III lesions were those in which the subchondral bone slanted into the epiphyseal growth plate in its lateral one-third rather than at its lateral edge. Obliquity of the superior surface of the talus was considered to be either present or absent. Normally, the surface is fiat; with obliquity the lateral aspect of the surface is more

superior than the medial aspect. This assessment was made to distinguish between a talus that was tilted in the angle mortise and one with obliquity of the superior articular surface. Tibia-Fibula Relationships: The tibia was measured from the plateau to the distal tip of the medial malleolus, and the fibula was measured from the most proximal portion of the head to the most distal tip of the lateral malleolus. The total fibula-tibia length ratio was calculated. This measurement was performed on all of the exostosis patients and on the 100 normal control legs from patients with fractures. The tibial measurement in this instance differed from that used on the Green-Anderson charts, as we wished to include the malleoli. The relationship of the lateral malleolus to the medial malleolus was assessed in 100 normal ankles and in 64 patients with hereditary multiple exostoses. The distal end of the fibula was observed to be distal to, at the same level as, or proximal to the distal end of the tibia (Fig. 25J). Because the proximal ends of the fibula and tibia are at different levels, it was necessary to assess both the tibiafibula length ratio and the relationship of the lateral to the medial malleolus. We wished to establish whether the differences at the ankle were due solely to altered position of the two bones or whether relative shortening of the fibula was also a factor. A correlation was also made between the relationship of the distal ends of the tibia and fibula and the degree of obliquity of the articular surface at the distal end of the tibia. The deformities of the lower limb noted in our series follow. Coxa Valga: The proximal end of the femur was assessed in 64 patients, and there were 16 cases of coxa valga (25%). On occasion, acetabular dysplasia was associated with the coxa valga. In 46 patients the head-neck-shaft relationship was normal despite the presence of exostoses. Coxa vara was not seen. Femora Vara: There were four instances of distal femora vara. Femora valga did not occur. Tibia Valga: This deformity occurred within the tibia; the term does not refer to femoral-tibial valgus angulation. The apex of the tibial valgus deformity was metaphyseal, just below the growth plate. The measurement of deformity refers to deviation into valgus. Normally, with no deformity the long axis of the medial tibial cortex is parallel to a line drawn perpendicular to the proximal tibial growth plate, that is, 0 ~ In the normal patient controls, 4 (8%) of 50 had a minimum deviation of 5 ~ and none had values greater than 5 ~ Thirty had no deviation (0 ~ and 16 had values between 1 and 4 ~ In the patients with hereditary multiple exostoses, 20 (33%) of the 60 tibiae had a tibia valga of 5 ~ or greater. In these 20, the tibia valga ranged from 5 to 25 ~ with a mean of 10~ Tibia vara was not seen. Obliquity of the Distal Tibial Epiphysis: This obliquity refers to the articular surface and subchondral bone of the distal end of the tibia and not to the epiphyseal growth plate. The growth plate itself was perpendicular to the long axis of

SECTION Xll ~ Review of Specific Skeletal Dysplasias the tibia in both normal and exostosis patients. A classification of the degree of obliquity and associated abnormality of the ankle mortise was outlined. Of 61 ankles assessed in the patients with hereditary multiple exostoses, 5 were type O (normal), 23 were type I, 20 were type II, and 13 were type III. Thus, 54% were moderately to markedly abnormal (types II and III). In type III ankles in exostosis patients, lateral talar subluxation was seen occasionally. Of the normal ankles, 73 were type O, 25 were type I, and only 2 were type II.

Intrinsic Obliquity of the Superior Talar Articular Surface: This term does not refer to a tilt of the talus but to an intrinsic obliquity of the superior talar articular surface. In the exostosis patients, the lateral region of the talar articular surface was more superior than the medial region in 42 (70%), whereas in 18 (30%) the surface was level. In the normal patients the lateral surface was superior in only 20 (20%) with the other 80 (80%) being level. Length and Distal Position Relationship of Tibia and Fibula: The fibula-tibia length ratio was 1.00 in the normal group of 58 legs (range = 0.98-1.03) and 0.94 in the exostosis group of 64 legs (range = 0.86-1.01). The numerical ratios were determined prior to any corrective osteotomies and prior to the one proximal tibial-fibular epiphyseal arrest that was done. This ratio documents the shortness of the fibula relative to the tibia. In the normal controls, the lateral malleolus was distal to the medial malleolus in all (100%) of the 100 ankles that were assessed. The fibular growth plate was at the level of the tibiotalar joint space. In the patients with hereditary multiple exostoses, the fibula was distal to the tibia in 28 (44%), at the same level in 23 (36%), and proximal to it in 13 (20%). There was a good but not absolute correlation between the obliquity of the articular surface of the distal end of the tibia and the relative positions of the two malleoli. Type II deformities predominated when the fibula equaled the tibia in length, and type III deformities predominated when the fibula was shorter than the tibia. Neither the coxa valga nor the femoral vara presented clinical problems. The valgus configuration of the proximal femur was seen in each of 50 hips in 25 patients with hereditary multiple exostoses. Weiner and Hoyt also described development of the proximal femur in HME (356). Tibia valga was considerable, and 20% of the patients had varus osteotomies of the proximal end of the tibia and fibula. This operation is always a source of potential difficulty due to the closeness of the vascular trifurcation and the peroneal and posterior tibial nerves to the osteotomy site. The risk is further increased in patients with hereditary multiple exostoses because the exostoses themselves frequently compromise the space available for the neurovascular structures. One transient peroneal nerve palsy occurred in this series. In HME the tibial osteotomy is performed at a more distal level than normal because of the metaphyseal widening, a closing wedge procedure with slight shortening is favored, and gentle debulking of the proximal fibula is considered based on

833

concern about the size, shape, and position of any exostoses there. Nawata et al. also showed tibial valgus angulation increasing with growth in HME (218). Lateral subluxation of the patella was also seen in a few instances. The degree and frequency of altered bone relationships at the ankle were documented. The ankle abnormalities that present clinically as a pes valgus are associated with shortness of the fibula relative to the tibia, the more proximal relationship of the lateral malleolus to the medial malleolus, obliquity of the distal tibial epiphyseal surface, and occasionally lateral subluxation of the talus. Partial compensation to help to form a more regular ankle joint was provided by developmental obliquity of the superior talar articular surface. The pes valgus was often roentgenographically marked, and two patients had a clinical problem severe enough to warrant a supramalleolar varus osteotomy at the distal end of the tibia and fibula in the first two decades of life. It would be expected that other patients with tibia valga plus a marked valgus deformity at the ankle may have difficulty with this during later years, but a lack of symptomatic and functional problems in the first two decades of life is generally seen. A certain degree of control can be expected from stapling of the distal medial tibial epiphyseal growth plate once there is early indication that a deformity appears likely to occur. Snearly and Peterson have reported on operative procedures for ankle deformity in nine patients (301). The indications for intervention were pain, angular deformity, and limitation of motion. Procedures used included removal of the osteochondroma, which eliminated discomfort and improved appearance but could not be counted on to correct the deformity, lengthening of the fibula, and asymmetric medial stapling of the distal tibial growth plate. Fibular lengthening was a one-stage step-cut lengthening in severe cases with gains of 12 and 17 mm. Asymmetric medial stapling allowed for a mean correction of tibial-talar tilt of 10.3 ~ in three cases. With this approach, distal tibial osteotomy had not been needed. c. Deformities of the Upper Limb. Deformities of the forearm in hereditary multiple extososes are characteristic with the degree of the deformity primarily related to the amount of shortening of the ulna referable to the length of the radius and the position of the exostoses of either bone. Characteristic deformity involves a relative shortening of the ulna with malformation of its distal end leading to an ulnar drift of the wrist and hand, obliquity of the distal radial articular surface with angulation toward the shortened ulnar side, bowing of the radius with the concavity toward the shortened ulna, and, in many instances, subluxation or dislocation of the radial head at the elbow and dorsal subluxation of the distal end of the ulna in a Madelung-like deformation (Fig. 25K). The tendency to radial head dislocation is increased in those in whom there is more extensive shortening of the ulna and limited to absent bowing of the radius. Pronation and supination are variably affected

834

CHAPTER 9 ~

Skeletal Dysplasias

depending on the presence and size of ulnar and radial exostoses, the degree of bowing of the radius, and the presence or absence of radial head dislocation. In our series there were 8 dislocations of the radial head in the 37 elbows that were assessed (22%) (279). In the 37 roentgenograms of an entire forearm that were available for review, the degree of deformity was normal or mild in 12 (40%), moderate in 6 (20%), and severe in 12 (40%). The distal end of the ulna invariably was involved more seriously than the distal radius. Measurements of pronation and supination were available in 26 forearms. Pronation and supination were normal or only mildly restricted (0-70 ~ or more) in 33 (63%) of 52 instances, moderately restricted (0-60 ~ in 11 (21%), and severely restricted (0-30 ~ in 8 (15%). These ranges of movement refer to pronation and supination separately and not to their total range. Usually either pronation or supination was affected, with the movement in the other direction fairly free. Pronation was the restricted direction of movement more frequently than supination by a margin of 2:1. Radial deviation was limited and ulnar deviation was increased. In those individuals for whom values were recorded, radial deviation averaged 12~ and ulnar deviation 45 ~. The deformities present at the various regions are quite consistent, the main difference being one of degree. Some progress slowly whereas others remain stationary. No deformity, once identified, improved with growth. Over half of the patients with hereditary multiple exostoses had moderate to severe deformity of the forearm, and one-third had significant restriction of either pronation or supination. The characteristic deformities of the upper extremity were treated surgically by excision of the dislocated radial head at the time when skeletal growth had terminated and by occasional osteotomy at the distal end of the radius. The excision of distal osteochondromas in an effort to increase rotation was sometimes highly effective but tended to result in limited improvement. No patient in this series underwent stapling of the lateral half of the distal radial epiphyseal growth plate to prevent ulnar drift. On one occasion, excision of the radial head in a forearm in which the distal end of the ulna was severely shortened and malformed resulted in marked instability, which eventually was treated by a midshaft radioulnar synostosis. Taniguchi has shown how the severity of the forearm lesions correlates with the overall severity of the HME (328). He assessed the forearm appearance in 41 patients and divided them into three groups: group 1, no involvement of the distal forearm (N = 8); group 2, involvement of the distal forearm without shortening of the radius or ulna (N = 11); and group 3, involvement of the distal forearm with shortening of the radius or ulna (N - 22). He then compared each group in relation to the number of lesions throughout the entire body, the age of onset, patient height, and the presence of valgus deformity of the ankle. Correlations show quite nicely that the mild, moderate, and severe appearances at the

distal forearm and wrist correlated with similar findings throughout the rest of the skeleton. The number of lesions in group 1 was 5.8 + 7.4, group 2 28.5 + 10.4, and group 3 45.7 _+ 10.2. Involvement of each region of the body also correlated well with the grouping system. For example, involvement in the proximal humerus was 31% in group 1, 77% in group 2, and 95% in group 3. In the proximal femur the percentages of involvement were 6%, 73%, and 95%. Excellent correlations were seen in all other regions of the body using this system. The age of onset, by which is meant the age at which the parents or physicians first noted exostoses, was 10.3 _+ 4.0 years in group 1, 4.6 + 4.5 years in group 2, and 2.8 _+ 1.8 years in group 3. It is very rare to diagnose this condition before 1 year of age, although detailed radiographic surveys beginning at the newborn period have not been done. Height correlations showed group 1 at 0.3 +_ 0.6 SD above the mean, group 2 at - 0 . 3 + 1.5 SD below the mean, and group 3 at - 1 . 1 _+ 1.0 SD below the mean. Correlations continued with valgus deformity of the ankle greater than 10~ showing no instances in group 1, 1 in group 2, and 12 in group 3. In addition, the only 3 cases of the dislocation of the radial head occurred in group 3. Considerable controversy persists regarding the longterm need for and value of upper extremity surgery to correct the deformities associated with hereditary multiple exostoses. Arms et al. performed a retrospective study on 37 patients who had had various forms of operative intervention relating to osteochondromas and associated deformities of the forearms (3). Their questionnaire sought specifically to determine patient responses to the value of various forearm interventions. The surgeries performed included 36 osteochondroma excisions, 6 radial head excisions, 5 distal radial hemi-epiphysiodeses, 2 distal radial osteotomies, and 4 ulnar lengthenings. Arms et al. concluded that early aggressive intervention of a surgical nature may not always be needed because most of the patients, after reaching skeletal maturity, functioned very well with minimal concerns regarding aesthetic appearance despite significant deformity. They concluded that simple excision of symptomatic osteochondromas or symptomatic dislocated radial heads at skeletal maturity would suffice for the large majority of patients. They were nonsupportive of early radical intervention during the growing years to prevent problems because they felt that problems at skeletal maturity were not overly great. Other studies had also concluded that, although appearance was markedly improved, function improved to a much less extent. Stanton and Hansen showed that meaningful function was preserved in the presence of great deformity on many occasions (318). With the increasing sophistication of upper extremity surgery, more patients are currently undergoing intervention. It remains extremely important to define the reasons for intervention and to be aware of the benefits and limitations of the procedures. Masada et al. have classified the forearm deformities in hereditary multiple exostoses into three types (192).

SECTION XII ~ Review of Specific Skeletal Dysplasias Type 1: The main area of osteochondroma formation is in the distal ulna. The ulna is relatively short with lateral bowing of the radius, the radial head located, tapering of the distal ulna, and ulnar tilt of the distal radial epiphysis. In their series 61% (22) forearms were of this type. Type 2: In addition to ulnar shorting, the radial head is dislocated but bowing of the radius is less severe than in the type 1 deformity because of the dislocation. In type 2A the radial head is dislocated because of an osteochondroma at the proximal metaphysis of the radius, whereas in type 2B there is no osteochondroma of the radius. Type 2A accounted for 6% of the forearms and type 2B for 14%. Type 3: The main osteochondroma formation is in the metaphysis of the distal radius and there is relative shortening of the radius. This was seen in 19% of patients. Masada et al. have summarized their recommended surgical procedures as follows. Type 1: Excision of osteochondroma, radial osteotomy, and immediate (one-stage) ulnar lengthening. Type 2A: Excision of osteochondroma, radial osteotomy, immediate ulnar lengthening, and excision of radial head. Type 2B: Excision of osteochondroma, radial osteotomy, and gradual ulnar lengthening. Type 3: Excision of osteochondroma only. The operative approaches to forearm deformities in hereditary multiple exostoses have been well-summarized by Fogel et al. (82), Masada et al. (192), and Peterson (240). The procedures used involve excision of the osteochondromas, lengthening of the ulna, corrective osteotomy of the distal radius, stapling of the distal physis of the radius (291), excision of the radial head, and open reduction of the dislocated radial head. The epiphyseal stapling of the distal radius involves a hemi-epiphyseal approach with stapling on the lateral radial side, allowing some growth medially to continue. The procedure used most commonly is excision of the osteochondroma. In some instances this can improve rotation and will also improve appearance in many. It will also tend to normalize the relationship of the radius to the ulna at either end of the forearm and minimize tendencies to subluxation and dislocation. Ulnar lengthening reduces the tethering force on the growing radius, thus encouraging it to continue to grow straight or to at least minimize its rate of curvature and also to protect against dislocation of the radial head. The ulnar lengthening will also minimize ulnar drift. It is evident that the relationship of the distal ulna to the carpus still will not be normal. Correction of the angular deformity of the distal radius can be achieved either by hemiepiphyseal lateral stapling or by wedge osteotomy (Fig. 25L). Hemi-epiphysiodesis is minimally invasive, but the result can be difficult to project because of uncertainty regarding how the tethering on the ulnar side has damaged the medial physis. Peterson recommends excision of the osteochondroma once the growth pattern on the involved bone clearly is being altered by its presence (240). Ulnar lengthening is generally performed if the radial-ulnar length discrepancy exceeds

835

1 cm. Up to 2 cm of length can be gained in a one-stage procedure, although removal of the soft tissue interosseous ligament tether is usually needed. Transverse osteotomies can be fixed with bone graft and plates, although occasionally oblique step-cut osteotomies are done. Length greater than 2 cm can be gained using gradual distraction lengthening with the Wagner and Orthofix apparatuses. In general, resection of the dislocated radial head is not recommended particularly if there is growth remaining, but efforts to relocate it generally are followed by stiffness. Efforts are directed therefore to early excision of osteochondromas at the proximal radial-ulnar area and minimization or prevention of the radial head dislocation by ulnar lengthening. Finidori et al. commented that the radial head was at high risk of dislocation when the affected ulna was greater than 1 cm shorter than the radius (80). Irani and Petrucelli agreed (133). When the deformities in the forearm are great or when surgical intervention has been less than fully effective, Rodgers and Hall have reported the creation of a one-bone forearm as an effective salvage procedure for recalcitrant forearm deformity (266). The radial-ulnar fusion is combined with removal of the proximal radius to allow the ulna to function alone as the elbow hinge and removal of the distal ulna to allow the radial-carpal joint to function singly. This surgical procedure can be combined with forearm lengthening. d. Lower Extremity Length Discrepancies. The femoraltibial limb length discrepancy measurements indicated a range from 0.1 to 4.0 cm (287). In 2 patients, femoral and tibial shortening were equal, in 20, femoral shortening was greater than tibial, and in 10, tibial shortening was greater than femoral shortening. On occasion, the shortening was limited to either the femur or the tibia. Of the 22 patients who had reached skeletal maturity, 11 (50%) had limb length discrepancies that were in the range for which limb length equalization would normally be recommended. Of these, however, only 5 (23%) actually had the procedure. The limb length discrepancies at the termination of growth in those who did not undergo growth-arrest procedures measured 2.1, 2.4, 2.3, 2.6, 3.1, and 3.5 cm. Many of the charts were not clear as to why surgery was not performed, but the following reasons were listed: subsequent planned correction of discrepancies with associated opening wedge osteotomy for deformity on the short side or closing wedge osteotomy on the long side; difficulty of performing epiphyseal arrests in regions in which large exostoses are present; clinical impression of an acceptable situation despite the roentgenographic measurement; and reluctance of patients and their families to undergo yet more procedures. When we take into consideration all 22 patients who had reached skeletal maturity and also the 7 who were old enough that it was possible to determine whether they would require an arrest, there were 29 patients, of whom 12 (41.2%) had discrepancies within the recommended range for operative intervention.

836

CHAPTER 9 ~

Skeletal Dysplasias

All five patients who were operated on had distal femoral epiphyseal arrests, and one had a proximal tibial-fibular epiphyseal arrest as well. The limb length discrepancies at the time that the arrest was performed and the eventual limb length discrepancy at the termination of growth were as follows: 2.9 to 1.2 cm; 2.4 to 1.3 cm; 2.3 to 0.5 cm; and 4.0 to 4.0 cm. There were no overcorrections, and although four of the five limbs were corrected into an acceptable range, all were short of perfect equalization. The last patient was operated on too late by all criteria. In the other four, however, the question arose as to whether growth anomalies in hereditary multiple exostoses might make prediction from the normal charts slightly unreliable. We therefore plotted femur: tibia length ratios along the appropriate percentile distribution in all of the patients with hereditary multiple exostoses and compared them with the standard charts. The mean value for this length ratio in the patients with hereditary multiple exostoses was 1.27, which was exactly the same as the value from the charts for normal subjects with the same size distribution. Reference to the records of each patient who underwent epiphyseal arrest indicated that there had been considerable difficulty in assessing the skeletal age from the roentgenograms of the wrists. The pattern of limb length discrepancy that can occur in this condition is variable. The discrepancy can remain unchanged for several years, it can increase at slow or moderate rates, which is the usual pattern, or it can, on occasion, decrease spontaneously. The growth study data did not support the belief that there is an increase in longitudinal growth in an affected bone following removal of an exostosis. In addition, no correlation was seen between the degree of shortening in a particular bone and the number or size of the exostoses present. Limb length discrepancies in our series were frequent, and in approximately one-half of the patients they were great enough to warrant epiphyseal arrest. These discrepancies point to the asymmetrical growth pattern in patients with hereditary multiple exostoses. The previously expressed belief that the growth retardation in a bone is directly related to the size and number of the exostoses is inaccurate. The discrepancies present in patients in our series were mild to moderate and were managed readily by appropriately timed epiphyseal arrests. Extremely careful observation is required, however, as the discrepancies can remain stable, increase at varying rates, or even, on occasion, spontaneously decrease. Corrective osteotomies can also alter limb length relationships. Some difficulties were encountered in determining the skeletal age accurately due to the associated wrist and knee anomalies, but the GreenAnderson charts were appropriate for predicted corrections in this condition. The limbs were more affected than the spine, both the femur and the tibia were involved, and the limb involvement was not invariably rhizomelic. e. Pressure Phenomena Caused by Exostoses. Pressure phenomena in relation to the spinal cord, nerves, vessels, and viscera can occur, but these are relatively infrequent. The pressure phenomena tend to become most apparent around

15 years of age or later as skeletal maturity is associated with thinning and eventual disappearance of the cartilage cap, leaving the underlying often pointed segments of bone more prominent. Pressure phenomena on nerves, spinal cord, adjacent vessels, and viscera were recognized early with many reports reviewed by Gibney as long ago as 1876 (88). Peripheral Nerve Compression-Stretching: The most commonly affected nerve is the peroneal due to expansion of proximal fibular exostoses (Fig. 25D) (45). Spinal Cord Compression: Spinal cord compression is possible due to exostoses growing within the canal from points of junction of the endochondral and intramembranous systems (183). Vascular Compromise: The most common anomaly is the formation of pseudo-aneurysms in postadolescent, early adult life, and the most commonly affected region is the posteromedial aspect of the distal femur producing lesions of the femoral or popliteal artery (Fig. 25M). Visceral Pressure: Virtually all internal viscera have been the subject of case reports of abnormal extrinsic pressure or rupture, including trachea, esophagus, all other parts of the gastrointestinal system, the urinary bladder, and the uterus. Bursa Formation: The most common and most benign sequel of an exostosis is bursal irritation. f. Fracture of Exostoses. Fracture of exostoses can occur. They occur most commonly at the distal femur and proximal tibia and usually heal uneventfully. g. Malignant Degeneration of Exostoses. Malignant degeneration to a chondrosarcoma of one of the exostoses is a well-known complication (130, 137, 304). The transformation almost invariably occurs in the adult age group especially after the third decade. Virtually all tumors are chondrosarcomas derived from remnants of the cartilage cap. Concern is raised about malignancy whenever there is increase in size of an exostosis after skeletal maturity, pain at the site of an exostosis, or radiographic evidence of increased bony radiolucency suggestive of local bony invasion. Some physicians recommend performing a bone scan at skeletal maturity and repeating it every few years to try to diagnose malignant transformation sooner than relying on clinical awareness. Approximately two-thirds of the exostoses that undergo malignant transformation are of the scapulae, proximal humerus, pelvis, or proximal femur with most of the others at the knee. The true incidence of malignant change has been difficult to determine because it requires a several-decade study of a large group of patients. Original estimations of 10-25% of transformation appear to be far too high, and more recent reports tend to place the rate as below 3% (30, 97).

Z. Metachondromatosis Metachondromatosis is rare autosomal dominant skeletal dysplasia characterized by (1) multiple metaphyseal juxtaphyseal exostoses that primarily involve the hands and feet,

SECTION XII 9 Review o f Specific Skeletal Dysplasias

point toward the joint, and may resolve spontaneously; (2) multiple enchondromas usually involving the iliac crest and long bone metaphyses of the lower extremity; (3) periarticular radiodensity representing spicules of bone formation in expanded juxtacortical enchondromas; and (4) frequent femoral head AVN (145, 186, 358). Histologic studies show a characteristic cartilage cap with endochondral bone formation on the inner surface, a sessile exostosis resembling a juxtacortical chondroma with spicules of endochondral bone formation, and intramedullary lesions of well-circumscribed islands of cartilage. Cases of AVN of the femoral head associated with sessile femoral neck expansile enchondromatous lesions have been described. Shortening and deformity of the long bones do not otherwise occur, and all patients with this disorder are of normal stature. The disorder was defined and named by Maroteaux in 1971, who stressed that in the majority of cases the disorder was noted because of deformation of the metacarpals, metatarsals, or phalanges of the fingers or toes (186). Management is by surgical removal of clinically bothersome exostoses and nonoperative containment or surgical treatment for the femoral head AVN as indicated. AA. Osteopetrosis 1, OVERVIEW Osteopetrosis is an inherited condition characterized by increased bone radiodensity and widened metaphyses throughout the skeleton due to a failure of normal osteoclastmediated bone and cartilage resorption. As with many hereditary skeletal disorders, it is not a single disease but a syndrome with several variants currently defined based on clinical criteria. There are three clinical groups: infantile malignant autosomal recessive, which is fatal within the first few years of life (in the absence of effective therapy); intermediate autosomal recessive, which appears during the first decade of life but does not follow a malignant course; and autosomal dominant, with full life expectancy but many orthopedic problems (282). The infantile variant shows a myelophthisic anemia, granulocytopenia, and thrombocytopenia, and patients eventually die from infection or bleeding or both. Neurological sequelae include cranial nerve compression (optic nerve, blindness; auditory nerve, deafness; facial nerve, paresis), hydrocephalus, convulsions, and mental retardation. Radiographs show uniform bone density without corticomedullary demarcation, broadened metaphyses, "bone within a bone" or endobone phenomena (tarsals, carpals, phalanges, vertebrae, and ilium), and thickened growth plates if there is superimposed tickets (Fig. 26A). Transverse pathological fractures occur, often followed by massive periosteal bone formation. Computed tomographic scans, magnetic resonance imaging, and bone scans provide specific information. Iliac crest bone biopsy is valuable to quantitate osteoclast and marrow changes by light and electron microscopy. Medical treatments involve high-dose calci-

837

triol to stimulate osteoclast differentiation and bone marrow transplantation to provide monocytic osteoclast precursors. Orthopedic problems in the intermediate and autosomal dominant forms include increased fractures, coxa vara, long bone bowing, hip and knee degenerative arthritis, and mandibular and long bone osteomyelitis. Cranial nerve compression also occurs. Osteotomy, plating, intramedullary rodding, and joint arthroplasty can be done but are difficult because of bone hardness. 2. TYPES OF OSTEOPETROSIS a. Human Variants. Infantile Malignant Autosomal Recessive Osteopetrosis: Most patients with infantile osteopetrosis are initially diagnosed within the first year of life. There are two modes of presentation. In one, the dense structure of the bones and the absence of a medullary canal lead to increased fragility and pathological fracture. When the involved individual is assessed for limb pain, the radiograph shows increased bone density and a pathological transverse or short oblique fracture. The second mode involves failure to thrive with frequent upper respiratory infections. When a detailed pediatric assessment is performed, a chest radiograph reveals the radiodense pattern of the fibs, vertebrae, and adjacent long bones, frequent concurrent pneumonia, and a hematologic picture characterized by anemia, granulocytopenia, and thrombocytopenia. Clinical examination reveals a marked hepatosplenomegaly. Further clinical manifestations of the condition also evident within the first year of life involve decreased function of the auditory and optic nerves, leading to progressive heating and vision loss. These latter two conditions are associated with narrowing and sclerosis of the auditory and optic canals. Virtually all affected children will die during the first few years of life because of progressive worsening of the hematologic picture, leading to massive problems with infection or bleeding or both. Intermediate Mild Autosomal Recessive Osteopetrosis: Patients with this form of osteopetrosis tend to be diagnosed toward the end of their first decade of life (141). Often a fracture brings about the diagnosis because the radiograph is characterized by increased density and decreased metaphyseal remodeling. At the initial diagnosis, the patient has some or all of the following conditions: mildly disproportionately (short limb) short stature, macrocephaly, recurrent fractures, mandibular osteomyelitis, dental abnormalities, and mild to moderate anemia. Hepatosplenomegaly is either slight or absent. Relatively few patients have been described, and virtually all patients have survived into adulthood. Mandibular osteomyelitis can be extremely troublesome. A subset of the intermediate form of osteopetrosis is associated with renal tubular acidosis, intracerebral calcification visible in plain radiographs, and a carbonic anhydrase II enzyme deficiency (298). Benign Autosomal Dominant Osteopetrosis: The autosomal dominant variant of osteopetrosis, sometimes referred to as the adult form, is a much more benign condition, and, in

838

CHAPTER 9 ~

Skeletal Dysplasias

F I G U R E 26 Characteristic findings in osteopetrosis are reviewed. (A) Radiographic findings of osteopetrosis include uniform radiodensity without corticomedullary demarcation, broadened metaphyses, and the "bone within a bone" or endobone phenomenon. The latter is seen most characteristically in the tarsals, carpals, phalanges, vertebrae, and ilium. These findings can be seen in the

SECTION XII ~ Review o f Specific Skeletal Dysplasias

F I G U R E 26 (continued) newborn or early months of life in the malignant recessive variant, but they also characterize the benign dominant forms. (B) Periosteal elevation often is seen in patients with osteopetrosis. This can occur in relation to undisplaced or unrecognized metaphyseal fractures and also with epiphyseal growth plate fracture separations. Bleeding often is

839

F I G U R E 26 (continued) magnified by the thrombocytopenia. (Bi) Note extensive humeral involvement. (Bii) Tibial specimen shows periosteal elevation (white arrows) associated with proximal and distal metaphyseal fractures (curved dark arrows). [Reprinted from Shapiro (1993), Clin. Orthop. Rel. Res. 294:34-44, 9 Lippincott Williams & Wilkins, with permission.] (C) Histopathologic findings are characteristic in osteopetrosis. (Ci) At the light microscopic level early in the disorder, the osteoclast number is increased markedly and the osteoclasts usually are quite large with far more nuclei than normal. (Cii) The osteoclasts (arrows) position themselves appropriately in relation to bone and cartilage but function poorly because those tissues remain unresorbed. In older cases, the osteoclast number dramatically diminishes. (Ciii) The characteristic matrix finding is the persistence of cartilage in the marrow (far right) with continuing new bone formation. In addition, woven bone also persists with lamellar bone deposition proceeding. This is seen in thickened periosteal bone at left. The straight arrow points to the lamellar bone and the curved arrow to woven bone. Failure of absorption leads to the density of the bones both histologically and on radiographic representation. (Civ) Sections can be found from the metaphyseal region in which cartilage cores persist surrounded by woven bone, with that mass then surrounded by lamellar bone. [Parts Ciii and Civ reprinted from (285), with permission.] (Cv) Both qualitative and quantitative ultrastructure of the osteoclast can be performed. The characteristic finding in most instances is a markedly diminished ruffled border adjacent to bone and cartilage indicating diminished function of the osteoclast. The cartilage is at top and the osteoclast at bottom. In (Cvi) A normal osteoclast shows an intense ruffled border reaction adjacent to the cartilage. [Parts Cv, Cvi reprinted from Shapiro et al. (1988), Calcif. Tiss. Int. 43:67-76, copyright notice of Springer Verlag, with permission.]

SECTION Xll ~ Review o f Specific Skeletal Dysplasias

general, patients have a full life expectancy (32-34). Hepatosplenomegaly is not seen. Almost half of the individuals diagnosed with this condition are asymptomatic. This number is determined on the basis of family studies initiated by discovery of a member with the condition. The major symptom is pathological fracture in approximately 40% of patients involved. Although extreme fragility with multiple fractures has been reported, in general, the fractures are single isolated occurrences. The fractures are often transverse, demonstrating the characteristics of fractures through pathologic bone. Bone pain, especially in the lumbar area, is reported in 25% of patients. Two other specific problems are cranial nerve palsies and osteomyelitis. A common focus of osteomyelitis is the mandible, although long bone infections also occur. Coxa vara is typical, and cases of anterolateral bowing of the femur also have been described. A gene for autosomal dominant osteopetrosis has been localized to chromosome lp21 (345). b. Animal Models. There are several naturally occurring animal models which have osteopetrosis. Those cited in most detail are murine variants, although less typical rabbit and chicken models also have been described. The murine models in the rat include the ia (incisor absent), tl (toothless), and op (osteopetrosis) and in the mouse the gl (gray-lethal), mi (microphthalmia), oc (osteosclerosis), and op (osteopetrosis) variants (77, 185). 3. HISTOPATHOLOGY The histological appearance of the skeletal tissues and the results of studies on experimental animals with a disease similar to human osteopetrosis are consistent with the conclusion that the skeletal lesions principally result from a marked decrease in the rate of bone and cartilage resorption with little or no change in the rate of bone and cartilage formation (205,285). Failure to resorb the calcified cartilage formed during endochondral ossification leads to progressive filling of the metaphyseal region and eventually of the marrow cavity of the diaphysis, with a tissue composed of cores of calcified cartilage surrounded by new bone. In the most severe cases the unresorbed tissue extends the full length and width of the bone, completely obliterating the marrow spaces and excluding the blood forming marrow cells. Failure of osteoclasts to resorb bone at the periphery of the metaphyseal-diaphyseal junction at the distal end of Ranvier's ossification groove leads to the widened, abnormally shaped metaphyses. In the diaphysis, failure of resorption inhibits normal remodeling and cylindrization and leads to a thicker cortex. a. Gross Morphology. All bones are affected. The metaphyses and diaphyses of the long bones are increased in diameter. The epiphyses are normal in shape. Coronal sections and specimen roentgenograms show that the secondary centers of ossification, although more radiodense than normal, are appropriately positioned and are of the same size and shape as the secondary centers of ossification of normal

841

children. The time of appearance of the secondary centers is normal. Healing epiphyseal displacements are seen. Cross sections of the diaphyses show the increased diameters to be due to an increased thickness of the true cortices as well as to the accumulation of a large amount of sub-periosteal new bone in which small amounts of red marrow tissue were seen. Compact osseous tissue is seen in the medulla rather than cancellous bone. Cortical-medullary demarcation is indistinct. Long bones show extensive diaphyseal new bone accretion. The calvarial diploE are increased in thickness, with little marrow space between them. b. Skeletal Histology, Cytology, and Ultrastructure Light Microscopy. A longitudinal section of an entire phalanx demonstrates well the contributions of the endochondral sequence and the intramembranous periosteal sequence to long bone development, because the synthesis pattern is preserved in the absence of effective resorption (285). c. Epiphysis and Metaphysis: The Epiphyses Are Normally Shaped. The cells in the articular cartilage and in the cartilage of the epiphyseal growth area are normal except for a few regions of the proliferative and hypertrophic zones of a few of the epiphyseal growth plates, which contain irregularly clumped chondrocytes between the vertical bars of extracellular matrix and irregular and somewhat widened columns of cells. Bundles of fibrous tissue extend between the hypertrophic zone and the zone of provisional calcification and interrupt the normally smooth, homogeneous appearance of the extracellular matrix. Except for these latter regions, the extracellular organic matrix of the articular cartilage and the epiphyseal growth cartilage appear to be normal. Because a number of type I epiphyseal fractures and fracture-separations can be observed clinically and roentgenographically, one cannot be certain whether these regions of irregularly dispersed columns of cells are an inherent part of osteopetrosis or whether they represent responses to the epiphyseal fractures. The latter seems more likely. The most striking and characteristic histological feature of osteopetrosis is observed at the metaphysis: the persistence of large amounts of calcified cartilage surrounded by bone, principally woven bone but in a number of instances lamellar bone. The cement lines are thicker and more prominent than usual. This endochondral tissue is present throughout the metaphysis and in many instances it fills the medullary canal of the diaphysis. The tissue remains organized as thickened primary trabeculae, consisting of a central core of calcified cartilage surrounded by new bone. Toward the diaphysis the trabeculae are densely packed, with virtually no marrow tissue present. In none of the major long bones studied do the regular columns of cartilage cells constituting the epiphyseal growth plate extend to the periphery of the epiphyseal cartilage. Moreover, at the periphery, the germinal layer of the epiphyseal cartilage is often observed to be immediately adjacent to newly formed endochondral bone, without the interposition of the regularly ordered columns of cartilage cells. This most likely represents a repair response to the

842

CHAPTER 9 9 Skeletal Dysplasias

epiphyseal fracture-separations rather than an inherent feature of osteopetrosis. d. Cortical and Sub-periosteal Bone. At the periphery of the metaphysis and diaphysis, the persisting remnant of the original cortical bone of the primary center of ossification can still be recognized, as it has not been resorbed. Peripheral to this, large masses of periosteal bone are seen of a thickness wider than the original central medullary cavity and center of primary ossification. The pattern formation within this sub-periosteal bone is not uniform. Adjacent to the primary center of ossification, true lamellar haversian bone is present. Peripherally, woven bone that is beginning to transform into true lamellar bone is seen. The woven bone looks normal under light microscopy, although there are some central areas of more primitively appearing bone. Adjacent to some areas of woven bone, however, thick accumulations of acellular cartilage are present, sometimes lying free and at other times surrounded by coarse and fine woven bone. Safranin O staining of this cartilage is normal. Scattered lamellar bone is also seen adjacent to cartilage and woven bone. The sub-periosteal accumulation of bone and cartilage is not as dense roentgenographically as the intramedullary tissue. Moderate vascularity and considerable fibrous tissue are seen histologically. The irregular histological pattern of bone and cartilage is not analogous to that seen in the normal reaction to fracture, when the cartilage is cellular and a more ordered sequence of repair is noted. The appearance of this tissue indicates that resorption has not occurred or has occurred only incompletely and that the regular temporal and spatial sequence of events has been disturbed. e. Joint Cavities: The Joint Cavities Appear Normal

Grossly and Histologically. f. Secondary Centers of Ossification: The Tissue in the Secondary Ossification Centers Has the Same Structure as That Seen in the Metaphyses. The central cores of cartilage persist and are surrounded by woven bone. Isolated spicules of mature haversian lamellar bone are seen. There are many osteoclasts in the secondary ossification centers.

g. Cell Populations (Cytology): Numerous Osteoclasts Are Present in Close Relationship to Both Cartilage and Woven Bone. On occasion, sections are seen in which spicules of bone and cartilage are surrounded entirely by huge osteoclasts. Osteoclast profiles with as many as 80 nuclei are noted. The osteoclasts are most abundant in the metaphysis adjacent to the growth plate and in the extracortical, subperiosteal region. They are infrequently seen in the lower metaphysis and diaphysis among the persisting thick cores of cartilage and bone. In the sub-periosteal tissue, plump osteoblasts cover the surface of both woven and young lamellar bone, and osteocytes are seen. The bone in the lower part of the metaphysis and diaphysis contains osteocytes, but osteoblasts are seen only occasionally. In the osteopetrotic iliac crest bone, there is a marked increase in the number of osteoclast profiles and marked ultrastructural abnormalities of the osteoclasts compared with those in normal bone obtained from the iliac crest of a

normal 2-year-old child. The osteoclasts from normal bone have many nuclei, an abundance of mitochondria, lysosomes, a Golgi apparatus, scattered ribosomes, and a wellstructured endoplasmic reticulum. Ruffled borders and a surrounding clear zone are frequently observed in the portion of the osteoclast adjacent to the bone surface. The ruffled border is composed of the cell membrane thrown into many folds. The clear zone, also part of the osteoclast cytoplasm, is peripheral to the ruffled border and is free to organelles. The lamina limitans, an electron dense, osmiophilic layer on the surface of bone, is also present. There is a close spatial relationship of the osteoclasts to the surfaces of bone and cartilage and to the blood vessels. Nineteen osteoclast profiles were counted in the normal cortical bone, 13 (68%) of which were adjacent to the bone and 6 (32%) of which were free. Six of the osteoclasts (32%) had ruffled borders and 10 (53%) had clear zones. In the osteopetrotic bone, 86 osteoclast profiles were examined. Due to the numerous osteoclasts present, profiles were quite easy to locate, whereas considerable sectioning was required to find such cells in the normal control. Sixtythree (73%) of the cells were on the bone or cartilage surface and 23 (27%) were free. Thus, the osteoclasts, as well as being more numerous, were still appropriately positioned to perform their resorptive functions. The most striking ultrastructural characteristic of osteoclasts in the osteopetrotic bone is the absence of ruffled borders and, except in a few instances, that of clear zones as well. Even osteoclasts that are adjacent to the surface of bone show no evidence of a change in their surface membranes. In a few instances, small cytoplasmic processes, which are often present along the portion of the cell away from the bone surface, are observed in that portion of the cell adjacent to the bone. The osteoclasts adjacent to cartilage have the same ultrastructural characteristics as those adjacent to bone. Osteoclasts in the osteopetrotic bone contain more nuclei than osteoclasts from normal bone, but the nuclei appear to be normal in shape, position, and appearance. Mitochondria, lysosomes, Golgi apparatus, and a well-developed endoplasmic reticulum are also present in the osteopetrotic osteoclasts. The endoplasmic reticulum, however, is situated in the portion of the osteoclast most distant from the surfaces of bone and cartilage. All of the organelles appear to be ultrastructurally normal. The lamina limitans, the electron dense, osmiophilic layer present on the surface of both bone and cartilage, is much more prominent in the osteopetrotic specimens. When present in cartilage, it does not inhibit the formation of bone matrix, which frequently overlies it. In certain sections large amounts of normal appearing collagen are present on the surface of the bone trabeculae. These are osteoid seams. 4. DIAGNOSTIC C~TERIA AND HEMATOLOGIC STUD.S Extensive hematologic abnormalities characterize the infantile malignant variant. The routine hemogram yields im-

SECTION Xll ~ Review o f Specific Skeletal Dysplasias

portant information concerning the compromised hematologic status. The patients have a myelophthisic anemia. The marrow forming regions of the bone are unable to function normally because they remain plugged with calcified cartilage and woven bone that persists due to failure of resorption. The eventual hematologic picture is characterized by anemia, granulocytopenia, and thrombocytopenia. The patients rely on extramedullary sources of hematopoiesis primarily from the enlarged liver and spleen for the production of blood cells. The anemia is associated with macrocytosis, reticulocytosis, circulating erythroblasts, and teardrop erythrocytes. Stress erythropoiesis leads to the production of fetal hemoglobin. The initial white blood cell count often is elevated with the differential showing a leukoerythroblastic reaction. Hepatosplenomegaly accompanies the extramedullary hematopoiesis, and the subsequent hypersplenism produces the thrombocytopenia, leukopenia, and hemolytic anemia. The susceptibility to infection has led to studies of neutrophil competence. Reeves et al. assessed monocytes and granulocytes in five osteopetrotic patients and suggested that the abnormal function of these cell types diminished host resistance (258). More recently, Beard et al. documented decreased neutrophil, monocyte, and granulocyte-macrophage colony responses to stimulation that pointed to expression of the osteopetrotic defect at or before the progenitor cell level (14). Mild to moderate anemia with evidence of extramedullary hematopoiesis is seen in the intermediate forms. There are no specific hematologic problems in patients with the benign autosomal dominant form. 5. BLOOD CHEMISTRY STUDIES In the infantile malignant variant at the time of diagnosis, the serum calcium, phosphorus, and alkaline phosphatase values are normal, but acid phosphatase is elevated. Development of a low calcium level is indicative of superimposed rickets, which is a well-documented phenomenon in some patients. The midmolecular parathyroid hormone level is usually normal or elevated slightly, but the levels of 1,25(OH)2-vitamin D are invariably markedly elevated. In the intermediate form, the calcium and phosphorus levels are also normal, and the acid phosphatase levels are markedly elevated. Blood chemistry values are normal in the autosomal dominant form except for the markedly elevated acid phosphatase levels. Bollerslev has provided a detailed study of several biochemical variables in two variants of the autosomal dominant condition (33). Acid phosphatase was increased markedly in what he refers to as type II; type I was normal.

6. IMAGING STUDIES: PLAIN RADIOGRAPHS AND SKELETAL SURVEY Radiographic features in the infantile form are illustrated. A skeletal survey reveals increased radiodensity of the bones throughout the skeleton. In the more advanced cases, there

843

is an absence of corticomedullary demarcation with uniform cortical density seen across the entire diameter of the bone. The metaphyses are wider than normal because of a failure of metaphyseal remodeling. This feature becomes more prominent after several months. In certain individuals, the "bone within a bone," or endobone phenomenon, can be seen particularly in the tarsal bones, vertebral bodies, phalanges of hands and feet, and pelvis (iliac wings). The metaphyseal regions often show alternating transverse radiodense and radiolucent bands, indicative of periods of exacerbation and remission. Thickening of the growth plates can be seen in patients who have tickets superimposed. Periosteal elevation caused by trauma and bleeding can be inferred from the longitudinal peri-diaphyseal alternating bands of radiodensity and radiolucency, often referred to as the "onion skin" phenomenon (Fig. 26B). Correlative radiographic and light microscopic pictures of osteopetrotic bone have been presented by several authors. In a rare autosomal recessive variant, osteopetrosis is associated with renal tubular acidosis, carbonic anhydrase II deficiency, and intracranial calcification. A characteristic finding in many adult autosomal dominant patients is the rugger-jersey spine appearance caused by thick radiodense accumulations at the superior and inferior vertebral body surfaces with relatively radiolucent areas in between. Bollerslev and Andersen have studied 35 patients with autosomal dominant osteopetrosis and described two types based on radiographic variations (33). Both types have osteosclerosis throughout the skeleton, but in type I patients there is prominent sclerosis and increased thickness of the cranial vault, diffuse sclerosis of the spine without preferential end plate thickening, and no bone within a bone (endobone) phenomenon in the pelvis. In type II patients, sclerosis of the skull is marked at the base, vertebrae show end plate thickening (rugger-jersey spine), and the endobone phenomenon is almost always seen in the pelvis. a. C T Scan. For diagnosis and determination of the effects of management, the CT scan assessment of the diameters of the auditory and optic canals is of great importance. It provides a baseline quantitative documentation of canal diameter that is helpful in observing patients over time, especially after treatments such as surgical decompression, bone marrow transplantation, or high-dose calcitriol. Optic canal tomography was used widely before the CT scan. b. Magnetic Resonance Imaging. Magnetic resonance imaging has been used in osteopetrosis primarily as an indicator of marrow activity. The infantile malignant forms demonstrate a complete lack of signal from the marrow in the vertebral bodies. The bodies appear "black" because of an absence of signal that contrasts sharply with the presence of signal from the adjacent intervertebral disks. Marrow elements are demonstrated, however, in benign forms and after successful bone marrow transplantation. c. Bone Scan. Bone scans have been used in osteopetrosis to show focal increased bone uptake in cases in which osteomyelitis and occult fractures are being considered (1).

844

CHAPTER 9 ~

Skeletal Dysplasias

A Tc-99m methylene diphosphonate scan in an adult with the benign dominant form shows increased uptake at sites of pathological fracture, but normal generalized uptake even with significant increased bone density on plain radiographs. Thus, periosteal reactivity can be assessed. When specific studies of the reticuloendothelial system are performed using Tc-99 human serum albumin millimicrospheres, a complete absence of bone marrow activity has been noted. The human serum albumin studies also document increased splenic uptake and hepatosplenomegaly. Because of relatively poor resolution, scintigraphy is not routinely performed in patients with osteopetrosis, although in selected cases it has value. It is expected that magnetic resonance imaging will be a more precise indicator of marrow function. 7. ILIAC CREST BONE BIOPSY Iliac crest bone biopsy is often performed to better assess the biological state of the condition and to monitor specific therapies. Shapiro et al. have documented a variable osteoclast appearance in nine patients with infantile malignant osteopetrosis studied between 1 and 12 months of age before any therapy (286). Because there is the risk of infection and bleeding, a cautious approach to biopsy is needed. The biopsies are performed in the operating theater under antibiotic coverage to minimize the risk of infection. Preoperative investigation involves a careful hematologic assessment. A biopsy is not performed if the platelet count is under 50,000. If the count is in that region or lower, a preoperative platelet transfusion is recommended as long as it does not compromise subsequent therapy such as bone marrow transplantation. Biopsy should be performed only after prior consultation with the hematology service. Iliac crest biopsy can be done either through a small open incision or through a percutaneous incision using a Bordier bone biopsy needle. In the author's institution, bone biopsy is performed to quantitate osteoclast number and bone surface coverage at the light microscopic level, to assess the ruffled border-clear zone relationship to bone and cartilage cores by transmission electron microscopy, and to assess the marrow contents (Fig. 26C). In the author's study, osteoclast numbers, size, and nucleation were always increased and the cells were always positioned appropriately in relation to bone and cartilage. In those patients with only a mild to moderate osteoclast increase, the marrow had an otherwise near-normal appearance with a good complement of hematopoietic cells. Light microscopic histomorphometry documented the percentage of bone and cartilage surfaces covered by osteoclasts. The controls from the areas of greatest osteoclast presence showed a 5% coverage. One osteopetrotic patient showed a coverage of 4.8% with all other patient values elevated from 7.6% to 27.9%. The authors have proposed that a rough index of 5 - 9 % surface coverage is indicative of a mild increase, 10-19% a moderate increase, and greater than 20% a marked increase. Quantitative electron microscopy showed the ruffled border-clear zone complex to be

absent or markedly diminished in seven of the nine patients. In two, however, osteoclast profiles had abundant ruffled border-clear zone complexes. Patients with the hyperosteoclastic bone marrow clinically were more severely affected. The bone marrow quantitation does not yet establish absolute diagnostic criteria but is helpful in monitoring the biological response to disease and therapy. 8. SPECIFIC ORTHOPEDIC CONCERNS IN OSTEOPETROSIS a. Fractures. Two types of fractures are seen in children

with infantile osteopetrosis. These involve either diaphyseal or metaphyseal fractures that are generally transverse and minimally displaced or epiphyseal fracture-separations that are also minimally displaced and often difficult to detect radiographically. Delay in the recognition of such fractures and the difficulty of immobilization in the small, sick infant often lead to extensive periosteal elevation and extracortical new bone formation. On occasion, the entire shaft from growth plate to growth plate is surrounded by exuberant periosteal new bone formation. The reaction may be enhanced by increased sub-periosteal bleeding associated with thrombocytopenia. The possibility of an occult osteomyelitis must also be considered. Fractures also occur with increased frequency in patients with intermediate autosomal recessive and benign autosomal dominant osteopetrosis. In general, they can be treated effectively by nonoperative means. The fractures heal although the time to healing is often prolonged. There is a high incidence of hip and proximal femoral fractures, particularly in the autosomal dominant group. These can undergo successful internal fixation, but comment is invariably made on the extreme hardness of the bone and the difficulty of the intervention. Both intramedullary rods and side plates have been used. Bollerslev and Andersen updated fracture patterns in autosomal dominant osteopetrosis (34). Of 35 patients studied, fractures occurred in 12. Only 2 of 20 had type I osteopetrosis, whereas 10 of the 15 fractures were in type II patients. A wide variety of bones were involved, including tibia, femur, patella, scaphoid, ribs, scapula, and transverse spinous processes. Ten of 12 patients had only 1 fracture, and each of the other 2 had only 2 fractures. Milgram and Jasty noted fractures in 9 of 21 patients. In 2 of the children, epiphyseal fracture-separations with abundant periosteal new bone formation had occurred. Compression fractures of the spine have not been reported. b. Coxa Vara. Many patients develop a coxa vara deformity, some in the first few years of life (32). This deformity is seen in the intermediate and autosomal dominant types and is the most common deformity in these groups by far. The coxa vara appears to be caused by stress fractures of the femoral neck, with gradual deformity ensuing. The deformity can be treated by osteotomy, but the hardness of the bone makes internal stabilization difficult.

SECTION XII ~ Review o f Specific Skeletal Dysplasias

c. Long Bone Deformity. Long bone deformity other than coxa vara is occasionally seen in the autosomal dominant type. There is lateral bowing, which usually involves the femur but has also occurred in the tibia, humerus, radius, and ulna. The deformity may be secondary to malunion after diaphyseal and metaphyseal fractures, but generally is of slow onset and is not associated with gross trauma. d. Osteomyelitis. Because of the diminished vascularity of the bones and impaired white cell function, osteomyelitis is seen frequently with osteopetrosis. The most common site of involvement is the mandible. The teeth are impaired frequently by osteopetrosis, and mandibular osteomyelitis is associated with dental caries. Long bone infections occur and must be treated vigorously because chronic foci have been established in many patients and are difficult to eradicate. e. Back Pain. Back pain is mentioned in most reports of benign autosomal dominant osteopetrosis. This rarely requires surgical intervention, but can be persistent. Bracing and conservative medication regimens often are required. f. Osteoarthritis. Degenerative arthritis of the hip and knee is seen with increased frequency in midadult life in patients with the autosomal dominant form (44). Total hip and total knee arthroplasties have been performed with good results, but the difficulty of reaming and instrumentation in association with the extremely hard bone and narrowed to absent intramedullary canals must be anticipated. It is theorized that the increased stiffness of the dense subchondral bone predisposes one to premature cartilage degeneration. A survey of physicians in the Pediatric Orthopedic Society of North America detailed results obtained in 79 patients with osteopetrosis, the vast majority having the autosomal dominant variant (4). The major orthopedic complication of the disorder was fracture with a high level of involvement of femoral neck, intertrochanteric proximal femur, subtrochanteric proximal femur, and femoral shaft. There were also many tibial fractures and lesser numbers of upper extremity fractures. The major hip deformity requiting treatment was coxa vara. In 6 patients with femoral neck fractures, the best results were obtained with closed or open reduction and internal fixation using either pins or compression screws. Neither infection nor delayed union occurred. On the other hand, those treated with cast therapy alone invariably developed varus and nonunion. Intertrochanteric fractures were best treated with internal fixation, although considerable difficulty due to the hardness of the bone was generally noted. Subtrochanteric fractures healed with both operative and nonoperative methods. Femoral shaft fractures generally were treated successfully with nonoperative methods. Intramedullary rodding was extremely difficult, and in each of 3 patients 3, 5, and 7 years of age further surgery was needed. Tibial fractures treated with casting alone tended to heal successfully. Coxa vara was best treated operatively with valgus osteotomy, preferably during childhood. Surgery was particularly valuable for coxa vara and the femoral neck fracture, but for most long bone fractures nonoperative management worked well.

845

9. SPECIFIC NEUROLOGIC CONCERNS IN OSTEOPETROSIS

Although optic and auditory nerve dysfunction has been recognized in osteopetrosis, more detailed studies indicate a broader spectrum of neurological involvement. In the infantile malignant variant, the most common neurological problems are optic atrophy with blindness, nystagmus, deafness, facial paresis, hydrocephalus, macrocephaly, and convulsions (169). The optic atrophy with blindness is believed to result primarily from direct compression of the optic nerve within the narrow optic canal. Other hypotheses have been advanced, however, including hydrocephalus, increased intracranial pressure, primary retinal degeneration, and primary demyelination of the optic nerve. It is now believed that there may be primary involvement of the central nervous system parenchyma, as well as changes secondary to skeletal abnormalities. A large number of patients with the infantile variant have delayed motor development, and mental retardation develops in some. Ventricular hypertrophy and cortical atrophy have been defined by CT scans. In one study, CT scans showed hydrocephalus in 5 of 6 infantile patients. Neurological sequelae also occur in the intermediate and benign autosomal dominant forms, but they are progressively less severe. In the benign form, about 20% of patients have readily detectable neurological manifestations, with the large majority being optic, trigeminal, facial, and auditory nerve compression syndromes. When specific studies looking for cranial nerve compression sequelae were performed on 14 patients, all but 1 had neurological problems involving at least one of these nerves. A neurosurgical procedure involving optic nerve decompression for osteopetrosis in early childhood has been reported to provide either improvement of visual capability or at least stabilization of function (100). Electrophysiologic studies such as visual evoked responses help to quantitate the sequelae of optic nerve compression and the results of decompression. Auditory evoked potentials help to document delays in nerve conduction associated with heating defects. Surgical decompression of the facial and auditory nerves with some good results has also been reported. 10. CURRENT SYSTEMIC MANAGEMENT OPTIONS FOR OSTEOPETROSIS

Treatment protocols aiming to cure osteopetrosis are based on current understanding of the cause of the condition. Because the failure of osteoclasts to resorb normal amounts of calcified cartilage and bone is believed to underlie the condition, therapeutic efforts have been directed at either stimulating host production of osteoclasts or providing an alternate source of osteoclasts to augment those not performing effectively. The osteoclast has its origin from circulating monocytes that fuse to form the multinucleated cell. a. High Dose CalcitrioL This approach is based on the experimental finding that high-dose calcitriol (1,25dihydroxyvitamin D) stimulates osteoclast formation and

846

CHAPTER 9 ~ Skeletal Dysplasias

bone resorption (147). The amounts given are extremely high, with several cases treated at dose levels of 1.5 micrograms per kilogram per day. It is important to note that, in infantile malignant osteopetrosis, the serum 1,25-dihydroxyvitamin D level is always markedly elevated, with a mean level of 193 pg/ml in 8 patients with the normal range 15-60. The massive pharmacologic doses are believed to induce bone resorption by stimulating quiescent osteoclasts or enhancing differentiation of precursors to the monocyte-osteoclast line. An integral part of this therapy is the use of a lowcalcium diet that itself may potentiate a resorptive phase. Cases indicating stabilization or reversal of key osteopetrotic parameters have been reported, but this approach is not truly curative. b. Recombinant Human Interferon Gamma. Key et al. utilized recombinant human interferon ~/-lb in 8 patients with osteopetrosis and subsequently expanded the study to 14 with the severe malignant infantile variant (148). Treatment was by subcutaneous injection. After 6 months of therapy, all 14 patients had decreases in trabecular bone area and increases in bone marrow space as determined by marrow imaging. This improvement was sustained in 11 patients treated for 18 months. The long-term therapy was shown to both increase bone absorption and hematopoiesis and improve leukocyte function. c. Bone Marrow Transplantation. This treatment evolved from experimental work by Walker in the gray-lethal and microphthalmic osteopetrotic mouse in which he demonstrated that splenic transplants and parabiosis with normal litter mates induced normal bone resorption and reversal of the osteopetrotic state by providing large numbers of osteoclast precursors (350, 351). The multinucleated osteoclast is formed by the fusion of circulating monocytes. Bone marrow transplant is the only curative therapy (50, 81,143,303). Successful allogenic human bone marrow transplant for osteopetrosis was reported in a 5-month-old child (50), and subsequent reports of successful transplantations have appeared. In a review of European experience, 9 cases of HLA-identical transplants were reported, 6 of which were surviving diseasefree, and 2 HLA-mismatched transplants were reported, 1 of which was surviving disease-free (81). Gerritsen et al. studied 65 patients receiving myeloablative pretreatment (87). Recipients of a human leukocyte antigen (HLA) identical bone marrow transplant had a probability for 5-year survival with osteoclast function of 79% and recipients of a phenotypically HLA-identical bone marrow graft from a related or an unrelated donor or 1 HLA-mismatched graft from a related donor had a probability for 5-year survival with osteoclast function of 38%, whereas those who received a graft from a HLA-haplotype mismatched related donor had a probability for a 5-year survival of only 13%. Osteoclast function developed in all patients with engraftment. In 25% of patients, however, recovery of osteoclast function was associated with severe hypercalcemia especially in those older than 2 years of age. Only some of the patients experi-

enced visual improvement in the group with impairment at the time of transplant even with engraftment. Solh et al. reported on 8 patients with malignant osteopetrosis who underwent bone marrow transplant (303). Median age at transplant was 9 months, with 6 patients receiving marrow from HLA-identical sibling donors, one from a phenotypically matched father, and one from a 1 antigen-mismatched father. Two patients died without engraftment. Three out of 6 who engrafted were alive and well at 48, 63, and 81 months. Those patients showed full bone marrow reconstitution. In the youngest patient, vision improved dramatically with successful bone marrow transplant. The best results overall were obtained with the best matches and in the youngest patients. Rappaport has commented on the use of bone marrow transplant, indicating that one problem (of many) is hypercalcemia in the face of the abrupt onset of massive bone resorption, and a review of the entire field of bone marrow transplantation provides details of possible benefits and also of extensive complications that can occur with these procedures (253). At present, bone marrow transplantation is considered only for the clear-cut infantile malignant form. High-dose calcitriol and bone marrow transplantation for osteopetrosis have been in use only for a few years in a few centers. These factors, the biological and clinical variability of osteopetrosis, and its infrequent occurrence have not yet allowed for well-documented consensus on the long-term results of systemic treatment. d. Additional Supportive Measures. Supportive measures can be used to stabilize patients before other therapies or to serve as palliation if patients fail to respond. The hematologic picture of symptomatic anemia, granulocytopenia, and thrombocytopenia can be treated by blood and platelet transfusions. Splenectomy can be done for hypersplenism, but other extramedullary sites of the reticuloendothelial system remain active. Steroids can minimize hematologic sequelae temporarily, raising hemoglobin and platelet levels while decreasing spleen size.

BB. Pycnodysostosis Pycnodysostosis is a generalized skeletal disorder of autosomal recessive inheritance characterized by short stature, increased radiodensity of bone (osteosclerosis), persisting open cranial sutures, obliquity of the mandibular angle, and defects of the terminal phalanges of the fingers (Fig. 27) (200). It has been recognized as a lysosomal disorder caused by cathepsin K deficiency (86). The bone density is not as great as in osteopetrosis, demonstrating cortical thickening but also showing narrow intramedullary canal persistence. Hematologic abnormalities do not occur, the patients are systemically well, and the distal fingers are short and bulbous with the terminal phalanges showing an absence of the distal two-thirds or at least marked shortening and tapering of the distal phalanges. The distal part of the clavicle is also tapered and shortened as is the distal ulna. The major orthopedic

SECTION Xll

~

Review of Specific Skeletal Dysplasias

847

F I G U R E 27 Pycnodysostosis is a rare skeletal dysplasia characterized by increased bone density. Radiographic findings in pycnodysostosis usually allow it to be distinguished from osteopetrosis. The bone density is not as great. (A) Cortical thickening is seen but medullary demarcation usually is evident. (B) There are open cranial sutures and a diminished obliquity of the mandibular angle. (C) Defects of the terminal phalanges of the fingers are seen, which are either absent, shortened, or tapered.

problem is frequent fracture with a tendency to either delayed union or nonunion. The fracture line is often transverse, a sign of underlying pathologic bone. Eventual healing occurs, but internal fixation and bone grafting may be needed.

CC. Osteogenesis Imperfecta 1. OVERVIEW Osteogenesis imperfecta (OI), one of the most common inherited connective tissue disorders, is characterized by in-

creased bone fragility causing multiple fractures and often marked deformity. It is clearly a skeletal dysplasia but not a chondrodystrophy. Molecular abnormalities are present in the type I collagen molecule, but a large number of different mutations have been found, which accounts for the wide variation in clinical severity (41, 54, 156). The cartilage model of the developing embryonic bones is normal but the bone tissue is abnormal. In the more severe variants, however, there is true growth retardation leading to short stature. This diminished stature is a combination of (1) spine and

848

CHAPTER 9 ~

Skeletal Dysplasias

limb shortening due to fracture and angular deformation, (2) damage to physeal tissue by periphyseal fracturing, and (3) systemic diminution of physeal stimulation (because shortening is not invariably associated with the first two factors). The clinical disease state is manifested in tissues in which the principal matrix protein is type I collagen (mainly bone, dentin, sclerae, and ligaments). The phenotypic manifestations are variable in severity, ranging from perinatal lethal forms with crumpled bones and severe deformity to clinically silent forms with subtle osteopenia and no deformity. In general, clinical subtypes represent a series of syndromes related to classes of molecular defects, each with a reasonably well-defined phenotypic pattern. 2. CLASSIFICATIONS The wide clinical variability in osteogenesis imperfecta has led to multiple attempts at classification. a. Congenita-Tarda. Looser in 1906 classified OI into two types on the basis of when the first fractures occurred: congenita (fractures at birth) and tarda (fractures after the perinatal period) (176). He noted that the prognosis in the congenita type was poor, with a high mortality rate. Seedorff, in 1949, further subclassified OI tarda into gravis (fracture occurs within the first year of life) and levis (fracture occurs after the first year of life), noting that tarda gravis was associated with the development of severe deformities and disability (278). b. Temporal-Radiographic Classification (Shapiro). We have described a temporal-radiographic classification of OI, which helps to combine valuable information derived from the older congenita-tarda approaches and the more recent Sillence approach (280). The prognostic indicators assessed in our study were the age at initial fracture and the radiographic appearance of the long bones and ribs at birth or at the time of initial fracture if this occurred after birth. Those patients who had intrauterine fractures and/or fractures at birth were defined as having osteogenesis imperfecta congenita (OIC), with this group divided into A (OIC-A) and B (OIC-B) subgroups based on the appearance of the long bones and ribs in newborn radiographs. Those who fractured initially after birth were defined as having osteogenesis imperfecta tarda (OIT), with this group divided into those who first fractured before walking, subgroup A (OIT-A), and those who first fractured after walking had begun, subgroup B (OIT-B). Consideration was not given as to whether a fracture was of the upper or lower extremity, and wormian bones of the skull were not considered as fractures. Osteogenesis Imperfecta Congenita A: There have been multiple intrauterine and birth fractures affecting virtually all bones. The extremities are shortened and deformed. The femurs particularly are markedly shortened, widened, and crumpled at birth, sometimes showing posterior bowing. Other radiologic findings include a small, bell-shaped rib cage with multiply fractured, short, and locally or extensively

widened ribs; tibias that are wide and present with 45-90 ~ anterolateral bowing, a high likelihood of short, broad humeri, and a caput membranaceum with wormian skull bones. Osteogenesis Imperfecta Congenita B: Intrauterine and/ or birth fractures occur, but at birth the long bones are of normal shape and length except where localized broadening is present in some in response to intrauterine fracture repair. Even those bones with extensive fracturing demonstrate regions of normal metaphyseal funnelization. The ribs are of normal length, narrow or normal in width, and show less extensive and often minimal fracturing. The rib cage assumes a normal conformation. There is a caput membranaceum with wormian skull bones. Osteogenesis Imperfecta Tarda A: These patients are born without fractures, but first fracture before or when beginning to walk. The long bones and ribs at the time of initial fracture are of normal shape and length but appear somewhat narrow and osteogenic. The rib cage is unremarkable. Osteogenesis Imperfecta Tarda B: These patients fracture first after walking has begun. The bones at the time of initial fracture are of normal shape, width, and length. In most, they are normal radiologically; in some, cortical thinning and minimal osteopenia are recognized. The rib cage is radiologically normal. c. Sillence Classification. The classification system currently used most widely was developed by Sillence in 1979 from a comprehensive survey of patients in Australia (292, 296). The Sillence classification divides OI patients into four types on the basis of multiple clinical, genetic, and radiologic features. Type I OI is the mildest and most common form. Inheritance is autosomal dominant, although new mutations are frequent. Type I is subclassified into the more common type A (without dentinogenesis imperfecta) and the less common type B (with dentinogenesis imperfecta). The sclerae are blue, and the first fractures usually occur in the preschool years after walking has begun. There is commonly absence of significant deformity, kyphoscoliosis is comparatively mild and uncommon, and stature is generally normal. Life expectancy is normal for patients with type IA OI and only marginally impaired for those with type IB. Type II OI is the lethal perinatal form. Many fetuses are stillborn and survivors are often born prematurely. The disorder is usually lethal within the first few weeks of life, but some affected infants survive for several months and a few live for 1 or more years. Death is generally due to respiratory failure, intracranial hemorrhage, or brain stem compression. The sclerae tend to be light gray or grayish blue. There are multiple intrauterine fractures, and the femurs, tibias, and ribs are short, broad, crumpled, and deformed. Inheritance was thought to be autosomal recessive; however, most cases appear to result from new dominant mutations in a proband of unaffected parents. Occasionally unaffected parents have multiple affected children; this is thought to result from parental mosaicism or to be due to a rare autosomal recessive form.

SECTION XII ~ Review of Specific Skeletal Dysplasias Type III OI is the severely deforming form, with fractures generally present at birth. The sclerae generally are normal in color. Frequent fractures and deformity are common, stature is typically severely shortened, and the spine is often deformed. Respiratory complications and dentinogenesis imperfecta are common. Inheritance is thought to be autosomal recessive; however, new dominant mutations are common, and a rare autosomal dominant variety exists. Life expectancy is decreased, but affected individuals live into adulthood. Early mortality is due to respiratory illness, injury with intracranial hemorrhage, and basilar invagination. Type IV OI is a moderately severe form with great phenotypic variation, but is usually intermediate in severity between type III and type I. This variant is infrequent, accounting for approximately 5% of cases. Sclerae are normal in color, short status is variable, dentinogenesis imperfecta is common, and fractures and deformity are relatively common. Inheritance is autosomal dominant. Life expectancy can be decreased depending on disease severity; however, a large percentage of patients function independently well into adulthood. Characteristic radiographic findings in the lethal perinatal form of OI are seen in Figs. 2B and 28Ai-28Aiii. 3. CLINICAL CHARACTERISTICS The musculoskeletal features of OI are variable in their extent and severity, depending on the clinical subtype and reflecting the underlying genotypic heterogeneity (76, 149, 151,153, 178, 299). Gross skeletal features can include short stature (with dwarfing in severe forms), kyphoscoliosis, and pectus excavatum. The skull is often misshapen, with a broad forehead, flattened posterior cranium, overhanging occiput, bulging calvaria, and triangular facial shape. Depending on the severity of disease, there may be marked long bone deformity with anterior bowing of the humerus, tibia, and fibula and anterolateral bowing of the femur, radius, and ulna. The hallmark of OI is bone fragility. As with the other phenotypic features of OI, the tendency to fracture is extremely variable, with manifestations ranging from innumerable fractures in utero and at birth to their virtual absence throughout life. The timing and number of fractures are included in some classification schemes. The more severe forms of OI are characterized by earlier and more numerous fractures. These fractures often occur after minor trauma. Fractures generally heal with abundant callus; however, the reparative bone is also abnormal, and fractures frequently lead to malunion and occasional pseudarthroses with resultant long bone deformity. The incidence of fractures decreases after puberty and rises again in women after menopause and in men after 60 years of age. The radiographic features of OI are also proportional to disease severity. The radiologic hallmark of OI is diffuse osteopenia associated with multiple fractures and deformities. Generalized osteopenia is seen in almost every case, and

849

there is often equal involvement of the appendicular and axial skeletons. The long bones of the lower extremities are usually more severely affected than those of the upper extremities. The long bones most often appear slender; however, they may show focal areas of thickened cortices secondary to callus buttressing or telescoping of fractures. The metaphyses of the long bones can be trumpet-shaped and cystic in appearance. In severe cases, "popcorn" calcifications appear in childhood in the metaphyseal-epiphyseal regions as displaced, fragmented physeal cartilage undergoes endochondral ossification. These calcifications commonly resolve after skeletal maturity, when all cartilage is transformed to bone. The vertebrae in OI often demonstrate flattening or are biconcave secondary to multiple microfractures. The overall incidence of spinal deformity in OI is approximately 60%, ranging from 90% for congenita forms to 10-40% for tarda forms. Thoracic scoliosis is the most common spinal deformity and is associated with osteoporosis, compression fractures, and ligamentous laxity. The spinal deformity is rarely isolated, and overall thoracic deformity associated with multiple rib fractures, scoliosis, molding of the soft thorax, and pectus excavatum or carinatum can be sufficiently severe to compromise respiratory function. The skull is characterized by wormian bones, but this finding is seen primarily in the more severe variants, congenita A and B types or Sillence II and III variants (Fig. 28B). The wormian bone appearance represents small, independent areas of primary ossification within membranous bones arranged in a mosaic pattern. The softened skull tends to develop a flattened occipital region as the affected child lies in the supine position; prominent frontal bossing and delayed closure of the fontanelles are also characteristic. Basilar invagination at the softened foramen magnum region can lead to cervicomedullary compression in the severe OIC A and B and Sillence II and III variants. Hyperplastic callus formation is rare but can occur in patients with OI. It often presents as pain, an enlarging mass, and erythema and can be difficult to distinguish from osteosarcoma radiographically, clinically, and even histologically. The articular cartilage appears to be normal. The physis often shows disorganization of the proliferative and hypertrophic zones with increased permeation of cartilage by metaphyseal blood vessels. The metaphysis is composed of a scanty, woven primary spongiosa. There is increased bone turnover as defined by tetracycline labeling studies. Dentinogenesis imperfecta is characterized by soft, translucent grayish colored teeth. The teeth are affected in a nonuniform manner with involvement usually greater in the primary teeth than in the secondary teeth. The enamel wears easily, and the teeth are carious, shortened, and susceptible to cracking. On X-ray films, the crowns are bulbous and there is obliteration of pulp chambers. About 30% of individuals with all types of OI have significant dental involvement. The Sillence classification scheme subclassifies type I

F I G U R E 28 Characteristic findings in osteogenesis imperfecta are illustrated. (A) Radiographic appearances of bones in lethal perinatal OI (Sillence II, OIC-A). (Ai) Rib cage at birth shows diminished lung fields and multiple rib fractures with broadened bone structure and deformities of clavicles, scapulae, and humeri. (Aii) Short, broad, crumpled femur and tibia with 90 ~ midshaft angulation.

C OSTEOGENESIS

IMPERFECTA

TEMPORAL-RADIOGRAPHIC DEATH

0.1. C O N G E N I T A

A (16)

NON-AMBULATORY

B (27)

~2

~

.

(21)

11

.

.

.

.

A

f/

GaI-O~

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

/

.

.

.

.

.

.

l

.

.

~ 7 33% .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

4 r/c/b. 9

33%

9 r/c/b. 14

67%

3 r/c/b. 21 100%

/

B

/

GaI-GIr

~l o.

1

(Martin

aJ=..r -L

~

6%

,1~16 59 %

.

f'

t I

GIrI

.

................

(21)

0.1. TARDA B

.

1

8%

-----,~. . . . . . . . . . .

0.1. TARDA A

AMBULATORY

,....,.~15 94% ...............

0.1. C O N G E N I T A

APPROACH

i'' 3 ....

!

9

"" OH o!

Ga!

Procollagen Suicide

Dendritic Fibrils F I G U R E 28 (continued) (Aiii) Specimen radiograph of femur. Note the central fracture, marked demineralization, crumpling, and absent cortical structure. Distal femur is at right. (B) Wormian bone formation of the cranium is seen in the more severe variants of osteogenesis imperfecta. It is seen infrequently in the tarda or benign autosomal dominant types. (C) Patient outcome based on temporal radiographic classification is shown. [Reprinted from Shapiro (1985), J. Pediatr. Orthop. 5:456-462, 9 Lippincott Williams & Wilkins, with permission.] (D) Structural alterations of collagen fibrils following a mutation are shown. [Reprinted from Prockop (1992), New Engl. J. Med. 326:540-546, copyright 9 1992 Massachusetts Medical Society. All rights reserved.] (E) The histopathology of OI bone provides valuable clues to the severity of the underlying disorder. In the most severe lethal perinatal variants, woven bone only is seen and even that is present in markedly diminished amounts (Ei). The bone tissue is hypercellular, which is indicative of the scanty matrix formation. In the most severe cases, isolated spicules of the bone are seen with little continuity apparent between them. In the disorders of moderate severity, woven bone persists but often is surrounded by lamellae of more mature bone (Eii-Ev). The greater the amount of lamellar bone, the stronger the underlying tissue. In the mildest

852

CHAPTER 9 ~ Skeletal Dysplasias

F I G U R E 28 (continued) benign dominant cases histologic structure is close to or may be indistinguishable from normal (Evi). (F) Multiple osteotomy with intramedullary rod fixation is resorted to in the presence of multiple fractures with increasing bowing deformity. In severe cases, only Rush rods or Steinmann pins can be used for fixation. (G) Scoliosis is a fairly frequent occurrence in OIC-B-OIT-A or the progressively deforming Sillence type III. Brace management is extremely difficult because of the relatively small size of the trunk and the softness of the ribs, which prevent comfortable and yet rigorous stabilization. Spinal fusion can be resorted to. Our preference is posterior spinal fusion with two rods with double sublaminar wires at each level and cross-links for rotary stabilization. In this patient, curve correction was from 90~ to 40~. [Reprinted from the Journal of the American Academy of Orthopaedic Surgeons, Volume 6(4), pp. 225-236, 9 1998 American Academy of Orthopaedic Surgeons, with permission.]

and type IV OI on the basis of the p r e s e n c e or a b s e n c e of d e n t i n o g e n e s i s imperfecta. D e n t a l t r e a t m e n t includes crowning, dentures, and i n t r a o s s e o u s implants. T h e blue sclerae in OI are the result of increased corneal translucency ( s e c o n d a r y to a b n o r m a l collagen), which re-

veals the u n d e r l y i n g uveal p i g m e n t and b l o o d vessels. This color changes with age, b e c o m i n g m o r e grayish in adulthood. T h e pericorneal region of the sclera is often white and opaque, resulting in a " S a t u r n ' s r i n g " appearance, and there m a y be opacities in the p e r i p h e r y of the cornea, giving an

SECTION XII ~ Review o f Specific Skeletal Dysplasias

F I G U R E 28 (continued)

853

854

CHAPTER 9 ~ Skeletal Dysplasias

arcus juvenilis appearance. The classic blue sclerae are most characteristic of the more benign autosomal dominant type I variants. The more severe type III progressively deforming types often have white or slate gray sclerae rather than blue sclerae. Hypermetropia is common. The skin in OI is often thin, translucent, and easily distensable due to collagen insufficiency of the dermal layer. Surgical scars commonly heal with widening. Ligamentous laxity is a characteristic feature of OI. Pes planus is the principal clinical manifestation, but other disorders of hyperlaxity, such as subluxating patellae and dysplastic hips, are occasionally seen. There is often a secondary muscular hypotonia and underdevelopment related to tendon or ligament anomalies and reduced activity. There is increased vascular fragility, which presents with bruising due to the weakened perivascular connective tissues. A small minority of patients demonstrate nonprogressive aortic root dilatation. Valvular disease, particularly mitral valve prolapse, is much less common than in Marfan syndrome but has been reported. The onset of hearing loss in OI, seen primarily in the type I benign autosomal dominant variant, begins in adolescence and becomes problematic for nearly 50% of affected adults. Heating loss can be conductive, sensorineural, or mixed. The extent of deafness is variable; however, loss in the high-frequency range is characteristic. Treatment usually involves prosthetic stapedial footplate replacement or stapedectomy. Other otologic findings include recurrent middle ear infections and sinusitis, tinnitus from stapedial fixation, vertigo from labyrinthine involvement, and speech delay. Low-pressure hydrocephalus is occasionally seen with severe forms of OI. The anterior fontanel remains open, and there is general dilation of the ventricles with cortical atrophy. The process usually is self-limiting and does not require shunting. Basilar impression occurs predominantly in Sillence type III and type IV OI. There is brain stem and spinal cord compression at the foramen magnum, resulting in progressive cerebellar disturbance and lower cranial decompression, which has variable results. Metabolic abnormalities, present to a variable extent in patients with OI, are characterized by hypermetabolism, heat intolerance, elevated body temperature (hyperthermia), increased sweating, resting tachypnea, and tachycardia. These findings are attributed to high metabolic activity and turnover of the connective tissue cells. 4. DIAGNOSIS a. Postnatal Diagnosis. The diagnosis of OI is based primarily on clinical and radiographic criteria. Fibroblast cell culture from skin biopsy specimens and DNA (blood) can detect collagen molecular abnormalities (see Section VI and XII.CC.7 below) in approximately 85% of patients with OI. The differential diagnosis of OI in infancy is relatively limited. Diagnosis of mild variants in childhood and adolescence can be more difficult due to less severe involvement.

Idiopathic juvenile osteoporosis is an extremely rare disorder with onset at puberty. It is characterized by generalized osteopenia and propensity to fracture. Spontaneous remission has been reported. It can be distinguished from OI by its onset in adolescence, the presence of normal sclerae and teeth, and a negative family history. Hypophosphatasia presents with multiple fractures, deformities, and osteopenia. Hypophosphatasia resembles OI with blue sclerae, bowing of the legs, and osteopenia; it can be differentiated from OI by abnormal laboratory findings such as the urine excretion product phosphoethanolamine and markedly decreased serum alkaline phosphatase levels. The distinction between mild forms of OI and child abuse can be difficult but is crucial to consider. Both may present with a propensity to fracture without a clear history of definite trauma. Classically, abuse presents with multiple fractures in different stages of healing, posterior rib fractures, metaphyseal comer fractures, and skull fractures. Other signs of abuse include bruises, bums, and retinal hemorrhages. b. Prenatal Diagnosis. Prenatal diagnosis of OI can be made on the basis of structural characteristics noted on fetal ultrasound, collagen molecular studies of cultured chronic villus cells, or genetic linkage studies with the use of collagen markers. Prenatal detection in pregnancies at risk for OI can provide valuable information for genetic counseling and obstetric management. Genetic counseling can provide information concerning the likely outcome of future pregnancies and the prognosis for individual affected children. In actuality, however, this is difficult because of the high incidence of spontaneous mutations and sporadic cases, the lack of diagnostic tests in the index case, and the lack of an accurate carrier test. Ultrasonography can effectively screen fetuses for severe forms of OI, but it remains difficult to detect mild forms. Sillence type II OI (OIC-A) can be recognized on ultrasound scans obtained at 16-20 weeks gestation by assessing femoral length adjusted for gestational age, extent of mineralization, evidence of fractures, skull echogenicity, and thoracic abnormalities. Midtrimester ultrasonography is usually useful in detecting type III OI by depiction of intrauterine fractures and deformity. Advances in transvaginal ultrasound may provide a means of first trimester prenatal diagnosis of severe forms of OI. If a certain biochemical defect of collagen or a specific mutation has been identified in an affected parent or sibling, prenatal detection can be accomplished by screening fetal tissue for the presence of that defect. This is performed by culturing chorionic villus cells and examining the electrophoretic properties of the collagen they produce. Amniotic fluid cells are, in general, not useful because most cells synthesize a variant of type I procollagen. Linkage studies performed with the use of collagen markers are currently the diagnostic investigation of choice for families with autosomal dominant OI, allowing genotyping of the fetus.

SECTION Xll ~ Review of Specific Skeletal Dysplasias 5. DETAILED ASSESSMENT OF CONSEQUENCES OF DIAGNOSIS

a. Osteogenesis Imperfecta Congenita A. Fifteen of the 16 patients died (94%) (280). Two were stillborn, 1 died at 4 hr, 6 between 1 and 6 weeks, 2 at 3 months, and the others at 4, 11, and 16 months and 3 years 9 months of age. One patient survived to 14 years of age, but never ffalked and was wheelchair-bound following intramedullary fixation of all four major lower extremity long bones. The patients who survived until 11 and 16 months of age had sequential radiographs that demonstrated progressive widening of the long bones and fibs. Radiographs in the 14-year survivor demonstrated the femurs to be short, wide, and crumpled at birth, whereas the tibias were broad and deformed with severe anterior angulation. The ribs were thin with only occasional local broadening due to repair; the rib cage demonstrated more normal size and shape. The long bones widened until 4 months of age, then began to diminish in width, and subsequently required intramedullary rods at 6 years. A caput membranaceum with wormian bone formation was noted in all patients. All had a negative family history with clinically normal parents. Two of the children with OIC-A were brother and sister, born 3 years apart, who died at 3 weeks and 3 months, respectively. There were no instances of parents with children with OIC-A giving birth to less seriously affected children with osteogenesis imperfecta. b. Osteogenesis Imperfecta Congenita B. Two patients (8%) died of respiratory problems at 3 and 5 months. Sixteen (59%) of the patients were wheelchair-bound, and 9 (33%) were ambulatory in some fashion. Eight of the 16 wheelchairbound patients had intramedullary rods placed in from one to four of the major lower extremity long bones. Five patients walked independently without requiting intramedullary rods or external assistive devices, although 3 required osteotomy or osteoclasis. Four patients walked using external assistive devices; 3 of these had stabilization with intramedullary rods. A caput membranaceum with wormian skull bone formation was noted in all patients. Some developed a normal appearing skull radiographically with time, whereas others persisted with delayed fontanelle closure and/or wormian bones several years after birth. Eighteen patients with early radiographs showing all bones each had evidence of healed or healing intrauterine fractures. A negative family history was documented in 96% (24 of 25 patients). The 1 individual with a positive family history had a father with osteogenesis imperfecta who walked with crutches after multiple surgical procedures, including femoral rod placement, and a brother who had an initial fracture at 2 weeks of age. This patient has had all four lower extremity long bones internally stabilized with rods but remains in a wheelchair at 10 years of age with long leg orthoses for protection.

855

c. Osteogenesis Imperfecta Tarda A. There were no deaths in this group. Seven patients (33%) ultimately were wheelchair-bound, whereas 14 (67%) were able to walk. Three of the 7 who were in wheelchairs had intramedullary rods. Five in the ambulatory group walked independently, whereas 9 used crutches or braces or had been stabilized with rods. In this latter group, 5 had rods placed in from one to four lower extremity long bones. Four of those with rods required other assistive devices to walk, and one walked without other assistance. The age at initial fracture did not correlate with eventual ambulatory capability. In the 5 who eventually walked independently, the ages at initial fracture were 4 weeks, 5 weeks, 6 weeks, 6 months, and 12 months. In those who ultimately were wheelchair-bound, the ages at initial fracture were 3, 4 (3 patients), 7, 9, and 10 months. A negative family history was documented in 89% (17 of 19). Of the 2 with a positive family history, one girl's father had OIC-B. The other patient's father had OIT-A. He had also fractured shortly after birth, and his family had a high incidence consistent with an autosomal dominant transmission. d. Osteogenesis Omperfecta Tarda B. All patients were ambulatory. There were no deaths. Of the 21 patients, 18 were able to walk independently, and 3 were ambulatory with crutches. Of these 3, one had rods placed in all four lower extremity long bones. The age at initial fracture varied from 14 months to 9 years of age. Sixteen (76%) patients had a positive family history of the autosomal dominant pattern. Of the 5 with a negative history, 3 of these were the only patients with OIT-B who did not walk independently. 6. SIGNIFICANCE OF VARIOUS CLASSIFICATION SCHEMES

The temporal-radiographic classification of osteogenesis imperfecta reflects the severity of involvement and allows the physician to provide early guidelines concerning survival and ambulation. In the OIC-A group, a virtually universal (94%) death rate occurs. The appearance of the femurs, tibias, fibs, and rib cage is characteristic. Distinction from the patient with OIC-B is possible, however, on the basis of the newborn radiographs. Such a distinction is extremely important as there is a significantly better prognosis in the OIC-B group, with 92% surviving. The two most important areas in differentiating infants with OIC-A from those with OIC-B are the rib cage and the femurs. Chest size and shape in the child with OIC-B are normal, as are the individual fibs, even if some fractures have occurred. Some OIC-B long bones at birth have healed intrauterine or new fractures, but they also show extensive normally shaped areas possessing normal metaphyseal funnelization. An infant with severe OIC-B is distinguished from a patient with OIC-A by these regions of normal metaphyseal funnelization. Figure 28C illustrates the progressively improved rates of survival and ambulatory capability as one moves from the OIC-A group to the OIT-B group.

856

CHAPTER 9 ~ Skeletal Dysplasias

A classification of congenita patients has been reported that documents a similar prognosis based on radiographs of long bones, ribs, skull, and vertebrae. The skull has not been included here because of the difficulty in quantifying wormian bones and because of the fact that caput membranaceum and wormian bones were present in all patients with congenital disease. Thus, the skull findings did not correlate with survival and ambulation. Retrospective classification into mild, moderate, and severe forms based on the degree of bone deformity several years after birth has been reported, as has subclassification of the tarda group into type I, with lower extremity bowing, and type II, without bowing. It thus appears premature to abandon totally the congenita-tarda approach, as has been suggested. Other studies have detailed the range of variation noted in the intrauterine period (188) and with the congenita types (308). The Sillence classification correlates well with the congenita-tarda temporal-radiographic classification. The congenita A group encompasses Sillence type II lethal perinatal patients, congenita B and tarda A generally are Sillence type III patients with progressive deformity, and tarda B represents the Sillence type I benign autosomal dominant patients. It is difficult to categorize patients with osteogenesis imperfecta by this or any approach based on purely clinical criteria. Classification by the Sillence approach is particularly difficult in the early phases of the disease, which is the time that parental anxiety is highest. It frequently requires a few years to classify individuals appropriately with the Sillence approach, as tooth and scleral findings either are not apparent at birth or change with time. There is also a lack of unanimity as to clinical genetic patterns. Although the Sillence approach attempts to take the heterogeneity into account, it is evident from biochemical findings on molecular abnormalities of collagen that heterogeneity is vastly more extensive than originally thought. It still appears to be practical to consider the age at the time of initial fracture and the radiographic appearance of the bones at initial fracture as providing important guidelines for counseling and management. This is important particularly because fractures represent the most debilitating feature of the disease. Our modification of the congenita-tarda approach is an index of disease severity. It is not intended to, nor can it, provide a rigid disease classification, which can only come with molecular definition. The term "congenita" is used to refer to fractures occurring in utero or at birth, whereas initial usage of the term by Looser (176) applied congenita only to intrauterine cases, with birth fractures defined under the tarda group. Seedorff (278) and King and Bobechko (149) have also defined congenita in this way, subdividing the tarda group into tarda gravis, with fractures occurring at birth or in the first year of life, and tarda levis, with fractures occurring initially after 1 year. The approach used here attempts to overcome the problem of having cases with fractures at birth referred to as

tarda cases, which is semantically awkward. More important, when radiographs in such cases were carefully reviewed, 18 of 18 patients with total body radiographs showed old healed or healing intrauterine fractures. This led to the congenita designation, with all of these patients considered as OIC-B. The terflporal-radiographic approach delineates well patients with a predominantly negative family history (OIC-A 100%, OIC-B 96%, and OIT-A 89%) from those with a predominantly positive history (OIT-B 76%). Those with a negative history, already considered to be a heterogeneous group by clinical and collagen studies, would have both autosomal dominant new mutations and autosomal recessive patterns of inheritance. The positive history group comprises the more benign autosomal dominant condition. OIC-B and OIT-A are not genetically distinct, as shown by three families in which one affected member was OIC-B and another OIT-A. This phenomenon was also noted by Seedorff. It is also noted that three of the five patients with OIT-B and a negative family history were the only three patients with OIT-B not walking independently, which might indicate a more severe genetic abnormality. The radiographs at birth or at the time of initial fracture, if this occurs after birth, are used as the classification baseline for this approach. Bones that are originally narrow can widen with the onset of a fracture-repair-fracture-repair cycle. Thus, broadened bones in a 4-month-old child might be less ominous than in a newborn in a situation in which the newborn's radiograph demonstrated long bones to be normal in terms of length, width, and shape. When initial radiographs in a congenita form are not available and short, broad bones are demonstrated at 3 - 4 months of age, the significant prognostic factor is whether the bone width and shape are progressing toward normal with time or continuing to widen with time. If the latter situation is seen, the prognosis becomes progressively dimmer. Ambulation in those having intramedullary rods progressively improved from OIC to OIT-B. Of the patients with OIC having rods, 9 remained wheelchair-bound and 3 walked with assistance, whereas of the patients with OIT, 3 were wheelchair-bound, 5 walked with assistance, and 1 walked unaided. 7. MOLECULAR BIOLOGY OF COLLAGEN DEFECTS Studies examining the skin of patients with OI in the 1960s and 1970s gave the first clear indication of collagen abnormalities. Subsequently, skin fibroblast cell culture studies have confirmed that the cells of affected individuals produce either decreased amounts of collagen or defective collagen. Detailed biochemical investigations have demonstrated heterogeneity of type I collagen defects. There has been increasing awareness of underlying collagen genetic and molecular abnormalities in OI but little investigation of the ways in which these abnormalities translate into structural and functional defects of bone. Virtually all patients

SECTION XII ~ Review of Specific Skeletal Dysplasias with OI have mutations in one of the two genes, COLIA1 and COLIA2, that encode the chains of type I collagen, but virtually all mutations are different (156, 250, 359, 365). In osteogenesis imperfecta there have now been well over 150 molecular defects of type I collagen detected, and it has been estimated that more than 200 mutations can be expected to saturate the mutation map (360). The clinical spectrum ranges from benign autosomal dominant variants (Sillence I) in which patients have few fractures, to progressively deforming types (Sillence III and some Sillence IV) with multiple fractures and an often nonambulatory status, to lethal perinatal variants (Sillence III) in which the paucity of bone formation is marked and individuals are stillborn or die within the first year of life (12). Molecular biology studies have defined more than 150 specific gene mutations that result in OI, and investigations to characterize the mechanisms that translate these mutations into the various observed phenotypes are ongoing (156, 366). Osteogenesis imperfecta is caused by mutations in the COLIA1 (18-kb size, located on the long arm of chromosome 17) and COLIA2 (38-kb size, located on the long arm of chromosome 7) genes that encode the two pro-oL1 (I) and one pro-et2 (I) chains of the type I procollagen trimer, respectively. These genes contain more than 50 exons to generate about 1450 amino acids of each chain. The formation of the essential triple-helical structure of procollagen I from these three pro-tx (I) chains depends on the presence of glycine at every third position in the 1014-residue triple-helical domain and is stabilized by the presence of hydroxyproline. The triple helix is propagated from the carboxyl-terminal end toward the amino-terminal end of the molecule. The mature molecule is then secreted from the cell into the extracellular matrix, where the amino-terminal and carboxylterminal propeptides are removed enzymatically and type I collagen fibrillogenesis occurs by self-assembly. More than 150 mutations of the COLIA1 and COLIA2 genes, including single-base-pair changes, deletions, insertions, premature stop codons, and splicing mutations, have been described, causing forms of OI ranging in phenotype from mild to lethal. The most frequent mutation types are single-base-pair substitutions in either of the two alleles that later encode for glycine in the triple-helical domain of the chain. The molecular basis of Sillence type I OI remains poorly understood. Cells from affected individuals largely demonstrate a quantitative defect of type I collagen; they synthesize and secrete about half the normal amount of type I procollagen. This is due to decreased synthesis of pro-ix 1 (I) chains; in general the pro-a2 (I) chain is normal. The mutation often occurs in one allele, resulting in about half the normal amount of the molecule, as type I procollagen must contain two pro-ct 1 (I) chains. The vast majority of infants with type II OI appear to be heterozygous for mutations that result in substitutions for glycine residues within the triple-helical domain of either the

857

pro-txl (I) or the pro-tx2 (I) chain. In some instances, exon skipping mutations in either gene can result in this phenotype. Types III and IV phenotypes can also result from heterozygosity for point mutations of glycine residues within the triple-helical domains of either chain of type I procollagen or from exon skipping mutations. There are several basic concepts concerning the nature of mutations and their phenotypic consequences. Quantitative mutations that decrease the expression and synthesis of normal type I procollagen molecules result in milder phenotypes, such as type I OI. Qualitative mutations that lead to structural aberrations and abnormal type I procollagen result in more severe phenotypes, such as types II and IIII. The severity of the phenotype reflects the location of the mutation within the chain, the nature of the mutation, and the chain in which the mutation occurs. In general, with point mutations, the phenotype becomes milder as the mutation is shifted toward the amino-terminal end of the chain, because the essential triple-helical structure is formed from the carboxyl-terminal end toward the amino-terminal end in a zipperlike helical coiling manner. With glycine substitution mutations, amino acid replacements with bulkier side chains such as arginine result in more severe phenotypes than those with smaller side chains such as serine and cystene. Mutations in the COLIA2 gene have milder consequences than similar mutations in the COLIA1 gene, as the type I procollagen molecule has two pro-txl (I) chains and one pro-ct2 (I) chain. The expression of these mutations in the observed phenotypic patterns is still understood incompletely. On the molecular level, these mutations can decrease the rate of synthesis, decrease the thermal stability of the triple-helical molecules (5), delay the rate of folding of the precursor procollagen molecule (5, 251), increase the level of aberrant posttranslational modification of procollagen chains (332), and impair the rate of export of molecules from cell to matrix. Mutated chains are incorporated into fibrils more slowly than normal collagen molecules, demonstrate an abnormal configuration, with more branching, shortening, or thickening, and impair mineralization by providing a mutated structural template for the incorporation of hydroxyapatite crystals. The end result of these molecular changes eventually translates into the microscopic and gross pathological features characteristic of OI. Figure 28D demonstrates how collagen mutations lead to structural fibrillar changes. 8. HISTOPATHOLOGY OF OSTEOGENESIS IMPERFECTA BONE The histologic structure of the bone differs markedly in the various types, is essential to an appreciation of the functional capacity of the bone, and thus warrants independent study to allow for meaningful correlation with molecular, microstructural-biomechanical, and clinical-radiographic abnormalities (109). The histopathologic features of OI vary depending on disease severity (40, 46, 47, 75). Overall, there

858

CHAPTER 9 ~ Skeletal Dysplasias

is a generalized decrease in bone tissue, with the bone structure in most cases demonstrating a mixture of woven and lamellar patterns. The histologic appearance of bone in OI patients follows the normal developmental and structural pattem but rarely achieves the fully compacted lamellar state. The more severe the involvement, the more immature the structural pattern (Fig. 28E). In the most severe forms of OI, the bone appears histologically as a woven bone matrix devoid of any organized lamellar pattern (343). There are plump osteoblasts crowded along prominent osteoid seams, large oval osteocytes surrounded by small amounts of matrix, and morphologically normal osteoclasts. Histological sections from patients with less severe OI show a definite tendency to lamellar bone formation. The lamellation and osteon formation can be rudimentary, partially compacted, or fully compacted in localized areas. Osteoid seams are prominent and there is hypercellularity, with larger than normal osteocytes and osteoblasts in the areas of woven bone. There is a correlation between the degree of osteonal maturity and bone strength as indicated by ambulatory status. Efforts are being made to correlate molecular defects with histologic changes (327). Bone has long been recognized in a biomechanical sense as a nonhomogeneous composite structure with mineral and organic phases. It is structured in an intricate hierarchical arraymmolecules, macromolecules, microfibril, fibrils, fibers, and so onmwith each level of structure serving as a template or scaffolding for the next larger level. Much is known about collagen deposition patterns in normal bone, but detailed structural analyses have not been performed in relation to the pathogenesis of OI bone and in relation to its molecular abnormalities. Cortical bone formation is initiated with collagen deposition in a randomly oriented pattem referred to as woven or coarse-fibered bone followed by its progressive replacement with a definitive lamellar structure until the matrix becomes well-compacted in osteonal systems. Jaffe (136) has defined coarse-fibered primary bone and fine-fibered lamellar bone, Weidenreich (355) defined woven-fibered bone, parallel-fibered bone, and lamellar bone, and Smith (300) defined woven-fibered bone primary osteons, secondary osteons, surface bone, and interstitial bone. A more detailed study of the three-dimensional patterning of collagen has been performed using polarization optics and both transmission and scanning electron microscopy. GiraudGuille (92) has defined a "twisted plywood" pattern of collagen deposition. Neville (222) has summarized the biology of fiber deposition patterns in many plant and animal tissues, including bone with the lamellar bone containing both an orthogonal and helicoidal architecture. Even the lacunarcanalicular system in bone, within which reside the osteocytes and osteocyte processes linking adjacent cells, is felt to have structural and biomechanical implications. Because the organization of collagen in OI bone is always imperfect with abnormal cortical bone development being the patho-

logic hallmark, knowledge of the mechanisms of matrix transformation is important. 9. ORTHOPEDIC MANAGEMENT

The overall goals of treatment of OI are to maximize comfort and function, minimize fracture and deformity, and allow the patient to achieve independence (153). Care obviously must be individualized, depending primarily on the severity of the disease and the age of the patient. The general approaches used in most centers seeing relatively large numbers of patients follow. Sillence type I OI (OIT-B), especially in its milder forms, has only a minimal impact on the patient and the role of the orthopedic surgeon is limited to conventional fracture care. Early death almost invariably occurs in type II OI (OIC-A) before any orthopedic intervention is indicated, whereas the fractures in those who survive longer simply are splinted for comfort. Type III and type IV OI present the greatest management challenges for the orthopedist in terms of fracture prevention, fracture management, limitation of deformity, and optimization of function. Most of the patients in the III-IV category of Sillence are OIC-B or OIT-A by the temporal-radiographic congenita-tarda approach. In the long-term series from which the latter classification was derived, approximately one-third of OIC-B patients walked, usually with treatments of rods, braces, or crutches, whereas two-thirds of OIT-A patients walked. The aim of therapy in these groups is to convert non-ambulatory patients to the ambulatory group while minimizing deformity and negative postfracture sequelae. The management of OI begins with early detection in utero for pregnancies at risk to guide obstetric management. Cesarean section is often recommended to minimize birth trauma. In the neonate with deforming OI (type III and some type IV), immediate life-threatening problems of respiratory insufficiency and intracranial hemorrhage are managed by neonatologists. In utero fractures have usually healed and recent fractures often do not require special treatment other than splinting. The parents are educated about handling of the infant and about the natural history of the disease. In infancy and childhood, physical therapists can help to optimize normal development as the infant develops trunk control and limb use. The orthopedist is involved in helping to obtain the greatest degree of mobility possible and in the treatment of fractures, which can be frequent. When fractures occur, we use as little immobilization as possible to minimize disuse osteopenia and muscle atrophy. Well-padded fiberglass or plaster splints with Ace bandages are favored over complete circumferential costs. In moderate and severe OI, the multiplicity of fractures, the underlying osteoporotic bone, and the abnormal mechanical stresses on malaligned bones can lead to further fractures and deformity, interfering with the ability to stand and walk. After the first 2 or 3 years of age in those who are showing a considerable propensity

SECTION Xll ~ Review of Specific Skeletal Dysplasias

to fracture and angulate, external support from orthoses may be necessary to optimize function. Deformity that is impairing function can be addressed surgically by multiple osteotomies and intramedullary fixation. Both bracing and intramedullary rods are used, however only, when fully independent function is failing because both methods decrease bone weight bearing and can predispose one to further osteopenia. Fracture rates almost always decrease dramatically after puberty in both males and females (206). The value of keeping the limbs well-aligned until this time is further heightened by this fact. The diminished rates of fracture are well-documented. They are not specifically due to increased caution with age but appear to represent a relative biological strengthening of the bone presumably due to diminished bone turnover, allowing the matrix synthesized to persist and be incorporated into structurally stronger units. There is no evidence of any change in collagen type synthesized with time. The fracture rate increases again in women after menopause and in men after age 60 years (231). Nonoperative treatment is the mainstay of orthopedic management of OI. Depending on disease severity, a comprehensive and progressive program of mobilization and bracing if needed is pursued for patients with ambulatory potential. The goal is to emphasize ultimate independent function. Lightweight bracing can be helpful for external structural support to promote stance and locomotion and for the prevention and treatment of fractures. Braces allow for earlier weight bearing postfracture. A program of appropriate seating and wheelchair locomotion is pursued for nonambulatory patients. Closed treatment methods are the mainstay of fracture management. Fractures generally heal, often with exuberant callus, but with the same abnormal bone quality. There is a subset of patients, however, who heal slowly and often persist for years with fibro-osseous unions, which can be quite firm. This appears to occur in less than 5%. The difficulty in fracture management lies in the prevention of angular and rotational deformities and the vicious circle associated with immobilization--further osteoporosis and muscle atrophy-and repeat fracture. Light-weight splints and braces are most often used with the emphasis on early mobilization. Operative intervention is indicated for recurrent fractures at the same site, deformity that impairs function, and femoral fractures after 10 years of age to prevent long-term immobilization with skeletal traction-hip spica treatments. In the latter group, intramedullary fixation, with or without a small side plate and long leg casting with or without a pelvic band with hip hinge can keep a child ambulatory during the healing phase. The optimal age for surgical intervention for deformity is controversial; some recommend early intervention with elongating rods, but traditional management involves accepting deformity from closed treatment until the patient is about 3 years old and then proceeding to corrective osteoto-

859

mies. Various techniques for deformity correction have been described, including closed osteoclasis with casting or splinting, closed osteoclasis with percutaneous intramedullary nailing, and multiple corrective osteotomies with open intramedullary nailing described initially by Sofield and Millar (302, 335). Fractures and deformity are much more common in the lower extremity than in the upper extremity in the OI III, progressively deforming variants. In addition, deformity is better tolerated in the upper extremity. As a result, bracing rarely if ever is required for the upper extremity. The bracing referred to is lower extremity, and virtually all osteotomies and surgical stabilizations involve femurs and tibias. Management must be highly individualized, taking into account personal experience, the severity of the patient's deformity, the diameter of the long bones, and the advantages and complications of each technique. Elongating intramedullary rods have a decreased replacement rate when compared with nonelongating rods; however, they are weaker, traverse the physis, are technically more demanding as they require a larger medullary canal and central placement, and are associated with complications related to disassembly and nonelongation (9, 85). More recent reports show a tendency to improved results with elongating rods (245) especially with a modification developed by Wilkinson et al. in Sheffield, England (363). General principles in the surgical management of OI include avoiding bone surgery in patients under age 2 years, avoiding plate-and-screw fixation in favor of intramedullary fixation, and using gentle techniques for muscle preservation and minimization of soft tissue bleeding. Bone holding clamps should be avoided or used with caution as they can crush fragile bone. Radiographic control is essential as the deformities are often three-dimensionally complex, necessitating different views. When multiple osteotomies are performed, the individual fragments should be as long and straight as possible. Placement of osteotomies in diaphyseal regions enhances stability with placement of intramedullary rods (Fig. 28F). Some bone shortening may be necessary when there are severe deformities, because the taut soft tissue structures on the concave side can be stretched excessively when deformity is corrected. Many patients with OI III requiting long bone osteotomy and intramedullary fixation have narrow bones, which only allow for Rush rod or even Steinmann pin fixation. Radiographs can be misleading concerning the medullary canal diameter because the diaphyses in those with severe involvement are often oval or flattened in one plane, further decreasing the diameter of rod that can be inserted. Violation of the growth plate should be avoided when possible. Immobilization with casting or braces until bone union is almost always necessary. Spinal deformity in patients with OI is especially difficult to manage (101). Truncal shortening of thoracolumbar spinal

860

CHAPTER 9 ~

Skeletal Dysplasias

segments can occur secondary to collapse of osteopenic vertebrae. If the patient is symptomatic, a soft anterior opening spinal orthosis is helpful. Scoliotic and kyphoscoliotic curves often progress rapidly. The soft, deformed rib cage and truncal shortening combine to render bracing relatively ineffective, but it can be helpful in mild forms of OI and as a sitting aid for the nonambulator. In milder forms of OI, bracing can be utilized for curves of 20-40 ~ or kyphosis greater than 40 ~ Spinal fusion has been recommended for scoliotic curves greater than 45 ~ to halt progression (Fig. 28G). For patients with more severe involvement, bracing is ineffective and can produce thoracic cage deformity; fusion is recommended for curves over 35 ~ as these curves are most often progressive and potentially severe. A higher incidence of complications and relatively lesser correction are noted from spinal fusion in OI because internal fixation is limited by poor bone quality, autogenous iliac crest bone graft is relatively ineffective, and patients have a tendency to increased bleeding. Worsening motor function progressing even to quadriparesis can occur in severe OI III and II variants due to cervicomedullary compression with basilar invagination, similar to what occurs in some skeletal dysplasias with occipitocervical deformity (274). The anesthesiologist should be aware of the increased risk of malignant hyperthermia in patients with OI. Anesthetic principles include avoiding the use of atropine, careful metabolic monitoring, not insulating the patient with large numbers of drapes, and stopping the operation with signs of rapidly increasing hyperthermia. Patients with OI may have platelet and coagulation abnormalities, and perivascular fragility due to collagen abnormalities may predispose to bleeding. Blood loss and insensible losses due to hypermetabolism should be carefully monitored.

10. MEDICAL MANAGEMENT At present there is no cure for OI. Multiple previous attempts to strengthen bones with fluoride and calcitonin were ineffective, and once collagen was identified as the molecular site of disease medical efforts temporarily slowed. Three current forms of intervention under investigation, however, appear to hold possible promise for the future. Glorieux et al. have reported on the value of cyclic intravenous administration of pamidronate, a bisphosphonate, for diminishing the fracture rate (by 1.7 per year), diminishing chronic bone pain, and increasing bone density (94). The study on 30 children 3-16 years of age with types III and IV OI relied on the drug's ability to inhibit resorption, thus increasing matrix mass. Another modality under investigation seeks to actually replace deficient collagen secreting bone cells by bone marrow transplant or marrow stromal cell-mesenchymal stem cell transplants (236, 249). These cells are precursors of osteoblasts and could theoretically synthesize normal collagen, thus eliminating the deficient gene pool. The final approach involving the suppression of mutant genes seeks to decrease

the expression of the mutant allele by introducing ribozymes into the cells to cleave the mutant gene product, leaving the normal gene product intact and bringing about a milder phenotype with norml but quantitatively less collagen.

DD. Trichorhinophalangeal Dysplasia The trichorhinophalangeal dysplasia syndrome has been increasingly recognized over the past few decades since its original description in 1966. The affected individuals have fine, sparse, and brittle hair, a pear-shaped, bulbus nose and long philtrum, and cone-shaped epiphyses in the phalanges. Most patients are also short and many show hip abnormalities characteristic of either unilateral or more often bilateral leg Perthes disease. The hip abnormalities, however, almost invariably are of the more severe gradient of the Perthes disorder and tend to lead to flattening and widening of the head with hinge abduction phenomena. Premature osteoarthritis occurs, and many patients have considerable discomfort even in adolescence. The disorder is autosomal dominant and it increasingly is suspected that it is more commonly present than has previously been recognized. Shortening of the fingers and toes generally accompanies the cone shaped epiphyses and may be particularly apparent in metacarpals and metatarsals. Cone-shaped epiphyses are present in many skeletal dysplasias, having been described in over 20 disorders, and thus must not be considered as diagnostic of TRP dysplasia alone. The most serious complication is the hip disorder, although the avascular necrosis is variable. It tends, however, to the more severe variants. Eventual changes of coxa magna, coxa plana, coxa vara, and a short neck are described. A wide variety of bone and ligament abnormalities have been reported, including joint laxity, scoliosis, dislocatable and even absent patellae, delayed tooth eruption, and spondylolisthesis. It is the hip disorder, however, that is most commonly seen and is the most severe skeletal aspect of the disorder. Progressive degenerative arthritis almost invariably occurs in the hips but can also affect the spine, knees, and elbows. It has been suggested that early diagnosis and more vigorous containment therapy might help to improve the end results. In one study, however, all patients eventually evolved to a Stulberg class IV deformity (59, 67).

XIII. A N E S T H E T I C I M P L I C A T I O N S IN T H E SKELETAL DYSPLASIAS Orthopedic and other surgical procedures frequently are indicated in patients with skeletal dysplasia. Although these can be performed safely in the vast majority of cases, there are specific structural anatomic and physiologic abnormalities associated with the skeletal dysplasias that must be carefully considered by the anesthesia team (25, 256).

References

A. Occipital and Cervical Structural Abnormalities Frequent problems can include foramen magnum stenosis, occipitalization of C 1, odontoid hypoplasia, and cervical kyphosis and stenosis. Manipulation of the cervical region during intubation and anesthesia in patients with these disorders can produce abnormal cord pressure and pyramidal tract signs. A routine preoperative series of cervical spine radiographs is warranted in the skeletal dysplastic patient, with further study involving CT scanning or MR imaging recommended for those in whom structural abnormalities are detected. Virtually all patients with achondroplasia have stenosis of the foramen magnum, with 5-10% showing some evidence of cervicomedullary compression. The patients with Morquio's disease all have abnormalities of the odontoid, and most patients with SED congenita also have such irregularities. In such children, positioning on the operating table is done with even greater care than normal, tracheal intubation is cautious with flexion-extension of the neck minimized, and in many instances fiber-optic intubation is warranted. In many patients with spondyloepiphyseal dysplasias the neck is short, which makes intubation more difficult, and many of these patients also have a pectus carinatum, which further augments the difficulty.

B. Airway Abnormalities Upper airway obstruction can occur not only due to cervicaloccipital anatomy but also due to many other disorders that can produce obstruction, including a large tongue, large tonsils and adenoids, stiffened temporomandibular joints, tracheal narrowing due to diminished transverse growth of the tracheal cartilage tings, and angular malposition of the maxilla and mandible due to facial growth abnormality. Patients with the mucopolysaccharidoses have difficult airway intubation. Other abnormalities described include laryngomalacia, laryngotracheal stenosis, micrognathia, and cleft palate.

C. Pulmonary Abnormalities Those with skeletal dysplasia also can have respiratory dysfunction from such additional factors as thoracic cage malformation with rib hypoplasia and progressive kyphosis, scoliosis, and thoracic lordosis. Structural abnormalities leading to thickening and narrowing of the walls of the trachea and bronchi, particularly in the mucopolysaccharidoses and in some other disorders with cartilage hypoplasia, can occur. Many of the thoracolumbar kyphoscoliotic conditions in the skeletal dysplasias tend to be more rigid than those seen in idiopathic disorders, such that respiratory compromise is potentially greater during surgery. Diaphragmatic motion can be further aggravated in those with kyphosco-

861

liosis and short stature by associated hepatosplenomegaly, particularly if it is massive.

D. Neurologic Abnormalities The three major neurologic abnormalities associated with skeletal dysplasias are hydrocephalus, cervicomedullary compression, and compressive spinal cord and nerve root syndromes. The latter two findings are associated with foramen magnum stenosis (invariable in achondroplasia), odontoid hypoplasia with cervical instability (seen in many disorders), generalized thoracolumbar spinal stenosis, lumbosacral spinal stenosis, and severe kyphosis or kyphoscoliosis usually at the thoracolumbar junction. Many patients with skeletal dysplasia have disproportionately large heads due to the fact that skull and facial bone are minimally to nonaffected because they are preformed in membrane and not in cartilage. Macrocephaly alone is not of medical or anesthetic concern. In those situations in which hydrocephaly is prominent, signs of elevated intracranial pressure may necessitate decompressive ventriculoperitoneal shunting as a primary procedure.

E. Abnormalities of Thermal Regulation Patients with osteogenesis imperfecta often show a generalized disturbance of energy metabolism, leading to a hypermetabolic state characterized by elevation of body temperature, elevated oxygen consumption, and diaphoresis. Under anesthesia the body temperature often rises even further, although in most instances these are not associated with true malignant hyperthermia with its muscle rigidity, arrhythmias, and acidosis. A few true episodes of malignant hyperthermia have been described, however, in OI.

F. Cardiac Abnormalities Cardiac abnormalities are not a component of the vast majority of skeletal dysplasias, but they are seen in those disorders of bone growth and development associated with metabolic abnormalities. The most common in this group are the mucopolysaccharidoses, in which abnormal glycosaminoglycan deposition can occur in heart muscle, coronary arteries, and valves. Cor pulmonale is seen frequently in those with severely restrictive lung disease particularly with the longstanding rigid kyphoscolioses.

References 1. Adams BK (1988) Scintigraphy in a patient with complicated osteopetrosis. Clin Nuc Med 14:323. 2. Anderson IJ, Goldberg RB, Marion RW, Upholt WB, Tsipouras P (1990) Spondyloepiphyseal dysplasia congenita:

862

CHAPTER 9 9

Skeletal Dysplasias

Genetic linkage to type II collagen (COL2A1). Am J Hum Genet 46:896-901. 3. Arms DM, Strecker WB, Manske PR, Schoenecker PL (1997) Management of forearm deformity in multiple hereditary osteochondromatosis. J Pediatr Orthop 17:450-454. 4. Armstrong DG, Newfield JT, Gillespie R (1999) Orthopedic management of osteopetrosis: results of a survey and review of the literature. J Pediatr Orthop 19:122-132. 5. Bachinger HP, Morris NP, Davis JM (1993) Thermal stability and folding of the collagen triple helix and the effects of mutations in osteogenesis imperfecta on the triple helix of type I collagen. Am J Med Genet 45:152-162. 6. Bailey JA, II (1970) Orthopedic aspects of achondroplasia. J Bone Joint Surg 52A:1285-1301. 7. Bailey JA, II (1971) Forms of dwarfism recognizable at birth. Clin Orthop Rel Res 76:150-159. 8. Bailey JA, II (1973) Disproportionate Short Stature. Diagnosis and Management. Philadelphia: WB Saunders. 9. Bailey RW, Dubow HI (1963) Studies of longitudinal bone growth resulting in an extensible nail. Surg Forum 14:455-458. 10. Ballo R, Briggs MD, Cohn DH, Knowlton RG, Beighton PH, Ramesar RS (1997) Multiple epiphyseal dysplasia, Ribbing type: A novel point mutation in the C O M P gene in a South African family. Am J Med Genet 68:396-400. 11. Beals RK, Edo WH Akeson, Bornstein R Glimcher MJ (1982) Orthopedic care for patients with skeletal dysplasia. AAOS symposium on heritable disorders of connective tissue. pp. 362-368, St. Louis: CV Mosby Co. 12. Beals RK, Horton W (1995) Skeletal dysplasias: An approach to diagnosis. J Am Acad Orthop Surg 3:174-181. 13. Beals RK, Lovrein EW (1976) Dyschondrosteosis and Madelung's deformity. Report of three kindreds and review of the literature. Clin Orthop Rel Res 116:24-28. 14. Beard CJ, Key L, Newburger PE, Ezekowitz AB, Arecci R, Miller B, Proto R, Ryan T, Anast C, Simons ER (1986) Neutrophil defect associated with malignant infantile osteopetrosis. J Lab Clin Med 108:498. 15. Beighton P (1978) Inherited Disorders of the Skeleton. Genetics in Medicine and Surgery Series, 2d ed. Livingstone: Churchill Livingstone. 16. Beighton P (1996) Osteochondrodysplasias in South Africa. Am J Med Genet 63:7-11. 17. Beighton R Geidion A, Gorlin R, Hall J, Horton B, Kozlowski K, Lachman R, Langer LO, Maroteaux R Poznanski A, Rimoin DL, Sillence D, Spranger J (1992) International classification of osteochondrodysplasias. Am J Med Genet 44:223-229. 18. Beighton R Kozlowski K (1980) Spondyloepimetaphyseal dysplasia with joint laxity and severe progressive kyphoscoliosis. Skeletal Radiol 5:205-212 19. Belin V, Cusin V, Viot G, Girlich D, Toutain A, Moncla A, Vekemans M, LeMerrer M, Munnich A, Cormier-Daire V (1998) SHOX mutations in dyschondrosteosis (Leri-Weill syndrome). Nature Genet 19:67-69. 20. Bell DF, Boyer MI, Armstrong PF (1992) The use of the Ilizarov technique in the correction of limb deformities associated with skeletal dysplasia. J Pediatr Orthop 12:283-290. 21. Bellus GA, Hefferon TW, Ortiz de Luna RI, Hecht JT, Horton WA, et al. (1995a) Achondroplasia is defined by recurrent G380R mutations of FGFR3. Am J Hum Genet 56:368-373.

22. Bellus GA, Mclntosh I, Smith EA, Aylsworth AS, Kaitila I, Horton WA, Greenhaw GA, Hecht JT, Francomano CA (1995b) A recurrent mutation in the tyrosine kinase domain of fibroblast growth factor receptor 3 causes hypochondroplasia. Nat Genet 10:357-359. 23. Bellus GA, Mclntosh I, Szabo J, Aylsworth A, Kaitila I, Francomano CA (1996) Hypochondroplasia: Molecular analysis of the fibroblast growth factor receptor 3 gene. Ann NY Acad Sci 785:182-187. 24. Bennett JT, McMurray SW (1990) Stickler syndrome. J Pediatr Orthop 10:760-763. 25. Berkowitz ID, Raja SN, Bender KS, Kopitis SE (1990) Dwarfs: Pathophysiology and anesthetic implications. Anesthesiology 73:739-759. 26. Bessel Hagen F (1891) Ueber Knochen-und Gelenkanomalieen, insbesondere bei partiellem Riesenwuchs und bei multiplen cartilaginaren Exostosen. Arch Klin Chir 41:420-466, 505-552, 749-823. 27. Bethem D, Winter RB, Lutter L (1980) Disorders of the spine in diastrophic dwarfism: A discussion of nine patients and review of the literature. J Bone Joint Surg 62A:529-536. 28. Bethem D, Winter RB, Lutter L, Moe JH, Bradford DS, Lonstein JE, Langer LO (1981) Spinal disorders of dwarfism. J Bone Joint Surg 63A:1412-1425. 29. Bethge JFJ (1963) Hereditare, multiple Exostosen and ihre pathogenetische Deutung. Archiv Orthop Unfall-Chir 54: 667-696. 30. Black B, Dooley J, Pyper A, Reed M (1993) Multiple hereditary exostoses: An epidemiologic study of an isolated community in Manitoba. Clin Orthop Rel Res 287:212-217. 31. Bogaert R, Wilkin D, Wilcox WR, Lachman R, Rimoin D, Cohn DH, Eyre DR (1994) Expression in cartilage of a 7amino-acid deletion in type II collagen from two unrelated individuals with Kniest dysplasia. Am J Hum Genet 55:11281136. 32. Bollerslev J (1989) Autosomal dominant osteopetrosis: Bone metabolism and epidemiological, clinical and hormonal aspects. Endocr Rev 10:45-67. 33. Bollerslev J, Anderson PE, Jr (1988) Radiological, biochemical and hereditary evidence of two types of autosomal-dominant osteopetrosis. Bone 9:7-13. 34. Bollerslev J, Anderson PE, Jr (1989) Fracture patterns in 2 types of autosomal-dominant osteopetrosis. Acta Orthop Scand 60:110-112. 35. Bonaventure J, Rousseau F, Legeai-Mallet L, LeMerrer M, Munnich A, Maroteaux P (1996) Common mutations in the fibroblast growth factor receptor 3 (FGFR3) gene account for achondroplasia, hypochondroplasia, and thanatophoric dwarfism. Am J Med Genet 63:148-154. 36. Bonnemann CG, Cox GF, Shapiro F, Wu J-J, Feener CA, Thompson TG, Anthony DC, Eyre DR, Darras BT, Kunkel LM (2000) A mutation in the alpha-3 chain of type IX collagen causes autosomal dominant multiple epiphyseal dysplasia with mild myopathy. Proc Natl Acad Sci USA 97: 1212-1217. 37. Briggs MD, Hoffman SMG, King LM, Olsen AS, Mohrenweiser H, Leroy JG, et al. (1995) Pseudoachondroplasia and multiple epiphyseal dysplasia due to mutations in the cartilage oliogometric matrix protein gene. Nature Genet 10: 330-336.

References 38. Briggs MD, Mortier GR, Cole WG, King LM, Golik SS, et al. (1998) Diverse mutations in the gene for cartilage oligomeric matrix protein in the pseudoachondroplasia-multiple epiphyseal dysplasia disease spectrum. Am J Hum Genet 62: 311-319. 39. Buckwalter JA, E1-Khoury GY, Flatt AE (1978) Dysplasia epiphysealis hemimelica of the ulna. Clin Orthop Rel Res 135:36-38. 40. Bullough PG, Davidson DD, Lorenzo JC (1981) The morbid anatomy of the skeleton in osteogenesis imperfecta. Clin Orthop Rel Res 159:42-57. 41. Byers PH, Steiner RD (1992) Osteogenesis imperfecta. Ann Rev Med 43:269-282. 42. Caffey J (1958) Achondroplasia of pelvis and lumbosacral spine: Some roentgenographic features. Am J Roent 80: 449-457. 43. Cai G, Nakayama M, Hiraki Y, Ozono K (1998) Mutational analysis of the D T D S T gene in a fetus with achondrogenesis type lB. Am J Med Genet 78:58-60. 44. Cameron HU, Dewar FP (1977) Degenerative osteoarthritis associated with osteopetrosis. Clin Orthop Rel Res 127: 148-149. 45. Cardelia JM, Dormans JP, Drummond DS, Davidson RS, Duhaime C, Sutton L (1995) Proximal fibula osteochondroma with associated peroneal nerve palsy: A review of six cases. J Pediatr Orthop 15:574-577. 46. Casella JP, Stamp TCB, Ali SY (1996) A morphological and ultrastructural study of bone in osteogenesis imperfecta. Calcif Tissue Int 58:155-165. 47. Cassella JP, Barber P, Catterall AC, Ali SY (1994) A morphometric analysis of osteoid collagen fibril diameter in osteogenesis imperfecta. Bone 15:329-334. 48. Chang S, Prados MD (1994) Identical twins with Ollier's disease and intracranial gliomas: Case report. Neurosurgery 34: 903-906. 49. Chitty LS, Altman DG, Henderson A, Campbell S (1994) Charts of fetal size: 4: femur length. Fetal and Neonatal Medicine. Br J Obstet Gyn 101:132-135. 50. Coccia PF Krivit W, Cervenka J, Clawson C, Kersey JH, et al. (1980) Successful bone marrow transplantation for infantile malignant osteopetrosis. N Engl J Med 302:701. 51. Cohen MM, Jr (1997) Short-limb skeletal dysplasias and craniosynostosis: What do they have in common? Pediatr Radiol 27:442-446. 52. Cohn DH, Briggs MD, King LM, Rimoin DL, Wilcox WR, Lachman RS, Knowlton RG (1996) Mutations in the cartilage oligomeric matrix protein (COMP) gene in pseudoachondroplasia and multiple epiphyseal dysplasia. Ann NY Acad Sci 785:188-194. 53. Cohn DH (1996) Mutations affecting multiple functional domains of FGFR3 cause different skeletal dysplasia: A personal retrospective in honor of John Wasmuth. Ann NY Acad Sci 785:160-163. 54. Cole WG (1993) Etiology and pathogenesis of heritable connective tissue diseases. J Pediatr Orthop 13:392-403. 55. Committee on Genetics (1995) Health supervision for children with achondroplasia. Pediatrics 95:443-461. 56. Connor JM, Horan FT, Beighton P (1983) Dysplasia epiphysialis hemimelica: A clinical and genetic study. J Bone Joint Surg 65B:350-354.

863

57. Cooper RR, Ponseti IV (1973) Metaphyseal dysostosis: Description of an ultrastructural defect in the epiphyseal plate chondrocytes. J Bone Joint Surg 55A:485-495. 58. Cooper RR, Ponseti IV, Maynard JA (1973) Pseudo-achondroplastic dwarfism: A rough-surfaced endoplasmic reticulum storage disorder. J Bone Joint Surg 55A:475-484. 59. Cope R, Beals RK, Bennett RM (1986) The trichorhinophalangeal dysplasia syndrome: Report of eight kindreds, with emphasis on hip complications, late presentations, and premature osteoarthrosis. J Pediatr Orthop 6:133-138. 60. Coscia MF, Bassett GS, Bowen JR, Ogilvie JW, Winter RB, Simonton SC (1989) Spinal abnormalities in camptomelic dysplasia. J Pediatr Orthop 9:6-14. 61. Costa T, Ramsby G, Cassia F, Peters KR, Soares J, Correa J, Quelce-Salgado A, Tsipouras P (1998) Grebe syndrome: Clinical and radiographic findings in affected individuals and heterozygous carders. Am J Med Genet 75:523-529. 62. Crossan JF, Wynne-Davies R, Fulford GE (1983) Bilateral failure of the capital femoral epiphysis: Bilateral Perthes disease, multiple epiphyseal dysplasia, pseudoachondroplasia, and spondyloepiphyseal dysplasia congenita and tarda. J Pediatr Orthop 3:297-301. 63. Cruz-Conde R, Amaya S, Valdivia P, Hernandez M, Calvo M (1984) Dysplasia epiphysealis hemimelica. J Pediatr Orthop 4:625-629. 64. D'Ambrosia R, Ferguson AB (1968) The formation of osteochondroma by epiphyseal cartilage transplantation. Clin Orthop Rel Res 61:103-115. 65. Deere M, Sanford T, Ferguson HL, Daniels K, Hecht JT (1998) Identification of twelve mutations in cartilage oligomeric matrix protein (COMP) in patients with pseudoachondroplasia. Am J Med Genet 80:510-513. 66. Diab M, W JJ, Shapiro F, Eyre D (1994) Abnormality of type IX collagen in a patient with diastrophic dysplasia. Am J Med Genet 49:402-409. 67. Dunbar JD, Sussman MD, Aiona MD (1995) Hip pathology in the trichorhinophalangeal syndrome. J Pediatr Orthop 15: 381-385. 68. Ehrenfried A (1915) Multiple cartilaginous exostoses--Hereditary deforming chondrodysplasia. J Am Med Assoc 64: 1642-1646. 69. Ehrenfried A (1917) Hereditary deforming chondrodysplasiam Multiple cartilaginous exostoses. J Am Med Assoc 68: 502-508. 70. Elejalde BR, de Elejalde MM (1986) The prenatal growth of the human body determined by the measurement of bones and organs by ultrasonography. Am J Med Genet 24: 575-598. 71. Eyre DR, Upton MP, Shapiro FD, Wilkinson RH, Vawter GF (1985) Nonexpression of cartilage type II collagen in a case of Langer-Saldino achondrogenesis. Am J Med Genet 39: 52-67. 72. Fairbank HAT (1949) Diaphysial aclasis. J Bone Joint Surg 31B:105-113. 73. Fairbank T (1947) Dysplasia epiphysialis multiplex. Br J Surg 34:225-232. 74. Fairbank T (1951) An Atlas of General Affections of the Skeleton. Edinburgh: E&S Livingstone. 75. Falvo KA, Bullough PG (1973) Osteogenesis imperfecta: A histometric analysis. J Bone Joint Surg 55A:275-286.

864

CHAPTER 9 9 S k e l e t a l D y s p l a s i a s

76. Falvo KA, Root L, Bullough PG (1974) Osteogenesis imperfecta: Clinical evaluation and management. J Bone Joint Surg 56A:783-793. 77. Felix R, Hofstetter W, Cecchini MG (1996) Recent developments in the understanding of the pathophysiology of osteopetrosis. Eur J Endocr 134:143-156. 78. Fernandes RJ, Wilkin DJ, Weis MA, Wilcox WR, Cohn DH, Rimoin DL, Eyre DR (1998) Incorporation of structurally defective type II collagen into cartilage matrix in Kniest chondrodysplasia. Arch Biochem Biophys 355:282-290. 79. Filly RA, Golbus MS, Carey JC, Hall JG (1981) Short-limited dwarfism: Ultrasonographic diagnosis by mensuration of fetal femoral length. Radiology 138:653-656. 80. Finidori G, Rigault P, Padovani JP (1979) Growth disturbance of the forearm in multiple exostoses and Ollier's disease (dyschondroplasia) Chir Pediatr 20:143-148. 81. Fischer A, Griscelli C, Friedrich W, Kubanek B. Levinsky R, Morgan G, et al. (1986) Bone marrow transplantation for immunodeficiencies and osteopetrosis, European survey 19681985. Lancet ii: 1080. 82. Fogel GR, McElfresh EC, Peterson HA, Wicklund PT (1984) Management of deformities of the forearm in children with multiple hereditary osteochondromas. J Bone Joint Surg 66A: 670-680. 83. Franco B, Meroni G, Parenti G, Levilliers J, Bernard L, Gebbia M, Cox L, Maroteux P, Sheffield L, Rappold GA, Andria G, Petit C, Ballablo A (1995) A cluster of sulfatase genes on Xp22.3: Mutations in chondrodysplasia punctata (CDPX) and implications for warfarin embryopathy. Cell 81:15-25. 84. Frezal J, LeMerrer M, Chauvet ML (1997) Osteochondrodysplasia, dysostoses, disorders of calcium metabolism, congenital malformations with skeletal involvement mapped on human chromosomes. Pediatr Radiol 27:366-387. 85. Gamble JG, Strudwick WJ, Rinsky LA, Bleck EE (1988) Complications of intramedullary rods in osteogenesis imperfecta: Bailey-Dubow rods versus nonelongating rods. J Pediatr Orthop 8:645-649. 86. Gelb BD, Shi GP, Chapman HA, Desnick RJ (1996) Pycnodysostosis, a lysosomal disease caused by cathepsin K deficiency. Science 273:1236-1238 87. Gerritsen EJA, Vossen JM, Fasth A, Friedrich W, Morgan G, Padmos A, et al. (1994) Bone marrow transplantation for autosomal recessive osteopetrosis. J Pediatr 125:896-902. 88. Gibney VP (1876) Hereditary multiple exostosis: Four cases with remarks. Am J Med Sci 72:73-80. 89. Gilbert EF, Yang SS, Langer L, Opitz JM, Roskamp JO, Heidelberger KP (1987) Pathologic changes of osteochondrodysplasia in infancy: A review. Path Ann Part 2 22:283-345. 90. Gilbert-Barnes E, Langer LO, Jr, Glasgow JFT, Nevin NC, Thomas PS (1978) Hypochondroplasia. Arch Dis Child 53: 868-872. 91. Gilbert-Barnes E, Langer LO, Opitz JM, Laxova R, Sotelo C (1996) Kniest dysplasia: radiologic, histopathological, and scanning electron microscopic findings. Am J Med Genet 63:34-45. 92. Giraud-Guille MM (1988) Twisted plywood architecture of collagen fibrils. Calcif Tiss Int 42:167-180. 93. Glasgow JFT, Nevin NC, Thomas PS (1978) Hypochondroplasia. Arch Dis Child 53:868-872. 94. Glorieux FH, Bishop NJ, Plotkin H, Chabot G, Lanoue G, Travers R (1998) Cyclic administration of pamidronate in

children with severe osteogenesis imperfecta. N Engl J Med 339:947-952. 95. Goldberg MJ (1987) The Dysmorphic Child: An Orthopedic Perspective. New York: Raven Press. 96. Gordienko IY, Grechanina EY, Sopko NI, Tarapurova EN, Mikchailets LP (1996) Prenatal diagnosis of osteochondrodysplasias in high risk pregnancy. Am J Med Genet 63:90-97. 97. Gordon SL, Buchanan JR, Ladda RL (1981) Hereditary multiple exostoses: A report of a kindred. J Med Genet 18: 428-430. 98. Grigoriadis AE, Wan ZQ, Wagner EF (1995) Fos and bone cell development: Lessons from a nuclear oncogene. Trends Genet 11:436-441. 99. Hahn YS, Engelhard HH, Naidich T, McLone DG (1989) Paraplegia resulting from thoracolumbar stenosis in a 7month-old achondroplastic dwarf. Pediatr Neurosci 15:39-43. 100. Haines SJ, Erickson DL, Wirtschafter JD (1988) Optic nerve decompression for osteopetrosis in early childhood. Neurosurgery 23:470. 101. Hanscom DA, Bloom BA (1988) The spine in osteogenesis imperfecta. Orthop Clin North Am 19:449-458. 102. Hastbacka J, de la Chapelle A, Kaitila I, Sistonen P, Weaver A, Lander E (1992) Linkage disequilibrium mapping in isolated founder populations: Diastrophic dysplasia in Finland. Nature Genet 2:204-211. 103. Hastbacka J, de la Chapelle A, Mahtani MM, Clines G, Reeve-Daly MP, Daly M, Hamilton B A, Kusumi K, Trivedi B, Weaver A, Coloma A, Lovett M, Buckler A, Kaitila I, Lander ES (1994) The diastrophic dysplasia gene encodes a novel sulfate transporter: Positional cloning by fine-structure linkage disequilibrium mapping. Cell 78:1073-1087. 104. Hastbacka J, Superti-Furga A, Wilcox WR, Rimoin DL, Cohn DH, Lander ES (1996) Sulfate transport in chondrodysplasia. Ann NY Acad Sci 785:131-136. 105. Hecht JT, Butler IJ (1990) Neurologic morbidity associated with achondroplasia. J Child Neurol 5:84-97. 106. Hecht JT, Francomano CA, Horton WA, Annegers JF (1987) Mortality in achondroplasia. Am J Hum Genet 41:454-464. 107. Hecht JT, Horton WA, Reid CS, Pyeritz RE, Chakraborty R (1989) Growth of the foramen magnum in achondroplasia. Am J Med Genet 32:528-535. 108. Hecht JT, Nelson LD, Croder E, Wang Y, Elder FBB, Harrison WR, Francomano CA, et al. (1995) Mutations in exon 17B of cartilage oligometric matrix protein (COMP) cause pseudoachondroplasia. Nature Genet 10:325-329. 109. Heise A (1944) On osteogenesis imperfecta with special reference to the pathologicoanatomical conditions. Acta Pathol Microbiol Scand 21:972-1000. 110. Hensinger RN (1977) Kyphosis secondary to skeletal dysplasias and metabolic disease. Clin Orthop Rel Res 128:113-128. 111. Hensinger RN, MacEwen GD (1982) Congenital anomalies of the spine. In: The Spine, 2nd ed, volume 1, eds. RR Rothman, FA Simeone. pp. 188-315, Philadelphia: WB Saunders. 112. Herring JA (1978) The spinal disorders in diastrophic dwarfism. J Bone Joint Surg 60A: 177-182. 113. Hobbins JC, Bracken MB, Mahoney MJ (1982) Diagnosis of fetal skeletal dysplasias with ultrasound. Am J Obstet Gyn 142:306-312. 114. Hollister DW, Lachman RS (1976) Diastrophic dwarfism. Clin Orthop Rel Res 114:61-69.

References

115. Hopwood JJ, Morris PC (1990) The mucopolysaccharidoses. Diagnosis, molecular genetics and treatment. Mol Biol Med 7:381-404. 116. Horton WA (1987) Achondrogenesis type II, abnormalities of extracellular matrix. Pediatr Res 22:324-329. 117. Horton WA (1993) In vitro chondrogenesis in human chondrodysplasias. Am J Med Genet 45:179-182. 118. Horton WA (1996) Evolution of the bone dysplasia family. Am J Med Genet 63:4-6. 119. Horton WA (1997) Advances in the genetics of human chondrodysplasias. Pediatr Radiol 27:419-421. 120. Horton WA (1997) Molecular genetics of human chondrodysplasias. Growth Genet Horm 13:49-56. 121. Horton WA, Campbell D, Machado MA, Aulthouse AL, Ahmed S, Ellard JT (1989) Tissue and cell studies of the growth plate in the chondrodysplasias. Am J Med Genet 34: 91-95. 122. Horton WA, Campbell D, Machado MA, Chou J (1989) Type II collagen screening in the human chondrodysplasias. Am J Med Genet 34:579-583. 123. Horton WA, Hall JG, Scott CI, Pyeritz RE, Rimoin DL (1982) Growth curves for height for diastrophic dysplasia, spondyloepiphyseal dysplasia congenita, and pseudoachondroplasia. Am J Dis Child 136:316-319. 124. Horton WA, Machado MA, Ellard J, Campbell D, Bartley J, Ramirez F, Vitale E, Lee B (1992) Characterization of a type II collagen gene (COL2A1) mutation identified in cultured chondrocytes from human hypochondrogenesis. Proc Nat Acad Sci USA 89:4583-4587. 125. Horton WA, Rotter JI, Rimoin DL, Scott CI, Hall JG (1978) Standard growth curves for achondroplasia. J Pediatr 93: 435-438. 126. Horton WA (1996) Progress in human chondrodysplasias: Molecular genetics. Ann NY Acad Sci 785:150-159. 127. Hume JB (1929-1930) The causation of multiple exostoses. Br J Surg 17:236-241. 128. Hunter AGW, Bankier A, Rogers JG, Sillence D, Scott CI (1998) Medical complications of achondroplasia: A multicenter patient review. J Med Genet 35:705-712. 129. Hunter AGW, Thompson MW (1976) Acromesomelic dwarfism: Description of a patient and comparison with previously reported cases. Hum Genet 34:107-113. 130. Huvos AG (1991) Multiple osteocartilaginous exostosis (hereditary multiple exostosis, diaphyseal aclasis) In Bone Tumors. Diagnosis, Treatment and Prognosis, 2nd ed. pp. 264-284, Philadelphia: WB Saunders. 131. Hwang SK, Park DM (1991) Induction of osteochondromas by periosteal resection. Orthopedics 14:809-812. 132. Ingram RR (1992) Early diagnosis of multiple epiphyseal dysplasia. J Pediatr Orthop 12:241-244. 133. Irani RN, Petrucelli RC (1993) Ulnar lengthening for negative ulnar variance in hereditary multiple osteochondromas. J Pediatr Orthop 1:143-147. 134. Jabs EW, Muller U, Li X, Ma L, Luo W, Haworth IS, Klisak I, et al. (1993) A mutation in the homeodomain of the human MSX2 gene in a family affected with autosomal dominant craniosynostosis. Cell 75:443-450. 135. Jacenko O, Olsen BR, Warman ML (1994) Invited editorial. Of mice and men: Heritable skeletal disorders. Am J Hum Genet 54:163-168.

865

136. Jaffe HL (1929) The structure of bone. With particular reference to its fibrillar nature and the relation of function to internal architecture. Arch Surg 19:24-52. 137. Jaffe HL (1943) Hereditary multiple exostosis. Arch Path 36: 335-357. 138. Jansen M (1928) Dissociation of bone growth. In: Robert Jones Birthday Volume. London: Humphrey Milford. 139. Jeune M, Larbre F, Carron R, Couette I (1953) La maladie congenitale des epiphyses pointillees ou calcinose fetale epiphysaire chondroystrophiante. Arch Fr Pediatr 10: 914-942. 140. Jones KL (1997) Smith's Recognizable Patterns of Human Malformation, 5th ed. Philadelphia: WB Saunders. 141. Kaibara N, Katsuki I, Hotokebucki T, Takagishi K (1982) Intermediate form of osteopetrosis with recessive inheritance. Skeletal Radiol 9:47. 142. Kaitila I, Korkko J, Marttinen E, Ala-Kokko L (1996) Phenotypic expressions of a Gly154Arg mutation in type II collage in two unrelated patients with spondyloepimetaphyseal dysplasia (SEMD) Am J Med Genet 63:111-122. 143. Kaplan FS, August CS, Fallon MD, Dalinka M, Axel L, Haddad JG (1988) Successful treatment of infantile malignant osteopetrosis by bone marrow transplantation. A case report. J Bone Joint Surg 70A:617. 144. Keith A (1920) Studies on the anatomical changes which accompany certain growth disorders of the human body. 1. The nature of the structural alterations in the disorder known as multiple exostoses. J Anat 54:101-115. 145. Keret D, Bassett GS (1990) Avascular necrosis of the capital femoral epiphysis in metachondromatosis. J Pediatr Orthop 10:658-661. 146. Keret D, Spatz DK, Caro PA, Mason DE (1992) Dysplasia epiphysealis hemimelica: Diagnosis and treatment. J Pediatr Orthop 12:365-372. 147. Key L, Carnes D, Cole, Holtrop M, Shavit-Bar Z, Shapiro F, Arceci R, Steinberg J, Gundberg C, Kahn A, Teitelbaum S, Anast C (1984) Treatment of congenital osteopetrosis with high-dose calcitriol. N Engl J Med 7:409-415. 148. Key LL, Jr, Rodriguiz RM, Willi SM, Wright NM, Hatcher HC, Eyre DR, Cure JK, Griffin PP, Ries WL (1995) Longterm treatment of osteopetrosis with recombinant human interferon gamma. N Engl J Med 332:1594-1599. 149. King JD, Bobechko WP (1971) Osteogenesis imperfecta. An orthopaedic description and surgical review. J Bone Joint Surg 53B:72-89. 150. King RE, Lovejoy JF, Jr (1973) Familial osteopetrosis with coxa vara. J Bone Joint Surg 55A:381. 151. Knaggs RL (1924) Osteogenesis imperfecta. Brit J Surg 737-759. 152. Knaggs RL (1927) Achondroplasia. Br J Surg 15:10-39. 153. Kocher MS, Shapiro F (1998) Osteogenesis Imperfecta. J Am Acad Orthop Surg 6:225-236. 154. Kopits SE (1976) Orthopedic complications of dwarfism. Clin Orthop Rel Res 114:153-179. 155. Kopits SE, Perovic MN, McKusick V, Robinson RA, Bailey JA (1972) Congenital atlantoaxial dislocations in various forms of dwarfism. J Bone Joint Surg 54A:1349-1350. 156. Kuivaniemi H, Tromp G, Prockop DJ (1991) Mutations in collagen genes: Causes of rare and some common diseases in humans. FASEB J 5:2052-2060.

866

CHAPTER 9 9 Skeletal Dysplasias

157. Kurtz AB, Filly RA, Wapner RJ, Golbus MS, Rifkin MR, Callen PW, Pasto ME (1986) In utero analysis of heterozygous achondroplasia: Variable time of onset as detected by femur length measurements. J Ultrasound Med 5:137-140. 158. Kurtz AB, Wapner RJ (1983) Ultrasonographic diagnosis of second-trimester skeletal dysplasias: A prospective analysis in a high-risk population: Research papers. J Ultrasound Med 2:99-106. 159. Lachman RS (1994) Fetal imaging in the skeletal dysplasias: Overview and experience. Pediatr Radio124:413-417. 160. Lachman RS (1997) The cervical spine in the skeletal dysplasias and associated disorders. Pediatr Radio127:402-408. 161. Lamy M, Maroteaux P (1960) Le nanisme diastrophique. La Presse Medicale 52:1977-1980. 162. Langenskiold A (1947) Normal and pathological bone growth in the light of the development of cartilaginous foci in chondrodysplasia. Acta Chir Scand 95:367-386. 163. Langer LO (1965) Diastrophic dwarfism in early infancy. Am J Roent 93:399-404. 164. Langer LO (1965) Dyschondrosteosis, a heritable bone dysplasia with characteristic roentgenographic features. Am J Roent 95:178-188. 165. Laville JM (1994) Knee deformities in Larsen's syndrome. J Pediatr Orthop 3:180-184. 166. Laville JM, Lakermance P, Limouzy F (1994) Larsen's syndrome: Review of the literature and analysis of thirty-eight cases. J Pediatr Orthop 14:63-73. 167. Le Guellec D, Mallein-Gerin F, Treilleux I, Bonaventure J, Peysson P, Herbage D (1994) Localization of the expression of types I, II, and III collagen genes in human normal and hypochondrogenesis cartilage canals. Histochem J 26: 695-704. 168. Lee B, Vissing H, Ramirez F, Rogers D, Rimoin D (1989) Identification of the molecular defect in a family with spondyloepiphyseal dysplasia. Science 244:978-980. 169. Lehman RAW, Reeves JD, Wilson WB, Wesenberg RL (1977) Neurologic complications of infantile osteopetrosis. Ann Neurol 2:378-384. 170. Lenormant C (1905) Les troubles de croissance. Dans les exostoses osteogeniques et ler chondromes des os. Rev Orthop 6:193-233. 171. Leri A, Weill J (1929) Une affection congenitale et symetrique du developpement osseux: La dyschondrosteose. Bull Mem Soc Med Hop Paris 53:1491-1494. 172. Lessoway VA, Schulzer M, Wittmann BK, Gagnon FA, Wilson RD (1998) Ultrasound fetal biometry charts for a North American Caucasian population. J Clin Ultrasound 26:433-453. 173. Leveuf MJ (1946) Le probleme de la croissance de l'os en longueur etudie a la lumiere de la maladie exostosique. Rev Orthop 32:5-14. 174. Li Y, Lacerda DA, Warman ML, Beier DR, Yoshioka H, Ninomiya Y, Oxford JT, Morris NP, Andrikopoulos K, Ramirez F, Wardel BB, Lifferth GD, Teuscher C, Woodward SR, Taylor BA, Seegmiller RE, Olsen BR (1995) A fibrillar collagen gene, COL11A1, is essential for skeletal morphogenesis. Cell 80:423-430. 175. Lipson SJ (1977) Dysplasia of the odontoid process in Morquio's syndrome causing quadriparesis. J Bone Joint Surg 59A:340-344.

176. Looser E (1906) Zur kenntnis der osteogenesis imperfecta congenita und tarda (sogenannte idiopathische osteopsathyrosis) Mitt aus den Grenz der Med Chir 15:160-207. 177. Lorincz AE, Hurst RE, Kolodny EH (1982) The early laboratory diagnosis of mucopolysaccharidoses. Ann Clin Lab Med 12:258-266. 178. Lovett RW (1906) Osteogenesis imperfecta. Brit Med J 2: 915-920. 179. Lutter LD, Langer LO (1977) Neurological symptoms in achondroplastic dwarfsmSurgical treatment. J Bone Joint Surg 59A:87-92. 180. Lutter LD, Longstein JE, Winter RB, Langer LO (1977) Anatomy of the achondroplastic lumbar canal. Clin Orthop Rel Res 126:139-142. 181. MacCallum WG (1915) Chondrodystrophia fetalis. Notes on the pathological changes in four cases. John Hopkins Hosp Bull 26:182-185. 182. Madelung O (1878) Die spontane subluxation der hand nach vorne. Verh Dtsch Gesell Chir 7:259. 183. Madigan R, Worrall T, McClain EJ (1974) Cervical cord compression in hereditary multiple exostosis. Review of the literature and report of a case. J Bone Joint Surg 56A:401-404. 184. Marie P (1900) L' achondroplasie. Dans 1'adolescence et 1'age adulte. Press Med 8:17-23. 185. Marks SC, Jr, Schneider GB (1978) Evidence for a relationship between lymphoid cells and osteoclasts: Bone resorption restored in ia (osteopetrotic) rats by lymphocytes. Monocytes and macrophages from a normal littermate. Am J Anat 152: 331-341. 186. Maroteaux P (1971) La metachondromatose. Z Kinderheilk 109:246-261. 187. Maroteaux P (1998) Bone dysplasias due to abnormalities in type II collagen. Rev Rhum (Engl) 65:221-224. 188. Maroteaux P, Frezal J, Cohen-Solal L, Bonaventure J (1986) Les formes antenatales de l'osteogenese imparfaite. Arch Fr Pediatr 43:235-241. 189. Maroteaux P, Lamy M (1958) La dysostose metaphysaire. Sem Hop Paris 34:1729-1735. 190. Maroteaux P, Lamy M (1964) Achondroplasia in man and animals. Clin Orthop Rel Res 33:91-103. 191. Maroteaux, P (1995) Les Maladies Osseuses De L'Enfant, 3rd ed. Paris: Flammarion. 192. Masada K, Tsuyuguchi Y, Kawai H, Kawabata H, Kawabata H, Noguchi K, Ono K (1989) Operations for forearm deformity caused by multiple osteochondromas. J Bone Joint Surg 71B:24-29. 193. Matsui Y, Yasui N, Kimura T, Tsumaki N, Kawabata H, Ochi T (1998) Genotype phenotype correlation in achondroplasia and hypochondroplasia. J Bone Joint Surg 80B:1052-1056. 194. Maynard JA, Cooper RR, Ponseti IV (1972) A unique rough surfaced endoplasmic reticulum inclusion in pseudoachondroplasia. Lab Invest 26:40-44. 195. Maynard JA, Cooper RR, Ponseti IV (1973) Morquio's disease (Mucopolysaccharidosis Type IV): Ultrastructure of epiphyseal plates. Lab Invest 28:194-205. 196. Maynard JA, Ippolito EG, Ponseti IV, Mickelson MR (1981) Histochemistry and ultrastructure of the growth plate in metaphyseal dysostosis: Further observations on the structure of the cartilage matrix. J Pediatr Orthop 1:161-169.

References 197. Maynard JA, Ippolito EG, Ponseti IV, Mickelson MR (1981) Histochemistry and ultrastructure of the growth plate in achondroplasia. J Bone Joint Surg 63A:969-980. 198. McKusick VA (1972) Heritable Disorders of Connective Tissue, 4th ed. St. Louis: CV Mosby. 199. McKusick VA, Amberger JS, Francomano CA (1996) Progress in medical genetics: Map-based gene discovery and the molecular pathology of skeletal dysplasias. Am J Med Genet 63:98-105. 200. Meredith SC, Simon MA, Laros GS, Jackson MA (1978) Pycnodysostosis: A clinical, pathological, and ultramicroscopic study of a case. J Bone Joint Surg 60A:1122-1130. 201. Micheli LJ, Hall JE, Watts HG (1976) Spinal instability in Larsen's syndrome. Report of three cases. J Bone Joint Surg 58A:562-565. 202. Mikles M, Stanton RP (1997) A review of Morquio syndrome. Am J Orthop August 533-540. 203. Miladi M, Elleuch MH, Sellami S, Douik M (1987) Dysplasie spondylo-epiphysaire tardive avec arthropathies progressives: Apropos de trois observations. Internat Orthop 11: 271-275. 204. Milgram JW (1983) The origins of osteochondromas and enchondromas. Clin Orthop Rel Res 174:264-284. 205. Milgram JW, Jasty M (1982) Osteopetrosis. A morphological study of twenty-one cases. J Bone Joint Surg 64A:912-929. 206. Moorefield WG, Jr, Miller GR (1980) Aftermath of osteogenesis imperfecta: The disease in adulthood. J Bone Joint Surg 62A:113-119. 207. Morquio L (1929) Sur une forme de dystrophie osseuse familiale. Arch Med Enf 32:129-140. 208. Moskowitz N, Carson B, Kopits S, Levitt R, Hart G (1989) Foramen magnum decompression in an infant with homozygous achondroplasia. J Neurosurg 70:126-128. 209. Mouchet A, Belot J (1926) La tarsomegalie. J Radiol Electrol 10:289-293. 210. Muenke M, Schell U (1995) Fibroblast growth factor receptor mutations in human skeletal disorders. Trends Genet 11: 308-313. 211. Muller E (1914) Uber hereditare multiple cartilaginare exostosen und ecchondrosen. Beitr Path Anat 57:232-281. 212. Mundlos S, Otto F, Mundlos C, et al. (1997) Mutations involving the transcription fractor CBFA1 cause cleidocranial dysplasia. Cell 89:773-779. 213. Mundlos S, Chan D, McGill J, Bateman JF (1996) An alpha (II) GLY913 to CYS substitution prevents the matrix incorporation of type II collagen which is replaced with type I and III collagens in cartilage from a patient with hypochondrogenesis. Am J Med Genet 63:129-136. 214. Mundlos S, Olsen BR (1997) Heritable diseases of the skeleton. Part I: Molecular insights into skeletal developmentTranscription factors and signaling pathways. FASEB J 11: 125-132. 215. Mundlos S, Olsen BR (1997) Heritable diseases of the skeleton. Part II: Molecular insights into skeletal developmentMatrix components and their homeostasis. FASEB J 11: 227-233. 216. Murray LW, Bautista J, James PL, Rimoin DL (1989) Type II collagen defects in the chondrodysplasias. 1. Spondyloepiphyseal dysplasias. Am J Hum Genet 45:5-15.

867

217. Nakayama Y, Takeno Y, Tsugu H, Tomonaga M (1994) Maffucci's syndrome associated with intracranial chordoma: Case report. Neurosurgery 34:907-909. 218. Nawata K, Teshima R, Minamizaki T, Yamamoto K (1995) Knee defomaities in multiple hereditary exostoses: A longitudinal radiographic study. Clin Orthop Rel Res 313: 194-199. 219. Nehme AE, Riseborough EJ, Tredwell SJ (1976) Skeletal growth and development of the achondroplastic dwarf. Clin Orthop Rel Res 116:8-23. 220. Nerlich AG, Freisinger P, Bonaventure J (1996) Radiological and histological variants of thanatophoric dysplasia reassociated with common mutations in FGFR-3. Am J Med Genet 63:155-160. 221. Neufeld E, Muenzes J (1995) The mucopolysaccharidoses. In: The Metabolic and Molecular Bases of Inherited Disease, eds C Scriver, A Beaudet, W Sly, D Valle. pp. 2465-2494, New York: McGraw-Hill. 222. Neville AC (1993) Biology of fibrous composites. Development Beyond the Cell Membrane. Cambridge, England: Cambridge Univ. Press. 223. Nogami H, Oohira A, Ozeki K, Oki T, Ogino T, Murachi S (1979) Ultrastructure of cartilage in heritable disorders of connective tissue. Clin Orthop Rel Res 143:251-259. 224. Noyes FR, Kivi LP (1972) Dysplasia epiphysealis hemimelica: A case simulating an intra-articular body. Clin Orthop Rel Res 86:175-177. 225. Ogden JA (1976) Multiple hereditary osteochondromata: Report of an early case. Clin Orthop Rel Res 116:48-60. 226. Oilier M (1898) Exostoses osteogeniques multiples. Lyon Med 88:484-485. 227. Oilier M (1900) Dyschondroplasie. Lyon Med 93:23-24. 228. Oilier M (1900) Sur une nouvelle affection: La dyschondroplasie. Rev De Chir 21:396. 229. Paassilta P, Lohiniva J, Annunen S, et al. (1999) COL9A3. A third locus for multiple epiphyseal dysplasia. Am J Hum Genet 64:1036-1044. 230. Parrot J (1878) Sur la malformation achondroplastique et le Dieu Ptah. Bull Soc Anthrop 1:296-308. 231. Paterson CR, McAllion S, Stellman JL (1984) Osteogenesis imperfecta after the menopause. N Engl J Med 310: 1694-1696. 232. Paterson DC, Morris LL, Binns GF, Kozlowski K (1989) Generalized enchondromatosis: A case report. J Bone Joint Surg 71A:133-140. 233. Pauli RM, Horton VK, Glinski LP, Reiser CA (1995) Prospective assessment of risks for cervicomedullary junction compression in infants with achondroplasia. Am J Hum Genet 56:732-744. 234. Pazzaglia UE, Beluffi G, Castello A, Coci A, Zatti G (1992) Bone changes of mucolipidosis II at different ages: Postmortem study of three cases. Clin Orthop Rel Res 276:283-290. 235. Penrose LS (1955) Parental age and mutation. Lancet ii: 312-313. 236. Pereira RF, Halford KW, O'Hara MD, Leeper DB, Sokolov BP, Pollard MD, Bagasra O, Prockop DJ (1995) Cultured adherent cells from marrow can serve as long-lasting precursor cells for bone, cartilage, and lung in irradiated mice. Proc Natl Acad Sci USA 92:4857-4861.

868

CHAPTER 9 ~ Skeletal Dysplasias

237. Peretti G, Memeo A, Paronzini A, Marzorati S (1995) Staged lengthening in the prevention of dwarfism in achondroplastic children: A preliminary report. J Pediatr Orthop 4: 58-64. 238. Peterson C, Morris LL, Binns GF, Kozlowski K (1989) Generalized enchondromatosis: A case report. J Bone Joint Surg 71A:133-140. 239. Peterson HA (1989) Multiple hereditary osteochondromata. Clin Orthop Rel Res 239:222-230. 240. Peterson HA (1994) Deformities and problems of the forearm in children with multiple hereditary osteochondromata. J Pediatr Orthop 14:92-100. 241. Peterson L, Parkin J, Nelson A (1982) Mucopolysaccharidosis type VII: A morphologic, cytochemical, and ultrastructural study of the blood and bone marrow. Am Soc Clin Path 78:544-548. 242. Pizzutillo PD, Osterkamp JA, Scott CI, Lee MS (1989) Atlantoaxial instability in mucopolysaccharidosis type VII. J Pediatr Orthop 9:76-78. 243. Ponseti IV (1970) Skeletal growth in achondroplasia. J Bone Joint Surg 52A:701-716. 244. Ponseti IV, Pedrini-Mille A, Pedrini V (1968) Histological and chemical analysis of human iliac crest cartilage. Observations on trunk growth. Calcif Tissue Res 2:197-213. 245. Porat S, Heller E, Seidman DS, Meyer S (1991) Functional results of operation in osteogenesis imperfecta: Elongating and nonelongating rods. J Pediatr Orthop 11:200-203. 246. Poussa M, Merikanto J, Ry6ppy S, Marttinen, Kaitila I (1991) The spine in diastrophic dysplasia. Spine 16:881-887. 247. Poznanski AK (1994) Punctate epiphyses: A radiological sign not a disease. Pediatr Radio124:418-424. 248. Prinster C, Carrera P, DelMaschio M, Weber G, Maghnie M, Vigone MC, Mora S, Tonini G, Rigon F, Beluffi G, Severi F, Chiumello G, Ferrari M (1998) Comparison of clinical radiological and molecular findings in hypochondroplasia. Am J Med Genet 75:109-112. 249. Prockop DJ (1997) Marrow stromal cells as stem cells for non-hematopoietic tissues. Science 276:71-74. 250. Prockop DJ, Kuivaniemi H, Tromp G (1994) Molecular basis of osteogenesis imperfecta and related disorders of bone. Clin Plast Surg 21:407-413 251. Raghunath M, Bruckner P, Steinmann B (1994) Delayed triple helix formation of mutant collagen from patients with osteogenesis imperfecta. J Mol Biol 236:940-949. 252. Rao SB, Roy DR (1994) Dysplasia epiphysealis hemimelica: Upper limb involvement with associated osteochondroma. Clin Orthop Rel Res 307:103-109. 253. Rappaport JM (1987) Bone marrow transplantation. In: Hematology of Infancy and Childhood, 3rd ed, volume 1, eds DG Nathan, FA Oski. p. 255, Philadelphia: WB Saunders. 254. Rasmussen SA, Bieber FR, Benacerraf BR, Lachman RS, Rimoin DL, Holmes LB (1996) Epidemiology of osteochondrodysplasias: Changing trends due to advances in prenatal diagnosis. Am J Med Genet 61:49-58. 255. Ray S, Bowen JR (1984) Orthopedic problems associated with survival in campomelic dysplasia. Clin Orthop Rel Res 185:77-82. 256. Redl G (1998) Anesthesia in handicapped children. Anaes (Suppl 2) 53:78-80.

257. Reeder MM (1993) Gamuts in bone, joint and spine radiology. Comprehensive Lists of Roentgen Differential Diagnosis. New York: Springer-Verlag. 258. Reeves JD, August CS, Humbert JR, Weston WL (1979) Host defense in infantile osteopetrosis. Pediatrics 64:202-206. 259. Ribbing S (1937) Studien uber hereditare, multiple epiphysenstorungen. Acta Radiol Suppl 34:77-107. 260. Rimoin DL (1995) Cervicomedullary junction compression in infants with achondroplasia: When to perform neurosurgical decompression. Am J Hum Genet 56:824-827. 261. Rimoin DL (1996) Molecular defects in the chondrodysplasias. Am J Med Genet 63:106-110. 262. Rimoin DL, et al. (1998) International nomenclature and classification of the osteochondrodysplasias (1997) Am J Med Genet 79:376-382. 263. Rimoin DL, Hollister DW, Lachman RS, Kaufman RL, McAlister WH, Rosenthal RE, Hughes GNF (1974) Histologic studies in the chondrodystrophies. Birth Defects: Original Article Series 10:274-295. 264. Rimoin DL, Hughes GN, Kaufman RL, Rosenthal RE, McAlister WH, Silberberg R (1970) Endochondral ossification in achondroplastic dwarfism. N Engl J Med 283:728-735. 265. Rimoin DL, Silberberg R, Hollister DW (1976) Chondro-osseous pathology in the chondrodystrophies. Clin Orthop Rel Res114:137-152 266. Rodgers WB, Hall JE (1993) One-bone forearm as salvage procedure for recalcitrant forearm deformity in hereditary multiple exostoses. J Pediatr Orthop 13:587-591. 267. Romero R, Pilu G, Jeanty P, Ghidini A, Hobbins JC (1988) Prenatal Diagnosis of Congenital Anomalies. Norwalk, CT: Appleton & Lange. 268. Roth SI, Jimenez JF, Husted S, Seibert JJ, Haynes DW (1982) The histopathology of camptomelia (Bent Limbs): A dyschondrogenesis. Clin Orthop Rel Res 167:152-159. 269. Rousseau F, Bonaventure J, Legeai-Mallet L, et al. (1994) Mutations in the gene encoding fibroblast growth factor receptor-3 in achondroplasia. Nature 371:252-253. 270. Rousseau F, Bonaventure J, Legeai-Mallet L, Schmidt H, Weissenbach J, Maroteaux P, Munnich A, LeMerrer M (1996) Clinical and genetic heterogeneity of hypochondroplasia. J Med Genet 33:749-752. 271. Rubin P (1964) Dynamic Classifications of Bone Dysplasias. Chicago: Yearbook Medical Publishers. 272. Ryken TC, Menezes AH (1994) Cervicomedullary compression in achondroplasia. J Neurosurg 81:43-48. 273. Ry6ppy S, Poussa M, Juhani M, Marttinen E, Kaitila I (1992) Foot deformities in diastrophic dysplasia. J Bone Joint Surg 74B: 441-444. 274. Sawin PD, Menezes AH (1997) Basilar invagination in osteogenesis imperfecta and related osteochondrodysplasias: Medical and surgical management. J Neurosurg 86:950-960. 275. Schantz K, Anderson JRPE, Justesen P (1988) Spondyloepiphyseal dysplasia tarda: Report of a family with autosomal dominant transmission. Acta Orthop Scand 6:716-719. 276. Schipani E, Langman CB, Parfitt AM, Jensen GS, Kikuchi S, Kooh SW, Kooh SW, Cole WG, Juppner H (1996) Constitutively activated receptors for parathyroid hormone and parathyroid hormone-related peptide in Jansen's metaphyseal chondrodysplasia. N Engl J Med 335:708-714.

References 277. Scott CI, Jr (1988) Dwarfism. Clinical Symposia, Ciba-Geigy 40:1-32. 278. Seedorff KS (1949) Osteogenesis imperfecta. A study of clinical features and heredity based on 55 Danish families comprising 180 affected members. Universitetsforlaget I Arhus. 279. Shapiro F (1982) Ollier's disease. An assessment of angular deformity, shortening, and pathological fracture in twentyone patients. J Bone Joint Surg 64A:95-103. 280. Shapiro F (1985) Consequences of an osteogenesis imperfecta diagnosis for survival and ambulation. J Pediatr Orthop 5:456-462. 281. Shapiro F (1992) Light and electron microscopic abnormalities in diastrophic dysplasia growth cartilage. Calcif Tissue Int 51:324-331. 282. Shapiro F (1993) Osteopetrosis. Current clinical considerations. Clin Orthop Rel Res 294:34-43. 283. Shapiro F (1998) Structural abnormalities of the epiphyses in skeletal dysplasias. Skeletal Growth and Development: Clinical Issues and Basic Science Advances, eds JA Buckwalter, MG Ehrlich, LJ Sandell, SB Trippel. pp. 471-490. Rosemont, IL: American Academy of Orthopedic Surgeons. 284. Shapiro F, Ellis R (1977) Experimental induction of osteochondromas. Orthop Trans 1:51 (abstract). 285. Shapiro F, Glimcher M, Holtrop ME, Tashjian AH, BrickleyParsons D, Kenzora JE (1980) Human osteopetrosis. A histological, ultrastructural, and biochemical study. J Bone Joint Surg 62A:384-400. 286. Shapiro F, Key LL, Anast C (1988) Variable osteoclast appearance in human infantile osteopetrosis. Calcif Tissue Int 43:67-76. 287. Shapiro F, Simon SH, Glimcher MJ (1979) Hereditary multiple exostoses. Anthropometric, roentgenographic, and clinical aspects. J Bone Joint Surg 61A:815-824. 288. Sharony R, Browne C, Lachman RS, Rimoin DL (1993) Prenatal diagnosis of the skeletal dysplasia. Am J Obstet Gyn 169:668-675. 289. Shears DJ, Vassaal HJ, Goodman FR, Palmer RW, Reardon W, Superti-Furga A, Scambler PJ, Winter RM (1998) Mutation and deletion of the pseudoautosomal gene S H O X cause Leri-Weill dyschondrosteosis. Nature Genet 19:70-73. 290. Shiang R, Thompson LM, Zhu YZ, Church DM, Fielder TJ, Bocian M, Winokur ST, Wasmuth JJ (1994) Mutations in the transmembrane domain of FGFR3 cause the most common genetic form of dwarfism, achondroplasia. Cell 78: 335-342. 291. Siffert RS, Levy RN (1965) Correction of wrist deformity in diaphyseal aclasis by stapling. J Bone Joint Surg 47A: 1378-1380. 292. Sillence D (1981) Osteogenesis imperfecta: An expanding panorama of variants. Clin Ortho Rel Res 159:11-25. 293. Sillence D, Worthington S, Dixon J, Osborn R, Kozlowski K (1997) Atelosteogenesis syndromes: A review, with comments on their pathogenesis. Pediatr Radio127:388-396. 294. Sillence DO, Horton WA, Rimoin DL (1979) Morphologic studies in the skeletal dysplasias. Am J Path 96:813-859. 295. Sillence DO, Rimoin DL, Lachman R (1978) Neonatal dwarfism. Pediatr Clin North Am 25:453-483. 296. Sillence DO, Senn A, Danks DM (1979) Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16:101-116.

869

297. Silveri CP, Kaplan FS, Fallon MD, Bayever E, August CS (1991) Hurler syndrome with special reference to histologic abnormalities of the growth plate. Clin Orthop Rel Res 269: 305-311. 298. Sly WS, Whyte ME Sundaram V (1985) Carbonic anhydrase II defciency in 12 families with the autosomal recessive syndrome of osteopetrosis with renal tubular acidosis and cerebral calcification. N Engl J Med 313:139. 299. Smars G (1961) Osteogenesis imperfecta in Sweden. Svenska Bokforlaget Stockholm. 300. Smith JW (1960) Collagen fiber patterns in lamellar bone. J Anat 94:329-344. 301. Snearly WN, Peterson HA (1989) Management of ankle deformities in multiple hereditary osteochondromata. J Pediatr Orthop 9:427-432. 302. Sofield HA, Millar EA (1959) Fragmentation, realignment and intramedullary rod fixation of deformities of the long bones in children: A ten year appraisal. J Bone Joint Surg 41A:1371-1391. 303. Solh H, et al. (1995) Bone marrow transplantation for infantile malignant osteopetrosis. J Pediatr Hematol Oncol 17: 350-355. 304. Solomon L (1974) Chondrosarcoma in hereditary multiple exostosis. SA Med J 48:671-674. 305. Solomon L (1963) Hereditary multiple exostosis. J Bone Joint Surg 45B:292-304. 306. Specht EE, Daentl DL (1975) Hypochondroplasia. Clin Orthop Rel Res 110:249-255. 307. Spranger J (1988) Bone dysplasia "families." Pathol Immunopathol Res 7:76-80. 308. Spranger J, Cremin B, Beighton P (1982) Osteogenesis imperfecta congenita. Features and prognosis of a heterogeneous condition. Pediatr Radiol 12:21-27. 309. Spranger J, Menger H, Mundlos S, Winterpacht A, Zabel B (1994) Kniest dysplasia is caused by dominant collagen II (COL2A1) mutations: Parental somatic mosaicism manifesting as stickler phenotype and mild spondyloepiphyseal dysplasia. Pediatr Radio124:431-435. 310. Spranger J, Wiedemann HR (1996) Dysplasia spondyloepiphysaria congenita. Helv Paed Acta 21:598-611. 311. Spranger J, Maroteaux P (1990) The lethal osteochondrodysplasias. Adv Hum Genet 19:1-103. 312. Spranger JW, Langer LO (1970) Spondyloepiphyseal dysplasia congenita. Radiology 94:313-322. 313. Spranger JW, Langer LO, Wiedemann HR (1974) Bone Dysplasias: An Atlas Of Constitutional Disorders of Skeletal Development. Philadelphia: WB Saunders. 314. Stanescu R, Stanescu V, Bordat C, Maroteaux P (1980) La dysplasie spondylo-epiphysaire congenitale et son heterogeneite. Arch Fr Pediatr 37:527-530. 315. Stanescu R, Stanescu V, Maroteaux P (1990) Homozygous achondroplasia: Morphologic and biochemical study of the cartilage. Am J Med Genet 37:412-421. 316. Stanescu R, Stanescu V, Muriel M, Maroteaux P (1993) Multiple epiphyseal dysplasia, Fairbank type: Morphologic and biochemical study of cartilage. Am J Med Genet 45:501-507. 317. Stanescu V, Stanescu R, Maroteaux P (1984) Pathogenic mechanisms in osteochondrodysplasias. J Bone Joint Surg 66A:817-836.

870

CHAPTER 9 ~ Skeletal Dysplasias

318. Stanton RE Hansen MO (1996) Function of the upper extremities in hereditary multiple exostoses. J Bone Joint Surg 78A: 568-573. 319. Steinbok P, Hall J, Flodmark O (1989) Hydrocephalus in achondroplasia: The possible role of intracranial venous hypertension. J Neurosurg 71:42-48. 320. Stelling FH (1973) The hip in heritable conditions of connective tissue. Clin Orthop Rel Res 90:33-49. 321. Stevens JM, Kendall BE, Crockard HA, Ransford A (1991) The odontoid process in Morquio-Brailsford's disease. The effects of occipitocervical fusion. J Bone Joint Surg 73B: 851-858. 322. Stover CN, Hayes JT, Hold JF (1963) Diastrophic dwarfism. Am J Roent 89:914-922. 323. Superti-Furga A, Hastbacka J, Rossi A, Van Der Harten JJ, Wilcox WR, Cohn DH, Rimoin DL, Steinmann B, Lander ES, Gitzelmann R (1996) A family of chondrodysplasias caused by mutations in the diastrophic dysplasia sulfate transporter gene and associated with impaired sulfation of proteoglycans. Ann NY Acad Sci 785:195-201. 324. Superti-Furga A, Jastbacka J, Wilcox WR, Cohn DH, et al. (1996) Achondrogenesis type IB is caused by mutations in the diastrophic dysplasia sulfate transporter gene. Nature Genet 12:100-102. 325. Superti-Furga A, Rossi A, Steinmann B, Gitzelmann R (1996) A chondrodysplasia family produced by mutations in the diastrophic dysplasia sulfate transporter gene: Genotype/ phenotype correlations. Am J Med Genet 63:144-147. 326. Svensson O, Aaro S (1988) Cervical instability in skeletal dysplasia: A report of 6 surgically fused cases. Acta Orthop Scand 59:66-70. 327. Sztrolovics R, Glorieux FH, Travers R, Van Der Rest M, Roughley PJ (1994) Osteogenesis imperfecta: Comparison of molecular defects with bone histological changes. Bone 15: 321-328. 328. Taniguchi K (1995) A practical classification system for multiple cartilaginous exostosis in children. J Pediatr Orthop 15: 585-591. 329. Tasker RC, Dundas I, Laverty A, Fletcher M, Lane R, Stocks J (1998) Distinct patterns of respiratory difficulty in young children with achondroplasia: A clinical sleep and lung function study. Arch Dis Child 79:99-108. 330. Tavormina PL, Shiang R, Thompson LM, Zhu YZ, et al. (1995) Thanatophoric dysplasia (types I and II) caused by distinct mutations in fibroblast growth factor receptor 3. Nature Genet 9:321-328. 331. Taybi H (1963) Diastrophic dwarfism. Radiology 80:1-10. 332. Tenni R, Valli M, Rossi A, Cetta G (1993) Possible role of over glycosylation in the type I collagen triple helical domain in the molecular pathogenesis of osteogenesis imperfecta. Am J Med Genet 45:252-256. 333. Thompson JN, Schaefer GB, Conley MC, Mascie-Taylor CGN (1986) Achondroplasia and parental age. N Engl J Med 314:521-522. 334. Thurston MN, Johnson DR, Kember NF (1985) Cell kinetics of growth cartilage of achondroplastic (cn) mice. J Anat 140: 425-434. 335. Tiley F, Albright JA (1973) Osteogenesis imperfecta: Treatment by multiple osteotomy and intramedullary rod insertion. J Bone Joint Surg 55A:701-713.

336. Tiller GE, Polumbo PA, Weis MA, Bogaert R, Lachman RS, Cohn DH, Rimoin DL, Eyre DR (1995) Dominant mutations in the type II collagen gene, COL2A1, produce spondyloepimetaphyseal dysplasia, Strudwick type. Nature Genet 11:87-89. 337. Tiller GE, Weis MA, Polumbo PA, Gruber HE, Rimoin DL, Cohn DH, Eyre DR (1995) An RNA-splicing mutation (G +51vs2~ in the type II collagen gene (COL2AI) in a family with spondyloepiphyseal dysplasia congenita. Am J Hum Genet 56:388-395. 338. Tolo VT (1997) Spinal deformity in skeletal dysplasia conditions. In: The Textbook of Spinal Surgery, 2nd ed, chapter 27, KH Bridwell, RL DeWald. pp. 307-320, Philadelphia: Lippincott-Raven. 339. Tretter AE, Saunders RC, Meyers CM, Dungan JS, Grumbach K, Sun CCJ, Campbell AB, Wulfsberg EA (1998) Antenatal diagnosis of lethal skeletal dysplasias. Am J Med Genet 75: 518-522. 340. Trevor D (1950) Tarso-epiphyseal aclasia: A congenital error of epiphyseal development. J Bone Joint Surg 32B:204-213. 341. Uematsu S, Wang H, Kopits SE, Hurko O (1994) Total craniospinal decompression in achondroplastic stenosis. Neurosurgery 35:250-258. 342. Vaara P, Peltonen J, Poussa M, Merikanto J, Nurminen M, Kaitila I, Ry6ppy S (1998) Development of the hip in diastrophic dysplasia. J Bone Joint Surg 80B:315-320. 343. Van der Harten HJ, Brons JTJ, Dijkstra PF, Meijer CJLM, Van Giejn HE Arts NFT, Niermeyer MF (1988) Perinatal lethal osteogenesis imperfecta: Radiological and pathological evaluation of seven prenatally diagnosed cases. Pediatr Pathol 8:233-252. 344. Van der Harten HJ, Brons JTJ, Dijkstra PF, Niermeyer MF, Meijer CJLM, Van Giejn HP, Arts NFT (1988) Achondrogenesis-hypochondrogenesis: The spectrum of chondrogenesis imperfecta: A radiological, ultrasonographic, and histopathologic study of 23 cases. Pediatr Pathol 8: 571-597. 345. Van Hul W, Bollerslev J, Gram J, Van Hul, Wuyts W, Benichou O, Vanhoenacker F, Willems PJ (1997) Localization of a gene for autosomal dominant osteopetrosis (AlbersSch6nberg disease) to chromosome l p21. Am J Hum Genet 61:363-369. 346. Van Mourik JBA, Hamel BCJ, Mariman ECM (1998) A large family with multiple epiphyseal dysplasia linked to COL9A2 gene. Am J Med Genet 77:234-240. 347. Vikkula M, Mariman ECM, Lui VCH, Zhidkova NI, Tiller GE, Goldring MB, Van Beersum SEC, Waal Malefijt MC, Van Den Hoogen FHJ, Ropers HH, Mayne R, Cheah KSE, Olsen BR, Mayne R, Cheah KSE, Olsen BR, Warman ML, Brunner HG (1995) Autosomal dominant and recessive osteochondrodysplasias associated with the COL11A2 locus. Cell 80:431-437. 348. Virchow R (1891) Uber multiple exostosen, mit vorlegung von praparaten. Berl Klin Woch 28:1082. 349. Walker BA, Scott CI, Hall JG, Murdoch JL, McKusick VA (1972) Diastrophic dwarfism. Medicine 51:41-59. 350. Walker DG (1973) Osteopetrosis cured by temporary parabiosis. Science 180:875. 351. Walker DG (1975) Bone resorption restored in osteopetrotic mice by transplants of normal bone marrow and spleen cells. Science 190:784-785.

References 352. Warman ML, Abbott M, Apte SS, Hefferon T, Mclntosh, Cohn DH, Hecht JT, Olsen BR, Francomano CA (1993) A type X collagen mutation causes Schmid metaphyseal chondrodysplasia. Nature Genet 5:79-82. 353. Wasylenko MJ, Wedge JH, Houston CS (1980) Metaphyseal chondrodysplasia, Schmid type. J Bone Joint Surg 62A: 660-664. 354. Webster MK, Donoghue DJ (1997) FGFR activation in skeletal disorders: Too much of a good thing. Trends Genet 13: 178-182. 355. Weidenreich F (1930) Das Knochengewebe. In: Handbuch der Mikroskopischen Anatomie des Menschen, ed W von Mollendorff, Die Gewebe, Teil 2, Sturtzgewebe Knochengewebe Skeletsystem. pp. 394-406, Berlin: Springer. 356. Weiner DS, Hoyt WA (1978) The development of the upper end of the femur in multiple hereditary exostosis. Clin Orthop Rel Res 137:187-190. 357. Weis MA, Wilkin DJ, Kim HJ, Wilcox WR, Lachman RS, Rimoin DL, Cohn DH, Eyre DR (1998) Structurally abnormal type II collagen in a severe form of Kneist dysplasia caused by an exon 24 skipping mutation. J Biol Chem 273:4761-4768. 358. Wenger DR, Birch J, Rathjen K, Tobin R, Billman G (1991) Metachondromatosis and avascular necrosis of the femoral head: A radiographic and histologic correlation. J Pediatr Orthop 11:294-300. 359. Wenstrup RJ, Lever LW, Phillips CL, Quarles LD (1993) Mutations in the C O L I A 2 gene of type I collagen that result in nonlethal forms of osteogenesis imperfecta. Am J Med Genet 45:228-232. 360. Wenstrup RJ, Willing MC, Starman BJ, Byers PH (1990) Distinct biochemical phenotypes predict clinical severity in nonlethal variants of osteogenesis imperfecta. Am J Hum Genet 46:975-982. 361. Whitley CB (1993) The Mucopolysaccharidoses. In: McKusick's Heritable Disorders of Connective Tissue, 5th ed, ed P Beighton. pp. 367-499, St. Louis: Mosby. 362. Wilkin DJ, Szabo JK, Cameron R, Henderson S, Bellus GA, Mack ML, Kaitila I, Loughlin J, Munnich A, Sykes B, Bonaventure J, Francomano CA (1998) Mutations in fibroblast growth-factor receptor 3 in sporadic cases of achondroplasia occur exclusively on the paternally derived chromosome. Am J Hum Genet 63:711-716.

871

363. Wilkinson JM, Scott BW, Clarke AM, Bell MJ (1988) Surgical stabilisation of the lower limb in osteogenesis imperfecta using the Sheffield telescopic intramedullary rod system. J Bone Joint Surg 80B:999-1004. 364. Williams BR, Cranley RE (1974) Morphologic observations on four cases of SED congenita. Birth Defects: Original Article Series 10:75-87. 365. Willing MC, Deschenes SP, Scott DA, Byers PH, Slayton RL, Pitts SH, Arikat H, Roberts EJ (1994) Osteogenesis imperfecta type 1: Molecular heterogeneity for COL1A1 null alleles of type lcollagen. Am J Hum Genet 55:638-647. 366. Willing MC, Pruchno CJ, Byers PH (1993) Molecular heterogeneity in osteogenesis imperfecta type I. Am J Med Genet 45:223-227. 367. Winter RB, Bloom BA (1990) Spine deformity in spondyloepimetaphyseal dysplasia. J Pediatr Orthop 10:535-539. 368. Winter RM (1995) Recent molecular advances in dysmorphology. Hum Mol Genet 4:1699-1704. 369. Winterpacht A, Superti-Furga A, Schwarze U, Stoss H, Steinmann B, Spranger J, Zabel B (1996) The deletion of six amino acids at the C-terminus of the alpha (II) chain causes overmodification of type II and type XI collagen: Further evidence for the association between small deletions in COL2A1 and Kniest dysplasia. J Med Genet 33:649-654. 370. Wynne-Davies R, et al. (1989) Instability of the upper cervical spine: Skeletal dysplasia group. Arch Dis Child 64: 283-288. 371. Wynne-Davies R, Gormley J (1985) The prevalence of skeletal dysplasias. An estimate of their minimum frequency and the number of patients requiting orthopaedic care. J Bone Joint Surg 67B:133-137. 372. Wynne-Davies R, Hall CM, Young ID (1986) Pseudoachondroplasia: Clinical diagnosis at different ages and comparison of autosomal dominant and recessive types. A review of 32 patients (26 kindreds) J Med Genet 23:425-434. 373. Yang SS, Kitchen E, Gilbert EF, Rimoin DL (1986) Histopathologic examination in osteochondrodysplasia. Arch Pathol Lab Med 110:10-12. 374. Young ID (1998) Achondroplasia: A case of neglect? Lancet 352:1950-1951.

CUArTER 1 0

Epiphyseal Involvement with Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders I. II. III.

Rickets

IV. Osteomyelitis and Septic Arthritis

Inflammatory Disorders

V. Hematologic Disorders

Neoplastic Disorders of Epiphyses

The epiphyses can be affected by a variety of acquired disorders either (1) primarily, (2) secondarily following original metaphyseal involvement, or (3) in association with systemic involvement. Epiphyseal involvement in the growing skeleton can superimpose skeletal growth sequelae onto the primary pathological process. The pathobiology and principles of orthopedic management of the major disorders in this category will be reviewed.

appropriate bone deposition in the metaphyseal area. This in turn leads to stunting of growth and bowing of the long bones through the abnormally softened physeal and metaphyseal regions. Jaffe defined a triad of changes in the endochondral bone sequence with rickets, involving (1) deficient calcification of the matrix of the proliferating cartilage, (2) an abnormal accumulation of cartilage (thickened growth plates), and (3) deposition of unmineralized osteoid. The longitudinal arrangement of the physeal cartilage cells is lost, vascular invasion is limited and poorly oriented, and the undermineralized chondroid and osteoid tissues mingle in disorganized fashion at the physeal-metaphyseal junction.

I. RICKETS A. Terminology

C. Experimental Models of Rickets

Rickets is a disorder of childhood characterized by the persistence of variable amounts of undermineralized or nonmineralized endochondral cartilage and osteoid. There are many causes of undermineralization, but the common final pathway during the growth years leads to a series of similar clinical, radiographic, and histopathological sequelae. Skeletal deformities are due to mechanical stresses on the undermineralized, weakened bones.

Once the biochemical defects in rickets became known, it was relatively easy to produce experimental models both to study the structural changes in growth plate cartilage and cortical and trabecular bone and to note the effects of varying biochemical and hormonal permutations on the disease process. 1. DODDS AND CAMERON

An excellent early study by Dodds and Cameron assessed the structure of the proximal tibial growth plate in rats made rachitic by dietary changes primarily involving a diet low in vitamin D (48). They noted the high degree of similarity between the histologic changes in rickets in human and rat bones. Description of normal growth plate structure indicates that the epiphyseal cartilage is composed of five zones: zone of reserve cells, zone of flattened cells and cell multiplication, zone of cell growth in which mitosis or cell proliferation has stopped but the cells themselves increase in size, zone of fully grown cells comprising the end of the rows at which the cells have attained their full size, and zone of cartilage removal in which marrow advancement, resorption of cartilage matrix, and deposition of bone on calcified cartilage cores occur. (The term hypertrophic cell, however, was

B. Pathogenesis of Rickets Mineralization of cartilage in the hypertrophic zone region of the growth plate is an integral part of the endochondral ossification process. It is stimulated by the presence and concentration of certain molecular constituents of the cartilage and by normal concentrations of circulating calcium and phosphorus. Disturbances of endochondral mineralization lead to deformities, particularly in growing bones, because mineralization of the lower portion of the growth plate is an integral part of the endochondral sequence, which underlies longitudinal bone growth. Poorly mineralized cartilage in the hypertrophic cell region of the growth plate prevents appropriate signaling to the advancing vascular and osteoprogenitor front from the metaphysis, which then leads to failure of 872

SECTION

not used to describe the fourth zone, which also was referred to as the zone of calcification because one of its outstanding characteristics was "the deposition of calcium salts in the longitudinal walls between columns of cells but not in the transverse walls between cells within the columns.") Dodds and Cameron recognized that "narrow tongues of marrow advance along the rows of cartilage cells, removing the uncalcified transverse wall and liberating the cells from their lacunae." The calcified portions of the matrix served a double purpose, guiding the advancing marrow and forming a base upon which the bone was deposited. In the zone of calcification, calcification advances steadily into the lower regions of the growth plate, keeping three or four cells ahead of the advancing marrow that removed the cartilage. In normal physeal cartilage there are five processes that maintain a constant relation to each other, these being cell multiplication, cell growth, calcification of the matrix, cartilage removal, and deposition of bone. The physeal cartilage maintains a constant thickness during growth, and the zones within it maintain nearly the same relative proportions. In rickets, one of the major changes is the thickening of the physeal cartilage, which can be noted both radiologically and histologically. When histologic samples of rachitic rats are examined, the increased thickness does not occur uniformly throughout all five levels, but is concentrated in zones 4 and 5. The zones of reserve cells, flattened cells and cell multiplication, and cell growth are normal. Major alterations, however, occur in zones 4 and 5, which are the zone of fully grown cells (the hypertrophic cell zone) and the zone of cartilage removal. The histologic preparations show marked thickening of the zone of fully grown cells, which in some instances is as much as 12 times its normal thickness. This is the region that normally would have been calcified, and due to the failure of calcification, cell and matrix resorption patterns are altered markedly. Cell multiplication and growth, however, continue above. The thickening of the cartilage, therefore, is due not to accelerated proliferation but due to its retarded mineralization and removal. There is a thickening of the zone of enlarged cells (zone of cell hypertrophy), which consists of uncalcified or poorly calcified cartilage. When a certain degree of thickening of the hypertrophic zone occurs, mechanical effects worsen the structural appearance with crushing or flattening of the cell columns seen followed by angulation or lateral deflection of the cartilage cell and matrix columns. This occurs adjacent to the metaphyseal bone and is often so marked that the usual linear alignment of the physeal tissues is positioned at fight angles to the longitudinal axis of the bone. This arrangement further confuses the once orderly vascular invasion from below. Comment is also made on the cupping or concave characteristic of the metaphyses at the ends of the bones in human rickets, an appearance also demonstrated in the rat model. In the normal bone the ends of the metaphyses are usually linear or slightly convex. Rather than being exclusively mechanical, as is widely suspected, the cupping defor-

I 9 Rickets

873

mity is due to the fact that physeal growth and calcification at the periphery are less affected than they are centrally, leading to the alteration of physeal shape. In other words, "cupping is produced by a greater retardation of calcification near the axis of the cartilage than near its margin." Eventually, the noncalcified cartilage is removed, but in a slow and irregular fashion. There is more removal at the periphery than centrally, leading to an irregularity in the thickness of the plate and in bone deposition patterns. The cartilage of the physis not only is thickened irregularly but also is invaded and resorbed irregularly. It is observed that "the absence of calcification thoroughly disorganizes the process of cartilage removal. The marrow no longer advances directly along the rows of cells, but instead advances by broad tongues as wide as two, three or more rows of cells." The process is disorganized, leaving isolated masses of cartilage with irregular formation of osteoid upon them. The rachitic type of cartilage removal does not result in normal trabecular bone, but rather a mass of isolated pieces of uncalcified cartilage embedded in osteoid and marrow. The general outlines of the structural changes in rickets are shown in Fig. 1. 2. SHOHL AND WOLBACH Shohl and Wolbach induced rickets in rats by altering their dietary intake of calcium and phosphorus (188). Histologic studies then showed the evolution of the condition at various dose and deprivation levels. The important histologic criteria of rickets result from disturbances in endochondral bone formation and from the failure of bone matrix or osteoid to calcify in all locations. The earliest manifestation of developing rickets is a moderate increase in the thickness of the physeal cartilage with an irregular border on the metaphyseal side and in particular the absence of capillary ingrowth. There is continued activity of the proliferative zone of the physis, but the normal differentiation into the hypertrophic zone does not occur. The initial abnormality is a combination of both the failure of formation of this hypertrophic cell layer and the absence of ingrowth of capillaries. Calcification of the cartilage matrix within the hypertrophic zone ceases, and osteoid accumulates around the capillaries of the metaphysis adjacent to the cartilage but does not undergo mineralization. In more advanced cases, the noncalcifled cartilage at the metaphyseal end of the physis becomes transversely rather than longitudinally oriented due to the mechanical effects of weight bearing on the weakened tissue. The osteoid increases in amount with time, and because it is noncalcified it is also molded in a transverse direction by the pressure of weight beating. With continuation of the tickets, the cancellous bone of the metaphysis actually disappears and there is in addition resorption of the adjacent cortical bone. Mild tickets is characterized by diminished penetration of the physeal cartilage by capillaries, marked irregular thickening of the physeal cartilage, persistence of cartilage cells in the metaphyseal bone, and abnormal stratification of the lower regions of the physeal cartilage. Moderate rickets is

874

CHAPTER 10 ~ Metabolic,

Inflammatory, Neoplastic, Infectious, and Hematoloftic Disorders

FIGURE 1 The general outlines of the structural changes in rickets are illustrated (48). Note the physeal widening concentrated in the middle part of the physis in severe rickets with less involvementat the periphery. This translatesradiographicallyinto cupping of the metaphyseal regions. [Reprinted from Dodds and Cameron (1934), Am. J. Anat. 55:135-165, 9 Lippincott Williams & Wilkins, with permission.] characterized by the almost complete absence of penetration of cartilage cells by capillaries, increased thickening of the physeal cartilage, incorporation of cartilage cell groups into bone trabeculae, which themselves are nonmineralized osteoid, and further disorganization of physeal cartilage. In the most severe variants, there is the complete absence of penetration of cartilage cells by capillaries, further thickening of the physeal cartilage, incorporation of cartilage cells into metaphyseal tissue (rather than its resorption), and large amounts of osteoid at the periphery of the metaphysis. There is failure of physeal cartilage cells to complete their cycle of growth, maturation, and degeneration (by which is meant hypertrophic change) and failure of cartilage and osteoid matrix to calcify. The proliferative activities of the cartilage cells, capillaries, and osteoblasts are not inhibited, but the smooth sequence of changes is disturbed markedly due to the failure of cartilage and osteoid mineralization.

D. Pathology of Human Nutritional Vitamin D Deficiency Rickets Park reviewed the pathology of rickets in a classic monograph (150). In the growing child the findings in the physeal cartilage are most sensitive, but abnormalities of both the endochondral and intramembranous bone sequences are generally seen. The first demonstrably pathological changes occur in the hypertrophic regions of the growth plate with a failure of normal calcium deposition, but newly forming bone of the cortices and trabeculae is also affected. The pri-

mary histopathological finding of the bone tissue is an increase in the thickness of the osteoid layer both within the cortex and surrounding the trabeculae of the marrow. In adults, for whom the disorder is referred to as osteomalacia, the presence of increased osteoid seams is the major histopathological finding because the physes have closed. In those children who are older and grow at a relatively slower rate than the younger child, the intramembranous bone osteoid is also a more prominent finding. Most observers now feel that bone resorption also occurs due to the secondary hyperparathyroidism, although the tendency is not to remove just the mineral but the matrix as well. Park commented on the irregularity of the distribution of osteoid throughout an individual bone and indeed throughout the entire skeleton due to the local mechanical conditions governing bone synthesis and resorption. The more evident changes in rickets in the young child are those at the physeal-metaphyseal junction. These are more marked the younger thechild, the more severe the tickets, and at those physes that experience the most extensive growth. The earliest changes seen involve decreased mineralization of the longitudinal cartilage trabeculae within the hypertrophic zone. With increased severity of this occurrence the adjacent bone formation, which normally occurs on the calcified cores, is also diminished because the osteoid cannot be mineralized and collapse of the weakened matrix leads to isolated and horizontally positioned cartilage-bone trabeculae of the physeal-metaphyseal region. The associated vascular invasion of the area then becomes disturbed

SECTION ! ~ Rickets

and orderly invasion stops. Growth of the cartilage within the proliferating zone continues, leading to thickening of the physis. The irregular vascular invasion leads to further disturbances in the normal endochondral sequence. Wolbach stated that "the first histological evidence of tickets is the absence in whole or in part of the layer of clear (hypertrophic) cells and the consequent absence of ingrowth of capillaries." Park noted that in the human there is generally patchy irregularity of calcification in the physis rather than abnormality in a uniform fashion along a common front. The next series of histopathological changes in tickets involves compression of the softened cartilage trabeculae and the adjacent cartilage-bone trabeculae that occurs with continued weight beating. In the most severe degrees of compression, the lower physeal and metaphyseal matrix lies horizontal rather than parallel to the long axis of the bone. Focal areas of abnormality are seen initially. Both in rat models of tickets and in human samples, flattening and compression of the trabeculae tend to occur centrally but not at the periphery. At the extreme periphery compression is unusual. Collapse at the physeal-metaphyseal (or cartilageshaft) junction is not universal but is dependent on the combination of rapidity of growth, severity of the rachitic process, and local pressure. Thus, it is seen chiefly at the faster growing ends of the long bones in young children. The cornpressed cells and matrices are limited to the lowermost parts of the physis and adjacent metaphysis and offer a formidable barrier to vascular invasion. The crushed tissues centrally prevent vascular invasion and enable the physeal cartilage to remain intact. Radiographically, it is the increased height of the physis in the longitudinal axis of the bone that first characterizes the disorder. Histopathologic studies show that this is due partially to the failure of mineralization of the hypertrophic zone but later is due to an increase in the proliferating zone height because it continues its growth but does not transform into a calcified region. Increased radiolucency is also due to deposition in the adjacent metaphysis of osteoid, which does not become fully mineralized. The proliferative cartilage increases in amount not due to more rapid growth but due to the fact that it accumulates passively. It is not transformed into an effective hypertrophic zone layer due to the abnormality of mineralization and centrally to the prevention of vascular invasion by the crushed and abnormally oriented cartilage-bone tissue interface. The findings are similar in many animal models and in the human. Vascular invasion of the lower margins of the physeal cartilage is deranged in terms of its normal pattern because of the failure of calcification of the cartilage matrix columns. The reason is 2-fold: one relates to the fact that vascularization appears to be triggered at least partially by mineralization of the matrix, whereas the other difficulty relates to the collapse of the cartilage matrix trabeculae such that the vascularization is made more difficult mechanically by the oblique and horizontal position of the cartilage and osteoid

875

seams. The periphery of the physeal cartilage appears to be somewhat protected, continuing relatively normal growth without a tendency to collapse at the lower levels. These descriptions lead to the histologic and radiologic example of cupping of the metaphyseal region, which is only partially a response to mechanical influences. The physis centrally is thicker such that the adjacent metaphyseal bone is concave in shape, more depressed centrally, and rising in a circumferential fashion toward the periphery. Tension thus becomes important, causing the lower part of the thickened physis centrally to give way such that the lateral part is pulled around and upward toward the outer surface of the shaft, leading to the cupping phenomenon. In tickets the normal vascular invasion from below into the hypertrophic zone is altered. For some time the physis thickens with little to no vascular invasion occurring. In severe tickets, however, vascular invasion of a highly pathologic and damaging type eventually breaks through. Rather than having a vessel delve into a single column of hypertrophic chondrocytes, there is a tendency for a relatively wide invasive track to be found involving three or four groupings of cartilage cells. In severe tickets, there is often an arc of dilated capillaries from the metaphyseal marrow attacking simultaneously dozens of cartilage columns. Park clearly points out the irregularity of vascular invasion of the proliferating cartilage in tickets. Compressed cartilage furnishes considerable obstruction to vascular penetration. Even slight rearrangements of the cartilage structure, in particular the linearity of chondrocyte and matrix orientation, make vascular invasion from below more difficult. If the vessels cannot advance in a linear fashion, they tend to disperse laterally, leading to a sinusoidal pattern of entrance in an irregular fashion. Vascular invasion of thickened, abnormally structured physeal cartilage in severe tickets, once established, comes from three sites. The first is from the adjacent metaphyseal bone. A variable number of bushlike vascular formations protrude into the cartilage. Some are thick and progress to the reserve zone. Some are thin and enter only the lower compressed layer. The buds have single individual stems. Branching occurs within the cartilage but adjacent vessels do not anastomose. With increased advancement into the cartilage the vessel systems become long and thin. This late occurring invasion is referred to as "en masse" invasion. Cartilage cells and adjacent matrices are both attacked and no osteoid or bone formation occurs. A tissue destructive effect occurs with the extreme, often sinusoidal dilatation of the invasive vascular tufts. The second vascular invasion is from the cartilage canals of the epiphysis, and the third is from the perichondrial vessels. The methods of invasion are exactly the same regardless of origin from metaphysis, epiphyseal cartilage, or perichondrial tissues. Histological specimens of thickened tickets physes in the severe chronic stage thus show the cartilage to be fiddled with vessels. When treatment begins, calcium is deposited both in the hypertrophic layer of cartilage and in the osteoid of the bone

876

C H A P T E R 10 9

Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

TABLE I S e q u e n c e o f H i s t o p a t h o l o g i c C h a n g e s in Nutritional Vitamin D Deficient Rickets ,

1. Deficient calcification Of cartilage in longitudinal septae of hypertrophic zone Of osteoid in zone of provisional calcification Of cortical and trabecular osteoid 2. Physeal thickening Initially of hypertrophic zone Also of proliferating zone (not converted to hypertrophic zone) Diminished to absent vascular invasion from metaphysis 3. Compression of central lower physeal tissues Physeal thickening concentrated in central four-fifths of physis Compression of lower physeal zone chondrocytes and matrix (mechanical) Oblique-horizontal orientation of cartilage-osteoid matrices Vascular invasion further diminished due to compressed, malaligned matrices 4. Cupping of metaphyseal ends of bones Physis thicker centrally than peripherally (bone formation more prominent at periphery leading to concave appearance) 5. Slowed growth 6. Pathologic vascular invasion of markedly thickened physes From vascular tufts from metaphyseal marrow From cartilage canals of epiphyses From perichondrial tissues (Pathologic invasion via long nonanastomosing vascular tufts; wider than normal; invading and destroying in sinusoidal pattern; not synthesizing bone or osteoid; transphyseal; fiddling thickened cartilage with sinusoidal channels) Repair with vitamin D therapy, calcium and phosphorus availability from stages 2 to 6, depending on time of diagnosis and treatment. Calcification begins initially in most normal, most recently synthesized part of physis and then spreads to older, more damaged regions ,,

trabeculi. The repair sequence with new calcification, however, tends to occur in the most recently synthesized hypertrophic zone, whereas the deeper area that has collapsed tends to repair later. The areas in which vessels are quite wide do not appear to be associated with early remineralization. Table I outlines the progressive histopathological changes in rickets.

E. Classification of Types of Rickets The process of mineralization is mainly guided by the serum concentrations of calcium and phosphate. Most forms of rickets can be classified into one of two types: those primarily associated with decreased serum levels of calcium (calciopenic rickets) and those primarily associated with decreased phosphate levels (phosphopenic rickets). The lack

TABLE II

O u t l i n e o f C a u s e s o f Rickets

I. Hypocalcemic-Calcipenic Rickets 1. Inadequate calcium intake 2. Vitamin D deficiency a. Nutritional vitamin D deficiency b. Malabsorption of vitamin D (intestinal, pancreatic, or hepatic disease) c. Absence of exposure to sunlight 3. Impaired 25-hydroxylation of vitamin D resulting from hepatic disease (biliary atresia, cirrhosis) 4. Impaired 1-hydroxylation of 25-hydroxyvitamin D a. Secondary to renal disease b. Due to autosomal recessive vitamin D dependent tickets, type I (inherited inability to 1-hydroxylate 25-hydroxyvitamin D with otherwise normal renal function) 5. Impaired target organ response to 1,25-dihydroxyvitamin D a. Vitamin D dependent tickets, type II (inherited failure of target organ response due to structural variations of receptors) b. Response impaired by anticonvulsants phenobarbital and dilantin 6. Increased turnover-losses of vitamin D a. Nephrotic syndrome (loss of vitamin D) b. Anticonvulsants (dilantin and phenobarbital) stimulate hepatic clearance of vitamin D metabolites II. Hypophosphatemic-Phosphopenic Rickets 1. Inadequate phosphate intake 2. Deficient renal tubular reabsorption of phosphate a. X-linked hypophosphatemic rickets b. Autosomal dominant and autosomal recessive hypophosphatemic rickets c. Hypophosphatemic rickets with hypocalciuria 3. Generalized disorders of renal tubular transport inducing hypophosphatemia a. Renal tubular acidosis b. Fanconi syndrome (multiple disorders characterized by phosphaturia, glycosuria, and aminoaciduria) 4. Tumor-induced hypophosphatemia a. Soft tissue and fibrous tumors produce factors inhibiting renal tubular reabsorption of phosphate (failure to control renal phosphate excretion also underlies renal osteodystrophy)

of calcium availability results either from a deficiency of calcium intake, which is rare, or more commonly from disorders of vitamin D metabolism or action. Lack of phosphate availability results from either excess phosphate loss or inadequate phosphate intake. Those disorders with diminished calcium generally respond well to vitamin D therapy, whereas effective treatment of phosphopenia requires phosphate replacement as part of the management program. Table II outlines the forms of tickets in these two broad categorizations. Subsequent discussion will concentrate on the three disorders most commonly seen in an orthopedic setting because

SECTION I 9 Rackets

877

Vitamin D deficiency it I l l l l l ! !I

II

!

!! !

!

,,J R e d u c e d Ca

absorption

i!

lilt

II

II I ! t!

Depressed extracellular Ca I|

....

vI~ Sec o l ~ d a l ~

/

||| IIii

'"'A

Depres~racellular

Defective mineralization of the bone in formation

h ~ r ~ e r p a r a ~ y r o l d~t s ~m

P

Increased bone resorption Demineralization of the existing bone

FIGURE 2

The biochemical consequences of a vitamin D deficiency rickets are shown (42). [Reprinted from David (1991), Nestle Nut. Workshop Ser. 21" 108, 9 Lippincott Williams & Wilkins, with permission.]

of the associated bone deformities: nutritional vitamin D deficient rickets, X-linked hypophosphatemic rickets, and renal osteodystrophy, previously referred to as renal rickets. Detailed reviews on calcium, phosphorus, vitamin D, and parathyroid hormone interactions include those of Chesney et al. (35), Cole et al. (41), and Reichel et al. (162). 1. N U T R I T I O N A L R I C K E T S

a. Pathophysiology Nutritional rickets due to a deficiency of vitamin D intake, though still rare, is being seen more frequently in the North American environment in association with breast-feeding and special dietary practices, including macrobiotic and other vegetarian diets. Children raised on unsupplemented breast milk in association with a strict vegetarian diet are prone to a rickets disorder. The vitamin D content of maternal milk is very low and the content in an infant diet in general is also very low. Minimal exposure to sunlight especially in densely populated northern urban environments during the winter also worsens vitamin D levels. The ultraviolet rays of the sun lead to endogenous synthesis of vitamin D, such that children with darkly pigmented skin living in northern environments are prone to develop rickets. For these reasons, the diagnosis of nutritional rickets usually is made in the later winter or early spring. Jacobsen et al. reported on eight patients with nutritional rickets, all of whom were black, who had been breastfed until 6-12 months of age and then were placed on a vegetarian diet (96). The situation can be worsened by other dietary deficiencies, particularly of calcium or phosphorus. The supplementary use of vitamin D following its discovery in the 1920s dramatically diminished the occurrence of tickets. The prevalence of nutritional rickets was extremely high prior to the discovery of vitamin D, with the prominent German pathologist Schmorl reporting an 80% rate of some

degree of rickets on autopsy findings of German children (181). The clinical onset of vitamin D deficient rickets is gradual and generally evolves between 4 and 18 months of age, during the time of greatest growth velocity. Congenital rickets can occur due to maternal vitamin D deficiency, and premature infants are more susceptible to vitamin D deficiency rickets. Vitamin D deficiency leads to reduced calcium absorption from the intestine and decreased circulating calcium levels. This in turn stimulates a secondary hyperparathyroidism. The serum calcium levels can be variable in vitamin D deficient rickets, primarily because of variable compensatory responses. Hypocalcemia is the most important biochemical abnormality, but it is seen in only approximately one-half of patients at the time of diagnosis. In the others the serum calcium levels are normal or only slightly diminished. Serum phosphate (P) levels tend to be low due to diminished tubular phosphate reabsorption and low urinary calcium excretion. Invariably there is a markedly increased serum alkaline phosphatase level due to the osteoblast activation in the calcium deficient environment. Serum parathyroid hormone levels also tend to increase with increasing severity of the rickets. Vitamin D levels tend to be below normal or in the low-normal range. The biochemical consequences of vitamin D deficient rickets are illustrated in Fig. 2. b. Pathogenesis o f Skeletal Abnormalities Decreased circulating and extracellular calcium leads to defective mineralization of newly forming bone, and the secondary hyperparathyroidism worsens this situation as it elevates the serum calcium by increasing bone resorption, which leads to further demineralization of existing bone. Due to the importance of mineralization in the normal development of the physis, major histopathological changes evolve in the epiphyseal, physeal, and metaphyseal regions with this disorder (Table I). There is absence or diminution of mineralization

878

CHAPTER 10 9 Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

in the cartilage matrix of the hypertrophic region, thickening of physeal cartilage, which continues to proliferate but is unable to proceed to a normal hypertrophic sequence in the absence of mineral, and cupping of the metaphyses with the concavity toward the epiphyseal side as the epiphysis is displaced and essentially is embraced peripherally by the undermineralized and poorly developed metaphyseal areas. Rickets, therefore, induces marked changes at the physealmetaphyseal region with considerable disorganization of the endochondral sequence noted. There is an accumulation of physeal cartilage, which loses its orderly columnation pattern. Marrow vessels do not tend to penetrate the hypertrophic cartilage because what would appear to be one of their major signals, mineralization of the cartilage matrix, has not occurred. Where metaphyseal vessels do penetrate, they do so at various angles and to various levels rather than being parallel to the long axis of the bone and well-controlled at the lower levels of mineralization. Irregular accumulation of cartilage persists in the metaphyseal region. The marrow tends to be fibrous rather than vascular. The increased width of the physeal-metaphyseal region is particularly evident clinically at those regions of the developing skeleton that are subcutaneous, including the costochondral junctions and the distal radial and ulnar epiphyses. The secondary centers of ossification are undermineralized and their appearance may be delayed. The skull is soft and misshapen due to the tendency of the child to lie on and flatten the posterior occipital region (rachitic craniotabes). The sutures are wider than normal and there is delayed closure of the fontanelles. There is increased protrusion of the sternum, localized prominence or beading of the chest wall along the lines of the costochondral junctions, and grooving of the ribs along the line of attachment of th~diaphragm referred to as Harrison's grooves. Demineralization and increased resorption also lead to the presence of unmineralized osteoid, referred to as osteomalacia, which preferentially weakens the metaphyseal regions in particular because they have the thinnest cortices and contributes to the bowing deformation of the long bones. c. Radiographic Changes Radiographically the appearance of a thickened growth plate with metaphyseal cupping or concavity is highly suggestive of tickets. The metaphyseal bone is also less sharply delineated at its physeal margin, being described as irregular, fuzzy, or frayed. The secondary ossification centers are less dense and may be delayed in appearance. This is particularly seen at the distal radius and ulna where growth is more extensive than at the proximal ends of these bones and where the physes are normally horizontal. With the institution of medical treatment, repair shortly becomes evident radiographically. The poorly mineralized and thickened cartilage of the hypertrophic zone region rapidly absorbs the calcium, and linear areas of radiodensity at the physeal-metaphyseal region are seen within weeks. Mineralization of the osteoid deposited by the periosteum then becomes apparent on X ray, and the general tendency of the bone is to return to normal density.

d. Deformities of Growth Bowing and angular deformities of the long bones of the lower extremities are common with upper extremity bones much less affected. At the knees, both bowleg (genu varum) and knock-knee (genu valgum) deformities can occur depending on the age at which the metabolic condition is present. A coxa vara deformity of the proximal femur is quite common and can be sufficiently severe that the neck of the femur forms a right angle with the long axis of the diaphysis. Due to the thickness and relatively poor mineralization of this region, deformities develop that further worsen the normal relationship between the epiphyseal and the metaphyseal-diaphyseal regions. Bowing occurs within the metaphyseal and diaphyseal regions, which are thinner than normal and more osteopenic (histologic osteomalacia) due to the presence of unmineralized osteoid and the failure of deep compaction of the Haversian systems. e. Responses to Medical Therapy If treatment is begun within the first 2 years of life, the metabolic abnormality is readily corrected and any bowing of the long bones will almost always remodel without the need for orthopedic intervention. David has outlined the temporal aspects of repair once therapy is begun with vitamin D and, if necessary, calcium supplements (42). The earliest clinical sign of effectiveness is rapid motor improvement with the mild myopathy responding to the appropriate levels of calcium and phosphorus. Plasma Ca, P, and PTH levels normalize within days or weeks, but the levels of 1, 25-dihydroxyvitamin D and alkaline phosphatase tend to remain elevated for months until final healing of tickets has been achieved. The first radiologic signs of healing rickets occur 2-4 weeks after vitamin D therapy is initiated, with the presence of linear dense deposits within the previously unmineralized cartilage at the lower end of the growth plate. These dense metaphyseal zones of calcification corresponding to the newly calcified cartilage may persist for 2-3 years until there is complete remodeling. Bowing of the long bones may take 4-5 years to disappear but should occur as long as there is no recurrence of the tickets condition. 2. FAMILIAL HYPOPHOSPHATEMIC RICKETS [X-LINKED HYPOPHOSPHATEMICRICKETS (XLH)] a. Overview and Pathophysiology This is now the most common type of rickets seen in growing children in developed industrial societies. It was specifically recognized as a distinct disorder by Albright et al. in 1937 and defined as rickets resistant to vitamin D therapy (vitamin D resistant tickets) (1). They noted 6 cases in which much more vitamin D was required for cure than was ordinarily needed for the prevention and cure of nutritional tickets. They described 1 patient in great detail who had been observed for 14 years. The patient with tickets resistant to vitamin D therapy had clinical and histologic evidence of a true tickets and hypophosphatemia that appeared to be due to a secondary hyperparathyroidism. The disorder of metabolism was corrected and the healing of tickets took place only with massive doses

SECTION I 9 Rickets of vitamin D, and therefore Albright et al. concluded that the disturbance was not a deficiency disease like ordinary rickets but rather a form of rickets due to an intrinsic resistance to the antirachitic action of vitamin D. Shortly afterward, Christensen defined the familial occurrence (38). Winters and associates recognized the X-linked dominant inheritance pattern and reviewed in detail the first 65 cases described to provide an overview of the disorder (233). It was then evident that vitamin D resistant tickets was a specific disorder characterized by a clinical, radiological, and biochemical picture similar to that seen in vitamin D deficiency tickets but progressing in spite of adequate antirachitic prophylaxis and healing only in response to very large doses of vitamin D (183). The level of serum inorganic phosphorus is the most sensitive index of the abnormality. Though both sexes are affected, females tend to have mild deformities or none at all, whereas males are affected far more seriously with moderate to severe deformities. It is most commonly transmitted as sex-linked dominant, but rarer modes of transmission such as autosomal dominant, sporadic new X-linked mutations, and autosomal recessive forms have been reported. Glorieux has pointed out that, as is characteristic for a sex-linked dominant condition, it is fully expressed in male hemizygotes and variably expressed in female heterozygotes (74). The phenotype of XLH is caused by a dual abnormality: a renal defect combining phosphate wasting and altered calcitriol synthesis and an osteoblast-osteocyte defect, which decreases the ability of these cells to control mineral deposition in the matrix. Undermineralization of physeal cartilage and osteoid is dependent ultimately on the decreased retention of phosphate by the kidneys due to a failure of its reabsorption in the distal tubular system. The hypophosphatemia also fails to stimulate the 1-hydroxylation of 25hydroxyvitamin D so that 1,25-dihydroxyvitamin D levels are lower than normal. The renal phosphate leak diagnosed by increased urinary phosphate excretion and the loss of the normal compensatory mechanism of 1,25-dihydroxyvitamin D formation combine to produce very low serum phosphate levels. The persistent hypophosphatemia is the best way to recognize affected individuals in combination with elevated serum alkaline phosphatase activity and normal serum calcium. Bicarbonate and PTH levels are generally normal. There is no aminoaciduria. There is decreased calcium absorption from the gastrointestinal tract but also low urinary calcium excretion. b. Gene Abnormality Progress has been made in discovering the genetic abnormality in XLH. Linkage of the hypophosphatemic HYP locus to DNA markers on chromosomal band Xp 22.1 was established (173). Subsequently, the candidate gene was identified that exhibits homology to a family of endopeptidase genes, members of which are involved in the degradation or activation of several peptide hormones (94). The gene, referred to as PEX, is composed of multiple exons. The PEX protein is a zinc metallopeptidase with phosphate regulating functions. Intragenic separate deletions

879

from 4 different families and 3 mutations were detected. Holm et al. found mutations in PEX in 9 of 22 unrelated HYP patients (87). Continuing study involved analysis of 99 HYP families for PEX gene mutations (174). Deletions, insertions, nonsense mutations, stop codons, and splice mutations occurred in 83% of families screened for in all 22 exons and in 51% of a separate set of families screened for 17 PEX gene exons. The wide range of mutations that align with regions required for protease activity suggests that PEX functions as a protease and may act by processing factors involved in bone mineral metabolism. This extensive mutation analysis thus revealed a range of defects in the PEX gene. A mouse model (hyp) identified in 1976 has been an invaluable tool to study the disease at the cellular and molecular levels (54). The hyp mouse has rickets, osteomalacia, hypophosphatemia, and defective renal tubular reabsorption of inorganic phosphorus. Mouse pex is homologous to human PEX. A study by Beck et al. found that p e x - P E X mRNA is expressed predominantly in human fetal and adult mouse calvaria and long bone (15). They concluded that p e x - P E X is a low-abundance transcript that is expressed predominantly in the bones of mice and humans and that a large deletion in the 3' region of the pex gene is present in the murine hyp homologue of X-linked hypophosphatemia. Strom et al. also noted that the mouse homologue ofpex was highly conserved between human and mouse (207). The 31 end of the gene was found to be d e l e t e ~ ~ y p m i c e . Thus, they concluded that the hyp mouse was not only a phenotypic model for X-linked phosphatemia in the human but also had similar genetic abnormalities. Another mouse model for the disorder is referred to as the Gy mouse, which is a mutation found among offspring of an irradiated female mouse (126). The Gy mice, in addition to hypophosphatemia and other findings similar to X-linked hypophosphatemic tickets, show inner ear abnormalities. PEX gene deletions were also found in the Gy mouse variants. c. Cell and Matrix Abnormalities in X L H Bone Studies by Engfeldt et al. on bone and cartilage from four patients with primary vitamin D resistant tickets showed important abnormalities (55). Histologic sections of the physis from the costochondral junction showed the characteristic pattern of tickets with thickening of the physeal cartilage, irregularity and persistence of the proliferating zone, and a disorganized and thickened hypertrophic zone. Vascular invasion from the metaphyseal region was highly irregular with widened blood vessels passing from the marrow cavity into the proliferative cartilage, but in a random fashion and not associated with new bone formation. Adjacent metaphyseal osteoid was extensive. Much of the work, however, concentrated on the appearance of the bone tissue itself both before and after vitamin D therapy. Engfeldt et al. concluded that, although increased mineralization of the bone could occur, there was no change in the abnormal pattern of the collagen. Both the pattern of bone deposition and the extent of its mineralization were abnormal. Whereas lamellar bone was

880

CHAPTER IO ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

seen by light and polarization microscopy, the collagen pattern of the sections was unusual, irregular, and consisted of short bundles running in different directions. On occasion the collagen fibers were organized into the normal Haversian systems but the tendency was for either disorganized lamellar bone or on occasion woven bone to be present. The appearance resembled that seen in Paget's disease. The osteocyte spaces or lacunae were often wide, irregular, and poorly defined. Pericellular mineralization was often minimal. Even within varying trabeculae there were widely varying areas of mineralization. Both the highly mineralized trabeculae and those showing a low mineral content displayed an apatite pattern with the same dimensions as that of normal bone. Numerous resorption cavities were present along with the variation in mineral content, the enlarged osteocyte lacunae, and the abnormal pattern of collagen deposition. Despite almost normal mineralization following massive doses of vitamin D, a distinct lack of normal compact bone formation was noted both histologically and microradiographically. This also was characterized by an enormous variability of the calcium content in different areas. The pattern was characteristic of a high-turnover pattern with extensive remodeling of the bone tissue with intense resorption and new deposition noted. Engfeldt et al. concluded that, although the high dose of vitamin D increased mineralization of the physis, it was unable to convert the structural abnormality of the bone tissue itself to a normal pattern. More recent studies have also implicated an actual osteoblast defect in X-linked hypophosphatemia based on detailed studies in the hyp mouse (50, 51). To test the hypothesis of an abnormal osteoblast function in XLH, periosteal osteoblasts isolated from normal and hyp mice were transplanted into normal and mutant mice. Impaired mineralization was seen in transplants of hyp cells into hyp mice as defined by excessively thick osteoid seams compared with transplants of normal cells into normal mice. When normal cells were transplanted into mutant mice, osteoid thickness was again markedly increased indicating that the extracellular host environment was critical for bone formation. Further studies demonstrated an inability for hyp osteoblasts to produce normal bone when placed in a normal environment, leading to the interpretation of an intrinsic cellular abnormality of the osteoblast as an important target for hyp mutations. There appear to be primary abnormalities in osteoblasts. Phosphate and vitamin D supplementation healed the mineralization defect at the epiphyseal plate region but not at the endosteal bone surfaces, suggesting that a bone resistant form of vitamin D is present. Lesions around the osteocyte lacunae never disappeared completely, indicating a direct abnormality of the osteocyte itself. d. Hypothetical Model of Disease Carpenter has utilized investigational findings in relation to XLH to produce a hypothetical model of the disorder (30). He feels that both human and murine studies suggest that a humoral factor mediates the defect in renal phosphate transport. The disor-

der, which is evident clinically in affecting the bones, is independent of disease expression at the kidney, which is shown by the increased circulating phosphate levels although the organ itself otherwise functions normally. Carpenter feels that the mineralized tissues may reflect a gene dose effect not present in the kidney. The hypophosphatemia is not associated with appropriate production of 1,25dihydroxyvitamin D. The abnormal gene in XLH appears to encode endopeptidase, but the human form is different from the murine form as it is underexpressed in hyp mice. He feels that renal intracellular phosphate is probably normal, which may mask the hypophosphatemic stimulus to renal 1,25dihydroxyvitamin D production. The important factors relating to phosphate movement across renal cell membranes are the existence of inorganic phosphate (Pi) regulating substances, which preserve the intracellular levels of the proximal renal tubular cell. When Pi is insufficient, a decrease in the factor would decrease the intracellular Pi, thereby mediating an increase in membrane transport of Pi and loL-hydroxylation of vitamin D. Unregulated persistence of such a factor during brief Pi depravation would disrupt appropriate renal Pi conservation and generation of 1,25dihydroxyvitamin D, which is the precise renal defect seen in XLH-hyp. Carpenter goes on to suggest that "the synthesis, processing or secretion of this putative humoral factor is rendered abnormal in XLH by mutations in the PEX gene such that normal regulation of the Pi homeostatic system is disrupted." e. Clinical and Radiographic Characteristics The disorder has some meaningful clinical variations from nutritional rickets. These include observations that the children are systemically well, there is no proximal muscle weakness, tetany does not occur, there is no bone pain, and the condition does not usually present clinically until 2-3 years of age or older. There is no involvement of the spine. The condition presents with hypophosphatemia, lower limb bowing deformities, and a slowed rate of growth (30). Characteristic radiographs are shown in Fig. 3. Bowing Deformities: The gradually developing deformity of the lower limbs involves proximal femoral bowing of a coxa vara type, anterolateral diaphyseal-metaphyseal femoral bowing, genu varum or valgum with proximal tibial bowing, and internal tibial torsion. Coxa vara deformity of the proximal end of the femur is a characteristic feature of severe XLH rickets, particularly in those instances in which growth and development have occurred over a few-year period before diagnosis or in which medical treatment is ineffective after diagnosis has been made. The bowing is associated with widening of the growth plate, thinning of trabecular bone due to the failure of appropriate mineralized osteoid deposition, and also thinning of the adjacent cortices. The physeal-metaphyseal radiologic changes are the same as those described earlier for nutritional tickets. There is thickening of physeal height, an irregular or frayed metaphyseal border, and cupping of the metaphyses. Physeal thickening

SECTION I 9 Rickets

881

F I G U R E 3 Characteristic radiographs in familial hypophosphatemic tickets. (A) Anteroposterior knee radiographs show widened distal femoral and proximal tibial physes. (B) Anteroposterior (Bi) and lateral (Bii) radiographs show characteristic lateral and anterior femoral bowing (arrows) with (Biii) relatively flexed position of distal femoral metaphysis-epiphysis. (C) Widened physis of distal tibias.

may be asymmetric with one side of the physis (for example, the medial half) thicker than the other. The bending is not caused by a slip of the proximal femoral epiphysis or by a fracture through the neck, but rather it is characterized by continuity of tissue with a tendency for it to bend at physeal cartilage, metaphyseal bone, and diaphyseal bone regions due to inadequate mineralization. Swelling of the wrist, an-

kle, and costochondral junctions is the same as that seen in the vitamin D deficient forms. Males are affected more severely than females with the sex-linked dominant inheritance. Dental abscesses are frequently seen due to abnormal tooth structure. Short Stature: Short stature is seen frequently in those affected and is almost always accompanied by the characteristic

882

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematolotlic Disorders

rachitic deformities. Normal stature is considered to range between 2 standard deviations above and below the mean. Most patients are below the 2nd standard deviation below the mean, but some are in the low-normal range. The shortened stature is based entirely on the shortened lower extremities and is not solely or even primarily due to the bowing deformities. No form of medical treatment has affected overall height even where it improves the rickets metabolically and radiographically. In several families studied by Burnett et at., no affected individual was taller than 65 in. (26). Seven affected males had a mean height of 58.6 in. compared to 19 nonaffected male family members with a mean height of 68.7 in. In females, 19 affected family members had a mean height of 56.8 in. and 26 nonaffected members had a mean height of 62.2 in. Similar findings were noted by Tapia et al. in a smaller series (213). The final mean height in 4 males was 58.9 in. (range = 52-63.5 in.) and in 7 females was 56.5 in. (range = 52-58.5 in). Similar height ranges were listed by Pedersen and McCarroll, with most adult patients between 51 and 59 in. and the tallest at 65 in. (153). Balsan and Tieder studied linear growth in a large group of XLH patients and noted no acceleration of growth with vitamin D or calcitriol regimens, but they saw "catchup growth" in 10 of 16 on loL-hydroxyvitamin D 3 (12). Stickler and Morganstern studied the heights of 52 patients aged at least 18 years with X-linked hypophosphatemic tickets (204). They concluded that there was no evidence that any form of treatment (vitamin D in high doses, vitamin D plus phosphate supplements, or calcitriol plus phosphate) had an effect on adult height. Stickler and Morganstern also noted that the complications of treatment such as renal failure due to vitamin D intoxication led to the question as to whether medical treatment had any benefit at all. They even considered the possibility of recommending no treatment for the primary disorder. Their criteria for diagnosis involved hypophosphatemia, a high alkaline phosphatase level, radiological evidence of tickets primarily in the legs, and no response to vitamin D in physiological doses. Stickler and Morganstern compared the treated patients to 8 other patients with the disorder, 4 of whom received no treatment and 4 of whom had been treated for 2 years or less. Regardless of the medical regimen used, the final height was not affected by treatment; in addition, neither the age at which treatment was started nor the duration of treatment had an effect on adult height. The 8 patients who had reached adulthood without any treatment or less than 2 years of treatment had a mean height of 3.06 standard deviations (SD) below the mean compared with 2.88 SD below the mean for 20 patients treated continuously for at least 8 years. There was also no correlation between treatment and the number of osteotomies the patients had undergone. f. Medical Treatment The medical treatment of XLH has continued to undergo changes due to the fact that no fully curative regimen has been defined. Whereas some studies have suggested that medical therapy is futile, most investi-

9 9

Calcitriol Phosphate salts

~ Assess

results

9 q 12 hours 9 5 doses / 24 hours

by:

9 Growth curve 9 Correction of deformities 9 Radiological healing 9 Serum alkaline phosphatase activity

9 Monitor

dosage

of:

Phosphate with" 9 Titration (A 1 5 mg/dl, 40-60 mins alter dose) 9 Clinical results Calcitriol by: 9 Urinary Ca (rag)/Creat (rag): < 0.3 q 3 months (on random morning sample) 9 Serum iPTH ~ 6 months

F I G U R E 4 Treatment profile for hypophosphatemic rickets is illustrated. [Reprinted from Glorieux et al. (1991), Nestle Nut. Workshop Ser. 21" 194, 9 Lippincott Williams & Wilkins, with permission.]

gators at present continue to favor it though fully recognizing that it is neither curative nor totally benign. Research discoveries in relation to the disorder have better defined the underlying pathophysiology. The initial approach to therapy was the use of massive doses of vitamin D (greater than 50,000 IU per day), which improved the tickets but led to hypercalcemia and renal damage in many instances as well as little if any improvement in the slow growth rate (1, 30, 73, 74, 204, 222). Stamp et al. noted that they were unable to prevent the recurrence of deformity or eliminate short stature in their patients (200). Pierce et al. showed poor control with high-dose vitamin D and a tendency to hypervitaminosis D toxicity and renal damage when therapy continued (158). The evolution of therapy has been well-summarized by Glorieux et al. (75). They developed a treatment regimen that used 1,25-dihydroxyvitamin D rather than vitamin D itself along with phosphate supplements to minimize phosphateinduced hyperparathyroidism (Fig. 4). They noted improvement in the growth rate, which they felt was primarily the result of the phosphate supplement. The overall growth period was not shortened and the adolescent growth spurt was of normal magnitude and duration. Metabolic control of the disorder was defined by a normal growth rate, normalization of the serum alkaline phosphatase level, and radiologic healing of tickets. These parameters are also important determinants of the timing for orthopedic intervention. The medical regimen involves calcitriol twice daily and phosphate salts at five separate times over a 24-hr period. The viewpoint of Stickler and Morgenstern (204) was supported to a considerable extent by Verge et al. (222) in a long-term study of 24 patients. A detailed assessment of the effects of medical therapy in XLH, however, commented both on its effects on growth and on the negative occurrence of nephrocalcinosis. With the calcitriol and phosphate treatment, there was a slight statistically significant increase in growth with the mean height SD score of - 1.08 as compared with - 2 . 0 5 in untreated historical controls. In 13 patients treated for at least 2 years, the increase in the mean height score was 0.33 from - 1 . 5 8 to -1.25. This too was statisti-

SECTION ! ~ Rickets cally significant. Actual height measurements were not provided. Of note, however, is the fact that in each group not all patients' rate of growth increased and some actually declined. The overall mean values, however, were indicative of slightly increased growth. There was much clearer evidence, however, of the negative sequelae in terms of the nephrocalcinosis detected on renal ultrasonography. Nineteen of 24 patients (79%) had nephrocalcinosis with the grade correlating significantly with the mean phosphate dose but not with the dose of vitamin D or the duration of therapy. The calcitriol and phosphate therapy helped to minimize lower limb deformity because few patients required surgical intervention in terms of osteotomy, whereas before the introduction of the combination therapy the majority of such patients required correction. The childhood therapy or a continuation of such therapy past skeletal maturity was of no benefit to the adult patient. With regard to height, the combination therapy had a slightly beneficial effect, but some previous studies had failed to demonstrate any beneficial effect. The absence of nephrocalcinosis in untreated adult relatives within the patient population supported the belief that nephrocalcinosis was a sequence of therapy rather than part of the natural history of the disorder. The radiographic response to medical treatment showing increased density of the lower extremity physes and adjacent metaphyses is similar to that observed following treatment for nutritional rickets. Because the XLH disorder is benign with a full life expectancy, any negative implications of therapy on renal function are of serious concern. The value in terms of bone strengthening and diminution of deformity is real, but close assessment for evolving renal problems is essential. Glorieux, however, strongly supports the need for treating patients with X-linked hypophosphatemic tickets (74). A combination of phosphate and calcitriol therapy was considered to be best, with patient compliance and close monitoring of renal and metabolic features considered to be mandatory. Glorieux felt that the diminution in the need for osteotomy was a major benefit with the other matters being controllable with close observation. Carpenter stresses that early treatment is likely to result in better outcomes (30). Children from affected families should be screened for abnormal serum and urine phosphorus levels and serum alkaline phosphatase activity within the first month of life and then again at 3 and 6 months. If subsequent radiographic examination is suggestive of the XLH diagnosis, therapy is started with calcitriol and phosphate. In Carpenter's unit the dosage range is quite wide depending on the severity of the tickets, the response to therapy, and the complications encountered. The three major complications of therapy for XLH are hypervitaminosis D, hyperparathyroidism, and soft tissue calcification particularly in the kidney. g. Orthopedic M a n a g e m e n t Virtually all operative treatment in XLH tickets is for femoral and tibial deformities. Upper extremity deformity, especially necessitating surgical treatment, is rarely if ever seen. Evans et al. defined three

883

levels of expectation in relation to deformities in XLH rickets (58). In those patients in whom the diagnosis is made within the first year of life, medical treatment will almost always cure or at least control the rickets such that deformities do not develop during the growing years. In those patients seen with mild deformities between 1 and 5 years of age, the institution of appropriate medical therapy should be sufficient to prevent the deformities from worsening to levels of clinical significance and to allow for spontaneous correction with growth as would be the case, for example, with instances of physiologic bowing of the lower extremities in this age group. In those patients greater than 5 years of age at the time of diagnosis, medical treatment is often not sufficient to allow for correction of the deformities, which have progressed to a considerable degree in many instances, and osteotomy must generally be used. Rubinovitch et al. also indicated that children under 6 years of age with milder deformities frequently corrected with growth and good metabolic control, whereas those above 6 years of age with severe deformity generally needed osteotomy (175). Pierce et al., reporting in 1964, expressed concern about the ability to correct deformity during the growing years and advocated brace management with osteotomy deferred until skeletal maturity (158). Tapia et al., in the same year, noted maintenance of correction gained by osteotomy in all 13 cases under adequate vitamin D therapy with the recurrence of deformity in 50% (9 of 18) of cases in which it was inadequate (213). Medical therapy is increasingly difficult to use in older children as the level of cooperation for the rigorous medical regimens tends to diminish in all but the most committed of families and patients. Replacement medical therapy must often be given 4 - 6 times per day and involves pills or liquid intake, which both are unpleasant to taste, and in fairly large volumes. In addition, the medical therapy in many instances is not sufficiently refined to lead to full correction of the abnormality even if faithfully adhered to. The second problem relates to the nature of the deformity. In the femur the bowing tends to involve the entire bone with an anterolateral bowing deformity seen. At the proximal end there is a coxa vara, in the diaphysis there is actual bowing due to the weakened state of the cortices as well as to the fact that during earlier time periods the metaphyseal regions were also bowed, and at the distal end metaphyseal and epiphyseal bowing is prominent due to the extensive stresses of the knee region and the fact that 70% of the growth of the femur occurs distally. Bowing of the tibia is not as marked generally but may involve the entire bone. The proximal tibial bowing can be either genu varum or genu valgum depending on the time during the growth period that the tickets is most pronounced and on the weight of child. There are two important considerations in attempting surgical correction of limb deformities in growing children with tickets. (1) Medical therapy must be well-established prior to the performance of corrective osteotomy. In severe nutritional tickets, which is unusual today, the disorder should be

884

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

cured metabolically prior to surgical intervention; vitamin D therapy restores the serum phosphorus and alkaline phosphatase levels to normal. In those with familial hypophosphatemic rickets or other variants, the most optimal form of medical management should have been achieved. If there are several years of growth remaining, correction of the deformity in the presence of continuing rickets will almost certainly have to be repeated as the subsequent growth in the presence of widened growth plates and weakened metaphyses will lead to recurrence of the deformity. There is also a marked tendency to delayed bone union, approaching nonunion on occasion, following osteotomy if the rickets is poorly controlled. In several papers over the past few decades, however, the osteotomies have been shown to unite regardless of the method of medical management. The impression is still gained, however, that healing can be delayed especially where medical control is suboptimal. (2) Close consideration must be given to the site of bowing, the alignment of both femurs and tibias, the timing of surgical interventions, and the method of immobilization postosteotomy. Three surgical approaches are considered for bone deformities: (1) single-level wedge osteotomy for focal angulation, (2) multilevel (two or more) osteotomy for bowing that is multiplanar and involves the entire bone from proximal end to distal end (physeal, metaphyseal, diaphyseal, metaphyseal, physeal bowing), and (3) asymmetric physeal stapling for focal deformity leading to fairly acute physeal-metaphyseal angulation. Single-level closing wedge or opening wedge osteotomies are performed when the deformity is reasonably localized, with stabilization usually with rigid AO plate instrumentation, although those expert in the use of external fixators have reported good results. The proximal femur is particularly difficult to correct in cases of moderate to advanced tickets, but it must be addressed or else reasonably good alignment achieved distal to it can be compromised by a persistent waddling gait. Even though the limbs appear to be straight clinically, patients can be less than fully satisfied by the mechanics of gait due to a persisting uncorrected proximal coxa vara deformity. For these reasons, approaches to the femur may require two or more levels of osteotomy followed by insertion of an intramedullary rod. The intramedullary rod is effective particularly in allowing for correction of the bowing of the proximal femur, which has led to the coxa vara deformity. Biplanar or multiplanar angular deformity is often present throughout the entire extent of both the femur and the tibia such that single-level osteotomy may not fully correct limb malalignment. Angular deformity, if focal, can also be corrected using asymmetric stapling. Osteotomies to correct femoral and tibial bowing have been studied in relation to the level of correction. In many instances, multiple osteotomies have been needed to gain full correction. Ferris et al. felt that diaphyseal osteotomies followed by intramedullary nailing for stabilization could be performed effectively at any age (64). The deformities

were multiplanar, involving flexion, lateral deviation, and rotational components. Wedge osteotomies of the diaphyseal segment at one level rarely led to full correction. Intramedullary nailing had many advantages, allowing for improved correction at two or more sites. Much of the deformity is concentrated in the metaphyseal region because of the closeness of the physis and because of the fact that metaphyseal bone bows more readily because of the thinness of the cortices. It is often essential to perform correction at the metaphyseal level, although fxation is a problem. The recommendation was to perform the metaphyseal correction close to skeletal maturity because the likelihood of recurrence is greater at this region the earlier the deformity is corrected. Eyres et al. favored intramedullary rodding with interlocking nails for rotational stability even when a single midshaft osteotomy was performed (60). They had experienced considerable difficulties with nonunion, delayed union, and recurrent deformity in osteotomies and asymmetric stapling in immature patients. Kanel and Price have shown the value of treating the multiplanar deformities with transverse osteotomies and use of a unilateral external fixation device (101). This approach allows for multiple osteotomies and correction through the metaphyseal regions. The patient remains ambulatory with crutches with no need for cast immobilization. They reported on 29 osteotomies in only 9 patients, a reflection of the fact that most patients require bilateral femoral and tibial osteotomies. All osteotomies healed. The external fixators remained on for an average of 90 days. In milder deformities asymmetric stapling toward the end of skeletal growth can be effective at distal femur or proximal and distal tibial physes. Examples of surgical corrective procedures for lower extremity deformities are shown in Fig. 5. 3. RENAL OSTEODYSTROPHY OR RENAL RICKETS

a. Overview and Pathophysiology Renal osteodystrophy refers to the secondary effects of progressive chronic renal failure on developing bone in children (77, 131, 177, 178). In the earlier literature it was referred to as renal rickets. Lucas recognized the correlation between renal and bone disease in 1883 (124). The most common causes of renal failure are renal dysplasia, obstructive uropathy, and reflux nephropathy in the congenital group and chronic glomerulonephritis and focal glomerular sclerosis in the acquired group (7, 14). The skeletal disorder is characterized by osteitis fibrosa, frank tickets or osteomalacia, and a series of typical clinical and radiographic abnormalities. These are due ultimately to a secondary hyperparathyroidism and include myopathy, growth slowdown, bone deformity, slipped epiphyses, and occasionally avascular necrosis. The kidney is responsible for the synthesis of 1,25-dihydroxyvitamin D3 (calcitriol), the most active metabolite of vitamin D. With renal disease the decreased ability of the kidney to synthesize calcitriol leads subsequently to the development of secondary hyperparathyroidism. Renal osteodystrophy inhibits

SECTION I 9 Rickets

885

FIGURE 5 Examplesof surgical corrective procedures for lower extremity deformities in familial hypophosphatemic rickets are seen. (A) In the young patient (5-6 years) a distal femoral valgus, extension osteotomy at one site may allow for good correction in anteroposterior (Ai) and lateral (Aii) projections. (B) By 10-12 years of age multilevel osteotomies are often needed for correction. (C) Distal angulation of the femur may still need final correction at skeletal maturity.

25-hydroxyvitamin D-loL-hydroxylase, leading to the impairment of calcitriol synthesis. The chronic renal failure syndrome is characterized metabolically by hyperphosphatemia, hypocalcemia, impaired renal calcitriol synthesis, and skeletal resistance to the calcemic action of parathyroid hormone. Radiologically this manifests as extensive skeletal demineralization and lytic destructive bone lesions resulting from the hyperparathyroidism. One of the characteristic findings of the disorder is a widening of the growth plates due to the failure of hypertrophic cartilage and metaphyseal bone calcification and the particular occurrence of slipped epiphyses throughout the skeleton (Fig. 6).

b. High-Turnover and Low-Turnover Bone Lesions in Renal Osteodystrophy The use of bone biopsy techniques in assessing bone structure particularly in relation to renal osteodystrophy has allowed for the distinction of two pattems of bone abnormality (91, 177). High-Turnover Bone Lesions: Bone abnormalities are defined as high-turnover when the response to high serum levels of parathyroid hormone is characterized by the simultaneous occurrence of active bone resorption and increased osteoblastic activity. The combined increase in osteoblastic synthesis and osteoclastic remodeling activity leads to the high turnover. Increased osteoclastic activity leads to bone resorption with an increase in the number and size of osteo-

clasts and diminished matrix. Fibrous tissue then accumulates adjacent to the bone trabeculae and within the marrow with complete fibrous replacement of individual trabeculae in severe cases. Although osteoblastic activity is increased, there is inability to mineralize the matrix leading to persistence of disordered osteoid and accumulations of fibrous tissue. These histologic findings are incorporated under the term osteitis fibrosa. Low-Turnover Bone Lesions: The histologic hallmark of low-turnover bone disease is osteomalacia. This refers simply to the presence of excess unmineralized osteoid, which accumulates in bone due to defective mineralization. The osteoid seams on trabecular bone within the marrow and within the cortex are thickened. When osteomalacia predominates, osteoblastic activity is markedly reduced. With continuation of a low-turnover state, there are normal or reduced amounts of osteoid synthesized, no tissue fibrosis, diminished numbers of both osteoblasts and osteoclasts, and little to no evidence of bone formation.

c. Bone Deformities in Childhood Renal Osteodystrophy:

Angular Deformity, Slipped Epiphyses, and Osteonecrosis Three major groups of orthopedic problems occur in the childhood end stage group with renal failure: characteristic angular deformities, slipped epiphyses, and osteonecrosis. The most common bone deformities involve a genu valgum

886

CHAPTER IO ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

FIGURE 6 Illustrationsfrom the work of Mehls et al. show the physeal and metaphysealchanges in a child with tickets due to vitamin D deficiency(centrally) and renal osteodystrophy(right). The proliferating and hypertrophic regions of the physis are widened in vitamin D deficient rickets (straight arrow), whereas the additional characteristic histologic finding in renal osteodystrophy is resorption of the lower regions of the physeal cartilage and their replacement with fibrous tissue and poorly organized and mineralized woven bone (curved arrow). [Reprinted from Mehls et al. (1980), Clin. Endoctin. Metab. 9:151-176, with permission.]

or genu varum, which are primarily due to metaphyseal deformity in either the distal femur or the proximal tibia or both. Angular metaphyseal deformities about the knee predominate, the direction of which appears to be dependent on the age at which the renal abnormality occurs. In those under 2.5 years of age in whom physiologic bowing predominates, metaphyseal distal femoral and proximal tibial varus worsen, whereas with disease worsening after this time the normal valgus increases to a pathologic range. Salusky has estimated that even with intensive vitamin D management of renal osteodystrophy, 20-25% of pediatric patients undergoing long-term dialysis require orthopedic procedures for deformities (177). Mehls has indicated that the majority of children who present with chronic renal failure are older than 10 years of age (131). The most common deformity is genu valgum. In a study by Davids et al., 11 of 12 patients with apparent genu valgum had tibiofemoral angles greater than 15 ~ (43). As well as assessing the tibiofemoral diaphyseal angle, they established measurements for a femoral alignment angle and a tibial alignment angle. In patients with chronic renal failure and renal osteodystrophy, the nature of the lower extremity angular deformity is believed to be related to the age at onset of renal failure. Those occurring at less than 3 years of age would tend to worsen the physiologic varus alignment, whereas those occurring after that age would worsen the valgus position. In Davids et al.'s series, all patients with genu valgum were older than 4 years of age at the onset of chronic renal failure. Oppenheimet al. identified 8 patients with significant valgus angulation of the proximal tibia, which they felt was physeal in origin (148). They likened

the radiographic changes to those seen in tibia vara with Blount's disease. Eleven of their 27 patients with renal osteodystrophy were treated for genu valgum. Apel et al. studied a pediatric renal transplant population in terms of skeletal abnormalities (7). They analyzed 130 patients who had undergone renal transplantation and thus had end stage renal failure or renal osteodystrophy. Osteonecroses were seen in 6 femoral heads, 1 lateral femoral condyle, and 3 tali. All cases occurred after transplantation. Slipped epiphyses were seen at the proximal humerus (1), distal radial epiphysis (4), proximal femoral capital epiphysis (8), and distal femoral epiphysis (1). The most common deformities were genu valgum (21), with other deformities being ankle valgus (9), coxa vara (7), genu varum (6), and ankle varum (2). Scoliosis was seen in 5 patients. The distal radial epiphyses slipped at an average age of 6.6 years, proximal humeral at 9 years, proximal femoral at 10 years, and distal femoral at 11.75 years. The coxa vara deformities were noted at an average age of 3.1 years, genu varum at 4 years, genu valgum at 11.6 years, ankle varus at 1.75 years, and ankle valgus at 8.2 years. Barrett and Papadimitirou reviewed skeletal disorders in children with renal failure, all of whom had biochemical and radiologic evidence of renal osteodystrophy (14). They reported on 16 patients out of 124 children with renal failure referred for management of skeletal problems. The nature and distribution of problems were similar to those of other reports. The most common deformity was genu valgum occurring in 11 of 16 children. Slipped capital femoral epiphysis was seen in 6, with all cases bilateral and all associated with genu valgum. Ankle valgus was seen in 4. The genu

SECTION

FIGURE 7 Characteristicradiograph shows bone changes in renal osteodystrophywith widenedphysis of proximal femur.

valgum was due to instability at the physis and bending of the metaphyseal regions. Other isolated deformities involved osteonecrosis of the distal femoral lateral condyle, osteochondritis dissecans, and progressive bowing of the femurs and tibias. Characteristic bone changes in renal osteodystrophy are illustrated in Fig. 7.

d. Pathogenesis of Bone Deformities in Renal Osteodystrophy Slipped Epiphyses: Epiphyseal slipping is the most common and most severe of the orthopedic sequelae of a childhood renal osteodystrophy. The mechanism appears to be somewhat different from that that occurs in idiopathic slipped capital femoral epiphysis. The slipped epiphyses in children with renal failure are a consequence of the general metabolic bone disease. Work by Mehls and colleagues has demonstrated that the slippage is secondary to physeal thickening and to associated osteitis fibrosa of the adjacent metaphysis (132). In children with uremia, the normal transformation of calcified cartilage at the lower margins of the endochondral sequence into metaphyseal bone is irregular. There is undermineralization of the hypertrophic cartilage and, thus, a thickening of a cartilage layer of the growth plate. The metaphyseal bone too is abnormal because the undifferentiated and newly differentiated osteoblast and its osteoid matrix suffer from a defect of mineralization. There is a tendency for dense fibrous tissue to be interposed between the growth cartilage and the adjacent metaphysis. This is weaker structurally than the normal tissues and provides a plane of cleavage for slipping of epiphyses. The vast majority of slipped epiphyses in renal osteodystrophy are not associated with trauma, have little pain, and appear to be transphyseal entities and not metaphyseal fractures. Many epiphyses can be the site of slipping in renal osteodystrophy. The slipping, which is virtually pain-free, occurs commonly at the proximal femur, and renal osteodystrophy must be included in the differential diagnosis of slipped capital femoral epiphysis particularly when the patient is less than 10 years of age. Of importance, however, is the fact that other epiphyses may also slip in renal osteodystrophy, for example, those of the proximal humerus, distal femur, proximal tibia, and distal tibia (214). Indeed, when nontraumatic

! 9 Rickets

887

slipping occurs at any epiphysis other than the proximal femur, the finding is essentially pathognomonic for a renal osteodystrophy condition. Other than an isolated report of slipped bilateral distal femoral epiphyses in scurvy, no other disorder has been implicated. The sites of epiphyseal slipping are age-related. Before the age of 5 or 6 years, slipping can be seen in both the upper and lower femoral epiphyses and in the distal tibial epiphyses. In older children the upper femoral epiphyses and the distal epiphyses of the radius and ulna are involved primarily. During and after puberty, slipping is seen frequently in the distal forearm bone epiphyses. When there is extremely severe osteitis fibrosa, epiphyseal slipping can occur irrespective of age in nearly all epiphyses. The slippage that occurs at the distal radial and ulnar epiphyses tends to be in a dorsolateral direction. Great caution must be exercised in deciding whether to perform surgical correction of slipped epiphyses. Extremely poor results have been noted when surgical correction has been undertaken before the metabolic bone disease has been cured or at least controlled. Although early intervention might be required on occasion for proximal femoral slippages, even here the extreme youth of the children and the marked osteopenia of the bone may preclude effective pinning. It is essential that appropriate treatment of the metabolic bone disease occurs prior to any orthopedic correction particularly when long bone osteotomy is projected. Immobilization also should be minimized because of the worsening of any osteoporosis. Swierstra et al. noted good results in three distal femoral slippages in young children using long leg casts followed by orthoses in conjunction with medical management (212). Slipped capital femoral epiphysis in renal osteodystrophy has also been reviewed in Chapter 5. Metaphyseal Fractures: Osteomalacia with poorly oriented metaphyseal trabeculae leads to considerable weakening of the metaphyseal regions of the long bones. If the condition persists and the child remains active, there can be gradual deformity mediated through the metaphyseal bone region. In other instances fracture can occur in these metaphyseal areas. It is extremely important to identify whether deformity in the region of the end of the long bone in renal osteodystrophy is due to slippage of the epiphyses or due to acute or chronic bending through the adjacent metaphyseal region. Osteitis Fibrosa: Mehls and colleagues have reviewed the specific differences in the physeal and periphyseal regions in renal osteodystrophy as compared with nutritional and vitamin D resistant hypophosphatemic tickets (131, 132). In early and middle stages of renal osteodystrophy, the zone of growth cartilage does not tend to be increased in height but rather is normal or even decreased. The early provisional calcification of the cartilage matrix is not defective. The abnormality is in the transition zone between the growth cartilage and the metaphysis, which is highly abnormal in the sense that vascular invasion is virtually absent. There is a tendency for the chondro-osteoid trabecular regions to be resorbed by

888

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

increased chondroclastic and osteoclastic activity. Growth cartilage is physically separated from the metaphyseal bone by a zone of fibrous tissue in advanced stages, and often woven bone forms within this tissue rather than forming on calcified cartilage trabeculae. Radiographically it may be difficult if not impossible to distinguish rachitic changes with thickened growth plate cartilage and osteitis fibrosa of the growth zones as both would be radiolucent. Woven bone rather than lamellar bone characterizes the osteitis fibrosa condition. Endosteal fibrosis is also seen. In high-turnover states the histologic response is fibro-osteoclastic, whereas in lowturnover states osteoporosis-osteomalacia predominates. Growth Retardation: Renal osteodystrophy is also associated with severe growth retardation. Bone age is noted to lag behind chronological age relatively early in the course of progressive renal disease. Linear growth unfortunately remains impaired even with renal replacement therapies involving either hemodialysis or peritoneal dialysis. There are significant relationships between growth hormone and bone and mineral metabolism with effects occurring on both the endochondral and intramembranous bone formation mechanisms. Mehls has noted specific changes in growth plate cartilage due to severe secondary hyperparathyroidism in children with chronic renal failure (131). Growth plate involvement in patients with osteitis fibrosa can lead to slipped epiphysis, bone deformities, and impaired linear growth. It is becoming apparent that the expression of the receptor for PTH and PTH-related peptide (PTHrP) and the production of PTHrP itself play a critical role in normal endochondral bone development. Deletion of the PTHrP gene in mice leads to early death and marked abnormalities of skeletal development. The growth plates are shortened and calcification of epiphyseal cartilage occurs prematurely. In addition, there is generalized overmineralization of the skeleton. Although serum growth hormone tends to be normal in chronic renal disease, tissue response to growth hormone is often diminished. These appear to relate to changes in receptor expression and in postreceptor events. Growth hormone (GH) actions are mediated through insulin-like growth factor-1 (IGF- 1). Various IGF binding proteins provide an additional mechanism by which tissue-specific actions of GH can be altered. Growth hormone acts directly on bone and on growth plate cartilage and is thus critical for postnatal skeletal growth and the induction of the endochondral ossification sequence. A proliferative response to growth hormone has been documented in chondrogenic cells from epiphyseal growth plate, and growth hormone is considered a major regulator of cartilage growth. One mechanism of its action is the stimulation of the synthesis of IGF-1 locally within the growth plate cartilage. The differentiation of prechondrocytes to chondrocytes is stimulated by both GH and IGF-1. In the epiphyseal cartilage, IGF-1 acts primarily in a paracrine manner to enhance chondrocyte proliferation. Osteonecrosis: Avascular necrosis is most common in renal osteodystrophy following renal transplantation, but it can

occur in the absence of transplantation. The femoral heads are the most common site in children but the distal femoral condyles also have frequent involvement. Mehls et al. reported on osteonecrosis of the femoral head in three patients in chronic renal failure in the absence of steroid therapy or transplantation (133). The AVN was unilateral in each case and was noted at 4, 8, and 9.5 years of age. Other studies documented osteonecrosis after renal transplantation in children. Stern and Watts noted osteonecrosis in 9 of 36 (25%) children following renal transplantation (203). Each case occurred in patients with a skeletal age greater than 10 years. There were no cases younger than this. Many patients had more than one site involved. AVN in 9 patients was in the femoral condyle (all bilateral), in 5 patients it involved the femoral head (bilateral in 4), in 4 patients it occurred in the talus (bilateral in 3), in 2 in the humeral head (bilateral in 1), and in 1 in both patellae. There was no correlation between the total dose of steroid and the development of osteonecrosis. On the other hand, the number of rejection phenomena correlated well with the development of osteonecrosis in that all patients in whom it occurred had at least one rejection episode, whereas in 9 patients who had no rejection episodes there were no instances of osteonecrosis. There were, however, 15 patients with one or more rejection episodes who did not develop the osteonecrosis. Uittenbogaart et al. also noted avascular necrosis in 11 of 171 (6%) recipients of renal allografts (219). They also noted no statistically significant difference in the total steroid dose received during the first posttransplant year between patients in whom necrosis developed and those in whom it did not. The sites involved were similar to other studies with frequent multiple bone involvement. The main sites of involvement were the femoral heads, the distal femoral condyles, and the talus. e. Orthopedic Treatment Control of the metabolic defect, by which is meant secondary hyperparathyroidism, should be optimal prior to any operative orthopedic treatment (147, 177). Crucial to a good result following any orthopedic intervention is appropriate assessment and control of the metabolic status in the perioperative period. Surgical complications can often be found to be due to metabolic instability. Assessment can include bone biopsy as well as radiographic and biochemical studies. The aim of therapy is to control the secondary hyperparathyroidism with effective vitamin D sterols, restrict dietary phosphate, and minimize the accumulation of toxic substances such as aluminum. Stabilization of the patient's metabolic profile can include dialysis, phosphate treatments, vitamin D replacement, correction of secondary hyperparathyroidism, and chelation therapy for aluminum intoxication. Oppenheim et al. indicated that, when the secondary hyperparathyroidism was well-controlled, orthopedic surgical procedures became much more feasible in renal osteodystrophy (147). A review of 24 patients with renal osteodystrophy who had undergone 41 orthopedic operations over a 6-year period indicated good results. There were no signifi-

SECTION ii 9 I n f l a m m a t o r y D i s o r d e r s

cant postoperative wound infections, and all osteotomies healed within an appropriate time. Surgical procedures included pinning for slipped capital femoral epiphysis. Genu valgum or other angular deformities did not resolve once the secondary hyperparathyroidism had been controlled. None of nine patients demonstrated spontaneous improvement at the distal femur or proximal tibia, although no details of age, severity of uremic involvement, or degree of deformity were given. Osteotomy was performed at either distal femur or proximal tibia with good healing. Total joint arthroplasty has often been needed early in adult life for the sequelae of proximal and distal femoral osteonecrosis.

II. I N F L A M M A T O R Y D I S O R D E R S A. Juvenile R h e u m a t o i d Arthritis 1. DISEASE PROFILE Juvenile rheumatoid arthritis (JRA) is an inflammatory disorder characterized by synovitis, which can involve either one joint alone or two or more joints simultaneously or at different times (5, 6, 33, 81, 179). The most severe variant is referred to as Still's disease, a multijoint synovitis associated with systemic symptoms involving pericarditis, lymphadenopathy, splenomegaly, and transient rash (5, 6, 179). More commonly, however, one notes involvement of a single joint with the inflammatory process. The most commonly affected joints in the child differ from those in the adult and involve the knee, subtalar joint of the foot, and hip. The etiology of childhood rheumatoid arthritis remains obscure. Many of the better known rheumatoid inflammatory indices also remain normal for several months to years, even in the face of acute synovial involvement. The erythrocyte sedimentation rate is invariably elevated, but other indices such as the rheumatoid factor, the C-reactive protein, and the ASLO titers can be normal. Swann has indicated that rheumatoid factor positive (seropositive) patients tend to show continuing synovitis, whereas periarticular contracture is greater in seronegative patients (211). Approximately 50% of patients undergoing surgical intervention during the growth years were seropositive. The disorder would appear to include several distinct processes. Recognizable subgroups include (1) systemic onset disease (20%), (2) rheumatoid factor negative polyarthritis (25%), (3) rheumatoid factor positive polyarthritis (5%), (4) pauciarthritis associated with antinuclear antibodies and chronic iridocyclitis (30-35%), and (5) pauciarthritis associated with sacroiliitis and HLA-B27 (10-15%) (179). Monoarticular onset occurs in approximately onethird of patients with most subsequently developing additional joint involvement a few months to years later. The monoarticular incidence in separate studies was 31% (32), 32% (52), and 39% (81). Attempts continue to provide rigid classification criteria. The American Rheumatologic Association defines the need

889

for objective arthritis in one or more joints for 6 consecutive weeks plus exclusion of other inflammatory disorders.

2. PATHOBIOLOGYOF RHEUMATOID ARTHRITIS The disorder begins with a synovitis, which is generally associated with an increase in joint fluid (synovial effusion) (27). The swelling and discomfort associated with this generally lead to flexion contractures, muscle atrophy, and, with time, increased periarticular fibrosis. Tendons around the joints can be involved with adhesions. Synovial involvement, which is long lasting, can be followed by synovial fibrosis. The epiphyses themselves are affected secondarily. If the inflammatory process lasts for several weeks or more and is not responsive to treatment, the first plain radiologic change is an osteopenia of the secondary ossification center and adjacent metaphyseal regions. With time, there is a creepage of synovial tissue across the articular surface, which leads to a subsequent degeneration of the articular cartilage. The characteristic fibrovascular material originating from the synovium and passing onto the cartilage surface is referred to as the pannus. With time, there is degeneration of the underlying articular cartilage and narrowing of the joint space. The osteopenia of the bone can be associated with small subchondral cysts. There are no other specific radiographic findings of a juvenile rheumatoid arthritis. The major joint problems relate to continuing joint effusion and synovial proliferation, which in itself can be uncomfortable and can lead to joint contractures and muscle weakness. Treatment is designed first to quiet the inflammatory process and second to rehabilitate the joint in terms of range of motion and associated muscle strength. One of the major orthopedic sequelae of a monoarticular juvenile rheumatoid arthritis in the growing child is the effect on the epiphyseal growth rate. The characteristic responses of the joint in juvenile rheumatoid arthritis have been assessed in our study on lower extremity length discrepancies. The increased vascularity associated with the inflammatory process leads to a stimulation of growth in the affected region and increased length on that side. This is particularly marked in those who have the inflammatory disorder prior to 10 years of age. In those who are affected initially after 10 years of age, the periphyseal vascular stimulation leads to premature cessation of growth and, thus, shortening of the involved limb. The increased vascularity of the joint region may lead to premature appearance of the secondary ossification centers in the involved epiphyses. The age of onset of juvenile rheumatoid arthritis is between 1.5 and 8 years age with a second group showing peak incidence between 11 and 14 years of age. In childhood rheumatoid arthritis, eye examinations are mandatory to assess for iridocyclitis or uveitis. A classic series of criteria defined by Ansell and Bywaters for definitive diagnosis required that the child have a polyarthritis affecting more than four joints for a minimum period of 3 months (5, 6). Synovial biopsy will demonstrate hyperplasia of the synovial

890

CHAPTER 10 9 Metabolic, Inflammatory, Neoplastic, Infectious, and Hemaroloyic Disorders

membrane, increased vascularization, and focal collections of lymphocytes and plasma cells. The high frequency of knee involvement is stressed. Other joints involved less frequently are the wrist, hand, elbow, shoulder, and cervical spine. In the severe variants of Still's disease, the cervical vertebrae may be affected showing limitation in motion and apophyseal joints that may become ankylosed. There is a fairly high incidence of recovery in juvenile rheumatoid arthritis, which differs from the adult disorder. In one large series of cases, nearly one-half of the affected children showed no residual lesions of the involved joints 5 years later. Other papers also report 50% of patients with complete recovery, 25% with a mild residual disability, and 25% further involved with worsening. The articular lesions of JRA are basically the same as those of rheumatoid arthritis in the adult. In the early part of this century, it became evident that a series of growth disorders could be associated with juvenile rheumatoid arthritis. These involved (1) a generalized lack of body growth, (2) a persistence of infantile proportions, and (3) asymmetry of growth. A common problem for the pediatric orthopedic surgeon is the occurrence of lower extremity length discrepancies due either to overgrowth in the first decade of life due to chronic synovitis particularly around the knee or to premature physeal closure also at the knee in the early part of the second decade (192, 228). Both occurrences are presumably due to increased vascular presence in the periphyseal region. Bywaters reported on the pathological aspects of juvenile chronic polyarthritis (27). In the child, the much thicker hyaline cartilage protected the subchondral bone from erosion. Bone overgrowth of the secondary ossification center was common. In later childhood, Harris arrest lines were seen. An excellent study of abnormal ossification patterns in the distal femoral and proximal tibial epiphyses of children with long-standing JRA was performed by Sundberg and Bratstrom (210). They assessed 63 cases ranging in age of onset from 6 months to 13 years. The most common time of onset was in the second year of life and the three highest years involved the ages from 2 to 5 years. With unilateral arthritis, they invariably found that the femoral epiphysis on the arthritic side was larger than that on the healthy side. They concluded that "the difference in size of the femoral epiphyses usually appears within one year of the onset of the disease in some groups and within the second year in others. It often amounts to 10-20% and in isolated instances it can rise to 30%." They are referring to the size of the secondary ossification centers. Differences in length were also seen. These growth abnormalities are magnified by the fact that the knee is the joint most commonly involved in JRA, whether it be pauciarticular, monoarticular, or systemic. After the knee, the wrist is one of the next most commonly involved joints in JRA. It is vulnerable particularly to involvement with systemic and polyarticular onset disorders. Growth defects in this region occur in the ulna and in the fourth and fifth metacarpal bones as a result of premature

epiphyseal fusion and ankylosis. The radiographic findings tend to be more marked than the clinical, and only 6% of patients in a large series showed clinical growth abnormalities. The shortened ulna can lead to ulnar deviation at the wrist, ulnar migration of the carpal bones, and subluxation. Brachydactyly can occur in the hand or foot. Bony overgrowth at the interphalangeal joints can occur. Hip growth abnormalities are the consequence of a chronic synovitis, which itself can lead to diminished walking and progressive hip flexion contractures. Often the proximal femoral epiphysis becomes enlarged (coxa magna) and decreased activity is associated with a coxa valga positioning. Only rarely does premature fusion of the proximal femoral capital epiphysis lead to a coxa vara deformity. 3. SKELETALABNORMALITIES

a. Lower Extremity Length Discrepancies We performed a retrospective study of patients with the monoarticular and pauciarticular forms of juvenile rheumatoid arthritis with particular reference to discrepancies in the lengths of the lower limbs (192). We studied all of the patients (100) with a diagnosis of juvenile rheumatoid arthritis who were referred to the Growth Study Clinic at Children's Hospital, Boston, from June 1941 to June 1979. Patients were referred when a limb length discrepancy had been detected by physical examination. Patients with widely disseminated rheumatoid arthritis were generally not referred to the clinic. Factors that might influence the development and treatment of limb length discrepancies in patients with juvenile rheumatoid arthritis were investigated. The sex distribution, age at onset of disease, joints involved, incidence of monoarticular and pauciarticular disease, disease activity, and treatment were recorded. The skeletal growth was studied, particularly assessing the extent of limb length discrepancies as related to the time of onset of the disease and the course of the discrepancy. A more general review of length discrepancies in JRA is presented in Chapter 8. In assessing the 100 patients, three groups were considered: group I, 36 patients who had been followed until skeletal maturity; group II, 15 patients who had been followed for 4 years or more but had not reached skeletal maturity; and group III, 49 patients who were followed for less than 4 years. Many patients in the third group had been referred for a single evaluation, but their records were used to supplement information. Disease Profile: There were 77 girls and 23 boys. In groups I and II the female-to-male ratio was 42:9 (4.7:1), which corresponds well with the predominant female-tomale ratio in monoarticular and pauciarticular arthritis noted previously. Seventy-two percent of the patients in the entire series and 80% of the patients in groups I and II had onset of the arthritis when they were less than 5 years old, as reported in previous series for more severe forms of juvenile rheumatoid arthritis. An additional 4% had onset of the disease between the ages of 5 and 9 years.

SECTION !1 ~ I n f l a m m a t o r y Disorders

Involvement was monoarticular slightly more often than it was pauciarticular both in the entire series and in groups I and II (a ratio of 1.7:1). Ninety percent of the patients had early involvement of the knee. Although this represents a higher incidence than normally recorded, patients usually were referred to the clinic after a discrepancy in limb lengths had been identified, a finding that is most common in patients who had involvement of the knee. In patients with bilateral disease, one limb generally was involved much more seriously than the other. The knee was involved in 84 of the 100 patients, 65 unilaterally and 19 bilaterally. The ankle was involved in 24 patients, 19 unilaterally and 5 bilaterally. The hip was involved unilaterally in 6 patients and the subtalar joint (foot) unilaterally in 4. For groups I and II, the analogous figures were knee involvement in 48 patients, 43 unilateral and 5 bilateral, and ankle involvement in 13, 10 unilateral and 3 bilateral. Two patients had unilateral subtalar (foot) involvement and 1 unilateral hip involvement. Medical treatment varied according to clinical symptoms and signs and consisted mostly of rest and courses of acetylsalicylic acid. Orthopedic treatment consisted of immobilization by plaster cast or traction and physical therapy. Only 2 patients had been treated with steroids and neither had an extensive limb length discrepancy. Only an occasional patient had an aspiration or synovectomy of the knee. In general, the interval between the onset of the patient's symptoms and the time of diagnosis of arthritis averaged 1.5-2 years, and at the time of diagnosis an initial stimulation of growth on the involved side usually was evident. If the disease exacerbations occurred frequently or were prolonged (after 6 months), and if during such episodes immobilization treatment was not protracted, regardless of the symptomatic severity of the disease, there tended to be growth stimulation on the involved side. When there was lengthy or frequent immobilization treatment during and sometimes after such exacerbations (for example, 1 year or more), there was inhibition of growth on the involved side. Skeletal Growth: Of the 51 patients included in groups I and II, whereas all had some length discrepancy, in 35 (70%) a length discrepancy of 1.5 cm or more developed during the study period. Twenty-one patients had a discrepancy between 2.0 and 2.9 cm and in 3 it was 3.0 cm or more. In patients with unilateral disease who had the onset of the disease before the age of 9 years, the involved side was almost invariably the longer one (39 of 40 cases). Those who had onset of the disease within the first 3 years of life tended to have a discrepancy greater than 1.5 cm (24 of 34 patients) relatively more often than children 3-8 years old, but overgrowth in the younger children never amounted to more than 3.0 cm. When the disease occurred initially after the patient was 9 years old (5 patients), the involved side usually became shorter (with one exception). Regardless of the age at onset, the major discrepancy that developed did so within the first 4 years after onset of

891

the disease. Thereafter, in the group I patients the discrepancy either increased very slowly (6 patients), remained unchanged (14 patients), or decreased spontaneously (12 patients). The other 4 patients, whose discrepancies continued to increase, were patients with late onset of the disease whose epiphyses fused prematurely on the involved side. Continuing involvement of the knee for several years in 1 patient resulted in a continuing increase with time until a prearrest discrepancy of 2.4 cm was reached, but such an occurrence was unusual. In the 12 patients whose discrepancies decreased, there was a gradual inhibition of growth in the involved limb over several years. In 7 of the 12, the discrepancy became clinically insignificant. The changes in discrepancy with time were 2.8 to 0.3, 2.5 to 1.2, 2.3 to 1.3, 2.2 to 0.4, 2.0 to 0.2, 1.8 to 1.4, and 1.5 to 0.9 cm. In one case the discrepancy decreased and reversed the side of shortness from 1.1 cm on the fight to 1.1 cm on the left. Epiphyseal Arrest: Of the 35 patients (groups I and II) having a discrepancy of 1.5 cm or more at some time during the period of assessment, 15 had epiphyseal arrests. Some of those who did not have arrests refused to have the procedure done despite recommendation by the physician. Twelve of the 15 patients who did have an epiphyseal arrest had the onset of disease before the age of 5 years. No patient who had onset of the disease before the age of 9 years had an epiphysiodesis for the inhibition of growth of the involved side. When onset of the disease occurred after the age of 9 years (5 patients), epiphyseal arrest was required on the uninvolved side in 3 patients and will be needed on the ipsilateral side in 1. Fourteen of the 15 epiphyseal arrests were performed on patients with disease at the knee. In patients with the monoarticular or pauciarticular form of juvenile rheumatoid arthritis who have predominant involvement of the major joints of one lower extremity, the length discrepancy occurs as a result of two factors: (1) stimulation of the epiphyseal growth plates, predominantly about the knee joint, during the time the disease is active and for some time afterward and (2) inhibition of the growth potential of the involved extremity. The two epiphyseal growth plates at the knee account for 70% of the growth potential of the lower extremity. They are sufficiently close to the synovial capsule to be affected by the hyperemia that occurs during the inflammatory process but are not adversely affected by the concomitant destructive process. Thus, because the knee is commonly involved in the type of arthritis under discussion and so often is involved unilaterally early in the course of the disease, great potential exists for growth stimulation of the involved lower limb. However, the severity of the disease often causes a decrease in the use of the involved extremity, either because of the patient's symptoms or because of the treatment. Such decreased use can explain the gradual reduction in epiphyseal stimulation and eventual inhibition of growth. When the patient has activity of the disease in early adolescence,

892

CHAPTER IO ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

consequent hyperemia about the knee could cause rapid and premature fusion of one or both of the growth plates, with a sudden decrease in growth of the involved lower limb. In this study, the most common and well-defined pattern of development and progression of length discrepancy was ipsilateral lengthening in patients who had onset of the disease before the age of 5 years. The major part of the discrepancy then occurred within the first few years of the disease. Thirty-nine of the 40 patients with unilateral involvement whose disease began before the age of 9 years showed this pattern. Early in the course of the disease the predominant factor is stimulation of the epiphyses about the involved knee joint. This rapid, early epiphyseal stimulation may not be attributable merely to the young age of the patient but may also be due to the rather long period of time it often takes to make the diagnosis and institute an adequate therapeutic program. In the older child it may be easier to make the diagnosis, and the child may be more compliant in restricting activities and in carrying out the treatment. This, of course, cannot be stated with certainty but is consistent with our results. If true, the importance of early recognition and treatment of the disease can be inferred. The pattern of development of the limb length discrepancy just described most often is followed by lack of a significant continuing increase in the discrepancy. Although occasionally (in 2 patients in this series) the discrepancy will increase over the ensuing years, in most patients it will decrease or remain unchanged with time. The thought that some combination of severity of disease activity, region involved, physician's overall treatment, and patient's compliance adds up to a predictable mode of progression of the length discrepancy is appealing, but factors vary from child to child so greatly that our data are inadequate to prove the point. However, this study does indicate clearly that, in the large majority of patients, significant changes in growth rates are long-term phenomena and are not due to single exacerbations of the disease or to failure of a therapeutic regimen over a short period of time. Rapid premature closure of the epiphyseal growth plates about the involved joint occurred at the end of growth in 4 children in this series. All had onset of the disease after the age of 9 years, and all showed shortening of the involved side. Their length discrepancies ranged from 1.9 to 5.9 cm, with the 2 patients who were 9-10 years old at onset of the disease showing larger discrepancies at skeletal maturity than the other 2 who were 11-12 years old at onset. Early epiphyseal growth plate fusion has been recognized as a complication of juvenile rheumatoid arthritis for many years, although the cause of this premature fusion is unknown. The phenomenon may also occur in patients with the more disseminated form of arthritis, with involvement of multiple joints more frequently than in patients with monoarticular and pauciarticular forms, but if so it usually does not produce a significant limb length discrepancy because of the relative symmetry of joint involvement.

We concluded that an alteration in growth of a lower extremity in a patient with monoarticular or pauciarticular juvenile rheumatoid arthritis is common. When a significant limb length discrepancy develops in a child, the knee joint almost always will have been involved. If the disease begins before the age of 9 years, the involved side will be longer. The maximum overgrowth in this series never exceeded 3.0 cm. In the initial few years following the onset of disease, most of the developing length discrepancy becomes evident. The initial discrepancy, noted within the first 3-4 years after onset, usually remains unchanged or decreases, but on occasion it can increase progressively. The decreases occurred slowly over several years, and in 12 (33%) of 36 patients in group I the decreasing discrepancy rendered epiphyseal arrest unnecessary. No evidence for rapid, definitive premature closure of the epiphyseal growth plate that significantly increased limb length discrepancies occurred in the group that had early onset of the disease. In 4 patients, each of whom had onset of the disease after the age of 9 years, rapid premature growth plate closure was evident, and 2 of them had discrepancies of 5.1 and 5.9 cm. The planning for epiphysiodesis by the Green-Anderson method proved satisfactory. b. Valgus Deformity at Knee The other common skeletal growth manifestation at the knee is a valgus deformity. This can be present during the first decade. An early report of treatment of valgus deformity of the knee in juvenile rheumatoid arthritis by asymmetric stapling was by Laine and Mikkelsen (115). They performed medial distal femoral stapling in 10 patients to treat valgus deformities between 10 and 25 ~. The affected leg was longer so that limb length discrepancy was also treated. The staples were left in until physiologic position was obtained either before or at skeletal maturity. If there was growth remaining, it appears that the staples were removed. In each instance except one, correction into the physiologic range was achieved and no varus overcorrection was noted. Not all patients were followed to skeletal maturity, but the paper indicated that asymmetric stapling was a relatively simple way to correct for angular deformity without resorting to osteotomy. Rydholm et al. also noted that the valgus deformity can be progressive requiting stapling of the physes in asymmetric fashion for correction (176). Their series reported on stapling for lower extremity length discrepancy and for valgus deformity. Seven patients underwent stapling for bilateral knee valgus deformity in polyarticular disease and 3 had the stapling for unilateral valgus deformity (2 polyarticular and 1 monoarticular). In all, 17 stapling procedures were performed for valgus angulation: stapling of the medial distal femur, 10; medial proximal tibia, 6; and both distal medial femur and proximal medial tibia, 1. The average age at surgery was 10.9 years (range = 5-16 years), the preoperative valgus deformity was 18.9 ~ (range = 10-35~ and the postoperative result was correction to 4.7 ~ (range = 0-20~ Two or three Blount staples per side were used. When angular deformity is seen in the rheumatoid knee, it almost invariably is

SECTION I! ~ Inflammatory Disorders

valgus. On occasion, there is an associated flexion contracture, although most of this is attributed to soft tissue involvement posteriorly rather than to physeal growth asymmetry. The valgus is attributed to some variable form of abnormal or asymmetric vascular stimulation. Brattstrom concluded that hyperemia in the femoral condyle is more pronounced medially than laterally because of asymmetry of the vascular supply (24). Other theories have involved accelerated metabolism with increased local temperature and oxygen tension, changes in regional blood flow following elevation of the intra-articular pressure, and hyperemia in the joint capsule and femoral metaphysis. c. Hip Joint I n v o l v e m e n t The frequency of hip involvement in JRA is approximately 20%. Hip involvement with JRA can also lead to abnormal growth phenomena. Kobayakawa et al. pointed out the occurrence of femoral head necrosis in JRA and the fact that its extent has been previously underappreciated (109). The early radiographic findings involve osteopenia and growth disturbance of the femoral capital epiphysis, with the trochanter continuing to grow and later findings showing premature closure of the proximal femoral capital epiphysis. A coxa valga deformation of the proximal femur with lateral subluxation and necrosis of the femoral head may also occur. This has been reported previously but generally in association with steroid therapy. Kobayakawa et al. found "obvious radiographic signs of necrosis of the femoral h e a d . . , in 10 hips in 6 children with a mean age of 3 (1-6) years at the onset of the disease." There were suspicious signs by radiography in 20 hips in 13 children and no radiographic signs of necrosis in 42 hips in 23 children. The mean age at the onset of hip pain was 3, 6, and 9 years, respectively. All of the children with necrosis had been on long-term (more than 3 months) treatment with corticosteroids, although 8 of 11 and 10 of 18 in the other two groups also had been treated with steroids. In addition to the known growth disturbances of the proximal femoral capital epiphysis and acetabulum, which are lateral subluxation of the femoral head, subchondral erosions, joint space narrowing, and acetabular protrusion, this group adds necrosis of the secondary ossification center to the list of sequelae. The changes were similar to those observed in Legg-Perthes disease. Synovial effusion with an increase in intracapsular blood pressure might compromise the blood supply by a tamponade mechanism. Hip traction might contribute to the increased intra-articular pressure when, in the presence of a synovitis, the hip was placed in the less desirable extended and internally rotated position. Flexion and internal rotation are known to decrease the intracapsular pressure. Long-standing changes in the hip in persisting JRA were also seen, however, in a second study of 36 children with no significant steroid therapy by Blane et al. (21). Sixty-eight patients with hip involvement in JRA were followed for at least 2 years. Approximately 50% (36/68) showed bone and cartilage changes at the hips. They defined changes as pri-

893

mary and secondary, with the primary disorders being due to inflammatory destruction of the hip with erosive arthritis and progressive joint space narrowing. The adaptive developmental changes included changes in position from mild valgus to subluxation to dislocation, flattening of the femoral capital epiphysis with wedging of the medial portion and squaring of the lateral portion, acetabular dysplasia, and secondary osteoarthritic changes involving acetabular sclerosis and femoral pseudo-cyst formation, joint space narrowing, and osteophyte formation. Patriquin et al. noted similar long-term radiographic changes of the hip in JRA in a smaller series (151). They felt that the changes were due to persisting synovitis of the hip joint, leading to ischemia of the femoral head secondary to joint distention by both fluid and synovial tissue. Changes noted included coxa magna, a short femoral neck, subluxation of the head combined with cystlike erosions of the femur and acetabulum, a widened femoral head in the absence of steroid therapy, and joint space narrowing due to destruction of the articular cartilage. d. Joint Contractures With severe and frequent joint involvement with poor response to medical management, there is a tendency for the patient to remain nonambulatory for extended periods of time. Joints assume the most comfortable position, invariably in flexion, and persistent contractures can develop. The most common patterns are knee flexion contractures, which can progress to associated posterior subluxation of the tibia, and flexion-adduction contractures of the hip. 4. SURGERY IN JUVENILE RHEUMATOID ARTHRITIS Surgery has been used in an effort to control the synovitis when medical treatment has failed, to correct angular deformities, and to equalize limb length discrepancies. In a rough estimate provided by a large rheumatology unit in Great Britain, approximately 10% of their JRA patients had open surgical procedures (211). Surgical rehabilitation in these patients is difficult because of the intra-articular synovitis, periarticular fibrosis, osteopenia, and muscle weakness. After medical regimens involving rest, physical therapy, splinting, pain medication, and anti-inflammatory medication have either failed or reached a plateau, surgical measures may be needed. Swann and colleagues strongly stress the value of soft tissue releases as the initial surgical treatment. These include soft tissue release procedures at the hips, knees, and other sites. At the hip, adductor and psoas tenotomies are performed. Much of the benefit appears to derive from decreasing the intra-articular pressure and diminishing the pain. If there has been angular deformity, osteotomy may be required. Osteotomy requires more prolonged immobilization postsurgery and is made more difficult by the invariable presence of osteoporosis particularly in a disorder sufficiently marked to lead to the deformity requiring the surgical intervention. Synovectomy is resorted to relatively infrequently in children in comparison to its

894

CHAPTER 10 9 Metabolic. Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

more extensive use in adult rheumatoid disease. The exception in the opinion of Swann is pauciarticular arthritis. In pauciarticular arthritis, the knee is frequently affected and tends to remain chronically active with warm, thickened synovium and excess fluid. This is an indication for synovectomy and has been found to be advantageous in a number of cases. Hip and knee destruction is occasionally so severe that total joint arthroplasty is resorted to late in the second decade of life.

the epiphyseal region because problems relate not only to the primary lesion but also to possible growth abnormalities due to their presence in growth-sensitive areas. There is relatively tittle primary involvement of the epiphyses with either benign or malignant neoplastic disorders. When there is neoplastic involvement of the epiphyses, however, it can be primary or secondary involving transphyseal spread from a focus of metaphyseal origin.

A. Primary Involvement of Epiphyseal Regions B. Pigmented Villonodular Synovitis Pigmented villonodular synovitis is a monoarticular proliferative process of the synovium characterized by a fibrous stroma (65, 97, 104). It can affect the synovium or the tenosynovium. The deposition of hemosiderin, histiocytic infiltrates, and giant cells occurs within the synovial membrane. The most commonly involved joint is the knee, with tenosynovial involvement most common in the hand and wrist. There are two basic patterns of involvement, one being localized or pedunculated and the other diffuse. The disorder usually presents with a swollen knee of gradual onset, mild discomfort, a boggy synovium, and the presence of reddishstained fluid on aspiration. It commonly affects adults in the third and fourth decades but is seen in children on occasion in the second decade, with most diagnosed around the time of fusion of the epiphysis or just afterward. The involvement of multiple joints has been reported in both children and adults but in less than 1% of cases. In two children, one had involvement of three joints and the other five (104). The plain X ray in the childhood group is almost invariably normal in relation to the bones and will show only the soft tissue swelling if great enough. Cysts or erosions of bone occur in the adult form. The synovial fluid has a deep xanthochromic hue to a bloody dark brown color and has the appearance of a chronic hemorrhagic effusion. The villi are long and tend to become interwoven with each other. On histology, the synovial lining cells are pigmented and 1-3 layers in depth. The treatment varies depending on the age of the patient and the extent of involvement. In localized pigmented villonodular synovitis, excision of the abnormal focus by either arthroscopic or open arthrotomy is recommended. Arthroscopic synovectomy is found to be slightly superior to open synovectomy in the diffuse form. Marginal excision is performed for localized disorders. Radiation therapy and radiation synovectomy have been used for the more diffuse variants but are rarely indicated today.

llI. NEOPLASTIC DISORDERS OF EPIPHYSES Neoplastic disorders of the developing skeleton are relatively common. Both benign and malignant disorders occur. A detailed description of these is beyond the scope of this book, but we will discuss briefly those disorders that involve

1. CHONDROBLASTOMA The one primary tumor of epiphyseal regions in the growing child is the chondroblastoma. This has its origin in the epiphyseal cartilage and is the most common primary tumor of this region of the developing bone (139). It was defined initially by Jaffe and Lichtenstein in 1942 as a benign disorder (98). Codman had drawn attention to the disorder previously as a variant of a giant cell tumor in the epiphyses of the proximal humerus (98). Fifty percent are at the distal femur or proximal tibia and two-thirds involve the lower extremities (92). The distal femur is the most common site in some series with the proximal tibia next, followed by the proximal humerus, distal tibia, and proximal femur. The most common areas in other series involve the proximal humerus (22%), distal femur (16%), proximal tibia (14%), and proximal femoral capital epiphysis and greater trochanteric epiphysis (8%). Seventy-five percent of the patients are teenagers and 80% are between 5 and 25 years of age (92). In a large series of 465 patients, 65% were between the ages of 10 and 20 years with 4% occurring in the first decade of life (139). Forty percent of chondroblastomas involve only the epiphysis, 55% extend into the adjacent metaphysis, 4% affect the metaphysis alone, and on occasion diaphyseal lesions have been reported (92). The lesion is lytic in radiographic appearance and will involve the secondary ossification center. Typically they are benign and nonaggressive and are treated by curettage with or without bone grafting. The tumors tend to be ovoid or rounded, sharply marginated, and in the 1- to 7-cm range in diameter. There is thickening of bone trabeculae around the lesion due to its slow and nonaggressive growth. Two tissue types are seen, one involving very cellular tissue with little matrix and the other in which the tissue is less cellular and surrounded with a chondroid matrix. The proportion of these tissue types varies from patient to patient and even in different areas of the same tumor. The chondroblastomas are almost always surrounded by at least a thin shell of reactive bone. With chondroblastomas, less than 50% of the total lesional tissue is cartilage or chondroid. The lesion contains chondroblasts and osteoclast-like giant cells. All chondroblastornas produce cartilage or chondroid tissue, and the lesion must contain islands of primitive chondroid, fibrochondroid, or hyaline cartilage. The lesion is eccentric in the epiphysis and usually involves less than one-half of its extent. As is frequent with many cartilage tumors, calcific densities within the tumor are often seen.

SECTION III ~ Neoplastic Disorders o f Epiphyses

Small punctate calcifications are evident in slightly more than one-half the cases. An open physis is usually present. It is extremely important to consider the age of the patient in diagnosing the lesion because the most common tumor of epiphyses in adult age groups is the neoplastic giant cell tumor, which occurs in older patients in whom the epiphyseal growth plate is closed. The giant cell tumor lacks calcification, occupies more than one-half of the epiphysis in most cases, and has no sclerotic border (139).

2. OSTEOIDOSTEOMA Intra-epiphyseal osteoid osteoma, although rare, has been documented on several occasions. The lesion is the same as those that present in the metaphyseal and diaphyseal regions. Diagnosis is delayed frequently because of the unusual position, and surgical removal is more problematic because of the location, which can be close to either the physis or the articular cartilage. Of 12 intra-epiphyseal osteoid osteomas described up to 1989, 7 were present in the distal femoral epiphysis, 2 in the distal tibial epiphysis, and 1 each in the distal radial, proximal tibial, and greater trochanteric epiphyses. The differential diagnosis includes chondroblastoma, enchondroma, a solitary bone cyst, eosinophilic granuloma, and subacute osteomyelitis. The age range is from 8 to 16 years. Ten of the cases were in males, with only 1 female and one report not mentioning gender. The time to diagnosis is often delayed not only by the nature of the lesion itself but also by its position. The range has varied from 2 to 24 months. The large majority of lesions have been in the lower extremity. The bone scan is extremely helpful in diagnosis and localization. Treatment has been by complete excision or by curettage in those regions close to the physis. Blair and Kube (19) reported an osteoid osteoma in a distal radial epiphysis in a 15-year-old boy, Beerman et al. (16) in a 14-year-old male's proximal tibia, Kendrick and Evarts (106) in the distal femur (no age given), Odaka et al. (144) the distal femoral epiphysis in a 9-year-old male, Simon and Beller (193) in a 12-year-old male in the distal tibial epiphysis, Sherman (187) involving one of the greater trochanteric apophyses of an 8-year-old girl, Seitz and Dick (186) in an 8-year-old girl's distal femur, Micheli and Jupiter (135) in the distal femoral epiphysis of a 15-year-old male, Iceton and Rang (95) in the distal femoral epiphysis of an 11-year-old male, and Kruger and Rock (114) in the distal femoral epiphysis of a 16-year-old male, and Van Horn and Karthaus (221) added two additional cases of epiphyseal osteoid osteoma, one involving an 11-year-old male with a lesion in the lateral part of the distal femoral epiphysis and the other a 15-year-old male with a lesion in the distal tibia. Mirra notes that, in rare instances, if the nidus is located near a growing epiphysis, it may cause an acceleration of bone growth (139). Overgrowth phenomena have not been reported with an intra-epiphyseal osteoid osteoma, but some authors have commented on this phenomenon with metaphyseal lesions particularly when the symptoms are longstanding prior to diagnosis in those with several years

895

of growth remaining. Habermann and Stern described an osteoid osteoma of the tibial metaphysis occurring in an 8-month-old male (82) with slight overgrowth. 3. EOSINOPHILIC GRANULOMA Eosinophilic granuloma is a benign disorder characterized radiographically by a lytic lesion of bone. The histopathology is characterized by two predominant cell types, eosinophils and mononuclear histiocytes. The disorder may be single or multiple, but occurs mainly in the shafts of long tubular bones or in the skull, ribs, or vertebrae. The occurrence in growing epiphyses is rare, but 10 cases were described up to 1985 (119, 220). Of the 10 patients reported, 5 had solitary and 4 multiple lesions. The lesions were radiolucent and well-circumscribed and on occasion demonstrated a thin bone rim of sclerosis within the secondary ossification center. They could be either central or peripheral. In those lesions described and detailed, the age range varied from 2.5 to 8 years with a mean of 5 years. Six of the 7 cases in which gender was indicated were male. In the 10 cases, the bones involved were the proximal femur (6), distal femur (2), proximal tibia (1), and distal tibia (1). This corresponds with the overall demographics of eosinophilic granuloma, which is found more frequently in males and has its peak incidence between the ages of 5 and 10 years with overall concentration in the first two decades of life. Leeson et al. described three additional cases (119). Their lesions involved a male 3 years 5 months of age with a lytic defect in the femoral head that appeared to extend across the physis into the proximal femoral metaphysis, a female 2 years 7 months of age with a circumscribed lytic lesion in the medial distal tibial epiphysis with subsequent transphyseal spread to the metaphysis, and a 7-year-old boy also with involvement of the proximal femoral epiphysis with a wellmarginated round, radiolucent lesion without transphyseal extension. Two months later, the lesion had crossed the physis into the neck region. Serial X rays indicate that the disorder begins in the secondary ossification center of the epiphysis and then with time passes in a transphyseal fashion toward the metaphysis. Leeson et al. reported that 5 of the 10 lesions in the literature demonstrated transphyseal extension with each of the 5 considered to be epiphyseal in origin. Treatment involves open biopsy because of the essential need to establish the diagnosis followed by a curettage. Bone grafting may be required for larger lesions (202). Clearly, efforts to protect the physis are mandatory.

B. Secondary Involvement of Epiphyseal Regions from Primary Metaphyseal Foci of Benign and Malignant Disorders 1. UNICAMERAL BONE CYST The most common site of the unicameral bone cyst is the proximal humeral metaphysis. In rare instances, apparently spontaneous passage of these lesions through physeal tissue into the epiphysis has been reported. Occasional examples

896

CHAPTER IO 9 Metabolic, intlammatory, Neoplastic, Infectious, and Hematologic Disorders

of epiphyseal cysts with growth deformity in the absence of any treatment have been reported. In these, however, the likely explanation is the occurrence of fracture of the cyst and adjacent physis allowing for either premature closure of the physis due to the fracture with subsequent spread of the cyst or spread of the cyst in relation to transphyseal discontinuities. Somewhat more common is damage to the presumably noninvolved physis during the course of treatment of active lesions fight against the physis, involving either curettage or injection of such materials as steroids. Following growth plate damage, the cyst subsequently can be seen to be present within the epiphysis. Cysts are described occasionally within the epiphyseal region after skeletal maturity, but in the absence of growth deformity they would be postulated to have occurred quite late. Nelson and Foster reported 5 cm of humeral shortening in a 15-year-old female with a large proximal humeral bone cyst, which had expanded into the epiphysis and was associated with an almost fully closed physis (142). They felt that a proximal humeral injury a few years previously represented a pathological fracture of the cyst with physeal damage. Cohen presented a case of premature proximal humeral physeal closure with a cyst treated nonoperatively and reviewed the pathogenesis extensively (40). Moed and La Mont presented three convincing cases of premature physeal closure of the proximal humerus, each associated with a fracture of the cyst but with no surgical intervention (104). Malawer and MarNe reported a unicameral bone cyst of the proximal femoral metaphysis communicating across the physis with the secondary ossification center of the epiphysis in a 14-year-old male (127). The growth plate was open. They reviewed previous reports, not including those of Moed and La Mont, and concluded that only 7 cases of epiphyseal involvement with a unicameral bone cyst had been described. Even in these most of the patients were older. Of the 8 lesions including their own, 3 involved the proximal femur and 1 the distal femur, with 2 each in the proximal humerus and proximal tibia. One cyst in a 6.5-yearold female in the proximal tibia subsequently healed with no shortening or deformity. The next youngest patient was a 15-year-old female whose proximal humeral physis was closed but who did have 2 cm of shortening, leading to the conclusion that the disorder was indeed present during the growing years. 2. ANEURYSMAL BONE CYST Aneurysmal bone cysts in skeletally immature patients almost invariably involve the metaphyseal regions, with transphyseal spread into the epiphysis quite rare. One reported case of spread from a metaphyseal focus through the growth plate into the secondary ossification center of the epiphysis was in a 10-year-old girl at the proximal fibula (130). Following removal of the entire proximal end of the fibula, gross hemi-dissection revealed blood-filled spaces with obvious cystic extension through the center of the growth plate to replace most of the fibular ossification center.

The growth plate itself had decreased organization in the hypertrophic zone. The central quarter of the plate had been destroyed but the periphery persisted, although it was somewhat attenuated and abnormal. The growth plate was relatively intact even immediately adjacent to the area being bridged by the cyst. Once within the secondary ossification center, however, the cyst lining had completely displaced the adjacent bone plate and the epiphyseal blood supply along the germinal zone of the physis. There was only a small area of transphyseal bone bridging. Reviews of large series generally show no lesions crossing the physis in skeletally immature patients. Carlson et al. reported 1 patient out of 26 with an aneurysmal bone cyst crossing the proximal femoral physis to extend into both the capital femoral and greater trochanteric epiphyses (29). Dyer et al. described physeal bridging in a metatarsal cyst (49). They also reviewed the English literature over a 25year period and found 223 cases of primary aneurysmal bone cyst in which the status of the growth plate was adequately described. In none of these was physeal bridging noted. 3. OSTEOGENIC SARCOMA Although the large majority of metaphyseal osteosarcomas had long been considered to remain within the metaphyses, there have been several well-documented cases of transphyseal spread. Simon and Bos reported a retrospective macroscopic and microscopic review of 26 macrosections of entire metaphyseal, growth plate, and epiphyseal tissues from patients with osteosarcoma who had been treated between 1926 and 1960 (191). Their study required the availability of at least one good macrosection of the whole tumor, which contained a well-visualized and active physis from a patient with high-grade metaphyseal osteosarcoma. The results in the 26 patients showed that in only 3 was the physis not perforated by tumor tissue. In 4 there was only microscopic perforation, but in 19 there was transphyseal spread evidenced by both microscopic and macroscopic examination. In some instances, there was more than one site of extension of the metaphyseal tumor into the ossific nucleus. Metaphyseal spread occurred through the center of the physis in 12 cases, the periphery in 9, and the region beneath the perichondrial ring into the margins of the ossific nucleus in 6. The tumor tissue invariably perforated the physis along with a series of vascular channels. In some instances, the physis was intact except for the transphyseal vascular penetration, but in others it was focally destroyed by the tumor mass. The authors concluded that the physis appeared to have a tendency to act as a barrier to tumor spread but did not restrain the tumor in most cases, such that epiphyseal extension of the tumor tissue was extremely common in general and was massive in more than one-half of the cases. These findings were similar to an assessment by Enneking and Kagan, who found that 21 of 28 skeletally immature patients had physeal perforation by osteosarcoma tissue in the proximal tibia and distal femur (56). They define an open physis as regularly oriented cartilaginous cells with the typ-

SECTION IV ~ Osteomyelitis and Septic Arthritis

ical zones of an epiphyseal plate and more than six cells comprising the zone of hypertrophy. They defined a closing plate as having histologic evidence of decreased cellular proliferation and hypertrophy with plate narrowing and some early transverse osseus trabeculae. Of 24 cases with open plates, only 7 showed no evidence of tumor in the epiphysis with 17 showing tumor, which was anywhere from microscopic to 100% invasion of the epiphysis with an average roughly calculated at 25 %. Of the 4 cases with closing plates, all showed involvement of the epiphysis with tumor. There were two patterns of transphyseal tumor spread. The most common was direct spread across the central portion of the growth plate with tumor proliferating through preexisting vascular channels. The adjacent cartilage was not destroyed in early instances but was at later time periods. The ability of the physis to at least temporarily contain the tumor was shown by the fact that in such cases the opposing metaphysis was almost entirely filled with tumor. On occasion, the tumor extended beneath the perichondrium, circumvented the plate, and penetrated the lateral margin of the epiphysis. Enneking and Kagan also pointed out that the then current imaging methods involving X rays and bone scans frequently were inaccurate in assessing transphyseal spread. Use of CT and MR imaging has improved assessment. A third study by Ghandur-Mnaymenh et al. also found transphyseal spread of osteosarcoma to be extremely common (68). In 14 cases of osteosarcoma of long bones in patients ranging in age from 10 to 19 years with open physes, the physeal cartilage was not crossed in only 2 cases. In 12 cases, the tumor crossed the cartilage plate partially or completely, remaining in the epiphysis in 7 and actually crossing the articular cartilage and involving the joint in 5. The incidence of transphyseal spread in the three published series (total of 64 cases) was actually 81.2%, leading to the conclusion that transphyseal spread of osteosarcoma is the rule rather than the exception. Histologic sections showed that increased vascularity on the metaphyseal side occurred initially followed by osteoclastic and chondroclastic activity at the physeal plate with subsequent tumor tissue invasion. Blood vessels appeared to be part of the tumor stroma, extending deeply into the cartilage with progressive advance of the osteosarcoma through the plate. When limb saving procedures are planned in association with chemotherapy, the importance of careful assessment of the possibility of transphyseal spread by CT and MR imaging is evident.

IV. O S T E O M Y E L I T I S A N D SEPTIC ARTHRITIS

A. Primary Subacute-Chronic Epiphyseal Osteomyelitis Epiphyseal osteomyelitis occurring as a primary entity is a well-defined, although infrequent occurrence. The common

897

site for long bone osteomyelitis in the growing child is metaphyseal. Episodes of primary epiphyseal osteomyelitis tend to be subacute or chronic in their presentation. As with most osteomyelitis disorders, they occur at those regions in which growth is most active, these being the distal femur, proximal tibia, distal tibia, and proximal humerus. They present with discomfort, a slightly decreased range of motion, a several-day to several-week history, and a low-grade temperature elevation. The primary abnormality on blood testing is an elevated ESR and perhaps a slightly elevated white blood cell count. There may be atrophy of the muscles surrounding the joint. The main radiographic finding is a focal, relatively circumscribed, and generally eccentric region of lysis within the secondary ossification center. Depending on the degree of chronicity, there may be some surrounding sclerosis, but this is usually not marked. King and Mayo reported 2 cases of primary subacute epiphyseal osteomyelitis, one in the distal femoral epiphysis and the other in the distal tibial epiphysis (108). Green et al. reported 8 cases, 1 of which was in the proximal femoral epiphysis, 5 in the distal femoral epiphysis, and 2 in the proximal tibial epiphysis (78). All of their patients were less than 5 years of age. Andrew and Porter reported 3 patients with localized primary subacute epiphyseal osteomyelitis (4). These were in a 4-year-old girl in the distal femoral epiphysis, a 3.5year-old girl in the distal femoral epiphysis, and a 9-year-old girl in the distal tibial epiphysis. In none of the previously mentioned 13 cases was there any long-term damage to the epiphysis or adjacent joint. Kramer et al. have also reported osteomyelitis of the distal femoral epiphysis in an 11-yearold boy with a slightly more acute onset (113). Published series show the same distribution of organisms as is found with the more common metaphyseal osteomyelitis (121). The most common infections are thus caused by Staphylococcus aureus with occasional examples of Hinfluenza and Streptococcus. Treatment is by curettage and intravenous antibiotics followed by oral antibiotics. Great care must be taken not to damage the physis surgically. These foci of infection tend to heal uneventfully with no serious long-term sequelae reported. Figure 8 shows how the epiphyseal foci of subacute or chronic osteomyelitis can be primary or secondary to transphyseal spread from the metaphysis.

B. Epiphyseal Osteomyelitis Secondary to Transphyseal Spread from Subacute or Chronic Metaphyseal Foci Epiphyseal osteomyelitis occurs more commonly secondary to a primary subacute or chronic metaphyseal osteomyelitis, which has undergone transphyseal spread (Fig. 8). The subacute or chronic metaphyseal focus is also referred to as a Brodie's abscess. Harris and Kircaldy-Willis demonstrated a subacute pyogenic abscess crossing the epiphyseal plate of the distal tibia in a 13-year-old child (85). Robertson published 8 cases of epiphyseal osteomyelitis, most of which

898

CHAPTER IO 9 Merabofic, Inflammatory, Neoplastic, Infectious, and Hematoloyic Disorders

Chronic/Subacute Epiphyseal Osteomyelitis

Secondary to Transphyseal/~~ Spread

Primary

FIGURE 8 Epiphysealinvolvement in subacute and chronic osteomyelitis can be either primary or secondaryto transphyseal spread from a metaphyseal focus.

appeared to have larger metaphyseal components (168). He defined involvement of 2 in the distal tibial epiphysis, 2 in the distal lateral femoral epiphysis, 2 in the distal fibular epiphysis, and 1 each in the distal ulnar and terminal phalangeal (finger) epiphyses. He clearly defined the involvement of epiphyseal cartilage and secondary ossification center bone. Gledhill reported transphyseal spread of a distal tibial metaphyseal subacute osteomyelitis across the physis into the epiphysis (72, 73). Examples of metaphyseal abscesses crossing the epiphyseal plate were also described by Gillespie et al.; 2 of 57 Brodie metaphyseal abscesses crossed the epiphyseal plate, whereas 2 of their 110 cases appeared to be primary within the epiphyseal bone (71). Roberts et al. described a subacute osteomyelitis of the lateral proximal femur that crossed the physis, passing into the greater trochanter in a 2.5-year-old girl (166). The relative infrequency of epiphyseal involvement in chronic and subacute osteomyelitis, however, is shown by the paper of Boriani, who reported a carefully documented series of 181 cases of Brodie's abscess with no cases in which the lesion involved the epiphyseal plate (23). Other detailed reviews of foci of chronic circumscribed osteomyelitis that traversed the epiphyseal plate have been presented. Kandel and Mankin presented 9 patients with subacute or chronic pyogenic abscesses of the long bones with circumscribed lesions in the metaphyseal regions (100). Three of these were found to traverse the epiphyseal plate, allowing for intra-epiphyseal involvement, but it was noted that in no case did any negative growth sequelae occur, as evidenced by the absence of stimulation, retardation, or angular deformity. Two of those 3 patients were sufficiently old at 15 and 16 years of age that growth sequelae would not have been expected, but they did have symptoms for 6 and

15 months, respectively, prior to diagnosis. Even when the epiphyseal plate was crossed in a 10-year-old child growth sequelae did not occur, which was attributed to the fact that it was only a narrow region of the central portion of the physis that had been damaged. Bogoch et al. presented 6 cases of localized transgression of the epiphyseal plate in patients with considerable growth remaining, with no subsequent growth disturbance seen (22). Surgery was performed in 4 of the cases, whereas medical treatment only was used in the other 2. Most episodes of transphyseal transgression occur near the end of growth such that the lack of growth sequelae can be understood. Four of the cases in this series, however, occurred in children less than 10 years of age. Jani and Remagen used the term primary chronic osteomyelitis to describe 3 cases of transphyseal spread of metaphyseal foci into the epiphysis, where again no negative growth sequelae occurred (99). Even though there may be large lytic defects in both the metaphyseal and epiphyseal bone, the passage through the physis itself appears to be through a relatively small and even narrow focus such that cure of the infection appears to leave the growth potential intact. The factors to be considered for making an assessment of growth potential following transphyseal spread involve the age of the patient, the position of the transphyseal spread, and the extent of the transphyseal defect. The first two are readily ascertainable, whereas the latter would require MR imaging.

C. Acute Neonatal-Infantile Osteomyelitis and Its Damaging Effects on Epiphyses Great damage to the epiphyses has been documented, with osteomyelitis of an acute nature in the infant. The epiphyses can be damaged extensively by acute infection of the bone, which occurs in a highly specific time frame, namely, the neonatal period particularly concentrated from birth to 4 weeks of age. The use of the term infantile to qualify the age at which osteomyelitis occurs is variable but generally refers to the first year of life. The younger the patients, in general, the worse the prognosis for epiphyseal damage. Damage occurs by two specific mechanisms, which are often combined in the same region (Fig. 9). The first involves transphyseal spread of infection from its metaphyseal localization through the physis and into the epiphysis in association with persisting open transphyseal vessels left over from the fetal developmental period. The second mechanism involves a septic arthritis in association with metaphyseal osteomyelitis, which can occur following the passage of organisms from the metaphysis to the joint in intracapsular epiphyses or from the epiphyses into the joint after transphyseal spread. On occasion, acute primary septic arthritis appears to occur in which the vascular spread of the organism is initially to the synovium and then into the joint. Each of these mechanisms will be referred to in the subsequent discussion, which will focus on septic arthritis of the hip of

SECTION IV 9 Osteomyelitis and Septic Arthritis

Acute Neonatal/Infantile Osteomyelitis with Proximal & Distal Epiphyseal Damage

Transphyseal Spread

decreased motion of the extremity. Radiographic changes of the bone were often dramatic at the time of initial presentation with marked femoral periosteal elevation, soft tissue swelling, localized bone resorption, and sub-periosteal new bone formation. In many instances the joint space was widened, and there was often frank subluxation or dislocation particularly at the hip and knee. They also reviewed several studies of neonatal osteomyelitis published between 1940 and 1965. Most of these identified the fact that multiple bones were involved, and three of the studies mentioned associated joint involvement (septic arthritis): Cass (31), in which joint involvement was seen in each of 3 cases, Lindblad et al. (122) with the joint involved in 12 of 29 cases, and Thomson and Lewis (216) with the joint involved in 2 of 4 cases. 1. ACUTE SEPTIC ARTHRITIS OF THE HIP OF INFANCY Septic arthritis of the hip of infancy is a rare but potentially devastating disorder, which can occur in one of three ways: (1) from spread of a metaphyseal focus of osteomyelitis through the physis into the femoral head and then into the joint; (2) from spread from an osteomyelitic focus within the neck directly into the joint cavity leading to an immediate septic arthritis; and (3) due to primary hematogenous joint infection without initial foci of adjacent osteomyelitis. Septic arthritis of the hip has been recognized for several decades to have severe growth sequelae. In the preantibiotic era reviews of 113 cases by Badgley et al. (10) and 65 cases by Slowick (195) showed major complications to include pathological dislocation of the hip joint, avascular necrosis (sequestration) of the femoral head, epiphyseal separation, resorption (disappearance) of the head, bony ankylosis, fibrous ankylosis, and shortening. These sequelae were seen throughout the childhood years. With improved antibiotics and earlier diagnosis, the most severe problems are now concentrated in those with neonatal sepsis and those affected within the first 2 or 3 years of life. The one region in which transphyseal involvement is commonly documented is at the proximal femur during the first year of life. The spread is attributed to the persistence of transphyseal vessels from the fetal period, which allow for communication between the metaphysis and the epiphysis. It is for this reason that primary proximal femoral osteomyelitis in the neonate is potentially so severe. Spread of infection can occur into the epiphysis, damaging the growth plate, and through the lateral walls of the neck passing directly into the joint, leading to a septic arthritis because the epiphysis at the proximal femur is entirely intracapsular. After the first year of life, there is little tendency to transphyseal passage of infection, although the risk of septic arthritis of the hip in association with proximal femoral osteomyelitis remains high because of the passage of organisms through the thin cortex of the neck, which leads to their immediate presence within the joint. a. G e n e r a l O v e r v i e w

,

:> Femoral Neck to Joint Cavity

___>

Femoral Head j to Joint C a v i t ~

y2 The long-termeffects of an acute neonatal-infantile osteomyelitis of the proximal and distal epiphyseal-metaphysealregions of the femur are shown.

FIGURE 9

infancy, neonatal involvement of the distal femoral epiphysis and knee joint, and sepsis of other growth regions. Definitive awareness of the problems of neonatal osteomyelitis in relation to epiphyseal damage accumulated only gradually. Prior to the development of antibiotics, neonatal osteomyelitis, by which is meant infection in the first 4 weeks of life, frequently resulted in death due to the severity of spread and frequent multibone involvement (10, 20, 79). Green and Shannon noted, however, in their series of infants with osteomyelitis under 2 years of age published in 1936, that there had been joint involvement in all but one instance in which there was eventual residual deformity (79). Weissberg et al. studied 17 infants, 1-28 days of age, with osteomyelitis and noted that joint involvement was present in 12 of the 17 and that eventually bone deformity was present in 10 patients, most with a significant functional deficit (225). In 14 of the 17 infants, the illness was not accompanied by systemic signs or symptoms and delay in diagnosis was usually seen because help was not sought until local signs became apparent. These involved swelling, discomfort, and

899

900

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hemaroloyic Disorders

The presence of transphyseal vessels in the human as a remnant of epiphyseal development remains somewhat poorly documented. There is considerable variability throughout vertebrates in terms of the persistence of these vessels. They are only infrequently seen in murine species and the rabbit but are commonly seen in the pig and lamb even several months after birth. Such vessels have been documented commonly in humans and other species in the late fetal and early postnatal time periods (186). The damage done to the epiphyseal apparatus of the femoral capital epiphysis by infection can be extensive and can occur rapidly over several hours to a few days in childhood septic arthritis of the hip. Diagnosis is frequently delayed for several days in those children affected in the first few months of life because they almost always are afebrile with little tendency to leukocytosis or elevated ESR values. The most common mode of presentation in this age group is listlessness and decreased use of the extremity on the involved side, referred to by some as "pseudo-paralysis." A study separated those having the disorder from 0 to 4 weeks of age from those from 1 to 3 years of age. The prognosis was clearly worse in the younger age group and also in those associated with osteomyelitis of the proximal femur. In a preantibiotic era 1936 study, Badgely reported a 12% mortality rate in septic arthritis of the hip with only 6.2% having normal hips (10). Blanche studied osteomyelitis in the first year of life in 25 patients in whom a bacteriologic diagnosis had been made between the years 1934 and 1950 (20). The age range was from 14 days to 9.5 months, but 23 of the 25 were seen within the first 2 months of life, 13 under 1 month of age, and 10 from 1 to 2 months of age. Many patients had no fever at all and the systemic temperature was rarely over 100~ Local symptoms predominated involving swelling, loss of function of an extremity, and discomfort with movement. In the 25 patients, 13 had a single focus of infection and 12 had multiple foci. The femur was the most common site by far with 22 of 50 known lesions present there. The next most common bone involved was the humerus with 7 followed by the tibia with 5 and lesser involvement of other bones. Proximal involvement of the femur was seen in 16, and of these there was an associated septic arthritis of the hip in 11 cases. Distal femoral involvement was frequently associated with growth damage to the distal physis and epiphysis. Intra-articular infection rapidly destroyed the cartilage and frequently caused disruption of the joint, pathological dislocation, and serious proximal femoral growth disturbances. The primary prognostic feature in terms of ultimate normal development and function is the quickness with which treatment is instituted. In the study, 16 neonates were affected from 0 to 4 weeks of age, of whom 11 were premature. In 14 of the 16 patients, the septic arthritis was secondary to an osteomyelitis close to the joint, with 10 patients having primary osteomyelitis of the femoral neck and

the other an osteomyelitis of the acetabulum. In the second group of 13 children affected between 1 month and 3 years of age, negative sequelae were secondary to avascular necrosis of the femoral head along with lytic damage of the articular cartilage, bone, and physeal cartilage. The greater trochanter appears to be relatively spared. Hallel and Salvatti defined results into three groups following septic arthritis of the hip in infancy (84). In group 1 there was a normal femoral head or a head only mildly deformed with a coxa magna. In group 2 there was a deformed small head-neck component with varus malposition. In group 3, there was complete absence of the head and neck due to lytic destruction. The femoral head may be in normal position, subluxed, or dislocated. Major predisposing features of neonatal septic arthritis were an indwelling umbilical catheter, prematurity, and septicemia. The earlier treatment began, the better the result. Treatment involves open arthrotomy combined with intravenous antibiotic coverage and maintenance of the hip in the reduced position using either a hip spica or Pavlik harness during the healing phase. Major negative sequelae from neonatal hip sepsis continue to be seen. Vidigal and Jacomo reported on 14 neonates with septic arthritis of the hip who had been diagnosed late, meaning that diagnosis was made and surgical drainage was carded out more than 4 days after the onset of symptoms with an average delay of 9 days (223). The age of their patients at admission averaged 17 days, with a range between 9 and 30 days. In 11 of the cases, the organism was staphylococcus aureus with one each of staphylococcus albus, enterobacter, and streptococcus hemolyticus. Radiographic features at the time of diagnosis include a normal appearance in a few, but often subluxation, dislocation, and dislocation with periosteal elevation of the proximal femur (associated osteomyelitis) or lysis of the acetabulum (osteomyelitis). Only one hip was normal at follow-up. Three patients died. The radiographic features of the femoral head seen after long-term follow-up for neonatal septic arthritis involved 7 in which the femoral head was destroyed and dislocated, 2 in which it was deformed and subluxated, 2 in which it was subluxated, 1 in which it was deformed and dislocated, 1 in which there was avascular necrosis and subluxation, and 1 that was normal.

b. Negative Growth Sequelae Following Septic Arthritis of the Hip A study by Choi et al. evaluated residual deformity and late treatment of 34 hips in 31 children who had septic arthritis when they were less than 1 year of age (36). They evolved a classification of the sequelae into four types, which incorporates well observations made over several decades including those of Hunka et al. (93). The end responses can be extremely variable due to the time that diagnosis is made, the age of the patient, and the type and effectiveness of the initial treatment used. The classification of the sequelae is illustrated in Fig. 10 and listed in Table III. An example of a damaged proximal femur is shown in Fig. 11.

SECTION IV ~ Osteomyelitis and Septic Arthritis

TABLE III

Type IA IB IIA IIB IIIA IIIB IVA IVB

FIGURE 10 The classification of the sequelae of septic arthritis of the hip in infancy developed by Choi et al. is shown. [Reprinted from (36), with permission.]

c. Treatment Approaches The long-term sequelae of septic arthritis of the hip can be extremely severe so that aggressive therapy is needed once the diagnosis is suspected. Open arthrotomy is preferable to joint aspiration for the hip. Arthrotomy releases intra-articular pressure, provides a channel through which subsequently developing pus can automatically escape, and allows for accurate bacteriological diagnosis. Maintenance of the femoral head in the fully reduced position within the acetabulum by hip spica immobilization minimizes the likelihood of a septic dislocation. Intravenous high-dose antibiotics are used to clear osseus and soft tissue infection. Treatments of the long-term sequelae have been described in several reports (10, 18, 34, 36, 59, 62, 123) and are outlined in Table IV.

901

Classification o f S e q u e l a e : Septic Arthritis o f t h e Hip in Infancy (Choi et al.) Sequelae

No residual deformity Mild coxa magna Coxa breva with deformed head Progressive coxa vara-valga Slipping of femoral neck with severe anteversion or retroversion Pseudarthrosis of femoral neck Destruction of femoral head and neck with small medial remnant of the neck Complete loss of femoral head and neck and no articulation of the hip

severe disability (197). Smith differentiated what he referred to as the "acute arthritis of infants" from any of the other recognized joint affections of childhood. The disorder occurs "in the first year of life and is characterized by the suddenness of its onset and the rapidity of its progress and termination." He concluded that "it is very dangerous to life and intensely destructive to the articular ends of the bones which of course at this period of life are largely cartilaginous." The disease "rarely produces anchylosis but leaves a child with a limb shortened by loss of part of the articular end of some bone and with a weakened flail-like joint." Smith's 21 cases all occurred within the first year of life with 8 in infants under 1 month of age and 19 of 21 under 6 months of age. The disease first attacked at the hip, shoulder, or knee and in the absence of effective treatment often involved more than one joint. At postmortem examination, he "found in all instances a considerable and rapid loss of substance in the articular end of one of the long bones entering into the joint affected." In some, the absorption or ulceration proceeded from the joint surface toward the deeper parts, whereas in others the destruction of tissue had commenced in an abscess

d. Pathoanatomy of Septic Arthritis of the Hip of Infancy Septic arthritis of the hip in infancy was referred to for many decades as Tom Smith disease after Smith's article in 1874 on acute arthritis in infants in which he described 21 patients, 13 of whom died and 8 of whom healed with

FIGURE 11 Radiographicexample of damage caused by neonatal septic arthritis of the hip. The head has been destroyed almost completely, the neck is short and wide, and greater trochanter overgrowth is marked.

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

902

TABLE IV Late Operative T r e a t m e n t s for Severe Sequelae o f Infantile Septic Arthritis of the Hip a ,,

1. Excision of damaged femoral head a. For necrotic, completely separated femoral head lying free in joint (septic epiphysiolysis) b. To remove projections and allow for repositioning procedures 2. Adductor-iliopsoas tenotomies a. For adduction, flexion contractures 3. Open reduction a. Femoral head into acetabulum, after septic pathologic dislocation b. Femoral neck into acetabulum, after severe head destruction c. Greater trochanter into acetabulum, trochanteric arthroplasty 4. Greater trochanteric interventions a. Trochanteric arthroplasty (see 3c) b. Distal transfer of greater trochanter c. Greater trochanteric epiphyseal arrest, all to increase hip stabilization, minimize Trendelenburg gait 5. Proximal femoral osteotomy a. Varus osteotomy, following lateral subluxation or partial premature lateral physeal closure b. Valgus osteotomy, to better position head into acetabulum following complete or medial premature physeal closure with persisting trochanteric growth, to support and align lower extremity in relation to the hip joint, disregarding shape and position of head but correcting the adduction position of femur c. Extension-derotation osteotomy, to correct additional deformities sometimes present with a or b 6. Acetabuloplasty a. Shelf, Pemberton, Salter, Chairi, other 7. Epiphyseal arrest a. For lower extremity length discrepancy b. Contralateral distal femur with or without proximal tibia 8. Ipsilateral tibial lengthening a. For lower extremity length discrepancy b. Femoral lengthening rarely done because of hip instability 9. Contralateral femoral shortening a. For severe lower extremity length discrepancy 10. Bone graft for proximal femoral pseudarthrosis a. Often accompanied by osteotomy 11. Hip arthrodesis a. For painful, unstable hip not amenable to preceding approaches ,,

,,

................................

aDerived from 10, 18, 36, 59, 62, 84, 93, and 123.

within the articular end of the bone, which then progressed outward. This often leaked into the joint by a small opening near the margin of the articular cartilage. Smith defined the term subarticular abscess referring to abscess cavities formed beneath the articular cartilage in either the cartilaginous or osseous structure of the end of the bone. He stressed that "in

many cases, the formation of a subarticular abscess in the bone must have been the first step in the joint affection since while the articular end of the bone was extensively excavated, the aperture through which the abscess had burst into the joint was a mere pinhole and though the joint contained pus, the articular cartilage was apparently healthy." Smith's description of the pathoanatomic findings remains valid today and underscores the need for rapid, accurate diagnosis and treatment. A brief summary of his descriptive findings at postmortem in several cases of septic arthritis of the hip follows. Case 3: The head and part of the neck of the fight femur were removed completely by ulceration, giving the appearance of destruction proceeding from the articular surface to the deeper part of the bone. In the opposite left hip, however, the process of destruction appeared to begin within the bone and pass toward the joint surface where the end of the bone contained a well-defined subarticular abscess opening by a small hole in the joint. The fight hip was filled with pus, the synovial membrane was thickened and vascular, the capsular ligament was perforated by abscess, there was no ligamentum teres, and the head of the femur had been completely removed. All of the acetabular cartilage had been destroyed. The left hip also contained pus, but the synovial membrane, articular cartilage, and ligaments appeared healthy. There was a small ragged hole at the margin of the articular cartilage, which led by a narrow sinus into a well-defined abscess cavity containing pus. This was situated partly in the ossifying cartilage and partly in the cancellous bone. Case 4: A huge right thigh abscess led to death 18 days after the patient presented for medical care. The fight hip joint was full of pus. The ligamentum teres had disappeared, the capsular ligament had ulcerated, the rim of the acetabulum was destroyed by ulceration, the synovial membrane was thickened and vascular, and the head and neck of the femur had disappeared completely. Case 5: There was left hip joint swelling and discomfort with an abscess found in the hip joint at postmortem. The ligamentum teres had disappeared, the capsule was opened by ulceration, the cartilage of the acetabulum totally was destroyed in places, and the head of the femur was absorbed, having lost approximately two-thirds of its structure. Case 6: The left hip was full of pus, the ligamentum teres had given way, and the cartilage, though intact, had lost its pearly appearance. Case 8: A patient, aged 7 weeks at presentation, died 6 weeks after the onset of symptoms with a fight hip abscess. Postmortem examination showed the synovium in the fight hip joint to be swollen and vascular, the ligamentum teres had disappeared, approximately one-fourth of the head of the femur had been resorbed, and the articular surface was covered with punctate ulcerations. One of the small ulcers passed through a minute sinus to a cavity in the cancellous bone just beneath the ossifying cartilage. This cavity, which was approximately 1 in. in its longest diameter was lined by

SECTION IV 9 Osteomyelitis and Septic Arthritis

a thick false membrane. The osseous nucleus and the cartilaginous head of the bone were similarly excavated and communicated by a pinhole opening in the articular cartilage with the cavity of the joint. The acetabulum was shallow, widened laterally with its sharp edge absorbed, and its cartilage gone. Case 9: There was a fight hip abscess with the hip joint full of pus. The capsule was widely opened by ulceration. The ligamentum teres had disappeared, and the acetabulum was denuded of cartilage. The head and upper part of the neck of the femur had been completely absorbed. Smith indicated, however, "that many children not only survive the attack, but recover with useful joints." His case 14 describes a 6-month-old female who developed a left hip abscess, which drained itself spontaneously. At 6 years of age, there was 1.5-2 in. of shortening, the thigh muscles were smaller, and the limb was everted in walking and somewhat flail. What remained of the head of the femur was dislocated and easily mobile. Smith described other cases with survival and reasonable function, although rarely with a normal joint. The problems were related to shortening, multiple scars where the abscesses had drained either spontaneously or by surgical intervention, tendency to a dislocated or dislocatable head of the femur, movements of the joint that were abnormally free, and considerable limping. 2. DISTURBED DISTAL FEMORAL EPIPHYSEAL GROWTH AFTER INFANTILE OSTEOMYELITIS

Although it has been recognized for a long time that infection in infancy can lead to negative growth sequelae of the bones, more recent studies of the pathoanatomy and clinical characteristics of the damage done have appeared. The vital anatomic finding underlying the possible spread of metaphyseal osteomyelitis through the physis and into the epiphysis directly in the newborn period is the persistence of transphyseal vessels during this time frame. These have been extensively documented in virtually all species with concentration in the fetal period and persistence for several months to years following birth. Neonatal osteomyelitis thus can damage epiphyses directly at the proximal femur as noted in the previous section and also in other large physes, with the two most commonly affected being the distal femur and proximal humerus. Infection at the distal femur has the next greatest tendency after the proximal femur to lead to major growth complications. Ogden and Lister performed a histopathological study of hip, shoulder, and knee from a child who died at 27 days of age in association with multifocal osteomyelitis (145). Studies of the distal femur epiphysis demonstrated severe destruction. Infectious material had led to a complete physeal separation. Metaphyseal trabecular bone was clearly affected and was being destroyed. The distal femoral growth plate had been completely destroyed and the hyaline cartilage of the epiphysis was being diffusely invaded. Serial sections demonstrated an inflammatory exudate that could be traced along the cartilage canals from the

903

metaphysis through the physis into the secondary ossification center of the epiphysis. The trabecular bone in the secondary ossification center was also being destroyed by the inflammatory process. Examples of distal femoral epiphyseal damage following neonatal infection were also reported by Banks et al. (13) and Halbstein (83). A study by Bergdahl et al. documented that neonatal osteomyelitis led to a high rate of negative long-term sequelae (17). Particular risk factors include large vessel and umbilical catheterization, a gestational age of less than 37 weeks or birth weight of less than 2500 g, and more generalized indications of systemic illness such as the respiratory distress syndrome, hyperbilirubinemia, perinatal asphyxia, and the need for an emergency Cesarean section. In the study of 40 patients with 71 sites of infection, the most commonly involved regions were the knee in 19 (27%), the hip in 17 (24%), the shoulder in 9 (13%), and the ankle in 6 (8%). Essentially any joint can be involved, but these four accounted for the highest involvement. Roberts pointed out that papers discussing osteomyelitis of the long bones in infancy in the preantibiotic era stressed high mortality, whereas information analyzed subsequent to that noted diminished mortality but increasing episodes of disturbed epiphyseal growth (167). He clarified the severity of growth problems at the knee with a study of 14 children who had bone infection present within 6 weeks of birth. One patient presented at 3 months. In 13 the distal femur was involved, whereas in only 2 was damage seen at the upper tibia. Multisite involvement of two or more sites was noted in 5 of the 15 patients. In spite of seemingly adequate treatment, growth sequelae were seen in virtually all instances. Major problems involved shortening, which was present in all, and angular deformity, which was present in most. Instability of the joint and diminution in the range of movement were seen frequently as well. Two major points are noted with neonatal infection in relation to the appearance of the epiphyses. (1) There is often considerable lysis followed by a delay in formation of the secondary ossification center particularly of the distal femoral epiphysis. Many have commented that the radiographic damage often looks greater than subsequent films at skeletal maturity indicate. The crucial factor is thus not the bony presence of the secondary ossification center but the presence, viability, and structure of the cartilage mass of the entire epiphysis. It is not infrequent after an infection that reossification occurs several years after the normal pattern if the cartilage model has managed to survive that. Often there is indication of the presence of the cartilage model based on the range of motion and function of the joint. At present this can be readily determined either by arthrography or by magnetic resonance imaging. (2) Growth may continue in a relatively normal fashion for several years after the infectious insult but then slow down well before skeletal maturation, leading to further shortening and angular deformity over the final years of growth. These patients should thus be followed to skeletal maturation. It is rare for the entire physis and

904

CHAPTER IO ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

epiphysis to be completely destroyed and corrective surgery and limb lengthening should lead to a reasonably functioning knee at skeletal maturity. Due to the fact that most of the growth of the lower extremity occurs at the knee the amount of shortening can be extensive. In Robert's series several patients had discrepancies between 10 and 14 cm, with many from 5 to 9 cm shorter on the involved side. Langenskiold reported several cases of growth disturbance of the distal femur after neonatal osteomyelitis (117). His cases also had onset of infection between 10 days and 4 weeks. The medial epiphysis was affected in 3 cases and the lateral in 4. The disorder led to partial closure of the growth plate seen in a range from 6 to 12 years of age. Growth and subsequent damage of the epiphysis were unpredictable on the basis of plain radiographic findings particularly during infancy and the early years of life. The maximum shortening documented by Langenskiold was variable depending on the nature of the involvement and ranged from 3 to as great as 19 cm. 3. SEPTIC ARTHRITIS OF JOINTS OTHER THAN THE HIP AND THEIR SEQUELAE IN RELATION TO EPIPHYSEAL DEVELOPMENT

In a large study of 102 cases of septic arthritis of childhood, the hip was most commonly affected both in the 0to 3-year-old age group and in the 3 years and older group (70). There were 42 joints involved in the 0- to 3-yearold group: hip, 18/42 or 43%; knee, 15/42 or 35%; ankle, 5/42 or 12%; wrist, 2/42 or 5%; and shoulder 1, elbow 1, 2% each. The same distribution continued in the older age group. a. Knee After the hip the next most frequently involved joint is the knee. Smith described the effects of septic arthritis of the knee in infancy as defined at postmortem examination (197). The medial condyle of the femur was involved with "a large ragged hole in the cartilage of the internal condyle, large enough to admit one's finger; this led into a deep excavated cavity in the bone and ossifying cartilage; this cavity occupied a large part of the articular part of the femur of which little remained but the shell." In another case, both knees were involved. Postmortem examination on the right showed the joint to be partially ankylosed and the synovial membrane thickened and of a pinkish color. The opposing surfaces of tibia and femur were firmly adherent by infected granulation tissue and "could not be separated by tearing." The articular surface of the femur was deeply excavated by ulceration. On hemisection, there was a small, circular cavity within the cartilaginous end of the femur, which led by a pinhole sinus through the condyle into the joint. The opposite knee was filled with pus. The synovial membrane was thickened and vascular, and the cartilage surface of the femur was irregularly absorbed to a depth of 0.5 in. or more. There was a fold on the anterior surface of the lateral condyle 0.25 in. wide and 0.5 in. deep lined by vascular tissue. Hemisection showed a cavity in the lower

end of the femur partly in the bony secondary center and partly in the ossifying cartilage. Septic arthritis of the knee continues to be described today. The distal femur suffers more growth-related sequelae than the proximal tibia. Angular deformities due to partial distal femoral physeal destruction are far more common than complete symmetric growth cessation. The most damaging growth sequelae continue to follow neonatal sepsis and to a lesser extent in those affected within the first 2 years of life. After this time, the tendency toward earlier diagnosis and more effective treatment minimize the negative sequelae. A large study of 96 septic knees in childhood demonstrated far more serious growth sequelae in patients who were 24 months of age or younger at the time of diagnosis (208). In 50 knees assessed in detail in those within 2 years of age, 26 had no deformity at follow-up and 24 did. Fifteen had a varus deformity averaging 14~ (range = 5-30~ and 9 were in valgus with an average of 22 ~ and a range from 5 to 40 ~ Seventeen of the knees had flexion contractures averaging 12~ with a range of 5-25 ~ Central growth retardation of the distal femoral physis was noted in almost every case in which metaphyseal changes were accompanied by deformity. The secondary ossification center was also delayed in appearance, or if present was partially lysed. The tibia was affected less often and less severely than the femur. b. Shoulder A shoulder septic arthritis was described by Smith with pus in the joint (197). The cartilage over the glenoid cavity had disappeared and the bone beneath was roughened. The head of the humerus was wasted, ulcerated, and deformed by absorption on hemisection. An irregular cavity was exposed between the ossifying cartilage and the shaft of the bone. Examples of septic arthritis of the shoulder with longterm sequelae continue to be described. A large series from Poland indicated that, in their referral hospital, 18% of septic joints of childhood were at the shoulder (120). In a review of 46 septic shoulders in 42 patients less than 18 months of age who were followed for an average of slightly under 7 years, only 7% of the humeral heads were entirely normal radiographically. Although length discrepancies did occur, the mean amount of shortening was only 2.4 cm after several years, but 1 patient experienced a 9-cm shortening. There were few complaints about the length either from functional or from cosmetic viewpoints. The initial appearance of the secondary ossification centers in the humeral head was slowed by infection, whereas in those patients in whom they were already present at the time of infection they usually disappeared for months to years. There was no specific pattern to the negative sequelae that occurred. There was a continuum of distortion of the shape of the humeral head from normal to knoblike, with various amounts of flattening, reorientation of the tilt of the head, and degrees of forking or saddling of the epiphysis. Schmidt et al. reviewed 9 children diagnosed with septic arthritis of the shoulder (180). Eight of the 9 were under

SECTION IV ~ Osteomyelitis and Septic Arthritis

18 months of age at the time of diagnosis. Involvement was felt to be due to transphyseal spread from metaphyseal vessels, secondary to metaphyseal perforation into an intraarticular part of the joint, and from the bicipital synovial sheath, which can be affected and which passes over the joint. Growth sequelae were relatively rare and all patients developed a satisfactory, pain-free range of motion. Even if varus of the proximal humerus and some shortening occur, these rarely present the clinical difficulties inherent with similar lower extremity damage. c. Wrist Septic arthritis of the wrist is extremely rare but can, on occasion, lead to negative growth sequelae (209). Damage can involve lysis of the secondary ossification center and destruction of all or part of the radial growth plate. d.

Long-Term Sequelae of Childhood Septic Arthritis

Involvement of major joints other than the hip with septic arthritis, particularly the knee, shoulder, and ankle, after the first year of life is due primarily to direct joint involvement rather than through spread from metaphyseal loci and transphyseal to epiphyseal spread. The physes are extracapsular, and transphyseal vessels in the human postnatal age group are infrequently documented after 1 year of age. There also is a tendency for septic arthritis in these joints to be diagnosed sooner than in the hip because they are more superficial and swelling and redness are more readily seen. After sepsis within the first 2 years of life particularly, even if it appears to have been treated effectively, it remains important to follow the growth pattern for several years and to skeletal maturity if possible. Peters et al. assessed several cases of neonatal joint sepsis other than the hip in which growth sequelae were not apparent for several years (154). It was often not clear retrospectively whether the damage resulted from a metaphyseal osteomyelitis or a primary septic arthritis. Four distal femurs suffered growth sequelae with partial growth plate arrest, leading to shortening and valgus deformation. There were four proximal humeral deformities, all varus, one distal radial shortening with ulnar prominence, and one distal humeral shortening. Growth plate arrest following sepsis often extended beyond the area of bone bridging, with the adjacent physeal tissue being fibrous rather than cartilaginous. Gillespie has analyzed cases of septic arthritis of major joints in childhood in terms of residual disability (70). In the hip, 17 of 41 cases (41%) had an unsatisfactory result, in the ankle, 2 of 13 cases (15%) were unsatisfactory, and in the knee, only 3 of 37 cases (8%) were unsatisfactory. Long-term assessment is also important because of the potential for radiographic improvement of the involved epiphyseal region several months to years after the infection. There is often lysis of the bone of the secondary ossification center and/or the adjacent metaphysis even though the cartilage model is intact. The infection can cause necrosis of bone with subsequent reabsorption but delayed new bone formation. In younger epiphyses in which the secondary center has not yet formed, vascular damage can lead to a marked delay

905

in formation of the center, but the cartilage model can continue to grow. Studies to assess the cartilaginous structure of an epiphyseal region using either arthrography or MR imaging are essential to assess growth potential when bone deficiency is seen by plain radiography.

D. Summary of Effects of Epiphyseal and Metaphyseal Infection on Epiphyseal Growth The three possible growth effects of juxtaepiphyseal pyogenic infections are stimulation of growth, retardation of growth, and angular deformity (189). The nature and extent of lower extremity length discrepancies in relation to septic arthritis of the hip and osteomyelitis of growing bones are summarized here and reviewed in greater depth in Chapter 8. 1. OSTEOMYELITIS

Overgrowth of a long bone that is the site of osteomyelitis was found in 18 of 35 patients (21%) in Wilson and McKeever's series (232). This almost always occurred when the osteomyelitis was metaphyseal and damage to the physis had not occurred. Trueta and Morgan (217) stated that overgrowth following osteomyelitis lasts until medullary recanalization occurs by which time the sequestra would have been resorbed and a normal vascular pattern would have been reestablished. The maximum overgrowth was 2.0 cm, but most patients had only a few millimeters. In chronic recurrent osteomyelitis of childhood, however, overgrowth will persist. When an infection is localized eccentrically beneath the epiphyseal plate, then growth stimulation can also be asymmetric and angular deformities may occur. The same percentage of patients had growth retardation when chronic infection damaged the physis. In the modem era, diagnosis of metaphyseal hematogenous osteomyelitis within a few days of onset followed by intravenous antibiotics and transcortical decompression drilling, if needed, has led to rapid cure, and growth sequelae are now encountered much less frequently. 2. SEPTIC ARTHRITIS OF THE HIP

Damage to the femoral capital epiphysis in septic arthritis can produce serious length discrepancies. In our series, such discrepancies tended to increase with time, but a type I pattern was seen in only 42% of the patients and most commonly when the infection had occurred relatively late, after the age of 7 or 8 years (185). An assessment of pattern development in this group was obscured somewhat more often than in other groups because of the necessity for early and often for frequent surgical intervention, although femoral osteotomy per se was performed only infrequently in growing children. Even with complete destruction of the physis, however, femoral shortening did not invariably become worse with time, particularly in the younger patients. When the greater trochanter overtakes the involved femoral head in height, the femur resumes a somewhat more regular growth pattern

906

CHAPTER 10 9 Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

because the greater trochanter and distal femoral physes are normal, thus accounting for the type II and III patterns seen. Studies of growth sequelae of septic arthritis of the hip of infancy by Betz et al. (18), Choi et al. (36), and Hallel and Salvati (84) documented mean length discrepancies of 3 3.5 cm from injury to the proximal femoral growth plate and an additional 2.5 cm if there was hip instability. The type IV growth pattern was limited almost exclusively to abnormalities of the proximal end of the femur such as occur with septic arthritis of the hip. In patients in whom damage was relatively mild, premature fusion of the proximal femoral capital physis has been noted years after the infectious insult. The premature fusion can be detected 2 or 3 years prior to skeletal maturation by the progressive change in the relationship of the level of the greater trochanteric physis to that of the proximal femoral capital physis. It is, therefore, extremely important to continue periodic assessment of these children by carefully monitoring the relationship of the head and neck to the greater trochanter until skeletal maturity, even if the discrepancy has been in a plateau phase for several years. Although the average increase in the late phase was only approximately 1 cm, this amount can convert a clinically insignificant discrepancy into one of 2.5 cm or more and thus warrants special consideration.

E. Tuberculosis 1. OVERVIEW Skeletal tuberculosis in the growing child commonly affects the spine and the epiphyseal regions of major long bones, where it presents as a combined epiphyseal osteomyelitis and septic arthritis (97). This disorder was particularly common until it became better controlled due to advances beginning early in the twentieth century, which involved the discovery of the tubercle bacillus by Koch, recognition of the spread of the disorder in association with poor living conditions involving crowded and unsanitary housing, and the discovery of effective antibiotic treatment initially involving streptomycin, isoniazid, and p-aminosalicylic acid. It is still seen in many regions of the world and even in relatively advanced industrial nations, particularly in children with other debilitating disorders. There has been a worrisome resurgence of tuberculosis over the past two decades partly due to increased drug resistance and HIVpositive and AIDS disorders. It has been estimated that there are over 1.3 million infected children globally under the age of 15 years. Approximately 1% of cases of tuberculosis have skeletal involvement. The spine is most commonly affected in childhood skeletal tuberculosis followed by the hip and knee joints (3, 97, 102, 110, 190, 218, 230, 231). In a review of 219 patients with skeletal tuberculosis in Malaysia seen between 1968 and 1976, the disorder was spread almost evenly over several age groups from the first to the sixth decade (190). From birth to age 19 years there were 67 patients (31%), and the

spine was affected in 35/67 (52%). In the entire series the spine was most commonly affected (128/219, 58%), followed by the hip (37/219, 17%), knee (17/219, 8%), ankle (13/219, 6%), and wrist (8/219, 4%). A markedly similar disease distribution was reported at varying ages in a Japanese study of 914 lesions: spinal 61%, hip 14%, knee 8%, and foot 5% (110). The onset of joint symptoms tends to be slow, with discomfort, swelling of the joint, periarticular muscle atrophy, low-grade temperature elevations, and decreased function. Involvement of the hip was particularly common, and indeed much of the early confusion involving what has come to be known as Legg-Perthes disease came because of the fact that many of the clinical symptoms were similar. Diagnosis is made by a positive tuberculin skin test and the demonstration of tubercle bacilli on either bone or synovial biopsies or following growth of the organism. Although skeletal tuberculosis is relatively common, it is still infrequent in the overall spectrum of tuberculosis. The primary tuberculous focus almost always occurs in the lung and heals in many leaving only a fibrous scar. The second stage involves a postprimary reinfection, which might then be associated with spread throughout the body. Involvement of any specific organ is referred to as the tertiary stage of tuberculosis. Primary foci of tuberculous infection are not known to appear in the skeletal system. Tuberculous infections in skeletal parts reach that area by the bloodstream from an already existing extraskeletal tuberculous disease focus. In children, the nidus of tuberculosis occurs through hematogenous spread. 2. AGE INCIDENCE Tuberculosis is rarely seen during the first year of life. Statistics presented by Jaffe indicate that "in the first half of the twentieth century.., in about 50% of all cases, the subjects were between 3 and 15 years of age" (97). The most common areas of involvement are the vertebral column, hip, and knee. The joints of the lower extremities are affected much more commonly than those of the upper extremities. The clinical characteristics of a tuberculous skeletal infection are much more indolent than those of a septic arthritis or osteomyelitis. Involvement is almost always monoarticular. 3. RADIOGRAPHIC CHANGES The radiographic characteristic of tuberculosis is a periarticular osteopenia with the presence of a lytic lesion within the secondary ossification center. Almost invariably there is an increase of joint fluid due to synovitis. Swelling of periarticular soft tissues is seen with joint involvement, which generally results in a combined epiphyseal osteomyelitis and septic tuberculous arthritis. There may be narrowing of the joint space. There is resorption of the subchondral bone and the adjacent trabeculae by granulation tissue, with minimal to absent bone reaction and no formation of sequestra. Osteoporosis can be marked. Wedge- or cone-shaped areas of destruction are often seen in epiphyseal ends of the bone.

SECTION IV 9 Osteomyelitis and Septic Arthritis

There is a tendency toward the involvement of bone on both sides of the joint. Sclerosis adjacent to the lytic lesions can be seen with prolonged, partially controlled involvement as can osteophyte formation. 4. HISTOPATHOLOGY Skeletal involvement is considered to occur following hematologic spread. Approximately 1% of tuberculosis cases show some skeletal localization (218). The tendency is for the disease to begin in the articular or epiphyseal end of a bone, break into the joint in which it becomes a septic arthritis, and then involve the synovial membrane. On occasion, the synovial membrane can become infected directly by hematogenous spread. In the long tubular bones, the tuberculous process usually originates at the epiphyseal subarticular end of the bone and spreads through the cartilage surface to involve the synovium. Tubercle formation occurs in the marrow, with the trabeculae of spongy bone affected secondarily. There are necrosis and resorption of the cancellous trabeculae with circumferential spread of the lesion. Inspection of joints often shows destruction of the articular cartilage, with the femoral head covered by a pannus of tuberculous granulation tissue. In the hip, the bone of the secondary ossification center is affected as is that of the acetabulum. Tuberculous synovitis is seen frequently, associated with inflammatory thickening of the periarticular connective tissue and fat. Synovial membranes are extremely thickened and the articular cavity will contain a large amount of necrotic material. Synovial involvement is associated with partial or complete erosion and destruction of the articular cartilage with the tuberculous granulation going both over the surface of the cartilage and along its subchondral region. Phemister has provided a detailed assessment of changes in tuberculous arthritis (157). The seat of primary infection is in the bone in some cases and in the synovial membrane in others. In the early stages, therefore, the articular cartilage is not affected and remains alive. It is attacked secondarily by tuberculous granulation tissue either along its free junctional surface with the bone and synovium or along the region of its attachment to subchondral bone. The initial destruction of cartilage is usually by granulations growing from the synovium onto the periphery of the cartilage where it is not immediately adjacent to cartilage from the opposite bone. The second region in which granulations develop is the peripheral portion of epiphyseal bone beneath the articular cartilage. A thin layer of subchondral granulation forms, which absorbs the subchondral bone and the deeper layer of cartilage, leading to detachment of overlying cartilage along with necrosis. Marginal undermining of the articular cartilage may also be seen by tuberculous granulations growing in from the synovium. The subchondral bone can be affected in two ways. In small focal areas, the bone may be broken down rapidly as the focus spreads, leaving a cavity filled with tuberculous granulations or necrotic debris. In other cases, large areas are involved and the tuberculous tissue of

907

the cancellous spaces undergoes early necrosis and caseation, leaving the dead bone trabeculae practically intact. The overlying cartilage is then affected secondarily due to subchondral bone collapse similar to what is seen in osteonecrosis where the cartilage is initially unaffected. Phemister concluded that secondary changes along the articular surfaces of the bones developed from spread of the tuberculous infection from the primary point of involvement. The most common region of tuberculous skeletal involvement after the spine and hip is the knee. Other areas of involvement are the tarsal, carpal, and elbow bones. Shoulder involvement is less frequently seen than the elbow. Involvement here, however, is in the head of the humerus, and the lesion spreads from there to the articular cavity. In joints, which go on to stabilize in the absence of full healing, fibrous ankylosis tends to develop. Cartilage is not regenerated, and in preantibiotic days there was often some degree of fibrous or bone ankylosis. In summary, in skeletal tuberculosis the tubercle is the characteristic pathological finding associated with tuberculous granulation tissue and, later on, caseation necrosis and eventual fibrosis. Except for the vertebral lesions, the majority of cases of skeletal tuberculosis involve a joint. Almost invariably there are an associated tuberculous synovitis and osteomyelitis of the epiphyseal region. Virtually all metaphyseal and epiphyseal foci sooner or later involve the adjacent joint. The epiphyseal focus extends into the adjacent joint at a margin near the synovial-cartilage junction. In association with synovial involvement, there is early invasion of the articular cortex at the joint margins. The tuberculous granulation tissue erodes through the edge of the cartilage to create a focus of advancing damage. There is subchondral granulation resulting in fibrous tissue deposition. Eventually, the zone of calcified cartilage and the noncalcified articular cartilage are slowly resorbed. Subchondral marrow fibrosis is seen. The articular cartilage is also resorbed from the surface beneath the granulation tissue creeping over it. Resorption and destruction of cartilage occur first at the margins of the joint and then move progressively onto the weight bearing surfaces. Within the bone itself, the dead trabeculae remain of normal thickness, whereas those at the periphery in the more viable regions tend to become thinner in association with decreased use and increased vascularity. The phrase "kissing" sequestra has been used for many decades to describe the pathological findings in tuberculous osteomyelitis and arthritis. These refer to secondary sequestra formed directly opposite each other in opposing epiphyses at sites of maximum pressure stress. In the preantibiotic era, there was very little, if any, tendency for tuberculous joints to heal with preservation of a useful range of motion. Healing occurred initially by a fibrous ankylosis, which in turn sometimes became a bony ankylosis. In most the healed joint became a mass of dense, fibrous adhesions with a fibrotic synovium. Figure 12 illustrates the pathophysiological changes of the joint and epiphyseal regions in the tuberculous disease.

908

CHAPTER IO 9 Metabolic, Inflammatory, Neoplastic, Infectious, and Hemarolottic Disorders

Tuberculous Arthritis Epiphyseal Osteomyelitis & Septic Arthritis

~:;:/ ~ii{ ~ ~ ~

1

~ _/~ ~

~

~--~i~ PrimaryEpiphyseal /~;:~~ Osteomyelitis(1) ~i~l Primary :.,~::/~: Synovitis(2) .

~ii~ ~ ~ 2~ :/ - -

~~ y

SecondaJ/ryArticula r CartilageDamage(3)

FIGURE 12 Tuberculous involvement of the joints is a combined epiphyseal osteomyelitisand tuberculous septic arthritis. The pathophysiologic changes are illustrated here.

5. TREATMENT Treatment for skeletal tuberculosis became effective with the development of antituberculous antibiotics (isoniazid, p-aminosalicylic acid, and streptomycin). Initially this was used in combination with splinting followed by physical therapy for range of motion of the joints. It shortly became evident that repair was improved with debridement of the fibrotic, necrotic infected tissue whether this was in the joint or within the adjacent bone. For children in the acute stages of the disorder, antibiotic therapy alone was often sufficient. Because development of the infection, however, tended to be indolent and discomfort was much less than that associated with pyogenic bacteria, many patients were already in the subacute or chronic phase when diagnosis was made. The combination of joint and bone debridement with antibiotic therapy and rehabilitation, with earlier focus on range of motion activities, tended to improve results greatly. Kondo and Yamada were early proponents of focal debridement (110). In large joints, all loose cartilaginous fragments, loose devitalized bone fragments, and necrotic or devitalized infectious tissues were removed surgically. They felt that the combined therapy greatly improved results. Debridement was especially indicated during the later chronic stages of the disorder particularly when abscesses or sinuses were present, which markedly limited access by systemic antibiotics. An extensive study also from Japan by Katayama et al. strongly supported the triple-therapy antibiotics along with surgical synovectomy and joint and bone debridement to physically remove all gross tuberculous foci particularly

in those regions that communicated with the joint space (102). Synovectomy alone was commonly resorted to within the knee joint. Katayama et al. felt that these treatments improved the likelihood of retention of joint motion, even though completely normal joints were rarely achieved with chronic disorders. Wilkinson also resorted increasingly to synovectomy of the knee and hip in combination with chemotherapy to improve joint healing and function (230, 231). Care was taken to distinguish the tuberculous joint from pyogenic arthritis even when there was extensive involvement with synovial swelling and lytic areas of bone because the articular cartilage tended to survive much better in tuberculous than in the pyogenic infection. Synovectomy and even removal of the overlying pannus layer in combination with chemotherapy led to some remarkable recoveries in tuberculous arthritis. Wilkinson recognized that in tuberculous arthritis both synovium and bone were involved by the time the disease had become recognizable clinically. On the basis of extensive clinical experience, he noted the pathological involvement at hip and knee to be definable into four groups according to the tissues involved (231). The first group involved those with disease of both synovial membrane and cartilage. The cartilage involvement tended to be around the periphery of the articular surface. The next group had disease of the synovial membrane, cartilage, and bone. This involved considerably advanced disease with frequent destruction of the joint and subluxation or dislocation of the femoral head in the hip. The third group involved disease of the synovial membrane and bone with little evidence of cartilage loss. The fourth group had mainly bony changes with no change in the joint found at operation. In those with disease of the synovium and cartilage, the synovium was always more markedly involved being thickened and vascular and encroaching on the periphery of the cartilage. In some cases the synovium spread over the cartilage almost to the center of the joint. Cartilage laceration also occurred over subchondral tuberculous foci with diseased tissue attacking from bone into cartilage from below. Pannus formation was also observed, spreading from the periphery onto the cartilage surface. Even when pannus covered the cartilage, the cartilage beneath was often intact. Disease of synovium, cartilage, and bone was seen, generally, when the patient was older and destruction more severe. Chemotherapy was combined increasingly with synovectomy and joint debridement, and many joints that initially appeared to be quite damaged were restored to reasonable function for several years. Debridement and curettage of adjacent bony foci frequently were needed. In terms of bone surgery, osteotomy was required in more severe cases and, although arthrodesis was resorted to early in the series at progressively later time periods, the combination of antibiotics and more vigorous and early surgery limited the number of arthrodeses needed. Wilkinson stressed, as had others, that there was a tendency for the articular cartilage to survive even in joints that appeared to

SECTION V ~ Hematologic Disorders

be severely affected by tuberculosis. Pannus was removed if possible, and if it was of relatively recent origin the cartilage showed an excellent tendency to regenerate. Synovectomy was helpful because synovial persistence in a thickened state delayed cartilage nutrition. In addition, the presence of pannus also delayed joint recovery. More recently ethambutol has replaced PAS as one of the triple-therapeutic drug treatments. Rifampicin has also been used. Many combinations of the drugs are used by differing practitioners, and some will use only one initially holding the others in reserve. Tuli has reviewed evidence that the uptake of drug at foci of tuberculous lesions in the skeleton is surprisingly high and stresses the value of the drug over surgical intervention. He supports the primacy of antitubercular drugs, repetitive active and assisted exercises for the involved joint from the very beginning, and rest in the functional position between exercises. Surgical intervention is reserved for those with chronic synovitis not responsive to medical management. Management is predicated to a great extent on economic and social factors. When there are no specific constraints to active therapy, it appears that optimal antibiotic use, early physical therapy to stress range of motion, and surgical debridement with any evidence of slowness of repair would be the optimal approach. 6. LONG-TERM FOLLOW-UP OF CHILDHOOD TUBERCULOUS JOINT INFECTION Chow and Yau studied 30 cases of tuberculosis of the knee followed for an average of 15 years, with the majority having developed the disease during childhood (37). In 75% of the patients, the infection occurred before the age of 18 years. The peak year of incidence was 4 years, with the large majority of cases occurring from 1 to 11 years of age. Treatment varied between conservative and operative, but all had chemotherapy. In those patients treated operatively, synovectomy and debridement followed by cast immobilization and bracing were the usual approach. On occasion, the destruction was sufficiently great that immobilization was continued to the point of ankylosis. Conservative treatment involved chemotherapy with casting and bracing alone. In patients in the childhood group, deformities commonly seen involved varus deformity of the knee, flexion contractures, and lower extremity shortening. In 6 instances there was varus deformity, which was mild in 2, simply leading to loss of the normal valgus angle, 10~ in 2 cases, and 30 ~ in 2 additional cases. The flexion contractures of the knee were seen in 12 cases at 10~ in 6, 20 ~ in 5, and 30 ~ in 1. Significant clinical shortening was demonstrated in 7 cases, with the amounts being 2.5 cm in 2, 5.0 cm in 4, and 7.5 cm in 1. Most of the abnormal deformities occurred in association with each other. There were 13 knees with no deformity and an otherwise excellent result. On occasion, extensive regeneration of the bone occurred even though much of the epiphyseal subchondral bone appeared lytic at diagnosis. At review, roughly 25% of the joints were radiologically nor-

909

mal. Of those with abnormalities, the most common were described as "congruous incongruity in which the joint space was well preserved although it was irregular in outline with a matching pattern of irregularity in the two opposing surfaces of the tibia and femur." Other X-ray abnormalities were osteophytes, osteoporosis, a smaller epiphysis, chondral calcinosis, abnormal patellae, and loose bodies. Of note was the fact that "as the child grew older, the previously destroyed bone and perhaps the overlying articular cartilage showed a remarkable power of reconstitution."

V. HEMATOLOGIC DISORDERS A. Hemophilia 1. OVERVIEW Hemophilia is a sex-linked recessive bleeding disorder caused by a deficiency of blood coagulation factors VIII or IX (47, 66, 90, 170, 227). It is an inherited tendency in males to bleed. Prior to the onset of effective therapy for this disorder, damage to the joints and epiphyseal ends of the bone was extensive, and treatment methods involving surgery and physical therapy were not curative. Lethal hemorrhage was frequently seen following major trauma or after surgery performed in which the diagnosis was not recognized; Koenig (111) and Key (107) reported several documented cases in their reviews of the disorder. The natural history of the disorder has changed significantly following factor replacement therapies over the past several decades, but considerable morbidity and some mortality persist. Larsson has documented the dramatic change in life expectancy with hemophilia in Sweden where detailed statistics have been available since the eighteenth century (118). The median life expectancy of patients with severe hemophilia increased from 11.4 years from 1831 to 1920, to the mid-20s from 1921 to 1960, to 56.8 years from 1961 to 1980. The AIDS crisis then led to a significant decline in median life expectancy in the United States and western Europe from 1981 to 1990 prior to recognition of its cause and avoidance of infected blood. Hemophilia A, the most common hereditary coagulation disorder, is seen throughout the world and has an incidence of 2 per 10,000 male births. It is an X-linked recessive disorder that affects males only and is transmitted by females. The disorder is the result of a new mutation in approximately 33% of new patients, so that a family history is often not obtainable. It is due to the absence, severe deficiency, or defective functioning of the plasma coagulation factor VIII (antihemophilic factor). Several different mutations cause the disorder, which accounts for the variable clinical severity. Hemophilia B (Christmas disease) is indistinguishable clinically from hemophilia A, is also transmitted as an X-linked disorder, and is the result of factor IX deficiency. It is the result of a new mutation in approximately 20% of new patients.

910

CHAPTER IO 9 Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

Hemophilia was described as an inherited, sex-linked disorder by John Otto of New Hampshire in 1803 (149). For several decades, however, no connection was made between the bleeding disorder and the joint swelling and degeneration. According to Koenig, it was not until the 1860s that this correlation even began to be considered (111). The factor VIII deficiency (hemophilia A) occurs in approximately 85% of cases and the factor IX deficiency in 15%. Three severities of disorder are defined for both factor VIII and IX deficiencies. Those with a mild disorder have 5-25% of factor activity in the blood, moderate 1-4% factor activity, and severe 0-1% factor activity. The distinction is extremely important because mild and moderate hemophilia A rarely lead to joint bleeding. It has been estimated that in hemophilia A the severe-moderate-mild distribution is 70%15%-15% and in hemophilia B it is 50%-30%-20% (47). When unusual bleeding is encountered in a male patient, laboratory screening tests include a normal platelet count and prothrombin time but a prolonged activated partial thromboplastin time. Specific factor assays are then needed to determine the deficiency differentiating hemophilia A from hemophilia B. Patients with von Willebrand's disease also have factor VIII deficiency due not to a defect in the X chromosome but to the indirect consequence of qualitative or quantitative changes in plasma von Willebrand factor (226). Factor VIII circulates in the plasma with von Willebrand factor in a noncovalent complex, but the two factors are distinct proteins. Much of the improved treatment of hemophilia comes from the molecular understanding. Some patients, however, remain resistant to therapy because of the presence of inhibitors, which are antibodies developing against exogenous factor VIII or IX (47, 53). These inhibitors are present in 2050% of people with severe hemophilia and are suspected when good responses to therapy are not achieved with the restoration of appropriate levels of factor. The clinical hallmarks of hemophilia A are joint and muscle hemorrhages, easy bruising, and prolonged hemorrhage after trauma or surgery, although there is no excessive bleeding after minor cuts or abrasions. Hemarthroses account for 75-85% of bleeding episodes in patients with severe hemophilia. Intra-articular bleeding is highly characteristic of factor VIII and IX deficiency and is not seen in other blood clotting factor deficiencies. The joints most frequently involved are the knees, elbows, and ankles, with occasional involvement of shoulders and hips. In a small series of 44 patients hospitalized with hemarthrosis, Eyring et al. reported an incidence of knee 62%, ankle 20%, elbow 15%, hip 2%, wrist 1%, and shoulder 0.5% (61). Houghton and Duthie reported joint involvement as follows: knee 328 (44%), elbow 190 (26%), ankle 106 (14%), sternoclavicle 33 (5%), shoulder 25 (3%), wrist 20 (35%), hip 19 (3%), and hand 17 (15%) (89). The ankle is the most common joint experiencing hemarthrosis in the second decade. Gamble et al. noted intra-articular bleeding to be most common in

the knee in those under 10 years of age, but in those between 11 and 19 years of age the frequency of hemarthrosis of the ankle was 2.5 times that of the knee (67). Hemarthroses tend to be noted when the child first begins to walk and generally are spontaneous. The joint bleeding is associated with discomfort and limitation of motion and then with inflammation and synovitis. This establishes a vicious cycle, which leads to more hemarthrosis. Ultimately, hemophilic arthropathy damages and destroys the cartilage, narrowing the joint space, and is associated with small subchondral bone cysts, osteopenia, and contractures. 2. GENE ABNORMALITIES The gene that codes for factor VIII is at the tip of the long arm of the X chromosome (47, 90, 105,227). It is one of the largest human genes, comprising nearly 186 kb and constituting nearly 0.1% of the X chromosome. The initial studies of mutations were performed by screening genomic DNA digested with restricted enzymes. By 1991 more than 150 different mutations had been identified in hemophilia A. With a few exceptions, deletions in the factor VIII gene cause a severe form of hemophilia A. Nonsense mutations also cause severe hemophilia. Over 80 point mutations and 70 gene alterations have been identified in the factor VIII gene, which underscores the wide range of clinical variability. Characterization of factor VIII mutations has been slow because of the extremely large size of the gene and the variable mutations, which are reflected in the fact that the hemophilia A is a clinically heterogeneous disorder. More recently, an inversion in the factor VIII gene at intron 22 was identified and found to account for as many as 40% of all severe hemophilia A gene abnormalities. The factor IX gene is smaller, comprising 34 kb (66). More than 300 unique mutations have been found; of families with moderate to severe disease, 95% have different mutations. No link between factor VIII and IX genes has been found. With pedigree analysis and laboratory studies involving plasma levels of factor VIII and von Willebrand factor, it is now possible to determine carder or noncarder status for women with greater than 95% accuracy in roughly 80% of women who are evaluated. DNA analysis has been used to determine carrier status. Ultrasound-assisted fetoscopy has made it possible to obtain fetal blood samples in the eighteenth to twentieth weeks of pregnancy, and prenatal diagnosis of hemophilia has become feasible with the advent of factor VIII immunoassays. More recently, prenatal diagnosis during the eighth to tenth weeks of pregnancy has become possible through DNA analysis of amniocytes or chorionicvillus material. 3. ORTHOPEDIC SEQUELAE OF CHILDHOOD HEMOPHILIA

a. Musculoskeletal Damage: Arthropathy, Pseudo-tumor, Neuropathy, Contractures, Volkmann's Ischemia, Fractures, and Ectopic Ossification. The major orthopedic prob-

SECTION V ~ Hematologic Disorders

lem in hemophilia is intra-articular bleeding with subsequent hemarthropathy (8, 45, 125, 163, 170, 198, 201). This is generally associated with muscle atrophy and a tendency to joint contractures. The most commonly involved joints are the knees, ankles, and elbows. Characteristic knee deformities accompanying hemophilic arthropathy involve a flexion contracture, genu valgum, and external rotation of the tibia. In long established cases, there is posterior subluxation of the tibia. The knees tend to be "knobby" due to irregular overgrowth of the epiphyses, overgrowth of the periphery of the patella, and periarticular muscle atrophy. The ankle shows diffuse osteopenia, collapse of the superior surface of the talus, a valgus tilt of the ankle joint, and a mild equinus deformity. The elbow develops a progressive flexion contracture. Other less common disorders involve bleeding surrounding and into the bone, primarily the proximal femur and ilium, which leads to the formation of a lytic lesion referred to as a hemophilic pseudo-tumor; bleeding into muscle and soft tissues, which can lead to neuropathies, particularly of the sciatic and femoral nerves and episodes of Volkmann's ischemia of the forearm and calf; fractures, generally of the distal one-half of the femur, associated with chronic knee arthropathy, stiffness, and osteopenia; and ectopic ossification, usually in the pelvis. Arthropathy is considered in detail in the following section, followed by the other less common disorders in the subsequent section. b. Hemophilic Arthropathy. Joint involvement is rare in the first year of life and generally begins in the second year with the onset of walking. The knee is most commonly affected followed by the ankle and elbow joints. Small joints of the fingers and toes are only rarely affected. In most hemophiliacs, a single joint tends to be affected repeatedly and is referred to as a target joint. Three Stages of Hemophilic Arthropathy (Koenig): Koenig in 1892 was one of the first to recognize that the joint abnormalities of hemophilia following the hemarthrosis represented a specific pathoanatomic entity, not simply a variant of rheumatologic or arthritic disease that just happened to occur in hemophiliacs. He divided the disorder into three stages (111). The first stage due to initial bleeding was referred to as the stage of true hemarthrosis, the second stage was the inflammatory stage with a panarthritis with reactive inflammation throughout the involved joint with proliferation of the synovial membrane, and the third stage was one of regression or wasting away of the joint with destruction of the joint surfaces, contractures causing permanent deformity, and ankylosis. Koenig recognized that the initial symptoms were often vague with apparently spontaneous but rapid onset of the hemarthrosis occurring without trauma, usually without associated pain and with a surprisingly good range of motion retained. The deterioration of function occurred gradually after several bleeding episodes. Aspiration yielded bloody fluid. Koenig was able to diagnosis a "bleeders' joint" of hemophilia "due to the patient's pallor, the

911

sudden appearance of the fluid effusion, lack of pain in the joint, and the freedom of movement." Depalma and Cotler distinguished four grades of hemophilic arthropathy (45). Grade 1: Intra-articular hemorrhages that have not resulted in any functional joint impairment. There are no or only minimal radiographic changes. Grade 2: The affected joint presents a slightly reduced range of motion but the joint space is well-preserved, there is no irregularity of the subchondral bone, and the spongy trabeculae are only somewhat more prominent. Grade 3: There is fixed deformity of the joint associated with periarticular muscle atrophy, pericapsular and capsular thickening, irregular articular surfaces, and subchondral cysts with marginal spur formation. Grade 4: The deformity is pronounced and fixed with marked periarticular muscle atrophy. The articular bone ends are irregular and deformed with increased radiolucency and narrowing of the joint space. With time, the bone ends, particularly at the knee, tend to flatten and widen and subluxations are common. Fibrous ankylosis may lead to bony ankylosis. Marginal exostoses are seen. Gross Pathologic Findings: Koenig, in his initial paper from 1892, described changes observed in three cases of open arthrotomy or postmortem examination (111). Joint aspiration in the early phase yielded blood. This was still found a few weeks after the onset of bleeding, but some blood clots were free in the joint whereas others were attached to the capsule. With time the capsule was thickened and its inner lining was discolored with a blood-colored fluid. Fibrous deposits were noted to occur both on the capsule and on part of the cartilage surface of the distal femur. Even at this stage, "the cartilage starts to change, to become fibrous, and the strangely sharp-edged defects described below start to develop." In the second stage, Koenig noted that, if no new bleeding had taken place, the contents of the joint are "not purely blood but bloody serous or purely serous with a slightly brown color." At arthrotomy in these joints a "great number of brownish synovial tufts was particularly noticeable." He then described the villous appearance of the synovium plus its red-brown, brown, or gray discoloration. He also noted changes in the cartilage. "All over it has lost its white color and its shine; it is discolored a dirty red-brown or gray-brown." Koenig noted the tendency for the cartilage to become fibrous. "The truly characteristic changes consist in strangely sharp-edged defects, small and big, deep and on the joint surface with gnawing defects to the cartilage in various spots of the joint but mostly at the same places where the fibrous deposits are found." There was a characteristic central position of the defects in which "we have found no actual changes on the edge of the cartilage surface as is the case in arthritic deformity." With histologic sections, there was granular blood pigment in both the cartilage and the synovium. The third stage he referred to as "regressive metamorphosis." Scarring and shrinking of the connective tissue set in with superadded mechanical influences. The joint continued to degenerate and became stiff and deformed.

912

CHAPTER IO ~ Metabolic, Inflammatory,

Neoplastic, Infectious, and Hematologic Disorders

FIGURES 13 and 14 Exampleofjoint damagecausedby severehemophilicarthropathyis shown.At right (Fig. 14) is the humeral head of a 27-year-oldhemophilic patient; this can be compared with the opposite normal nonaffected humeral head at left (Fig. 13). [Reprinted from DePalma and Cotler (1956), Clin. Orthop. Rel. Res. 8:163-190, 9 LippincottWilliams & Wilkins, with permission.]

Koenig went on to describe the fibrous and then bony ankylosis and the fibrosis and tightening of the capsule. Flexion contractures and slight valgus deformation of the knee joint followed. Key provided the earliest and most detailed description of the hemophilic joint in English (107). At operation (for a mistaken diagnosis of traumatic arthritis), it was noted that the synovial membrane was unusually dark in color. Blood without clots escaped once the arthrotomy had been made. The synovial membrane was dark chocolate in color and the joint cavity was filled with hypertrophied synovial folds and villi. The articular cartilage on femoral, tibial, and patellar surfaces was yellowish brown in color, and there were many areas in which the cartilage either was thinned or had disappeared entirely. The underlying bone was covered by a chocolate-colored connective tissue. The areas of cartilage erosion were irregular in size and contour but were sharply demarcated as described by Koenig. Some of the synovial villi were pedunculated, whereas others were sessile. Many of the synovial folds branched extensively, giving areas of a mosslike structure. The villi were markedly friable. DePalma and Cotler provided an excellent description of the gross pathology (45). Postmortem study of two shoulders showed one side to be perfectly normal and the other with marked abnormalities (Figs. 13 and 14). The affected side was totally degenerated. There was no trace of hyaline cartilage, and the subchondral bone was eburnated and covered with numerous craterlike defects varying in size and shape. Subchondral cystic regions were seen that communicated with the joint cavity. Many of the defects in the cancellous bone had a smooth, firm, compact inner surface, indicating

that the contents of the cyst were under great pressure. There was considerable erosion of both cartilage and bone elements along the superior and posterior borders. The soft tissues of the joint were discolored. The synovial membrane was thin with only minimal villus formation noted at this stage. The ligaments were thicker and broader than normal. In the knee joint, the entire synovial lining was discolored in varying shades of reddish brown. The entire thickness of both menisci and the quadriceps tendon was permeated by blood pigments. The articular cartilage was similarly discolored, swollen, wavy, and shredded. Numerous fine villi covered its surface. The cartilage was markedly pitted and also exhibited clefting, but there were no areas of erosion to the level of the subchondral bone. There was no eburnation. The synovial membrane and subsynovial tissues were thickened markedly. Numerous fine villi extended from the synovium. The intercondylar notch was filled with thickened hyperplastic synovial tissue. The extent of joint damage in hemophilic arthropathy is shown in Fig. 14. Pathogenesis of Hemophilic Arthropathy: Clinical swelling of a joint is rarely documented before 12 months of age, and most of the bleeding episodes become prominent at 4 - 5 years of age. It is usual that the first massive hemarthrosis can be resorbed with no apparent sequelae. It is the recurrent intra-articular bleeding that becomes problematic. The three-stage progression of the disease outlined by Koenig remains accurate today: hemarthrosis, inflammatory panarthritis, and regressive stages (111). The repeated insuits of numerous and frequent hemorrhages can be either massive or small and subclinical. When the stages of chronicity begin, the synovium is involved first followed by the

SECTION V ~ Hematologic Disorders

cartilage and then the bone. The subsequent synovial hyperplasia serves as a deterrent to effective resorption of blood from the joint cavity with repeated episodes. Bleeding occurs more frequently with even minimal trauma because the thickened synovium is hypervascular. Organization of the subsynovial granulation tissue further worsens the process by producing dense scar tissue. Pannus ensues covering the periphery of the articular cartilage and leading to underlying cartilage degeneration. Occasional osteophytes are seen at the periphery. The subchondral and cancellous bone tends to atrophy. There is coalescence of small marrow spaces forming larger cystic spaces. There is proliferating vascular connective tissue in some of the cysts, whereas in other areas there is blood clot associated with intraosseous bleeding. Most cysts eventually gain communication with the joint cavity as the overlying cartilage degenerates. Articular cartilage degeneration is an invariable part of the picture, with marked fibrillation of the collagen fibers and in some areas focal full thickness defects. In some cartilage areas, cellular elements have disappeared almost completely. In areas in which the entire thickness of the cartilage has been destroyed, surface tissue is replaced by vascular granulation tissue extending from the marrow. With cartilage destruction the joint space narrows. In the final or third phase of Koenig the bleeding episodes diminish because the synovium has become so thickened and sclerotic that vessel involution has occurred. By this time joint cartilage can be nonexistent, and bone on bone crepitus is very painful, motion is limited, contracture deformity is great, and function is markedly diminished. Histopathologic Findings: The early synovial reaction to intra-articular bleeding resembles the changes occurring with rheumatoid arthritis. Initial hemarthroses are quickly resorbed with no clinical sequelae. With multiple bleeds, however, resorption is slower and persisting intra-articular blood and blood clot lead to the development of progressively severe changes. The synovial lesion is characterized by hemosiderin deposition and fibrovascular proliferation. There is inflammation, reactive proliferation, and brownish discoloration of the synovial membrane, which eventually lead to osteoarticular changes. The synovial membrane with time becomes hypertrophic and thickened with villous proliferation, increased vascularity, and round cell infiltrates. The blood products are resorbed by the synovium, but it is the overload of iron that appears to be most damaging to the synovial cells. Hemosiderin accumulates within synovial villi. Synovial cells that absorb too much iron disintegrate, releasing lysosomal enzymes that destroy articular cartilage and further irritate the synovium. Chondrocytes are also damaged directly by the iron. The synovium gradually becomes fibrotic and, in end stage disease, loses the capacity for excess bleeding and joint distention. These changes are soon followed by articular cartilage degeneration from superficial to deep. A brownish pannus covers the cartilage from the periphery. If full thickness focal regions of cartilage are lost, the subchondral bone becomes exposed and thickened. Subchondral cysts form and there

913

FIGURE 15 Outlineof early (A) and late (B) arthropathic changes in hemophilia from the work of Speer (198). c/j, synovitis; f/h, subchondral cyst; g, chondrolysis,fibrillation; i, osteochondral cyst; k, bone sclerosis; 1, widened notch. [Reprintedfrom Speer (1984), Clin. Orthop. Rel. Res. 185: 250-265, 9 LippincottWilliams& Wilkins, with permission.]

may be collapse of the articular surfaces. Subchondral cysts are often found to begin with an intraosseous hemorrhage that breaks down the atrophic bony trabeculae and can reach giant proportions. Coalescence and collapse of these defects produce marked deformities of the bone ends and flattening and incongruity of the articular surfaces. The cysts are formed by destruction of the overlying subchondral bone and cartilage by vascular connective tissue originating in the marrow spaces immediately adjacent to the subchondral bone. Cartilage surface pitting and focal erosion occur. The entire hypervascular process is associated with osteopenia of the underlying trabeculae. The capsule of the joint becomes thickened. The subchondral cystic cavities communicate at times with the joint space and contain fresh clot as well as fibrous tissue. Key defined the microscopic anatomy with findings similar to those reviewed earlier (107). The synovium was thrown into innumerable folds and villi. The subsynovial tissues were thickened by cellular proliferation and edema and filled with granules of yellowish-brown pigment. The changes were seen universally throughout the joint lining. The surface synovial cell layer was from 4 to 10 cells in thickness. Many of them had pigment granules. There were also phagocytic cells or macrophages loaded with pigment. The subsynovial layer was markedly thickened and composed of a markedly vascularized component, increased collagenous tissue, and macrophagic cells. Beyond this the capsular tissues were markedly thickened and fibrotic. The margin of the articular cartilage was invaded at its periphery by the vascular synovial and subsynovial tissue. The surface of much of the cartilage had undergone fibrosis with little evidence of hyperplasia. The underlying bone was extremely atrophic, and in many places the cartilage virtually rested upon the bone marrow and the cancellous bone of the epiphysis. Speer has delineated well the early pathogenesis of hemophilic arthropathy particularly in relation to the evolution of the subchondral cysts (198) (Fig. 15). His study was based on patients undergoing synovectomy of the knee. Six of the 9 patients were in the younger age group showing early stage changes, but all had had more than two episodes

914

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

of hemarthrosis in the affected knee and chronic synovitis of greater than 6 months duration that was not responding to therapy. The changes described are similar to those reviewed earlier. Speer noted osteochondral defects at central load bearing areas of the femoral condyles with the surrounding articular cartilage relatively spared. The flattened or gently curvilinear space at the bottom of each osteochondral defect was 3-7 mm below the surface of the joint and was invariably depressed below the contour of the normal subchondral bone plate. Fluid-filled osteochondral cysts were seen in several patients, the surface layer of which was continuous with superficial zones of the articular cartilage. The cysts, when unroofed, contained a dark viscous fluid with no evidence of bone or cartilage. Although a pannus was seen, it was not as vascular or as advanced as that in rheumatoid arthritis. In addition, at the osteochondral junction, the fibrovascular pannus appeared to invade the articular cartilage from the surface, but the marrow cavity itself did not show cystic or fibrous changes in contrast to early rheumatoid arthritis. Eventually the fibrous pannus extended to the margin of the osteochondral defects in the central parts of the condyles. At the bases of the osteochondral defects, there were thickened, sclerotic, laminated bone plates depressed below the plate of the normal subchondral bone. It was Speer's feeling that "a subchondral cyst evolves as an expanding hematoma that aggressively destroys the subchondral bone and the overlying articular cartilage before eventually communicating with the joint space. The subchondral cyst progressed to an osteochondral cyst and then to an osteochondral defect." Because these cysts occurred in the intercondylar notch in the central portion of the condyles, they related to the characteristic radiographic findings. Speer defined five mechanisms by which the articular cartilage was destroyed in hemophilic arthropathy: (1) enzymatic digestion of articular cartilage due to the production of proteolytic enzymes including collagenases during the chronic synovitis stage; (2) direct cellular destruction of articular chondrocytes by phagocytic and other cells; (3) direct damage to articular chondrocytes by "blood or other tissue products;" (4) mechanical factors of load pressure, abrasion, and abnormal loading secondary to contractures and osteopenia; and (5) changes resulting from the subchondral hemorrhagic cysts. Speer states that "the cyst of hemophilic arthropathy is the result of subchondral hemorrhage." An inciting event for the subchondral cyst was fracture of the subchondral bone plate with a resulting subchondral hemorrhage. The osteopenic bone would increase the likelihood of fracture. The hemophilic arthropathy begins in the center of the condyle and moves progressively outward, which is different from the progression of rheumatoid arthritis from the periphery toward the center. Due to its position, the subchondral cyst expands with its hematoma by direct bone and articular cartilage tissue destruction until rupture into the joint space occurs. The subchondral cyst thus paves the way for eventual joint destruction as the articular cartilage, which is initially

intact overlying it, subsequently collapses. In summary, the evolution of hemophilic arthropathy begins in the synovium with acute and chronic synovitis. Subchondral cysts develop following subchondral hemorrhage with osteolysis and/or collapse resulting in flattening of the condyles and widening of the intercondylar notch. Fibrillation and wear of the joint surface then occur on both a chemical and mechanical basis. The subchondral cysts expand with lysis of overlying cartilage to form an osteochondral cyst. This can be associated with full thickness articular cartilage loss. In the later stages, chronic synovial fibrovascular pannus grows to undermine the osteochondral junction of articular cartilage peripherally. Bone sclerosis occurs at the base of the osteochondral defects. There is further widening of the intercondylar notch due to further subchondral osteolysis and/or collapse with overlying articular cartilage loss. Stein and Duthie reported on chronic end stage hemophilic arthropathy, analyzing 44 tissue specimens from 39 adult patients during reconstructive joint surgery (201). The mean age of patients having hip, knee, and ankle procedures ranged from 20.5 years for knee synovectomies to 41.5 years of age for total hip replacements. Specimens were obtained from 21 knees, 15 hips, 6 ankles, 1 foot, and 1 wrist. Their assessments understandably showed extensive damage of both the hemophilic synovium and articular cartilage. The synovial lining became progressively fibrotic and the hyaline cartilage disintegrated and eventually disappeared. Both mechanical and chemical processes were considered to cause degeneration of cells, but enzymatic processes in particular appeared to be primarily responsible for degradation of the cartilage matrix. Synovial changes worsened with age. In those less than 20 years of age there was marked villous proliferation, whereas in all adult patients the synovium was progressively replaced showing marked thinness, an increasing tendency to avascularity, and eventual fibrous thickening. In the younger age groups the lining cells of the synovium were heavily filled with hemosiderin. There was inflammation around the synovial hemosiderin deposits. Subsynovial layers were almost totally free of hemosiderin. The cartilage was understandably thinned and degenerated. Fibrous tissue invaded from the synovial regions at the periphery of the joint and from the underlying subchondral bone and marrow. Pyknotic chondrocytes were frequently seen. The cellularity of the cartilage was increased in some areas with chondrocyte clones but in general was decreased with destruction and disappearance of damaged chondrocytes. Connective tissue often covered and partially replaced the superficial layer of the original articular cartilage. Histochemical stains revealed hemosiderin in both the superficial synovial membrane and the chondrocytes of the articular cartilage. Stein and Duthie concluded that the synovial membrane underwent extensive fibrosis with time, with scar tissue replacing the entire thickness including the lining layer. The architecture of articular cartilage was severely disrupted. Chondrocytes degenerated. At one stage they were irregu-

SECTION V ~ Hematologic Disorders

REPEATED HAEMARTHROSES

J/

SUBCHONDRAL HAEMORRHAGES

ACTIVATED PLASMIN

BLEEDS CHRONIC EPIPHYSEAL EASILY SYNOVITIS~ OVERGROWTH DUE TO

HYPERTROPHIC SYNOVIAL MEMBRANE

1 DESTRUCTION OF BIOLOGICAL "SHOCK ABSORBER"

ACCUMULATION OF INTRACELLULAR ~ IRON DEPOSITS / ~ SYNOVIAL FIBROSIS

LIBERATION OF LYSOSOMAL ENZYMES (CATHEPSIN~, D) FURTHER INFLAMMATORY RESPONSE

~

l

cartilage :eikdct:n

ARTICULAR "-* CARTILAGE BREAKDOWN

~ ~=

FIGURE 16 The sequel of changes in the pathogenesis of chronic hemophilic arthropathyas outlinedby Stein and Duthie is shown [reprinted from (201), with permission]. larly dispersed in the matrix and aggregated in cell clusters referred to as clones. With time many areas of matrix were acellular. Fibrous tissue then invaded the cartilage both from the subchondral bone and from the periphery of the synovial lining. Iron deposits were seen both in the synovial cells and even within chondrocytes. All cells containing iron deposits eventually showed degeneration and disintegration. Mechanical, chemical, and particular enzymatic effects of the hemophilic arthropathy led to articular cartilage breakdown. This is illustrated in Fig. 16 from their work. Radiologic Findings: The striking radiographic changes of hemophilic arthropathy were recognized early and in great detail (28, 88). Some felt that the radiographic features of hemophilia, once it entered the nonacute phases, were sufficiently characteristic to allow diagnosis to be made radiographically. These included generalized osteopenia, markedly increased density of the synovial tissues, enlargement of the epiphyseal bone in comparison to the normal side, narrow joint spaces, roughened articular surfaces, cystic lesions in the subchondral region, and widening and deepening of the intercondylar notch of the femur. Several observers had noted clinical evidence of joint involvement in hemophilia before the age of 10 years, with Thomas reporting 89% of 98 hemophilic patients so involved (215). In Thomas' early study of 98 hemophiliacs prior to the era of blood and factor replacement, 78% gave a history of joint involvement and 61% had permanent deformities. The distribution of joints affected was knees 68%, ankles 56%, el-

915

bows 53%, hips 16%, fingers 15%, wrists 5%, and toes 2%. In the earliest stages with acute hemarthrosis, soft tissue swelling could be noted but the radiograph was otherwise unremarkable. As the arthropathy worsened with recurrent bleeds, the soft tissue hypertrophy led to increased density radiographically, and this was often increased because of the high iron content of the synovial and subsynovial tissues. The joint space was often reduced because of degeneration of the articular cartilage. The articular surfaces of the bones became irregular and there were focal subchondral bone cystic defects noted. Petersson et al. based their classification on radiographic changes previously described in the literature, including intra-articular effusion, periarticular soft tissue thickening, periarticular soft tissue calcification, synovial thickening and increased density, enlargement of the secondary ossification centers of the epiphyses, periarticular osteoporosis, narrowed joint spaces, irregularity of subchondral surfaces, subchondral sclerosis, subchondral cysts, incongruence between articular surfaces, erosions of joint margins, and Harris lines (155, 156). They then formed their radiologic classification of changes by determining which of the preceding criteria could be quantified in some form radiologically, and each change was allotted 0, 1, or 2 points according to severity. The sum of points allotted to a given joint at each examination led to its score. The radiologic evaluation classification recommended by the orthopedic advisory committee of the World Federation of Hemophilia is shown in Table V. A widened intercondylar notch of the distal femur seen on the anteroposterior radiograph is a characteristic finding in hemophilia after recurrent hemarthroses. There can also be broadening of the radial head at the elbow. Other radiographic findings included soft tissue swelling, osteopenia, and overgrowth of the epiphysis. Later bone changes included subchondral cysts, marginal erosions, subchondral surface irregularity, widening of the intercondylar notch of the distal femur, squaring of the patella, enlargement of the radial head, and widening of the trochlear notch of the olecranon. The latest stages are characterized by articular cartilage joint space narrowing, flexion contractures, and angular deformity. An additional radiographic classification has been defined by Arnold and Hilgartner (8). Stage 1: There are no skeletal abnormalities visible on radiographs, but there is soft tissue swelling due to hemarthrosis or periarticular bleeding. Stage 2: Subacute hemarthropathy. There is osteoporosis particularly in the epiphyses along with overgrowth of the epiphyses, maintenance of joint integrity with no narrowing, and no bone cysts. Stage 3: Joint disorganization is evident radiologically but with no significant narrowing of the cartilage space. Subchondral cysts, squaring of the patella, denser synovium (hemosiderin deposit), widening of the intercondylar notch of the distal femur, and ulnar trochlear notch widening are characteristic findings. Stage 4: Narrowing of the joint space with articular cartilage destruction. Stage 3 changes are magnified. Stage 5: Fibrous joint contracture,

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematoloyic Disorders

916

Radiologic Evaluation Recommended by the Orthopedic Advisory Committee of the WFH

TABLE V

Finding

Score (points)

Absent Present Absent Present Absent Partly involved Totally involved Absent Joint space > 1 mm Joint space < 1 mm Absent 1 cyst > 1 cyst Absent Present Absent Slight Pronounced Absent Slight Pronounced

0 1 0 1 0 1 2 0 1 2 0 1 2 0 1 0 1 2 0 1 2

Type of change Osteoporosis Enlarged epiphysis Irregular subchondral surface

Narrowing of joint space

Subchondral cyst formation

Erosions of joint margins Gross incongruence of articulating bone ends Joint deformity (angulation and/or displacement between articulating bones)

apossible point score: 0-13 points.

loss of joint space, extensive enlargement of the epiphyses, and disorganization of the joint structures are seen. Radiologic changes are illustrated in Fig. 17. The radiologic changes reflect the cumulative changes following large and smaller subclinical hemorrhages that have occurred. DePalma and Cotler noted irregular sub-periosteal ossification following sub-periosteal hemorrhages in some patients (45). In addition to the other epiphyseal changes described frequently, they commented specifically on premature closure of part or all of the epiphyseal plate, leading to shortening of the bone or angular deformities. DePalma and Cotler did not note a discrepancy in bone length that could be attributed to overstimulation at the epiphyseal plate to the extent that it led to an operable limb length discrepancy. We have noted very few hemophilia patients in the growth study series, despite the fact that a large hemophilia program has existed in the hospital for some time. One of the earliest summaries of accumulated information on the radiology of developing joints suffering from hemophilia was by Caffey and Schlesinger (28). As well as providing a review of earlier scattered observations, this review summarized an era in which no effective treatment for the bleeding was available. Radiologically there was advanced maturation of the bony epiphyses, enlargement of the epiphyses, and irregular and asymmetrical ossification particularly in the hip joint. They described five children who

also showed both advanced and abnormal epiphyseal development. Measurements indicated that all of the epiphyses in the involved joints were larger than their counterparts on the opposite nonaffected side. The sizes of the secondary ossification centers in the various axes were measured. There was early awareness of the acceleration of epiphyseal maturation in those with hemophilic hemarthrosis, and Holmes and Ruggles commented on the squaring and enlargement of the epiphyses in the disorder (88). There was also recognition of Perthes-like changes in the femoral capital epiphysis with pronounced hip bleeds.

Joint Evaluation Grading: Semiquantitative Criteria Established by World Federation of Hemophilia: The World Federation of Hemophilia has established some semiquantitative standards for joint evaluation (69). They consider four parameters involving pain, bleeding, physical examination, and the radiologic evaluation. The scoring system for pain goes from 0 to 3, bleeding from 0 to 3, physical examination from 0 to 12, and radiologic evaluation from 0 to 13. Following the score, any patient requiting aids to ambulation is subcategorized as B for the use of a brace or orthosis, C for cane, CR for crutches, and WC for wheelchair. Pain Scores: A score of 0 is registered if there is no pain, no functional deficit, and no analgesic use except for acute hemarthrosis. A score of 1 indicates mild pain that does not interfere with occupation or activities of daily living (ADL) and may require occasional nonnarcotic analgesic. Moderate pain is given a score of 2 when there is partial or occasional interference with occupation or ADL, use of nonnarcotic medications, and occasional narcotic medication use. A score of 3 for severe pain involves discomfort that interferes with occupation or ADL and requires frequent use of nonnarcotic and narcotic medications. Bleeding Scores: A score of 0 is given if there is no bleeding, a score of 1 if there are no major or 1-3 minor criteria, a score of 2 if there are 1-2 major or 4 - 6 minor criteria, and a score of 3 if there are 3 or more major or 7 or more minor criteria. The guidelines for a minor grading are mild pain, minimal swelling, minimal restriction of motion, and resolution within 24 hr of treatment. Guidelines for major criteria include pain, effusion, limitation of motion, and failure to respond within 24 hr. The number of occurrences is registered in terms of hemarthroses per year. Physical Examination Scores: Physical examination includes grading of swelling 0 - 2 or more with a subgradation for chronic persisting synovitis, muscle atrophy 0 or 1, axial deformity 0-2, crepitus on motion 0 or 1, range of motion 0-2, flexion contracture 0-2, and instability 0-2. Radiographic Scores: The specific criteria here have been listed in Table V. These four groups of criteria are listed in full in the paper by Gilbert published in Seminars in Hematology (69). c. Other Musculoskeletal Damage: Intramuscular Hemorrhage followed by Contractures, Ischemic Muscle Necrosis, and Neuropathies. Ischemia of the calf muscles

F I G U R E 17 Radiographic changes in patients with hemophilic arthropathy are shown. (A) Anteroposterior views of affected right elbow (Ai) and nonaffected left elbow (Aii) show decreased joint space (arrow) and premature fusion of proximal radial physis on the affected side. (B) Anteroposterior knee radiograph (Bi) shows widened intercondylar notch (arrows), joint space narrowing, and irregular subchondral bone. In a different patient, (Bii) changes at skeletal maturity show markedly diminished medial femoral-tibial joint space due to cartilage destruction. (C) Anteroposterior ankle radiograph (Ci) shows joint space narrowing and both distal tibial epiphyseal and talar cysts (lytic areas). Oblique radiograph (Cii) shows joint space narrowing and talar subchondral lysis (arrow).

918

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

can occur following intramuscular bleeding in which a Volkmann-like compartment syndrome evolves. One of the occasional sequelae of such an occurrence is tightness of the gastrosoleus muscle and associated Achilles tendon shortening. This can be effectively repaired with the Achilles tendon lengthening. The most common area of muscle bleeding in some series is within the iliopsoas muscle mass. This can be recurrent and leads to difficulty because of the problems immobilizing the muscle. In one study of muscle bleeding in hemophilia, the sites included the quadriceps muscle (44%), posterior calf muscles (35%), hip adductors (7%), and anterior leg muscles (7%) (9). In the study by Houghton and Duthie, over a 10-year period, muscle hemorrhage was noted in the iliopsoas in 53 (31%), thigh in 40 (24%), calf in 34 (20%), forearm in 32 (19%), shoulder girdle in 7 (4%), and buttock in 4 (2%) (89). The study involved 170 episodes. The complications of muscle hemorrhage involved limb contractures, which were present in 20% of severe calf and forearm bleeds. The other complication was neurological with nerve palsies. In the 170 patients with muscle bleeds, 43 (25%) had an associated peripheral nerve lesion. This was usually a neuropraxia or axonotmesis and, with the exception of the sciatic nerve, full recovery invariably occurred. There were 30 femoral nerve palsies complicating 53 iliacus muscle hemorrhages, and virtually all of these recovered although return to full function often took 6 months. Katz et al. completed a more detailed review of peripheral nerve lesions in hemophilia over a longer period of time, also from the Nuffield Orthopaedic Center in Oxford, England (103). They identified 88 lesions in 54 patients over a 24-year period. The femoral nerve was most commonly involved. There were 61 lesions in which long-term follow-up was obtained. In 30 (49%) of the lesions the nerve had full motor and sensory recovery, in 21 (34%) there was a residual sensory deficit, and in 10 (16%) both a persistent motor and sensory deficit was noted. Patients with antibodies to factor VIII were significantly less likely to recover full motor or sensory function, and the time to full recovery in these patients was significantly longer. The most common nerve involved was the femoral, which was affected in 31 of 61 instances (51%). The next most common nerves involved were the median (10, 16%), ulnar (7, 11%), sciatic (4, 7%), and radial (5%). The most common cause of nerve palsy was intramuscular hemorrhage. Full motor recovery was more likely than full sensory recovery. Fractures: Hemophiliacs are not considered to have an increased incidence of fractures, and with appropriate treatment any fractures that occur unite without complications within the anticipated time. Skeletal traction is contraindicated, but open reduction and internal fixation are often most helpful. In the series of Houghton and Duthie there were 23 fractures over a 10-year period (89). The most common were those of the femoral shaft (10) and tibial shaft (4). Hemophilic Pseudo-tumor: A hemophilic pseudo-tumor is a slowly progressive hemorrhage that increases in size in

a confined space (169). The most common site is the pelvis and proximal thigh, and the lesions may originate within bone, in the periosteal region, or occasionally in adjacent soft tissues. Because a pseudo-tumor is an expanding hematoma or hemorrhage it must not be biopsied or aspirated for diagnosis. No malignant change of a pseudo-tumor has been documented, although they can be confused with malignancy initially. Pseudo-tumors have always been infrequent in hemophilia and with improvements with medical therapy they are becoming even less common. They are dramatic lesions, however, and have the potential to lead to disastrous results with inappropriate intervention; hence, a clear understanding of them is essential. Rodriguez-Merchan reported on 17 pseudo-tumors within bone (169). Eight were in the femur, 2 in the pelvis, 2 in the radius, 2 in the hand, and 1 each in the tibia, humerus, and big toe. The pseudo-tumor is a progressive cystic swelling produced by recurrent bleeding involving muscle and accompanied by bone involvement. The most common sites are the proximal long tubular bones in which repeated and unresolved intramuscular hematomas may lead to encapsulation and calcification with a progressive enlargement of the mass and erosion of the adjacent bone. The pseudo-tumor is an encapsulated hematoma. The hemorrhage is chiefly extraosseus with extensive subperiosteal bleeding and reactive new bone formation. The mass expands both externally and internally causing extensive destruction of the adjacent bone. They are most commonly proximal particularly around the femur and pelvis where they start in the soft tissues and erode bone secondarily. They tend to develop slowly over many years. They occur in adults and do not respond to conservative treatment. Calcification within the mass is common. Biopsy is absolutely contraindicated because of the danger of severe and perhaps lethal hemorrhage. Many of these are simply observed and appear to progress slowly over several years with minimal damage. In those in whom there is discomfort, unacceptable bone and joint disruption, or associated neurological problems, treatment is mandatory. Extensive preoperative workup with multiple imaging techniques is essential. It is extremely important to outline the associated vascularity. The best approach, if possible, is to remove the entire mass en bloc with clear attention to its vascularity. Efforts to curette and pack the lesion are contraindicated. 4. CURRENT TREATMENT CONSIDERATIONS a. Medical Treatment. The treatment of hemophilia became effective in 1940 when whole blood transfusions were utilized to replace the missing factor VIII and IX components (47, 66, 90, 129, 227). Treatment shortly evolved into the use of fresh-frozen plasma. In 1964, cryoprecipitate was developed; it represents the protein that precipitates in freshfrozen plasma when it is thawed at 4~ which is rich in factor VIII and fibrinogen. In 1970, specific factor concentrate began to be used. Each of these methods, however,

SECTION V ~ Hematologic Disorders

919

FIGURE 18 Massiveextent of a hemophilicpseudo-tumoris shown. (A) Radiograph of pelvis shows extensiveinvolvementof left iliac bone (P). (B) Magnetic resonance image (cross section) shows extensive bone lysis and soft tissue mass (P) both anterior and posterior to iliac wing.

required the use of human blood, a situation that became devastating in terms of complications with transmission of the AIDS virus. At present, both factor VIII concentrates and cryoprecipitate are prepared from donors screened carefully for the human immunodeficiency virus (HIV). In addition, all factor VIII preparations are processed further to diminish the likelihood of any viral contaminants. Methods to destroy viruses are based on terminal heating of the lyophilized product at 80~ heating in solution at 60~ (pasteurization) in a suspension containing various organic solvents, or adding an organic solvent and a detergent during the manufacturing process. Heating serves to inactivate heat-sensitive viruses such as HIV-1. Other preparations involving solventdetergent-extracted factor VIII remove coated viruses including HIV-1 and hepatitis viruses. Because factor VIII concentrates are obtained from screened donors and have other antiviral steps performed, their use is recommended today over cryoprecipitate. The risk of transmission of hepatitis B and C viruses has also diminished because of the use of virucidal methods, but it has not been abolished completely. Many clinics turn increasingly to even safer factor VIII concentrates using either monoclonal factor VIII concentrates prepared using affinity chromatography or recombinant factor VIII, which is a synthetic product prepared in Chinese hamster ovary cells or baby hamster kidney cells. The monoclonal or recombinant forms of factor VIII have essentially no risk of hepatitis, AIDS, or hemolysis. Replacement therapy in hemophilia B includes the infusion of factor IX concentrate or fresh-frozen plasma. Cryoprecipitate and factor VIII concentrate play no role. The hemophilia community was hit particularly hard during the time period when AIDS began to flourish but its recognition within blood bank samples was neither appreciated nor specifically screened. In some clinics, approximately 90% of patients with hemophilia who had been

multiply transfused with untreated products became HIVpositive. One study showed 55% of patients with hemophilia infected with HIV-1 from 1979 through 1985, with even larger percentages of this population infected with hepatitis C. A study of mortality in patients with hemophilia between 1986 and 1992 showed that nearly one-third of deaths (27%) were related to HIV infection and only one-fifth were related to hemorrhage. Prior to the AIDS epidemic, excess mortality among hemophiliacs had always been due to intracranial hemorrhage. The magnitude of the problem was also illustrated by Rodriguez-Merchan in a review of statistics from Spain (171). Of the 435 hemophiliacs in his unit, 257 (59%) had HIV infection caused by therapy with human plasma. Of these HIV-positive patients, 95 (37%) had already developed full-blown AIDS. In Spain roughly 70% of hemophilic patients who received pooled, untreated factor VIII preparations became HIV-positive, with 90% of severe hemophiliacs so affected. One estimate of the occurrence of AIDS in hemophilia in North America listed 6% of the entire hemophilia population as having or having had AIDS. As of June 1992, more than 1600 patients with hemophilia A and no other risk factors had been given a diagnosis of AIDS. Greene et al. demonstrated that effective surgery could still be performed in hemophilic AIDS patients and listed preoperative evaluation criteria (80). AIDS was first identified in 1982 in a patient with hemophilia who had received transfusions. This number has diminished dramatically in new patients over the past few years. All blood has been screened for the HIV virus, and manufactured monoclonal coagulation factors increasingly are used, thus bypassing completely the need for human blood. Fresh-frozen plasma (FFP) contains all coagulation factors but is rarely used. Single donors with a low risk of hepatitis or other viruses continue to provide blood products needed for some cases. Cryoprecipitate contains factor VIII

CHAPTER 10 ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

920

TABLE VI

Brief Outline o f Factor T r e a t m e n t for Musculoskeletal Bleeding in Hemophilia A a n d Ba Hemophilia A

Hemophilia B

Initial dose

Repeat dose

Initial dose

Repeat dose

Site or situation

(U/kg)

(U/kg)

(U/kg)

(U/kg)

Acute hemarthrosis Early

15-30+

Rarely needed

20

Rarely needed

25-50+ 25-50+

25 ql2h 25 ql2h, often for several days

40-80+ Mild to moderate: 20-40 Severe: 80

20-25 q24h 30 q24h, often for several days

25 q-12h or 25 followed by continuous fusion IV 3-4 UNg/hr

80

40 q24h or loading dose plus continuous IV fusion

Late Intramuscular hemorrhage

Major surgery or trauma

50

Desired range of factor levels

Early (less than 3 hr from onset): raise factor level to 25-30%, repeat if needed clinically Late: raise factor level to 50-100% Major: calf, forearm, iliopsoas, hip; raise factor level to 100%, maintain several days to several weeks, taper range to 30-50% for additional periods Minor: other areas, superficial bleeds, raise factor level to 50%, usually for 2-3 days only Initially 100%; 50% early heeling; 30% late heeling

aAbstracted from Boston Hemophilia Center guidelines, 2/98, B. Ewenstein, E. Neufeld, J. Gorlin; see also D. DiMichele, ref 47; + lower ranges represent "textbook" responses of 1 U/kg = 2% factor increase.

and fibrinogen and still can be used in hemophilia A and von Willebrand's disease. Each bag is also manufactured from one donor with a low risk of hepatitis or other viruses. The most commonly used treatment now is genetically engineered recombinant factor VIII or plasma-derived, monoclonal-antibody-purified factor VIII for hemophilia A. Desmopressin (DDAVP) increases the plasma levels of factor VIII and von Willebrand factor and can be used for nontransfusional treatment of patients with mild or moderate hemophilia and von Willebrand's disease (128,227). An outline of factor replacement in varying bleeding and operative situations is shown in Table VI. b. Early Intervention for Bleeding Episodes. The most common approach to medical management in North America has involved early intervention for bleeding episodes with increased factor given at the onset of discomfort and, if possible, even prior to the observation of joint swelling. This approach is often referred to as an on-demand treatment. This treatment involves the use of home infusion of factor VIII without the need to be seen initially at a hospital or a clinic. Most clinics and families use this approach, which is based on patient and parent awareness of the earliest symptoms of disorder and easy communication between families and comprehensive hemophilia clinic personnel. c. Prophylactic Factor VIII Coverage. Over the past several years, a treatment approach involving continual prophylactic use of factor VIII has become more popular particularly in western Europe (2, 143, 182). This approach seeks to maintain the factor VIII levels in a therapeutic range of 1-5% at all times rather than reacting to the occurrence of

joint bleeding. This level is chosen because moderate severity hemophiliacs rarely develop significant arthropathy. The prophylactic treatment converts severe patients to a moderate or even mild grading. The infusion of factor VIII is done three times weekly and factor IX twice weekly. Trials in Europe extending as far back as three decades have been positive. A major Swedish study of prophylactic treatment for hemophilia A and B in 60 severe hemophiliacs (52 A, 8 B) has been reported (143). Treatment was started when the boys were 1-2 years of age with a regimen of 24-40 IU of factor VIII kg -1 three times weekly for hemophilia A (greater than 2000 IU kg -1 annually) and 25-40 IU of factor IX kg-1 twice weekly for hemophilia B. Of those subjects aged 3-17 years, 29 out of 35 individuals had joint scores of 0 (normal). Factor VIII and IX concentrations were designed not to fall below 1% of normal. The study concluded that it appeared to be possible to prevent hemophilic arthropathy by giving effective continuous prophylaxis from an early age and preventing factor VIII and IX concentrations from falling below 1% of normal. The treatment served to convert hemophilia from a severe to a moderate form, thus preventing joint bleeds and hemophilic arthropathy. The best results were obtained when treatment began at the age of 1-2 years with factor VIII given in dosages of approximately 3000 IU kg -1 annually. Studies have also been performed in Germany, with one study analyzing 70 patients treated actively when bleeding occurred and 17 receiving factor VIII regularly on a prophylactic regimen for a mean of 4.5 years (182). Both pediatric and adult patients were included. The mean age of the pro-

SECTION V ~ Hematologic Disorders

phylactic group was 18.9 years. The prophylactic group missed fewer days of work due to joint bleeding. There was less differential in relation to joint pain, but the on-demand therapy patients experienced more pain with increasing age. The age of the patient was considered to be more important than treatment strategy in the development of joint symptoms. To be most effective, prophylactic regimens should be started in the very young prior to any occurrence of joint pathology. Because of the extremely high cost of prophylactic treatment, the German group felt that specific indications for its use should be developed rather than simply offering the treatment to all hemophilic patients with severe involvement. A longitudinal study of orthopedic outcomes for severe factor VIII deficient hemophilia was performed by Aledort et al. (2). Twenty-one international hemophilia centers were used to assess patients over a 6-year period. Physical and radiologic examinations of ankle, knee, and elbow involvement were assessed. The study population involved those with severe factor VIII deficiency (less than 1% factor) under the age of 24 years without inhibitors. Some patients were treated on a prophylactic regimen, but most had an ondemand response for any joint bleeding. The authors concluded that year-long prophylaxis significantly reduced the rate at which joints deteriorated by both physical examination and radiographic criteria. Patients on prophylaxis had significantly fewer days lost from work or school and fewer days spent in the hospital. Aledort et al. concluded that higher doses of factor in themselves did not produce improved orthopedic outcomes but that full-time prophylaxis was likely to produce the best orthopedic outcome. The most critical factor for a good orthopedic outcome was diminution in the number of joint bleeds. Full-time prophylaxis had been assessed in 66 of 477 patients with appropriate physical examination data and 53 of 323 patients with full radiographic data. These assessments led the authors to conclude that prophylaxis significantly reduced the rate at which joints deteriorated and that it did so by dramatically reducing the number of bleeds, thus minimizing joint damage. The number of days lost from work or school was significantly lower in the entire population studied for those who received > 2000 units of factor VIII/kg per patient per year. Patients treated with on-demand therapy progressed significantly more than those on prophylaxis. The obvious observation was made, however, that those on prophylaxis required significantly more factor than those on on-demand therapy. The authors felt that their findings strongly supported the previously mentioned Swedish experience that a high dose of regular prophylaxis starting at a very young age can maintain hemophilic joints in an essentially normal range for prolonged periods of time. Prophylaxis, however, could decrease the rate of deterioration of patients' joints if begun after some damage was already present. If prophylaxis was not performed, doses of factor VIII from 25 to 40 units/kg per bleeding episode were best in reducing the progression of arthropathy. This approach is currently reserved for those

921

in the severe category, partly because it is enormously expensive and the need to maintain comfortable, infection-free continuous intravenous access can be problematic. d. Inhibitors. Despite continuing improvement in the effectiveness of treatment for hemophilia A, the development of factor VIII inhibitors remains a serious problem. Factor VIII is a foreign protein for most patients with severe hemophilia. Because the antibody develops against factor VIII itself the fact that monoclonal purified concentrates are used increasingly does not change the development of the inhibitor effect. Hoyer has summarized the various strategies used for patients with inhibitors (90). Possible reasons for high numbers of inhibitor patients include the increased use of factor VIII, for example, in prophylactic regimens, and perhaps stronger antigenicity of clotting factors. A study by Ehrenforth et al. clearly delineated the severity of the problem, although they did not feel that antibody formation was greater using the monoclonal recombinant factor VIII (53). In a long-term study, they assessed 63 children with hemophilia A and 17 with hemophilia B. Only those with severe and moderate forms were analyzed because very few patients with mild hemophilia of either type ever develop inhibitors. Factor VIII inhibitors developed in 15 of 46 (33%) hemophilia A patients. The percentage was greater in severe hemophilia, with 14 of 27 (52%) developing inhibitors compared with only 1 of 19 with moderate hemophilia. If patients with mild hemophilia were also taken into account, the proportion affected was still 24% in the entire group (15 of 63). Inhibitors developed quickly with the mean number of exposure days only 11.7. Inhibitors developed by age 1 year in 33% of the affected patients, by the age of 2.5 years in 73%, and by the age of 5.2 years in all who were destined to develop them. In no patient did an inhibitor first develop after age 6 years. Virtually all patients with factor VIII deficiency who developed inhibitors were those in whom factor VIII activity was less than 3% (53). Because previous studies tended to include all patients the number of patients affected with inhibitors was relatively lower. If one limits the study to severe and some moderate patients, the number correspondingly increases. New inhibitor formation occurred in 24% of all hemophilic patients, in 33% of those with factor VIII levels of 5% or less, and in 52% of severe hemophilic patients with factor levels less than 1%. In larger studies inhibitors rarely develop after the age of 11 years, and the greatest risk of inhibitor formation is before the age of 5 years. Data in this study are higher than those reported elsewhere. The general consensus is that in moderate and severe hemophilia A inhibitors develop in 20-33% of patients. Inhibitors are much less common in factor IX deficiency. Previous reports indicated that inhibitors to factor IX developed in only about 3% of all patients with hemophilia B or 7-10% of those with severe disease. e. Pain M a n a g e m e n t . The best form of pain management in hemophilia is the use of factor VIII or IX replacement plus splinting of the involved joint. Analgesics are

922

CHAPTER IO

~

Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

often helpful. The use of aspirin is contraindicated because of its tendency to further limit platelet adhesiveness and particularly to predispose one to gastrointestinal bleeding. Nonnarcotic analgesics such as acetaminophen are favored particularly in the childhood and adolescent age groups. On occasion, acetaminophen and codeine or propoxyphene can be added. If chronic hemophilic arthropathy causes discomfort because of the arthritic component, nonsteroidal medications can be used. Trilisate is preferred, but ibuprofen and naproxen can also be used with close observation for a change in bleeding patterns. f. Orthopedic Management. The orthopedic management of hemophilia has been modified and aided greatly by the use of factors VIII and IX. Factor VIII and IX concentrates treated to inactivate viruses or produced by recombinant technology in viral-free environments prevent or control bleeding in most patients. Whether factor VIII is used as a prophylactic or immediately upon the onset of joint symptoms, the results have markedly improved the management profile from those described two and three decades ago. The mainstay of orthopedic management for intra-articular, periarticular, or intramuscular bleeding is rest of the involved limb, factor replacement, and gentle reinstitution of motion and muscle strengthening once the bleeding symptoms are controlled. Joint aspiration plays virtually no role currently in the management of the hemophilic patient. The nature of the immobilization during the acute phase is dependent on the joint involved as well as the age and activity level of the patient. We never use solid circumferential casts in the upper extremity in which the elbow is most commonly involved, preferring either a sling in a particularly cooperative patient or a sling and bivalved long arm cast in one who is less cooperative. Increasingly, a bivalved cylinder or long arm cast is used when splinting and/or factor VIII replacement has been ineffective. Immobilization of the knee involves the use of a knee immobilizer or bivalved long leg cylinder cast and crutches, and at the ankle either a posterior splint, a bivalved short leg fiberglass cast, or a lightweight air splint is used along with crutches. Immobilization is recommended for a few days followed by active range of motion exercises and eventually by muscle strengthening as part of the rehabilitation process. The ankle has proven to be the most difficult joint to maintain at a high functioning level. Ribbans and Phillips stated that there was no evidence that the pattern of ankle function had improved despite advances in medical therapy over the previous four decades (164). g. Recalcitrant Joints. On occasion, the return of normal function with full resolution of synovial thickening and joint effusion is difficult and more intensive or prolonged mechanical treatment measures are needed. It remains extremely important to quieten the synovitis such that splints or bivalved casts must be used frequently. It is not appropriate, in our opinion, to rely completely on factor replacement to treat a hemophilic hemarthrosis. If there is a recurrence of

bleeding after seemingly quietening the disorder, then further protection particularly at the ankle area can be achieved with a solid ankle AFO followed by a hinged AFO to allow for joint motion in a controlled plane. In situations in which the synovitis and bleeding are recurrent even with seemingly excellent control of factor VIII, two possible modalities have been used to control the synovitis. h. Surgical Synovectomy. Prior to the availability of factor VIII concentrate in the late 1960s, major reconstructive surgery was simply not attempted for patients with hemophilia. Once factor replacement became feasible, however, orthopedic surgical intervention was used widely to help alleviate some of the damage to involved joints. Synovectomy has played a major role in the surgical treatment of hemophilia ever since factor VIII and IX replacement became feasible. It is resorted to when recurrent joint hemarthroses prove to be difficult or impossible to control with factor replacement and nonoperative orthopedic measures. In many instances, the synovitis reaches a degree of magnitude that appears manageable only by synovectomy. Removal of the synovium from a hemophilic joint is a major debulking procedure that removes a thickened, hypertrophic, hypervascular mass of tissue, which is releasing toxic enzymes such as cathepsin D into the joint environment, serving as a source of repeat hemorrhage due to its extreme friability, and showing ineffectiveness in resorption. The procedure is not without its complications, however, and considerable controversy continues in regard to indications for it. Some clinics resort to it after two or three episodes of poorly controlled hemarthrosis, whereas others wait for several episodes until the synovium is so hypertrophied that resolution of a spontaneous nature would appear impossible. The procedure itself can induce considerable bleeding and fibrosis such that postoperative factor and rehabilitation management must be extremely rigorous. There must be maximal hemophilic control. Rehabilitation must also be maximized. The use of the continuous passive motion apparatus has been extremely helpful particularly in relation to synovectomy at the knee. With the increasing use of arthroscopic synovectomy, a further decision must be made between open and closed procedures. Much of the choice is dependent upon the expertise of the surgeon, but we have been more impressed, both in personal experience and in reviews of the literature, with open synovectomy for the knee as well as the elbow, hip, and ankle, although the closed procedure in expert hands has much to recommend it. A lengthy midline or medial parapatellar incision is used for the knee. Depending on the extent of synovial hypertrophy, a posterior incision may also be needed. In some adolescent patients a Baker's cyst is present, which virtually mandates an additional posterior incision for removal. Use of the continuous passive motion apparatus for 10 days to 2 weeks postsurgery has been extremely helpful in minimizing postoperative stiffness and eliminating the need for manipulations with all the risks they entail.

SECTION V ~ Hematologic Disorders

Synovectomy was first used widely for hemophilic arthropathy in the mid-1960s, with early reports by Storti et al. of Italy noting positive results (206). Storti and Ascari reported that they had performed 63 synovectomies for hemophilia A and B between 1966 and 1975 (205). By removing the highly vascularized synovial tissue they sought to prevent or at least minimize further bleeding. The synovectomy was thus viewed as a "hemostatic" procedure. The operation was performed when the synovial membrane had undergone a series of hypertrophic changes after repeated hemorrhages such that it was highly prone to further bleeding with minimal to no trauma. The increased vascularization was accompanied by proliferation of the synovial lining into innumerable villi, each with a central vascular stem. A selfperpetuating cycle was established in which hemarthrosis led to synovial hypertrophy, which itself was highly vascularized and led to recurrent hemorrhages with even more minimal trauma. Review of the results following knee synovectomy showed a significant diminution of recurrence of hemarthrosis episodes. In 48 knee joints (94%) the markedly reduced recurrence of hemarthrosis was observed. In 34 (66%) the hemorrhage had been completely stopped; whereas in 14 (28%) there were occasional episodes of mild knee swelling of short duration, these being much less severe than the preoperative episodes. Storti and Ascari also noted increases in joint function with knee joint mobility improved in 28 cases (55%), unchanged in 17 cases (33%), and worsened in only 6 cases (12%). In their discussion, they reviewed results from several European clinics during the same time period also showing good to excellent results from the procedure and the associated medical control. Luck and Kasper reported on knee synovectomy and debridement, either open or arthroscopic, in 62 patients (125). Results reported in open and closed arthroscopic synovectomy for the knee have been roughly comparable. In general, there is perhaps slightly increased range of motion maintained after the arthroscopic approach, but most feel that open visualization allows for a more effective removal of synovial tissue. Rodriguez-Merchan et al. performed 18 open and 9 arthroscopic synovectomies and felt that the procedure by both methods significantly reduced bleeding episodes (172). Arthroscopic synovectomy can lead to good short-term and long-term results (229). It was recognized, however, that these were still essentially palliative procedures, which appeared to slow but did not halt deterioration of the joints. One of the problems following knee synovectomy has been stiffness. In a study by Montane et al. completed before development of the concept of postoperative continuous passive motion, knee synovectomy in 10 of 13 patients was followed by a mean loss of 41% of motion (141). De Gnore and Wilson reported on 34 synovectomies: 16 of the knee, 15 of the elbow with radial head excision, and 3 of the ankle (44). The average number of bleeding episodes preoperatively was 26 per yearin the target joint, and with an average follow-up of 4.8 years there was a marked diminution in the

923

number of bleeding episodes after surgery to 4 per year. The continuous passive motion machine was utilized in the first week postsynovectomy. The general consensus of opinion is that synovectomy relieves pain, decreases swelling, and diminishes the number of bleeding episodes per year. It is clearly not curative and one of the earlier complications was joint stiffness. Post et al. reported on a 5-year follow-up of 12 knee and 4 elbow synovectomies with the previously mentioned impressions (161). In their small group of patients, elbow motion was reduced significantly. Favorable results were also reported in open synovectomy of the elbow by Balc'h et al. (11). They reported on 23 elbow synovectomies in 18 patients 8-25 years of age. Those elbows operated after skeletal maturity frequently had associated resection of the radial head. The authors noted a significant improvement in mobility for pronationsupination in 9 elbows and for flexion-extension in 14. Episodes of bleeding were diminished markedly. They strongly recommended the open synovectomy performed through a single lateral incision in those hemophiliacs in whom the nonoperative treatment had failed. i. R a d i a t i o n S y n o v e c t o m y . Radiation synovectomy has been used in some centers but has not received governmental approval for unrestricted use in the pediatric age group in the United States. Radioactive substances are injected into the joint to damage and/or kill synovial cells once they are absorbed into them (63, 134, 194). Radiation synovectomy has been reported using radionuclides such as yttrium-90, gold-198, phosphorus-32, and Rhenium-186. Merchan et al. reported on a long-term series of results with the use of gold-198 in the treatment of hemophilic synovitis of the knee (134). They described treatments in 38 males with a minimum follow-up of 13 years and representative of treatment between 1974 and 1976. One hundred affected joints in 64 patients were injected with a single dose of gold-198. Follow-up was limited to those who had survived and had not had subsequent surgical synovectomy. The results based on the difference between the clinical joint score of the year before injection and the year of review were good in 8 (21%), fair in 23 (60%), and poor in 7 cases (19%). There were more good results, 31%, in those with a stage I (milder) grading than in those with a stage II (more advanced) grading at the time of intervention, in whom only a 14% good rating was achieved. Good and fair results were also more common in those whose symptoms were less than 1 year before injection compared with those greater than 1 year. Merchan et al. concluded that there was a degree of effectiveness that might subsequently be improved with other radioactive agents or different protocols. Fernandez-Palazzi et al. briefly reviewed the use of radiation synovectomy for hemophilic hemarthrosis (63). They reported on 50 procedures in 43 of their patients, whose ages ranged from 6 to 43 years. They concluded that there was a significant reduction of hemarthroses, immense diminution of the amount of

924

CHAPTER IO ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

antihemophilic factor needed, and a markedly diminished period of treatment needed for the patient because basically only an injection was required. Fernandez-Palazzi et al. estimated that good results had been obtained in 80% of patients, which was felt to be similar to the results with the more invasive surgical synovectomy procedures. A possible problem with the technique is concern that the damaging effects of the radioactivity are not strictly limited to the synovium but could clearly also affect the articular chondrocytes that already are abnormal because of the repeated episodes of hemarthrosis. Although this has not been studied in detail, it remains a primary biologic concern. j. Soft Tissue Releases and Lengthenings. Once a joint has been effectively quietened either with medical and physical therapy alone or with the association of synovectomy, persistent tightness and rigidity can be improved with soft tissue releases and tendon lengthenings. Those areas most frequently approached by these techniques are the posterior knee and particularly the ankle area where Achilles tendon lengthening with or without posterior ankle capsulotomy can be most beneficial. k. Osteotomy. In many patients, angular deformity occurs due to the relative overgrowth of one side of a physis in relation to the other, which can cause valgus, varus, or flexion deformities particularly at the major lower extremity joints, the knee and ankle. Fibrous contractures also contribute to the malalignment. In those instances in which rehabilitation physical therapy is not effective or in which bony deformity is too great, osteotomy has been shown to improve alignment. The most common joint deformity is a flexion contracture of the knee, often associated with a valgus, external rotation deformity. Distal femoral supracondylar osteotomy may be required especially if 25 ~ of flexion persists with nonoperative management. The hyperextension osteotomy should always be accompanied by some bone shortening to prevent neurovascular bundle stretching. Good results have been reported at the distal femur (196), proximal tibia and fibula, and also distal tibia and fibula in the supramalleolar region (152). 1. Arthrodesis. Once operative therapy became feasible with appropriate factor VIII and factor IX coverage, many joints had been so damaged that arthrodesis alone appeared to offer effective relief. This allowed for the correction of angular deformity and elimination of pain. The joint fused most commonly and most effectively has been the ankle. With the increasing effectiveness of total hip and total knee arthroplasty, these two joints now rarely are considered for fusion. When bleeding occurred in the ankle in the second decade, there was a tendency for relatively rapid degeneration of the joint to occur with flattening of the talus, markedly diminished joint motion, overgrowth of the medial side of the distal tibial epiphysis, and a valgus alignment of the ankle joint. Gamble et al. protected the ankle by casts and orthoses, but they felt that compliance with treatment in this age group was poor (67). In those joints not responsive to

treatment, synovectomy was apparently not used and with degeneration of the joint there was a relatively early resort to arthrodesis. The results were good in terms of stabilizing the ankle, correcting equinus deformity, and eliminating discomfort and recurrent episodes of bleeding. The most common approach is still to perform a synovectomy, and reports of this in hemophilia have been made. m. Adult Surgical Treatments. Although treatment of the adult hemophiliac is beyond the scope of this presentation, it should be noted that total hip and total knee arthroplasty have been highly effective in adult patients with destroyed joints. The complication rate is much higher than in otherwise uncomplicated joint arthroplasties, but relief of discomfort and improvement of range of motion are most beneficial to the patients. Other procedures in the adult with severe joint destruction have involved synovectomies, excision of the radial head at the elbow, fusion of the ankle and knee, and patellectomy, meniscectomy, and debridement of the knee. An excellent review of 168 procedures over a 20-year period in end stage adult hemophilia was presented by Luck and Kasper (125).

B. Von Willebrand Disease Von Willebrand disease is a mild bleeding disorder, which must be distinguished from hemophilia (165). It is now recognized to be the most common congenital bleeding disorder, affecting as many as 1% of the population. Von Willebrand factor plays two major roles in normal hemostasis (226). It is the cofactor for platelet adhesion and the carrier protein for factor VIII. When abnormal yon Willebrand factor fails to bind to factor VIII, the clinical picture resembles hemophilia. It is very rare, however, that the clinical bleeding disorder is anything more than mild. Most of the clinical manifestations of yon Willebrand disease relate to superficial or postoperative bleeding such as epistaxis, menorrhagia, and cutaneous hemorrhage. There are three types of yon Willebrand disease (vWD). (1) Type 1 vWD. Most common type, affecting 70-80% of patients. Caused by partial shortage of vW factor protein. (2) Type 2A vWD. Affects 15-30% of people with vWD. Caused by absence of an important part of the vW factor protein, high-molecular-weight multimers (HMWM), which are needed to help form blood clots. (3) Type 2B vWD. Affects 15-30% of people with vWD. The vW factor protein does not work properly and binds to platelets in the wrong way. This can lead to both a shortage of vWD protein and a shortage of platelets. (4) Type 3 vWD. This is the rarest form of vWD, affecting only 1 person per million. The blood does not clot properly due to an almost complete shortage of vW factor protein. These patients also may have a shortage of factor VIII. Treatment for vWD in the milder variants is by desmopressin available as a highly concentrated nasal spray, although it also can be used as an injection (DDAVP-

SECTION V ~ Hematologic Disorders

desmopressin acetate). In the more severe forms or when surgery is needed, the treatment of choice is replacement therapy with a factor concentrate. The only vW factor protein concentrate approved for use in the United States today for vWD is Humate-P antihemophilic factor-von Willebrand factor complex (human) manufactured by Aventis Behring (165).

C. Hemoglobinopathies: Sickle Cell Anemia and Thalassemia 1. OVERVIEW The hemoglobinopathies can lead to bone growth problems primarily due to osteonecrosis, which is a characteristic of sickle cell anemia, and premature fusion of physes, which is a characteristic of thalassemia. In these diseases autosomal dominant mutations of the hemoglobin gene lead to disorders of hemoglobin structure and a tendency to premature breakdown and change of cell shape from a round concave disk to a pointed ellipse or sickle-shaped cell (25, 46, 116). The three most important abnormal hemoglobin genes are the sickle gene hemoglobin S, hemoglobin C, and the thalassemia gene. When an abnormal hemoglobin gene is inherited from both parents, the patient will show the clinical features of a hemoglobinopathy. When an abnormal hemoglobin gene is inherited from only one parent, there are usually no clinical manifestations and the involved patient is said to have the sickle trait. Hemoglobin C abnormalities tend to be mild. Sickle cell disease is a hemoglobinopathy that causes sickling of the red cells, leading to thromboembolic infarcts in bone. There is increased hemolytic destruction of inefficient sickle red cells, and subsequent erythroblastic activity increases the volume of the marrow cavity, leading to the appearance of increased lysis radiographically. The normal fetal hemoglobin (HbAF) is gradually replaced by sickle cell hemoglobin (HbSS) within the first year of life. The normal hemoglobin genotype is HbAA. When sickle hemoglobin (hemoglobin S) is deoxygenated, the replacement of [36glutamic acid with valine results in a hydrophobic interaction with another hemoglobin molecule, triggering an aggregation into large polymers (25). The polymerization of deoxygenated hemoglobin S is the primary event in the pathogenesis of sickle cell disease, resulting in distortion of the shape of the red cell and a marked decrease in its deformability. These rigid cells are then responsible for the vasoocclusive phenomena that characterize the disease. 2. CLINICAL CHARACTERISTICS Those with severe sickle cell disease, generally referred to as sickle cell anemia, are affected by multiple episodes of pain and by sickling crises, which lead to vaso-occlusive disorders. These in turn can lead to stroke, acute chest syndromes, impaired neuropsychological function, and premature death. Sickle cell disease is present when the abnor-

925

mal hemoglobin gene is inherited from both parents (HbSS). When the gene is inherited from only one parent, there are usually no clinical manifestations and the patient is said to carry the sickle trait only (HbAS), which has no clinical significance in relation to the skeleton. The disorder is far more problematic in homozygous HbSS than in heterozygous HbAS or HbSC. Acute osteomyelitis is common, and often the early stages cannot be distinguished clinically from the vaso-occlusive crises of bone infarcts. Chronic osteomyelitis can occur. Salmonella is the major organism involved with osteomyelitis in this disorder, accounting for 65-80% of cases. Other problems involve leg ulcers and growth retardation particularly in the early years of life in those with active crises. This rarely leads to short stature because in many skeletal maturation is delayed, allowing normal height to be reached. A large study of 3578 patients with sickle cell disease found the number of pain episodes per year to correlate well with frequency of death in patients greater than 20 years of age, with those with high rates of pain episodes tending to die earlier than those with low rates (160). High rates were associated with high hematocrit and low fetal hemoglobin levels. Mortality was greatest in sickle cell anemia and sickle [3(0)-thalassemia and lower in hemoglobin SC disease and sickle [3(+)-thalassemia. The fetal hemoglobin level had a strong influence on the pain rate without a threshold effect. Increments in the fetal hemoglobin level were beneficial even when the level was low. The presence of fetal hemoglobin in effect converted the severity of the sickle cell disorder to those with the milder hemoglobin SC disease. Thalassemia is a form of hemolytic anemia and sickle cell disease common in those of Italian and Greek descent and previously was referred to as Mediterranean anemia. The severe variant referred to as sickle [3-thalassemia (S [3thalassemia) results from inheritance of a sickle (S) gene from one parent and a [3-thalassemia gene from the other. In the " + " type some normal [3-chain is produced, and in the "0" type no normal [3-globin is produced. Even in the severe forms, bony changes are only slight during the first year of life. Bony abnormalities frequently affect the femurs and vertebral bodies. The vertebral bodies are osteopenic and may be flattened with prominent cupping due to displacement of the adjacent intervertebral disks. The cortices are thin and the trabeculae are markedly reduced in number. On occasion, there is shortening of the spine without deformity, but instances of scoliosis or kyphosis also occur. The cortex of the long bones tends to be thinned and the marrow cavities are widened and prominent. Angular bone deformities occur due to the structural weakness. Premature fusion of the epiphyses of long tubular bones has been noted in severe cases of thalassemia. When this is asymmetric in relation to left and right femurs or tibias, length discrepancies occur. When asymmetric involvement occurs within the same epiphysis, angular deformity also is seen. The head of the humerus is often found to be tilted medially into a varus position.

926

CHAPTER 10 9 Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

3. MOLECULAR AND OTHER APPROACHES TO THERAPY

It was felt that therapy with drugs such as hydroxyurea, which increased the production of fetal hemoglobin, could reduce the pain rate and by implication improve survival (159, 160). Molecular approaches to therapy currently being investigated include chemical inhibition of hemoglobin S polymerization, reduction of the intracellular hemoglobin concentration, and pharmacological induction of hemoglobin F because F is a very potent inhibitor of the polymerization of deoxyhemoglobin S (25, 57). Hemoglobin E therefore, inhibits sickling and drugs that could increase its synthesis would be expected to benefit patients with sickle cell disease. Hydroxyurea has been used widely to enhance hemoglobin F production in patients with sickle cell anemia. Bone marrow transplantation has been used effectively to cure sickle cell anemia, but the complications of this treatment can be devastating and considerable concern exists as to its advisability for this disorder, (159, 224). At a minimum, only severely symptomatic cases of sickle cell disease are considered in most centers (sickle cell anemia, HbSS; sickle cell hemoglobin C disease, HbSC; and sickle cell [3-thalassemia). 4. PATHOPHYSIOLOGY OF BONE DISEASE IN HEMOGLOBINOPATHIES

Jaffe noted that the more specific changes in the bone result from (1) hyperplasia of the erythroblastic elements of the bone marrow; (2) plugging or thrombosis of local blood vessels by masses of sickle cells; and (3) a complicating infection, usually salmonella, involving bones at the site of infarcts (97). In a patient with sickle crisis there may be an anemia, but it is the thrombotic episode with fever and leukocytosis along with pain that simulates an acute bone or joint infection. When a bone is the main site of thrombosis acute osteomyelitis is suspected, and if in a joint acute arthritis may be suspected. Two pathological processes are involved: (1) sickling of the red cells results in thrombotic infarcts in bone, which cause pain, crises, and in some cases, osteomyelitis; and (2) increased destruction of inefficient sickle red cells produces hemolysis. Pathologic occurrences in bones in sickle cell disease are due to blockage of the microcirculation, which causes infarction of the bone marrow and adjacent bony trabeculae. The marrow infarcts occur during generalized vaso-occlusive crises and are the main cause of pain. Bone marrow aspirated from areas of infarct during crises actually shows necrotic hemopoietic cells. The earliest skeletal symptoms in crises often occur in the digits and are referred to as dactylitis or the hand and foot syndrome. This occurs in approximately 60% of patients presenting between the ages of 9 months and 4 years with painful tender swelling of the hands and feet along with fever and leukocytosis. The discomfort is caused by expansion of the bone at the metaphyseal regions. Bone and joint pain then occurs between 3 years of age and skeletal maturity.

Diggs has pointed out that the bone marrow in sickle cell anemia is hypercellular and contains little fat (46). Even in infants the red marrow fills all of the marrow spaces, including the small bones of the hands and feet. In older children and adults, the marrow recedes from the tubular bones of the hands and feet but persists in the carpal and tarsal bones and the shafts of the long bones. The cellular marrow extends into the widened Haversian and Volkmann canals within the cortices. There is widening of the medullary cavities and intertrabecular spaces and thinning of both trabecular and cortical bone. Avascular necrosis is most common in the femoral head and to a lesser extent in the humeral head. It is the femoral lesion that is most disabling. Due to necrosis of the bone and the poor ability to repair, an involved femoral head almost invariably proceeds to complete degeneration with the need for early arthroplasty. Necrosis also tends to occur in the vertebral bodies, which not only causes considerable back pain but also flattening of the vertebrae due to osteoporosis. The pathoanatomy is characterized by a hypercellular marrow, superimposed on which is vascular occlusion due to sickled erythrocytes and increased viscosity of the blood, chronic stasis, and hypoxia. Focal areas of ischemia thus occur. Chung and Ralston have reviewed the stages of femoral head necrosis in sickle cell anemia (39). The bone changes, as noted by Diggs, fall into two categories: (1) those due to erythroid hyperplasia, and (2) those due to thrombosis and infarction. They have summarized the changes extremely well. Hyperplastic changes involve the following: the vertebral bodies (osteoporosis with biconcave cupping and bulging of the disks); the skull (loss of trabecular definition, thinning of the outer table, hair-on-end appearance, localized osteoporosis, lamellated new bone); the fiat bones (osteoporosis, widening of the trabecular pattern, cortical thinning); and the long bones (osteoporosis, widening of the medullary canal, thinning of the cortex, and widening of the diaphyseal trabeculae). The changes associated with thrombosis and infarction are seen in the vertebra (massive infarction with collapse); the small tubular bones (periostosis, cortical destruction); the flat bones (patchy sclerosis); the shafts of long bones (cortex thickened and sclerosed, patchy and irregular segmental necrosis); and the epiphyses (retardation of growth, acceleration of growth, and necrosis and collapse).

a. Femoral Head Osteonecrosis. The major problem referable to the epiphyses is avascular necrosis with the disorder becoming apparent clinically between the ages of 10 and 15 years (46, 76, 86, 112). The head of the femur is affected most commonly after puberty. Milner et al. studied a large group of 2590 sickle disease patients who were over 5 years of age at entry (136, 137). Hip radiographs were taken at least twice several years apart in the large group followed for an average of 5.6 years. The overall prevalence of osteonecrosis of the femoral head in sickle cell disease is about 10%. Patients with the hemoglobin SS genotype + oLthalassemia were at the greatest risk for osteonecrosis. The

References patients with hemoglobin SS genotype alone had approximately the same extent of involvement as those with the hemoglobin SC genotype. Intermediate rates of osteonecrosis were seen in those with hemoglobin S 13-thalassemia of either type. Osteonecrosis was occasionally found in patients as young as 5 years old. It was further noted that 3% of the patients under 15 years of age had osteonecrosis. Virtually none of these, however, were of the thalassemia or hemoglobin SC variants. For the entire series the prevalence of osteonecrosis of the hip was 9.7%. This was 10% in SS, 11.9% in S 13 0, 9.1% in SC, and 6.9% in S 13 +. There was progressive involvement the older the patient. In those from 5 to 9 years of age involvement was only 1.3%, and in those from 10 to 14 years of age it was 4.6%. At each decade following 15 years of age the rate of involvement increased. When disease was present it was bilateral in 54.2%. The estimated median age at diagnosis was 28 years for the most severe hemoglobin SS 13-thalassemia group, 36 years for those with hemoglobin SS alone, and 40 years for those with hemoglobin SC. In approximately half of the patients in whom osteonecrosis developed during the time of the study, no pain or limitation of motion was initially noted. A high percentage of patients went on to hip arthroplasty. The lesions appear like those of Legg-Perthes disease, although those affected are somewhat older than those with Perthes. The disorder begins with subchondral sclerosis, progressing to a Pertheslike necrosis and finally to total collapse of the femoral head with degenerative changes of osteoarthritis occurring 5 - 1 0 years later regardless of the age of the patient. Early management appears to have no beneficial effect. Avascular necrosis of the hip generally leads to a disabling disorder necessitating relatively early intervention with hip arthroplasty. When osteonecrosis occurs in the femoral capital epiphysis prior to skeletal maturity, healing can occur but distortion of the head similar to that in Legg-Perthes disorder is seen almost invariably. This is particularly true because the osteonecrosis occurs later than in patients with Perthes disease and generally is not seen until just before skeletal maturity. The patients do not tend to do well. In a study by Hernigou et al. assessing 95 affected hips in 52 patients with sickle cell disease, whose osteonecrosis of the femoral head occurred between the ages of 7 and 15 years, 80% of the hips were painful and showed permanent damage with decreased mobility, limb length discrepancy, and an abnormal gait at a mean follow-up of 19 years (86). When patients were evaluated at an average age of 31 years, 15 hips (16%) had already undergone surgery and 60 (63%) had major problems due to pain. Premature closure of the physis was seen frequently. The premature closure, however, either was across the entire physis leading to a short femoral neck with no major change in the head-neck-shaft angle or involved only the medial portion of the physis leading to broadening of the femoral neck and development of a varus deformity. Lateral physeal arrest leading to valgus angulation of the femoral neck was never noted. The closure also

927

occurred sufficiently late that it did not appear to have any effect on development of the acetabulum. b. H u m e r a l Head Osteonecrosis. The head of the humerus can also be affected by osteonecrosis (136, 138). Osteonecrosis was studied in a large group of 2524 patients in a multicenter study. The patients had HbSS (1737), HbSC (510), sickle cell 13 (+) thalassemia (139), and sickle cell 13 (0) thalassemia (138). The overall prevalence of humeral head necrosis was 5.6%. It was bilateral in two-thirds of those involved. Osteonecrosis was seen in each of the variants in approximately the same amounts. There was a gradual increase in osteonecrosis the older the patient. The prevalence was 1.2% of those between 5 and 9 years of age at initial assessment, 2.6% in those between 10 and 14 years of age, and 3.8% in those between 15 and 24 years of age. The prevalence increased with each advancing age of adult assessment. Those slowest to develop the disorder were in the sickle cell 13 (+) thalassemia group, none of whom had the disorder under the age of 25 years. The disorder, however, was much less debilitating than that that occurred in the femoral head. At the time of diagnosis, 80% of the patients were completely asymptomatic. Pain and limitation of motion developed in most at a later date, but this was mild to moderate and only one patient in the entire study had undergone shoulder arthroplasty for severe symptoms.

References 1. Albright F, Butler AM, Bloomberg E (1937) Rickets resistant to vitamin D therapy. Am J Dis Child 54:529-547. 2. Aledort LM, Haschmeyer RH, Petterson H (1994) A longitudinal study of orthopedic outcomes for severe factor-VIIIdeficient haemophiliacs. J Intern Med 236:391-399. 3. Ambekar AP, Joshipura JCN (1973) Classification and treatment of tuberculosis of the hip joint: Analysis of 104 cases. Ind J Surg 35:309-315. 4. AndrewTA, Porter K (1985) Primary subacute epiphyseal osteomyelitis: A report of three cases. J Pediatr Orthop 5:155-157. 5. Ansell BM, Bywaters EGL (1959) Prognosis in Still's disease. Bull Rheum Dis 9:189-192. 6. Ansell BM, Bywaters EGL (1962) Diagnosis of "probable" Still's disease and its outcome. Ann Rheum Dis 21:253-262. 7. Apel DM, Millar EA, Moel DI (1989) Skeletal disorders in a pediatric renal transplant population. J Pediatr Orthop 9: 505-511. 8. Arnold WB, Hilgartner MW (1977) Hemophilic arthropathy: Current concepts of pathogenesis and management. J Bone Joint Surg 59A:287-305. 9. Aronstam A, Browne RS, Wassef M, Hamad Z (1983) The clinical features of early bleeding into the muscles of the lower limb in severe hemophiliacs. J Bone Joint Surg 65B: 19-23. 10. Badgley CE, Yglesias L, Perham WS, Snyder CH, Arbor A (1936) Study of the end results in 113 cases of septic hips. J Bone Joint Surg 4:1047-1061. l l. Balc'h TL, Ebelin M, Laurian Y, Lambert T, Verroust E Larrieu M (1987) Synovectomy of the elbow in young hemophilic patients. J Bone Joint Surg 69A:264-269.

928

CHAPTER IO ~ Metabolic, Inflammatory, Neoplastic, Infectious, and Hematologic Disorders

12. Balsan S, Tieder M (1990) Linear growth in patients with hypophosphatemic vitamin D-resistant rickets: Influence of treatment regimen and parent height. J Pediatr 116:365-371. 13. Banks SW, Krigsten W, Compere EL (1940) Regeneration of epiphyseal centers of ossification. J Am Med Assoc 114: 23-27. 14. Barrett IA, Papadimitriou DG (1996) Skeletal disorders in children with renal failure. J Pediatr Orthop 16:264-272. 15. Beck L, Soumounou Y, Martel J, Krischnamurthy G, Gauthier C, Goodyer CG, Tenehouse HS (1997) Pex~EX tissue distribution and evidence for a deletion in the 3 ~region of the pex gene in X-linked hypophosphatemic mice. J Clin Invest 99: 1200-1209. 16. Beerman PJ, Crowe JE, Summer TE, Roberts JE (1981) Case report 164. Osteoid osteoma of ossification center of tibia. Skeletal Radiol 7:71-74. 17. Bergdahl S, Ekengren K, Eriksson M (1985) Neonatal hematogenous osteomyelitis: Risk factors for long-term sequelae. J Pediatr Orthop 5:564-568. 18. Betz RR, Cooperman DR, Wopperer JM, Sutherland RD, White JJ, Schaaf HW, Aschliman MR, Choi IH, Bowen JR, Gillespie R (1990) Late sequelae of septic arthritis of the hip in infancy and childhood. J Pediatr Orthop 10:365-372. 19. Blair WF, Kube WJ (1977) Osteoid osteoma in a distal radial epiphysis: A case report. Clin Orthop Rel Res 126:160-161. 20. Blanche DW (1952) Osteomyelitis in infants. J Bone Joint Surg 34A:71-84. 21. Blane CE, Ragsdale CG, Hensinger RN (1987) Late effects of JRA on the hip. J Pediatr Orthop 7:677-680. 22. Bogoch E, Thompson G, Salter RB (1984) Foci of chronic circumscribed osteomyelitis (Brodie's abscess) that traverse the epiphyseal plate. J Pediatr Orthop 4:162-169. 23. Boriani S (1980) Brodie's abscess: A study of 181 cases, with special reference to radiographic diagnostic criteria. Ital J Orthop Traum 6:373-383. 24. Brattstrom M (1963) Asymmetry of ossification and rate of growth of long bones in children with unilateral juvenile gonarthritis. Acta Rheum Scand 9:102-115. 25. Bunn HF (1997) Pathogenesis and treatment of sickle cell disease. New Eng J Med 337:762-769. 26. Burnett CH, Dent CE, Harper C, Warland BJ (1964) Vitamin D-resistant tickets. Analysis of twenty-four pedigrees with hereditary and sporadic cases. Am J Med 36:222-232. 27. Bywaters EGL (1977) Pathologic aspects of juvenile chronic polyarthritis. Arth Rheum 20(Supp):271-276. 28. Caffey J, Schlesinger ER (1940) Certain effects of hemophilia on the growing skeleton. Some roentgenographic observations of overgrowth and dysgenesis of the epiphyses associated with chronic hemarthrosis. J Pediatr 16:549-565. 29. Carlson UH, Wilkinson RH, Bhakkaviziam A (1972) Aneurysmal bone cysts in children. Am J Roentgenol Radium Ther Nucl Med 116:644-650. 30. Carpenter TO (1997) New perspectives on the biology and treatment of X-linked hypophosphatemic tickets. Pediatr Clin North Am 44:443-465. 31. Cass JN (1940) Staphylococcus aureus infection of the long bones in the newly born. Arch Dis Child 15:55-60. 32. Cassidy JT, Brody GL, Martel W (1967) Monarticular juvenile rheumatoid arthritis. J Pediatr 70:867-875.

33. Cassidy JT, Petty RE (1995) Juvenile rheumatoid arthritis. In: Textbook of Pediatric Rheumatology, 3rd ed. pp. 133-223, Philadelphia: WB Saunders Company. 34. Cheng JCY, Aguilar J, Leung PC (1995) Hip reconstruction for femoral head loss from septic arthritis in children: A preliminary report. Clin Orthop Rel Res 314:214-224. 35. Chesney RW, Rosen JF, DeLuca HF (1983) Disorders of calcium metabolism in children. In: Recent Progress in Pediatric Endocrinology, eds G Chiumello, M Sperling. pp. 5-24, New York: Raven Press. 36. Choi IH, Pizzutillo PD, Bowen JR, Dragann R, Malhis T (1990) Sequelae and reconstruction after septic arthritis of the hip in infants. J Bone Joint Surg 72A:1150-1165. 37. Chow SP, Yau A (1980) Tuberculosis of the knee--A longterm follow-up. Internat Orthop 4:87-92. 38. Christensen JF (1940) Three familial cases of atypical late tickets. Acta Paed 28:247-270. 39. Chung SMK, Ralston EL (1969) Necrosis of the femoral head associated with sickle-cell anemia and its genetic variants. J Bone Joint Surg 51:33-58. 40. Cohen J (1977) Unicameral bone cysts: A current synthesis of reported cases. Orthop Clin North Am 8:715-736. 41. Cole DEC, Carpenter TO, Goltzman D (1989) Calcium homeostasis and disorders of bone and mineral metabolism. In: Pediatric Endocrinology, 2nd ed, eds R Collu, JR Ducharme, JH Cuyda. New York: Raven Press. 42. David L (1991) Common vitamin D-deficiency rickets. Nestle Nut Workshop Ser 21:107-119. 43. Davids JR, Fisher R, Lum G, VonGlinski S (1992) Angular deformity of the lower extremity in children with renal osteodystrophy. J Pediatr Orthop 12:291-299. 44. DeGnore LT, Wilson FC (1989) Surgical management of hemophilic arthropathy. In: Instructional Course Lectures, volume 38. pp. 383-388, Park Ridge, IL: The American Academy of Orthopedic Surgeons. 45. DePalma AF, Cotler J (1956) Hemophilic arthropathy. Clin Orthop Rel Res 8:163-190. 46. Diggs LW (1967) Bone and joint lesions in sickle-cell disease. Clin Orthop Rel Res 52:119-143. 47. DiMichele D (1996) Hemophilia 1996: A new approach to an old disease. Pediatr Clin N Am 43:709-736. 48. Dodds GS, Cameron HC (1934) Studies on experimental tickets in rats I. Structural modifications of the epiphyseal cartilages in the tibia and other bones. Am J Anat 55:135-165. 49. Dyer R, Stelling CB, Fechner RE (1981) Epiphyseal extension of an aneurysmal bone cyst. Am J Roentgenol 137: 172-173. 50. Ecarot B, Glorieux FH, Desbarats M, Travers R, Labelle L (1992) Defective bone formation by hyp mouse bone cells transplanted into normal mice: Evidence in favor of an intrinsic osteoblast defect. J Bone Min Res 7:215-220. 51. Ecarot-Charrier B, Glorieux FH, Travers R, Desbarats M, Bouchard F, Hinek A (1988) Defective bone formation by transplanted hyp mouse bone cells into normal mice. Endocrin 123:768-773. 52. Edstrom G (1958) Rheumatoid arthritis and Still's disease in children. Arthrit Rheum 1:497-504. 53. Ehrenforth S, Kreuz W, Scharrer I, Linde R, Funk M, Gungor T, Krackhardt B, Kornhuber B (1992) Incidence of develop-

References ment of factor VIII and factor IX inhibitors in haemophiliacs. Lancet 339:594-598. 54. Eicher EM, Southard JL, Scriver CR, Glorieux GH (1976) Hypophosphatemia: Mouse model for human familial hypophosphatemic (vitamin D-resistant) tickets. Proc Natl Acad Sci 73:4667-4671. 55. Engfeldt B, Zetterstrom R, Winberg J (1956) Primary vitamin D-resistant tickets. J Bone Joint Surg 38A:1323-1334. 56. Enneking WF, Kagan A, II (1978) Transphyseal extension of osteosarcoma: Incidence, mechanism, and implications. Cancer 41:1526-1537. 57. Epstein FH (1997) Pathogenesis and treatment of sickle cell disease. New Eng J Med 337:762-769. 58. Evans GA, Arulanantham K, Gage JR (1980) Primary hypophosphatemic tickets: Effect of oral phosphate and vitamin D on growth and surgical treatment. J Bone Joint Surg 62A: 1130-1138. 59. Eyre-Brook AL (1960) Septic arthritis of the hip and osteomyelitis of the upper end of the femur in infants. J Bone Joint Surg 42B:11-20. 60. Eyres KS, Brown J, Douglas DL (1993) Osteotomy and intramedullary nailing for the correction of progressive deformity in vitamin D-resistant hypophosphataemic tickets. J R Coil Surg Edinb 38:50-54. 61. Eyring EJ, Bjornson DRB, Close JR (1965) Management of hemophilia in children. Clin Orthop Rel Res 40:95-112. 62. Fabry G, Meire E (1983) Septic arthritis of the hip in children: Poor results after late and inadequate treatment. J Pediatr Orthop 3:461-466. 63. Fernandez-Palazzi F, Rivas S, Cibeira JL, Dib O, Viso R (1996) Radioactive synoviorthesis in hemophilic hemarthrosis. Clin Orthop Rel Res 328:14-18. 64. Ferris B, Walker C, Jackson A, Kirwan E (1991) The orthopedic management of hypophosphataemic tickets. J Pediatr Orthop 11:367-373. 65. Flandry F, Hughston JC (1987) Pigmented villonodular synovitis. J Bone Joint Surg 69A:942-949. 66. Furie B (1995) Vitamin K-dependent blood coagulation proteins: Normal function and clinical disorders. In: Blood: Principles and Practice of Hematology, eds RJ Handin, TP Stossel, SE Lux. pp. 1181-1203, Philadelphia: LB Lippincott. 67. Gamble JG, Bellah J, Rinsky LA, Glader B (1991) Arthropathy of the ankle in hemophilia. J Bone Joint Surg 73A: 1008-1015. 68. Ghandur-Mnaymneh L, Mnaymneh WA, Puls S (1983) The incidence and mechanism of transphyseal spread of osteosarcoma of long bones. Clin Orthop Rel Res 177:210-215. 69. Gilbert MS (1993) Prophylaxis: Musculoskeletal evaluation. Semin Hematol 30:3-6. 70. Gillespie R (1973) Septic arthritis of childhood. Clin Orthop Rel Res 96:152-159. 71. Gillespie WJ, Moore TE, Mayo KM (1986) Subacute pyogenic osteomyelitis. Orthopaedics 9:1565-1570. 72. Gledhill RB (1973) Subacute osteomyelitis in children. Clin Orthop Rel Res 96:57-69. 73. Gledhill RB (1973) Various phases of pediatric osteomyelitis. In: Instructional Course Lectures, volume 22. pp. 245-269, Chicago: American Academy of Orthopaedic Surgeons.

929

74. Glorieux FH (1991) Rickets, the continuing challenge. New Eng J Med 325:1875-1977. 75. Glorieux FH, Chabot G, Tau C (1991) Familial hypophosphatemic tickets: Pathophysiology and medical management. Nestle Nut Workshop Ser 21:185-199. 76. Golding JSR (1973) Conditions of the hip associated with hemoglobinopathies. Clin Orthop Rel Res 90:22-28. 77. Goodman WG, Salusky IB (1995) Growth hormone and calcitriol as modifiers of bone formation in renal osteodystrophy. Kidney Internat 48:657-665. 78. Green NE, Beauchamp RD, Griffin PP (1981) Primary subacute osteomyelitis. J Bone Joint Surg 63A:107-114. 79. Green WT, Shannon JG (1936) Osteomyelitis of infants. Arch Surg 32:462-493. 80. Greene WB, DeGnore LT, White GC (1990) Orthopedic procedures and prognosis in hemophilic patients who are seropositive for human immunodeficiency virus. J Bone Joint Surg 72A:2-11. 81. Grokoest AW, Snyder AI, Schlaeger R (1962) Juvenile Rheumatoid Arthritis. Boston: Little Brown & Company. 82. Habermann ET, Stern RE (1974) Osteoid-osteoma of the tibia in an eight-month-old boy. J Bone Joint Surg 56A: 633-636. 83. Halbstein BM (1967) Bone regeneration in infantile osteomyelitis. J Bone Joint Surg 49A:149-152. 84. Hallel T, Salvati EA (1978) Septic arthritis of the hip in infancy: End result study. Clin Orthop Rel Res 132:115-128. 85. Harris HN, Kircaldy-Willis WH (1965) Primary subacute pyogenic osteomyelitis. J Bone Joint Surg 47B:526-532. 86. Hernigou P, Galacteros F, Bachir D, Goutallier D (1991) Deformities of the hip in adults who have sickle-cell disease and had avascular necrosis in childhood: A natural history of fifty-two patients. J Bone Joint Surg 73A:81-92. 87. Holm IA, Huang X, Kunkel LM (1997) Mutational analysis of the PEX gene in patients with X-linked hypophosphatemic tickets. Am J Hum Genet 60:790-797. 88. Holmes GW, Ruggles HE (1936) In: Roentgen Interpretation, 5th ed. Philadelphia: Lea and Febiger. 89. Houghton GR, Duthie RB (1979) Orthopedic problems in hemophilia. Clin Orthop Rel Res 138:197-216. 90. Hoyer LW (1994) Hemophilia A. New Eng J Med 330: 38-48. 91. Hruska KA, Teitelbaum SL (1995) Renal osteodystrophy. New Eng J Med 333:166-174. 92. Hudson TM (1987) Chondroblastoma. In: RadiologicPathologic Correlation of Musculoskeletal Lesions, Chapter 12. pp. 127-131, Baltimore: Williams & Wilkins. 93. Hunka L, Said SE, MacKenzie DA, Rogala EJ, Cruess RL (1982) Classification and surgical management of the severe sequelae of septic hips in children. Clin Orthop Rel Res 171: 30-36. 94. HYP Consortium (1995) A gene (PEX) with homologies to endopeptidases is mutated in patients with X-linked hypophosphatemic tickets. Nature Genet 11:130-136. 95. Iceton J, Rang M (1986) An osteoid osteoma in an open distal femoral epiphysis: A case report. Clin Orthop Rel Res 206: 162-165. 96. Jacobsen ST, Hull CK, Crawford AH (1986) Nutritional rickets. J Pediatr Orthop 6:713-716.

930

CHAPTER IO 9

Metabolic, Inflammatory, Neoplastic, Infectious, and Hematoloyic Disorders

97. Jaffe HL (1972) Metabolic, Degenerative, and Inflammatory Diseases of Bones and Joints. Philadelphia: Lea and Febiger. 98. Jaffe HL, Lichtenstein L (1942) Benign chondroblastoma of bone. A reinterpretation of the so-called calcifying or chondromatous giant cell tumor. Am J Pathol 18:969-991. 99. Jani L, Remagen W (1983) Primary chronic osteomyelitis. Internat Orthop 7:79-83. 100. Kandel SN, Mankin HJ (1973) Pyogenic abscess of the long bones in children. Clin Orthop Rel Res 96:108-117. 101. Kanel JS, Price CT (1995) Unilateral external fixation for corrective osteotomies in patients with hypophosphatemic tickets. J Pediatr Orthop 15:232-235. 102. Katayama R, Itami Y, Marumo E (1962) Treatment ofhip and knee joint tuberculosis. J Bone Joint Surg 44A:897-916. 103. Katz SG, Nelson IW, Atkins RM, Duthie RB (1991) Peripheral nerve lesions in hemophilia. J Bone Joint Surg 73A: 1016-1019. 104. Kay RM, Eckardt JJ, Mirra JM (1996) Multifocal pigmented villonodular synovitis in a child: A case report. Clin Orthop Rel Res 322:194-197. 105. Kazazian HH (1993) The molecular basis of hemophilia A and the present status of carrier and antenatal diagnosis of the disease. Thromb Haemostas 70:60-62. 106. Kendrick JI, Evarts CM (1967) Osteoid osteoma. A critical analysis of 40 tumours. Clin Orthop 54:51-59. 107. Key JA (1932) Hemophilic arthritis. Ann Surg 95:198-225. 108. King DM, Mayo KM (1969) Subacute haematogenous osteomyelitis. J Bone Joint Surg 51B:458-463. 109. Kobayakawa M, Rydholm U, Wingstrand H, Pettersson H, Lidgren L (1989) Femoral head necrosis in juvenile chronic arthritis. Acta Orthop Scand 60:164-169. 110. Kondo E, Yamada K (1957) End results of focal debridement in bone and joint tuberculosis and its indications. J Bone Joint Surg 39A:27-31. 111. Konig F (1892) Die Gelenkerkrankungen bei Blutern mit besonderer Berucksichtigung der diagnose. Klin Vortrage 36:233-243 [Diseases of the joints in bleeders, especially with regard to the diagnosis. Clin Orthop Rel Res 52:5-11 (translation, 1967)]. 112. Konotey-Ahulu FID (1974) The sickle cell diseases. Arch Intern Med 133:611-619. 113. Kramer SJ, Post J, Sussman M (1986) Acute hematogenous osteomyelitis of the epiphysis. J Pediatr Orthop 6:493-495. 114. Kruger GD, Rock MG (1986) Osteoid osteoma of the distal femoral epiphysis: A case report. Clin Orthop Rel Res 222: 203-209. 115. Laine H, Mikkelsen OA (1968) Epiphyseal stapling in juvenile rheumatoid gonarthritis. Acta Rheum Scand: 14:317-322. 116. Lane PA (1996) Sickle cell disease. Ped Clin North Am 43: 639-665. 117. Langenskiold A (1984) Growth disturbance after osteomyelitis of femoral condyles in infants. Acta Orthop Scand 55: 1-13. 118. Larsson SA (1985) Life expectancy of Swedish hemophiliacs, 1831-1980. Br J Haemato159:593-602. 119. Leeson MC, Smith A, Carter JR, Makley JT (1985) Eosinophilic granuloma of bone in the growing epiphysis. J Pediatr Orthop 5:147-150.

120. Lejman T, Strong M, Michno P, Hayman M (1995) Septic arthritis of the shoulder during the first 18 months of life. J Pediatr Orthop 15:172-175. 121. Lew DE Waldvogel FA (1997) Osteomyelitis. New Eng J Med 336:999-1007. 122. Lindblad B, Ekengren K, Aurelius G (1965) The prognosis of acute haematogenous osteomyelitis and its complications during early infancy after the advent of antibiotics. Acta Paediat Scand 54:24-32. 123. Lloyd-Roberts GC (1960) Suppurative arthritis of infancy: Some observations upon prognosis and management. J Bone Joint Surg 42B:706-720. 124. Lucas RC (1883) Form of late tickets associated with albuminuria, tickets of adolescents. Lancet i:993-994. 125. Luck JV, Jr, Kasper CK (1989) Surgical management of advanced hemophilic arthropathy: An overview of 20 years experience. Clin Orthop Rel Res 242:60-82. 126. Lyon MF, Scriver CR, Baker LRI, Tenenhouse HS, Kronick J, Mandla S (1986) The Gy mutation: Another cause of X-linked hypophosphatemia in mouse. Proc Natl Acad Sci 83:4899-4903. 127. Malawer MM, Markle BR (1982) Unicameral bone cyst with epiphyseal involvement: Clinicoanatomic analysis. J Pediatr Orthop 2:71-79. 128. Mannucci PM (1990) Desmopressin: A nontransfusional hemostatic agent. Ann Rev Med 41:55-64. 129. Mannucci PM (1993) Modem treatment of hemophilia: From the shadows toward the light. Thromb Haemostas 70:17-23. 130. McCarthy SM, Ogden JA (1982) Epiphyseal extension of an aneurysmal bone cyst. J Pediatr Orthop 2:171-175. 131. Mehls O (1984) Renal osteodystrophy in children: Etiology and clinical aspects. In: End Stage Renal Disease in Children, eds RN Fine, AB Gruskin. pp. 227-250, Philadelphia: WB Saunders. 132. Mehls O, Ritz E, Kreusser W, Krempien B (1980) Renal osteodystrophy in uremic children. Clin Endocrinol Metab 9:151-176. 133. Mehls O, Ritz E, Oppermann HC (1981) Femoral head necrosis in uraemic children without steroid treatment or transplantation. J Pediatr 99:926-929. 134. Merchan ECR, Magallon M, Martin-Villar J, Galinod E, Ortega F, Pardo JA (1993) Long-term follow-up of haemophilic arthropathy treated by Au-198 radiation synovectomy. Internat Orthop 17:120-124. 135. Micheli LJ, Jupiter J (1978) Osteoid osteoma as a cause of knee pain in the young athlete: A case study. Am J Sport Med 6:199-209. 136. Milner PF, Joe C, Burke GJ (1994) Bone and joint disease. In: Sickle Cell Disease: Basic Principles and Clinical Practice, eds SH Embury, RP Hebbel, N Mohandas, MH Steinberg. pp. 645-661, New York: Raven Press. 137. Milner PF, Kraus AP, Sebes JI, Sleeper LA, Dukes KA, Embury SH, Bellevue R, Koshy M, Moohr JW, Smith J (1991) Sickle cell disease as a cause of osteonecrosis of the femoral head. New Eng J Med 325:1476-1481. 138. Milner PF, Kraus AP, Sebes JI, Sleeper LA, Dukes KA, Embury SH, Bellevue R, Koshy M, Moohr JW, Smith J (1993) Osteonecrosis of the humeral head in sickle cell disease. Clin Orthop Rel Res 289:136-143.

References 139. Mirra JM (1980) Chondroblastoma. In: Bone Tumors. Diagnosis and Treatment. pp. 219-226, Philadelphia: JB Lippincott. 140. Moed BR, LaMont RL (1982) Unicameral bone cyst complicated by growth retardation. J Bone Joint Surg 64A: 1379-1381. 141. Montane I, McCollough NC, Lian ECY (1986) Synovectomy of the knee for hemophilic arthropathy. J Bone Joint Surg 68A:210-216. 142. Nelson JP, Foster RJ (1976) Solitary bone cyst with epiphyseal involvement: A case report. Clin Orthop Rel Res 118: 147-150. 143. Nilsson IM, Berntorp E, Lofqvist T, Pettersson H (1992) Twenty-five years' experience of prophylactic treatment in severe hemophilia A and B. J Intern Med 232:25-32. 144. Odaka T, Koshino T, Saito T (1987) Intra-articular epiphyseal osteoid osteoma of the distal femur. J Pediatr Orthop 7: 331-333. 145. Ogden JA, Lister G (1975) The pathology of neonatal osteomyelitis. Pediatrics 55:474-478. 146. Onuba O (1993) Bone disorders in sickle-cell disease. Internat Orthop 17:397-399. 147. Oppenheim WL, Salusky IB, Kaplan D, Fine RN (1990) Renal osteodystrophy in children. In: Metabolic Bone Disease in Children, eds S Costello, L Finberg. pp. 197-229, New York: Marcel Dekker. 148. Oppenheim WL, Shayestehfar S, Salusky IB (1992) Tibial physeal changes in renal osteodystrophy: Lateral Blount's disease. J Pediatr Orthop 12:774-779. 149. Otto JC (1803) An account of a hemorrhagic disposition existing in certain families. Med Repos 6:1-4. 150. Park EA (1939) Observations of the pathology of tickets with particular reference to the changes at the cartilage-shaft junctions of the growing bones. Bull New York Acad Med 15: 495-543. 151. Patriquin HB, Camerlain M, Trias A (1984) Late sequelae of juvenile rheumatoid arthritis of the hip: A follow-up study into adulthood. Pediatr Radiol 14:151-157. 152. Pearce MS, Smith MA, Savidge GF (1994) Supramalleolar tibial osteotomy for hemophilic arthropathy of the ankle. J Bone Joint Surg 76B:947-950. 153. Pedersen HE, McCarroll HR (1951) Vitamin D-resistant tickets. J Bone Joint Surg 33A:203-218. 154. Peters W, Irving J, Letts M (1992) Long-term effects of neonatal bone and joint infection on adjacent growth plates. J Pediatr Orthop 12:806-810. 155. Pettersson H (1994) Can joint damage be quantified? Sem Hematol, Suppl 2, 31:1-4. 156. Pettersson H, Ahlberg A, Nilsson IM (1980) A radiologic classification of hemophilic arthropathy. Clin Orthop Rel Res 149:153-159. 157. Phemister DB (1925) Changes in the articular surfaces in tuberculous arthritis. J Bone Joint Surg 7:835-846. 158. Pierce DS, Wallace WM, Herndon CH (1964) Long-term treatment of vitamin D-resistant tickets. J Bone Joint Surg 46A:978-997. 159. Platt OS, Guinan EC (1996) Bone marrow transplantation in sickle cell anemia--The dilemma of choice. New Eng J Med 335:426-428.

931

160. Platt OS, Thorington BD, Brambilla DJ, Milner PF, Rosse WF, Vichinsky E, Kinney TR (1991) Pain in sickle cell disease: Rates and risk factors. New Eng J Med 325:11-16. 161. Post M, Watts G, Telfer M (1986) Synovectomy in hemophilic arthropathy: A retrospective review of 17 cases. Clin Orthop Rel Res 202:139-146. 162. Reichel H, Koeffler HP, Norman AW (1989) The role of the vitamin D endocrine system in health and disease. New Eng J Med 320:980-922. 163. Reinecke, Wohlwill (1929) Uber hamophile Gelenkerkrankung. Arch Klin Chir 154:425-479. 164. Ribbans WJ, Phillips AM (1996) Hemophilic ankle arthropathy. Clin Orthop Rel Res 328:39-45. 165. Rick ME (1994) Diagnosis and management of von Willebrand's syndrome. Med Clin North Am 78:609-623. 166. Roberts JM, Drummond DS, Breed AL, Chesney J (1982) Subacute hematogenous osteomyelitis in children: A retrospective study. J Pediatr Orthop 2:249-254. 167. Roberts PH (1970) Disturbed epiphyseal growth at the knee after osteomyelitis in infancy. J Bone Joint Surg 52B: 692-703. 168. Robertson DE (1967) Primary acute and subacute localized osteomyelitis and osteochondritis in children. Can J Surg 10: 408-413. 169. Rodriguez-Merchan EC (1995) The haemophilic pseudotumour. Internat Orthop 19:255-260. 170. Rodriguez-Merchan EC (1996) Effects of hemophilia on articulations of children and adults. Clin Orthop Rel Res 328:7-13. 171. Rodriguez-Merchan EC (1998) Management of the orthopedic complications of haemophilia. J Bone Joint Surg 80B: 191-196. 172. Rodriguez-Merchan EC, Galindo E, Ladreda JMM, Pardo JA (1994) Surgical synovectomy in haemophilic arthropathy of the knee. Internat Orthop 18:38-41. 173. Rowe PSN, Goulding JN, Francis F, Oudet C, Econs MJ, Hanauer A, et al. (1996) The gene for X-linked hypophosphataemic tickets maps to a 200-300 kb region in Xp22. 1, and is located on a single YAC containing a putative vitamin D response element (VDRE). Hum Genet 97:345-352. 174. Rowe PSN, Oudet CL, Francis F, Sinding C, Pannetier S, Econs MJ, et al. (1997) Distribution of mutations in the P E X gene in families with X-linked hypophosphataemic tickets (HYP). Hum Mol Genet 6:539-549. 175. Rubinovitch M, Said SE, Glorieux FH, Cruess RL, Rogala E (1988) Principles and results of corrective lower limb osteotomies for patients with vitamin D-resistant hypophosphatemic tickets. Clin Orthop Rel Res 237:264-270. 176. Rydholm U, Brattstrom H, Bylander B, Lidgren L (1987) Stapling of the knee in juvenile chronic arthritis. J Pediatr Orthop 7:63-68. 177. Salusky IB (1995) Bone and mineral metabolism in childhood end-stage renal disease. Pediatr Clin North Am 42: 1531-1550. 178. Salusky IB, Goodman WG (1995) Growth hormone and calcitriol as modifiers of bone formation in renal osteodystrophy. Kidney Internat 48:657-665. 179. Schaller JG (1984) Chronic arthritis in children: Juvenile rheumatoid arthritis. Clin Orthop Rel Res 182:79-89.

932

CHAPTER 10 ~ Metabolic, ln[lammatory, Neoplastic, Infectious, and Hematoloyic Disorders

180. Schmidt D, Mubarak S, Gelberman R (1981) Septic shoulders in children. J Pediatr Orthop 1:67-72. 181. Schmorl G (1909) Die pathologische anatomie der rachitischen Knochenerkrankung mit besonderer Berucksichtigung ihrer histolgie und pathogenese. Engeb Inn Med Kinderheil 4:403-454. 182. Schramm W (1993) Experience with prophylaxis in Germany. Sem Hemato130:12-15. 183. Scriver CR, Tenenhouse HS, Glorieux FH (1991) X-linked hypophosphatemia: An appreciation of a classic paper and a survey of progress since 1958. Medicine (Baltimore) 70: 218-228. 184. Seitz WH, Dick HM (1978) Case report: Intraepiphyseal osteoid osteoma of the distal femur in an 8-year-old girl. J Bone Joint Surg 3:505-507. 185. Shapiro F (1982) Developmental patterns in lower-extremity length discrepancies. J Bone Joint Surg 64A:639-651. 186. Shapiro F (1998) Epiphyseal and physeal cartilage vascularization: A light microscopic and tritiated-thymidine autoradiographic study of cartilage canals in newborn and young postnatal rabbit bone. Anat Rec 252:140-148. 187. Sherman MS (1947) Osteoid osteoma. Review of the literature and report of 30 cases. J Bone Joint Surg 29:918-930. 188. Shohl AT, Wolbach SB (1936) Rickets in rats. The effect of low calcium-high phosphorus diets at various levels and ratios uponthe production of tickets and tetany. J Nutrit 11:275-291. 189. Siffert RS (1957) The effect of juxta-epiphyseal pyogenic infection on epiphyseal growth. Clin Orthop Rel Res 10: 131-139. 190. Silva JF (1980) A review of patients with skeletal tuberculosis treated at the University Hospital, Kuala Lumpur. Internat Orthop 4:79-81. 191. Simon MA, Bos GD (1980) Epiphyseal extension of metaphyseal osteosarcoma in skeletally immature individuals. J Bone Joint Surg 62A: 195-204. 192. Simon S, Whiffen J, Shapiro F (1981) Leg-length discrepancies in monoarticular and pauciarticular juvenile rheumatoid arthritis. J Bone Joint Surg 63A:209-215. 193. Simon WH, Belier ML (1975) Intracapsular epiphyseal osteoid osteoma of the ankle joint. A case report. Clin Orthop Rel Res 108:200-203. 194. Sledge CB, Atcher RW, Shortkroff S, Anderson RJ, Bloomer WD, Hurson BJ (1984) Intra-articular radiation synovectomy. Clin Orthop Rel Res 182:37-40. 195. Slowick FA (1935) Purulent infections of the hip joint. An analysis of sixty cases. New Eng J Med 212:672-676. 196. Smith MA, Urquhart DR, Savidge GF (1981) The surgical management of varus deformity in hemophilic arthropathy of the knee. J Bone Joint Surg 63B:261-265. 197. Smith T (1874) The acute arthritis of infants. Saint Barth Hosp Rep 10:189-204. 198. Speer DP (1984) Early pathogenesis of hemophilic arthropathy: Evolution of the subchondral cyst. Clin Orthop Rel Res 185:250-265. 199. Springfield DS, Capanna R, Gherlinzoni F, Picci P, Campanacci M (1985) Chondroblastoma: A review of seventy cases. J Bone Joint Surg 67A:748-755. 200. Stamp WG, Whitesides TE, Field MH, Scheer GE (1964) Treatment of vitamin D-resistant tickets. J Bone Joint Surg 46A:965-977.

201. Stein H, Duthie RB (1981) The pathogenesis of chronic haemophilic arthropathy. J Bone Joint Surg 63B:601-609. 202. Stern MB, Cassidy R, Mirra J (1976) Eosinophilic granuloma of the proximal tibial epiphysis. Clin Orthop Rel Res 118: 153-156. 203. Stern PJ, Watts HG (1979) Osteonecrosis after renal transplantation in children. J Bone Joint Surg 61A:851-856. 204. Stickler GB, Morgenstern BZ (1989) Hypophosphataemic tickets: Final height and clinical symptoms in adults. Lancet ii:902-905. 205. Storti E, Ascari E (1975) Surgical and chemical synovectomy. Ann NY Acad Sci 240:316-327. 206. Storti E, Traldi A, Tosatti E, Davoli PG (1969) Synovectomy, a new approach to haemophilic arthropathy. Acta Haemat 41: 193-205. 207. Strom TM, Francis F, Lorenz B, Boddrich A, Econs MJ, Lehrach H, Meitinger T (1997) Pex gene deletions in Gy and hyp mice provide mouse models for X-linked hypophosphatemia. Hum Mol Genet 6:165-171. 208. Strong M, Lejman T, Michno P, Hayman M (1994) Sequelae from septic arthritis of the knee during the first two years of life. J Pediatr Orthop 14:745-751. 209. Strong M, Lejman T, Michno P (1995) Septic arthritis of the wrist in infancy. J Pediatr Orthop 15:152-156. 210. Sundberg J, Brattstrom M (1965) Juvenile rheumatoid gona~hritis II. Disturbance of ossification and growth. Acta Rheum Scand 11:279-290. 211. Swann M (1990) The surgery of juvenile chronic arthritis: An overview. Clin Orthop Rel Res 259:70-75. 212. Swierstra BA, Diepstraten FM, Heyden BJ (1993) Distal femoral physiolysis in renal osteodystrophy: Successful nonoperative treatment of 3 cases followed for 5 years. Acta Orthop Scand 3:382-384. 213. Tapia J, Stearns G, Ponseti IV (1964) Vitamin D-resistant tickets. J Bone Joint Surg 46A:935-958. 214. Tebor GB, Ehrlich MG, Herrin J (1983) Slippage of the distal tibial epiphysis. J Pediatr Orthop 3:211-215. 215. Thomas HB (1936) Some orthopedic findings in ninety-eight cases of hemophilia. J Bone Joint Surg 18:140-147. 216. Thomson J, Lewis IC (1950) Osteomyelitis in the newborn. Arch Dis Child 25:273-279. 217. Trueta J, Morgan JD (1954) Late results in the treatment of one hundred cases of acute haematogenous osteomyelitis. Brit J Surg 41:449-457. 218. Tuli SM (1987) Judicious management of tuberculosis of bones, joints, and spine. U Penn Orthop J 3:31-40. 219. Uittenbogaart CH, Isaacson AS, Stanley P, Pennisi AJ, Malekzadeh MH, Ettenger RB, Fine RN (1978) Aseptic necrosis after renal transplantation in children. Am J Dis Child 132:765-767. 220. Usui M, Matsuno T, Kobayashi M, Yagi T, Sasaki T, Ishii S (1983) Eosinophilic granuloma of the growing epiphysis. Clin Orthop Rel Res 176:201-205. 221. VanHorn JR, Karthaus RP (1989) Epiphyseal osteoid osteoma: Two case reports. Acta Orthop Scand 5:625-627. 222. Verge CF, Lam A, Simpson JM, Cowell CT, Howard NV, Silink M (1991) Effects of therapy in X-linked hypophosphatemic tickets. New Eng J Med 325:1843-1878. 223. Vidigal EC, Jacomo AD (1994) Early diagnosis of septic arthritis of the hip in neonates. Internat Orthop 18:189-192.

References 224. Walters MC, Patience M, Leisenring W, Eckman JR, Scott JP, Mentzer WC, Davies SC, Ohene-Frempong K, Bernaudin F, Matthews DC, Storb R, Sullivan KM (1996) Bone marrow transplantation for sickle cell disease. New Eng J Med 335: 369-376. 225. Weissberg ED, Smith AL, Smith DH (1974) Clinical features of neonatal osteomyelitis. Pediatr 53:505-510. 226. Werner EJ (1996) Von Willebrand disease in children and adolescents. Ped Clin North Am 43:683-707. 227. White GC (1995) Coagulation factors V and VIII: Normal function and clinical disorders. In: Blood: Principles and Practice of Hematology, eds RJ Handin, TP Stossel, SE Lux. pp. 1151-1174, Philadelphia: JB Lippincott. 228. White PH (1990) Growth abnormalities in children with juvenile rheumatoid arthritis. Clin Orthop Rel Res 259:46-50.

933

229. Wiedel JD (1996) Arthroscopic synovectomy of the knee in hemophilia: 10-15 year follow-up. Clin Orthop Rel Res 328:46-53. 230. Wilkinson MC (1962) Partial synovectomy in the treatment of tuberculosis of the knee. J Bone Joint Surg 44B :34-41. 231. Wilkinson MC (1969) Tuberculosis of the hip and knee treated by chemotherapy, synovectomy, and debridement: A follow-up study. J Bone Joint Surg 51A:1343-1359. 232. Wilson JC, McKeever FM (1936) Bone growth disturbance following hematogenous acute osteomyelitis. J Am Med Assoc 107:1188-1193. 233. Winters RW, Graham JB, Williams TF, McFalls VW, Bumett CH (1958) A genetic study of familial hypophosphatemia and vitamin D-resistant tickets with a review of the literature. Medicine 37:97-142.

Index

A Abduction devices, LCP treatment, 344-346 Acetabula acute epiphyseal growth plate fractureseparation, 574-575 childhood hip, 167 corrective procedures, CDH, 208-212 development hip reduction treatment, 232-233 limbus removal in infancy, 233-234 stages, 159-160 dislocated hip, 180 experimental femoral head displacement, 190-192 Fairbank's studies, 176 femoral head-acetabular repair index, 330-333 femoral osteotomies, DDH, 213 growth after surgery, 234 hip development, 155 LCP disease deformation, 296 head index, 330-331 reconstruction, 362-363 responses, 316-318 shelf arthroplasty, 363 Acetabular dysplasia, 196, 789 Acetabular index, hip position radiography, 214,217 Achondrogenesis, 759, 769, 771 Achondroplasia clinical and radiographic characteristics, 795-796 craniocervical junction abnormalities, 798-799 developmental milestones, 799 histopathologic changes, 773 macrocephaly-hydrocephalus, 798 morbidity and mortality, 796-798 orthopedic aspects, 800-801 overview, 794-795 spinal abnormalities, 799-800 Acquired coxa vara, 377 Acrodysplasias, 815-816 Acromelic dysplasia, 785, 815-816

Acromesomelic dysplasia, 813-815 Acute on chronic slip, SCFE, 396 Acute epiphyseal growth plate fractureseparation distal femur, 576-582 distal fibula, 591 distal humerus, 562, 566-570 distal radius, 570-574 distal tibia, 585-591 distal ulna, 574 ligament damage, 592 management principles, 560-561 metacarpals, 574 phalanges, 574 proximal femur, 575-576 proximal fibula, 591 proximal humerus, 561-562 proximal radius, 570 proximal tibia, 582-585 Roentgen stereophotogrammetry, 591-592 triradiate acetabular cartilage, 574-575 Acute slip, SCFE, 396 Adductor muscle, congenitaldevelopmental hip abnormalities, 196 Adolescent hip dysplasia, 239 Adolescent tibia vara clinical profile, 493-495 deformity, 495 femoral varus association, 496 pathoanatomy, 496-497 physeal height details, 495-496 radiographic assessments, 496 terminology, 493 treatment, 497 AER, s e e Apical ectodermal ridge Age distal femoral fracture-separations, 580 epiphyseal growth plate fractureseparation occurrence, 558-559 infantile tibia vara deformity after osteotomy, 488-490 osteochondritis dissecans, 468, 477-478 role in LCP disease, 362 SCFE, 398, 400-402 tuberculosis incidence, 906

935

AI, s e e Acetabular index Airways, skeletal dysplasias, 861 Ambulation hip MR imaging, 259 LCP treatment, 345-346 Anesthesia, skeletal dysplasias, 860-861 Aneurysmal bone cyst, 896 Angiomatous lesions, 646-647 Angioplastic disorders, lower extremity length discrepancy, 641-642 Anisomelia, lower extremity length discrepancies, 639-640 Ankle, skeletal dysplasias, 793-794 Anterior tibial spine, growth plate fractureseparation, 582-584 Apert syndrome, 816 Apical ectodermal ridge bone patterning, 53 definition, 58 limb development control, 74 Apoptosis, hypertrophic chondrocytes, 37 ARSE, s e e Arylsulfatase Arteries, DDH, 240-243 Arteriovenous fistula, surgical induction, 672-673 Arthritis, s e e Juvenile rheumatoid arthritis; Septic arthritis Arthrodesis, hemophilia, 924 Arthrography coxa vara, 378 hip position, 218-221 index for LCP classification, 333 knee joint in infantile tibia vara, 486 proximal femur cartilage model, 311-312 Arthropathy, s e e Hemophilic arthropathy Articular cartilage development, 51-53 diaphysis lengthening effect, 692 pin penetration, in chronic SCFE treatment, 413-414 Arylsulfatase, skeletal dysplasias, 751 Ascending cervical arteries, DDH, 240-243 Asphyxiating thoracic dysplasia, 761,773 Atelosteogenesis, 759

936

Index

Atlantoaxial instability, cervical spine in skeletal dysplasia, 779-780 Authors, notable research studies achondrogenesis I, Parenti-Fracaro, 769 achondrogenesis II, Langer-Saldino, 769, 771 bone development Belchier, John, 5-7 Broca, 13 Duhamel, Henri-Louis, 7 Flourens, 8 Hales, Stephen, 5-7 Howship, 7-8 Hunter, John, 7 Kolliker, 13 Nesbitt, Robert, 5-7 Oilier, 15 Retterer, 15 bone as tissue or organ, Virchow, 14-15 congenital dislocation of hip Barlow, 204-205 Hilgenreiner, 203-204 Ortolani, 204 Putti, 204 Severin, 205-206 Von Rosen, 204-205 congenital limb deficiences, Frantz and O' Reilly, 615-617 developmental dysplasia of hip Badgley, 178-179 Bennett, 175 Brodhurst, 165-166 Bucholz, 252 Calandriello, 179-180 Clarke, 172 Cruveilhier, 164 Deutschlander, 176 Dunn, 181-183 Dupuytren, 162-163 Fairbank, 176-177 Goldsmith, 184-186 Graf, 221,223,225,229 Hirohashi, 253 Howorth, 179 Ippolito, 186 Kalamachi, 252 Keetley, 172-173 Kirmisson, 169-170 Kuhlmann, 252-253 Lance, 176 Le Damany, 173-175 Leveuf, 177-178 Loeffler, 176 Ludloff, 175-176 MacEwen, 252 Massie, 251-252 McKibbin, 183-184 Ogden, 252 Palletta, 161-162 Pavlik, 249-250

Ponseti, 186 Ralis, 183-184 Reeves, 166 Robert, 252 Roser, 164 Sainton, 166-169 Salter, 181 Scaglietti, 179-180 Sedillot, 163-164 Seringe, 252 Somerville, 186-187 Stanisavljevic, 180-181 Tonnis, 252-253 Verneuil, 164-165 Walker, 184-186 Werndorf, 176 diaphysis lengthening Abbott, 676-679 Anderson, 679-680 Cauchoix and Morel et al., 681 Codivilla, 675 D'Aubigne and Dubousset, 681 distraction osteogenesis, 688-695 Judet technique, 682-683 Kawamura, 681-682 Le Coeur, 680-681 Ombredanne, 675 Putti, 675-676 Wagner technique, 683-685 epiphyseal cartilage, Mueller, Heinrich, 13-14 epiphyseal growth plate, Dodds, 17-18 epiphyseal growth plate fractureseparations Aitken, 531-532 Aitken, Steinert, Weber, and Morscher, 536 Bergenfeldt, 531 Bidder, 526 Bret and Curtillet, 526 Broca, 523,535-536 Carothers and Crenshaw, 537 Cornil and Coudray, 526-528 Foucher, 523-525 Gueretin, 521-522 Hutchinson, 528-529 Jouon, 529 Morscher, 534 Nove-Josserand, 526 Ogden, 534-535 Oilier, 525-526 pathophysiology, Shapiro, 537-539 Peterson, 535 Poland, 529-530 Rognetta, 520-521 Salter and Harris, 532 Smith, 522 Steinert, 533 Vogt and Bruns, 525

Weber, 533-534 epiphyseal transplantation Farine, 719 Haas, 716-717 Harris et al., 718-719 Silfverskiold, 719 hemihypertrophy, Trelat and Monod, 640 hemophilic arthropathy, Koenig, 911 hip dislocation Bost, 202 Farrell, 202 Gill, 202 Howorth, 202 Leveuf, 201-202 hip position radiography, Tonnis, 214 infantile tibia vara B lount, 480-481,484 Golding and McNeil-Smith, 486 Jaffe, 485-486 Lamy and Weissman, 486 Langenskiold, 481,484-485 Sloane, Sloane, and Gold, 485-486 innominate osteotomy, Salter, 208-210 Legg-Calve-Perthes disease Axhausen, 282-283 Bell, 290 Bennett, 286-287 Birmingham splint, 345-346 Boldero, 299 Brailsford, 298-299 Brotherton and McKibbin, 345 Butel, Borgi, and Oberlin grading system, 335 Caffey, 299 Catterall, 293-295,304, 324-326 Delchef, 284 Dolman, 290 Evans, 343 Eyre-Brook, 330, 342 Ferguson, 285-286 Freeman, 292 Freund, 298 Gage, 304 Gall, 286-287 Green, Beauchamp, and Griffin, 332 Haythorn, 287 Heitzmann, 283 Herndon and Heyman, 330-331 Herring et al., 327-328 Hirohashi et al., 328-329 Hirrayama, 289-290 Hosoya, 292-293 Howorth, 285-286 Inoue, 292-293 Jansater, 333 Jensen, 292 Jonaster, 287-289 Kelly et al., 343 Kemp, 299

Index

Kendig, Evans, and Bohr, 344-345 Konjetzny, 283-284 Kotani, 289-290 Larsen, 290 Lauritzen, 292 Lippmann, 284 McKibbin, 290-292 Meyer, 331-332 Mizuno, 289-290, 292 Moiler, Fleming, 337 Mose concentric circle template method, 330 Nagassaka, 284-285 O'Garra, 324 O'Hara et al., 343 Ono, 292-293 Perpich et al., 343 Phemister, 282 Pike and Gossling, 342-343 Ponseti, 289, 295 Ralis, 290-292 Reiman, 290 Riedel, 283 Rockemer, 284 Salter-Thompson subchondral fracture, 326-327 Schwarz, 282 Shigeno and Evans, 333 Simazu, 289-290 Sjovall, 330 Stulberg, 333-335 Sundt, 330, 337-338 Takaoka, 292-293 Vernon-Roberts, 292 Walter, 283 Wiberg, 331 Yoshioka, 292-293 Zemansky, 281-282, 285 lower extremity length discrepancies Gill and Abbott, 654-655 Goidanich and Campanacci, 646-647 Green and Anderson, 655-658 Hatcher, 653 Hechard and Carlioz, 658-659 Menelaus, 658 Moseley, 658 White, 653-654 Wilson and Thompson, 654 Osgood-Schlatter disease Ehrenborg and Engfeldt, 498, 503-504 Hulting, 498 Krause et al., 507 Lazerte and Rapp, 501-502 Ogden, Hempton, and Southwick, 498-500 Uhry, 502-504 Woolfrey and Chandler, 506-507 osteochondritis dissecans Axhausen, 470

Barth and Kappis, 471 Konjetzny, 471 Paget, Teale, and Koenig, 466-467 osteogenesis imperfecta Shapiro, 848 Sillence, 848-849 pericapsular osteotomy of ilium, Pemberton, 210-211 physeal distraction DePablos et al., 704 Monitcelli and Spinelli, 704-705 Peltonen, 704 physeal transplantation Ring, 718 primary and secondary bone, Gegenbaur, 15 proximal femoral focal deficiency Aitken, 437 Amstutz, 437-438 Fixsen and Lloyd-Roberts, 438-439 Gillespie and Torode, 439-440 Haminishi, 440 Kalamchi, Cowell, and Kim, 440 Lange, Schoenecker, and Baker, 439 Pappas, 440 proximal femur Harris lines, O'Brien, 234 rickets Dodds and Cameron, 872-873 Shohl and Wolbach, 873-874 short rib-polydactyly syndromes Majewski, 772 Saldino-Noonan, 772 skeletal dysplasias Jeune, 761 Majewski, 761 Naumoff, 761 Saldino-Noonan short rib syndromes, 760-761 slipped capital femoral epiphysis Agamanolis et al., 388 Balensweig, 386 Cruess, 388 Elmslie, 385 Frangenheim, 384 Grashey, 385 Haedke, 384 Hofmeister, 384-385 Howorth, 386 Ingram, 389 Jerre, 389 Kallio et al., 397 Key, 386 Kleinberg and Buchman, 385 Kocher, 384 Lacroix and Verbrugge, 387 Lance et al., 389 Loder et al., 397 Mueller, 384 Ponseti, 387-388

937

Ponseti and Barta, 388 Rammstedt, 385-386 Schlessinger, 384 Sprengel, 385 Sutro, 386 Waldenstrom, 388 Walters and Simon, 413-414 teratological congenital dislocation of hip Cautru, 171 Grosse, 170-171 Kirmisson, 171 LePage, 170-171 Potocki, 171-172 Autoradiography, cell proliferation studies, 96 Avascular necrosis childhood CDH-DDH, 258 DDH treatment basic problem, 246-247 clinical and research studies, 247-249 closed reduction age, 249 Denis Browne splint, 254 extreme immobilization positioning, 258-262 Frejka abduction pillow, 253 milder sequelae, 251 Pavlik harness, 253-254 Pavlik's functional treatment, 249-250 prereduction traction benefits, 250-251 recent studies, 254-258 treatment principles, 247 as SCFE complication, 431-432 Avulsion fractures, proximal tibial tuberosity, 584-585 Axes, bone patterning mechanisms, 55-58 signaling regions, 53 tissue patterning models, 53-55 B Back pain, osteopetrosis, 845 Beckwith-Wiedemann syndrome, lower extremity length discrepancies, 646 Bed rest, LCP treatment, 343 Benign autosomal dominant osteopetrosis, 837, 841 Bilateral abduction brace, LCP treatment, 346-349 Bilateral abduction-internal rotation cast, LCP treatment, 346 Bilaterality, SCFE, 402-403 Biologic non-diplacement osteotomy, LCP treatment, 344-345 Biopsy, iliac crest bone, osteopetrosis, 844 Birmingham splint, LCP treatment, 345-346

938

Index

Births birth order, hip dysplasia, 194 epiphyseal growth plate fractureseparations, 528, 559-560 Bleeding hemophilia, 920 hemophilic arthropathy, 916 Blood chemistry, osteopetrosis, 843 Blood supply epiphyseal, s e e Epiphyseal blood supply epiphyseal growth plate fractureseparation, 540, 542-543, 548 proximal femur, DDH treatment, 239-245 Blount's disease, s e e Infantile tibia vara Blount stapling technique, limb shortening, 664-667 Body side SCFE predominance, 403 shortening in SCFE, 434 Body temperature, skeletal dysplasias, 861 Body weight, SCFE diagnosis, 400 Bone bridge epiphyseal growth plate fractureseparation, 596-597 lower extremity length discrepancies, 706-707 Bone cells cartilage cell transformation, 20 gap junction links, 19 Bone cysts aneurysmal cysts, 896 LCP disease, 314-316 unicameral cysts, 895-896 Bone disease, hemoglobinopathies, 926-927 Bone grafts chronic SCFE treatment, 418 LCP treatment, 360 SCFE treatment, 406-407 Bone growth early studies, 5-8 MR imaging, 135-136 Bone lesions, renal osteodystrophy, 885 Bone marrow imaging characteristics, 135 transplantation, osteopetrosis treatment, 846 Bone matrix, X-linked hypophosphatemic tickets, 879-880 Bone morphogenetic proteins, limb axes, 80-81 Bone patterning mechanisms, 55-58 signaling regions, 53 tissue patterning models, 53-55 Bones childhood renal osteodystrophy, 885-887 developing, s e e Developing bone

epiphyseal growth plate stimulation, 674-675 formation, 4-5, 16-19 hemangiomas, lower extremity length discrepancies, 643 imaging characteristics, 135 LCP disease, 278-279, 287-288 neurofibromatosis, 648 osteogenesis imperfecta, 857-858 renal osteodystrophy, 887-888 tissue vs. organ, 14-15 Bone scan, osteopetrosis, 843-844 Bone sialoproteins, 92, 95 Bowing, s e e Genu varum Breech position congenital-developmental hip abnormality effects, 193-194 young rabbit hip dislocation, 192 BSP, s e e Bone sialoproteins Burns, lower extremity length discrepancies, 652 C Caffey's disease, lower extremity length discrepancies, 652-653 Calcitriol, osteopetrosis management, 845-846 Calcium-phosphorus mineral phase, mineralization, 94 Caliper, LCP treatment, 343 Campomelic dysplasia, 760, 772 Capital epiphysis, epiphyseal growth plate fracture-separation, 575-576 Capsule dislocated hip, 180 Fairbank's studies, 177 laxity in hip abnormalities, 195-196 Caput index, LCP classification, 333 Cardiac system, skeletal dysplasias, 861 Cartilage achondrogenesis II, electron microscopy, 771 change in LCP disease, 288 MR imaging, 133-134 necrosis, SCFE, 388-389 Retterer's studies, 15-16 skeletal dysplasias, 753-754 Cartilage canals DDH treatment, 245-246 epiphyseal intrinsic blood supply, 47-50 secondary ossification center formation, 50-51 skeletal dysplasias, 767 Cartilage cells, 20-21 Cartilage collagens, epiphyseal tissue, 88-89 Cartilage-hair metaphyseal dysplasia, 812 Cartilage hyperplasie, Retterer's studies, 16 Cartilage matrix, skeletal dysplasias, 768

Cartilage matrix protein, epiphyseal tissue, 91 Cartilage models, LCP disease femoral head, 307-311 proximal femur, 311-312 Cartilage oligomeric matrix protein, skeletal dysplasia, 751 Cartilaginous epiphyses, skeletal dysplasias, 767 Catheters, abnormal lower leg growth, 627 CBFA1, skeletal dysplasias, 751-752 CDH, s e e Congenital dislocation of hip CDK, s e e Congenital dislocation of knee Cell death cartilage, 20-21 epiphyseal growth plate fractureseparation, 548-549 Cells bone forming, 19-20 bone patterning, 57-58 LCP disease, 281-295 proliferation, physeal cartilage, 96 skeletal dysplasias, 767-768 X-linked hypophosphatemic rickets, 879-880 Cell surface heparan sulfate, 751 Cell surface proteoglycans, 92 Cerebral vasculature, hemihypertrophy association, 641 Cericothoracic spinal stenosis, in achondroplasia, 799 Cervical arteries, DDH, 240-243 Cervical osteoplasty, 407-408 Cervical spina bifida occulta, 780 Cervical spinal stenosis, 781 Cheilectomy, LCP treatment, 363 Chiari osteotomy, LCP treatment, 363 Chick, limb bud, H o x c genes, 78 Childhood congenital dislocation of hip adult osteoarthritis, 198-199 avascular necrosis, 258 Childhood hip acetabulum, 167 anatomy, 166-167 congenital luxation of femur, 167-168 femoral-acetabular articulation, 167 pathoanatomy, 168-169 Childhood infantile tibia vara, 491-493 Childhood osteochondritis dissecans, 478-479 Childhood renal osteodystrophy, 885-887 Childhood septic arthritis, 905 Children's Hospital, Boston distal femoral fracture-separation studies, 579-582 SCFE epidemiologic data, 398-400 Chondroblastoma, 894-895 Chondroblasts, bone formation, 18-19 Chondrocytes bone formation, 18-19

Index

epiphyseal growth plate fractureseparation, 548-549 epiphyseal and physeal cartilage, 17-18 hypertrophic, s e e Hypertrophic chondrocytes physeal, metabolism, 96 suspensions, implantation, limb length discrepancy, 719 Chondrodysplasias characteristics, 759-760, 804 Grebe chondrodysplasia, 815 histopathological changes, 769-774, 777 lethal chondrodysplasias, 769-773 nonlethal chondrodysplasias, 773-774, 777 Chondrolysis, as SCFE complication, 432-434 Chondrosarcoma, hereditary multiple exostoses, 836 Chromosomes, skeletal dysplasias, 744-749 Chrondrolysis, SCFE, 388-389 Chronic repetitive activity knee lesions, 147 resulting physeal separation, 595-596 Chronic slipped capital femoral epiphysis treatments chronic slip, 396 femoral head-neck epiphysiodesis with bone graft, 418 femoral head position change, 419-430 growth plate fusion, 409-410, 413-419 hip spica casting, 418-419 pinning treatments, 409-410, 413-415, 417-418 Circulatory disturbances, distal femur osteochondritis dissecans, 469-470 Circumflex femoral arteries, DDH, 240 Clavicle, skeletal dysplasia, 785 Cleidocranial dysostosis, 813 Closed reduction CDH treatment, 205-206 chronic SCFE treatment, 419 DDH avascular necrosis, 249 femoral head vascularity, 261-262 early infancy hip development, 232 hip acetabular development, 232-233 CMP, s e e Cartilage matrix protein Collagen epiphyseal tissue, 82-89 mineralization, 93-94 osteogenesis imperfecta, 856-857 skeletal dysplasias, 750-754 COMP, s e e Cartilage oligomeric matrix protein Compensatory osteotomies, chronic SCFE treatment, 425-428

Compressive stress, epiphyseal plates, 103-105 Computerized tomography coxa vara, 378 DDH studies, 142 distal tibia and fibula disorders, 147 epiphyses, 132 hip structure, 230-231 LCP pathology, 303 lower extremity length discrepancies, 611 osteopetrosis, 843 upper extremities, 147-148 Concentric circle template method, LCP classification, 330 Congenital arteriovenous fistula, 646 Congenital coxa vara, 377 Congenital-developmental hip abnormalities acetabular dysplasia, 196 adductor muscle tightness, 196 capsular laxity, 195-196 ethnic considerations, 194 extrauterine postnatal environment, 194 genetic considerations, 194 idiopathic dysplasia, 197 intrauterine environment effects, 193-194 overview, 195 proximal femoral dysplasia, 196 secondary change worsening, 197-198 sex incidence, 192-193 side of instability, 193 soft tissue deformation, 196-197 stabilization without treatment, 194-195 teratologic dysplasia, 197 treatment, avascular necrosis, 250-251 Congenital dislocation of hip acetabular corrective procedures, 208-212 acetabular-proximal femoral osteotomies, 213 characteristics, 789 childhood adult osteoarthritis, 198-199 avascular necrosis, 258 Dunn's studies, 181-183 experimental reproduction, 190-192 Hilgenreiner's studies, 203-204 open reduction treatment, 206-207 Ortolani, 204 Pavlik harness, 206 proximal femoral osteotomies, 212-213 Putti, 204 radiographic classification system, 205-206 Stanisavljevic's studies, 181 Von Rosen and Barlow, 204-205 Congenital dislocation of knee classification, 507

939

clinical profile, 507 diagnostic considerations, 509 pathoanatomy, 508-509 treatment approaches, 509-510 Congenital hip subluxation, 180-181 Congenital limb deficiences classification, 615-617 dysmelia, 617-618 fibular hemimelia, 620-623 international terminology, 618 posteromedial bowing, 623-625 proximal femoral focal deficiency, 618 short femur, 618-620 tibial hemimelia, 623 Congenital lower extremity length discrepancies dysmelia, 617-618 fibular hemimelia, 620-623 Frantz and O'Reilly classification, 615-617 international terminology, 618 posteromedial bowing, 623-625 proximal femoral focal deficiency, 618 short femur, 618-620 tibial hemimelia, 623 Congenital luxation, child femur, 167-168 Congenital proximal tibial-fibular synostosis, 511 Congenital pseudoarthrosis, tibia, neurofibromatosis, 648-649 Congenital short femur, 442 Contralateral hip abnormalities, LCP disease, 279 Cord-root compression, cervical spine skeletal dysplasia, 781 Cortical bone, osteopetrosis, 842 Couches cartilaginous, 13 Couches chondroid, 13 Couches chondrospongioid, 13 Couches spongioid, 13 Coxa magna avascular necrosis in DDH treatment, 251 LCP disease, 307-309 Coxa plana, LCP disease, 309 Coxa valga, hereditary multiple exostoses, 832 Coxa vara acquired type, 377 causes, 376-378 clinical presentation, 378 congenital limb deficiences, 618-620 congenital type, 377 deformities, 376-377 imaging assessments, 378 infantile, s e e Infantile coxa vara LCP disease, 309-311 osteopetrosis, 844 skeletal dysplasias, 789 terminology, 376

940

Index

Craniocervical junction, achondroplasia, 798-799 Craniofacial dysostosis, 815-816 Craniosynostosis, campomelic dysplasia, 772 Crouzon syndrome, 816 Crown-rump length, hip development, 154 Crushing, mechanical, chondrocytes, 548-549 Crutches, LCP treatment, 343 CT, s e e Computerized tomography Cuneiform osteotomy, chronic SCFE treatment, 419-425 Cutis marmorata telangiectatica congenita, 646 Cysts, s e e Bone cysts Cytoarchitecture, growth plate, skeletal dysplasias, 766-767 D Databases, skeletal gene, 59-74 DDH, s e e Developmental dysplasia of hip Dedifferentiation, hypertrophic chondrocytes, 19-20 Deep femoral arteries, DDH, 240 Denis Browne splint, DDH treatment, 254 Determination wave mechanism, bone patterning, 57 Developing bone articular cartilage, 51-53 axes patterning mechanisms, 55-58 signaling regions, 53 tissue patterning models, 53-55 deformity, 107-109 diaphyseal bone formation, 41-44 early studies, 5-8 embryogenesis, 3-4 endochondral ossification, 4 epiphyseal tissue blood supply, 46-51 cartilage collagens, 88-89 cell surface proteoglycans, 92 collagen, 82-87 collagen groups, 87-88 endochondral sequence, 92-93 glycoproteins, 90-92 growth, 96-99 long bones, 111-118 noncollagenous proteins, 90-92 overview, 4, 21, 118-119 proteoglycans, 89-90 histological studies, 16-19 hypertrophic chondrocyte fate, 19-21 intramembranous ossification, 4-5 joint development, 44-46 light microscopic studies, 13-16 limb development apical ectodermal ridge, 58

dorsal nonridge ectoderm, 58 embryology, 8-11 gene control overview, 74 homeobox genes, 74-75, 77-78 matrix metalloproteinases, 81-82 polarizing region, 58 progress zone, 58 signaling molecules, 78-81 TIME 81-82 mechanical stresses abnormal pressure responses, 107-109 compressive and tensile stresses, 103-105 normal responses, 99-103 pressure effects, 109-111 skeletal development effects, 105-107 mineralization, 93-96 perichondrial ossification groove of Ranvier, 5, 39 periosteum relationships, 39-41 physis, structure and function, 25, 29, 31, 34-37 secondary ossification center formation, 21-25 skeletal gene database, 59-74 Developmental dysplasia of hip 3 months of age, 236-237 6 months of age, 237-238 12 months of age, 238 18 months of age, 238 18 months to 4.5 years of age, 238-239 after 5 years of age, 239 associated terminology, 153-154 avascular necrosis basic problem, 245-247 clinical studies, 247-249 closed reduction age, 249 Denis Browne splint, 254 extreme immobilization positioning, 258-262 Frejka abduction pillow, 253 milder sequelae, 251 Pavlik harness, 253-254 Pavlik's functional treatment, 249-250 prereduction traction benefits, 250-251 research studies, 247-249, 251-253 treatment principles, 247 early clinical-pathoanatomic descriptions, 161-170, 172-176 early treatment result reviews, 200-201 epiphyseal blood supply, 245-246 experimental reproduction, 190-192 extrinsic causes, 189-190 imaging, 142-143 intrinsic causes, 187-189

later clinical-pathoanatomic descriptions, 176-179 mid-twentieth century results, 201-203 newborns, 221,223,225,229-230, 235-236 overview, 187, 234-235 preadolescent-adolescent age, 239 proximal femur blood supply ascending cervical arteries, 240-243 deep and circumflex femoral arteries, 240 general pattern, 239-240 intracartilaginous-intraosseous vessels, 243-244 lateral and medial circumflex arteries, 240-243 ligamentum teres vascularity, 240 vascular pattern changes, 244-245 recent clinical-pathoanatomic descriptions, 179-187 teratological CDH, 170-172 treatment in 1800s and early 1900s, 199- 200 Diaphyseal periosteum characteristics, 39-41 elevation and stripping, 673 Diaphysis ipsilateral fracture, premature physeal closure, 594-595 lengthening, Ollier's disease, 820 lower extremity length discrepancy femoral diaphysis, 636-639 shortening, 669-672 tibial diaphysis, 639 lower extremity length discrepancy, lengthening concerns in complex abnormalities, 700-701 distraction osteogenesis, 685, 688-695 early clinical approaches, 675-682 humeral lengthening, 701 intramedullary rod lengthening, 695-698 longitudinal growth, 698-700 rigid fixator results, 682-685 woven and lamellar bone, 41-44 Diastrophic dysplasia characteristics, 805-807 histopathologic changes, 774 Diastrophic dysplasia sulfate transporter, 751,753 Digital dysplasia syndromes, 816 Diminished stature, LCP disease, 278-279 Distal femur acute epiphyseal growth plate fractureseparation, 576-582, 592 developmental abnormalities, 442 epiphyseal growth after infantile osteomyelitis, 903-904

Index

epiphyses, 362, 462-465 Ollier's disease, 818-819 osteochondritis dissecans age of occurrence, 468 causes, 468-470 childhood OD, treatments, 478-479 current understanding of disease, 474-475 disease profile, 465-466 healing factors, 477-478 original descriptions, 466-467 pathogenesis and pathoanatomy, 470-474 radiography and imaging, 475-477 site of occurrence, 468 stages, 467-468 radiographic characteristics, 116 stress analysis, 103 tilt in adolescent tibia vara, 495-496 Distal fibula acute epiphyseal growth plate fractureseparation, 591-592 disorders, imaging, 147 radiographic characteristics, 117 Distal humerus acute epiphyseal growth plate fractureseparation, 562, 566-570 radiographic characteristics, 112-115 Distal radius acute epiphyseal growth plate fractureseparation, 570-574 radiographic characteristics, 115 Distal tibia acute epiphyseal growth plate fractureseparation intra-articular component, 591 overview, 585-586 physis closure pattern, 586 Roentgen stereophotogrammetry, 591-592 transitional fractures, 586-589 type III medial fracture-separations, 589 type III medial malleolus, 589 type IV fracture, 589-591 associated disorders, 147 epiphysis, hereditary multiple exostoses, 832-833 radiographic characteristics, 116-117 Distal ulna acute epiphyseal growth plate fractureseparation, 574 radiographic characteristics, 115 Distraction osteogenesis, diaphysis lengthening articular cartilage effect, 692 associated research, 693-695 cell and matrix deposition patterns, 695 clinical techniques, 685, 688-690 muscle strength effect, 691-692

nerve function effect, 692-693 technique comparison, 690-691 Dorsal nonridge ectoderm, 58 Drilling, LCP treatment, 360 DTDST, s e e Diastrophic dysplasia sulfate transporter Dyggve-Melchior-Claussen dysplasia, 808 Dyschondrosteosis, 813 Dysmelia, congenital limb deficiences, 617-618 Dysplasia epiphysealis hemimelica characteristics, 803-804 lower extremity length discrepancies, 626 Dysplasias acetabular dysplasia, 196, 789 acrodysplasias, 815-816 acromelic dysplasia, 785, 815-816 acromesomelic dysplasia, 813-815 adolescent hip dysplasia, 239 asphyxiating thoracic dysplasia, 761, 773 campomelic dysplasia, 760, 772 cartilage-hair metaphyseal dysplasia, 812 CDH, s e e Congenital dislocation of hip chondrodysplasias, 759-760, 769-774, 777, 804, 815 DDH, s e e Developmental dysplasia of hip diastrophic dysplasia, 774, 805-807 digital dysplasia syndromes, 816 Dyggve-Melchior-Claussen dysplasia, 808 Grebe chondrodysplasia, 815 hip, natural history, 198 Hunter-Thompson acromesomelic dysplasia, 815 idiopathic dysplasia, 197 Jansen metaphyseal dysplasia, 812 Kniest dysplasia, 774, 804 lethal chondrodysplasias, 769-773 lethal thoracic dysplasia, 761 McKusick metaphyseal dysplasia, 812 mesomelic dysplasias, 813 metaphyseal dysplasia, 774, 812 metatropic dysplasia, 774, 804 multiple epiphyseal dysplasia, 774, 777, 801-803 nonlethal chondrodysplasias, 773-774, 777 proximal femoral dysplasia, 196 rhizomelic chondrodysplasia punctata, 772 Schmid metaphyseal dysplasia, 812 skeletal, s e e Skeletal dysplasias Smith-McCort dysplasia, 808 spondyloepimetaphyseal dysplasia, 8O4-8O5

941

spondyloepiphyseal dysplasia, 774, 807-808 spondyloepiphyseal dysplasia congenita, 807 spondyloepiphyseal dysplasia tarda, 807-808 spondylometaphyseal dysplasia, 774, 812-813 teratologic dysplasia, 197 thanatophoric dysplasia, 756-758, 771-772 trichorhinophalangeal dysplasia, 860 E Elbow, imaging, 148 Electron microscopy, achondrogenesis II cartilage, 771 Ellis-Van Creveld syndrome, 815 Embryonic development acetabulum, 159-160 basic theories, 3-4 femur, 158-159 limbs, 8-11 Enchondromatosis, lower extremity length discrepancies, 625-626 Endochondral ossification hypertrophic chondrocytes, 36-37 mechanism, 4 Endochondral sequence epiphyseal tissue, 92-93 mineralization, 93-96 Endocrine disorders, SCFE, 392-394 Eosinophilic granuloma, 895 Epidemiology LCP disease, 277-281 SCFE, 398-404 Epigenesis, embryogenesis theory, 3-4 Epiphyseal arrest, Ollier's disease, 820 Epiphyseal blood supply cartilage canals, 50-51 DDH treatment, 245-246 dual physeal supply, 46-47 extrinsic supply, 47 intrinsic supply, 47-50 transphyseal communicating cartilage canals, 50 Epiphyseal cartilage blood supply, DDH treatment, 245-246 chondrocyte shape, 17-18 early descriptions, 13-14 intrinsic blood supply, 47-50 skeletal dysplasias, 767 Epiphyseal extrusion index, LCP classification, 332 Epiphyseal growth plate fractureseparation acute, s e e Acute epiphyseal growth plate fracture-separation blood supply, 540, 542-543 chronic repetitive activity, 595-596

942

Index

Epiphyseal growth plate fractureseparation ( c o n t i n u e d ) clinical profile, 556-560 deformity pathogenesis, 548-549 early clinical descriptions, 519-520 genum recurvatum, 596 growth deformity pathogenesis, 548-549 incidence, 558 joint capsule position, 539-540 later studies, 528-530 MR imaging, 549-556 negative sequelae management, 596-597 other notable studies, 547-548 pathoanatomic studies, 523-528 pathologic, 592-594 pathophysiologic classification, 537-539 physeal cartilage, 543-545 premature physeal closure, 594-595 radiographic clinical approaches, 531-537 seminal studies, 520-522 shape, 541 slow research acceptance, 522-523 stability, 540-541 Epiphyseal index, LCP classification, 330 Epiphyseal-metaphyseal junction abnormalities, skeletal dysplasias, 768-769 shaping abnormalities, 762 Epiphyseal osteomyelitis acute neonatal-infantile type, 898-905 knee, 147 primary subacute-chronic type, 897 secondary to transphyseal spread, 897-898 tuberculosis, 906-909 Epiphyseal quotient, LCP classification, 330-331 Epiphyseal tissue, developing bone blood supply, 46-51 cartilage collagens, 88-89 cell surface proteoglycans, 92 collagen, 82-87 collagen groups, 87-88 endochondral sequence, 92-93 glycoproteins, 90-92 growth, 96-99 long bones, 111-118 noncollagenous proteins, 90-92 overview, 4, 21, 118-119 proteoglycans, 89-90 Epiphysiodesis, limb shortening, 667-668 Epiphysis blood supply, 46-47 cartilage collagens, 88-89 cell surface proteoglycans, 92 collagen, 82-87

collagen groups, 87-88 definition, 4 endochondral sequence, 92-93 formation and evolution, 118-119 fusion, knee, 632-633 glycoproteins, 90-92 growth cell proliferation, 96 freezing effects, 652 growth quantity, 97-98 Harris lines, 98-99 infection effect, 905-906 kinetics, 96-97 mechanical stress responses, 109-111 physeal chondrocyte metabolism, 96 slowdown, 98-99 stimulation in limb lengthening, 672-675 imaging growth and ossification, 135-137, 142 lower extremities, 142-147 technical aspects, 129-135 upper extremities, 147-148 lower extremity length discrepancies bone bridge resection, 706-707 chondrocyte suspension implantation, 719 focal physeal implant models, 715 free autogenous iliac crest physeal grafts, 708, 712-714 interpositional materials, 708 physeal reconstruction, 715 physes and epiphyses transplantation, 715-719 premature physeal closure treatment, 719-720 vascularized autogenous epiphyseal iliac crest grafts, 714-715 mechanical stress responses abnormal pressure responses, 107-109 compressive and tensile stresses, 103-105 normal responses, 99-103 skeletal development effects, 105-107 neoplastic disorders aneurysmal bone cyst, 896 chondroblastoma, 894-895 eosinophilic granuloma, 895 osteogenic sarcoma, 896-897 osteoid osteoma, 895 unicameral bone cyst, 895-896 noncollagenous proteins, 90-92 osteopetrosis, 841-842 periosteum relationship, 39-41 proteoglycans, 89-90 proximal femoral acute epiphyseal growth plate fractureseparation, 575-576

LCP disease, 304 proximal fibula disorders, 511-512 proximal tibial, radiographic developmental variants, 462-465 secondary ossification center formation, 21-25 shaping abnormalities, 762 skeletal dysplasias, 761-762, 767-768 slipped capital femoral, s e e Slipped capital femoral epiphysis Ethnicity, hip dysplasia, 194 EVC, s e e Ellis-Van Creveld syndrome Exostoses cord-root compression, 781 hereditary multiple exostoses, 831-832 skeletal dysplasias, 793 External fixator, intramedullary rod lengthening, 695-697 Extra-articular ligaments, hip development, 156 Extracellular matrix, epiphyseal tissue cartilage collagens, 88-89 cell surface proteoglycans, 92 collagen, 82-87 collagen groups, 87-88 endochondral sequence, 92-93 glycoproteins, 90-92 noncollagenous proteins, 90-92 proteoglycans, 89-90 Extrauterine postnatal environment, hip dysplasia, 194 Extremities lower, discrepancies, s e e Lower extremity length discrepancies skeletal dysplasia, 785-787 upper hereditary multiple exostoses, 833-835 Ollier's disease, 819 skeletal dysplasias, 794 F FACIT collagens, epiphyseal tissue, 88 Factor VIII, prophylactic coverage in hemophilia, 920-921 Familial hypophosphatemic rickets, s e e X-linked hypophosphatemic rickets Femoral-acetabular articulation, childhood hip, 167 Femoral arteries, DDH, 240-243 Femoral catheters, neonate abnormal lower leg growth, 627 Femoral diaphysis, lower extremity length discrepancies, 636-639 Femoral head abnormality in skeletal dysplasias, 789-790 chronic SCFE treatment, 419-430 LCP disease, 307-311, 319, 343-349, 351,353-355, 358-360

Index

osteonecrosis, hemoglobinopathies, 926-927 pathogenesis, LCP disease, 296 postreduction, MR imaging, 231-232 vascularity, after DDH treatment, 261-262 Femoral head-acetabular repair index, LCP classification, 330-333 Femoral head-neck epiphysiodesis, in chronic SCFE treatment, 418 Femoral neck, LCP disease, 313-316 Femoral-tibial diaphyseal angle, infantile tibia vara, 486-487 Femora vara, hereditary multiple exostoses, 832 Femur child, congenital luxation, 167-168 congenital limb deficiences, 618-620 developmental abnormalities, 436-442 developmental stages, 158-159 diaphyseal lengthening, 698 discrepancies in LCP disease, 321-323 dislocated hip, 180 distal, s e e Distal femur experimental head displacement, 190-192 Fairbank's studies, 176-177 LCP disease, shortening, 319-321 proximal, s e e Proximal femur shortening, lower extremity length discrepancies, 671-672 Fetus acetabulum and femur, 158-160 epiphyses, physeal reconstruction, 715 FGFR, s e e Fibroblast growth factor receptor FGFs, s e e Fibroblast growth factors Fibroblast growth factor receptor, skeletal dysplasias, 750, 753-754 Fibroblast growth factors, limb axes development, 79-80 Fibrochondrogenesis, histopathology, 772-773 Fibula bowing, posteromedial, 623-625 childhood proximal tibial metaphyseal fractures, 511 distal, s e e Distal fibula hemimelia, 620-623 hypoplasia, 511-512 lower extremity length discrepancies, 672 physes, radiographic characteristics, 117 proximal, s e e Proximal fibula Flexion contractures, skeletal dysplasias, 793 Focal physeal bone bridge resection, infantile tibia vara, 491 Focal physeal implants, animal models, 715

Foot, skeletal dysplasias, 794 Forearm, skeletal dysplasias, 785-787 Fractures avulsion, proximal tibial tuberosity, 584-585 childhood nonphyseal fractures, 557-558 physeal fractures, 557-558 proximal tibial metaphyseal fractures, 510-511 tibial metaphyseal fractures, 510-511 epiphyseal growth plate fractureseparations, 528, 545-547 hereditary multiple exostoses, 836 juvenile, Tillaux, 587-588 metaphyseal, renal osteodystrophy, 887 osteopetrosis, 844 pathological, Ollier's disease, 820 Salter-Harris type, 511,559 triplane, distal tibial epiphyseal fractureseparation, 588-589 Freezing, lower extremity length discrepancies, 652 Frejka abduction pillow, DDH treatment, 253 G Gadolinium, MR imaging, 132 Gait asymmetry, lower extremity length discrepancies, 610 Gap junctions, bone cell linking, 19 Gender epiphyseal growth plate fractureseparation incidence, 558 hip dislocations, 192-193 SCFE, 398, 400 Genes arylsulfatase, 751 hemophilia, 910 hip dysplasia, 194 H o x , 74-75, 77-78 limb development control, 58, 74-75, 77-78 osteochondritis dissecans of distal femur, 469 S H O X , 751 skeletal gene database, 59-74 S O X 9 , 751 X-linked hypophosphatemic tickets, 879 Genum recurvatum, after skeletal traction, 596 Genu valgum childhood, 462 skeletal dysplasias, 790-792 Genu varum physiologic, 462-465 posteromedial fibular and tibial, 623-625 skeletal dysplasias, 790-792 X-linked hypophosphatemic tickets, 880-881

943

GLA protein, epiphyseal tissue, 91 Glenoid labrum, hip development, 154-155 Gluteus medius, dislocated hip, 179-180 Glycoproteins, epiphyseal tissue, 90-92 Grafts, s e e Bone grafts Greater trochanter, LCP disease responses, 316 surgical intervention, 362 Grebe chondrodysplasia, 815 Green-Phemister technique, limb shortening, 664 Groove of Ranvier, skeletal dysplasias, 761-762 Growth disturbance lines, s e e Harris lines Growth plates fracture, s e e Epiphyseal growth plate fracture-separation fusion, in chronic SCFE treatment, 409-410, 413-419 lower extremity length discrepancies, 663-669 medial, childhood proximal tibial metaphyseal fractures, 511 proximal femoral growth plate, 117-118, 383 proximal tibial growth plate, 117-118, 582 radiographic characteristics, 117-118 skeletal dysplasias, 766 transplantation, 708, 712-715 H HA, s e e Hydroxyapatite Harness LCP treatment, 343 Pavlik harness, 206, 253-254 Harris lines epiphyseal growth, 98-99 proximal femur, 234 Head, deformity mechanism, 289-292 Head-neck area, SCFE, 382-383 Hemangiomas, lower extremity length discrepancies, 642-643 Hematologic disorders hemophilia, s e e Hemophilia osteopetrosis, 842-843 von Willebrand disease, 924-925 Hemiatrophy, lower extremity length discrepancies, 639-640 Hemihypertrophy, lower extremity length discrepancies Beckwith-Wiedemann syndrome, 646 cerebral vasculature abnormalities, 641 initial delineation, 6,40 Klippel~.Igri~iman~a~yndrome, 643-645 neoplasin associafi'on, 640-641 overview, 639.?--640 Parkes Weber~yndrome, 64~5 Proteus syndrome, 645-646

944

Index

Hemihypertrophy, lower extremity length discrepancies ( c o n t i n u e d ) Silver-Russell syndrome association, 641 Hemiparetic cerebral palsy, 629-630 Hemoglobinopathies bone disease pathophysiology, 926-927 clinical characteristics, 925 overview, 925 therapeutic approaches, 926 Hemophilia adult surgical treatment, 924 arthrodesis, 924 bleeding intervention, 920 childhood hemophilia, 910-918 gene abnormalities, 910 hemoglobinopathies, 925-927 inhibitors, 921 lower extremity length discrepancies, 651 medical treatment, 918-920 orthopedic management, 922 osteotomy, 924 overview, 909-910 pain management, 921-922 prophylactic factor VIII, 920-921 radiation synovectomy, 923-924 recalcitrant joints, 922 soft tissue manipulation, 924 surgical synovectomy, 922-923 Hemophilic arthropathy gross pathology, 911-912 histopathology, 913-915 joint evaluation, 916 pain and bleeding, 916 pathogenesis, 912-913 radiology, 915-916 stages, 911 Hereditary arthro-ophthalmopathy, 813 Hereditary multiple exostosis clinical problems, 831-836 deformities, 832-835 general stature, 831 histopathology, 821 overview, 512, 625,821 pathogenesis, 821,825,827-831 Hinge abduction, LCP treatment, 364 Hip childhood, s e e Childhood hip computerized tomography, 230-231 development acetabulum, 159-160, 232-234 early infancy, 232 etiologies, 160-161 femur, 158-159 general aspects, 154-158 normal and abnormal, 230 dislocations, 164-165, 198 disorders, imaging, 142-145 growth, 234 infancy, acute septic arthritis, 899-903

juvenile rheumatoid arthritis, 893 osteoarthritis, 608 position, 214, 217-221 septic arthritis, 630-631,905-906 skeletal dysplasias, 787-790 subluxation, natural history, 198 Hip spica cast, LCP treatment, 343 DDH immobilization, 261-262 SCFE treatment, 404-405 Hip subluxation congenital hip, 180-181 knee extension, 190 Histogenesis, bone formation, 16-17 Histology bone formation, 16-19 hip MR imaging, 259 physeal distraction, 704-705 skeletal, osteopetrosis, 841 transphyseal bone bridge, 549-552 Histopathology hemophilic arthropathy, 913-915 hereditary multiple exostoses, 821 LCP disease, 285,297 lethal chondrodysplasias, 769-773 nonlethal chondrodysplasias, 773-774, 777 Osgood-Schlatter disease, 503-504 osteogenesis imperfecta bone, 857-858 osteopetrosis, 841-842 proximal femoral focal deficiency, 441 skeletal dysplasias, 762, 766-769 tuberculosis, 907 Homeobox genes, limb development, 74-75, 77-78 Homozygous achondroplasia, 758-759, 772 Hormones, laxity, young rabbits, 192 H o x genes, s e e Homeobox genes Humerus acute epiphyseal growth plate fractureseparation, 561-562, 566-570 distal, s e e Distal humerus lengthening, 701 osteonecrosis in hemoglobinopathies, 927 proximal, s e e Proximal humerus radiographic characteristics, 111-115 Hunter syndrome, 810 Hunter-Thompson acromesomelic dysplasia, 815 Hurler syndrome, 810 Hyaline cartilage, imaging characteristics, 134 Hydroxyapatite, mineralization, 94 Hyperabduction, hip MR imaging, 259 Hyperparathyroidism, SCFE, 394 Hypertension, LCP disease, 280 Hypertrophic cell zone, Retterer's studies, 16

Hypertrophic chondrocytes apoptosis, 37 endochondral bone development role, 37 endochondral mechanism, 11 endochondral ossification, 36 enlargement, 36-37 light microscopic studies, 19-21 physis, 29 Hypochondrogenesis, 759, 771 Hypochondroplasia, 773,801 Hypophosphatasia, 760, 773 Hypophosphatemic rickets, pathophysiology, 878-879 Hypoplasia, fibula, 511-512 Hypothyroidism, 393 I Idiopathic dysplasia, congenitaldevelopmental hip abnormalities, 197 IGF, s e e Insulin-like growth factor Iliac crest bone biopsy, osteopetrosis, 844 free autogenous physeal grafts, 708, 712-714 vascularized autogenous epiphyseal grafts, 714-715 Iliopsoas muscle groups, dislocated hip, 179-180 Imaging, s e e Magnetic resonance imaging; Radiography Immobilization DDH treatment, 258-262 hip ischemia, 258-262 knee, 632-633 Implants epiphyseal growth plate stimulation, 674-675 focal physeal implants, 715 lower extremity length discrepancy treatment, 719 Indian hedgehog, limb axes development, 78-79 Infantile coxa vara characteristics, 377 clinical-radiographic correlations, 443-444, 448-449 deformity pathomechanics, 447-448 management, 449-451 pathoanatomy, 444-446 radiographic change, 446-447 terminology, 443 Infantile cortical hyperostosis, 652-653 Infantile malignant autosomal recessive osteopetrosis, 837 Infantile paralysis, hip dislocation, 164-165 Infantile tibia vara childhood, adult sequelae, 491-493 clinical and radiographic profile, 480-481

Index

deformities, 481-484, 488-490 femoral varus-tibia vara association, 496 imaging, 147,486-487 management, 487-488 pathoanatomy, 484-486 spontaneous correction, 490 surgical treatment, 490-491 terminology, 479-480 Infarction, LCP disease, 280-281 Infection childhood tuberculosis joints, 909 effect on epiphyseal growth, 905-906 Inflammatory disorders juvenile rheumatoid arthritis, 889-894 pigmented villonodular synovitis, 894 Inflammatory-pathologic theory, childhood hip, 167-168 Innominate osteotomy hip dislocation treatment, 208-210 LCP treatment, 355, 358-359 Insulin-like growth factor, limb axes, 81 Interferon ~/, osteopetrosis treatment, 846 Intermediate mild autosomal recessive osteopetrosis, 837 Interpedicular narrowing, achondroplasia, 799-800 Interpositional materials, epiphyses operation, 708 Interterritorial matrix, hypertrophic chondrocytes, 31, 34 Intertrochanteric compensatory osteotomies, 427-428 Intracartilaginous-intraosseous vessels, 243-244 Intramembranous ossification, 4-5 Intramedullary rod, diaphysis lengthening, 695-698 Intrauterine environment, congenitaldevelopmental hip, 193-194 Ipsilateral diaphysis, premature physeal closure, 594-595 Irradiation, childhood tumor, physeal damage, 635-636 Ischial weight bearing brace, LCP treatment, 343

J Jansen metaphyseal dysplasia, 812 Joints childhood tuberculosis infection, 909 contractures, juvenile rheumatoid arthritis, 893 development, 44-46 epiphyseal growth plate fractureseparation, 539-540 hemophilia, 922 hemophilic arthropathy, 916 hip development, 155 osteopetrosis, 842

septic arthritis, 904-905 skeletal dysplasias, 787 stabilization in limb length abnormalities, 700-701 Joint surface quotient, LCP classification, 331-332 Juvenile fracture, Tillaux, 587-588 Juvenile renal osteodystrophy, slipped epiphyses, 395 Juvenile rheumatoid arthritis lower extremity length discrepancies, 649-650 pathobiology, 889-890 profile, 889 skeletal abnormalities, 890-893 surgery, 893-894 K Klippel-Trenaunay syndrome, 643-645 Klippel-Trenaunay-Weber syndrome, 647-648 Knee disorders, imaging, 145-147 extension, hip dislocation, 190 infantile tibia vara, arthrography, 486 juvenile rheumatoid arthritis, 892-893 lower extremity immobilization, 632-633 septic arthritis, 904 skeletal dysplasias, 790-794 synovial hemangioma, 651 Kniest dysplasia, 774, 804 Knock-knee, s e e Genu valgum KTW, s e e Klippel-Trenaunay-Weber syndrome K-wire, proximal tibia placement, 596 Kyphoscoliosis, skeletal dysplasia, 782-783 Kyphosis, skeletal dysplasia, 782-783

L Lamellar bone, diaphyseal bone formation, 41-44 Langenskiold grade, infantile tibia vara deformity after osteotomy, 488 Larsen's syndrome, 816 Lateral circumflex femoral arteries, DDH, 240-243 Lateral epiphyseal lysis, Gage-Catterall sign, 304 Lateral humeral condyle, epiphyseal growth plate fracture-separation, 562, 566 Lateral-proximal neck convexity, Gage sign-Catterall sign, 304 LCP, s e e Legg-Calve-Perthes disease Legg-Calve-Perthes disease acetabular deformation, 296 acetabulum responses, 316-318 age of occurrence, 323-324

945

basic considerations, 323 Calve's study, 273-275 cartilage model, proximal femur, arthrography, 311-312 classification, skeletal maturity, 329-336 contralateral hip abnormalities, 279 definition, 272 delayed bone maturation role, 278-279 diminished stature role, 278-279 epidemiologic features, 277-278 femoral head cartilage model, 307-311 femoral head containment, 343-349, 351,353-355,358-359 femoral head non-containment, 360 femoral neck anteversion, 316 femoral shortening, 319-321 general features, 276-277 greater trochanter responses, 316 hinge abduction, 319 imaging, 143-145 imperfect healing, 319 incipient to residual stages, 297 late-stage surgical intervention, 362-364 Legg's study, 273 lower extremity length discrepancies, 321-323, 651-652 metaphysis response, 313-316 multiple infarctions, 280-281 nonhereditary disorder, 279 nonoperative, noncontainment treatment times, 361-362 non-weight-bearing treatment, 341-343 overview, 272-273 pathology, 281-295,300-303 Perthes' study, 275 physis response, 312-313 plain radiographic classifications, 297-299, 324-329 proximal femoral epiphysis nutrition, 304 radiodensity-radiolucency areas, 306-307 relation to age, 362 residual phase, remodeling, 318 sagging rope sign, 313 secondary ossification center, 304-306, 360 Sourdat's study, 276 subchondral fracture, 304-305 transient synovitis, 279-280 treatment approaches, 336-341, 360-361,364-368 venous hypertension role, 280 Waldenstrom' study, 275-276 Lesions angiomatous lesions, 646-647 bone, renal osteodystrophy, 885 childhood OD treatment, 479 knee, chronic repetitive trauma, 147

946

Index

Lethal chondrodysplasias, 769-773 Lethal thoracic dysplasia, 761 Ligamentum teres DDH, 240 dislocated hip, 180 hip development, 156 Light microscopy abnormal cell appearance, skeletal dysplasias, 767 bone development studies, 13-16 hypertrophic chondrocytes, 19-21 osteopetrosis ultrastructure, 841 Limb axes bone morphogenetic proteins, 80-81 fibroblast growth factors, 79-80 insulin-like growth factor, 81 parathyroid hormone, 81 parathyroid hormone receptor protein, 81 PTH/PTH/aR/AP receptor, 81 signaling molecules, 78-79 transforming growth factors-[3, 80 vitamin D, 81 Wnt 7A, 80 Limb bud chick, Hoxc genes, 78 hip development, 154-155 Limb development apical ectodermal ridge, 58 dorsal nonridge ectoderm, 58 gene control, 74-75, 77-78 long bones, 9-11 matrix metalloproteinases, 81-82 polarizing region, 58 progress zone, 58 timing and staging, 8-9 TIMP, 81-82 Limb lengthening lower extremity length discrepancies concerns in complex abnormalities, 700-701 distraction osteogenesis, 685, 688-695 early clinical approaches, 675-682 epiphyseal growth plate stimulation, 672-675 humeral lengthening, 701 intermedullary rod lengthening, 695-698 longitudinal growth, 698-700 rigid fixator results, 682-685 transiliac lengthening, 705-706 transphyseal lengthening, 701-705 skeletal dysplasias, 794 Limb shortening, lower extremity length discrepancies diaphyseal shortening, 669-672 epiphysiodesis, 667-668 growth plate therapeutic arrest, 663-667 metaphyseal shortening osteotomies, 669

partial therapeutic growth plate arrest, 668-669 Limbus CDH patients, infant acetabular development, 233-234 dislocated hip, 180 Long bones Broca's studies, 13 embryonic development, 9-11 epiphysis, radiographic characteristics, 112-118 formation, intramembranous ossification, 4-5 growth rate in skeletal dysplasia, 785-787 Kolliker's studies, 13 lower extremity length discrepancies, 653 osteopetrosis, 845 perichondrial ossification groove of Ranvier, 5 Salter-Harris type fracture distribution, 559 Low back pain, lower extremity length discrepancies, 608-610 Lower extremity length discrepancies angiomatous lesion classification, 646-647 associated angioplastic disorders, 641-642 Caffey's disease, 652-653 causes, 611-612 childhood tumor irradiation, 635-636 clinical effects, 608-610 clinically significant discrepancies, 606 congenital, see Congenital lower extremity length discrepancies destroyed physes, 627 developmental patterns, 612-615, 661-663 diaphysis lengthening, 675-685, 688-701 epiphyses operation, 706-708, 712-720 equal limb length, 606-608 external causes, 652 fractured femoral diaphysis, 636-639 fractured tibial diaphysis, 639 hemangiomas, 642-643 hemiatrophy, 639-640 hemihypertrophy, 639-641 hemiparetic cerebral palsy, 629-630 hemophilia, 651 hereditary multiple exostoses, 835-836 juvenile rheumatoid arthritis, 649-650, 890-892 knee joint synovial hemangioma, 651 LCP disease, 321-323, 362, 651-652 limb length determination, 610-611 limb lengthening, 672-675,701-706 limb shortening, 663-672

management considerations, 663 meningococcemia, 634-635 neonate catheter effects, 627 neurofibromatosis, 648-649 Ollier's disease, 819-820 osteomyelitis, 633-634 poliomyelitis, 627-629 premature epiphyseal fusion at knee, 632-633 SCFE, 652 septic arthritis of hip, 630-631 skeletal dysplasias, 625-627, 785 skeletal maturity, 653-661 thalassemia, 650-651 tuberculosis, 631-632 vascular malformations, 643-648 Lower limb, hereditary multiple exostoses, 832-833 Lumbar lordosis, 783 Lumbar spinal stenosis, 783 Lysosomal enzyme, 751 M Macrocephaly-hydrocephalus, 798 Maffucci syndrome, 626, 821 Magnetic resonance imaging bone and marrow, 135 cartilage, 133-134 coxa vara, 378 distal tibia and fibula disorders, 147 epiphyseal growth plate fractureseparation, 549-556 epiphyses, 130-132 femoral head postreduction, 231-232 hip disorders, 142-145 hip ischemia, extreme positioning, 258-262 infantile tibia vara, 486 knee disorders, 145-147 LCP pathology, 302-303 normal and abnormal bone growth, 135-137, 142 osteopetrosis, 843 upper extremities, 147-148 Malformations, childhood hip, 168 Malignant degeneration, hereditary multiple exostoses, 836 Malignant transformation, Ollier's disease, 820-821 Maroteaux-Lamy syndrome, 812 Matrix metalloproteinases, limb development, 81-82 McKusick metaphyseal dysplasia, 812 Mechanical stress, developing bone and epiphyses responses abnormal pressure responses, 107-109 compressive and tensile stresses, 103-105 normal responses, 99-103 pressure effects, 109-111

Index

skeletal development effects, 105-107 Medial circumflex femoral arteries, DDH, 240-243 Medial gap, hip, radiographic measurement, 214 Medial growth plate, childhood proximal tibial metaphyseal fractures, 511 Medial humeral condyle, epiphyseal growth plate fracture-separation, 569-570 Medial malleolus, type III fracture, 589 Medial physeal slope, infantile tibia vara, 487 Medial tibial articular surface, infantile tibia vara, 491 Melorheostosis, lower extremity length discrepancies, 627 Meningococcemia, lower extremity length discrepancies, 634-635 Mesomelic dysplasias, 813 Mesomelic shortening, skeletal dysplasia, 785 Metacarpals, acute epiphyseal growth plate fracture-separation, 574 Metaphyseal-diaphyseal angle, infantile tibia vara, 487 Metaphyseal dysplasia, 774 Metaphyseal fractures, renal osteodystrophy, 887 Metaphyseal periosteum, elevation and stripping, 673 Metaphyseal shortening osteotomies, 669 Metaphysis infection, effect on epiphyseal growth, 905-906 osteopetrosis, 841-842 periosteum relationship, 39-41 response in LCP disease, 313-316 skeletal dysplasias, 761-762, 766 Metatarsal growth plate, radiographic characteristics, 117-118 Metatropic dysplasia, 774, 804 Micromelic shortening, skeletal dysplasia, 785 Microscopy electron, achondrogenesis II cartilage, 771 light, s e e Light microscopy Midcervical kyphosis, cervical spine skeletal dysplasia, 780 Mineralization endochondral sequence, 17-18, 93-96 hypertrophic chondrocytes, 31, 34 MMPs, s e e Matrix metalloproteinases Models bone patterning, 55-57 cartilage, LCP disease, 307-312 focal physeal implants, 715 osteopetrosis, 841 rabbit, distraction osteogenesis, 695

rickets, 872-874 tissue patteming, 53-55 X-linked hypophosphatemic rickets, 880 Morbidity, achondroplasia, 796-798 Morphology, osteopetrosis, 841 Morquio syndrome, 810 Mortality, achondroplasia, 796-798 MPS, s e e Mucopolysaccharidoses MPSIH, s e e Hurler syndrome MPSII, s e e Hunter syndrome MPSIII, s e e Sanfilippo syndrome MPSIS, s e e Scheie syndrome MPSIV, s e e Morquio syndrome MPSVI, s e e Maroteaux-Lamy syndrome MPSVII, s e e Sly syndrome MRI, s e e Magnetic resonance imaging Mucolipidosis, 777 Mucopolysaccharidoses, 777, 808-812 Multiple epiphyseal dysplasia, 774, 777, 801-803 Muscles adductor, congenital-developmental hip abnormalities, 196 Fairbank's studies, 177 iliopsoas, dislocated hip, 179-180 strength, distraction osteogenesis, 691-692 Mutations, skeletal dysplasias, 744, 749-752 Myelomeningocele, 592-593 Myogenesis, childhood hip dislocation, 168 N Necrosis avascular, s e e Avascular necrosis osteonecrosis, s e e Osteonecrosis SCFE, 388-389 secondary ossification centers, 360 Neonates epiphyses, physeal reconstruction, 715 post-catheter abnormal lower leg growth, 627 proximal femoral physeal separation, 576 Neoplasia, hemihypertrophy association, 640-641 Neoplastic disorders, epiphyses aneurysmal bone cyst, 896 chondroblastoma, 894-895 eosinophilic granuloma, 895 osteogenic sarcoma, 896-897 osteoid osteoma, 895 unicameral bone cyst, 895-896 Neurofibromatosis, lower extremity length discrepancies, 648-649 Neurogenesis, childhood hip dislocation, 168 Neurology concerns in osteopetrosis, 845

947

diaphysis lengthening effect, 692-693 skeletal dysplasias, 861 Newborns DDH diagnosis, 235-236 fractures in distal femoral epiphysis, 579 Noncollagenous proteins, epiphyseal tissue, 90-92 Nonhereditary disorder, LCP disease, 279 Nonlethal chondrodysplasias, 773-774, 777 Nonphyseal fractures, distribution in childhood, 557-558 Nutritional rickets, deformities, 878 O OA, s e e Osteoarthritis Obesity, SCFE, 382, 403 OD, s e e Osteochondritis dissecans Odontoid process, cervical spine abnormalities, 779-780 OI, s e e Osteogenesis imperfecta Ollier's disease clinical sequelae, 818 deformities, 818-819 disease profile, 816-818 distal femoral deformities, 818-819 lower extremity length discrepancies, 625-626, 819-820 malignant transformation, 820-821 pathological fractures, 820 terminology, 816 tibial deformities, 818-819 upper extremity, 819 Open reduction CDH treatment, 206-207 chronic SCFE treatment, 419-425 hip acetabular development, 232-233 Organs, bone as, 14-15 Orthoroentgenograms, lower extremity length discrepancies, 611 Osgood-Schlatter disease adults, 506-507 clinical and radiologic features, 504 complications, 506 general treatment approach, 504-505 pathoanatomic changes, 501-504 pathophysiology, 497-498 surgical treatment, 505-506 terminology, 497 tibial tuberosity development, 498-501 Ossification endochondral ossification, 4, 36-37 intramembranous, mechanism, 4-5 MR imaging, 136 perichondrial groove of Ranvier, 5, 39 secondary centers, s e e Secondary ossification centers Osteitis fibrosa, renal osteodystrophy, 887-888

948

Index

Osteoarthritis adult, as SCFE complication, 434 hip, lower extremity length discrepancies, 608 osteopetrosis, 845 post-childhood CDH, 198-199 Osteoblasts, bone formation, 18 Osteocalcin, epiphyseal tissue, 91-92 Osteochondral fragment, medial malleolus fracture, 589 Osteochondritis dissecans age of occurrence, 468 causes, 468-470 childhood OD, treatments, 478-479 current understanding of disease, 474-475 disease profile, 465-466 healing factors, 477-478 imaging, 147 lesion in LCP disease, 319 original descriptions, 466-467 pathogenesis and pathoanatomy, 470-474 radiography and imaging, 475-477 site of occurrence, 468 stages, 467-468 Osteoclasts bone formation, 19 osteopetrosis, 842 Osteocytes bone formation, 18 diaphyseal bone formation, 43 Osteogenesis imperfecta bone, histopathology, 857-858 characteristics, 759 classification, 848-849, 855-856 clinical characteristics, 849, 852, 854 collagen defects, 856-857 diagnosis, 854-855 medical management, 860 orthopedic management, 858-860 overview, 847-848 Osteogenic sarcoma, 896-897 Osteoid osteoma, 895 Osteomyelitis epiphyseal, see Epiphyseal osteomyelitis infection effects, 905 lower extremity length discrepancies, 633-634 osteopetrosis, 845 Osteonecrosis hemoglobinopathies, 926-927 renal osteodystrophy, 885-888 Osteopetrosis blood chemistry studies, 843 diagnostic criteria, 842-843 hematologic studies, 842-843 histopathology, 841-842 iliac crest bone biopsy, 844 imaging studies, 843-844

neurologic concerns, 845 orthopedic concerns, 844-845 overview, 837 systemic management, 845-846 types, 837, 841 Osteoprogenitor cells, epiphyseal growth plate fracture-separation, 548 Osteotomy chronic SCFE treatment, 419-428 DDH, 212-213 hemophilia, 924 hip acetabular development, 232-233 hip dislocation treatment, 208-210 ilium, 210-211 infantile tibia vara, 488-490 LCP treatment, 344-345,349, 351, 353-355,358-359, 363-364 lower extremity length discrepancies, 669 proximal femoral growth, 234 P Pain hemophilia, 921-922 hemophilic arthropathy, 916 LCP treatment, 340-341 Parathyroid hormone limb axes, 81 mineralization, 94 Parathyroid hormone receptor protein, 81 Parathyroid hormone related peptide receptor, 751 Parkes Weber syndrome, lower extremity length discrepancies, 645 Parks-Harris line, MR imaging, 137, 142 Patellar dislocation, skeletal dysplasias, 793 Pathoanatomy adolescent tibia vara, 496-497 childhood hip, 168-169 congenital dislocation of knee, 508-509 epiphyseal growth plate fractureseparations, 523-528, 531-535 infantile coxa vara, 444-446 infantile tibia vara, 484-486 Osgood-Schlatter disease, 501-504 osteochondritis dissecans of distal femur, 470-474 proximal femoral focal deficiency, 441 SCFE, 383-392 septic arthritis of infant hip, 901-903 Pathobiology, rheumatoid arthritis, 889-890 Pathogenesis congenital luxation of femur, 167-168 developing bone deformity, 107-109 epiphyseal growth plate fractureseparation, 548-549 femoral head, LCP disease, 296 hemophilic arthropathy, 912-913

hereditary multiple exostoses, 821,825, 827-831 infantile tibia vara deformity, 481-484 LCP disease, venous hypertension role, 280 nutritional rickets skeletal abnormality, 877-878 osteochondritis dissecans of distal femur, 470-474 renal osteodystrophy bone deformities, 887-888 rickets, 872 SCFE, 389-392 skeletal dysplasias, 769 Pathological fractures, Ollier's disease, 820 Pathology hemophilic arthropathy, 911-912 LCP disease, 281-295 vitamin D deficiency rickets, 874-876 Pathomechanics, infantile coxa vara deformity, 447-448 Pathophysiology epiphyseal growth plate fractureseparation blood supply, 539-548 classifications, 537-539 deformity pathogenesis, 548-549 MR imaging, 549-556 familial hypophosphatemic rickets, 878-879 hemoglobinopathy bone disease, 926-927 nutritional rickets, 877 Osgood-Schlatter disease, 497-498 renal rickets, 884-885 Patterning, bone mechanisms, 55-58 signaling regions, 53 tissue patterning models, 53-55 Pavlik harness avascular necrosis in DDH treatment, 253-254 CDH treatment, 206 Percutaneous fixation, in chronic SCFE treatment, 415-417 Percutaneous technique, limb shortening, 667 Pericapsular osteotomy, ilium, 210-211 Perichondrial ossification groove of Ranvier epiphyseal growth plate, 39 mechanism, 5 Perichondrium, imaging characteristics, 134 Periosteum diaphysis, epiphysis, and metaphysis, 39-41,673 Nesbitt's studies, 6 Periphyseal tissues, groove of Ranvier, 761-762

Index

Pfeiffer syndrome, 816 PFFD, s e e Proximal femoral focal deficiency Phalangeal growth plate, radiographic characteristics, 117-118 Phalanges, acute epiphyseal growth plate fracture-separation, 574 Phemister technique, limb shortening, 664 Physeal cartilage cell proliferation, 96 chondrocyte shape, 17-18 epiphyseal growth plate fractureseparation direct damage, 543-545 fracture pathway, 545-547 skeletal dysplasias, 767 Physeal chondrocytes, metabolism, 96 Physeal distraction histology, 704-705 premature closure treatment, 719-720 Physeal fracture, distribution in childhood, 557-558 Physis closure avascular necrosis in DDH treatment, 251 distraction treatment, 719-720 growth plate transplantation treatment, 708, 712-715 premature, LCP disease, 320 radiographic characteristics, 117 damage, childhood tumor irradiation, 635-636 destroyed, lower extremity length discrepancies, 627 height, adolescent tibia vara, 495-496 response in LCP disease, 312-313 skeletal dysplasias, 768 structure and function, 25, 29, 31, 34-37 transplantation, 715-718 Pigmented villonodular synovitis, 894 Pillar, classification in LCP disease, 327-328 Pinning, chronic SCFE treatment articular cartilage penetration danger, 413-414 femoral head position change, 429 in situ studies, 417-418 stabilization pins, 414-415 transphyseal pinning, 409-410, 413 Plain radiography, coxa vara, 378 Poland syndrome, 816 Poliomyelitis, lower extremity length discrepancies, 627-629 Polydactyly syndromes, 816 Positional information models, bone patterning, 55-57 Preformationism, embryogenesis theory, 3-4

Prepattern mechanism, bone patterning, 57 Pressure epiphyseal growth effects, 109-111 phenomena in hereditary multiple exostoses, 836 physes responses, abnormal pressure, 107-109

Progress zone, 58 Prophylactic pinning, chronic SCFE treatment, 430-431 Proteoglycans cell surface heparan sulfate, 751 cell surface type, 92 epiphyseal tissue, 89-90 mineralization, 94 Proteus syndrome, lower extremity length discrepancies, 645-646 Proximal femoral dysplasia, 196 Proximal femoral epiphysis, LCP disease, 304 Proximal femoral focal deficiency classifications, 437-440 clinical characteristics, 440-441 congenital limb deficiencies, 618 imaging, 142 pathoanatomic studies, 441 treatment options, 441-442 Proximal femoral growth plate radiographic characteristics, 117-118 SCFE, 383 Proximal femoral osteotomies, DDH, 212-213 Proximal femoral valgus osteotomy, LCP treatment, 363-364 Proximal femur acute epiphyseal growth plate fractureseparation, 575-576 blood supply, DDH treatment ascending cervical arteries, 240-243 deep and circumflex femoral arteries, 240 general pattern, 239-240 intracartilaginous-intraosseous vessels, 243-244 lateral and medial circumflex arteries, 240-243 ligamentum teres vascularity, 240 vascular pattern changes, 244-245 cartilage model, LCP disease, arthrography, 311-312 growth after osteotomy, 234 Harris lines, 234 radiographic characteristics, 115-116 stress analysis, 103-105 Proximal femur varus-derotation osteotomy, LCP, 349, 351,353-355 Proximal fibula acute epiphyseal growth plate fractureseparation, 591 elongation, 511

949

epiphysis disorders, 511-512 radiographic characteristics, 117 Proximal humerus acute epiphyseal growth plate fractureseparation, 561-562 radiographic characteristics, 111-112 Proximal radius acute epiphyseal growth plate fractureseparation, 570 radiographic characteristics, 115 Proximal tibia acute epiphyseal growth, plate fractureseparation anterior tibial spine, 582-584 avulsion fractures of tibial tuberosity, 584-585 ligament damage, 592 growth plate, radiographic characteristics, 117-118 infantile tibia vara, metaphysealdiaphyseal angle, 487 K-wire, placement, 596 metaphyseal fractures, childhood deformities, 510-511 physeal arrest, infantile tibia vara treatment, 490-491 physis, secondary proximal fibular overgrowth, 512 radiographic characteristics, 116 stress analysis, 103 Proximal tibial-fibular osteotomy, infantile tibia vara treatment, 490 Proximal ulna, radiographic characteristics, 115 Pseudo-achondroplasia, 777, 808 PTH, s e e Parathyroid hormone PTH/PTHrP receptor, limb axes, 81 PTHRP, s e e Parathyroid hormone receptor protein PTHRPR, s e e Parathyroid hormone related peptide receptor Pulmonary system, skeletal dysplasias, 861 Pycnodysostosis, 846-847 R

Rabbit breech malposition, 192 distraction osteogenesis model, 695 femoral head displacement, 190-192 knee extension, 190 Radiation synovectomy, hemophilia, 923-924 Radiodensity, LCP disease, 305-307 Radiography achondroplasia, 795-796 adolescent tibia vara, 496 CDH studies, 205-206 epiphyseal growth plate fractureseparation, 531-537 hip position, 214, 217-218

950

Index

Radiography ( c o n t i n u e d ) infantile coxa vara, 443-444, 446-447, 449-451 infantile tibia vara, 480-481 LCP disease, 297-299, 324-329, 364-366 long bone epiphyses, 111-118, 462-465,475-477 lower extremity length discrepancies, 611 nutritional tickets, 878 osteopetrosis, 843-844 plain, coxa vara, 378 SCFE severity, 396 skeletal dysplasia diagnosis, 743-744 tuberculosis, 906-907 X-linked hypophosphatemic tickets, 880-882 Radiology equal lower extremity limb lengths, 607-608 hemophilic arthropathy, 915-916 Osgood-Schlatter disease, 504 Radiolucency, LCP disease, 306-307, 314-316 Radiotherapy, in predisposition to SCFE, 394-395 Radius acute epiphyseal growth plate fractureseparation, 570-574 radiographic characteristics, 115 Renal osteodystrophy bone deformity pathogenesis, 887-888 bone lesions, 885 childhood bone deformities, 885-887 orthopedic treatment, 888-889 pathophysiology, 884-885 Renal tickets, s e e Renal osteodystrophy Retinacula of Weitbrecht, hip development, 155-156 Retinoids, limb axes development, 78 Rheumatoid arthritis, s e e Juvenile rheumatoid arthritis Rhizomelic chondrodysplasia punctata, 772 Rhizomelic shortening, skeletal dysplasia, 785 Rickets experimental models, 872-874 nutritional tickets, 877-878 pathogenesis, 872 renal tickets, 884-889 terminology, 872 vitamin D deficiency, pathology, 874-876 X-linked hypophosphatemic, s e e X-linked hypophosphatemic tickets Rigid fixators, diaphysis lengthening, 682-685 Roentgen stereophotogrammetry, growth plate fracture-separations, 591-592

S Sagging rope sign, radiographic LCP disease characteristic, 313 Salter-Harris type fractures childhood proximal tibial metaphyseal fractures, 511 epiphyseal growth plate fractureseparation, 559 Sanfilippo syndrome, 810 Scanograms, lower extremity length discrepancies, 611 SCFE, s e e Slipped capital femoral epiphysis Scheie syndrome, 810 Schmid metaphyseal dysplasia, 812 Sciatica, lower extremity length discrepancies, 608-610 Scintigraphy LCP pathology, 300-302 skeleton, 132-133 Scoliosis achondroplasia, 800 skeletal dysplasia, 782-783 Screening, SCFE, 403-404 Secondary ossification centers formation, 21-25, 50-51 hip position radiography, 217-218 LCP treatment, 304-306, 360 osteopetrosis, 842 skeletal dysplasias, 767 Septic arthritis childhood, long-term sequelae, 905 hip, 630-631,905-906 infant hip negative growth sequelae, 900 overview, 899-900 pathoanatomy, 901-903 treatment approaches, 901 non-hip joints, 904-905 tuberculosis, 906-909 Short rib syndromes characteristics, 760-761 polydactyly type, 772 Short stature homeobox, skeletal dysplasias, 751 Shortwave diathermy, limb lengthening, 673-674 Shoulder imaging, 147-148 septic arthritis, 904-905 SHOX, s e e Short stature homeobox Sickle cell anemia, clinical characteristics, 925 Signaling molecules, limb axes bone morphogenetic proteins, 80-81 fibroblast growth factors, 79-80 Indian hedgehog, 78-79 insulin-like growth factor, 81 parathyroid hormone, 81 parathyroid hormone receptor protein, 81

PTH/PTHrP receptor, 81 retinoids, 78 sonic hedgehog, 78-79 transforming growth factors- [3, 80 vitamin D, 81 Wnt 7A, 80 Signaling regions, bone development patterning, 53 Silver-Russel syndrome, hemihypertrophy association, 641 Skeletal dysplasias achondroplasia, 794-801 acromelic dysplasia, 815-816 acromesomelic dysplasia, 813-815 affected developmental cycle, 754-755 anesthetic considerations, 860-861 associated coxa vara, 378 chondrodysplasia punctata, 804 chromosome abnormalities, 744-749 classification approaches, 733-734, 737-738 cleidocranial dysostosis, 813 clinical examination, 740-743 developing epiphyses and metaphyses structures, 761-762 diastrophic dysplasia, 805-807 Dyggve-Melchior-Claussen dysplasia, 808 dyschondrosteosis, 813 dysplasia epiphysealis hemimelica, 803-804 hereditary arthro-ophthalmopathy, 813 hereditary multiple exostoses, 821,825, 827-836 histopathologic classification, 762, 766-769 hypochondroplasia, 801 Kniest dysplasia, 804 laboratory studies, 743-744 Larsen's syndrome, 816 lethal chondrodysplasias, 769-773 lethal perinatal achondrogenesis, 759 asphyxiating thoracic dystrophy, 761 atelosteogenesis, 759 campomelic dysplasia, 760 chondrodysplasia punctata, 759-760 diagnostic profile, 755-756 homozygous achondroplasia, 758-759 hypochondrogenesis, 759 hypophosphatasia, 760 osteogenesis imperfecta, 759 short rib syndromes, 760-761 thanatophoric dysplasia, 756-758 limb lengthening, 794 lower extremity length discrepancies, 625-627 Maffucci syndrome, 821 mesomelic dysplasias, 813

Index

metaphyseal dysplasia, 812 metatropic dysplasia, 804 molecular function defects, 752-754 mucopolysaccharidoses, 808-812 multiple epiphyseal dysplasia, 801-803 mutation families, 744, 749-752 nonlethal chondrodysplasias, 773-774, 777 Ollier's disease, 816-821 orthopedic deformities ankle abnormalities, 793-794 cervical spine abnormalities, 779-781 clavicle abnormalities, 785 extremity abnormalities, 785-787 foot abnormalities, 794 hip abnormalities, 787-790 knee abnormalities, 790-794 lumbar lordosis, 783 lumbar spinal stenosis, 783 overview, 778-779 skull abnormalities, 785 thoracolumbar spine abnormalities, 782-783 upper extremity abnormalities, 794 osteogenesis imperfecta, 847-849, 852, 854-860 osteopetrosis, 837, 841-846 overview, 738-740 pathogenesis, 769 prenatal assessment, 740 prevalence, 738 pseudo-achondroplasia, 808 pycnodysostosis, 846-847 radiographic examinations, 743 Smith-McCort dysplasia, 808 spondyloepimetaphyseal dysplasia, 804-805 spondyloepiphyseal dysplasia, 807-808 spondylometaphyseal dysplasia, 812-813 terminology, 733 trichorhinophalangeal dysplasia, 860 Skeleton development, mechanical stress effects, 105-107 gene database, 59-74 hip development, 156 histology, osteopetrosis, 841 juvenile rheumatoid arthritis, 890-893 maturity, LCP disease adult responses to childhood disorder, 335-336 Butel, Borgi, and Oberlin grading system, 335 femoral head-acetabular repair indices, 330-333 general considerations, 329-330 Stulberg classification, 333-335 Sundt classification, 330

maturity, lower extremity length discrepancies long bone growth percentage, 653 methods of determination, 659-661 prediction systems, 653-659 maturity, prior diaphysis lengthening, 698-700 nutritional rickets, 877-878 osteopetrosis, imaging studies, 843-844 prenatal sonographic evaluation, 129-130 scintigraphy, 132-133 ultrasonography, 129 Skin traction, LCP treatment, 343 Skull, skeletal dysplasia, 785 Sling, LCP treatment, 343 Slipped capital femoral epiphysis anatomic-physiological features, 382-383 classifications, 396-398 clinical awareness and description, 379-381 complications adult osteoarthritis, 434 avascular necrosis, 431-432 chondrolysis, 432-434 involved side shortening, 434 stiffness, 434 diagnostic imaging studies, 404-408 epidemiologic characteristics Children's Hospital, Boston, 398-400 overview, 398 school screening recommendations, 403-404 study comparisons, 400-403 imaging, 145 initial theories, 381-382 juvenile renal osteodystrophy, 395 long-term follow-up studies, 434-436 lower extremity length discrepancies, 652 medical disorder predisposition, 383 obesity, 382 pathoanatomy, 383-389 pathogenesis and pathoanatomy, 389-392 predisposing medical disorders, 392-395 terminology, 378-379 therapy goals, 404 treatment, 404-409, 430-431 Slipped epiphyses, renal osteodystrophy, 885-887 Sly syndrome, 812 Smith-McCort dysplasia, 808 Smith-Peterson nail, SCFE treatment, 407 Soft tissues childhood proximal tibial metaphyseal fractures, 511 congenital-developmental hip abnormalities, 196-197

951

manipulation in hemophilia, 924 PFFD, 441 Sonic hedgehog, limb axes development, 78 Sonography indices, normal and abnormal hip development, 230 prenatal skeleton, 129-130 SOX9, skeletal dysplasias, 751,754 Spine abnormalities in achondroplasia, 799-800 skeletal dysplasias, 861 Splints Birmingham splint, LCP treatment, 345-346 Denis Browne splint, DDH treatment, 254 Spondyloepimetaphyseal dysplasia, 804-805 Spondyloepiphyseal dysplasia, 774, 807-808 Spondylometaphyseal dysplasia, 774, 812-813 Stickler syndrome, 813 Stiffness, as SCFE complication, 434 Stress compressive, epiphyseal plates, 103-105 mechanical, see Mechanical stress resulting physeal separation, 595-596 tensile, epiphyseal plates, 103-105 Structural proteins, skeletal dysplasias, 752-753 Subluxation congenital hip, 180-181 hip, knee extension, 190 LCP disease, 309 Sub-periosteal bone, osteopetrosis, 842 Superior gap, hip, radiographic measurement, 214 Superior talar articular surface, hereditary multiple exostoses, 833 Surgery acetabular growth, 234 adult hemophilia, 924 childhood OD treatment, 479 epiphyseal growth plate fractureseparation treatment, 530 induced arteriovenous fistula, 672-673 infantile tibia vara treatment, 490-491 juvenile rheumatoid arthritis, 893-894 LCP disease, 349, 351,353-355, 358-359, 362-364 Osgood-Schlatter disease treatment, 505-506 SCFE treatment, 404-405 Surgical synovectomy, hemophilia, 922-923 Survival, hypertrophic chondrocytes, 19-20

952

Index

Sympathectomy, limb lengthening, 672 Syndactyly, 816 Synovectomy, hemophilia, 922-924 Synovial hemangioma, knee joint, 651 Synovium, hip development, 155 T Teleoroentgenograms, lower extremity length discrepancies, 611 Tensile stress, epiphyseal plates, 103-105 Teratologic dysplasia, congenitaldevelopmental hip abnormalities, 197 TGF-[L s e e Transforming growth factors-[3 Thalassemia, 650-651,925 Thanatophoric dysplasia, 756-758, 771-772 Therapeutic arrest, lower extremity length discrepancies, 663-669 Thoracolumbar kyphosis, in achondroplasia, 800 Thoracolumbar spine, skeletal dysplasia, 782-783 Tibia congenital pseudoarthrosis, 648-649 diaphyseal lengthening, 698 distal, s e e Distal tibia LCP disease, 321-323 lower extremity length discrepancies, 672 metaphyseal fractures in childhood, 510-511 Ollier's disease, 818-819 Osgood-Schlatter disease, 498-501 proximal, s e e Proximal tibia Tibia-fibula relationship, hereditary multiple exostoses, 832 Tibial bowing, posteromedial, 623-625 Tibial diaphysis, lower extremity length discrepancies, 639 Tibial hemimelia, 623 Tibial metaphyseal fractures, childhood, 510-511 Tibial tubercle chronic traumatic apophysitis, s e e Osgood-Schlatter disease Tibia valga, hereditary multiple exostoses, 832 Tibia vara, imaging, 147 Tillaux, juvenile fracture, 587-588 TIMPs, s e e Tissue inhibitor of matrix metalloproteinases Tissue inhibitor of matrix metalloproteinases, limb development, 81-82 Tissues bone as, 14-15 epiphyseal, s e e Epiphyseal tissue fetal and postnatal epiphyses, 715 LCP disease, 281-295

patterning models, 53-55 periphyseal, groove of Ranvier, skeletal dysplasias, 761-762 soft, s e e Soft tissues Tissue spongieux, early description, 13 Traction avascular necrosis in DDH treatment, 250-251 genum recurvatum, 596 skin, LCP treatment, 343 Transcription factors, skeletal dysplasias, 751-752, 754 Transformation, cartilage cells to bone cells, 20 Transforming growth factors-[3, limb axes, 80 Transient synovitis, LCP disease, 279-280 Transiliac process, lengthening, 705-706 Transphyseal bone bridge, epiphyseal growth plate fracture-separation, 549-552 Transphyseal chondrodiatasis, epiphyseal growth plate fracture-separation, 597 Transphyseal communicating cartilage canals, blood supply, 50 Transphyseal drilling, SCFE treatment, 406 Transphyseal fracture, epiphyseal growth plate fracture-separations, 528 Transphysis, lengthening in lower extremity length discrepancies, 701-705 Transplantation bone marrow, osteopetrosis treatment, 846 growth plates, 708, 712-715 lower extremity length discrepancies, 715-719 Trauma chronic repetitive, knee lesions, 147 distal femur osteochondritis dissecans, 468-469 distal tibia and fibula, 147 Traumatic births, epiphyseal growth plate fracture-separations, 528 Traumatic theory of congenital dislocation of hip, 167 Trichorhinophalangeal dysplasia, 860 Triplane fracture, distal tibial epiphyseal fracture-separation, 588-589 Trochanteric height, LCP classification, 333 Trochanteric overgrowth, avascular necrosis in DDH treatment, 251 Tuberculosis age incidence, 906 childhood, joint infection, 909 histopathology, 907

lower extremity length discrepancies, 631-632 overview, 906 radiography, 906-907 treatment, 908-909 Tumors, irradiation in childhood, physeal damage, 635-636 U Ulna acute epiphyseal growth plate fractureseparation, 574 radiographic characteristics, 115 Ultrasonography coxa vara, 378 LCP pathology, 302 lower extremity length discrepancies, 611 newborn DDH, 221,223, 225, 229-230 SCFE, 397 skeleton, 129 Umbilical catheters, neonate abnormal lower leg growth, 627 Undifferentiated mesenchymal cells, bone formation, 18 Unicameral bone cyst, 895-896 Upper extremities hereditary multiple exostoses, 833-835 Ollier's disease, 819 skeletal dysplasias, 794 V Valgus deformity childhood proximal tibial metaphyseal fractures, 510-511 knee in juvenile rheumatoid arthritis, 892-893 SCFE, 397-398 Vascularity cartilaginous epiphyses, skeletal dysplasias, 767 endochondral sequence, 17-18 femoral head after DDH treatment, 261-262 postreduction, 231-232 malformations, lower extremity length discrepancies Beckwith-Wiedemann syndrome, 646 congenital arteriovenous fistula, 646 cutis marmorata telangiectatica congenita, 646 Klippel-Trenaunay syndrome, 643-645 Klippel-Trenaunay-Weber syndrome, 647-648 Parkes Weber syndrome, 645 Proteus syndrome, 645-646

Index

953

Venous hypertension, LCP disease, 280 Vitamin D limb axes, 81 mineralization, 94 Von Willebrand disease, 924-925

Woven bone, diaphyseal bone formation, 41-44 Wrist imaging, 148 septic arthritis, 905

gene abnormality, 879 hypothetical model, 880 medical treatment, 882-883 orthopedic management, 883-884 pathophysiology, 878-879

W Weight, infantile tibia vara deformity after osteotomy, 488 Weight relieving caliper, LCP treatment, 343 Wnt 7A, limb axes, 80

X XLH, s e e X-linked hypophosphatemic rickets X-linked hypophosphatemic rickets cell and matrix abnormalities, 879-880 characteristics, 880-882

Z Zone of polarizing activity definition, 58 gene control of limb development, 75 sonic hedgehog, 79 ZPA, s e e Zone of polarizing activity