Gastroenterol Clin N Am 37 (2008) xi–xii
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Preface
Nicholas J. Talley, MD, PhD, FRACP, FRCP, FACP Guest Editor
I
t is fascinating to watch the evolution of medical practice and the introduction of new disease entities in medicine. In this issue of Gastroenterology Clinics of North America, a number of examples of relatively newly recognized diseases are comprehensively reviewed. The goal of this issue is to bring clinicians up to date on the topic of eosinophil and autoimmune gut diseases, where knowledge is rapidly advancing and clinical practice has or may be changed as a consequence. We are now recognizing the potential role of tissue eosinophilia in many gastrointestinal diseases, both new and old. Eosinophilic esophagitis was first reported in 1978 in children, but this entity is now clearly recognized to be increasing in incidence in both pediatric and adult gastroenterology practice for reasons that are as yet unclear. How much of this disease is because of acid reflux and how much is explained by aeroallergens remains controversial, but treatment algorithms have advanced, as is summarized for both adult and pediatric patients in the first two articles in this issue. The next article covers an important clinical entity of unexplained cause, namely eosinophilic gastroenteritis. Treatment observations in patients with this rare entity have usefully been applied in other eosinophilic diseases of the gut, and the topic is knowledgeably summarized here. The next article reviews the wide differential diagnosis of gut eosinophilia, and provides information on entities that clinicians need to at least consider in the differential diagnosis when they encounter any patient with gut eosinophilia. The final article in the eosinophil section concerns a potentially brand new eosinophilic gut disease. Functional gastrointestinal disorders are remarkably common conditions that are not considered to have a pathologic correlate. Functional dyspepsia is one of these disorders affecting up to 1 in 10 people, but notably, recent observations have linked the condition to duodenal
0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.03.003
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
xii
PREFACE
eosinophilia in both adults and children. While still somewhat controversial, this area is reviewed here because, if the hypothesis is correct, the work should lead to novel diagnostic and treatment strategies for patients with functional dyspepsia. Autoimmune diseases of the gastrointestinal tract may often be missed in practice unless actively sought, in part because many of these diseases have only recently been identified. The recognition that enteric autoantibodies can induce motility disorders of the gut has been one of the areas to advance, and not all of these patients have a malignancy. When to screen patients and what antibodies to consider are the topics reviewed in the first article in this section. Celiac disease is well known to be associated with extraintestinal autoimmune diseases (as well as small intestinal eosinophilia), and the mechanisms are starting to be elucidated. For these reasons, an article on the topic is included in this issue. The relationship between autoantibodies and inflammatory bowel disease has received widespread attention in recent years, and commercially available panels of autoantibodies are now available. The value of this tool needs to be critically considered, and this is discussed in the next article in the issue. Autoimmune pancreatitis is a comparatively newly recognized entity with clear-cut clinical features and a characteristic responsiveness to steroid therapy that sets it apart from other pancreatic diseases, and this topic is presented. Finally, autoimmune hepatitis and a newer entity, autoimmune cholangitis, are each covered in the autoimmune diseases section. This issue represents a comprehensive synthesis of a disparate body of important literature. I am grateful to all of the internationally recognized experts in the field who agreed to contribute; I applaud their clear and careful synthesis of the published material and their diligence in preparing these articles. The contributions all are of the highest quality, and I commend them to you whether you are a practicing gastroenterologist or an expert in the field. I hope you will enjoy reading these articles as much as I have enjoyed writing or editing them. Nicholas J. Talley, MD, PhD, FRACP, FRCP, FACP Division of Gastroenterology & Hepatology Mayo Clinic College of Medicine 4500 San Pablo Road Jacksonville, FL 32224, USA E-mail address:
[email protected]
Gastroenterol Clin N Am 37 (2008) 307–332
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Gut Eosinophilia in Food Allergy and Systemic and Autoimmune Diseases Nicholas J. Talley, MD, PhD, FRACP, FRCP, FACP Division of Gastroenterology and Hepatology, Davis Building E-6, Mayo Clinic Campus, Mayo Clinic, 4500 San Pablo Road, Jacksonville, FL 32224, USA
E
osinophilic gastroenteritis is a rare disease characterized by striking tissue eosinophilia in any layer of the gut wall; however, many diseases can cause increased gut eosinophilia. Allergic reactions to food are an important cause of gut eosinophilia. Not all adverse reactions to food are IgE mediated, and most cases of IgE-mediated food allergy do not have eosinophilic gastroenteritis. Parasitic, bacterial, and viral pathogens as well as certain systemic diseases such as vasculitis can cause gut eosinophilia. These heterogeneous conditions are reviewed in this article. The following search strategy was applied. Ovid MEDLINE and Pub Med (no time limit) on-line databases were searched to identify the published literature in the English language or translations when available. An initial search was performed using the terms eosinophilic gastroenteritis, gastrointestinal eosinophilia, and gut eosinophilia. These terms were then coupled with other secondary search terms, that is, food allergy, cow milk protein allergy, drugs/medications, parasites, bacteria, virus, transplant, graft versus host disease (GVHD), polyps, hypereosinophilic syndrome, and specific types of vasculitis. Relevant published articles and, when important, abstracts were included and referenced herein. Due to the paucity of literature, case reports were also carefully considered and included if deemed appropriate. All levels of evidence were considered, but more emphasis was placed on randomized trials in this review. When evidence is lacking, this is stated explicitly. FOOD ALLERGY The presence of increased eosinophils in the gastrointestinal tract may occur as an allergic reaction to certain foods. Food allergy and food intolerance are the two major groups of adverse reactions. Food intolerances (non-allergic food hypersensitivities) are adverse responses caused by some unique physiologic characteristic of the host, such as metabolic disorders (eg, enzymatic or transport
E-mail address:
[email protected] 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.008
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
308
TALLEY
deficiencies), whereas food hypersensitivity/allergy is an adverse immunologic reaction that might be due to IgE- or non–IgE-mediated immune mechanisms. Epidemiology of Food Allergy Unfortunately, data on the prevalence of true allergic reactions to food are lacking, but probably no more than 1% to 2% of the population is affected. Food intolerance and several other disorders result in symptoms mimicking food allergy, which makes it particularly difficult to calculate the actual number of persons with true food allergies. Both food allergy and intolerance are reported more commonly in children than adults. The frequency of self-reported symptoms from food intolerance is estimated to range from 4.9% [1] to 20% [2], the vast majority of which are non-immunogenic, with lactose intolerance being the most frequently reported [3]. Peanut or tree nut and seafood and egg allergies are the best documented IgE-mediated food allergies in the United States. Studies have estimated that 1.3% of Americans are allergic to peanuts or tree nuts [4], and seafood allergies affect 3.5% to 4% of the population [5]. Egg is another common allergen, especially in childhood, with an estimated cumulative prevalence of 2% to 6% by age 2 to 5 years [6]. Altman and Chiaramonte [7] reported on average 1.17 household members with food allergy in their survey of 5000 demographically representative American households. Milk, various fruits and vegetables, seafood, chocolate, eggs, and peanuts were the individual foods most frequently reported in this survey. Pathologic Mechanisms The gut-associated lymphoid system includes organized lymphoid tissue such as mesenteric lymph nodes, Peyer’s patches, and lymphoid follicles within the lamina propria. This system comes into direct contact with food antigens, which may explain why as many as 50% of food allergic disorders present with gastrointestinal manifestations. Even though 2% of ingested food antigens are absorbed, tolerance develops through a variety of mechanisms (eg, T-cell anergy or induction of regulatory T cells) that protect against the development of allergy. In a sensitized host, a food antigen can bind to IgE, activating and releasing several potent mediators and cytokines and leading to accumulation of interleukin-5 (IL-5), eosinophils, basophils, IgE, and mast cells in intestinal biopsy specimens of patients with food allergy [8]. On the other hand, food protein–induced proctocolitis presents in the first few months of life. The eosinophils may have a critical role independent of the presence of IgE or immediate hypersensitivity. It has been suggested that maternal dietary protein is complexed in some fashion with breast milk immunoglobulin (IgA) or another immunologic component, initiating an inflammatory response [9]. Risk Factors Many food proteins have been identified as potential allergens (Table 1). Certain allergens, such as peanuts, have been suggested to be associated with an increased risk of sensitization to other foods. Schafer and colleagues [2] in their study reported that more than 70% individuals were sensitized to more than
GUT EOSINOPHILIA
309
Table 1 Food proteins that have been identified as potential allergens Age group
Most common foods implicated in food allergy
Infants Children Adults
Cow’s milk, soya proteins Cow’s milk, eggs, peanuts, soy, wheat, tree nuts, fish and shellfish Peanuts, tree nuts, fish and shellfish, chocolate, eggs
one allergen, with the highest concomitant sensitization being to hazelnuts, peanuts, and celery. Food allergy is the result of exposure of a susceptible individual to an allergen, but the influence of various genetic and environmental factors may increase this risk. Over the years multiple studies have tried to identify a possible relationship between certain genes and manifestations of food allergy. One gene of particular interest is a polymorphism in the promoter region for CD14, which is the receptor for lipopolysaccharide. Woo and colleagues [10] studied 175 asthmatic and 77 food-allergic patients of varying age and suggested a significant association between the -159T allele, one of the alleles in that region, and the presence of food allergy and non-atopic asthma. Similarly, in another study by Hand and colleagues [11], the relationship between HLAs and nut allergy was investigated in 84 individuals with a clinical history. The frequency of HLA-B*07 (28.57%) and DRB1*11 (15.48%) was increased in the nut-allergic patients when compared with atopic controls (12.20% and 3.66%, respectively). Sicherer and colleagues [12] studied 70 twin pairs with at least one member with peanut allergy. Among the monozygotic pairs (n ¼ 14), nine were concordant for peanut allergy (pair-wise concordance, 64.3%); and among dizygotic pairs (n ¼ 44), three were concordant for peanut allergy (pair-wise concordance, 6.8%; P < .0001). The heritability of peanut allergy was estimated at 81.6% (95% CI, 41.6–99.7). The significantly higher concordance rate of peanut allergy among monozygotic twins suggested a strong genetic influence on peanut allergy. Although statistically significant, these findings appear to be of limited clinical significance, and, to date, no gene has been linked conclusively to food allergy [13]. The search for genes strongly associated with food allergy continues. Maternal exposure to certain foods during pregnancy and lactation is a risk factor in the development of food allergy in children. Frank and colleagues [14] reviewed 25 peanut-sensitized and 18 control children (allergy to milk but not to peanuts) and concluded that mothers who consumed peanuts more than once a week were more likely to have a peanut-allergic child than mothers who consumed peanuts less than once a week. A connection between food allergy and natural rubber latex allergy has been recognized in the literature [15–17]. Kim and Hussain [18] reported 49 potential allergic reactions to foods in 29 of 115 (21.1%) patients with a confirmed type I latex allergy. Foods responsible for these reactions included bananas (18%), avocado (16%), shellfish (12%), fish (8%), kiwi (12%), and tomato (6%).
310
TALLEY
Clinical Presentation As mentioned earlier, food allergy can broadly be classified into IgE- and non– IgE-mediated processes, and the symptoms range from minor impairments to life-threatening shock reactions. Each subclass is discussed individually in the following sections. IgE-mediated reactions In food-allergic individuals, IgE is produced against naturally occurring food components, primarily glycoprotein, that usually retain their allergenicity after heating or proteolysis. Adults tend to be allergic to fish, crustaceans, peanuts, and tree nuts, whereas children tend to be allergic to cow’s milk, egg white, wheat, and soy. These reactions are rapid in onset and comprise both intestinal and extraintestinal manifestations. Gastrointestinal anaphylaxis and oral allergy syndrome are the most important symptom complexes. Similar mechanisms have been suggested in 10% to 15% cases of infantile colic [19]. Gastrointestinal anaphylaxis. Gastrointestinal anaphylaxis is a common form of IgE-mediated food allergy. Yocum and colleagues [20] reviewed the medical records of Olmsted County inhabitants and reported an annual incidence of food anaphylaxis of 7.6 cases per 100,000 person-years and a food anaphylaxis occurrence rate of 10.8 cases per 100,000 person-years. Its pathophysiology is well understood. The antigen cross-links with IgE antibody and activates immunoglobulin receptors on inflammatory cells. This activation causes a release of multiple inflammatory mediators that, in turn, increase vascular permeability and cause smooth muscle contractions [21]. Gastrointestinal symptoms of nausea, a painful abdomen, edema of the lips and tongue, nausea, vomiting and diarrhea usually develop within minutes of ingesting the offending agent; however, the onset of diarrhea and colic may be delayed for up to 6 hours. Other organ systems are also involved, including the respiratory (shortness of breath, wheezing, laryngeal edema), skin (hives, pruritus, angioedema), and cardiovascular system (hypotension, tachycardia, and circulatory collapse). A biphasic anaphylactic reaction has been described in the literature with a recurrence of symptoms after a period of recovery. Its incidence is highly variable, ranging from a low of 1% to a high of 20% of episodes [22]. The nature and severity of the symptoms, the time of onset after administration of antigen, and the history of a previous biphasic reaction have all been described as risk factors in various studies [23,24]. The severity of the late-phase reaction is highly variable, and events have ranged from mild to severe with rare fatalities. The diagnosis is mostly clinical, and a careful history of ingestion of any particular food is of great importance. Recurrences of symptoms after food re-challenge and evidence of food-specific IgE antibodies (high serum IgE levels and positive skin prick test) may increase the diagnostic yield [25], although the former is usually contraindicated in patients with a clear-cut history of anaphylaxis. Anaphylaxis is a medical emergency. The treatment protocol should include the administration of epinephrine, intravenous fluids, antihistamines,
GUT EOSINOPHILIA
311
bronchodilators, or corticosteroids [26]. A 24-hour period of posttreatment observation is ideal for biphasic reactions. Avoidance of offending agent is the best preventative strategy; however, accidental or contaminant exposure to the allergen is possible, and epinephrine self-administration injections should be prescribed to all patients at risk for anaphylaxis. Oral allergy syndrome. Oral food allergy syndrome, also known as pollen food allergy syndrome, is a clinical term used for a form of contact hypersensitivity that occurs when a susceptible individual ingests a food that has cross-reactive antigens to pollens that they are sensitive to. Schafer and colleagues [2] surveyed 1537 patients with questionnaires and reported 659 (42.9%) patients with oral symptoms. The frequency varied with different types of food consumed, with fruits (72.3%) and nuts (68%) being the most common culprits. An association between pollen and oral allergy syndrome has been well recognized [27,28], and some patients have oral allergy syndrome symptoms only during the pollen season when anti–birch-specific IgE titers rise [29]. Symptoms usually occur within minutes after the offending agent comes in contact with oral mucosa. The most common symptoms are mild itching and swelling of the lips and mouth. Severe angioedema of the pharyngeal mucosa causing life-threatening emergencies has been reported [30]. Diagnosis is based on a careful history that should include details about seasonal patterns and pollen and specific food exposures. The development of symptoms with raw fruits and vegetables and not to cooked food strongly points toward the diagnosis. Physical examination is once again invaluable and should include a careful inspection of the oral cavity and pharynx for rash, urticaria, or angioedema. A positive skin prick test (prick-to-prick skin test) is the preferred method of testing in most cases. Prick-to-prick testing is performed by inserting the needle into the fruit, withdrawing it, and then immediately pricking the patient’s cleaned skin. Anhoej and colleagues [31] compared the skin prick test and basophil histamine release tests of 36 patients with grass or birch allergies and 17 control subjects. All of the subjects were skin prick tested and had basophil histamine release tests done with fresh fruits and various extracts of hazelnut, apple, and melon. The diagnosis of oral allergy syndrome was confirmed by oral challenges. In addition, histamine release to recombinant Bet v 1 and Bet v 2 and recombinant Phl p 1, Phl p 2, and Phl p 5 was performed. The skin prick test showed an almost optimal diagnostic value, with a satisfactory sensitivity (>89%) and excellent negative predictive value with fresh fruits. Oral challenge did not result in severe systemic reactions, and no systemic reactions were observed with skin prick tests with fresh fruits. When the skin prick test with the culprit food item cannot be performed, histamine release is a diagnostic alternative, although it is still limited to tertiary research centers. There is no recommended treatment protocol for oral allergy syndrome. Generally, avoidance of the offending food is recommended. Patient education about early recognition of the symptoms in the event of accidental ingestion
312
TALLEY
and epinephrine self-administration training are advised in patients who have rare severe systemic reactions. Therapies of unproven benefit for pollen food allergy syndrome include prophylactic administration of H1 antihistamines, immunotherapy for pollinosis, and anti-IgE therapy [32–34]. Infantile colic. Infantile colic is an ill-defined condition of infancy characterized by paroxysmal fussing and excessive crying. The most commonly used definition of colic was coined in 1954 by Wessel and colleagues [35] who described it by using a ‘‘rule of three’’: crying for more than 3 hours per day, for more than 3 days per week, and for more than 3 weeks in an infant that is well fed and otherwise healthy. During this time, infants cry an average of 2.2 hours per day, peaking at 6 weeks of age and gradually decreasing [36]. Various studies have reported the incidence of infantile colic to be 5% to 25% of infants [37,38]. Gastrointestinal, psychosocial, and neurodevelopmental disorders have been suggested as the cause of colic [39]. IgE-mediated hypersensitivity has been proposed as a pathogenic factor, possibly in 10% to 15% of colicky infants [19]. The child is usually brought to the physician by concerned parents seeking advice, and a careful history and physical examination are imperative to determine whether there is an organic cause for the crying or to relieve parental fears and allow for a diagnosis of colic. Documentation of the frequency and quantity of spitting up is necessary to satisfy the diagnostic criteria of infantile colic and to rule out gastroesophageal reflux or pyloric stenosis [40]. Colic usually resolves in 60% of infants by 3 months of age and in 90% of infants by 4 months of age [41]. Multiple strategies have been employed in management of the infant with prolonged or excessive crying [42], including hypoallergenic (protein hydrolysate) formula (for formula fed infants), a lowallergen maternal diet (for breastfeeding mothers), a dicyclomine trial, and reduced environmental stress on the child. Five trials have studied the effect of eliminating cow’s milk protein on excessive crying [43–47]. Two reports studying hypoallergenic (protein hydrolysate) formula in nearly 130 infants found an effect size of 0.22 (95% CI, 0.10–0.34). A comparison of breast milk with standard cow’s milk in infants who were already weaned showed no significant differences. The anticholinergic drugs dicyclomine and dicycloverine showed a clear benefit in the treatment of excessive crying. The pooled results showed a clinically significant improvement (effect size of 0.46 [95%CI, 0.33–0.60]) [42]. Dicyclomine is effective in treating infantile colic, but 5% of the treated infants had side effects. The manufacturer reports breathing difficulties, seizures, syncope, asphyxia, muscular hypotonia, and coma as side effects [48]. Taubman [49] found that increasing parental responsiveness decreased crying significantly from 2.09 1.07 h/d to 1.19 0.60 h/d. Non–IgE-mediated reactions Non–IgE-mediated hypersensitivities are believed to be the result of abnormal antigen processing or cell-mediated mechanisms and include dietary protein-induced eosinophilic proctocolitis, dietary protein-induced enterocolitis,
GUT EOSINOPHILIA
313
dietary protein-induced enteropathy, and celiac disease and gliadin-sensitive enteropathy. Celiac disease and gliadin-sensitive enteropathy leading to malabsorption with associated dermatitis herpetiformis, a chronic blistering skin disorder, is a classic gastrointestinal food allergy that is not IgE mediated and believed to be the result of various cell-mediated mechanisms. Celiac disease is reviewed elsewhere in this issue. Dietary (food) protein-induced enteropathy. Dietary protein-induced enteropathy is a symptom complex characterized by malabsorption, failure to thrive, diarrhea, emesis, and hypoproteinemia [50]. Although it is predominantly a disease of infancy, residual symptoms may persist to school age [51]. It is usually related to an immunologic reaction to cow’s milk protein but has also been associated with sensitivities to soy, egg, wheat, rice, chicken, and fish. Damage to the intestinal mucosa by allergic inflammation leads to villous architectural distortion and mild eosinophilic infiltration causing malabsorption and osmotic diarrhea [50]. Beyer and colleagues [52] demonstrated milk-specific lymphocytes in duodenal biopsies of 60% of infants with milk protein–induced enteropathy. Clinical and histologic features share similarities with celiac disease, which should be considered as the main differential diagnosis; however, unlike in celiac disease, loss of protein sensitivity occurs by the age of 6 to 18 months. Colitis features such as mucus and gross or microscopic hematochezia are usually absent [53]. For an accurate clinical diagnosis, challenge with the offending food after a demonstrated response to cow’s milk elimination is critical. When available, serial small intestinal bowel biopsies related to elimination and challenge are also important [50]. Dietary protein-induced enterocolitis. Food protein-induced enterocolitis syndrome (FPIES) is a gastrointestinal allergic inflammation of the small intestine and colon in young infants aged 1 week to 3 months. Although cow’s milk and soy are considered the main causative allergens, other solid foods, including vegetables, cereals, fish, and poultry meats, have been reported as allergens [54]. FPIES usually presents with profuse diarrhea, vomiting, dehydration, acidosis, transient methemoglobinemia, and failure to thrive. Anemia, occult positive stools, and gastrointestinal hemorrhage may be the presenting symptoms [15]. The pathophysiology of this disorder remains unclear. Chung and colleagues [55] performed immunohistochemical staining for TGF-b1, type 1 and 2 TGF-b receptors, and TNF-a on duodenal biopsy specimens of 28 infants diagnosed with FPIES by means of clinical criteria and challenge test results. Duodenal villous atrophy was associated with increased tissue staining for TNF-a, whereas expression of duodenal TGF-b was reduced in patients with FPIES. Stools contain occult blood and leukocytes, predominantly neutrophils and eosinophils. Jejunal biopsies classically reveal flattened villi, edema, and increased numbers of lymphocytes, eosinophils, and mast cell infiltration [53].
314
TALLEY
Complete resolution of symptoms within 72 hours with elimination of the causative allergen and recurrence with oral challenge is diagnostic [56]. Dietary protein-induced eosinophilic proctocolitis. Food-induced eosinophilic proctocolitis appears in the first 2 months of life with blood-tinged stools. Initially reported in the literature as a disorder of breast-fed infants [57], it has increasingly been reported in infants receiving cow’s milk, soy, and hydrolysate formula [58]. It is unclear why the allergic inflammation is limited only to the lower colon, and the role of eosinophils in this process has not been clearly defined. The infants typically have blood-streaked, normal-to-soft stools at 2 to 8 weeks of age. The age of the infant can range from 2 days to 3 months [9]. Weight gain and growth are normal. Endoscopic biopsies reveal increased eosinophils and T-cell infiltration of the colonic mucosa [59]. Stool cultures are negative for bacteria or Clostridium difficile by definition. Complete resolution of symptoms occurs with 3 to 4 days of elimination of the offending protein from the diet [9]. Bleeding and fecal leukocytes may clear within days; the endoscopic and histologic healing can take several weeks. Rechallenge with the offending protein usually provokes a recurrence of bleeding; however, the infant can tolerate an unrestricted diet after 9 months of age. Diagnostic Evaluation Laboratory tests Food-specific serum IgE antibody testing is now commercially available. Levels above the clinical cut-off indicate a more than 95% likelihood of experiencing an allergic reaction and can be used to compliment skin prick testing [60]. The positive predictive values for the three major food allergens (ie, egg, milk, and peanut) are 95% or greater [60]; however, a definitive diagnosis of gastrointestinal food allergy still relies on standard double-blind, placebo-controlled food challenges [61]. Skin puncture test Skin prick or puncture testing is performed by introducing the food allergen to cutaneous mast cells. If food-specific IgE antibody is present on the surface of the patient’s mast cells, the cells will degranulate, releasing histamine and other mediators that cause localized cutaneous swelling (ie, a wheal). Vasodilatation also develops as a result of an axonal reflex (ie, a flare). Commercial extracts from foods with stable proteins, such as peanuts, milk, egg, tree nuts, fish, and shellfish, are generally better in terms of specific IgE antibody production. Extracts from fruits, vegetables, and other foods containing labile proteins may be altered during processing and can produce variable results [62]. The general sensitivity and specificity of skin prick testing for the diagnosis of food allergy are often estimated to be greater than 90% and approximately 50%, respectively [63]. This test provides rapid detection of sensitization, and negative responses essentially confirm the absence of IgE-mediated allergy (negative predictive value over 95%); however, a positive test response does
GUT EOSINOPHILIA
315
not indicate that the particular food is involved [64]. The utility of the skin prick test can be greatly enhanced if it is used under correct clinical settings (ie, with a high pretest probability), and it should not be used as a screening test due to low specificity. Case series have shown that infants with milk or soy enterocolitis have negative skin prick tests, serum food-specific IgE tests, or both [65]. Endoscopic biopsies Upper or lower gastrointestinal tract biopsies are not diagnostic. The small intestinal mucosa may show chronic villous atrophy similar to celiac disease; however, eosinophilic infiltrates in the rectal biopsy strongly suggest dietary protein-induced eosinophilic proctocolitis. Bischoff and colleagues [66] developed a new approach to improve the diagnostic yield of endoscopy called colonoscopic allergen provocation (COLAP). During this test, cecal mucosa of 70 adult patients with abdominal symptoms suspected to be related to food allergy and of five healthy volunteers were challenged endoscopically with three food antigen extracts, a buffer control, and a positive control (histamine). The mucosal weal and flare reaction was registered semiquantitatively 20 minutes after challenge, and tissue biopsy specimens were examined for mast cell and eosinophilic activation. The COLAP test was positive to at least one food antigen in 54 of 70 patients (77%), whereas no reaction in response to antigen was found in healthy volunteers. Management Diet Avoiding the identified food allergen is the cornerstone of treatment in the management of food allergy and may improve the likelihood that tolerance will develop with time, especially to cow’s milk, egg, and soy [56]. This avoidance is particularly important in peanut allergy, in which even tiny traces of allergens can initiate anaphylaxis. Nevertheless, the practicality of such elimination diets is limited. They necessitate well-trained counselors, time, and a great deal of motivation on the part of the affected persons. The response to an elimination diet is variable in adults, and relapses are common, which are more likely due to an accidental ingestion of a previously identified but hidden common food allergen (eg, milk, egg) rather than a reaction to a new food. An elimination diet can lead to malnutrition, especially in children if large numbers of food items are involved for prolonged periods of time [67]. Several studies have shown that milk and peanut protein are secreted into the breast milk of lactating women after maternal ingestion of these foods [68,69]. On the basis of this observation, the American Academy of Pediatrics, the European Society for Pediatric Allergology and Clinical Immunology, and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition all recommend exclusion of identified causal protein from the maternal diet if the infant is affected [70,71]. Newly implemented food labeling laws clearly indicate major food allergens, and patients should learn to understand these labels to avoid accidental
316
TALLEY
exposure to allergens and foods with cross-reacting antigens. The Food Allergy Network is a useful source of information on food allergies and anaphylaxis (http://www.foodallergy.org). If the patient fails to respond to diet elimination or if all of the provocative foods are not clearly identifiable, pharmacologic therapy may be required. Sodium cromoglycate This drug prevents the release of toxic mast cell mediators. It also alters the mucosal permeability of the gastrointestinal tract by reducing direct absorption of antigens [72]. In a clinical trial conducted in the early 1970s, Freier and Berger [73] used oral sodium cromoglycate in a dose of 50 mg every 6 hours in four infants with milk protein intolerance. All four were re-challenged with milk and remained asymptomatic. Symptoms recurred in one child after cessation of treatment. Subsequently, Kocoshis and Gryboski [74] in a double-blind, clinical, placebo-controlled trial involving 14 children with milk allergy and concomitant allergies to one or more foods evaluated the efficacy of orally given cromolyn sodium versus placebo. Treatment was begun while the children were receiving elimination diets; they were then challenged with specific antigen after 48 hours of drug administration. Crossover took place if the initial agent was ineffective. Cromolyn was effective in 11 of 13 children compared with three of nine in the placebo group. In a case-control study, Patriarca and colleagues [75] used oral desensitization treatment over a period of 6 to 8 months in combination with sodium cromoglycate (250–500 mg before each meal) in 47 patients with a clinical or biochemical diagnosis of allergies to various food products (positive prickby-prick test and high IgE levels plus urticaria/angioedema, erythema with pruritus, rhinitis, rhinorrhea, vomiting or diarrhea with abdominal pain or general malaise, collapse or loss of consciousness). Desensitization was successful (based on a negative skin prick test, decrease in specific IgE, and clinical improvement) in 45 of the 54 treatments (83%). Although oral desensitization treatment in combination with oral cromoglycate may be an alternate approach in a patient with food allergy when avoidance has failed, this is not the standard of care. Molkhou and Dupont [76] treated 16 children and four adults suffering from food allergy or food intolerance with a 1-mg daily dose of ketotifen, another mast cell stabilizer. Gastrointestinal permeability was measured in 5 of 20 patients using mannitol and lactulose. In five individuals, food ingestion resulted in a significant rise of the lactulose to mannitol urinary ratio, and the administration of ketotifen resulted in a normalization of the mannitol to lactulose urinary ratio. Patients with urticaria with or without angioedema, gastrointestinal symptoms, asthma, and oropharynx pruritus with edema of the lips appeared to be protected from food allergy in this study. Suzuki and colleagues [77] successfully treated a 10-year-old boy with ketotifen and oral disodium cromoglycate who was suffering from protein-losing enteropathy, eosinophilic gastroenteritis, and increased allergen-specific IgE antibodies to some food items.
GUT EOSINOPHILIA
317
Corticosteroids Systemic corticosteroids remain the first-line therapy for management of eosinophilic gastroenteritis in acutely symptomatic patients or those who fail or cannot tolerate dietary restrictions because they provide rapid and effective relief of symptoms within a few days to weeks. Most experts recommend doses similar to those used in inflammatory bowel disease (1–2 mg/kg/d) orally for 8 weeks and tapered over 6 to 8 weeks to induce remission [78,79]. Although topical corticosteroids have been proven effective in eosinophilic esophagitis [80], there is no evidence in the literature to establish their usefulness in gastrointestinal food allergy. Leukotriene receptor antagonists Due to various side effects associated with the prolonged use of corticosteroid therapy, multiple steroid-sparing modalities have been tried with variable results. The cysteinyl leukotrienes (LTC4, LTD4, LTE4) are potent inflammatory eicosanoids released from various cells, including mast cells and eosinophils, that bind to cysteinyl leukotriene receptors (CysLT) found in the human airway and cause a number of airway actions, including bronchoconstriction, mucous secretion, vascular permeability, and eosinophil recruitment. Montelukast binds selectively, competitively, and with high affinity to these receptors (CysLT1), blocking the eosinophilic chemotactic and other proinflammatory actions. There have been anecdotal case reports [81,82] of successful treatment of food allergy with montelukast. Vanderhoof and colleagues [83] observed marked improvement in the symptoms of eight children ranging in age from 2 to 17 years with gastrointestinal eosinophilia unresponsive to standard therapies, including diet and cow’s milk protein restriction. The optimal dose is not documented in the literature. Despite its use for weeks to months with gradual tapering of the steroids, tissue eosinophilia may persist [84]. Probiotics Probiotics are microbial cell preparations or components of microbial cells that have a beneficial effect on the health and well-being of the host [85]. Most of these agents belong to the genera lactobacilli or bifidobacteria, appear safe for human use, and can bind to and colonize human intestine [86,87]. The concept of replenishing gut flora is not new. The Russian scientist Eli Metchnikoff at the beginning of the twentieth century first suggested that it would be possible to modify the gut flora and to replace harmful microbes by useful microbes [88,89]. Loskutova [90] in 1985 reported the improvement of food allergy symptoms with the administration of a mixture containing propionibacteria and Lactobacillus acidophilus. Multiple studies since then have yielded variable results. Wheeler and colleagues [91] in a randomized crossover study compared the immune parameters of participants who received 16 oz of yogurt fermented with Lactobacillus bulgaricus and Streptococcus thermophilus versus 16 oz of milk per day and reported no significant improvements in any parameter. In contrast, Majamaa
318
TALLEY
and Isolauri [92] in a prospective trial showed a clinical and biologic response of atopic eczema and cow’s milk protein allergy by using cow’s milk elimination without (n ¼ 14) and with (n ¼ 13) the addition of lactobacillus GG (5 108 colony-forming units/g formula) in an extensively hydrolyzed whey formula. A significant drop was noted in stool inflammatory markers in infants treated with the extensively hydrolyzed whey formula fortified with lactobacillus GG when compared with the extensively hydrolyzed whey formula alone. Currently, clinical data supporting the efficacy of probiotic therapy in the management of food allergy are limited, and rigorous scientific effort is required to elucidate the characteristics of distinct probiotic strains and to determine their safety and efficacy in food allergy.
INFECTIONS Parasites Due to improved hygiene conditions in developed countries, human worm infestations have almost been eradicated, except in travelers, exotic food adventurers, and immigrants from underdeveloped countries. Although eosinophils kill and phagocytose bacteria, they are unable to clear a bacterial infection in the absence of neutrophils. Their primary function is considered to be defensive against organisms that are too large to be phagocytosed, particularly parasitic helminthes [93]. Eosinophilia induced by helminthic infection is mainly dependent on IL-5 generated by TH2 lymphocytes. Carbohydrate ligands expressed on the parasite surface attract and activate eosinophils that, in turn, release their toxic granules on the parasite surface. The killed parasite is eventually phagocytosed by macrophages. Studies showing how eosinophils protect against schistosomiasis support this concept [94]. Helminthic infections characteristically are associated with peripheral eosinophilia reflecting an immunologic response to tissue migration. Gastrointestinal eosinophilia has been associated with Ancylostoma, Strongyloides (Figs. 1 and 2), Enterobius vermicularis, Eustoma rotundatum, Trichuris trichiura, Gnathostoma spinigerum, Anisakis simplex, Trichinella spiralis, Ascaris suum, Schistosoma, Necator americanus, Isospora belli, and Toxocara canis (Table 2) [95–117]. Treatment with anthelmentic drugs results in resolution of symptoms and tissue eosinophilia in most cases. Ancylostoma caninum, a nematode belonging to the Ascaridia superfamily and also known as dog hookworm, can cause prolonged and persistent gastrointestinal eosinophilia that can be confused with eosinophilic gastroenteritis. A caninum has been increasingly reported to cause eosinophilic enteritis in Australia [95] and the United States [96]. Enteric infection with A caninum is a leading cause of human eosinophilic enteritis in northeastern Australia [118]. Human infection depends on the worm prevalence in dogs, suitable ecology, and appropriate behavior by both canines and humans. Larvae of A caninum classically enter a human host by skin penetration, causing a classic creeping eruption, but infection by oral ingestion is also possible. These larvae probably remain dormant in skeletal muscles and create no symptoms;
GUT EOSINOPHILIA
319
Fig. 1. Duodenal biopsy showing Strongyloides. Abundant eosinophils and chronic inflammatory cells are also visible (hematoxylin-eosin stain, original magnification 40).
however, in some individuals, larvae may reach the gut and mature into adult worms [97]. Clinical features are nonspecific, and a history of potential exposure is rare in those with gut eosinophilia from this worm. Abdominal pain, diarrhea, anemia, and intestinal obstruction are common presenting symptoms. Gastrointestinal bleeding is uncommon. Blood eosinophilia and high serum IgE levels are present but nonspecific. Indirect ELISA and Western blots can detect antibodies to antigens of adult A caninum and are helpful in the right clinical setting. Colonoscopy often reveals focal mucosal inflammation and aphthous ulcers of the terminal ileum, cecum, and colon with marked eosinophilic infiltration on biopsy; finding the worm itself is rare [98]. Clinical symptoms and peripheral eosinophilia rapidly improve with mebendazole, albendazole, or ivermectin [99]. In
Fig. 2. Strongyloides stercoralis within the glandular cells of duodenal mucosa. The underlying lamina propria showed abundant eosinophils (hematoxylin-eosin, original magnification 40).
TALLEY
320
Table 2 Infections characteristically associated with gastrointestinal eosinophilia Site of eosinophilia
Parasite
Reference
Esophagitis Gastroenteritis
Gnathostoma spinigerum Anisakis simplex Enterobius vermicularis Trichuris trichiura Toxocara canis Anisakis simplex Trichinella spiralis (rat) Schistosoma mansoni (rat) Necator americanus Sarcocystis hominis Isospora belli Strongyloides stercoralis Enterobius vermicularis Eustoma rotundatum Anisakis simplex Enterobius vermicularis Ascaris suum Trichuris trichiura Strongyloides stercoralis Ancylostoma duodenale
[95] [96] [97] [98] [99,100] [101] [102] [103] [104,105] [106] [107] [108] [109] [110] [111,112] [113] [114] [115] [116] [117]
Enteritis
Terminal ileitis
Colitis/proctitis
dog owners with unexplained eosinophilic ileocolitis, an empiric trial of antihelminthic therapy is reasonable. Helicobacter pylori In healthy individuals, eosinophils are normally detectable in stomach and intestine but not in the esophagus. The normal cut-offs by quantification microscopy are poorly defined, but less than 5 eosinophils per high-power field is probably within normal limits in health. There have been a few case reports of the coexistence of H pylori and eosinophilic infiltration of gastric mucosa [119,120]. McGovern and colleagues [100] demonstrated increased eosinophilic infiltration and degranulation in the gastric mucosa in the presence of H pylori. Although there is no conclusive evidence of a causal relationship, screening and eradication of H pylori have resulted in complete resolution of eosinophilic gastroenteritis in a few cases [119,120]. Cytomegalovirus Whittington and Whittington [101] proposed cytomegalovirus (CMV) as the possible etiology in 2 of 17 children evaluated for eosinophilic gastroenteritis. CMV antigenemia seroconversion has been reported in children post liver transplant in whom gastrointestinal eosinophilia developed [102]. Takeyama and colleagues [103] reported CMV infection and eosinophilic gastroenteritis in an immunocompetent child who experienced transient protein losing enteropathy and peripheral eosinophilia. Anti-CMV IgG and IgM
GUT EOSINOPHILIA
321
antibodies were positive. Gastric, duodenal, and colonic biopsies revealed eosinophilic infiltration of the lamina propria and cytomegalic cells (large cells containing eosinophilic intranuclear inclusions and, frequently, basophilic intracytoplasmic inclusions) that were immunohistochemically positive for CMV. The patient was treated with transfusions of albumin, oral famotidine, oxatomide (an H1 histamine blocker), and cromoglycate (a mast cell stabilizer) but not corticosteroids and eventually recovered. Repeat biopsies were negative. The presence of intranuclear and intracytoplasmic inclusions in a biopsy specimen should raise suspicion, and appropriate testing should be conducted to rule out this rare presentation. MEDICATIONS The overall incidence of peripheral eosinophilia from a drug is probably less than 0.1%. O’Donovan [104] reported the first case of gold-induced colitis, and since then, over 30 well-documented cases have been reported in the literature. The onset of enterocolitis was delayed, and the clinical findings included abdominal pain, nausea, vomiting, watery diarrhea, or bloody stools. Colonoscopy may reveal friable ulcerated mucosa. Pathologic examination of colonic biopsies may show mucosal hemorrhage and ulcerations with inflammatory and, in some cases, a predominantly eosinophilic infiltration of the gastrointestinal mucosa [105,106]. Despite causing significant morbidity due to prolonged diarrhea and severe protein losing enteropathy, no cases of mortality has been reported in the literature since 1980 [106]. Several other drugs (Table 3) have been reported to cause eosinophilic gastroenteritis even though they do not share any common chemical or pharmacologic properties [121–128]. Withdrawal of the drug and rapid clinical and histologic improvement is the only possible way to confirm a casual association. There have been no reported treatment trials to date in the literature. TRANSPLANTATION Liver Transplant Dhawan and colleagues [107] in 1992 first described eosinophilic gastroenteritis in two patients post liver transplant. Treatment of acute rejection with bolus Table 3 Drugs that have been reported to cause eosinophilic gastroenteritis Drugs implicated in causing gastrointestinal eosinophilia
Reference
Azathioprine Gemfibrozil Enalapril Carbamazepine Clofazimine Co-trimoxazole Non-steroidal anti-inflammatory drugs Tacrolimus
[121] [122] [123] [124] [125] [126] [127] [128]
322
TALLEY
steroids has been postulated to be the possible mechanism because it prompts an imbalance in the Th1/Th2 profile promoting eosinophilia. Since then, this entity has been increasingly reported in the literature. In a retrospective study, Lee and colleagues [108] reported eosinophilic colitis in 14 of 38 (37%) children post liver transplant. These patients had a significantly higher incidence of peripheral eosinophilia during the first 2 months, and the diagnosis was made at a mean of 7 months after transplantation. Twelve of the 14 patients were fed hypoallergenic formula, and major allergens, such as eggs, milk, soy, and seafood, were not given after weaning from milk. All of the patients showed significant improvement with food restriction treatment. Romero and colleagues [109] retrospectively analyzed the medical records of all pediatric liver transplant recipients over a period of 3 years and identified 15 of 53 (28%) patients in whom peripheral eosinophilia developed at a mean of 417 240 days after transplantation. Peripheral eosinophilia was defined as an eosinophilic count greater than 10% of the total white cell count. Six (11%) of these patients underwent upper or lower gastrointestinal endoscopies for various symptoms including nausea, vomiting, diarrhea, and blood in the stools and were diagnosed with eosinophilic gastroenteritis, which was defined as greater than 10 eosinophils per high-power field in the lamina propria. Affected patients were younger and had more rejection episodes. Moreover, three of six patients with eosinophilic gastrointestinal involvement were re-transplanted compared with 1 of 38 (2.5%) in the group that had a normal peripheral eosinophilic count, a significant difference. A younger age at transplantation, frequent rejection episodes, tacrolimusrelated immunosuppressant, peripheral eosinophilia during the first 2 months after transplant, and Epstein-Barr viremia are proposed risk factors in liver transplant patients in whom eosinophilic gastroenteritis develops. Eosinophilic gastroenteritis should be considered in the differential diagnosis of posttransplant patients who experience vomiting, diarrhea, hematochezia, and abdominal pain. It should also be considered when peripheral eosinophilia is detected or when Epstein-Barr virus seroconversion develops during the first 2 months following transplantation. There is no standard treatment protocol, and all of the patients in the previously mentioned studies were symptomatically treated. Graft-Versus-Host Disease Acute graft-versus-host disease (GVHD) is characterized by inflammation of the skin, liver, and gut occurring before day 100 after allogenic bone marrow transplantation. In contrast, chronic GVHD is a more pleotropic disease, usually occurring after day 100 after bone marrow transplantation. Daneshpouy and colleagues [110] evaluated 93 patients with gastrointestinal predominant GVHD and reported 36 patients (38.7%) with tissue eosinophilia, especially in the duodenal lamina propria. IL-5 is the main cytokine that stimulates eosinophil production in the bone marrow, and its levels were found to be high in all duodenal biopsy specimens. McNeel and colleagues [111] also
GUT EOSINOPHILIA
323
reported eosinophilia with gastrointestinal symptoms in two GVHD patients who underwent allogenic bone marrow transplantation. Interestingly, both patients responded to high-dose prednisone with complete resolution of symptoms. INFLAMMATORY FIBROID POLYPS An inflammatory fibroid polyp is a solitary non-encapsulated, usually submucosal, polypoid or sessile lesion characterized by edematous connective tissue and numerous small blood vessels with diffuse eosinophilic inflammatory infiltrate. It was first described in 1949 by Vanek [112] as a ‘‘gastric submucosal granuloma with eosinophilic infiltration.’’ In the literature, they also have been referred to as an inflammatory pseudotumor, localized eosinophilic granuloma, hemangiopericytoma, neurofibroma, or submucosal granuloma. In 1953, Helwig and Ranier [113] introduced the term inflammatory fibroid polyp, which has now been largely accepted in the literature. This polyp is seen in all age groups and has been reported in all parts of the gastrointestinal tract, including the esophagus [114,115], stomach [116,117], and small [118,129,130] and large intestines [131,132]. It may induce a variety of symptoms depending on its size, including pain, diarrhea, obstruction, bleeding, and intussusception. The size ranges from 0.7 to 20 cm, and, in larger polyps, the overlying mucosa is usually ulcerated [119]. Microscopically, the stroma is infiltrated by eosinophils to a varying degree, ranging from a few to dense aggregates. Fibroblasts in the stroma can gather in a nodular configuration with an onionskin-like arrangement and may form a concentric aggregation around small vessels and inflammatory infiltrates. Due to the presence of dense eosinophilic infiltration, Suen and Burton [120] in 1979 suggested that these polyps may be a variant of eosinophilic gastroenteritis, but they are now considered a separate and distant clinicopathologic entity of unknown etiology. Treatment is almost always resection, either endoscopic or surgical; however, gastric fibroid polyp elimination has also been reported following Helicobacter pylori eradication therapy [133]. HYPEREOSINOPHILIC SYNDROME The hypereosinophilic syndromes (HES) are disorders marked by the sustained overproduction of eosinophils characterized clinically by damage to multiple organs due to eosinophilic infiltration and mediator release [134]. Chusid and colleagues [135] suggested three defining features for clinical diagnosis: (1) blood eosinophilia of greater than 1500/lL is present for more than 6 months; (2) there are no other apparent etiologies for eosinophilia, such as parasitic infection or allergic disease; and (3) signs and symptoms of end-organ dysfunction are present. IL-5 has been implicated in the pathogenesis of HES due to its unique ability among the hematopoietins to cause selective eosinophilic granulocytosis both in vitro and in vivo [136]. Some HES patients have features of a myeloproliferative disorder with a complex chromosomal abnormality, leading to a fusion of
324
TALLEY
the Fip1-like 1 (FIP1L1) gene to the platelet-derived growth factor receptoralpha (PDGFRA) gene generated by a deletion on chromosome 4q12 and resulting in the production of the constitutively activated tyrosine kinase FIP1L1-PDGFR-alpha. A distinctive part of the syndrome is its marked predilection to damage specific organs, with cardiac disease being the most serious. The gastrointestinal tract is involved in 20% patients, including eosinophilic gastritis, enterocolitis, cholecystitis, chronic active hepatitis, and the BuddChiari syndrome from hepatic vein obstruction. Left upper quadrant pain can also occur from splenomegaly and splenic infracts [137]. HES should be suspected in patients with persistent high-grade eosinophilia in the absence of identifiable causes. Extra-intestinal manifestations rule out primary eosinophilic gastroenteritis. Involvement of medium-to-small vessels may suggest Churg-Strauss syndrome which shares many similar characteristics. Eosinophilic leukemia is important to consider in the differential diagnosis. Establishing clonality and the presence of chromosomal abnormalities consistent with eosinophilic leukemia assists in identifying this diagnosis. Asymptomatic patients do not need treatment, and close follow-up is indicated. Prednisone remains the mainstay of treatment. Many other treatment modalities have been suggested for the treatment of HES. Imatinib mesylate is now the first-line treatment for patients with HES and myeloproliferative features [138]. Alpha interferon is indicated in patients who fail steroid treatment. Bone marrow transplantation should be considered in patients who have FIP1L1/PDGFRA-associated HES that fails imatinib therapy. VASCULITIS AND CONNECTIVE TISSUE DISEASES Systemic Sclerosis Systemic sclerosis is a chronic multisystem disorder of unknown etiology characterized clinically by thickening of the skin caused by the accumulation of connective tissue and by structural and functional abnormalities of visceral organs, including the gastrointestinal tract, lungs, heart, and kidneys. Eosinophilic infiltrates of the esophagus, stomach, and small bowel lamina propria in a bandlike fashion were described by DeSchryver-Kecskemeti and Clouse [139] in six patients with scleroderma, dermatomyositis, and polymyositis. Buchman and colleagues [121] reported a similar case of a 47-year-old black man who presented with nausea, vomiting, a painful abdomen, and unspecified connective tissue disease. No peripheral eosinophilia was noted. Large numbers of mast cells and eosinophilic infiltration were also seen in the lamina propria. Systemic Lupus Erythematosus Gastrointestinal involvement is relatively common in patients with systemic lupus erythematosus, with a reported prevalence ranging from 0.9% to 28% [122], including primary vasculitis, bowel infarction, protein losing enteropathy, intestinal pseudo-obstruction, fat malabsorption, and infectious diarrhea. Barbie and colleagues [123] and Sunkureddi and colleagues [140] have reported
GUT EOSINOPHILIA
325
two well-documented cases of eosinophilic gastroenteritis in patients with diagnosed systemic lupus erythematosus. Nevertheless, at this point, it is not possible to conclude there is a casual link between these two entities. Churg-Strauss Vasculitis Churg-Strauss syndrome is a granulomatous small-vessel vasculitis characterized by allergic rhinitis, asthma, and prominent peripheral blood eosinophilia [124]. In 1951, Churg and Strauss [125] first described the syndrome in 13 patients who had asthma, eosinophilia, granulomatous inflammation, necrotizing systemic vasculitis, and necrotizing glomerulonephritis. Although the exact etiology of Churg-Strauss syndrome is unknown, the disease is most likely due to an autoimmune process [126]. The lungs and skin are the two organs most commonly involved; however, any organ system can be involved in advanced disease. The gastrointestinal manifestations of Churg-Strauss syndrome are diverse and include the features of eosinophilic gastroenteritis as well as mesenteric vasculitis. Abdominal pain, diarrhea, ulcers, and lower gastrointestinal tract bleeding may be seen in as many as 33% patients [127]. Aoyagi and colleagues [128], reviewed gastrointestinal findings in nine patients with Churg-Strauss syndrome. Ulcers, erosions, edema, or small nodules were present in the stomach (5/8), duodenum (4/8), small intestine (3/5), and colon (6/8). Endoscopic biopsy specimens demonstrated eosinophilic infiltration in the stomach (3/5), duodenum (1/4), small intestine (1/4), and colon (3/6), but there were no specific findings of vasculitis. The presence of significant gastrointestinal disease may be one of the strongest indicators of a poor prognosis [141]. OTHER UNUSUAL GASTROINTESTINAL EOSINOPHILIC DISEASES Physicians and health care professionals from all over the world continue to report unusual associations with gastrointestinal eosinophilia. Pseudotumoral Enterocolitis Male and colleagues [142] in 1983 described a patient who presented with abdominal pain, bloody diarrhea, and massive peripheral eosinophilia. After failed treatment with corticosteroids, she underwent colectomy and symptoms improved. Five years later, the disease recurred in the rectum and small bowel with massive granuloma formation. Extensive work-up failed to reveal any obvious cause, and after multiple failed corticosteroid treatments, she succumbed to this illness. At necropsy, the small bowel contained intramural polypoid tumors. Pseudotumours were also present on the mesentery. All of the tumors were heavily infiltrated by mature eosinophils. Paraneoplastic Syndrome Stefanini and colleagues [143] described a 39-year-old man with abdominal pain, diarrhea, and peripheral eosinophilia. Jejunal biopsy revealed diffuse eosinophilic infiltration of the lamina propria. His symptoms subsided with corticosteroid treatment. Two months later, symptoms recurred with
326
TALLEY
supraclavicular lymphadenopathy. Anaplastic large cell lung carcinoma was diagnosed, and the possibility of eosinophilic gastroenteritis arising from a paraneoplastic syndrome was suggested. Malignant T-Cell Lymphoma Shepherd and colleagues [144] in 1986 published a case series of 28 patients diagnosed with primary T-cell lymphoma of the gastrointestinal tract and histologic distinctive massive tissue eosinophilia. Abdominal pain, weight loss, nausea and vomiting were the chief symptoms. All of these patients were treated with local excision of the tumor. Radiofrequency ablation was used in seven patients, and chemotherapy was attempted in nine patients; however, these modalities remained palliative only, and the mean survival time was 11 months. Acknowledgment The author thanks Dr. Murli Krishna from the Department of Pathology, Mayo Clinic, Jacksonville, Florida, for providing histopathologic slides. References [1] Strobel S. Epidemiology of food sensitivity in childhood–with special reference to cow’s milk allergy in infancy. Monogr Allergy 1993;31:119–30. [2] Schafer T, Bohler E, Ruhdorfer S, et al. Epidemiology of food allergy/food intolerance in adults: associations with other manifestations of atopy. Allergy 2001;56(12):1172–9. [3] Young E, Stoneham MD, Petruckevitch A, et al. A population study of food intolerance. Lancet 1994;343:1127–30. [4] Sicherer SH, Munoz-Furlong A, Burks AW, et al. Prevalence of peanut and tree nut allergy in the US determined by a random digit dial telephone survey. J Allergy Clin Immunol 1999;103(4):559–62. [5] Munoz-Furlong A, Sampson HA, Sicherer SH. Prevalence of self-reported seafood allergy in the US. J Allergy Clin Immunol 2004;113(Suppl):S100. [6] Eggesbo M, Botten G, Halvorsen R, et al. The prevalence of allergy to egg: a populationbased study in young children. Allergy 2001;56(5):403–11. [7] Altman DR, Chiaramonte LT. Public perception of food allergy. J Allergy Clin Immunol 1996;97(6):1247–51. [8] Vandezande LM, Wallaert B, Desreumaux P, et al. Interleukin-5 immunoreactivity and mRNA expression in gut mucosa from patients with food allergy. Clin Exp Allergy 1999;29(5):652–9. [9] Laker AM. Food-induced eosinophilic proctocolitis. J Pediatr Gastroenterol Nutr 2000;30(Suppl):S58–60. [10] Woo JG, Assa’ad A, Heizer AB, et al. The -159 C–>T polymorphism of CD14 is associated with nonatopic asthma and food allergy. J Allergy Clin Immunol 2003;112(2):438–44. [11] Hand S, Darke C, Thompson J, et al. Human leucocyte antigen polymorphisms in nut-allergic patients in South Wales. Clin Exp Allergy 2004;34(5):720–4. [12] Sicherer SH, Furlong TJ, Maes HH, et al. Genetics of peanut allergy: a twin study. J Allergy Clin Immunol 2000;106(1 Pt 1):53–6. [13] Bjorksten B. Genetic and environmental risk factors for the development of food allergy. Curr Opin Allergy Clin Immunol 2005;5(3):249–53. [14] Frank L, Marian A, Visser M, et al. Exposure to peanuts in utero and in infancy and the development of sensitization to peanut allergens in young children. Pediatr Allergy Immunol 1999;10(1):27–32.
GUT EOSINOPHILIA
327
[15] Heine RG. Pathophysiology, diagnosis and treatment of food protein-induced gastrointestinal diseases. Curr Opin Allergy Clin Immunol 2004;4(3):221–9. [16] Perkin JE. The latex and food allergy connection. J Am Diet Assoc 2000;100(11):1381–4. [17] Raulf-Heimsoth M, Kespohl S, Crespo JF, et al. Natural rubber latex and chestnut allergy: cross-reactivity or co-sensitization? Allergy 2007;62(11):1277–81. [18] Kim KT, Hussain H. Prevalence of food allergy in 137 latex-allergic patients. Allergy Asthma Proc 1999;20(2):95–7. [19] Sampson HA. Infantile colic and food allergy: fact or fiction? J Pediatr 1989;115(4):583–4. [20] Yocum MW, Butterfield JH, Klein JS, et al. Epidemiology of anaphylaxis in Olmsted County: a population-based study. J Allergy Clin Immunol 1999;104(2 Pt 1):452–6. [21] Worobec A, Metcalfe DD. Anaphylactic syndrome. In: Austen KF, Frank MM, Atkinson JP, et al, editors. Samter’s immunologic diseases. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 825–36. [22] Lieberman P. Biphasic anaphylactic reactions. Ann Allergy Asthma Immunol 2005;95(3): 217–26 [quiz: 226, 258]. [23] Ellis AK, Day JH. Incidence and characteristics of biphasic anaphylaxis: a prospective evaluation of 103 patients. Ann Allergy Asthma Immunol 2007;98(1):64–9. [24] Douglas DM, Sukenick E, Andrade WP, et al. Biphasic systemic anaphylaxis: an inpatient and outpatient study. J Allergy Clin Immunol 1994;93(6):977–85. [25] Yocum MW, Khan DA. Assessment of patients who have experienced anaphylaxis: a 3-year survey. Mayo Clin Proc 1994;69(1):16–23. [26] The diagnosis and management of anaphylaxis: an updated practice parameter. J Allergy Clin Immunol 2005;115(3 Suppl 2):S483–523. [27] Osterballe M, Hansen TK, Mortz CG, et al. The clinical relevance of sensitization to pollenrelated fruits and vegetables in unselected pollen-sensitized adults. Allergy 2005;60(2): 218–25. [28] Asero R, Massironi F, Velati C. Detection of prognostic factors for oral allergy syndrome in patients with birch pollen hypersensitivity. J Allergy Clin Immunol 1996;97(2): 611–6. [29] Crimi E, Voltolini S, Gianiorio P, et al. Effect of seasonal exposure to pollen on specific bronchial sensitivity in allergic patients. J Allergy Clin Immunol 1990;85(6):1014–9. [30] Pastorello EA, Ortolani C. Oral allergy syndrome. In: Metcalfe DD, Sampson HA, Simon RA, editors. Food allergy: adverse reactions to food and food additives. Cambridge (UK): Blackwell Science; 1997. p. 221–34. [31] Anhoej C, Backer V, Nolte H. Diagnostic evaluation of grass- and birch-allergic patients with oral allergy syndrome. Allergy 2001;56(6):548–52. [32] Bindslev-Jensen C, Vibits A, Stahl Skov P, et al. Oral allergy syndrome: the effect of astemizole. Allergy 1991;46(8):610–3. [33] Asero R. Effects of birch pollen-specific immunotherapy on apple allergy in birch pollenhypersensitive patients. Clin Exp Allergy 1998;28(11):1368–73. [34] Bucher X, Pichler WJ, Dahinden CA, et al. Effect of tree pollen specific, subcutaneous immunotherapy on the oral allergy syndrome to apple and hazelnut. Allergy 2004;59(12): 1272–6. [35] Wessel MA, Cobb JC, Jackson EB, et al. Paroxysmal fussing in infancy, sometimes called colic. Pediatrics 1954;14(5):421–35. [36] Brazelton TB. Crying in infancy. Pediatrics 1962;29:579–88. [37] Lucassen PL, Assendelft WJ, van Eijk JT, et al. Systematic review of the occurrence of infantile colic in the community. Arch Dis Child 2001;84(5):398–403. [38] Clifford TJ, Campbell MK, Speechley KN, et al. Sequelae of infant colic: evidence of transient infant distress and absence of lasting effects on maternal mental health. Arch Pediatr Adolesc Med 2002;156(12):1183–8. [39] Roberts DM, Ostapchuk M, O’Brien JG. Infantile colic. Am Fam Physician 2004;70(4): 735–40.
328
TALLEY
[40] Reust CE, Blake RL Jr. Diagnostic workup before diagnosing colic. Arch Fam Med 2000;9(3):282–3. [41] Parker S. Colic. In: Parker S, Zuckerman B, Augustyn M, editors. Developmental and behavioral pediatrics: a handbook for primary care. Philadelphia: Lippincott Williams & Wilkins; 2005. p. 158. [42] Lucassen PL, Assendelft WJ, Gubbels JW, et al. Effectiveness of treatments for infantile colic: systematic review. BMJ 1998;316(7144):1563–9. [43] Evans RW, Fergusson DM, Allardyce RA, et al. Maternal diet and infantile colic in breastfed infants. Lancet 1981;1(8234):1340–2. [44] Lothe L, Lindberg T, Jakobsson I. Cow’s milk formula as a cause of infantile colic: a doubleblind study. Pediatrics 1982;70(1):7–10. [45] Campbell JP. Dietary treatment of infant colic: a double-blind study. J R Coll Gen Pract 1989;39(318):11–4. [46] Forsyth BW. Colic and the effect of changing formulas: a double-blind, multiple-crossover study. J Pediatr 1989;115(4):521–6. [47] Hill DJ, Hudson IL, Sheffield LJ, et al. A low allergen diet is a significant intervention in infantile colic: results of a community-based study. J Allergy Clin Immunol 1995;96 (6 Pt 1):886–92. [48] Williams J, Watkins-Jones R. Dicyclomine: worrying symptoms associated with its use in some small babies. Br Med J (Clin Res Ed) 1984;288(6421):901. [49] Taubman B. Parental counseling compared with elimination of cow’s milk or soy milk protein for the treatment of infant colic syndrome: a randomized trial. Pediatrics 1988;81(6):756–61. [50] Walker-Smith JA. Cow milk-sensitive enteropathy: predisposing factors and treatment. J Pediatr 1992;121:S111–5. [51] Kokkonen J, Haapalahti V, Laurila K, et al. Cow’s milk protein-sensitive enteropathy at school age. J Pediatr 2001;139(6):797–803. [52] Beyer K, Castro R, Birnbaum A, et al. Human milk-specific mucosal lymphocytes of the gastrointestinal tract display a TH2 cytokine profile. J Allergy Clin Immunol 2002;109(4): 707–13. [53] Sampson HA. Food allergies. In: Feldman M, Scharschmidt BF, Sleisenger MH, editors. Sleisenger and Fordtran’s gastrointestinal and liver disease: pathophysiology, diagnosis and management. Philadelphia:: WB Saunders; 2006. p. 433–4. [54] Nowak-Wegrzyn A, Sampson HA, Wood RA, et al. Food protein-induced enterocolitis syndrome caused by solid food proteins. Pediatrics 2003;111(4 Pt 1):829–35. [55] Chung HL, Hwang JB, Park JJ, et al. Expression of transforming growth factor beta-1, transforming growth factor type I and II receptors, and TNF-alpha in the mucosa of the small intestine in infants with food protein-induced enterocolitis syndrome. J Allergy Clin Immunol 2002;109(1):150–4. [56] Sicherer SH, Eigenmann PA, Sampson HA. Clinical features of food protein-induced enterocolitis syndrome. J Pediatr 1998;133(2):214–9. [57] Lake AM, Whitington PF, Hamilton SR. Dietary protein-induced colitis in breast-fed infants. J Pediatr 1982;101(6):906–10. [58] Odze RD, Wershil BK, Leichtner AM, et al. Allergic colitis in infants. J Pediatr 1995;126(2): 163–70. [59] Ormala T, Rintala R, Savilahti E. T cells of the colonic mucosa in patients with infantile colitis. J Pediatr Gastroenterol Nutr 2001;33(2):133–8. [60] Sampson HA. Utility of food-specific IgE concentrations in predicting symptomatic food allergy. J Allergy Clin Immunol 2001;107(5):891–6. [61] Bindslev-Jensen C, Ballmer-Weber BK, Bengtsson U, et al. Standardization of food challenges in patients with immediate reactions to foods–position paper from the European Academy of Allergology and Clinical Immunology. Allergy 2004;59(7): 690–7.
GUT EOSINOPHILIA
329
[62] Ortolani C, Ispano M, Pastorello EA, et al. Comparison of results of skin prick tests (with fresh foods and commercial food extracts) and RAST in 100 patients with oral allergy syndrome. J Allergy Clin Immunol 1989;83(3):683–90. [63] American College of Allergy, AI. Food allergy: a practice parameter. Ann Allergy Asthma Immunol 2006;96(3 Suppl 2):S1–68. [64] Sicherer SH, Sampson HA. 9. Food allergy. J Allergy Clin Immunol 2006;117(2 Suppl): S470–5. [65] Sicherer SH. Food protein-induced enterocolitis syndrome: case presentations and management lessons. J Allergy Clin Immunol 2005;115(1):149–56. [66] Bischoff SC, Mayer J, Wedemeyer J, et al. Colonoscopic allergen provocation (COLAP): a new diagnostic approach for gastrointestinal food allergy. Gut 1997;40(6):745–53. [67] Mofidi S. Nutritional management of pediatric food hypersensitivity. Pediatrics 2003;111(6 Pt 3):1645–53. [68] Jarvinen KM, Makinen-Kiljunen S, Suomalainen H. Cow’s milk challenge through human milk evokes immune responses in infants with cow’s milk allergy. J Pediatr 1999;135(4): 506–12. [69] Vadas P, Wai Y, Burks W, et al. Detection of peanut allergens in breast milk of lactating women. JAMA 2001;285(13):1746–8. [70] American Academy of Pediatrics. Committee on Nutrition. Hypoallergenic infant formulas. Pediatrics 2000;106(2 Pt 1):346–9. [71] Host A, Koletzko B, Dreborg S, et al. Dietary products used in infants for treatment and prevention of food allergy. Joint Statement of the European Society for Paediatric Allergology and Clinical Immunology (ESPACI) Committee on Hypoallergenic Formulas and the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) Committee on Nutrition. Arch Dis Child 1999;81(1):80–4. [72] Edwards AM. Oral sodium cromoglycate: its use in the management of food allergy. Clin Exp Allergy 1995;25(Suppl 1)):31–3. [73] Freier S, Berger H. Disodium cromoglycate in gastrointestinal protein intolerance. Lancet 1973;1(7809):913–5. [74] Kocoshis S, Gryboski JD. Use of cromolyn in combined gastrointestinal allergy. JAMA 1979;242(11):1169–73. [75] Patriarca G, Nucera E, Roncallo C, et al. Oral desensitizing treatment in food allergy: clinical and immunological results. Aliment Pharmacol Ther 2003;17(3):459–65. [76] Molkhou P, Dupont C. Ketotifen in prevention and therapy of food allergy. Ann Allergy 1987;59(5 Pt 2):187–93. [77] Suzuki J, Kawasaki Y, Nozawa R, et al. Oral disodium cromoglycate and ketotifen for a patient with eosinophilic gastroenteritis, food allergy and protein-losing enteropathy. Asian Pac J Allergy Immunol 2003;21(3):193–7. [78] Khan S. Eosinophilic gastroenteritis. Best Pract Res Clin Gastroenterol 2005;19(2): 177–98. [79] Lee CM, Changchien CS, Chen PC, et al. Eosinophilic gastroenteritis: 10 years experience. Am J Gastroenterol 1993;88(1):70–4. [80] Arora AS, Perrault J, Smyrk TC. Topical corticosteroid treatment of dysphagia due to eosinophilic esophagitis in adults. Mayo Clin Proc 2003;78(7):830–5. [81] Neustrom MR, Friesen C. Treatment of eosinophilic gastroenteritis with montelukast. J Allergy Clin Immunol 1999;104(2 Pt 1):506. [82] Schwartz DA, Pardi DS, Murray JA. Use of montelukast as steroid-sparing agent for recurrent eosinophilic gastroenteritis. Dig Dis Sci 2001;46(8):1787–90. [83] Vanderhoof JA, Young RJ, Hanner TL, et al. Montelukast: use in pediatric patients with eosinophilic gastrointestinal disease. J Pediatr Gastroenterol Nutr 2003;36(2):293–4. [84] Daikh BE, Ryan CK, Schwartz RH. Montelukast reduces peripheral blood eosinophilia but not tissue eosinophilia or symptoms in a patient with eosinophilic gastroenteritis and esophageal stricture. Ann Allergy Asthma Immunol 2003;90(1):23–7.
330
TALLEY
[85] Kalliomaki M, Salminen S, Poussa T, et al. Probiotics and prevention of atopic disease: 4-year follow-up of a randomised placebo-controlled trial. Lancet 2003;361(9372): 1869–71. [86] Aguirre M, Collins MD. Lactic acid bacteria and human clinical infection. J Appl Bacteriol 1993;75(2):95–107. [87] Salminen S, Von Wright A, Morelli L, et al. Demonstration of safety of probiotics—a review. Int J Food Microbiol 1998;44(1–2):93–106. [88] Metchnikoff E, Metchnikoff LL, Mitchell PC. The prolongation of life: optimistic studies. Paris: Essais Optimists; 1907. [89] Heyman M, Corthier G, Petit A, et al. Intestinal absorption of macromolecules during viral enteritis: an experimental study on rotavirus-infected conventional and germ-free mice. Pediatr Res 1987;22(1):72–8. [90] Loskutova IE. Effectiveness of using Maliutka and Malysh adapted propionic-acidophilus mixtures in the combined treatment of congenital hypotrophy. Vopr Pitan 1985;(3): 17–20 [in Russian]. [91] Wheeler JG, Bogle ML, Shema SJ, et al. Impact of dietary yogurt on immune function. Am J Med Sci 1997;313(2):120–3. [92] Majamaa H, Isolauri E. Probiotics: a novel approach in the management of food allergy. J Allergy Clin Immunol 1997;99(2):179–85. [93] Weller PF. The immunobiology of eosinophils. N Engl J Med 1991;324(16):1110–8. [94] Hagan P, Wilkins HA, Blumenthal UJ, et al. Eosinophilia and resistance to Schistosoma haematobium in man. Parasite Immunol 1985;7(6):625–32. [95] Croese J, Loukas A, Opdebeeck J, et al. Human enteric infection with canine hookworms. Ann Intern Med 1994;120(5):369–74. [96] Schad GA. Hookworms: pets to humans. Ann Intern Med 1994;120(5):434–5. [97] Walker NI, Croese J, Clouston AD, et al. Eosinophilic enteritis in northeastern Australia: pathology, association with Ancylostoma caninum, and implications. Am J Surg Pathol 1995;19(3):328–37. [98] Prociv P, Croese J. Human enteric infection with Ancylostoma caninum: hookworms reappraised in the light of a ‘‘new’’ zoonosis. Acta Trop 1996;62(1):23–44. [99] Caumes E, Carriere J, Datry A, et al. A randomized trial of ivermectin versus albendazole for the treatment of cutaneous larva migrans. Am J Trop Med Hyg 1993;49(5): 641–4. [100] McGovern TW, Talley NJ, Kephart GM, et al. Eosinophil infiltration and degranulation in Helicobacter pylori-associated chronic gastritis. Dig Dis Sci 1991;36(4):435–40. [101] Whitington PF, Whitington GL. Eosinophilic gastroenteropathy in childhood. J Pediatr Gastroenterol Nutr 1988;7(3):379–85. [102] Bogers J, Moreels T, De Man J, et al. Schistosoma mansoni infection causing diffuse enteric inflammation and damage of the enteric nervous system in the mouse small intestine. Neurogastroenterol Motil 2000;12(5):431–40. [103] Takeyama J, Abukawa D, Miura K. Eosinophilic gastroenteritis with cytomegalovirus infection in an immunocompetent child. World J Gastroenterol 2007;13(34): 4653–4. [104] O’Donovan WJ. Gold induced diarrhoea. Br J Dermatol 1929;41:17. [105] Jackson CW, Haboubi NY, Whorwell PJ, et al. Gold induced enterocolitis. Gut 1986;27(4):452–6. [106] Michet CJ Jr, Rakela J, Luthra HS. Auranofin-associated colitis and eosinophilia. Mayo Clin Proc 1987;62(2):142–4. [107] Dhawan A, Seemayer TA, Pinsinski C, et al. Posttransplant eosinophilic gastroenteritis in children. Liver Transpl Surg 1997;3(6):591–3. [108] Lee JH, Park HY, Choe YH, et al. The development of eosinophilic colitis after liver transplantation in children. Pediatr Transplant 2007;11(5):518–23.
GUT EOSINOPHILIA
331
[109] Romero R, Abramowsky CR, Pillen T, et al. Peripheral eosinophilia and eosinophilic gastroenteritis after pediatric liver transplantation. Pediatr Transplant 2003;7(6): 484–8. [110] Daneshpouy M, Socie G, Lemann M, et al. Activated eosinophils in upper gastrointestinal tract of patients with graft-versus-host disease. Blood 2002;99(8):3033–40. [111] McNeel D, Rubio MT, Damaj G, et al. Hypereosinophilia as a presenting sign of acute graft-versus-host disease after allogeneic bone marrow transplantation. Transplantation 2002;4(12):1797–800. [112] Vanek J. Gastric submucosal granuloma with eosinophilic infiltration. Am J Pathol 1949;25(3):397–411. [113] Helwig EB, Ranier A. Inflammatory fibroid polyps of the stomach. Surg Gynecol Obstet 1953;96(3):335–67. [114] Costa PM, Marques A, Tavora Oliveira E, et al. Inflammatory fibroid polyp of the esophagus. Dis Esophagus 2000;13(1):75–9. [115] Simmons MZ, Cho KC, Houghton JM, et al. Inflammatory fibroid polyp of the esophagus in an HIV-infected individual: case study. Dysphagia 1995;10(1):59–61. [116] Guerra Bautista JA, Ibanez Delgado F, Hernandez de la Torre Bustillo JM, et al. Inflammatory fibroid polyp of the stomach. Rev Esp Enferm Dig 2006;98(6):482–3 [in Spanish]. [117] Hirasaki S, Matsubara M, Ikeda F, et al. Gastric inflammatory fibroid polyp treated with Helicobacter pylori eradication therapy. Intern Med 2007;46(12):855–8. [118] Kim JS, Kwon SY, Byun KS, et al. Jejunal inflammatory fibroid polyp presenting as intussusception–a case report with review of the literature. Korean J Intern Med 1994;9(1):51–4. [119] Blackshaw AJ, Levison DA. Eosinophilic infiltrates of the gastrointestinal tract. J Clin Pathol 1986;39(1):1–7. [120] Suen KC, Burton JD. The spectrum of eosinophilic infiltration of the gastrointestinal tract and its relationship to other disorders of angiitis and granulomatosis. Hum Pathol 1979;10(1):31–43. [121] Buchman AL, Wolf D, Gramlich T. Eosinophilic gastrojejunitis associated with connective tissue disease. South Med J 1996;89(3):327–30. [122] Hallegua DS, Wallace DJ. Gastrointestinal manifestations of systemic lupus erythematosus. Curr Opin Rheumatol 2000;12(5):379–85. [123] Barbie DA, Mangi AA, Lauwers GY. Eosinophilic gastroenteritis associated with systemic lupus erythematosus. J Clin Gastroenterol 2004;38(10):883–6. [124] Noth I, Strek ME, Leff AR. Churg-Strauss syndrome. Lancet 2003;361(9357):587–94. [125] Churg J, Strauss L. Allergic granulomatosis, allergic angiitis, and periarteritis nodosa. Am J Pathol 1951;27(2):277–301. [126] Hellmich B, Ehlers S, Csernok E, et al. Update on the pathogenesis of Churg-Strauss syndrome. Clin Exp Rheumatol 2003;21(6 Suppl 32):S69–77. [127] Pagnoux C, Mahr A, Cohen P, et al. Presentation and outcome of gastrointestinal involvement in systemic necrotizing vasculitides: analysis of 62 patients with polyarteritis nodosa, microscopic polyangiitis, Wegener granulomatosis, Churg-Strauss syndrome, or rheumatoid arthritis-associated vasculitis. Medicine (Baltimore) 2005;84(2):115–28. [128] Aoyagi K, Yamamoto C, Maeda K, et al. Endoscopic and clinicopathological features of gastrointestinal involvement in Churg-Strauss syndrome. Gastrointest Endosc 2004;59(5): W1687. [129] Wysocki AP, Taylor G, Windsor JA. Inflammatory fibroid polyps of the duodenum: a review of the literature. Dig Surg 2007;24(3):162–8. [130] Saiji E, Ayadi-Kaddour A, Ben Slama S, et al. Inflammatory fibroid polyp of the ileum presenting as intussusception: a case report in an adolescent. Tunis Med 2006;84(7): 454–7 [in French]. [131] Ng C, Lam KY, Gupta TS, et al. Inflammatory fibroid polyp of the caecum in a patient with neurofibromatosis. Ann Acad Med Singapore 2004;33(6):797–9.
332
TALLEY
[132] de la Plaza R, Picardo AL, Cuberes R, et al. Inflammatory fibroid polyps of the large intestine. Dig Dis Sci 1999;44(9):1810–6. [133] Papadopoulos AA, Tzathas C, Polymeros D, et al. Symptomatic eosinophilic gastritis cured with Helicobacter pylori eradication. Gut 2005;54(12):1822. [134] Weller PF, Bubley GJ. The idiopathic hypereosinophilic syndrome. Blood 1994;83(10): 2759–79. [135] Chusid MJ, Dale DC, West BC, et al. The hypereosinophilic syndrome: analysis of fourteen cases with review of the literature. Medicine (Baltimore) 1975;54(1):1–27. [136] Enokihara H, Kajitani H, Nagashima S, et al. Interleukin-5 activity in sera from patients with eosinophilia. Br J Haematol 1990;75(4):458–62. [137] Spry CJ, Davies J, Tai PC, et al. Clinical features of fifteen patients with the hypereosinophilic syndrome. Q J Med 1983;52(205):1–22. [138] Baccarani M, Cilloni D, Rondoni M, et al. The efficacy of imatinib mesylate in patients with FIP1L1-PDGFRalpha-positive hypereosinophilic syndrome. Results of a multicenter prospective study. Haematologica 2007;92(9):1173–9. [139] DeSchryver-Kecskemeti K, Clouse RE. A previously unrecognized subgroup of ‘‘eosinophilic gastroenteritis’’: association with connective tissue diseases. Am J Surg Pathol 1984;8(3):171–80. [140] Sunkureddi PR, Luu N, Xiao SY, et al. Eosinophilic enteritis with systemic lupus erythematosus. South Med J 2005;98:1049–52. [141] Guillevin L, Cohen P, Gayraud M, et al. Churg-Strauss syndrome: clinical study and longterm follow-up of 96 patients. Medicine (Baltimore) 1999;78(1):26–37. [142] Male PJ, de Toledo F, Widgren S, et al. Pseudotumoral enterocolitis and massive eosinophilia. Gut 1983;24:345–50. [143] Stefanini GF, Addolorato G, Marsigli L, et al. Eosinophilic gastroenteritis in a patient with large-cell anaplastic lung carcinoma: a paraneoplastic syndrome? Ital J Gastroenterol 1994;26(7):354–6. [144] Shepherd NA, Blackshaw AJ, Hall PA, et al. Malignant lymphoma with eosinophilia of the gastrointestinal tract. Histopathology 1987;11(2):115–30.
Gastroenterol Clin N Am 37 (2008) 333–348
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Eosinophilic Gastroenteritis Seema Khan, MDa,*, Susan R. Orenstein, MDb a
Thomas Jefferson University Medical School, Division of Pediatric Gastroenterology and Nutrition, Alfred I. DuPont Hospital for Children, 1600 Rockland Road, Wilmington, DE 19803, USA b Pediatric Gastroenterology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15260, USA
E
osinophilic gastroenteritis (EG) is commonly regarded as a synonym for eosinophilic gastrointestinal disorders (EGID), referring to a broad spectrum of clinical presentations produced by prominent eosinophilic infiltration through a variable depth of one or more gastrointestinal sites, and thus including eosinophilic esophagitis, gastritis, enteritis, and proctocolitis. Alternatively, the term has been used to refer to eosinophilic inflammation limited exclusively to the stomach and small intestine. To avoid redundancy in this issue of Gastroenterology Clinics of North America, this article is restricted to a focused and updated review of eosinophilic gastroenteritis as limited to the latter definition. EPIDEMIOLOGY Because of a notable scarcity of published data on the epidemiology, it is difficult to determine the true burden of EG. An English language literature review does not reveal a well-analyzed published case series describing primary EG in the last 3 years, suggesting that it is an uncommon and perhaps difficult-todiagnose disorder. Nevertheless, several case reports of EG have emerged recently in different parts of the world [1–5]. The diagnosis affects most ethnicities, and all ages of both genders [6–10]. It is diagnosed most frequently in the third decade of life [6,7,10,11]. A personal or family history of allergic disorders, such as asthma, hay fever, or eczema, is present in 60% to 70% of patients [12]. A report from India on EG documented only 7 young adults during a 10year period [10], and a hospital in China identified 15 patients who had EG, including 2 children, during 18 years [6]. Talley and colleagues [7] reported one of the largest case series of EG by characterizing 40 patients diagnosed during a 30-year period.
*Corresponding author. E-mail address:
[email protected] (S. Khan). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.003
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
334
KHAN & ORENSTEIN
PATHOGENESIS EG may be primary, also known as allergic, or secondary to one of many conditions that provoke eosinophilic inflammation (see later section on differential diagnosis), with parasitic infections being the leading cause. The cause cannot be precisely determined in a small but important minority of patients, who are regarded as having idiopathic EG. Eosinophils and Key Mediators EG, like other allergic diseases, is characterized predominantly by a Th-2 type of inflammatory response. This response, in concert with eosinophils and their granular proteins and eotaxin-1, a selective chemoattractant, are all critical to the pathogenesis of EG. Eosinophils, the hallmark of the pathology of EG, normally reside in the gastrointestinal tract lining beyond the esophagus, and are believed to have a role in host defense (eg, against parasitic infections). Reported normal densities of eosinophils vary throughout the length and depth of the gastrointestinal tract (see later discussion of diagnostic evaluation). Eosinophils function as antigen-presenting cells and mediate inflammatory effects by releasing preformed granular proteins, such as eosinophil cationic protein (ECP), eosinophil derived neurotoxin (EDN), eosinophil peroxidase, and major basic protein (MBP) [12,13]. Extracellular deposits of ECP and MBP have been demonstrated in the small intestine of patients who have EG; electron microscopy has provided further evidence for eosinophil degranulation in duodenal mucosal samples from patients who have EG [14,15]. Eosinophils also secrete Th-2–type proinflammatory cytokines, such as interleukin (IL)–3, IL-4, IL-5, IL-18, and transforming growth factor, and lipid mediators that are cytotoxic to human intestinal epithelium [12,16,17]. Further new and exciting insights into the mechanism of intestinal mucosal eosinophilia have been made possible through the study of the BioBreeding (BB) lymphopenic (lyp) rat model [18]. The lyp gene encodes a molecule known as GTPase of the immunity-associated protein, which may protect lymphocytes against apoptosis. Homozygosity of the lyp gene mutation in this murine model is associated with intense eosinophilia of the small and large intestinal mucosa, an overall Th-2 phenotype expressed by cell cultures derived from lymph nodes and spleen, with increased levels of IL-4, IL-5, and IL-13 RNA, and increased serum levels of IgE. Furthermore, the rats develop clinical features of wasting, bloating, intestinal distention, splenomegaly, and mesenteric adenopathy. An interesting feature observed even before the histopathology findings was the presence of an autoantibody, BB rat serum IgG, bound to intestinal cells that resembled fibroblasts. This finding has led to the speculation that a novel autoantigen in the subepithelial layer may provoke mucosal eosinophilic invasion. Allergy Excess circulating eosinophils are considered a hallmark of atopy, and hence it is no surprise that the frequency of allergies in patients who have EG is 25% to
EOSINOPHILIC GASTROENTERITIS
335
75%. Peripheral eosinophilia is present in 50% of those who have both EG and atopic diseases, such as asthma, eczema, and hay fever [19–21]. EG-associated allergies are reported variably as mediated by either IgE or non-IgE mechanisms. Mice challenged by oral antigens produce correlates of human EG disease: antigen-specific serum antibodies, prominent eosinophilic intestinal infiltrates, and clinical manifestations, including diarrhea, gastromegaly, dysmotility, and cachexia [22]. Genetics/Familial Genetic susceptibility is also now being implicated in the pathogenesis, as evidenced by the presence of a family history of an EGID in 10% of patients [19]. CLINICAL PRESENTATIONS Patients who have EG have heterogeneous clinical presentations, some of which overlap with more familiar diagnoses, such as functional GI disorders and inflammatory bowel disease. Presenting symptoms may be vague and nonspecific (eg, recurrent abdominal pain [23]) or dramatic (eg, gastric ulcer perforation [24]). Symptom onset may be at any age, but most patients are diagnosed as young adults, between the third and fifth decades of life. Often patients experience symptoms for many months and undergo an elaborate evaluation before the diagnosis of EG is correctly established. The natural history of EG is not well studied, although the course is usually characterized by waxing and waning symptoms. Patients who have EG have selective involvement of the stomach (26%– 81%) and small intestine (28%–100%) [7,11,25]. As alluded to earlier, however, EG is popularly used to refer to those cases of prominent gastric or small intestinal eosinophilia that may have concurrent, although less prominent, eosinophilia of the esophagus, large intestine, or rectum. It is debated, for example, whether the occasional patient presenting with apparent eosinophilic esophagitis (EE) and marked eosinophilic inflammation extending to other segments of the gastrointestinal tract represents primary EE or EE as part of EGID. The most common presenting complaint is abdominal pain, reported by two thirds of patients in most case series, followed by the symptoms of nausea, vomiting, and diarrhea [7,20,26]. Classification The Klein classification describes EG clinical presentations in terms of the variable depth of eosinophilic infiltration: mucosal, submucosal, and serosal subtypes [27]. Mucosal The mucosal subtype is the most common one (25%–100%), perhaps because of the accessibility to diagnosis by routine endoscopy and biopsies. Patients who have mucosal EG present with common, albeit nonspecific, complaints of abdominal pain, nausea, vomiting, diarrhea, occult gastrointestinal bleeding,
336
KHAN & ORENSTEIN
anemia, weight loss, or protein-losing enteropathy [1,8,21]. Because of their nonspecific nature, these clinical presentations may be confused with irritable bowel syndrome, dyspepsia, pancreatitis, acute appendicitis, or inflammatory bowel disease. Intestinal loss of blood and protein is a unique presentation inchildren who have allergic EG, and is suspected to be attributable to increased intestinal permeability induced by eosinophilic inflammation. Frequent coexisting findings in patients who have mucosal EG are atopy and high serum IgE levels. Muscular This subtype of EG is diagnosed in 13% to 70% of all EG cases, and is best known to present with a clinical picture of gastric outlet or intestinal obstruction [2,4,28]. Abdominal pain is usually characterized as colicky. EG presenting as gastric outlet obstruction can mimic hypertrophic pyloric stenosis, and has been successfully managed with a hypoallergenic formula rather than surgery [9]. Intestinal obstruction attributable to enteric strictures, although rare, occurs most frequently at the level of the jejunum, and has been described in children and adults [2,4,29,30]. Serosal The serosal layer is involved in 12% to 40% of cases of EG. Most commonly affected are adults who present classically with ascites. Other notable findings in these patients are significant bloating, a higher level of peripheral eosinophilia, and a better response to steroids [7,31–34]. DIFFERENTIAL DIAGNOSIS A broad range of conditions are associated with gastrointestinal mucosal eosinophilia and the following is a brief review of some of the important differential diagnoses. Infections Parasitic infestations Among parasites, helminths are most characteristically associated with peripheral eosinophilia, probably as a reflection of an immunologic response to their tissue migration. Tissue eosinophilia also may be found with several other parasites, including hookworms (Ancylostoma caninum), pinworms (Enterobius vermicularis), Eustoma rotundatum, Giardia lamblia, Anisakis, Trichinella spiralis, Ascaris, Trichuris, Schistosomiasis, Toxocara canis, and Strongyloides stercoralis. Diagnosis should be pursued in the appropriate clinical situation using studies such as stool for ova and parasites, stool giardia antigen, serology, paracentesis (in the case of ascites), duodenal aspirate, and endoscopy with biopsies [35–40]. Helicobacter pylori H pylori gastritis has been observed in patients who have diffuse intestinal eosinophilia in the absence of food allergies. The significance of these coexisting findings, beyond coincidence, is not clear [41].
EOSINOPHILIC GASTROENTERITIS
337
Cytomegalovirus Cytomegalovirus (CMV) gastroenteritis evident by serology and pathology, the latter also consistent with EG, has been reported in an immunocompetent child presenting with protein-losing enteropathy [42]. Whether the allergic mucosa predisposed to reactivation of latent CMV or CMV triggered prominent intestinal eosinophilia could not be elucidated from this report. Interestingly, this patient had complete resolution of clinicopathologic features without specific antiviral or antiallergy therapy within 4 weeks. Medications Several medications have been implicated in producing gastrointestinal eosinophilia as an allergic response. Examples include interferon [43], gemfibrozil [44], enalapril [45], carbamazepine [46,47], clofazimine [48] and co-trimoxazole [49]. Tissue eosinophilia is usually reversible with cessation of the medication. Connective Tissue Disease and Vasculitis Various connective tissue disorders (eg, scleroderma, dermatomyositis, lupus) and vasculitis (eg, Churg-Strauss syndrome, polyarteritis nodosa) are associated with fluctuating peripheral and gastrointestinal eosinophilia [50–53]. The presence of specific clinicopathologic and autoimmune markers is helpful in their differentiation from primary EG. Inflammatory Fibroid Polyps Rarely diagnosed, these benign localized polyps, also known as fibroma, inflammatory pseudotumor, submucosal granuloma, or localized EG, originate in the submucosa and are accompanied by a variable eosinophilic infiltrate. These lesions are most commonly located in the stomach (35% of cases) and the small bowel (50%), and come to attention because of obstructive presentations. Surgical excision is a cure for symptomatic patients [54,55]. Hypereosinophilia Syndrome Hypereosinophilia syndrome is a rare heterogeneous disorder with features of unexplained marked peripheral hypereosinophilia (>1500 cells/lL for more than 6 consecutive months) despite an extensive evaluation, and presence of organ damage or dysfunction related to hypereosinophilia [56–58]. The heart, skin, and central nervous system are the major targets, whereas occasional intestinal involvement is reported. Recently, hypereosinophilia syndrome has been classified as either myeloproliferative or lymphocytic indicating an underlying hematologic basis for these variants. Inflammatory Bowel Disease Peripheral and intestinal eosinophilia is often noted in patients who have irritable bowel disease, in the context of classic clinicopathologic features that allow differentiation from primary EG and other conditions. Transplantation Recipients of solid organ transplantation may develop intestinal eosinophilia as a consequence of immunosuppression, an imbalance of Th-1/Th-2 lymphocytes,
338
KHAN & ORENSTEIN
and de novo food allergies. A substantial number of patients are on immunosuppressive regimens composed of tacrolimus, a calcineurin inhibitor that is strongly implicated in inducing intestinal eosinophilia. A high total IgE and food allergen–specific IgE levels are observed in this subgroup of patients and dietary therapy has proved to be satisfactory [59–61]. DIAGNOSTIC EVALUATION Diagnosis of EG requires suspecting the disease, excluding other disorders in the differential diagnoses, confirming the definitive diagnosis, and assessing for potential complications. Currently accepted diagnostic criteria are the presence of gastrointestinal symptoms, an intense eosinophilic infiltrate on histopathologic examination, and exclusion of other causes of intestinal eosinophilia. Tests considered useful in the evaluation of EG and its differentials are presented in Table 1 [62]. A single diagnostic algorithm may not be universally applicable, because, for example, the screening evaluation may follow the demonstration of tissue eosinophilia. The following aspects of evaluation usually complement each other. History and Physical Examination The history and physical examination should be aimed at eliciting information pertaining to food-related adverse effects, stigmata of atopic diseases (wheezing, eczema, rhinitis), and evidence of malnutrition (edema, anemia, failure to thrive). Laboratory In the context of gastrointestinal symptoms, peripheral eosinophilia, present in 50% to 100% of those who have EG, is indeed a useful clue to EG, but is not definitive. Peripheral eosinophil concentrations fluctuate and may reflect the effects of circadian rhythms. Tests on stool and duodenal aspirates for parasites (particularly helminths) are strongly recommended to exclude secondary causes of EG, particularly in high-risk geographic areas. Other potentially useful laboratory investigations are directed toward evaluation for anemia (complete blood count), hypoalbuminemia (serum albumin), enteric protein losses (stool a-1 antitrypsin), autoimmune associations (autoantibodies), the eosinophilic intestinal inflammatory process (eosinophils and their remnants—ie, Charcot-Leyden crystals in stools), and eosinophilic ascites (paracentesis). In the near future, serum, stool, and urine assays of active eosinophil inflammation (eg, ECP, EDN) may be used for follow-up and response to therapy. Allergy Evaluation Commonly available tests for allergy include skin prick tests (SPT) and in vitro quantitative CAP-fluorescent enzyme immunoassay (CAP-FEIA), formerly known as radioallergosorbent test (RAST), which detects allergen-specific IgE antibody. The results of both should be interpreted with caution because of low sensitivity and a high rate of false-positive results [63,64]. It has also
EOSINOPHILIC GASTROENTERITIS
339
Table 1 Diagnostic evaluation of eosinophilic gastroenteritis Evaluation
Significance
Complete blood count with differential Total protein and albumin
Anemia, absolute eosinophilia Hypoalbuminemia, edema, protein-losing enteropathy, ascites, malnutrition Protein-losing enteropathy Protein-losing enteropathy EG diagnosis and follow-up EG diagnosis and follow-up
Quantitative immunoglobulins Stool a-1 antitrypsin Stool eosinophils Stool for eosinophil-derived granular proteins (eg, ECP) Stool giardia antigen, ova and parasite Radiology Barium contrast studies Computerized tomography Ultrasound Food Allergy Serum CAP-FEIA Skin prick tests Skin patch tests Endoscopy Gross Microscopy
Surgery and pathology Paracentesis Miscellaneous Liver enzymes Pancreatic enzymes Erythrocyte sedimentation rate C-reactive protein Tissue transglutaminase antibody Autoimmune antibodies Stool calprotectin, lactoferrin
Exclusion of parasite-induced EG Thickening of folds, stenoses/strictures, inflammation, lymphadenopathy; exclusion of differentials (eg, appendicitis) Ascites IgE-mediated allergies IgE-mediated allergies Non–IgE-mediated allergies Erythema, erosions, ulcers, polyps, nodules Dense eosinophilic infiltrates, epithelial and glandular invasion, abscesses; exclusion of differentials Establishing diagnosis, relief of obstruction Eosinophilia in serosal EG Evaluation of differentials and multi-organ involvement
Abbreviation: CAP-FEIA, CAP-fluorescent enzyme immunoassay. Data from Khan S. Eosinophilic gastroenteritis. In: Gupte S, editor. Section IV: Stomach, duodenum and intestine (Textbook of Pediatric Gastroenterology, Hepatology and Nutrition). New Delhi (India): Peepee; in press.
been shown that the probability of a true positive food challenge, and hence the positive predictive value, is high if allergen-specific IgE to a few select foods (milk, soy, egg, wheat, peanut, and fish) exceeds certain values [65]. The high negative predictive value of SPT is useful in confirming the absence of IgE-mediated reactions, if good quality food extracts are used. In the evaluation of non–IgE-mediated food allergies, patch testing is now available in Europe and North America, but has not been studied specifically in EG, and remains of limited usefulness because of lack of standardized
340
KHAN & ORENSTEIN
criteria [66]. Double-blind placebo-controlled food challenges are not practical in the clinical setting, and may also have limited usefulness in EG, because delayed hypersensitivity reactions may not be apparent for a few days. In children strongly suspected to have allergic EG, therefore, the clinician is often faced with the challenge of balancing the benefits of eliminating offending foods against the risks of unnecessarily restricted diets caused by the institution of time-limited dietary trials. Radiographic Evaluation Barium contrast studies in patients who have EG may reveal irregular gastric or small intestinal folds, a string sign in gastric outlet obstruction, or strictures [9,43]. Ultrasound of the abdomen is useful in detecting serosal EG and ascites [67]. Deep layer infiltration and intestinal wall thickening may also be appreciated on computerized tomography [68]. White blood cell Tc-99m scintigraphy may demonstrate inflammation in EG, but without differentiating EG from other inflammatory causes [69]. Endoscopy and Pathology The macroscopic features of EG are few, but include erythema, whitish specks, focal erosions, ulcerations, thickening of folds, polyps, nodules (Fig. 1), and friability [5,42]. A fair number of cases are not associated with any visible mucosal abnormalities (Fig. 2A). Histopathology is the gold standard for diagnosis (Figs. 3 and 2B), but the precise criteria differentiating normal from pathologic states remain a matter of debate. The factors currently taken into consideration for differentiating
Fig. 1. The endoscopic appearance of the gastric mucosa in a patient who had eosinophilic gastroenteritis. The diffuse mucosal nodularity, as shown here, mimics H pylori nodular gastritis.
EOSINOPHILIC GASTROENTERITIS
341
Fig. 2. Eosinophilic duodenitis in a patient who had grossly normal duodenum (A) depicts histologic features of normal villous surface, hypercellular lamina propria, predominanteosinophilic inflammation, and eosinophilic microabscesses (B).
normal gastrointestinal tissue from any type of EGID are: eosinophil density, location along the gastrointestinal tract (ie, within the normal digestive tract, the highest number of eosinophils—up to 68 per high power field—are found in the appendix and cecum, and the lowest—none—in the esophagus), eosinophil distribution through the wall depth (eg, it is unusual to find eosinophils infiltrating the normal epithelium and crypts, or forming superficial aggregates or abscesses), eosinophil degranulation, and the absence of features pathognomonic for other diseases (eg, granulomata in Crohn disease) [70,71]. Normal eosinophil levels were determined retrospectively in 28 children who did not have apparent pathology, most of whom underwent an endoscopy for abdominal pain [71]. The mean (maximum) number of eosinophils per high power
Fig. 3. A photomicrograph of a gastric mucosal biopsy from the same patient shows hypercellularity of the lamina propria, up to 20 eosinophils per high power field with focal epithelial involvement, and subepithelial collagen. A Diff-Quick stain for H pylori was negative.
342
KHAN & ORENSTEIN
field in the lamina propria of the gastric antrum and duodenum were 1.9 1.3 (8) and 9.6 5.3 (26), respectively. The diagnosis may be missed despite endoscopy and biopsies because of patchy disease distribution, and also because of sparing of the mucosa in some forms of muscular and serosal EG. The diagnosis of muscular or serosal EG is particularly challenging and requires diligence. The diagnosis of these subtypes can be confirmed on pathology specimens obtained at laparoscopy or laparotomy [4,72]. Capsule endoscopy and balloon enteroscopy may prove useful in patients who have EG, but the inability to procure biopsies is likely to limit the value of these techniques in the diagnosis of EG [73–75]. MANAGEMENT Various treatments may be beneficial in EG [76]. Our current knowledge of the treatment of EG is derived virtually exclusively from small studies and anecdotal experience. Evidence for the true efficacy of any therapy suffers from the lack of well-designed and controlled studies. Nevertheless, corticosteroids and dietary therapy are two treatment interventions for which the data are convincingly favorable. Emerging treatments in development are biologics and selective anti-eosinophil agents. Diet The particular EG presentation, the patient’s age, and the expected compliance may be the most important determinants of the type of dietary therapy. Dietary therapy assumes the form of either an elimination diet or an elemental diet; both benefit from guidance by allergy evaluation. To ensure success with dietary therapy, it is important to support the patients and parents with relevant educational sources and dietary consultation. Elemental diets are indicated in those who have multiple food allergies, and produce improvement of symptoms and histology in 4 to 9 weeks [5,8]. The efficacy and safety of long-term dietary therapy has not been studied in EG. Moreover, these diets are aversive because of their poor palatability and impractical because of the high cost and extreme restriction. Directing food elimination by way of routinely available allergy testing (CAP-RAST and SPT) may be effective but does not always produce a favorable response in EG, probably because the immunologic reactions to foods in EG are mediated by non-IgE as well as IgE mechanisms. A food diary monitoring routine food consumption and its relation to adverse reactions may be used to decide the extent of food elimination. The empiric removal of the six major food allergens (milk, soy, wheat, egg, nuts, peanut, seafood), termed ‘‘the six food elimination diet,’’ has been shown to be effective and safe in EE [77]. It is a practical empiric approach in all patients who have primary EG who do not have obvious food allergies on allergy testing. Caution should be used in the implementation of an extremely restricted or elemental diet for longer than 6 to 8 weeks. In those assessed clinically as responders, a new food may be introduced every 5 to 7 days with continued
EOSINOPHILIC GASTROENTERITIS
343
vigilant follow-up. Repeat endoscopic evaluation is not routinely indicated, but may be advantageous in selected cases to guide complex treatment decisions. Steroids Systemic steroids produce symptomatic and histologic improvement in EG, regardless of type, within a few days to weeks of initiation [21,31,76]. Steroids are indicated in those patients who have EG who have failed to respond to, or declined, dietary therapy, and in those who have severe clinical presentations. The usual dosage, taper, and duration of therapy are equivalent to the treatment in inflammatory bowel disease. The use of 1 to 2 mg/kg/d for at least a month induces remission, and is then tapered over 2 to 3 months. The tapering and discontinuation may result in disease relapse, necessitating repeated use of steroids. Patients who have a disease course marked by chronic symptoms and relapses are candidates for steroid-sparing options to minimize the multitude of steroid-related serious side effects. These side effects include fluid and electrolyte imbalance, hyperglycemia, cushingoid state, growth suppression, bone demineralization, pituitary and adrenocortical hyporesponsiveness, and posterior subcapsular cataracts. Non–enteric-coated budesonide, with its extensive first pass metabolism and relatively favorable side effect profile, is a potentially safe option in those who have EG affecting the ileocecum and right colon [2,78]. Fluticasone may be used as a topical alternative to systemic steroids in treating EE [79,80] presenting concurrently with EG. Mast Cell Inhibitors A limited number of case reports favor the use of oral disodium cromoglycate and ketotifen as treatment options in EG [81,82]. Antihistamines Current evidence does not support the use of antihistamines (H-1 receptor antagonists) in EG, except in the patient who has concurrent environmental allergies. Leukotriene Receptor Antagonists Montelukast is a selective and competitive antagonist of leukotriene Cys-LT1 receptors expressed on bronchial smooth muscle cells and eosinophils. It is therefore an attractive steroid-sparing option in EG, acting by blocking the inflammatory effects of eosinophils. Studies report mixed results in induction of clinical and histologic remission, however [33,83]. Biologic Therapies Anti–interleukin-5 (mepolizumab) Early experience with humanized monoclonal antibody against IL-5 are limited to a small number of patients who had hypereosinophilia syndrome and EE, but may be seen as encouraging for future application in EG [84,85].
344
KHAN & ORENSTEIN
Anti-IgE therapy (omalizumab) Omalizumab is a humanized anti-IgE monoclonal Ab that has been shown to be effective in allergic asthma and allergic rhinitis. There is new evidence to support its use in EG. It was administered as a subcutaneous infusion every 2 weeks for eight doses in nine patients (12–76 years) who had EG. Omalizumab use was associated with decreased absolute eosinophil counts, a nonsignificant reduction in tissue eosinophilia, lowered IgE levels, and improved symptom scores [86]. Novel and Emerging Treatment Agents Several new therapeutic options are under investigation. These anti-eosinophil agents include eosinophil selective adhesion molecules, a monoclonal eotaxin antibody (CAT-213), and agents to enhance eosinophil apoptosis [87–89]. Surgery Surgery is sometimes necessary to relieve, and occasionally cure, symptoms in patients who have obstructive EG presentations [4,90]. Close postoperative follow-up and consideration of adjuvant medical or dietary therapy are important because of the potential for disease recurrence. NATURAL HISTORY There is a dearth of information about the long-term course of patients who have treated or untreated EG. Most experts regard primary EG as a chronic disorder characterized by relapses and remissions, and hence the recommendations for close follow-up and perhaps repeated endoscopic surveillance in selected cases. The natural history of EG has not been well described, but clinical studies point to a chronic relapsing course for patients who have this diagnosis. The long-term follow-up (2.5 to 5.5 years) of six children who had EG and protein-losing enteropathy initially treated successfully with an elemental diet has been described recently: despite some liberalization of the diet, clinical remission could be maintained only by continued dietary restrictions [8]. This study emphasizes the chronic nature of EG and the importance of long-term follow-up. Larger, prospective, multicenter studies are needed to investigate further the pathogenesis, safe and effective management options, natural history, and long-term outcome of patients who have EG. SUMMARY EG is an uncommon, yet important, entity in the spectrum of primary EGIDs, selectively affecting the stomach and small intestine with an eosinophilic inflammatory process. Multiple clinical presentations are recognized because of the variability in the location and depth of eosinophilic infiltration. History of atopy and allergies is present in 25% to 75% of cases. Recent investigations providing an insight into the pathogenesis of EG support a critical role for allergens, eosinophils, Th-2–type cytokines, and eotaxin-1 in mediating eosinophilic inflammation. The diagnosis is confirmed by demonstrating prominent tissue eosinophilia on histopathology. Treatment strategies include the use of
EOSINOPHILIC GASTROENTERITIS
345
restricted diets, corticosteroids, leukotriene receptor antagonists, mast cell stabilizers, and antibodies against IL5 and IgE. Many unanswered questions remain with regard to the natural history, optimal duration of therapy, safer steroid-sparing long-term treatment agents, and the means of reliable and noninvasive follow-up. References [1] Mendez Sanchez IM, Rivera Irigoin R, Ubina Aznar E, et al. Distinct clinical presentations of a single medical entity: eosinophilic enteritis. Gastroenterol Hepatol 2007;30(1):19–21 [in Spanish]. [2] Elsing C, Placke J, Gross-Weege W. Budesonide for the treatment of obstructive eosinophilic jejunitis. Z Gastroenterol 2007;45(2):187–9. [3] Shin WG, Park CH, Lee YS, et al. Eosinophilic enteritis presenting as intussusception in adult. Korean J Intern Med 2007;22(1):13–7. [4] Yun MY, Cho YU, Park IS, et al. Eosinophilic gastroenteritis presenting as small bowel obstruction: a case report and review of the literature. World J Gastroenterol 2007;13(11):1758–60. [5] Chehade M, Sicherer SH, Magid MS, et al. Multiple exudative ulcers and pseudopolyps in allergic eosinophilic gastroenteritis that responded to dietary therapy. J Pediatr Gastroenterol Nutr 2007;45(3):354–7. [6] Chen MJ, Chu CH, Lin SC, et al. Eosinophilic gastroenteritis: clinical experience with 15 patients. World J Gastroenterol 2003;9(12):2813–6. [7] Talley NJ, Shorter RG, Phillips SF, et al. Eosinophilic gastroenteritis: a clinicopathological study of patients with disease of the mucosa, muscle layer, and subserosal tissues. Gut 1990;31(1):54–8. [8] Chehade M, Magid MS, Mofidi S, et al. Allergic eosinophilic gastroenteritis with proteinlosing enteropathy: intestinal pathology, clinical course, and long-term follow-up. J Pediatr Gastroenterol Nutr 2006;42(5):516–21. [9] Khan S, Orenstein SR. Eosinophilic gastroenteritis masquerading as pyloric stenosis. Clin Pediatr (Phila) 2000;39(1):55–7. [10] Venkataraman S, Ramakrishna BS, Mathan M, et al. Eosinophilic gastroenteritis—an Indian experience. Indian J Gastroenterol 1998;17(4):148–9. [11] Lee CM, Changchien CS, Chen PC, et al. Eosinophilic gastroenteritis: 10 years experience. Am J Gastroenterol 1993;88(1):70–4. [12] Rothenberg ME. Eosinophilic gastrointestinal disorders (EGID). J Allergy Clin Immunol 2004;113(1):11–28, quiz 29. [13] Jacobsen EA, Taranova AG, Lee NA, et al. Eosinophils: singularly destructive effector cells or purveyors of immunoregulation? J Allergy Clin Immunol 2007;119(6):1313–20. [14] Torpier G, Colombel JF, Mathieu-Chandelier C, et al. Eosinophilic gastroenteritis: ultrastructural evidence for a selective release of eosinophil major basic protein. Clin Exp Immunol 1988;74(3):404–8. [15] Keshavarzian A, Saverymuttu SH, Tai PC, et al. Activated eosinophils in familial eosinophilic gastroenteritis. Gastroenterology 1985;88(4):1041–9. [16] Furuta GT. Emerging questions regarding eosinophil’s role in the esophago-gastrointestinal tract. Curr Opin Gastroenterol 2006;22(6):658–63. [17] Hogan SP, Rothenberg ME. Eosinophil function in eosinophil-associated gastrointestinal disorders. Curr Allergy Asthma Rep 2006;6(1):65–71. [18] Cousins L, Graham M, Tooze R, et al. Eosinophilic bowel disease controlled by the BB ratderived lymphopenia/Gimap5 gene. Gastroenterology 2006;131(5):1475–85. [19] Guajardo JR, Plotnick LM, Fende JM, et al. Eosinophil-associated gastrointestinal disorders: a world-wide-web based registry. J Pediatr 2002;141(4):576–81. [20] Khan S, Kandula L, Orenstein SR. Educational clinical case series in pediatric allergy and immunology. Pediatr Allergy Immunol 2007;18(7):629–39.
346
KHAN & ORENSTEIN
[21] Mendez-Sanchez N, Chavez-Tapia NC, Vazquez-Elizondo G, et al. Eosinophilic gastroenteritis: a review. Dig Dis Sci 2007;52(11):2904–11. [22] Hogan SP, Mishra A, Brandt EB, et al. A pathological function for eotaxin and eosinophils in eosinophilic gastrointestinal inflammation. Nat Immunol 2001;2(4):353–60. [23] Kokkonen J, Ruuska T, Karttunen TJ, et al. Mucosal pathology of the foregut associated with food allergy and recurrent abdominal pains in children. Acta Paediatr 2001; 90(1):16–21. [24] Siaw EK, Sayed K, Jackson RJ. Eosinophilic gastroenteritis presenting as acute gastric perforation. J Pediatr Gastroenterol Nutr 2006;43(5):691–4. [25] Redondo-Cerezo E, Cabello MJ, Gonzalez Y, et al. Eosinophilic gastroenteritis: our recent experience: one-year experience of atypical onset of an uncommon disease. Scand J Gastroenterol 2001;36(12):1358–60. [26] Naylor AR. Eosinophilic gastroenteritis. Scott Med J 1990;35(6):163–5. [27] Klein NC, Hargrove RL, Sleisenger MH, et al. Eosinophilic gastroenteritis. Medicine (Baltimore) 1970;49(4):299–319. [28] Tursi A, Rella G, Inchingolo CD, et al. Gastric outlet obstruction due to gastroduodenal eosinophilic gastroenteritis. Endoscopy 2007;39(Suppl 1):E184. [29] Tan HL, Sithasanan N, Foley P, et al. The successful medical management of severe duodenal strictures secondary to eosinophilic gastroenteritis in an infant. Pediatr Surg Int 2003;19(7):562–3. [30] Karande T, Oak SN, Trivedi A, et al. Proximal jejunal obstruction due to eosinophilic gastroenteritis. J Postgrad Med 1996;42(4):121–3. [31] Zhou HB, Chen JM, Du Q. Eosinophilic gastroenteritis with ascites and hepatic dysfunction. World J Gastroenterol 2007;13(8):1303–5. [32] Fenoglio LM, Benedetti V, Rossi C, et al. Eosinophilic gastroenteritis with ascites: a case report and review of the literature. Dig Dis Sci 2003;48(5):1013–20. [33] Urek MC, Kujundzic M, Banic M, et al. Leukotriene receptor antagonists as potential steroid sparing agents in a patient with serosal eosinophilic gastroenteritis. Gut 2006;55(9): 1363–4. [34] Mazokopakis E, Vrentzos G, Spanakis E, et al. A case of eosinophilic gastroenteritis with severe peripheral eosinophilia. Mil Med 2006;171(4):331–2. [35] Chira O, Badea R, Dumitrascu D, et al. Eosinophilic ascites in a patient with toxocara canis infection. A case report. Rom J Gastroenterol 2005;14(4):397–400. [36] Bahgat MA, El Gindy AE, Mahmoud LA, et al. Evaluation of the role of Ancylostoma caninum in humans as a cause of acute and recurrent abdominal pain. J Egypt Soc Parasitol 1999;29(3):873–82. [37] Tsibouris P, Galeas T, Moussia M, et al. Two cases of eosinophilic gastroenteritis and malabsorption due to Enterobious vermicularis. Dig Dis Sci 2005;50(12):2389–92. [38] Kocabay G, Gul E, Cagatay A, et al. Is eosinophilic gastroenteritis a primary disease or a secondary developing entity due to parasitosis? South Med J 2006;99(8):901. [39] Hong ST, Lim HS, Kim DH, et al. A case of gastroenteritis associated with gastric trichuriasis. J Korean Med Sci 2003;18(3):429–32. [40] Repiso Ortega A, Alcantara Torres M, Gonzalez de Frutos C, et al. [Gastrointestinal anisakiasis. Study of a series of 25 patients]. Gastroenterol Hepatol 2003;26(6):341–6 [in Spanish]. [41] Papadopoulos AA, Tzathas C, Polymeros D, et al. Symptomatic eosinophilic gastritis cured with Helicobacter pylori eradication. Gut 2005;54(12):1822. [42] Takeyama J, Abukawa D, Miura K. Eosinophilic gastroenteritis with cytomegalovirus infection in an immunocompetent child. World J Gastroenterol 2007;13(34):4653–4. [43] Kakumitsu S, Shijo H, Akiyoshi N, et al. Eosinophilic enteritis observed during alpha-interferon therapy for chronic hepatitis C. J Gastroenterol 2000;35(7):548–51. [44] Lee JY, Medellin MV, Tumpkin C. Allergic reaction to gemfibrozil manifesting as eosinophilic gastroenteritis. South Med J 2000;93(8):807–8.
EOSINOPHILIC GASTROENTERITIS
347
[45] Barak N, Hart J, Sitrin MD. Enalapril-induced eosinophilic gastroenteritis. J Clin Gastroenterol 2001;33(2):157–8. [46] Atkinson RJ, Dennis G, Cross SS, et al. Eosinophilic colitis complicating anti-epileptic hypersensitivity syndrome: an indication for colonoscopy? Gastrointest Endosc 2004;60(6): 1034–6. [47] Shakeer VK, Devi SR, Chettupuzha AP, et al. Carbamazepine-induced eosinophilic enteritis. Indian J Gastroenterol 2002;21(3):114–5. [48] Ravi S, Holubka J, Veneri R, et al. Clofazimine-induced eosinophilic gastroenteritis in AIDS. Am J Gastroenterol 1993;88(4):612–3. [49] Morimoto T, Sato T, Matsuoka A, et al. Trimethoprim-sulfamethoxazole-induced hypersensitivity syndrome associated with reactivation of human herpesvirus-6. Intern Med 2006; 45(2):101–5. [50] Sunkureddi PR, Luu N, Xiao SY, et al. Eosinophilic enteritis with systemic lupus erythematosus. South Med J 2005;98(10):1049–52. [51] Barbie DA, Mangi AA, Lauwers GY. Eosinophilic gastroenteritis associated with systemic lupus erythematosus. J Clin Gastroenterol 2004;38(10):883–6. [52] Schwake L, Stremmel W, Sergi C. Eosinophilic enterocolitis in a patient with rheumatoid arthritis. J Clin Gastroenterol 2002;34(4):487–8. [53] Kimura T, Nakaoka Y, Yoshida K, et al. Churg-Strauss syndrome diagnosed and followed with gastrointestinal fiberscopic studies and electroneuromyography. Intern Med 1998; 37(7):646–50. [54] Blackshaw AJ, Levison DA. Eosinophilic infiltrates of the gastrointestinal tract. J Clin Pathol 1986;39(1):1–7. [55] Makhlouf HR, Sobin LH. Inflammatory myofibroblastic tumors (inflammatory pseudotumors) of the gastrointestinal tract: how closely are they related to inflammatory fibroid polyps? Hum Pathol 2002;33(3):307–15. [56] Fletcher S, Bain B. Diagnosis and treatment of hypereosinophilic syndromes. Curr Opin Hematol 2007;14(1):37–42. [57] Roufosse FE, Goldman M, Cogan E. Hypereosinophilic syndromes. Orphanet J Rare Dis 2007;2:37. [58] Sheikh J, Weller PF. Clinical overview of hypereosinophilic syndromes. Immunol Allergy Clin North Am 2007;27(3):333–55. [59] Romero R, Abramowsky CR, Pillen T, et al. Peripheral eosinophilia and eosinophilic gastroenteritis after pediatric liver transplantation. Pediatr Transplant 2003;7(6):484–8. [60] Saeed SA, Integlia MJ, Pleskow RG, et al. Tacrolimus-associated eosinophilic gastroenterocolitis in pediatric liver transplant recipients: role of potential food allergies in pathogenesis. Pediatr Transplant 2006;10(6):730–5. [61] Lee JH, Park HY, Choe YH, et al. The development of eosinophilic colitis after liver transplantation in children. Pediatr Transplant 2007;11(5):518–23. [62] Khan S. Eosinophilic gastroenteritis. In: Gupte S, editor. Section IV; Stomach, duodenum and intestine (Textbook of pediatric gastroenterology, hepatology and nutrition). New Delhi: Peepee; in press. [63] Wood RA, Segall N, Ahlstedt S, et al. Accuracy of IgE antibody laboratory results. Ann Allergy Asthma Immunol 2007;99(1):34–41. [64] Bock SA. Diagnostic evaluation. Pediatrics 2003;111(6 Pt 3):1638–44. [65] Sampson HA, Ho DG. Relationship between food-specific IgE concentrations and the risk of positive food challenges in children and adolescents. J Allergy Clin Immunol 1997;100(4): 444–51. [66] Spergel JM, Brown-Whitehorn T, Beausoleil JL, et al. Predictive values for skin prick test and atopy patch test for eosinophilic esophagitis. J Allergy Clin Immunol 2007;119(2):509–11. [67] Buljevac M, Urek MC, Stoos-Veic T. Sonography in diagnosis and follow-up of serosal eosinophilic gastroenteritis treated with corticosteroid. J Clin Ultrasound 2005;33(1): 43–6.
348
KHAN & ORENSTEIN
[68] Sandrasegaran K, Rajesh A, Maglinte DD. Eosinophilic gastroenteritis presenting as acute abdomen. Emerg Radiol 2006;13(3):151–4. [69] Imai E, Kaminaga T, Kawasugi K, et al. The usefulness of 99mTc-hexamethylpropyleneamineoxime white blood cell scintigraphy in a patient with eosinophilic gastroenteritis. Ann Nucl Med 2003;17(7):601–3. [70] Lowichik A, Weinberg AG. A quantitative evaluation of mucosal eosinophils in the pediatric gastrointestinal tract. Mod Pathol 1996;9(2):110–4. [71] DeBrosse CW, Case JW, Putnam PE, et al. Quantity and distribution of eosinophils in the gastrointestinal tract of children. Pediatr Dev Pathol 2006;9(3):210–8. [72] Alexander P, Jacob S, Paul V. Laparoscopy in eosinophilic jejunitis presenting as subacute bowel obstruction: a case report. Trop Gastroenterol 2003;24(2):97–8. [73] Pungpapong S, Stark ME, Cangemi JR. Protein-losing enteropathy from eosinophilic enteritis diagnosed by wireless capsule endoscopy and double-balloon enteroscopy. Gastrointest Endosc 2007;65(6):917–8 [discussion: 918]. [74] Kim N, Kim JW, Hwang JH, et al. Visualization of jejunal bleeding by capsule endoscopy in a case of eosinophilic enteritis. Korean J Intern Med 2005;20(1):63–7. [75] Chen YY, Su WW, Soon MS, et al. Eosinophilic jejunitis presenting with acute abdomen: the usefulness of double-balloon enteroscopy. Gastrointest Endosc 2006;63(3):532–4. [76] Foroughi S, Prussin C. Clinical management of eosinophilic gastrointestinal disorders. Curr Allergy Asthma Rep 2005;5(4):259–61. [77] Kagalwalla AF, Sentongo TA, Ritz S, et al. Effect of six-food elimination diet on clinical and histologic outcomes in eosinophilic esophagitis. Clin Gastroenterol Hepatol 2006;4(9): 1097–102. [78] Siewert E, Lammert F, Koppitz P, et al. Eosinophilic gastroenteritis with severe protein-losing enteropathy: successful treatment with budesonide. Dig Liver Dis 2006;38(1):55–9. [79] Teitelbaum JE, Fox VL, Twarog FJ, et al. Eosinophilic esophagitis in children: immunopathological analysis and response to fluticasone propionate. Gastroenterology 2002;122(5): 1216–25. [80] Remedios M, Campbell C, Jones DM, et al. Eosinophilic esophagitis in adults: clinical, endoscopic, histologic findings, and response to treatment with fluticasone propionate. Gastrointest Endosc 2006;63(1):3–12. [81] Bolukbas FF, Bolukbas C, Uzunkoy A, et al. A dramatic response to ketotifen in a case of eosinophilic gastroenteritis mimicking abdominal emergency. Dig Dis Sci 2004; 49(11–12):1782–5. [82] Suzuki J, Kawasaki Y, Nozawa R, et al. Oral disodium cromoglycate and ketotifen for a patient with eosinophilic gastroenteritis, food allergy and protein-losing enteropathy. Asian Pac J Allergy Immunol 2003;21(3):193–7. [83] Quack I, Sellin L, Buchner NJ, et al. Eosinophilic gastroenteritis in a young girl—long term remission under Montelukast. BMC Gastroenterol 2005;5:24. [84] Garrett JK, Jameson SC, Thomson B, et al. Anti-interleukin-5 (mepolizumab) therapy for hypereosinophilic syndromes. J Allergy Clin Immunol 2004;113(1):115–9. [85] Stein ML, Collins MH, Villanueva JM, et al. Anti-IL-5 (mepolizumab) therapy for eosinophilic esophagitis. J Allergy Clin Immunol 2006;118(6):1312–9. [86] Foroughi S, Foster B, Kim N, et al. Anti-IgE treatment of eosinophil-associated gastrointestinal disorders. J Allergy Clin Immunol 2007;120(3):594–601. [87] Nutku-Bilir E, Hudson SA, Bochner BS. Interleukin-5 priming of human eosinophils alters siglec-8 mediated apoptosis pathways. Am J Respir Cell Mol Biol 2008;38(1):121–4. [88] Ackerman SJ, Bochner BS. Mechanisms of eosinophilia in the pathogenesis of hypereosinophilic disorders. Immunol Allergy Clin North Am 2007;27(3):357–75. [89] Bochner BS. Verdict in the case of therapies versus eosinophils: the jury is still out. J Allergy Clin Immunol 2004;113(1):3–9, quiz 10. [90] Uenishi T, Sakata C, Tanaka S, et al. Eosinophilic enteritis presenting as acute intestinal obstruction: a case report and review of the literature. Dig Surg 2003;20(4):326–9.
Gastroenterol Clin N Am 37 (2008) 349–368
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Eosinophilic Esophagitis in Adults Ganapathy A. Prasad, MD, MSa,*, Nicholas J. Talley, MD, PhD, FRACP, FRCP, FACPb a
Division of Gastroenterology and Hepatology, Alfred Main, GI Diagnostic Unit, Mayo Clinic College of Medicine, 200 First Street SW, Rochester, MN 55905, USA b Division of Gastroenterology and Hepatology, Davis Building, 6th Floor East/B, Mayo Clinic College of Medicine, 4500 San Pablo Road, Jacksonville, FL 32224, USA
E
osinophilic esophagitis (EE) is a disease characterized by eosinophilic infiltration of the esophageal mucosa, and is associated with a clinical syndrome of dysphagia and food impaction (FI) in adults. First described in 1978 [1], this syndrome is being recognized increasingly in the developed world, with multiple case series reported from the United States, Europe, and Australia during the past decade [2–9]. It is unclear if the increasing recognition of the disease is due to a truly increasing incidence or is a reflection of increased recognition by pathologists and clinicians. Endoscopic features suggestive of EE have also been described, including the presence of rings, longitudinal furrows, and mucosal fragility, although a proportion of patients may have normal-appearing mucosa. Recently, the specificity of these endoscopic findings has been questioned [10]. Diagnosis is established by the presence of eosinophilic infiltration of the esophageal mucosa; different thresholds have been used by various investigators, ranging from more than 15 eosinophils/ high-power field (HPF) to more than 24 eosinophils/HPF. A recent consensus statement proposed the use of more than 15 eosinophils/HPF as the diagnostic criteria for EE in the proper clinical context [11]. The cause of EE remains unclear, with allergic (allergies to food or aeroallergens) and immunologic mechanisms being proposed [12]. Successful treatment in adults has been reported with systemic and swallowed topical steroid preparations (in case series and randomized controlled trials) [2,13] and with oral leukotriene inhibitors in small case series [14]. Endoscopic dilation alone, and in combination with medications, has also been reported as a treatment modality in patients who have EE [15,16]. The natural history and clinical course of EE in adults is thought to be characterized by recurrent symptoms [17,18] but remains poorly defined. EPIDEMIOLOGY Data on the epidemiology of EE in adults remain scarce, particularly in the United States. Straumann and Simon [9] reported the increasing prevalence *Corresponding author. E-mail address:
[email protected] (G.A. Prasad). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.03.002
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
350
PRASAD & TALLEY
of EE diagnosed in adults in Olten County, Switzerland. Olten County is a defined geographic area with a stable population of 100,000, with a single gastroenterologist and a single pathologist providing care. The investigators provided estimates from 1989 to 2004. They reported an average annual incidence of 1.48 cases/100,000 population (range, 0–6) with a marked increase in cases during the past few years of this time period. The prevalence of EE progressively increased over this time period, to 23 cases/100,000 population. The investigators considered this figure to be an underestimate, given that it represented only severely symptomatic patients. The first truly population-based estimate of esophageal eosinophilia was provided from northern Sweden [7], where 1000 randomly selected subjects from two counties who had completed an abdominal symptom questionnaire underwent endoscopy with biopsies (taken 2 cm above, and at, the gastroesophageal junction). Biopsies were taken by three trained gastroenterologists and were interpreted by two gastrointestinal pathologists. The investigators classified patients into those with definite EE (20 eosinophils/HPF), probable EE (15–19 eosinophils/HPF), and possible EE (5–14 eosinophils/HPF). They found that 4 patients had definite EE, 7 had probable EE, and 25 had possible EE. In addition, they reported that 48 patients in all had eosinophils present in their esophageal biopsies; hence, the prevalence of EE by the current consensus definition of EE (>15 eosinophils/HPF) was 1.1%. Of the 4 patients who had definite EE in this study, 3 had symptoms of gastroesophageal reflux disease (GERD) and 1 was completely asymptomatic. Given the cross-sectional nature of this study, estimates of the incidence of EE could not be assessed. Secular trends in the epidemiology of EE were assessed in Olmsted County, Minnesota, during the past 3 decades (1976–2006) in a retrospective study. All cases of EE diagnosed between 1976 and 2006 were identified using the resources of the Rochester Epidemiology Project. All esophageal biopsies with any evidence of eosinophilic infiltration were reviewed by a single pathologist. Patients presenting with unexplained FI needing endoscopic therapy were included as probable surrogates of EE, to maximize case identification. The clinical course of all patients was also defined using medical records and, prospectively, by using a telephone questionnaire. A total of 3456 patient charts were reviewed; 82 patients who had EE and 80 patients who had idiopathic FI were identified. The incidence of EE increased significantly during the past 3 decades (from 0.86 cases [95% CI, 0.15, 1.56]/100,000 population/year from 1976 to 1985, to 8.78 cases [95% CI, 7.19, 10.37]/100,000 population/year from 1996 to 2006). The prevalence of EE was 104.7 cases (87.5, 122.0)/100,000 population as of 1/1/2007 in Olmsted County. In this study, the prevalence and incidence of EE appear to be higher than previously reported. The incidence of EE has increased significantly during the past 3 decades, perhaps indicating the presence of an as-yet-unidentified cause of EE in the environment. Estimates of the prevalence and incidence of EE in pediatric patients have been reported by investigators from Hamilton County, Ohio. The incidence ranged from 0.9 cases/10,000 population in 2000 to 1.28 cases/10,000
EOSINOPHILIC ESOPHAGITIS IN ADULTS
351
in 2003, with the prevalence in 2003 estimated to be 4.29 cases/10,000 population; these estimates were higher than estimates of Crohn’s disease in the pediatric population [19]. Lower estimates of prevalence were reported in a pediatric population from western Australia by Cherian and colleagues [20] (0.89 cases/10,000). A common feature among all case series reported in the literature is the male preponderance of cases and the diagnosis of the disease in younger patients. Other factors influencing EE, such as racial predilection or socioeconomic status, have yet to be evaluated. Natural History The natural history of EE in adults was perhaps best described by Straumann and colleagues [16], who followed 30 adults who had EE (>24 eosinophils/ HPF) with periodic dilations alone during a mean of 7.2 years. They reported that quality of life was affected in a severe manner in only 1 patient and in a minor manner in 15. Nutritional status was not compromised in any patient. Of the 30 patients, only 11 were treated with endoscopic dilation; of these, 4 had repeated dilations and 7 required only a single dilation. The remaining 19 patients were monitored only. Patients who had peripheral blood eosinophilia and severe endoscopic abnormalities were more likely to relapse than those who did not. Although fibrosis of the lamina propria was noted in later biopsies, no progression to deeper layers or to other parts of the gastrointestinal tract was observed. No patient developed esophageal carcinoma or hypereosinophilic syndrome. Hence, EE does not appear to influence life expectancy. In another study, reported from the Mayo Clinic in abstract form [21], follow-up of 10 patients who had EE 3 years after diagnosis using a mailed questionnaire revealed recurrent solid food dysphagia occurring at least once a week in 60% of patients. All had received retreatment with inhaled steroids, and 70% had received endoscopic dilation despite medical treatment. The clinical course of patients who had EE in Olmsted County was characterized by recurrent symptoms in 40% of patients responding to medical or endoscopic treatment, with EE also appearing to be a recurrent relapsing disease in a substantial proportion of patients. PATHOGENESIS Debate continues on the etiopathogenesis of EE, with allergic and immunologic mechanisms being proposed and investigated [22–25], based on the substantial proportion of adult patients who have associated food, inhalant, and seasonal allergies (40%–60%) [12,26] (although this appears to be somewhat lower than in children, where up to 80% have been reported to have an allergic diathesis [27–29]). Specific IgE testing, skin prick testing, and atopy patch testing have been reported to be positive in variable (32%–80%) proportions of adults and children with EE [25,30]. On the basis of allergy testing results, it is thought that EE may be a result of IgE-mediated and non–IgE-mediated reactions [31]. In addition, the relationship between EE and GERD remains to be defined [32], given the presence of esophageal eosinophilia in patients who have GERD (albeit to a lesser degree).
352
PRASAD & TALLEY
Two possible mechanisms of antigen sensitization have been proposed in EE. One is an initial esophageal sensitization to a food antigen, resulting in EE when esophageal re-exposure to the same antigen results; an alternative hypothesis is that initial bronchial sensitization followed by esophageal reexposure leads to EE. This latter mechanism was supported by experiments conducted by Mishra and colleagues [22], when they exposed mice to Aspergillus fumigatus intranasally and intragastrically. Only mice exposed to intragastric A fumigatus developed esophageal eosinophilia. A bronchial esophageal connection was also supported by experiments in which airway delivery of interleukin (IL)-13 promoted the development of EE [33]. EE appears to be associated with a Th2-type immune response; increased levels of eosinophil-active Th2 cytokines (such as IL-4, IL-5, and IL-13) and mast cells have been described in the esophagus in patients who have EE. Experimental models of EE can be induced in mice by means of allergen exposure, especially in the respiratory tract after mucosal or epicutaneous sensitization, and by means of overexpression of Th2 cytokines (IL-5 and IL-13) [22,23,34]. The esophagus is normally devoid of eosinophils. In patients who have EE, recruitment of eosinophils to the esophagus is thought to be mediated by IL-5 and eotaxin. IL-5 is a cytokine produced by T-helper (type 2) lymphocytes, which can prime eosinophils to react to chemoattractants such as eotaxin (an eosinophil chemotactic factor, also known to be overexpressed in patients who have EE Ref. [35]) (Fig. 1). IL-13 has also been shown to be crucial in the development of EE. In sensitized individuals, allergen exposure can lead to IgEmediated mast cell degranulation, which leads to the production of chemokines, histamine, and eosinophilic chemoattractants. These factors then induce eosinophil migration and degranulation, releasing products such as major basic protein (MBP), eosinophil cationic protein, and eosinophil-derived neurotoxin. These products can cause tissue damage, edema, and chronic inflammation, which, if prolonged, can lead to fibrosis [16]. In addition, these cationic proteins released by eosinophils set up a positive feedback loop by causing further mast cell degranulation, which leads to greater eosinophil recruitment to the esophagus and further release of eosinophil granule products [31]. The effect of eosinophil products such as MBP on smooth muscle has also been described, with muscarinic M2 receptors mediating smooth muscle contraction. This finding may explain the ability of eosinophilic infiltration of the esophagus to lead to dysphagia and FI, analogous to bronchoconstriction in asthma [28,36]. Familial clustering of EE has been reported by some investigators [9,19,37,38]. Whether this clustering reflects a true genetic predilection or a common environmental exposure remains unclear. Support for a genetic basis of EE comes from a study that found that the gene encoding for eotaxin 3 was the most highly induced gene in EE patients when compared with healthy controls [35]. Seasonal variations in the diagnosis and severity of symptoms of EE have been reported in adults [27] and children [39], rendering support to the hypothesis that environmental allergens play a role in the pathogenesis of EE.
EOSINOPHILIC ESOPHAGITIS IN ADULTS
353
Fig. 1. A proposed mode to explain the molecular and cellular mechanisms involved in EE pathogenesis, eotaxin-3–associated eosinophil recruitment, and treatment. Aeroallergen, food allergen, and skin sensitization have been implicated in EE pathogenesis. Elemental diet, glucocorticoids, and anti–IL-5 treatments improve the microscopic features of EE acting at different levels on the disease pathogenesis. Proton pomp inhibitors (PPI) inhibit Hþ secretion by parietal cells of the stomach and only partially improve EE features. Hyperplasic epithelial cells of the esophagus overexpress eotaxin-3, likely in response to IL-13. Eotaxin-3 overexpression allows chemoattraction of CCR3þ cells. Inheritance of EE disease suggests a genetic predisposition. A single-nucleotide polymorphism (SNP) in the exotoxin-3 gene has been associated with EE. The linkage disequilibrium pattern (ID’I) of the SNPs in the eotoxin-3 gene, based on the genotype data in a European population from the HapMap database, is presented. ID’I is the measure of the correlation between two SNPs; it ranges from 0 (in linkage equilibrium) to 1 (in strong linkage disequilibrium). Except for the SNP pair þ76 and 5419 with an ID’I of 0.71, the other SNP pairs (shown in black) are in strong linkage disequilibrium (ID’I ¼ 1). As such, the genotype of one SNP speaks for the others. (From Blanchard C, Wang N, Rothenberg ME. Eosinophilic esophagitis: pathogenesis, genetics, and therapy. J Allergy Clin Immunol 2006;118:1054; with permission.)
CLINICAL MANIFESTATIONS Patients who have EE tend be young men, with a male gender preponderance (a male/female ratio of 3:1 was reported in one meta-analysis Ref. [8]). Most patients are diagnosed in the third or fourth decade of life. The most common presenting symptom in adults is dysphagia to solids (60%–90%). FI is also a common presenting manifestation, with 50% to 60% of patients reporting episodes. Two prospective studies have reported that 50% to 55% of patients presenting to the emergency room with FI have EE, based on esophageal biopsies [4,40]. In addition, symptoms of GERD (heartburn and acid regurgitation) are also common in adults who have EE, from 24% in a meta-analysis [8] to 50% in a prospective study of patients who had dysphagia [10]. Some patients who have EE may also present with acid reflux symptoms that do not respond to medical treatment with proton pump inhibitors. Despite the presence of acid reflux symptoms, most patients who have EE (82%) have normal or negative
354
PRASAD & TALLEY
ambulatory pH studies [11]. Chest pain, abdominal pain, diarrhea, and weight loss are some of the other symptoms reported by patients diagnosed with EE. This pattern of presenting symptoms in adults differs from that of children, who present with more nonspecific symptoms such as poor feeding, vomiting, regurgitation, and failure to thrive. Older children and adolescents present with symptoms similar to those of adults. Symptoms in adults are compared with those in children in Box 1. Ten percent to fifty percent of studies in adults have reported varying degrees of peripheral eosinophilia [16,36,41]. The degree of eosinophilia reported has been modest and is complicated by studies using different thresholds to define eosinophilia. A recent meta-analysis reported that 30% of patients (82/266 patients from 19 studies) had peripheral eosinophilia [8]. DIAGNOSIS Endoscopic Manifestations Several typical endoscopic manifestations of EE have been described in case series of EE, including esophageal rings, which lead to ‘‘trachealization’’ or ‘‘felinization’’ of the esophageal mucosa (Fig. 2A); raised white specks, which may represent eosinophilic microabscesses (Fig. 2B); longitudinal furrows (Fig. 2C); whitish exudates; and ‘‘crepe paper’’ mucosa, which refers to a friable mucosa that tears by the mere passage of the endoscope. Strictures can be located in the proximal, middle, or lower third of the esophagus. Case series also describe a substantial proportion of patients who have EE as having normal esophageal mucosa on endoscopy (10%–33%) [8,11,17]. A prospective study on outpatients presenting with dysphagia found that 10% of patients Box 1: Comparison of presenting symptoms in adults and children with eosinophilic esophagitis Adults Dysphagia Food impaction Heartburn/acid reflux not responsive to medical treatment Chest pain Children Feeding aversion Vomiting, regurgitation Failure to thrive GERD symptoms not responsive to medical treatment Abdominal pain Chest pain Dysphagia, food impaction
EOSINOPHILIC ESOPHAGITIS IN ADULTS
355
Fig. 2. (A) Endoscopy showing multiple rings in the esophagus. (B) Endoscopy showing multiple rings with white spots in the esophagus. (C) Endoscopy showing linear furrows in the esophagus.
who had a ‘‘normal endoscopy’’ had evidence of EE on midesophageal biopsies, supporting the argument for taking esophageal biopsies in all patients who have dysphagia and a ‘‘normal’’ endoscopic examination [10]. The same study also found that the typical endoscopic manifestations of EE have a poor predictive value for a diagnosis of EE, with only 38% of patients who had typical endoscopic features suggestive of EE meeting the histologic criteria for EE on midesophageal biopsies. However, the presence of more than one endoscopic feature of EE increased the likelihood of EE on esophageal biopsies in the same study [10]. Esophageal Biopsies Site The recent consensus statement on the diagnosis and management of EE states that ‘‘multiple biopsy specimens should be obtained from different esophageal locations along the length of the esophagus’’ to confirm a diagnosis of EE in all patients who have suspected EE, ‘‘regardless of the gross appearance of the esophageal mucosa’’ [11]. Multiple prior studies have based their diagnosis of EE on biopsies obtained from either the proximal or the distal esophagus.
356
PRASAD & TALLEY
These studies have demonstrated a marked variability in the degree of eosinophilia in biopsies taken from different levels of the esophagus, with levels from the distal esophagus being numerically greater than those from the proximal esophagus [17,42,43]. Gonsalves and colleagues [42] performed a subgroup analysis in 20 out of 61 patients who had proximal and distal biopsies available; they found that 4 patients (20%) met the histologic criteria for EE only on distal biopsies. In contrast, in a large pediatric study, proximal esophageal biopsies demonstrated 100% sensitivity in making a diagnosis of EE, when compared with distal esophageal biopsies. Number In an elegant study, investigators modeled the sensitivity of diagnosis by varying the number of biopsies [42], with a diagnostic threshold of 15 eosinophils/HPF. The sensitivity of one biopsy specimen was 55% and increased to 100% with five biopsy specimens (Fig. 3), which led the consensus statement on the diagnosis of EE to recommend that ‘‘multiple’’ biopsy specimens be taken. At the authors’ institution, they recommend that at least four biopsies be taken from the esophagus at different levels, to maximize sensitivity. Despite multiple biopsies, false-negative results may occur because of patchy distribution of the eosinophilic infiltrate, other medications the patient may be taking (such as systemic or inhaled steroids and leukotriene inhibitors), or involvement of deeper esophageal layers (muscularis propria) by the eosinophilic inflammation [44]. Processing Bouin’s fluid was found to reduce the sensitivity of eosinophil detection in one study [45] and hence, it is recommended that mucosal biopsy samples be fixed in an alternative preservative (such as formalin).
Fig. 3. Variable sensitivity of diagnosis, with varying number of biopsies from the esophagus with different diagnostic thresholds. Bx, biopsy. (From Gonsalves N, Policarpio-Nicolas M, Zhang Q, et al. Histopatholgic variability and endoscopic correlates in adults with eosinophilic esophagitis. Gastrointest Endosc 2006;64:313; with permission.)
EOSINOPHILIC ESOPHAGITIS IN ADULTS
357
Pathology The diagnostic sine qua non of EE is eosinophilic infiltration of the esophageal mucosa (Fig. 4). The diagnostic threshold for EE has varied in different studies, given the lack of consensus on the precise threshold. The most commonly used thresholds are 15 [17,42,46,47], 20 [2,10,14,48], and 24 [9,13,20,35,49] eosinophils/HPF. In addition, variability has been enhanced by the use of mean, median, and maximal eosinophil counts in the diagnostic threshold, and by variation in the method of counting eosinophils, with some studies targeting the most densely infiltrated areas and other studies counting a set number of HPFs in a given area [50]. In addition, as pointed out by Dellon and colleagues [50], because of significant variation in eosinophil counts introduced by the use of different microscopes, the size of a single HPF can vary from 0.12 mm2 to 0.44 mm2. Given this background of significant variability, the North American consensus statement states that ‘‘intraepithelial eosinophils should be counted in the most intensely inflamed HPF of a biopsy ( 400) to generate a peak count.’’ A peak count of greater than or equal to 15 eosinophils/HPF in at least one HPF, in the proper clinical context, was adopted in this statement as the diagnostic criterion for EE [11]. Other pathologic features that are not pathognomonic of EE but have been recognized to be characteristic of EE include eosinophilic microabscesses (defined as the presence of four or more eosinophils in a cluster), which are pathologic correlates of endoscopic mucosal specks and plaques; preferential superficial layering of the eosinophilic infiltrate in the upper half of the esophageal mucosa; basal zone hyperplasia (defined as the basal zone occupying greater than 20% of the epithelium, as seen in Fig. 5); papillary lengthening; and epithelial edema. MBP has been proposed as a surrogate marker of eosinophil degranulation, with increased deposition of this protein being reported in the esophageal mucosa of patients who have EE compared with those who have GERD
Fig. 4. Eosinophilic infiltration of the esophageal squamous mucosa.
358
PRASAD & TALLEY
Fig. 5. Cross-section showing full thickness of the epithelium, demonstrating the striking basilar hyperplasia (elongated subepithelial papillae, thickened basal zone), commonly seen in EE.
[4,41]. It may be speculated that identification of increased levels of MBP in esophageal biopsies could increase the sensitivity of the diagnosis of EE. However, acquisition of biopsies and biopsy processing alone may cause eosinophil degranulation, releasing MBP, and possibly leading to a decrease in specificity. Other Tests Ambulatory pH testing Multiple studies have reported on the results of pH monitoring in patients who have EE. In a meta-analysis, results were normal in 90% of patients [8]. Esophageal pH testing may be helpful in excluding abnormal distal esophageal reflux because it is occasionally difficult to distinguish between EE and GERD using symptoms, endoscopy, and biopsies. An alternative approach, as suggested by the recent consensus statement on EE, is to perform an upper endoscopy after 6 to 8 weeks of twice-a-day proton pump inhibitor therapy [11]. Esophageal manometry Results of manometry in patients who have EE have been reported by a few investigators [14,51]. Most patients have normal results, with the most common abnormal findings being nonspecific peristaltic abnormalities [11] and incomplete lower esophageal sphincter relaxation [8]. Esophageal manometry does not appear to provide useful information in patients who have EE. Radiology Barium esophageal studies may demonstrate the presence of proximal and distal esophageal strictures, in addition to revealing diffuse narrowing of the esophagus [51–53]. A pediatric study reported the presence of EE in 8 of 18 children with Schatzki’s rings [54]. Hence, radiologic studies may be helpful before endoscopy in patients who have prominent dysphagia with a high likelihood of
EOSINOPHILIC ESOPHAGITIS IN ADULTS
359
strictures and may provide valuable information in terms of planning endoscopy and possible endoscopic dilation. ALLERGIC EVALUATION Most studies evaluating the role of allergies in EE have been conducted in children. In this article, the authors attempt to summarize the results of studies performed in adults. Allergic responses have been implicated in the pathogenesis of EE, based on multiple immunologic studies. Food allergies have been documented in a substantial proportion of patients who have EE [28]; this proportion may be higher in pediatric patients than in adult patients. Despite positive tests for food allergies, however, some patients who have EE may not respond to withdrawal of the foods that cause a positive food allergy test [55]. Based on these studies and the coexistence of other allergic diseases such as allergic rhinitis, asthma, and skin or food allergies in patients who have EE, it has been recommended that patients who have EE be evaluated by an allergist [11]. Methods of allergy testing include determination of peripheral eosinophil counts (modestly elevated in some studies), assessment of specific IgE testing, skin prick testing, and skin patch testing. In addition, some studies have correlated the extent of peripheral and tissue eosinophilia with the severity and persistence of symptoms [16,56]. Eosinophil products such as eosinophil-derived neurotoxin may have potential as biomarkers of EE. Although total IgE levels have been reported to be elevated in adult patients who have EE, the significant confounding influence of coexisting atopic conditions makes the values difficult to interpret. Food-specific radioallergosorbent testing has also been reported in patients who have EE [57]. The results of these tests appear to have poor predictive value regarding the response of symptoms to the avoidance of the foods associated with positive tests. In vitro food allergy testing (using IgE levels) is not currently recommended [11]. Skin prick testing involves pricking the skin with a bifurcated needle through a drop of commercially available allergen extract; a wheal-and-flare reaction is measured after 15 minutes. Most of the studies reporting the results of this test in EE patients have been from pediatric patients [30]. Close to two thirds of patients have positive reactions to one or more foods; common foods include peanuts, eggs, soy, cow milk, and wheat. The significance of a positive skin test remains to be determined fully, however, because treatment correlates of a positive test are unclear [58]. Skin patch testing involves prolonged contact of the allergen to the skin surface and, unlike skin prick testing (which assesses type 1 or immediate hypersensitivity), it assesses delayed (or Th2-mediated) hypersensitivity. As with skin prick testing, most studies have reported results with skin patch testing in children, with most pediatric patients who have EE testing positive [25,59]. Spergel and colleagues [59,60] reported that 77% of patients who had EE had resolution of eosinophilic inflammation on biopsies when their diets were modified on the basis of positive skin tests and patch tests. The most commonly identified foods were milk, eggs, soy, chicken, and wheat. These results were
360
PRASAD & TALLEY
further confirmed when the investigators reintroduced the foods that tested positive, which led to recurrent esophageal eosinophilia. Because these results are from only one center, the consensus statement recommended the use of patch testing for EE in research settings until these results are confirmed by other investigators. The diagnostic criteria for EE, as per the recent American Gastroenterological Association consensus statement, are summarized as follows [11]: Clinical symptoms of esophageal dysfunction 15 eosinophils in one HPF Lack of responsiveness to high-dose proton pump inhibitor therapy Normal pH monitoring of the distal esophagus
TREATMENT Medical Treatment Corticosteroids Because corticosteroids are able to control inflammation in other organs, systemic [51,61] and locally active corticosteroids have been used in EE, with success. Systemic corticosteroids are able to induce symptomatic and histologic remission. Their effectiveness has not been compared with other treatments. Doses used have been comparable to those used in other inflammatory diseases such as inflammatory bowel disease (1–2 mg/kg/d, tapered over 4–6 weeks). However, given the potential for significant systemic toxicity with prolonged use, and the availability of other effective alternatives, their use is now recommended only in patients who have severe disease manifestations necessitating rapid symptom relief, such as severe dysphagia associated with dehydration and weight loss, perhaps associated with refractory esophageal strictures. Given the toxicities associated with prolonged systemic steroid administration, maintenance therapy with systemic corticosteroids is not recommended because this is, unfortunately, associated with recurrent disease in a large proportion of patients. Recent reports have documented the ability of other immunosuppressants (such as azathioprine and 6 mercaptopurine) to exert a ‘‘steroid-sparing’’ effect, analogous to their use in inflammatory bowel disease, allowing the weaning of steroids and maintaining clinical and histologic remission [62]. Faubion and colleagues [48] first reported the ability of swallowed topical corticosteroids (fluticasone propionate, 880 lg/d, or beclomethasone) to induce symptomatic remission in children who have EE. This symptomatic remission has been subsequently reported in adult patients [2,6]. Doses used in adults range from 440 to 500 lg/d of fluticasone propionate twice a day for 4 to 6 weeks. Patients were asked to use the inhalers without a spacer, and to swallow afterwards. Patients were also instructed not to eat or drink for 30 minutes after swallowing the medication. Symptomatic remission was induced in all patients except one. Relapse of symptoms was noted in almost 50% of patients, most of whom were retreated with inhaled corticosteroids. A randomized placebo-controlled trial of inhaled fluticasone propionate in pediatric patients who
EOSINOPHILIC ESOPHAGITIS IN ADULTS
361
had EE showed that patients who had EE who received fluticasone were more likely to achieve symptomatic and histologic remission (50%) than those who received placebo (9%) (P ¼ .047) [13]. Younger patients who have a nonallergic phenotype appeared to be more likely to respond to treatment. An alternative approach has been reported recently in the literature using swallowed budesonide suspension mixed in sucralose (a nonabsorbable sugar) to increase fluid viscosity; safety (no adverse effects with unaffected morning cortisol levels) and efficacy (80% histologic response) was reported in 20 children who had EE [63]. The most common adverse effect associated with inhaled topical corticosteroids appears to be oral candidiasis, perhaps as a result of the drug being deposited in the pharynx; oral candidiasis does not appear to be associated with the use of the viscous budesonide preparation. Inhaled topical corticosteroids may be associated with a risk for bone loss at higher doses (typically more than 750 lg/d) for prolonged periods [12]; this finding has not been reported in patients who have EE. Leukotriene receptor antagonists Because of the postulated involvement of eosinophil-derived leukotrienes in the inflammatory cascade causing EE, leukotriene receptor antagonists have been used in patients who have EE, with some success. However, the evidence is based on one case series only. Attwood and colleagues [14] reported the treatment of eight patients with montelukast, with six reporting complete symptomatic relief (the remainder experienced improvement in dysphagia). The doses of montelukast used were 10 to 40 mg/d. Histologic improvement was not observed. The investigators also reported continuing maintenance treatment with montelukast for a median duration of 14 months, with persistent symptomatic remission in these patients. The drug was reasonably well tolerated, with one patient developing nausea and myalgia at 40 mg/d. Contradictory data were reported by Gupta and colleagues [64], who measured cysteinyl leukotriene (eosinophil chemoattractants) levels in pediatric patients who had and did not have EE. They found that levels were comparable between the two groups and did not correlate with the degree of eosinophilic esophageal inflammation. The consensus statement on EE does not recommend treatment with leukotriene inhibitors at this time. Endoscopic Treatment Several earlier studies have reported the successful treatment of EE with dilations alone; these have mostly been in adult patients [16,54,65–67]. In one study, patients were managed with dilations alone (most needing only one dilation over a mean follow-up of 7 years), with minimal effect on quality of life [16]. Concern about mucosal friability leading to deep mucosal tears (sometimes with the mere passage of the endoscope) and perforation, in addition to recurrence of symptoms, remains a barrier to the widespread use of dilations as a primary means of treatment of EE. Even patients who do not have perforation have been reported to develop significant chest pain needing narcotic analgesics and hospitalization, raising the possibility of small, self-contained
362
PRASAD & TALLEY
perforations in these patients [66]. A recent audit of complications associated with endoscopy in EE from a single medical center reported an alarmingly high rate of endoscopic complications; 31% of patients (11/31) who had EE had these complications. The investigators reported seven mucosal lacerations, three perforations, and one emesis-induced rupture. They also attempted to identify the factors associated with an increased risk for complications and found that the presence of a stricture, a longer duration of symptoms, and a greater density of eosinophilic infiltration predicted an increased risk for complications. Because the density of eosinophilic infiltration cannot be determined before endoscopy, they suggested that the performance of endoscopy and subsequent dilatation be deferred until biopsy specimens are reviewed or treatment is completed [68]. Given this concern about perforation in patients who have EE, it has been recommended that patients who have esophageal eosinophilia have a diagnostic endoscopy with biopsies followed by medical treatment before endoscopic dilation, whenever possible. It remains unknown, however, whether dilation following medical treatment is safer than dilation upfront. It is imperative to recognize that dilation may be necessary in patients who are refractory to medical management with underlying strictures. No data are currently available on the durability of symptomatic response following endoscopic dilation alone in patients who have EE. Dietary Treatment Given the strong circumstantial evidence of the role of food allergy in the pathogenesis of EE, an argument can be made for the removal of offending foods/ antigens as a strategy for the treatment of EE. Indeed, substantial evidence of the success of this strategy exists in the pediatric population. In children, success rates of 77% to 98% in inducing clinical and histologic remission have been reported through the elimination of all potential food antigens by using an elemental amino acid formula [55,69]; removing foods most likely to be associated with eosinophilia (such as dairy, eggs, wheat, soy, peanuts, and fish) [58]; and eliminating specific foods indicated by allergy testing by skin prick or skin patch testing [25,59]. Despite the success of dietary therapy in the pediatric population, instituting dietary changes, particularly on a long-term basis, is challenging in adults (given the poor tolerability of these significantly restricted diets), which explains the emphasis on anti-inflammatory medications and endoscopic approaches in treating EE in adult patients. A recent abstract reported the first clinical prospective trial of dietary modification in adults who had EE [70]. Adults with newly diagnosed EE (diagnosed with more than 20 eosinophils/ HPF on esophageal biopsies), or those with relapsing symptoms and a prior diagnosis of EE, were recruited. All patients underwent 24-hour pH testing and were unresponsive to twice-a-day proton pump inhibitor therapy. Following skin prick testing for food and aeroallergens, patients completed a 6-week trial of a six-food elimination diet (milk, soy, eggs, wheat, nuts, and seafood were all excluded from the diet) in addition to any foods that tested positive
EOSINOPHILIC ESOPHAGITIS IN ADULTS
363
on skin prick testing. Repeat endoscopy with biopsies was performed after 6 weeks of the special diet, and remission was defined as the presence of fewer than 5 eosinophils/HPF on esophageal biopsies. Nine patients entered the trial, of which three had completed the trial at the time of publication of this article. Of the three patients, one had a complete response, one had a partial response (defined as significant resolution of esophageal eosinophilia without meeting the criteria for response), and one had no response. Symptom scores of dysphagia improved by 30% after dietary treatment. In addition, markers of epithelial proliferation (Ki-67 and p63 staining) also decreased significantly. Further detailed results of this study and other trials are needed before any recommendations can be made about dietary modifications in adults. Given the limited tolerability of exclusion diets in adults, this approach may be more appropriate for patients who remain refractory to medical and endoscopic therapy. Acid Suppression Acid reflux is also associated with esophageal eosinophilia and hence, remains in the differential diagnosis of eosinophilic infiltration of the esophagus. A recent elegant review speculated on the possible association between EE and GERD, raising four possible hypotheses to explain this interaction between acid and eosinophils in the esophagus: (1) GERD causes esophageal injury that results in a mild eosinophilic infiltration; (2) GERD and EE coexist but are unrelated; (3) EE contributes to, or causes, GERD; or (4) GERD contributes to, or causes, EE [32]. Studies in children appear to show that, although GERD may be associated with eosinophilic infiltration in the esophagus, the degree of esophageal eosinophilia is mild (<5 eosinophils/HPF) and does not reach the levels associated with a diagnosis of EE. In addition, studies in children who have EE have reported normal pH studies in most patients who have EE [71–73]. Another study correlated the likelihood of response to acid blockade and found that patients who had fewer than 3 eosinophils/HPF were more likely to respond than those who had more than 20 eosinophils/HPF [71]. Studies in adults provide more confusing results, with not only a substantial proportion of patients (23% in a meta-analysis) meeting the histologic criteria for EE reporting symptoms of GERD [10], but also a fair proportion of patients (as high as 42% in one study Ref. [6]) meeting the histologic criteria for EE having positive pH studies. This confusion is further confounded by reports of patients meeting the histologic and endoscopic criteria for EE, who respond to highdose acid-suppressive therapy [65,74]. Biologic Agents With greater understanding of the pathogenesis of EE, it is evident that certain cytokines, such as IL-5, play a crucial role in the production, recruitment, and activation of eosinophils in the esophagus. Preclinical studies have found that monoclonal antibodies to IL-5 or IL-5 gene deletion block the induction of EE in mice. It has also been shown that monthly administration of intravenous monoclonal antibody to IL-5 in a patient who had hypereosinophilic syndrome along with esophageal eosinophilia resulted in a resolution of esophageal and
364
PRASAD & TALLEY
systemic eosinophilia within 3 months [75], which raises the possibility of similar agents being tried in patients who have EE who are resistant to conventional medications. UNRESOLVED ISSUES Several issues remain unresolved in the diagnosis and management of EE because of limited available evidence. Controversy continues about whether the observed increase in cases of EE reflects a true increase in the incidence of EE [9,19] or is merely a reflection of increased recognition by endoscopists and pathologists and an increasing use of endoscopy [76]. Studies reaching both conclusions have appeared in the literature, although most of the available evidence appears to favor a true increase in the incidence of EE. Data are lacking on secular trends in the incidence of EE in a population-based setting in adults from the United States. The interaction of GERD with esophageal eosinophilia remains to be defined fully. The overlap of findings (histologic and endoscopic) in patients who have GERD and EE opens up many theories [32] on the interaction between acid reflux and EE. Carefully performed, controlled studies are needed to delineate clearly the relationship between EE and GERD; it possible that an overlap exists between the two diseases, with the clearer distinctions at both ends of the spectrum. The role of food allergy testing and dietary therapy in adults, with exclusion of foods likely to be inciting an allergic response in the esophagus, also remains unclear. The number and location of biopsies needed for the diagnosis of EE also remains uncertain, although more evidence exists for the number of biopsies needed [42] than for their location (proximal versus distal). Even though the intensity of eosinophilic infiltration varies across the esophagus, the need for biopsies at multiple sites in the esophagus, although endorsed by the consensus statement on EE [11], remains to be proved in an evidence-based fashion; the largest pediatric study indicated that midesophageal biopsies alone missed only one case of EE when compared with distal esophageal biopsies [13]. The cost effectiveness of multiple biopsies at multiple locations also remains to be determined. The best treatment of EE in adults also remains undefined, with the only randomized study showing efficacy for inhaled/swallowed fluticasone being reported in children [13]. Studies reporting the efficacy of inhaled topical steroids (fluticasone and budesonide) in adults are all uncontrolled case series with unclear end points. An important issue that is yet to be defined is the end point of treatment in EE. Is symptomatic improvement sufficient to classify patients as responders or is a stricter end point of histologic remission (fewer than 5 eosinophils/HPF) necessary? The answer to this question perhaps depends on another unanswered question: What is the significance of persistent esophageal eosinophilia, in the absence of symptoms? Does it act as a harbinger of symptoms and relapse and does it lead to long-term alterations in esophageal function? Treating asymptomatic or minimally symptomatic esophageal eosinophilia with currently available treatments can lead to impaired quality
EOSINOPHILIC ESOPHAGITIS IN ADULTS
365
of life (especially with elimination diets) and adverse effects (oral candidiasis and suppression of the pituitary hypothalamic axis with high-dose topical steroids, especially in children). Finally, whether the best method of monitoring patients who have EE who have responded to treatment is clinically, with symptom assessment alone; endoscopically, with assessment of the mucosa visually and with biopsies; or radiologically also remains to be delineated in an evidence-based fashion. References [1] Landres RT, Kuster GG, Strum WB. Eosinophilic esophagitis in a patient with vigorous achalasia. Gastroenterology 1978;74:1298. [2] Arora AS, Perrault J, Smyrk TC. Topical corticosteroid treatment of dysphagia due to eosinophilic esophagitis in adults. Mayo Clin Proc 2003;78:830. [3] Chehade M, Sampson HA. Epidemiology and etiology of eosinophilic esophagitis. Gastrointest Endosc Clin N Am 2008;18:33. [4] Desai TK, Stecevic V, Chang CH, et al. Association of eosinophilic inflammation with esophageal food impaction in adults. Gastrointest Endosc 2005;61:795. [5] Liacouras CA, Markowitz JE. Eosinophilic esophagitis: a subset of eosinophilic gastroenteritis. Curr Gastroenterol Rep 1999;1:253. [6] Remedios M, Campbell C, Jones DM, et al. Eosinophilic esophagitis in adults: clinical, endoscopic, histologic findings, and response to treatment with fluticasone propionate. Gastrointest Endosc 2006;63:3. [7] Ronkainen J, Talley NJ, Aro P, et al. Prevalence of oesophageal eosinophils and eosinophilic oesophagitis in adults: the population-based Kalixanda study. Gut 2007;56:615. [8] Sgouros SN, Bergele C, Mantides A. Eosinophilic esophagitis in adults: a systematic review. Eur J Gastroenterol Hepatol 2006;18:211. [9] Straumann A, Simon HU. Eosinophilic esophagitis: escalating epidemiology? J Allergy Clin Immunol 2005;115:418. [10] Prasad GA, Talley NJ, Romero Y, et al. Prevalence and predictive factors of eosinophilic esophagitis in patients presenting with dysphagia: a prospective study. Am J Gastroenterol 2007;102:2627. [11] Furuta GT, Liacouras CA, Collins MH, et al. Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007;133:1342. [12] Arora AS, Yamazaki K. Eosinophilic esophagitis: asthma of the esophagus? Clin Gastroenterol Hepatol 2004;2:523. [13] Konikoff MR, Noel RJ, Blanchard C, et al. A randomized, double-blind, placebo-controlled trial of fluticasone propionate for pediatric eosinophilic esophagitis. Gastroenterology 2006;131:1381. [14] Attwood SE, Lewis CJ, Bronder CS, et al. Eosinophilic oesophagitis: a novel treatment using Montelukast. Gut 2003;52:181. [15] Gawrieh S, Shaker R. Treatment options for eosinophilic esophagitis: montelukast. Curr Gastroenterol Rep 2004;6:190. [16] Straumann A, Spichtin HP, Grize L, et al. Natural history of primary eosinophilic esophagitis: a follow-up of 30 adult patients for up to 11.5 years. Gastroenterology 2003;125: 1660. [17] Liacouras CA, Spergel JM, Ruchelli E, et al. Eosinophilic esophagitis: a 10-year experience in 381 children. Clin Gastroenterol Hepatol 2005;3:1198. [18] Straumann A. The natural history and complications of eosinophilic esophagitis. Gastrointest Endosc Clin N Am 2008;18:99. [19] Noel RJ, Putnam PE, Rothenberg ME. Eosinophilic esophagitis. N Engl J Med 2004;351: 940.
366
PRASAD & TALLEY
[20] Cherian S, Smith NM, Forbes DA. Rapidly increasing prevalence of eosinophilic oesophagitis in Western Australia. Arch Dis Child 2006;91:1000. [21] Helou Ef SJ, Arora A. Three-year follow-up of topical corticosteroid treatment for eosinophilic esophagitis in adults. Am J Gastroenterol 2005;100:S47. [22] Mishra A, Hogan SP, Brandt EB, et al. An etiological role for aeroallergens and eosinophils in experimental esophagitis. J Clin Invest 2001;107:83. [23] Mishra A, Rothenberg ME. Intratracheal IL-13 induces eosinophilic esophagitis by an IL-5, eotaxin-1, and STAT6-dependent mechanism. Gastroenterology 2003;125:1419. [24] Mishra A, Schlotman J, Wang M, et al. Critical role for adaptive T cell immunity in experimental eosinophilic esophagitis in mice. J Leukoc Biol 2007;81:916. [25] Spergel JM, Beausoleil JL, Mascarenhas M, et al. The use of skin prick tests and patch tests to identify causative foods in eosinophilic esophagitis. J Allergy Clin Immunol 2002;109:363. [26] Simon D, Marti H, Heer P, et al. Eosinophilic esophagitis is frequently associated with IgEmediated allergic airway diseases. J Allergy Clin Immunol 2005;115:1090. [27] Fogg MI, Ruchelli E, Spergel JM. Pollen and eosinophilic esophagitis. J Allergy Clin Immunol 2003;112:796. [28] Rothenberg ME, Mishra A, Collins MH, et al. Pathogenesis and clinical features of eosinophilic esophagitis. J Allergy Clin Immunol 2001;108:891. [29] Teitelbaum JE, Fox VL, Twarog FJ, et al. Eosinophilic esophagitis in children: immunopathological analysis and response to fluticasone propionate. Gastroenterology 2002;122: 1216. [30] Dauer EH, Freese DK, El-Youssef M, et al. Clinical characteristics of eosinophilic esophagitis in children. Ann Otol Rhinol Laryngol 2005;114:827. [31] Swoger JM, Weiler CR, Arora AS. Eosinophilic esophagitis: is it all allergies? Mayo Clin Proc 2007;82:1541. [32] Spechler SJ, Genta RM, Souza RF. Thoughts on the complex relationship between gastroesophageal reflux disease and eosinophilic esophagitis. Am J Gastroenterol 2007;102:1301. [33] Thompson DM, Arora AS, Romero Y, et al. Eosinophilic esophagitis: its role in aerodigestive tract disorders. Otolaryngol Clin North Am 2006;39:205. [34] Mishra A, Hogan SP, Brandt EB, et al. IL-5 promotes eosinophil trafficking to the esophagus. J Immunol 2002;168:2464. [35] Blanchard C, Wang N, Stringer KF, et al. Eotaxin-3 and a uniquely conserved geneexpression profile in eosinophilic esophagitis. J Clin Invest 2006;116:536. [36] Straumann A, Bauer M, Fischer B, et al. Idiopathic eosinophilic esophagitis is associated with a T(H)2-type allergic inflammatory response. J Allergy Clin Immunol 2001;108:954. [37] Meyer GW. Eosinophilic esophagitis in a father and a daughter. Gastrointest Endosc 2005;61:932. [38] Patel SM, Falchuk KR. Three brothers with dysphagia caused by eosinophilic esophagitis. Gastrointest Endosc 2005;61:165. [39] Wang FY, Gupta SK, Fitzgerald JF. Is there a seasonal variation in the incidence or intensity of allergic eosinophilic esophagitis in newly diagnosed children? J Clin Gastroenterol 2007;41:451. [40] Kerlin P, Jones D, Remedios M, et al. Prevalence of eosinophilic esophagitis in adults with food bolus obstruction of the esophagus. J Clin Gastroenterol 2007;41:356. [41] Parfitt JR, Gregor JC, Suskin NG, et al. Eosinophilic esophagitis in adults: distinguishing features from gastroesophageal reflux disease: a study of 41 patients. Mod Pathol 2006;19:90. [42] Gonsalves N, Policarpio-Nicolas M, Zhang Q, et al. Histopathologic variability and endoscopic correlates in adults with eosinophilic esophagitis. Gastrointest Endosc 2006;64:313. [43] Noel RJ, Putnam PE, Collins MH, et al. Clinical and immunopathologic effects of swallowed fluticasone for eosinophilic esophagitis. Clin Gastroenterol Hepatol 2004;2:568.
EOSINOPHILIC ESOPHAGITIS IN ADULTS
367
[44] Fox VL, Nurko S, Teitelbaum JE, et al. High-resolution EUS in children with eosinophilic ‘‘allergic’’ esophagitis. Gastrointest Endosc 2003;57:30. [45] Potter JW, Saeian K, Staff D, et al. Eosinophilic esophagitis in adults: an emerging problem with unique esophageal features. Gastrointest Endosc 2004;59:355. [46] Fox VL, Nurko S, Furuta GT. Eosinophilic esophagitis: it’s not just kid’s stuff. Gastrointest Endosc 2002;56:260. [47] Liacouras CA. Eosinophilic esophagitis in children and adults. J Pediatr Gastroenterol Nutr 2003;37(Suppl 1):S23. [48] Faubion WA Jr, Perrault J, Burgart LJ, et al. Treatment of eosinophilic esophagitis with inhaled corticosteroids. J Pediatr Gastroenterol Nutr 1998;27:90. [49] Blanchard C, Wang N, Rothenberg ME. Eosinophilic esophagitis: pathogenesis, genetics, and therapy. J Allergy Clin Immunol 2006;118:1054. [50] Dellon ES, Aderoju A, Woosley JT, et al. Variability in diagnostic criteria for eosinophilic esophagitis: a systematic review. Am J Gastroenterol 2007;102:2300. [51] Vitellas KM, Bennett WF, Bova JG, et al. Idiopathic eosinophilic esophagitis. Radiology 1993;186:789. [52] Khan S, Orenstein SR, Di Lorenzo C, et al. Eosinophilic esophagitis: strictures, impactions, dysphagia. Dig Dis Sci 2003;48:22. [53] Vitellas KM, Bennett WF, Bova JG, et al. Radiographic manifestations of eosinophilic gastroenteritis. Abdom Imaging 1995;20:406. [54] Nurko S, Teitelbaum JE, Husain K, et al. Association of Schatzki ring with eosinophilic esophagitis in children. J Pediatr Gastroenterol Nutr 2004;38:436. [55] Markowitz JE, Spergel JM, Ruchelli E, et al. Elemental diet is an effective treatment for eosinophilic esophagitis in children and adolescents. Am J Gastroenterol 2003;98:777. [56] Konikoff MR, Blanchard C, Kirby C, et al. Potential of blood eosinophils, eosinophil-derived neurotoxin, and eotaxin-3 as biomarkers of eosinophilic esophagitis. Clin Gastroenterol Hepatol 2006;4:1328. [57] Croese J, Fairley SK, Masson JW, et al. Clinical and endoscopic features of eosinophilic esophagitis in adults. Gastrointest Endosc 2003;58:516. [58] Kagalwalla AF, Sentongo TA, Ritz S, et al. Effect of six-food elimination diet on clinical and histologic outcomes in eosinophilic esophagitis. Clin Gastroenterol Hepatol 2006;4:1097. [59] Spergel JM, Andrews T, Brown-Whitehorn TF, et al. Treatment of eosinophilic esophagitis with specific food elimination diet directed by a combination of skin prick and patch tests. Ann Allergy Asthma Immunol 2005;95:336. [60] Spergel JM, Brown-Whitehorn T, Beausoleil JL, et al. Predictive values for skin prick test and atopy patch test for eosinophilic esophagitis. J Allergy Clin Immunol 2007;119:509. [61] Liacouras CA, Wenner WJ, Brown K, et al. Primary eosinophilic esophagitis in children: successful treatment with oral corticosteroids. J Pediatr Gastroenterol Nutr 1998;26:380. [62] Netzer P, Gschossmann JM, Straumann A, et al. Corticosteroid-dependent eosinophilic oesophagitis: azathioprine and 6-mercaptopurine can induce and maintain long-term remission. Eur J Gastroenterol Hepatol 2007;19:865. [63] Aceves SS, Bastian JF, Newbury RO, et al. Oral viscous budesonide: a potential new therapy for eosinophilic esophagitis in children. Am J Gastroenterol 2007;102:2271. [64] Gupta SK, Peters-Golden M, Fitzgerald JF, et al. Cysteinyl leukotriene levels in esophageal mucosal biopsies of children with eosinophilic inflammation: are they all the same? Am J Gastroenterol 2006;101:1125. [65] Morrow JB, Vargo JJ, Goldblum JR, et al. The ringed esophagus: histological features of GERD. Am J Gastroenterol 2001;96:984. [66] Vasilopoulos S, Murphy P, Auerbach A, et al. The small-caliber esophagus: an unappreciated cause of dysphagia for solids in patients with eosinophilic esophagitis. Gastrointest Endosc 2002;55:99. [67] Vasilopoulos S, Shaker R. Defiant dysphagia: small-caliber esophagus and refractory benign esophageal strictures. Curr Gastroenterol Rep 2001;3:225.
368
PRASAD & TALLEY
[68] Cohen MS, Kaufman AB, Palazzo JP, et al. An audit of endoscopic complications in adult eosinophilic esophagitis. Clin Gastroenterol Hepatol 2007;5:1149. [69] Kelly KJ, Lazenby AJ, Rowe PC, et al. Eosinophilic esophagitis attributed to gastroesophageal reflux: improvement with an amino acid-based formula. Gastroenterology 1995;109:1503. [70] Gonsalves N, Ritz S, Yang G, et al. A prospective clinical trial of allergy testing and food elimination diet in adults with eosinophilic esophagitis. Gastroenterology 2007;132:A6. [71] Ruchelli E, Wenner W, Voytek T, et al. Severity of esophageal eosinophilia predicts response to conventional gastroesophageal reflux therapy. Pediatr Dev Pathol 1999;2:15. [72] Steiner SJ, Gupta SK, Croffie JM, et al. Correlation between number of eosinophils and reflux index on same day esophageal biopsy and 24 hour esophageal pH monitoring. Am J Gastroenterol 2004;99:801. [73] Steiner SJ, Kernek KM, Fitzgerald JF. Severity of basal cell hyperplasia differs in reflux versus eosinophilic esophagitis. J Pediatr Gastroenterol Nutr 2006;42:506. [74] Ngo P, Furuta GT, Antonioli DA, et al. Eosinophils in the esophagus–peptic or allergic eosinophilic esophagitis? Case series of three patients with esophageal eosinophilia. Am J Gastroenterol 2006;101:1666. [75] Garrett JK, Jameson SC, Thomson B, et al. Anti-interleukin-5 (mepolizumab) therapy for hypereosinophilic syndromes. J Allergy Clin Immunol 2004;113:115. [76] Vanderheyden AD, Petras RE, DeYoung BR, et al. Emerging eosinophilic (allergic) esophagitis: increased incidence or increased recognition? Arch Pathol Lab Med 2007;131:777.
Gastroenterol Clin N Am 37 (2008) 369–381
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Eosinophilic Esophagitis in Children: Clinical Manifestations Philip E. Putnam, MD, FAAP Cincinnati Center for Eosinophilic Disorders, Division of Gastroenterology, Hepatology, and Nutrition, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Avenue, ML 2010, Cincinnati, OH 45229, USA
D
uring the past decade, the increasing number of recognized cases of eosinophilic esophagitis in children and adults has resulted in a dramatic expansion of the medical literature surrounding it. Clinical and basic research has contributed to a better, but still incomplete, body of knowledge regarding its clinical and histologic manifestations, as well as its immunologic and genetic pathogenesis. This article provides a broad framework for recognizing the remarkable variety of clinical manifestations of eosinophilic esophagitis in children. There is not a convenient ‘‘one size fits all’’ description of the presentation of eosinophilic esophagitis, and therefore physicians must consider it as part of the differential diagnosis in many different clinical situations. Prompt, accurate diagnosis is required for timely management, symptom control, and prevention of untoward sequelae from the disorder. Much of the medical literature related to eosinophilic esophagitis exists as case reports, case series, and reviews from centers in North America. The larger case series form the basis for the current understanding of the clinical manifestations of eosinophilic esophagitis in children and come from the tertiary referral centers that focus on diagnosis and management of pediatric eosinophilic esophagitis [1–6]. In addition, the recent publication of diagnostic guidelines will allow standardization for clinical care and future research studies [7]. Although most of these works were not designed primarily as epidemiologic studies, the demographics of the identified patients support several common findings. For reasons that remain unclear, approximately 70% of children who have eosinophilic esophagitis are males. A series in the United States also has confirmed the clinical impression that the majority (94.4%) of patients are white [2].
This article appeared previously in Gastrointestinal Endoscopy Clinics of North America 18:1, 2008; with permission.
E-mail address:
[email protected] 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.009
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
370
PUTNAM
In southwest Ohio, the incidence of eosinophilic esophagitis in children has been estimated at 1.25 cases per 10,000 children per year [6]. Formal epidemiologic studies encompassing the broader population are needed to understand more precisely the risk for developing eosinophilic esophagitis in North America and abroad. SYMPTOMS OF EOSINOPHILIC ESOPHAGITIS The most frequent symptoms of eosinophilic esophagitis are nonspecific and include dysphagia, pain, and vomiting. The age at which the symptoms develop varies considerably [6]. Feeding difficulty was the predominant reason for referral and evaluation of infants and toddlers ultimately diagnosed as having eosinophilic esophagitis. Preschool and school-aged children commonly complained of abdominal pain or vomiting, whereas dysphagia generally appeared as a primary complaint in the preadolescent years. In that analysis the chief complaint was recorded, but most patients had more than one symptom attributable to the disorder. Dysphagia Dysphagia caused by eosinophilic esophagitis is described variously by affected individuals and is notoriously intermittent. Age and ability to communicate effectively influence how a child describes his or her own difficulty swallowing. Some note that food ‘‘goes down slowly,’’ whereas others say that food gets stuck temporarily somewhere in the throat or chest before proceeding down. Some report difficulty initiating the swallow while the bolus is still in the mouth or recognize anxiety promoted by prior episodes of food impaction. Retrospective analyses note that dysphagia was present for more than 2 years before diagnosis in some patients [8,9]. As detailed in the article by Katzka, elsewhere in this issue, dysphagia has been the predominant presenting symptom in adults. A careful history in patients who complain of dysphagia reveals that they have learned to compensate by eating slowly, drinking after each bite, taking small bites, chewing excessively, or avoiding specific food consistencies that are problematic such as meat or bread. This phenomenon may help to explain the intermittent experience of dysphagia, because the failure to compensate at a particular meal or even for a single bite may allow the symptom to manifest and ‘‘reset’’ the compensatory mechanisms. Food impaction requiring endoscopic removal as the initial contact between an adolescent and the medical community is a fairly common presentation for eosinophilic esophagitis. Often the dysphagia has been long standing but mild and intermittent. Many individuals report years of symptoms that were ignored, compensated for, attributed to ‘‘taking too big a bite’’, or that were blamed on ‘‘not chewing food well enough.’’ The author and colleagues also have diagnosed eosinophilic esophagitis when called to remove from the esophagus impacted but nonobstructing foreign bodies that normally would have passed, such as coins. It is good practice to obtain mucosal biopsies from the esophagus, remote from the site of the
EOSINOPHILIC ESOPHAGITIS IN CHILDREN
371
impaction, after endoscopic removal of a foreign body, especially food impactions. Endoscopic evidence for eosinophilic esophagitis (thickening, furrowing, or pinpoint or diffuse white exudate) (Figs. 1 and 2) is frequently present in affected individuals at the time of the impaction, but because the endoscopic appearance may be influenced by the foreign body on the one hand, and because up to one third of patients who have eosinophilic esophagitis may have a normal-appearing esophagus at endoscopy on the other hand [4], mucosal biopsy is warranted after extracting a foreign body in children. Less common but clearly problematic, gastroenterologic consultation has been requested for some individuals after long periods of psychiatric evaluation and treatment of chronic dysphagia without radiographic evidence for dysmotility or stricture. The authors have seen children ultimately proved to have eosinophilic esophagitis whose dysphagia has been attributed to underlying anxiety, obsessive-compulsive disorder, eating disorder, or pervasive developmental disorder. One boy was diagnosed with bulimia at age 10 years because he vomited after eating, although he had symptoms since infancy. Dysphagia in patients who have eosinophilic esophagitis can be caused by mechanical obstruction of the esophagus and can manifest endoscopically and/or radiographically as a ring (Fig. 3) focal stricture, Schatzki ring, or long-segment narrowing (so-called ‘‘small-caliber esophagus’’) [10]. It is more common, however, for children to complain of dysphagia in the absence of radiographic narrowing of the esophagus [9]. Nurko and colleagues [8] examined the esophageal histology proximal to radiographically diagnosed Schatzki rings in 18 children and discovered that eosinophilic esophagitis was present in 8 of them. Notably, endoscopic evidence for Schatzki ring was not present in the children who had histologic eosinophilic esophagitis, suggesting that focal transient spasm accounted for the radiographic finding.
Fig. 1. Endoscopic appearance of furrowing.
372
PUTNAM
Fig. 2. Endoscopic appearance of exudate.
It is tempting to attribute dysphagia to the degree of inflammation, duration of disease, degree of lamina propria fibrosis, or the impact of chronic inflammation on deeper layers of the esophagus. The deep layers of the esophagus are affected, as evidenced by the thickening of the entire wall of the esophagus demonstrated by endoscopic ultrasound [11]. Nonspecific manometric abnormalities were observed in 60% of studied adult patients who had eosinophilic esophagitis [12]. In that study, parallel improvement in manometric and histologic findings was observed. It is tempting to speculate that adolescents and adults who have had longstanding untreated disease develop dysphagia as a consequence of persistent inflammation-induced damage including lamina propria fibrosis or remodeling, but because the age of disease onset often is unknown in that group, it has not been possible to associate dysphagia directly with the duration of inflammation in many individuals [13,14]. Stricture is a feared complication of eosinophilic esophagitis and has been described as early as 7 months of age [9]. Although the strictures are amenable to dilatation, esophageal mucosa with eosinophilic inflammation is more prone to tear during dilatation than the mucosa overlying a peptic stricture, and perforation of the esophagus has been described as a complication of attempted dilatation in adults [10,15,16]. Many clinicians managing adults who have eosinophilic esophagitis–associated dysphagia now recommend aggressive medical management of the inflammation before attempting dilatation unless clinical circumstances dictate more immediate relief of the obstruction [17]. Prospective study has not confirmed that this approach reduces the number of dilatations or the risk of perforation, but this approach is reasonable until refuted. Similarly, dogmatic historical concepts generated by experience in dilating peptic strictures should be abandoned and replaced by recommendations for gentle, gradual dilatation only when necessitated by failure of medical therapy for eosinophilic esophagitis.
EOSINOPHILIC ESOPHAGITIS IN CHILDREN
373
Fig. 3. (Top) Radiographic appearance and (bottom) endoscopic and radiographic appearance of the ringed proximal esophagus.
Infants and preverbal toddlers who have eosinophilic esophagitis may present with feeding difficulties that manifest as gagging, choking, food refusal/ aversion, or vomiting [18]. It is not well understood whether these problems are consequences of dysphagia, nausea, anorexia, pain, fear or anxiety after untoward experiences with eating, or a combination of these factors. Clinicians participating in the evaluation of young children who have feeding disorders must be knowledgeable in the evaluation and treatment options for eosinophilic esophagitis and have a low threshold for early evaluation and intervention. Because there are overlapping issues, concurrent therapy to account for all of the child’s conditions (gastroesophageal reflux disease [GERD], behavioral feeding aversion, and eosinophilic esophagitis) may be necessary.
374
PUTNAM
Vomiting Vomiting as a primary complaint among pediatric patients who have eosinophilic esophagitis occurs most often in young children. It can mimic reflux, manifest as effortless regurgitation, but eosinophilic esophagitis is seldom diagnosed in the first 6 months of life when GERD is common. Accurate diagnosis of eosinophilic esophagitis in very young children who vomit is often delayed by months to years, during which time treatment for symptoms attributed to GERD is attempted. In the author’s experience, parents describe vomiting more often than effortless regurgitation. Circumstances surrounding episodes of vomiting vary. Vomiting can be part of an overt and immediate allergic response to a recently ingested food. In that case, the family generally recognizes that there is a consistent reaction to that food and chooses to avoid it. Immediate vomiting may be associated with other manifestations of an adverse immune response such as hives, diarrhea, pain, or even anaphylaxis. Some children who have eosinophilic esophagitis and multiple food allergies have a history of developing different constellations of symptoms in response to various foods, up to and including food protein–induced enterocolitis syndrome. Chronic, episodic, unpredictable vomiting of variable severity also occurs in patients who have eosinophilic esophagitis, and it is not clinically distinguishable from other causes of emesis. This vomiting does not occur proximate to exposure to a particular food, which precludes devising an appropriate diet based solely on history. In the first year of life, clues that eosinophilic esophagitis, rather than reflux, is the underlying culprit include early onset of vomiting (within days to weeks of birth) associated with eczema, vomiting associated with irritability that does not respond to acid suppression, and onset of vomiting in the second half of the first year of life during the introduction of solid foods. After infancy, vomiting is more chronic, intermittent, and not readily associated with particular foods, although it is usually meal related. Vomiting to the point of excess calorie loss resulting in failure to gain weight or weight loss is possible but uncommon. Rarely, nausea without vomiting is the sole presenting manifestation of eosinophilic esophagitis. Although usually present with other symptoms, it has been an isolated complaint in a few individuals. Nausea has been responsive to standard treatment of the esophagitis in most individuals, but the author and colleagues have seen other children who have remarkably recalcitrant nausea as a persistent complaint after resolution of active eosinophilic inflammation. Retching to remove a piece of food that is stuck may be called ‘‘vomiting’’ by some patients and needs to be distinguished from emesis of gastric contents. Retching can be recognized as such by careful history. The patient will report that something got stuck and he or she retched to remove or otherwise dislodge it. Pain Pain is a frequent but not universal complaint among individuals who have untreated eosinophilic esophagitis. Children who have pain associated with
EOSINOPHILIC ESOPHAGITIS IN CHILDREN
375
eosinophilic esophagitis report chest, epigastric, and/or abdominal pain. Overt odynophagia is not typical. Some individuals have been evaluated the emergency department for episodic crushing substernal pain concerning for heart disease. Those individuals have been remarkably free of symptoms between pain attacks, which have been unpredictable in frequency and which lack an obvious inciting event. The pain may be described as heartburn and may respond to acid suppression. When histologic eosinophilic inflammation persists despite effective therapy for GERD, a clinical diagnosis of eosinophilic esophagitis is established. A remarkable number of children indicate that their pain is in the epigastrium or the periumbilical area. They often have difficulty describing the pain in more sophisticated terms. It is imperative that these complaints promote appropriate investigation, particularly when the broader clinical picture is concerning, such as a boy who has asthma or eczema or a history of food allergy. The absence of complaints referable to the chest should not preclude consideration of esophageal disease. Associated Conditions Communicative children who have eosinophilic esophagitis may present with a single symptom, but careful questioning often elicits additional symptoms that are attributable to the inflammation. In addition, many children who have eosinophilic esophagitis have other medical conditions with symptoms that may overlap, contribute to, or serve to confuse the issue at presentation. For example, chest pain may accompany episodes of bronchospasm in a child who has asthma, and that phenomenon may divert attention away from other potential causes of chest pain, such as esophagitis (reflux or otherwise). The profile of individuals who have eosinophilic esophagitis is quite varied, but there are notable associations. Because definitions and rigor in establishing a formal diagnosis vary, the precise frequency of associated conditions is subject to interpretation. Nevertheless, eosinophilic esophagitis coexists or is a parallel phenomenon associated with asthma, eczema, and environmental and food allergies in up to 75% of children [2,19]. The association of eosinophilic esophagitis with other conditions is becoming more apparent from experience. The epidemiology and immunobiology are as yet unexplored. For example, esophageal eosinophilia has been seen in the Cincinnati Center for Eosinophilic Disorders in a handful of children who have other gastrointestinal conditions such as Helicobacter pylori gastritis, Crohn’s disease, or celiac disease. Esophageal involvement can be seen in patients who have eosinophilic gastroenteritis. It is inappropriate to make a clinical diagnosis of eosinophilic esophagitis when substantial eosinophilic inflammation of stomach, small intestinal, or colon is present concurrently or sporadically in a patient who has had eosinophilic infiltration of the esophagus. Studies have examined the number of eosinophils in normal gastrointestinal mucosa, but formal histologic and clinical definitions of eosinophilic gastroenteritis are yet to be developed [20,21]. As
376
PUTNAM
such, separating patients who have eosinophilic esophagitis who happen to have mild and clinically irrelevant mucosal eosinophilia distal to the esophagus from cases of eosinophilic gastroenteritis with esophageal involvement can be challenging. Diarrhea, anemia, hypoalbuminemia, failure to thrive, and gastrointestinal bleeding are not typical features of isolated eosinophilic esophagitis, and individuals who display them should be evaluated appropriately, even if a provisional diagnosis of eosinophilic esophagitis has been made. A number of children who have underlying neurologic or neurodevelopmental conditions have been diagnosed as having eosinophilic esophagitis. Eosinophilic esophagitis has been seen in children who also have intractable seizures, cerebral palsy, Chiari malformation, pervasive developmental disorder, sensory integration disorder, or migraine. Hypersensitivity to antiepileptic drugs has been implicated in the development of eosinophilic esophagitis, so it is important to account for medication use in this group [22]. Otherwise, no obvious, direct, cause-and-effect relationship between these neurologic conditions and eosinophilic esophagitis has been discerned to date. The epidemiology of eosinophilic esophagitis in children who have concurrent neurodevelopmental conditions has not been explored formally. It is not clear whether the prevalence of eosinophilic esophagitis is indeed higher in that population of children or whether clinicians have become more thorough in evaluating them. GERD is said to occur frequently in this population as well, but it is no longer safe to assume that GERD is responsible for nonspecific upper gastrointestinal symptoms. Similarly, eosinophilic esophagitis has been discovered in children who have other syndromes, without comment on the epidemiology. A case report notes the association of eosinophilic esophagitis in Rubenstein-Taybi syndrome [23]. The author and colleagues have seen eosinophilic esophagitis in children who have coloboma, heart anomalies, choanal atresia, retardation of growth and development, and genital and ear anomalies (CHARGE syndrome), vertebral, anal, tracheo-esophageal, and radial limb or renal anomalies (VATER syndrome), Pierre-Robin syndrome, Klinefelter’s syndrome, Moebius syndrome, and Pfeiffer syndrome. EOSINOPHILIC ESOPHAGITIS AND DISEASE OF THE RESPIRATORY TRACT Eosinophilic esophagitis has been discovered in children presenting with airway and respiratory complaints, who may or may not have concomitant gastrointestinal symptoms [24,25]. Although the prevalence of eosinophilic esophagitis in children presenting with primary airway disease is unknown, it must be considered in children who present with stridor, adenoidal hypertrophy, recurrent croup, recurrent pneumonia, or aspiration. The diagnosis has been made in children who have airway anomalies such as subglottic stenosis or laryngeal cleft [26]. No criteria yet exist on which to base recommendations for formal diagnostic evaluation for eosinophilic esophagitis in patients who have these presenting
EOSINOPHILIC ESOPHAGITIS IN CHILDREN
377
symptoms. At the Cincinnati Center for Eosinophilic Disorders, more than 50 children who have eosinophilic esophagitis have been seen in the tertiary, multidisciplinary clinic dedicated to the evaluation and treatment of significant airway problems resulting in tracheostomy tube dependence. Most have had subglottic stenosis caused by prolonged intubation after premature birth, and none were suspected of having or diagnosed as having eosinophilic esophagitis before evaluation in this institution. Many had failed prior attempts at surgical laryngotracheoplasty. Because the mechanisms involved are not established, it is not clear whether there is any direct relationship between the esophagitis and the recurrent stenosis, but the association and concern has been sufficient for the author and colleagues to consider evaluation for and treatment of eosinophilic esophagitis in this population warranted before committing to elective airway surgery until prospective studies further examine the pathogenesis. DO DIFFERENT PHENOTYPES OF EOSINOPHILIC ESOPHAGITIS EXIST? Studies that describe the symptom complex in children who have eosinophilic esophagitis have generally lumped all individuals with defined histology together, and the disorder is considered as a single entity. Those who have followed a large number of children with eosinophilic esophagitis, however, suspect that clinically distinct patterns occur. Whether these presentations are simply at opposite ends of a continuum or deserve a separate diagnosis or classification scheme is not established. A particular single nucleotide polymorphism in the eotaxin-3 gene is present in some individuals with eosinophilic esophagitis, but it is not clear how that information can or should be used in clinical practice at this time [27]. In clinical practice, it is possible to categorize children loosely into clinical subtypes or phenotypes. For example, there is a group of children in which the individuals are extraordinarily atopic and exhibit a constellation of eosinophilic esophagitis, eczema, chronic rhinitis, asthma, repeated respiratory infections, and multiple food allergies. These children often present in early infancy with vomiting and irritability. Eczema, chronic nasal congestion, recurrent bronchospasm, and repeated upper respiratory infection become apparent in the first 2 years of life. In some, the eczema has been severe. Although the gastrointestinal symptoms respond to dietary restriction (most often requiring an elemental diet with an amino acid-based formula), the asthma is generally an independent phenomenon requiring standard therapy and does not resolve as a result of diet change alone. In this group, a clinical diagnosis of GERD has been made almost universally, but the children do not respond completely to antireflux therapy. Some of these children have been extraordinarily sensitive to food and exhibit recurrent eosinophilic esophagitis with virtually any food added to the diet. At the other end of the spectrum is group of children who have eosinophilic esophagitis in the absence of atopy, eczema, or asthma. Evidence for food allergies is lacking by formal allergy testing, and the children do not respond to
378
PUTNAM
dietary therapy even with an elemental diet. This group of patients generally responds well to topical (swallowed) steroids (usually fluticasone or budesonide) as mainstay of chronic therapy. Recurrence after discontinuation of steroids is nearly universal, although follow-up of these patients is still relatively brief (<10 years for children). If these presentations are considered as a spectrum, the middle of the spectrum is populated by children who have eosinophilic esophagitis and food allergies, often with asthma, and/or mild eczema. They develop eosinophilic esophagitis with several foods but can be maintained on an elimination diet, avoiding those foods that trigger the inflammation. This group seems to be large and to incorporate the majority of children who have eosinophilic esophagitis who have been evaluated in the Cincinnati Center for Eosinophilic Disorders. FAMILIAL EOSINOPHILIC ESOPHAGITIS Eosinophilic esophagitis has been demonstrated in more than one individual within many families [28]. The occurrence across generations suggests a genetic rather than environmental pathogenesis, but only now are the details being explored. Careful assessment of the family history may identify other individuals who already have been diagnosed, but the author and colleagues frequently hear about adults with longstanding dysphagia, perhaps requiring dilatation, attributed to reflux or Schatzki ring. Siblings who exhibit symptoms deserve early endoscopic evaluation to establish a diagnosis. RESPONSE TO THERAPY Much of what has been learned about children who have eosinophilic esophagitis has come directly from studies designed to observe the response of the inflammatory process to therapy, be it dietary antigen elimination or steroids [4,19,29–33]. By convention, eosinophilic esophagitis is distinguished from reflux by the failure to respond to aggressive antireflux therapy. In the landmark description of eosinophilic esophagitis in children, the complete response of otherwise recalcitrant eosinophilic esophagitis to an elemental diet was a defining feature [34]. Now, the same parameters are used to make a formal diagnosis (ie, failure to respond to antireflux therapy) and to distinguish the food allergic variant (ie, response to diet). Specifically, complete clinical and histologic resolution of active eosinophilic esophagitis in response to diet therapy either with specific antigen avoidance or total antigen elimination (eg, diet replaced by an elemental formula) supports a diagnosis of the allergic variant of eosinophilic esophagitis. In theory, patients who do not respond to an elemental diet still may have an allergic variant, although the assumed nonfood environmental antigens involved remain unclear. Because up to 98% of children who have eosinophilic esophagitis respond to an elemental diet, the vast majority of affected children would seem to have the disease as a manifestation of adverse response to dietary antigens [32].
EOSINOPHILIC ESOPHAGITIS IN CHILDREN
379
CONTROVERSIES IN EOSINOPHILIC ESOPHAGITIS One of the great difficulties in managing children who have eosinophilic esophagitis has been the absence of symptoms in treated individuals who nevertheless have active esophageal eosinophilia. Symptoms typically abate with therapy (diet or medication), but some patients still have unaltered or only minimally improved esophageal histology. Another common scenario is for the esophagitis to return without overt symptoms during attempts at reintroducing food to the diet or upon withdrawal of steroid therapy [2,4]. This phenomenon has generated considerable debate as to whether the symptom(s) or the histology should be treated. Because only endoscopy with biopsy has been an adequate method for reassessment in the absence of symptoms, one of the concerns has been the risk and cost of the number of endoscopies that need to be performed during the lifetime of an individual undergoing food elimination/reintroduction. Early work that examined genetic, biochemical, and/or immunologic markers for the disease failed to discover factors with sufficient predictive value to be used routinely in place of endoscopy and histology [27,35,36]. The debate has not been resolved officially, although expert consensus now supports treating the histology irrespective of the absence of symptoms, because of the known longer-term risks of esophageal remodeling, fibrosis, and potential stricture formation [17]. References [1] Orenstein SR, Shalaby TM, Di Lorenzo C, et al. The spectrum of pediatric eosinophilic esophagitis beyond infancy: a clinical series of 30 children. Am J Gastroenterol 2000;95(6):1422–30. [2] Assa’ad AH, Putnam PE, Collins MH, et al. Pediatric patients with eosinophilic esophagitis: an 8-year follow-up. J Allergy Clin Immunol 2007;119(3):731–8. [3] Aceves SS, Newbury RO, Dohil R, et al. Distinguishing eosinophilic esophagitis in pediatric patients: clinical, endoscopic, and histologic features of an emerging disorder. J Clin Gastroenterol 2007;41(3):252–6. [4] Liacouras CA, Spergel JM, Ruchelli E, et al. Eosinophilic esophagitis: a 10-year experience in 381 children. Clin Gastroenterol Hepatol 2005;3(12):1198–206. [5] Fox VL, Nurko S, Furuta GT. Eosinophilic esophagitis: it’s not just kid’s stuff. Gastrointest Endosc 2002;56(2):260–70. [6] Noel RJ, Putnam PE, Rothenberg ME. Eosinophilic esophagitis. N Engl J Med 2004;351(9): 940–1. [7] Furuta GT, Liacouras CA, Collins MH, et al. Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment. Gastroenterology 2007;133(4):1342–63. [8] Nurko S, Teitelbaum JE, Husain K, et al. Association of Schatzki ring with eosinophilic esophagitis in children. J Pediatr Gastroenterol Nutr 2004;38(4):436–41. [9] Khan S, Orenstein SR, Di Lorenzo C, et al. Eosinophilic esophagitis: strictures, impactions, dysphagia. Dig Dis Sci 2003;48(1):22–9. [10] Vasilopoulos S, Murphy P, Auerbach A, et al. The small-caliber esophagus: an unappreciated cause of dysphagia for solids in patients with eosinophilic esophagitis. Gastrointest Endosc 2002;55(1):99–106. [11] Fox VL, Nurko S, Teitelbaum JE, et al. High-resolution EUS in children with eosinophilic ‘‘allergic’’ esophagitis. Gastrointest Endosc 2003;57(1):30–6.
380
PUTNAM
[12] Lucendo AJ, Castillo P, Martı´n-Cha´varri S, et al. Manometric findings in adult eosinophilic oesophagitis: a study of 12 cases. Eur J Gastroenterol Hepatol 2007;19(5):417–24. [13] Aceves SS, Newbury RO, Dohil R, et al. Esophageal remodeling in pediatric eosinophilic esophagitis. J Allergy Clin Immunol 2007;119(1):206–12. [14] Straumann A, Spichtin HP, Grize L, et al. Natural history of primary eosinophilic esophagitis: a follow-up of 30 adult patients for up to 11.5 years. Gastroenterology 2003;125(6):1660–9. [15] Eisenbach C, Merle U, Schirmacher P, et al. Perforation of the esophagus after dilation treatment for dysphagia in a patient with eosinophilic esophagitis. Endoscopy 2006 [Epub ahead of print]. [16] Lucendo AJ, De Rezende L. Endoscopic dilation in eosinophilic esophagitis: a treatment strategy associated with a high risk of perforation. Endoscopy 2007;39(4):376. [17] Liacouras CA, Bonis P, Putnam P, et al. FIGER summary. J Pediatr Gastro Nutr 2007;45: 370–91. [18] Pentiuk SP, Miller CK, Kaul A. Eosinophilic esophagitis in infants and toddlers. Dysphagia 2007;22(1):44–8, Epub 2006 Oct 6. [19] Spergel JM, Andrews T, Brown-Whitehorn TF, et al. Treatment of eosinophilic esophagitis with specific food elimination diet directed by a combination of skin prick and patch tests. Ann Allergy Asthma Immunol 2005;95(4):336–43. [20] DeBrosse CW, Case JW, Putnam PE, et al. Quantity and distribution of eosinophils in the gastrointestinal tract of children. Pediatr Dev Pathol 2006;9(3):210–8. [21] Lowichik A, Weinberg AG. A quantitative evaluation of mucosal eosinophils in the pediatric gastrointestinal tract. Mod Pathol 1996;9(2):110–4. [22] Balatsinou C, Milano A, Caldarella MP, et al. Eosinophilic esophagitis is a component of the anticonvulsant hypersensitivity syndrome: description of two cases. Dig Liver Dis 2007 [Epub ahead of print]. [23] Noble A, Drouin E, Faure C. Eosinophilic esophagitis and gastritis in Rubinstein-Taybi syndrome. J Pediatr Gastroenterol Nutr 2007;44(4):498–500. [24] Mandell DL, Yellon RF. Synchronous airway lesions and esophagitis in young patients undergoing adenoidectomy. Arch Otolaryngol Head Neck Surg 2007;133(4):375–8, 13. [25] Dauer EH, Ponikau JU, Smyrk TC, et al. Airway manifestations of pediatric eosinophilic esophagitis: a clinical and histopathologic report of an emerging association. Ann Otol Rhinol Laryngol 2006;115(7):507–17. [26] Goldstein NA, Putnam PE, Dohar JE. Laryngeal cleft and eosinophilic gastroenteritis: report of 2 cases. Arch Otolaryngol Head Neck Surg 2000;126(2):227–30. [27] Blanchard C, Wang N, Stringer KF, et al. Eotaxin-3 and a uniquely conserved gene-expression profile in eosinophilic esophagitis. J Clin Invest 2006;116(2):536–47. [28] Zink DA, Amin M, Gebara S, et al. Familial dysphagia and eosinophilia. Gastrointest Endosc 2007;65(2):330–4. [29] Konikoff MR, Noel RJ, Blanchard C, et al. A randomized double-blind placebo-controlled trial of fluticasone propionate for pediatric eosinophilic esophagitis. Gastroenterology 2006;131(5):1381–91. [30] Noel RJ, Putnam PE, Collins MH, et al. Clinical and Immunopathologic effects of swallowed Fluticasone for eosinophilic esophagitis. Clin Gastroenterol Hepatol 2004;2(7):568–75. [31] Kagalwalla AF, Shah A, Ritz S, et al. Cow’s milk protein-induced eosinophilic esophagitis in a child with gluten-sensitive enteropathy. J Pediatr Gastroenterol Nutr 2007;44(3):386–8. [32] Markowitz JE, Spergel JM, Ruchelli E, et al. Elemental diet is an effective treatment for eosinophilic esophagitis in children and adolescents. Am J Gastroenterol 2003;98(4):777–82. [33] Teitelbaum JE, Fox VL, Twarog FJ, et al. Eosinophilic esophagitis in children: immunopathological analysis and response to fluticasone propionate. Gastroenterology 2002;122(5): 1216–25. [34] Kelly KJ, Lazenby AJ, Rowe PC, et al. Eosinophilic esophagitis attributed to gastroesophageal reflux: improvement with an amino acid-based formula. Gastroenterology 1995;109(5):1503–12.
EOSINOPHILIC ESOPHAGITIS IN CHILDREN
381
[35] Konikoff MR, Blanchard C, Kirby C, et al. Potential of blood eosinophils, eosinophil-derived neurotoxin, and eotaxin-3 as biomarkers of eosinophilic esophagitis. Clin Gastroenterol Hepatol 2006;4(11):1328–36, Epub 2006 Oct 23. [36] Gupta SK, Fitzgerald JF, Kondratyuk T, et al. Cytokine expression in normal and inflamed esophageal mucosa: a study into the pathogenesis of allergic eosinophilic esophagitis. J Pediatr Gastroenterol Nutr 2006;42(1):22–6.
Gastroenterol Clin N Am 37 (2008) 383–395
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Functional Gastrointestinal Disorders and the Potential Role of Eosinophils Marjorie M. Walker, BMedSci, BMBS, FRCPatha,*, Nicholas J. Talley, MD, PhD, FRACP, FRCP, FACPb a
Department of Histopathology, St. Mary’s Campus, Imperial College London, Norfolk Place, London W2 1PG, United Kingdom b Department of Internal Medicine, Mayo Clinic, 4500 San Pablo Road, Jacksonville, FL 32224, USA
FUNCTIONAL GASTROINTESTINAL DISORDERS The Rome III classification of the functional gastrointestinal disorders (FGIDs) defines 28 adult and 17 pediatric symptom-based gut conditions that have no established pathology [1]. These disorders place a considerable burden on individual patients, because quality of life is poorer in patients who have FGIDs compared with health and is comparable to organic GI disease [2]. There is also a major societal burden because of high direct and indirect costs [3]. Currently, diagnosis is based on symptoms or by exclusion, because there is no objective diagnostic test for these disorders; by definition it is necessary to exclude other diseases, specifically evidence of an inflammatory, anatomic, metabolic, or neoplastic process that explains the patient’s symptoms [1]. Abdominal pain, disordered defecation, and meal-related discomfort are common symptomatic themes for a range of FGIDs. Treatment remains empirically based on relieving symptoms rather than treating abnormal pathophysiology. The current accepted conceptual model of disease reflects an interaction between psychosocial factors, including stress and anxiety, that impinge on gut physiology by way of the brain–gut axis to produce visceral hypersensitivity or dysmotility. These signals interact by way of the gut enteric nervous system (ENS) and track back to the central nervous system (CNS) [1]. It is possible that this viscious cycle originates in gut mucosa, where an insult triggers a feedback loop, although how often this is the case remains uncertain. Key barriers in attempting to define an organic pathology for FGIDs include a lack of understanding of how symptoms are triggered and how any feedback loop is maintained. The pathophysiology of the FGIDs has been studied extensively to try to identify a common link for symptom expression. It is currently assumed that *Corresponding author. E-mail address:
[email protected] (M.M. Walker). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.007
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
384
WALKER & TALLEY
there is not a simple or universal pathway to these differing outcomes, but a complex interaction of genetic predisposition, visceral hypersensitivity, and psychosocial distress in a primed central and enteric nervous system [4,5]. EOSINOPHILS AND FUNCTIONAL GASTROINTESTINAL DISORDERS The gut is an active immune organ that is the first line of defense for many insults, such as allergens in food, drugs, viruses, bacteria, and parasites. The authors surmise that altered immune function, by way of infection or hypersensitivity on a background of genetic predisposition and environment, triggers symptoms (Fig. 1). This hypothesis suggests that components of the hypersensitivity immune cascade may play a vital role in FGIDs. A key component of this pathway is the eosinophil, and this review explores a potential role for eosinophils and associated components of this pathway in FGIDs, specifically irritable bowel syndrome (IBS) and functional dyspepsia (FD), which have been principally studied. EOSINOPHILS IN THE GASTROINTESTINAL TRACT Eosinophils are a normal constituent throughout the gastrointestinal tract (GIT), except in the squamous mucosa of the esophagus. These cells have a defined role in physiologic and pathologic function. In the gut mature eosinophils originate from hemopoietic stem cells in the bone marrow; recruitment is regulated by interleukin (IL)–5, IL-3, and granulocyte macrophage colony-stimulating factor (GM-CSF) to induce migration to the gut. Eosinophils are effector cells, which release secretory granules and have an immunoregulatory function by way of cytokines and release and presentation of antigens [6]. WHAT IS THE NORMAL EOSINOPHIL COUNT IN THE GUT? Eosinophils have a characteristic morphology with a bright eosinophilic cytoplasm and a bilobed nucleus. They are easily recognizable in biopsies and
Fig. 1. A role for eosinophils in FGIDs? The interaction of psychosocial distress, environmental factors, and a genetic predisposition with a hypersensitive immune system in the gut may be maintained by eosinophil accumulation in the gut, by way of signaling to the ENS and then the CNS in functional bowel disorders.
FUNCTIONAL GASTROINTESTINAL DISORDERS AND EOSINOPHILS
385
therefore counts of eosinophils can be performed on routine hematoxylin and eosin stained sections. Counting of eosinophils is critical at all sites to discriminate normal from a pathologic state. In the esophagus, eosinophils are evoked in acid reflux and to distinguish this from eosinophilic esophagitis, the cutoff level is conventionally considered to be at least 15 per high power field (HPF), although variable cutoffs have been used [7]. In gastric mucosa, eosinophilia is often seen in Helicobacter pylori infection but typically not in functional dyspepsia [8]. In H pylori–negative adult subjects, the mean gastric eosinophil counts were 10/5HPF in antral and body mucosa [8]. In the duodenum, eosinophilia has been noted in children and adults who have functional dyspepsia, defined as greater than 10/HPF for children [9] and greater than 22/5HPF with or without clusters of eosinophils at the base of glands for adults [8]. Normal duodenal counts were defined by less than 10/HPF in children and 19/5HPF in adults in these studies based on control values [8,9]. In the cecum, eosinophils are more numerous in the normal state, and up to 40/HPF have been suggested to be normal at this site [10]. Colonic eosinophilia has also been associated with IBS, inflammatory bowel disease (IBD), and food allergy in children, and mean eosinophil numbers per crypt (standard deviation) have been reported to be 34.8 (17.1) in IBD, 21.3 (8.8) in IBS, and 25.4 (16.7) in food allergy [11]. In this study there was no variation between colonic sites. Geographic variation in colonic eosinophilia in routine biopsies, in which eosinophils in the intercryptal space were counted, has been noted in the United States, where those in southern states had a higher count than the northern states [12]. Few studies have consistently counted in the same manner in normal or diseased subjects so establishing true normal values is difficult. There are relatively few studies to adequately address the normal counts throughout the gut, particularly the upper gut. Because subtle changes in numbers may be significant [13], it is important therefore for histopathologists to count eosinophils in the gastrointestinal sites and have data on local normal controls to define abnormality. EOSINOPHILS, ATOPY, ALLERGY, AND FUNCTIONAL GASTROINTESTINAL DISORDERS It has been recognized that gastrointestinal symptoms often occur in children who have allergy and atopy [14,15], and IBS is significantly more prevalent in adult patients who have allergic diseases, such as asthma and allergic rhinitis [16]. Key cells in the allergy cascade are eosinophils. Small bowel biopsies from patients who have asthma and allergic rhinitis show accumulation of eosinophils, T-cells, mast cells, macrophages, and increased expression of the proallergic cytokines IL-4 and IL-5, which are also observed in the inflammatory reaction in the airways [17–19]. Allergy to foodstuffs is a common patient perception. In a study from Germany, 32% of patients attending outpatients complained of adverse reactions to food as a cause of their abdominal symptoms. In 14%, the diagnosis
386
WALKER & TALLEY
of intestinal food allergy was suspected by an elevated total IgE, specific IgE against food antigens, peripheral eosinophilia, responsiveness to cromoglycate, and clinical signs of atopic disease. In only 3% (one tenth of the original cohort), however, was the diagnosis confirmed by endoscopic allergen provocation or elimination diet and re-challenge [20]. Eosinophils in the lower intestine of patients who have documented food allergy show characteristic features with regard to morphology, distribution, and functional behavior to IgE receptor stimulation. By immunohistochemical analysis of eosinophil peroxidase, there was a highly significant increase of eosinophils throughout the lower GI tract in samples obtained from patients who had food hypersensitivity–related diarrhea compared with control subjects [21]. Although there is reasonable evidence for a role for food allergy in provoking GI symptoms in eosinophilic gastroenteritides, there is emerging evidence that food allergy may be also important in FGIDs. A recent systematic review of this subject concluded that food allergy may be a factor contributing to the development of some FGIDs and food-specific IgE antibodies can mediate hypersensitivity reactions in patients who have IBS by sensitizing mucosal mast cells [22]. It has also been shown in clinical trials with exclusion diets that there is the potential to reduce symptoms in a subgroup of patients who have IBS with a clear history of adverse food reactions. These specific foods included wheat, beef, pork, and lamb in a study from the United Kingdom [23]. A second study from Australia of patients who had IBS and FD suggested wheat, crab, and shrimp allergy in IBS, whereas patients who had FD had high antibody titers to soybean and egg [24]. It was concluded that these findings did not contribute to the pathogenesis of FD and IBS; however, the method did not include biopsy of the GI tract to evaluate eosinophils and mast cells. EOSINOPHILS, INFECTION, AND FUNCTIONAL GASTROINTESTINAL DISORDERS Bacterial infections have been implicated as a trigger of FGIDs, particularly in the development of IBS, in which a specific postinfectious type is best recognized [25]. Postinfectious FD also seems probable, because acute onset of FD has been described following Salmonella infection in Spain [26]. Another study noted that 17% of patients who had acute-onset FD describe a preceding infective episode [27]. The bacterial species responsible for postinfectious IBS include Salmonella, Shigella, and Campylobacter jejuni [28]. These patients have an initial response with an increase in the CD3 and CD8 T lymphocyte population and a striking increase in serotonin containing enterochromaffin cells [29]. Although these changes may decline in most subjects, those who have persistent symptoms may continue to have persistent but subtle cellular changes. Patients who develop postinfectious IBS when assessed up to 3 months following infection still had a 20% increase in enterochromaffin cells, which contain serotonin [29].
FUNCTIONAL GASTROINTESTINAL DISORDERS AND EOSINOPHILS
387
In FD, meta-analyses of trials show a small benefit of eradication of the gastric bacteria H pylori. In dyspepsia in the absence of alarm features and in FD, a policy of test and treat for H pylori is generally recommended; the number needed to successfully treat FD is 15, which suggests that this is not the sole cause of FD [30] and no consistent disturbance of upper GI motility or sensory disturbances have been found to be linked to H pylori infection [31]. Few studies have addressed the immune pathways in FGIDs. Notably, in eosinophil infiltration can be a major component after bacterial infection. Gastric eosinophil infiltration has been reported to be significantly higher in patients who have serum CagA positivity for H pylori [32]. A study of subjects who have FD showed gastric eosinophilia was predominant only in H pylori infection [8]. In colonic and rectal biopsies of patients who had IBS, eosinophils were not noted to be increased compared with controls, although these were not typed as to whether these were postinfection cases [33]. Although not all patients who have FGIDs have overt atopy, allergy, or history of infection, we have postulated that this type of inflammatory mechanism, and in particular the eosinophil, may provide potential answers to the pathophysiology of certain FGIDs [8]. EOSINOPHILS IN THE DUODENUM IN FUNCTIONAL DYSPEPSIA AND IRRITABLE BOWEL SYNDROME It is postulated that the pathophysiology of FD is linked to duodenal dysmotility with distension [34] and sensitivity to luminal contents [35]. Biopsy of the duodenum to ascertain pathology is, however, not routine practice and there are few studies that have focused on duodenal histology as a potential target for investigation. A study of pediatric patients who had dyspepsia with duodenal biopsies found that in 71% had duodenal eosinophilia. Following treatment with H1 and H2 antagonist therapy, on repeat biopsy this was reduced and abdominal pain decreased or disappeared [9]. We have shown that duodenal eosinophilia is associated with FD in adults [8]. In our population-based study of 1000 normal subjects, not actively seeking health care, the prevalence of reflux symptoms, dyspeptic symptoms, and irritable bowel symptoms was 40%, 38%, and 30% respectively [36]. There was a significant duodenal eosinophilia in subjects who had FD compared with controls and it was also noted in subjects who had FD that eosinophils formed clusters at the base of duodenal mucosal crypts [8]. These clusters of eosinophils in the duodenal bulb and second part were present more often in subjects who had FD compared with controls (respectively 51% versus 21%; 62% versus 12.5%). Furthermore, degranulation of eosinophils was seen only in subjects who had FD in a subset analysis (7 of 15 cases, 47% versus none in the 5 controls). In subjects who had IBS from the same cohort, however, duodenal eosinophilia was not observed, suggesting that duodenal eosinophilia is linked to FD but not IBS [37]. EOSINOPHIL FUNCTION IN THE GUT Eosinophils store active immunologic mediators in two intracellular compartments, large granules and small secretory granules. These active mediators
388
WALKER & TALLEY
are cationic proteins, growth factors, cytokines, chemokines, lipid mediators, and neuromodulators. The large granules contain four primary cationic proteins: eosinophil peroxidase (EPO), major basic protein (MBP), eosinophil cationic protein (ECP), and eosinophil-derived neurotoxin (EDN); these are cytotoxic secretory products [38]. The ability of eosinophils to show diversity in disease, for example by differing modes of degranulation [38], may be key to a role in the FGIDs. In a study of FD, eosinophils accumulated at the base of crypts and degranulation of MBP was shown by immunocytochemistry in those who had dyspepsia, but not in control subjects [8]. The interaction of eosinophils with other cells in the hypersensitivity cascade may be important in their role in FGIDs. In allergic inflammation, a combination of Th2 cell and cytokines (IL-4, IL-5, and IL13) and mast cell activity seems to be responsible for the initiation of eosinophil activity [39]. This response is sustained because it has been shown that tissue eosinophils can survive for 2 weeks in vitro [40]. IL5 and eotaxin selectively regulate eosinophil trafficking [41]. The Th2 cytokines (tumor necrosis factor, IL4, and IL13) up-regulate eotaxin production [42]. Eotaxin is produced locally and recruits eosinophils to the intestine with accumulation in Peyer patches [43]. Although eotaxin-1 has been noted to play a role in eosinophilic esophagitis in mice, eotaxin-1–deficient mice only developed modestly attenuated disease [44,45]. Esophageal eotaxin-3 mRNA and protein levels strongly correlated with tissue eosinophilia and mastocytosis in eosinophilic esophagitis in human subjects who have characteristic gene arrays, and a single-nucleotide polymorphism in the human eotaxin-3 gene was associated with disease susceptibility [46]. This susceptibility to eosinophil accumulation because of a genetic predisposition may provide a link to a genetic susceptibility to FGIDs. Twin studies support a genetic predisposition to FGIDs, although FD has not been adequately evaluated [47–49]. An association between the CC genotype of GNb3 and functional dyspepsia has been observed in two independent studies, which may reflect an altered immunophenotype in subjects prone to dyspepsia [50–52]. Eosinophils also secrete growth factors, including nerve growth factor (NGF), which maintains sympathetic neurons and promotes mast cell survival and activation. NGF is also produced by mast cells and has a role in resolution of tissue healing. [53]. Other major mediators secreted by eosinophils are the lipid mediators, leukotrienes (LTC) and platelet-activating factors (PAF), chemokines, and cytokines [54]. EOSINOPHIL INTERACTIONS IN GUT MUCOSA Eosinophils and T Cells Eosinophils communicate with T cells in a bidirectional manner and activate T cells by serving as antigen-presenting cells (APCs). Eosinophils can also regulate T-cell polarization through synthesis of indoleamine 2,3-dioxygenase (IDO). IDO catalyzes the oxidative metabolism of L-tryptophan to N-formylkynurenine (KYN), a regulator of Th1/Th2 balance [54]. KYN is subsequently
FUNCTIONAL GASTROINTESTINAL DISORDERS AND EOSINOPHILS
389
converted to a range of metabolites, including xanthurenic acid and the neurotoxin quinolinic acid. IDO activity is induced in response to interferon-c and requires heme as a cofactor. Recent studies have suggested that, in vivo, IDO expression by APCs is responsible for inhibiting T-cell proliferation, and may be involved in immune tolerance [55]. In a study of subjects who had FD, the number of duodenal intraepithelial lymphocytes (IELs) was similar in patients who had FD and healthy controls. A careful analysis of activation markers on IELs showed CD95/Fas and HLADR were expressed by a significantly smaller percentage of duodenal CD3þ IELs in H pylori–negative patients who had FD defined according to Rome II criteria [56]. These lower ratios of CD95/Fas and HLA-DR expressing duodenal IELs in dyspeptic patients may reflect the presence of an altered population of primed lymphocytes [57]. Possibly these may be regulated by the increased eosinophils as seen in FD [8]. Eosinophils and Mast Cells Mast cells are bone marrow–derived granulocytes, also normally present in the small bowel, which contribute to a variety of immune responses. Mast cells are activated in stress [58,59], without allergic degranulation [60]. This finding may provide an all-important link with stress-related symptoms of FGIDs. Mast cells have been studied in FGIDs, particularly IBS, wherein they are increased in the jejunum [61], ileum [62], and cecum [33], and lie in close proximity to enteric nerves in colonic and rectal mucosa [63]. In FD, gastric mast cells may be increased in both H pylori–negative and –positive patients [64]. Mast cells can be activated through eosinophil-derived MBP by way of a similar pathway to the neuropeptide substance P mast cell activation pathway. MBP and substance P may both directly activate the mast cell Gi3 protein that controls degranulation [65]. Eosinophils express various growth factors, including stem cell factor (SCF), which promote mast cell survival, proliferation, maturation, and mediator release. It is speculated that the production of SCF by eosinophils is stimulated by mast cell chymase [66]; eosinophil-derived NGF further influences mast cell growth and survival by way of a specific TrkA NGF receptor expressed by mast cells [67]. Mast cells and eosinophils are thus codependent; initially mast cells can induce eosinophils into the mucosa, and these in turn activate mast cells by way of mediators and growth factors, causing proliferation, maturation, and degranulation. Mast cells in the presence of eosinophil-derived growth factors demonstrate resilience to death induced by death receptors [68]. The mast cell has been principally studied in IBS. The mast cells are not only increased but show greater proximity to nerves, and this has been positively correlated with the severity and frequency of patients’ abdominal pain [69]. Mast cells express receptors for the endocrine and paracrine messengers released from associated nerve endings and also have the ability to produce, store, metabolize, and release relevant molecules that modulate neural responses [69]. Substance P, released from nerve endings, activates mast cells.
390
WALKER & TALLEY
The mast cells release a mediator, which in turn results in neural activation, promoting intestinal secretion [70]. Nonactivated human mast cells that have not formed synaptic associations with nerve fibers do not respond to neuropeptides, such as substance P [71]. An interaction between mast cells and the CNS in rats has been demonstrated by pairing an audiovisual cue with an antigen challenge, thereby degranulating mast cell protease II. This interaction occurred predominantly in the intestinal mucosal mast cells [72]. In mast cell hyperplasia and inflammation, afferent vagal nerves penetrate jejunal mucosa and synapse with intestinal mucosal mast cells [73]. If truncal vagotomy is performed, the concentration of mucosal mast cells decreases and, conversely, stimulation of the cervical vagus nerve leads to an increase in mast cell concentration and histamine content of jejunal mucosa [74]. This finding provides important evidence for a CNS link to hypersensitivity (Fig. 2). Eosinophils and Serotonin The serotonin (5-hydroxytryptamine) signaling system is also implicated in functional bowel disorders [75] and the interplay of serotonin and enteric nerves may induce symptoms of IBS. Serotonin is the predominant neurotransmitter of the ENS and facilitates the effector components: muscle, epithelial secretion, enterochromaffin cells, and vasculature. It is also the principal communicator molecule between the ENS and the CNS. The gut secretes 95% of serotonin within the body, compared with 3% in the brain and 2% taken up
Fig. 2. Eosinophil function. Eosinophils produce lipid mediators, platelet-activating factor and leukotrienes (PAF, LTC), cytokines and chemokines (CXC), cytotoxic secretory products, eosinophil peroxidase (EPO), major basic protein (MBP), eosinophil cationic protein (ECP), eosinophil-derived neurotoxin (EDN), nerve growth factor (NGF), and stem cell factor (SCF). MBP, SCF, and NGF regulate mast cell interaction. Eosinophils activate T cells by serving as antigen-presenting cells (APC) and regulate T cells through synthesis of indoleamine 2,3dioxygenase (IDO) by way of metabolism of L-tryptophan to N-formylkynurenine (KYN), a regulator of Th1/Th2 balance. The interaction of mast cells and eosinophils with neural pathways may produce pain and dysmotility.
FUNCTIONAL GASTROINTESTINAL DISORDERS AND EOSINOPHILS
391
by platelets. [76]. Enterochromaffin cells in the gut epithelium from duodenum to rectum secrete serotonin. Serotonin is a potent eosinophil-active chemoattractant that can function additively with eotaxin [77]. Serotonin also induces mast cell adhesion and migration [78] and mast cells also secrete serotonin [79]. It is therefore likely that serotonin secreted by mast cells modulates dysmotility. On release, serotonin binds to receptors on the intrinsic afferent neurons in the mucosa leading to peristalsis and epithelial secretion [80]. Binding to 5HT3 receptors on efferent sensory neurons increases pain (and nausea) perception, which occurs by way of the vagus nerve, thus acting as another coordinating link between the CNS and the ENS (Fig. 3). EOSINOPHILS IN GASTROINTESTINAL DISEASE Eosinophil accumulation in the gastrointestinal tract is commonly seen in parasite infestation, drug reactions, specific eosinophilic syndromes, IBD, gastroesophageal reflux disease, and celiac disease [81]. A specific group, the eosinophilic syndromes of the gastrointestinal tract (eosinophilic esophagitis, eosinophilic gastritis, and gastroenteritis) are hypersensitivity disorders in which allergy is a prominent feature, and they often are associated with other allergy states, such as atopy [81,82]. These syndromes are independent of blood eosinophilia. In eosinophilic syndromes large numbers of eosinophils in the mucosa cause damage, and it is the accrual of cells at mucosal sites that is key to maintaining the balance between good and harm. There may be an overlap between eosinophilic syndromes and FGIDs that depends on eosinophil load. The evidence from studies of functional dyspepsia indicating a subtle elevation of duodenal eosinophils [8,9] suggests that this functional disorder in some cases may be another eosinophil-mediated disease.
Fig. 3. Interaction of serotonin, eosinophils, and mast cells in the gut. Serotonin is a potent eosinophil-active chemoattractant and induces mast cell adhesion and migration. Mast cells also secrete serotonin. Serotonin secreted by mast cells increases gut dysmotility and binds to receptors on afferent neurons leading to pain, by way of the vagus nerve, acting as a coordinating link between the CNS and the ENS.
392
WALKER & TALLEY
SUMMARY We propose that the neural–mast cell–eosinophil interaction may cause abdominal pain or meal-related symptoms characteristic of functional disease. The trigger is likely a pathogen, food, infection, or other allergen in the gut mucosa. This trigger evokes eosinophils, mast cells, and other components to cascade to up-regulate serotonin and modulation of the ENS directly and thus the CNS. Further study of the putative role of eosinophils in FGIDs by molecular techniques and histopathology is worth pursuing. References [1] Drossman DA. The functional gastrointestinal disorders and the Rome III process. Gastroenterology 2006;130:1377–90. [2] Simren M, Svedlund J, Posserud I, et al. Health-related quality of life in patients attending a gastroenterology outpatient clinic: functional disorders versus organic diseases. Clin Gastroenterol Hepatol 2006;4:187–95. [3] Nyrop KA, Palsson OS, Levy RL, et al. Costs of health care for irritable bowel syndrome, chronic constipation, functional diarrhoea and functional abdominal pain. Aliment Pharmacol Ther 2007;26:237–48. [4] Mayer EA, Naliboff BD, Craig AD. Neuroimaging of the brain-gut axis: from basic understanding to treatment of functional GI disorders. Gastroenterology 2006;131:1925–42. [5] Jones MP, Dilley JB, Drossman D, et al. Brain-gut connections in functional GI disorders: anatomic and physiologic relationships. Neurogastroenterol Motil 2006;18:91–103. [6] Goldstein RA, Paul WE, Metcalfe DD, et al. NIH conference. Asthma. Ann Intern Med 1994;121:698–708. [7] Rothenberg ME, Mishra A, Collins MH, et al. Pathogenesis and clinical features of eosinophilic esophagitis. J Allergy Clin Immunol 2001;108:891–4. [8] Talley NJ, Walker MM, Aro P, et al. Non-ulcer dyspepsia and duodenal eosinophilia: an adult endoscopic population-based case-control study. Clin Gastroenterol Hepatol 2007;5:1175–83. [9] Friesen CA, Sandridge L, Andre L, et al. Mucosal eosinophilia and response to H1/H2 antagonist and cromolyn therapy in pediatric dyspepsia. Clin Pediatr (Phila) 2006;45: 143–7. [10] Lowichik A, Weinberg A. A quantitative evaluation of mucosal eosinophils in the pediatric gastrointestinal tract. Mod Pathol 1996;9:110–4. [11] Pensabene L, Brundler MA, Bank JM, et al. Evaluation of mucosal eosinophils in the pediatric colon. Dig Dis Sci 2005;50:221–9. [12] Pascal RR, Gramlich TL, Parker KM, et al. Geographic variations in eosinophil concentration in normal colonic mucosa. Mod Pathol 1997;10:363–5. [13] Rothenberg ME, Cohen MB. An eosinophil hypothesis for functional dyspepsia. Clin Gastroenterol Hepatol 2007;5:1147–8. [14] Caffarelli C, Deriu FM, Terzi V, et al. Gastrointestinal symptoms in patients with asthma. Arch Dis Child 2000;82:131–5. [15] Caffarelli C, Cavagni G, Deriu FM, et al. Gastrointestinal symptoms in atopic eczema. Arch Dis Child 1998;78:230–4. [16] Powell N, Huntley B, Beech T, et al. Increased prevalence of gastrointestinal symptoms in patients with allergic disease. Postgrad Med J 2007;83:182–6. [17] Powell N, Huntley B, Knight W, et al. Dyspepsia is associated with allergic rhinitis. J Allergy Clin Immunol 2002;109(Suppl 1):S100. [18] Wallaert B, Desreumaux P, Coplin MC, et al. Immunoreactivity for interleukin 3 and 5 granulocyte/macrophage colony-stimulating factor of intestinal mucosa in bronchial asthma. J Exp Med 1995;182:1897–904.
FUNCTIONAL GASTROINTESTINAL DISORDERS AND EOSINOPHILS
393
[19] Pires GV, Souza HS, Elia CC, et al. Small bowel of patients with asthma and allergic rhinitis: absence of inflammation despite the presence of major cellular components of allergic inflammation. Allergy Asthma Proc 2004;25:253–9. [20] Bischoff SC. Prevalence of adverse reactions to food in patients with gastrointestinal disease. Allergy 1996;51:811–8. [21] Schwab D. Functional and morphologic characterization of eosinophils in the lower intestinal mucosa of patients with food allergy. Am J Gastroenterol 2003;98:1525–34. [22] Park M-I, Camilleri M. Is there a role of food allergy in irritable bowel syndrome and functional dyspepsia? A systematic review. Neurogastroenterol Motil 2006;18:595–607. [23] Zar S, Benson MJ, Kumar D. Food-specific serum IgG4 and IgE titers to common food antigens in irritable bowel syndrome. Am J Gastroenterol 2005;100:1550–7. [24] Zuo XL, Li YQ, Li WJ, et al. Alterations of food antigen-specific serum immunoglobulins G and E antibodies in patients with irritable bowel syndrome and functional dyspepsia. Clin Exp Allergy 2007;37:823–30. [25] Spiller RC, Jenkins D, Thornley JP, et al. Increased rectal mucosal enteroendocrine cells, T. lymphocytes, and increased gut permeability following acute Campylobacter enteritis and in post-dysenteric irritable bowel syndrome. Gut 2000;47:804–11. [26] Mearin F, Pe´rez-Oliveras M, Perello´ A, et al. Dyspepsia and irritable bowel syndrome after a Salmonella gastroenteritis outbreak: one-year follow-up cohort study. Gastroenterology 2005;129:98–104. [27] Tack J, Demedts I, Dehondt G, et al. Clinical and pathophysiological characteristics of acuteonset functional dyspepsia. Gastroenterology 2002;122:1738–47. [28] Spiller RC. Role of infection in irritable bowel syndrome. J Gastroenterol 2007;42(Suppl 17):41–7. [29] Dunlop SP, Jenkins D, Neal KR, et al. Relative importance of enterochromaffin cell hyperplasia, anxiety, and depression in postinfectious IBS. Gastroenterology 2003;125:1651–9. [30] Available at: http://guidance.nice.org.uk/CG17/guidance/pdf/English. Accessed November 11, 2007. [31] Tack J. Pathophysiology and treatment of functional dyspepsia. Gastroenterology 2004;127:1239–55. [32] Magalha ˜ es AF, Carvalhaes A, Natan-Eisig J, et al. CagA status and Helicobacter pylori eradication among dyspeptic patients. Gastroenterol Hepatol 2005;28:441–4. [33] O’Sullivan M, Clayton N, Breslin NP, et al. Increased mast cells in the irritable bowel syndrome. Neurogastroenterol Motil 2000;12:449–57. [34] Holtmann G, Geobell H, Talley NJ. Impaired small intestinal peristaltic reflexes and sensory thresholds are independent functional disturbances in patients with chronic unexplained dyspepsia. Am J Gastroenterol 1996;91:485–91. [35] Lee KJ, Demarchi B, Demedts I, et al. A pilot study on duodenal acid exposure and its relationship to symptoms in functional dyspepsia with prominent nausea. Am J Gastroenterol 2004;99:1765–73. [36] Aro P, Ronkainen J, Storskrubb T, et al. Valid symptom reporting at upper endoscopy in a random sample of the Swedish adult general population: the Kalixanda study. Scand J Gastroenterol 2004;39:1280–8. [37] Talley NJ, Walker MM, Aro P, et al. Duodenal eosinophilia and mast cell infiltration in functional dyspepsia (FD) and irritable bowel syndrome (IBS): duodenal eosinophilia a biomarker for FD but not IBS in adults? Gastroenterology 2007;132:A73. [38] Moqbel R, Coughlin JJ. Differential secretion of cytokines. Sci STKE 2006;338:pe26. [39] Miyamasu M, Misaki Y, Yamaguchi M, et al. Regulation of human eotaxin generation by Th1-/Th2-derived cytokines. Int Arch Allergy Immunol 2000;122(Suppl 1):54–8. [40] Rothenberg ME, Owen WF Jr, Silberstein DS, et al. Eosinophils co-cultured with endothelial cells have increased survival and functional properties. Science 1987;237:645–7. [41] Rankin SM, Conroy DM, Williams TJ. Eotaxin and eosinophil recruitment: implications for human disease. Mol Med Today 2000;6:20–7.
394
WALKER & TALLEY
[42] Zimmermann N, Hershey GK, Foster PS, et al. Chemokines in asthma: cooperative interaction between chemokines and IL-13. J Allergy Clin Immunol 2003;111:227–42. [43] Hogan SP, Mishra A, Brandt EB, et al. A pathological function for eotaxin and eosinophils in eosinophilic gastrointestinal inflammation. Nat Immunol 2001;2:353–60. [44] Mishra A, Hogan S, Brandt EB, et al. An etiological role for aeroallergens and eosinophils in experimental esophagitis. J Clin Invest 2001;107:83–90. [45] Mishra A, Rothenberg ME. Intratracheal IL-13 induces eosinophilic esophagitis by an IL-5, eotaxin-1, and STAT6-dependent mechanism. Gastroenterology 2003;125:1419–27. [46] Blanchard C, Wang N, Stringer KF, et al. Eotaxin-3 and a uniquely conserved geneexpression profile in eosinophilic esophagitis. J Clin Invest 2006;116:536–47. [47] Morris-Yates A, Talley NJ, Boyce PM, et al. Evidence of a genetic contribution to functional bowel disorder. Am J Gastroenterol 1998;93:1311–7. [48] Levy RL, Jones KR, Whitehead WE, et al. Irritable bowel syndrome in twins: heredity and social learning both contribute to etiology. Gastroenterology 2001;121:799–804. [49] Lembo A, Zaman M, Jones M, et al. Influence of genetics on irritable bowel syndrome, gastro-oesophageal reflux and dyspepsia: a twin study. Aliment Pharmacol Ther 2007;1(25):1343–50. [50] Camilleri CE, Carlson PJ, Camilleri M, et al. A study of candidate genotypes associated with dyspepsia in a U.S. community. Am J Gastroenterol 2006;101:581–92. [51] Holtmann G, Siffert W, Haag S, et al. G-protein beta 3 subunit 825 CC genotype is associated with unexplained (functional) dyspepsia. Gastroenterology 2004;126:971–9. [52] Holtmann G, Talley NJ. Hypothesis driven research and molecular mechanisms in functional dyspepsia: the beginning of a beautiful friendship in research and practice? Am J Gastroenterol 2006;101:593–5. [53] Micera A, Puxeddu I, Aloe L, et al. New insights on the involvement of nerve growth factor in allergic inflammation and fibrosis. Cytokine Growth Factor Rev 2003;14:369–74. [54] Rothenberg ME, Hogan SP. The eosinophil. Annu Rev Immunol 2006;24:147–74. [55] von Bubnoff D, Hanau D, Wenzel J, et al. Indoleamine 2,3-dioxygenase-expressing antigenpresenting cells and peripheral T-cell tolerance: another piece to the atopic puzzle? J Allergy Clin Immunol 2003;112:854–60. [56] Gargala G, Lecleire S, Franc¸ois A, et al. Duodenal intraepithelial T lymphocytes in patients with functional dyspepsia. World J Gastroenterol 2007;13:2333–8. [57] Morimoto Y, Hizuta A, Ding EX, et al. Functional expression of Fas and Fas ligand on human intestinal intraepithelial lymphocytes. Clin Exp Immunol 1999;116:84–9. [58] Eutamene H, Theodorou V, Fioramonti J, et al. Acute stress modulates the histamine content of mast cells in the gastrointestinal tract through interleukin-1 and corticotropin-releasing factor release in rats. J Physiol 2003;553:959–66. [59] Santos J, Guilarte M, Alonso C, et al. Review—Pathogenesis of irritable bowel syndrome: the mast cell connection. Scand J Gastroenterol 2005;40:129–40. [60] Theoharides TC, Cochrane DE. Critical role of mast cells in inflammatory diseases and the effect of acute stress. J Neuroimmunol 2004;146:1–12. [61] Guilarte M, Santos J, de Torres I, et al. Diarrhoea-predominant IBS patients show mast cell activation and hyperplasia in the jejunum. Gut 2007;56:203–9. [62] Weston AP, Biddle WL, Bhatia PS, et al. Terminal ileal mucosal mast cells in irritable bowel syndrome. Dig Dis Sci 1993;38:1590–5. [63] Barbara G, Stanghellini V, De Giorgio R, et al. Activated mast cells in proximity to colonic nerves correlate with abdominal pain in irritable bowel syndrome. Gastroenterology 2004;126:693–702. [64] Hall W, Buckley M, Crotty P, et al. Gastric mucosal mast cells are increased in Helicobacter pylori-negative functional dyspepsia. Clin Gastroenterol Hepatol 2003;1:363–9. [65] Aridor M, Rajmilevich G, Beaven MA, et al. Activation of exocytosis by the heterotrimeric G-protein Gi3. Science 1993;262:1569–72.
FUNCTIONAL GASTROINTESTINAL DISORDERS AND EOSINOPHILS
395
[66] Piliponsky AM, Gleich GJ, Bar I, et al. Review—Effects of eosinophils on mast cells: a new pathway for the perpetuation of allergic inflammation. Mol Immunol 2001;38:1369–72. [67] Horigome K, Bullock ED, Johnson EM Jr. Effect of nerve growth factor on rat peritoneal mast cells, survival promotion and immediate- early gene induction. J Biol Chem 1994;269: 2695–702. [68] Piliponsky AM, Daigle I, Simon H-U, et al. Expression and functionality of death receptors on the human mast cell line HMC-1. J Allergy Clin Immunol 2000;105:S63. [69] Spiller R, Campbell E. Post-infectious irritable bowel syndrome. Curr Opin Gastroenterol 2006;22:13–7. [70] Suzuki R, Furuno T, McKay DM, et al. Direct neurite–mast cell communication in vitro occurs via the neuropeptide substance P. J Immunol 1999;163:2410–5. [71] Bischoff SC, Schwengberg S, Lorentz A, et al. Substance P and other neuropeptides do not induce mediator release in isolated human intestinal mast cells. Neurogastroenterol Motil 2004;16:185–93. [72] Stead RH, Colley EC, Wang B, et al. Vagal influences over mast cells. Auton Neurosci 2006;125:53–61. [73] Williams RM, Berthoud HR, Stead RH. Vagal afferent nerve fibres contact mast cells in rat small intestinal mucosa. Neuroimmunomodulation 1997;4:266–70. [74] Gottwald T, Lhotak S, Stead RH. Effect of truncal vagotomy and capsaicin on mast cells and IgA- positive plasma cells in rat cells in rat jejunal mucosa. Neurogastroenterol Motil 1997;9:25–32. [75] Gershon MD, Tack J. The serotonin signaling system: from basic understanding to drug development for functional GI disorders. Gastroenterology 2007;132:397–414. [76] Gershon MD. Serotonin and its implication for the management of irritable bowel syndrome. Rev Gastroenterol Disord 2003;3:S25–34. [77] Boehme SA, Lio FM, Sikora L, et al. Cutting edge: serotonin is a chemotactic factor for eosinophils and functions additively with eotaxin. J Immunol 2004;173:3599–603. [78] Kushnir-Sukhov NM, Gilfillan AM, Coleman JW, et al. 5-hydroxytryptamine induces mast cell adhesion and migration. J Immunol 2006;177:6422–32. [79] Kushnir-Sukhov NM, Brown JM, et al. Human mast cells are capable of serotonin synthesis and release. J Allergy Clin Immunol 2007;119:498–9. [80] Gill RK, Saksena S, Tyagi S, et al. Serotonin inhibits Naþ/Hþ exchange activity via 5-HT4 receptors and activation of PKC alpha in human intestinal epithelial cells. Gastroenterology 2005;128:962–74. [81] Rothenberg ME. Eosinophilic gastrointestinal disorders (EGID). J Allergy Clin Immunol 2004;113:11–28. [82] Liacouras CA, Spergel JM, Ruchelli E, et al. Eosinophilic esophagitis: a 10-year experience in 381 children. Clin Gastroenterol Hepatol 2005;3:1198–206.
Gastroenterol Clin N Am 37 (2008) 397–410
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Enteric Autoantibodies and Gut Motility Disorders Purna Kashyap, MBBSa,b,d, Gianrico Farrugia, MDa,b,c,d,* a
Enteric NeuroScience Program, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA Division of Gastroenterology and Hepatology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA c Department of Physiology and Biomedical Engineering, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA d Miles and Shirley Fiterman Center for Digestive Diseases, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA b
A
small proportion of patients with occult or established neoplasms develop a gastrointestinal motility disorder, referred to as paraneoplastic dysmotility. The diagnosis of paraneoplastic dysmotility requires the onset of gastrointestinal dysmotility associated with the presence of a tumor and specific serum antibodies. In these patients, a humoral immune response involving circulating anti-neuronal antibodies is commonly seen. The exact mechanism by which these antibodies are generated is unclear. They are known to target onconeural antigens shared by enteric neurons and tumor cells, suggesting that the antibody is generated against the tumor antigen with the enteric neuron as the ‘‘innocent’’ bystander [1]. The antigens for these antibodies may be localized to the nucleus, plasma membrane, or cytoplasm. ANTIBODIES ASSOCIATED WITH PARANEOPLASTIC AND IDIOPATHIC DYSMOTILITY ANNA-1 (Anti-Hu) The most common neuronal autoantibody associated with paraneoplastic dysmotility is the type 1 anti-neuronal nuclear antibody (ANNA-1) [1,2]. ANNA-1 recognizes the nuclear protein Hu, which belongs to a family of conserved RNA binding proteins that includes HuC, HuD, HuR, and Hel-N1. These proteins are expressed in the neurons of the central, peripheral, and enteric nervous system, with the exception of HuR, which is ubiquitously expressed in
This work was supported by grants DK57061, DK52766, and P01 DK 68055 from the National Institutes of Health.
*Corresponding author. Division of Gastroenterology and Hepatology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. E-mail address:
[email protected] (G. Farrugia). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.005
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
398
KASHYAP & FARRUGIA
proliferating cells [3]. The tumor that most commonly expresses ANNA-1 is small cell lung cancer [4]. Other tumors that may express ANNA-1 include breast, prostate, ovarian carcinomas, and lymphomas [5]. Antibodies to ANNA-1 are consequently most commonly found in patients with small cell lung cancer with associated paraneoplastic gastrointestinal dysmotility. Although there is a strong association between the presence of ANNA-1 in the setting of a gastrointestinal motility disorder and the presence of an occult or manifest tumor, the exact mechanism by which ANNA-1 antibodies cause enteric neuronal dysfunction remains unclear, because the proteins to which the antibody is directed are not expressed on the cell membrane. Nevertheless, there is some evidence that the antibodies may directly influence motility. A preliminary study in guinea pig ileum has suggested that anti-Hu antibodies impair the ascending excitatory reflex and therefore peristalsis. Enteric neuronal degeneration has also been reported in patients with paraneoplastic dysmotility as a possible pathogenetic mechanism [6]. Anti-HuD–positive sera from patients with a paraneoplastic gut dysmotility disorder as well as commercial anti-HuD antibodies have been shown to induce apoptosis in a human neuroblastoma cell line (SH-Sy5Y) as well as guinea pig cultured myenteric neurons. It was further demonstrated that the apoptosis was dependent on mitochondria, as evidenced by the specific activation of effector caspase-3 and the cytochrome c–dependent proapoptotic messenger apaf-1 [7]. Mitochondrial dysfunction leading to subsequent neuronal injury is well described and has also been implicated in dorsal root ganglion apoptosis in streptozocin-induced diabetes in rats [8]. Pardi and colleagues [9] described a patient with sudden onset of gastroparesis and small bowel dysfunction and the presence of high circulating levels of ANNA-1. This patient was subsequently found to have decreased and disorganized interstitial cells of Cajal networks and a small cell lung cancer expressing c-Kit, also expressed on interstitial cells of Cajal. Another nuclear autoantigen associated with disease is Ri, expressed in neurons of the central nervous system, small cell lung cancer, and some breast cancer cells [10]. Formation of type 2 anti-neuronal nuclear antibodies (ANNA-2 or anti-Ri) is far less common than the formation of anti-Hu and is usually associated with neurologic symptoms from midbrain, brain stem, cerebellar, or spinal cord dysfunction [11]. ANNA-2 has not been associated with gastrointestinal dysmotility. Calcium Channel Antibodies The second most commonly reported antibodies in patients with paraneoplastic dysmotility target voltage-activated calcium channels. Calcium channels were originally classified based on pharmacology as L, N, P/Q, R, and T channels, a classification still used today. This classification corresponds to the current accepted nomenclature that classifies voltage-gated Ca2þ channels into Cav1.1-Cav1.4 (L-type Ca2þ channels), Cav2.1 (P/Q), Cav2.2 (N), Cav2.3 (R), and Cav3.1-Cav3.3 (T) based on the amino acid sequence of the alpha 1 subunit (the pore-forming subunit) of the channel. P- or Q-type calcium
ENTERIC AUTOANTIBODIES AND GUT MOTILITY DISORDERS
399
ion channels regulate acetylcholine release at the neuromuscular junction as well as central neurotransmission. N-type calcium channels are particularly involved in cerebrocortical, cerebellar, spinal, and autonomic neurotransmission. Both channel types are expressed in small cell lung cancer and are common targets of autoantibodies in such patients. These antibodies are predominantly seen in patients with Lambert-Eaton myasthenic syndrome in association with small cell lung cancer [12]. Antibodies to P/Q- and N-type calcium channels are found in some patients with paraneoplastic dysmotility, and their presence should trigger a targeted search for an occult malignancy; however, these antibodies are less frequently found when compared with ANNA-1 antibodies, and their association with the eventual finding of a malignancy in the setting of a gastrointestinal motility disorder is not as strong as for ANNA-1. These antibodies may coexist with ANNA-1. Nicotinic Acetylcholine Receptors Another class of autoantibodies associated with gastrointestinal dysmotility is directed against neuronal nicotinic acetylcholine receptors. Antibodies directed toward the extracellular segment of acetylcholine receptor protein in the postsynaptic membrane of skeletal muscle are found in patients with myasthenia gravis associated with thymic epithelial tumors [13]. Neuronal nicotinic acetylcholine receptors are also present on neurons in the sympathetic and parasympathetic nervous systems as well as the enteric nervous system. Antibodies targeting this protein can disrupt cholinergic synaptic transmission leading to autonomic failure. These antibodies are seen in both idiopathic and paraneoplastic forms of autonomic neuropathy resulting in autoimmune autonomic neuropathy [14]. Patients with ganglionic receptor blocking antibodies often manifest with symptoms of gastrointestinal dysmotility, abnormal pupillary response, and subacute onset of autonomic neuropathy [15]. This antibody is of use to differentiate autoimmune autonomic neuropathy from degenerative autonomic neuropathy, which is often more insidious and not usually associated with gastrointestinal manifestations. It is important to make this distinction because the diagnosis affects the prognosis and therapy. Degenerative autonomic neuropathy is a slowly progressive disorder, whereas autoimmune autonomic neuropathy can be life threatening and often responds to therapy such as immunomodulators [15]. Neuronal nicotinic acetylcholine receptor antibodies are most likely directly pathogenic because their levels correspond to the severity of autonomic dysfunction, and because a decrease in their level is accompanied by clinical improvement [15]. Symptoms of autonomic failure can also be induced experimentally by passive transfer of antibodies. Mice injected with rabbit IgG containing ganglionic acetylcholine receptor antibodies develop gastrointestinal dysmotility and autonomic dysfunction. Similar results are obtained by injecting mice with sera from patients with ganglionic acetylcholine receptor antibody [16]. Purkinje Cell Cytoplasmic Autoantibody, Type 1 Purkinje cell cytoplasmic autoantibody, type 1 (PCA-1) (sometimes called ‘‘anti-Yo’’) targets a neuronal signal transduction protein Cdr. The antibody
400
KASHYAP & FARRUGIA
was originally defined as a marker of paraneoplastic cerebellar degeneration related to ovarian or breast carcinoma with remarkably limited metastasis [17,18]. In vitro, its Cdr antigen, a prominent cytoplasmic component of large neurons in the central and autonomic/enteric nervous systems [19], has been shown to promote neuronal apoptosis and degeneration by inhibiting c-myc transcriptional activity [20]. Paraneoplastic gastrointestinal dysmotility has been documented in a minority of PCA-1–seropositive patients (with and without cerebellar ataxia) in association with gynecologic or breast carcinoma [21]. Other Autoantibodies Antibodies have also been detected against other cytoplasmic antigens such as amphiphysin, present on the cytoplasmic side of the synaptic vesicle membrane [22]. Anti-striational autoantibody targets the skeletal muscle proteins actinin, alpha actin, myosin, titin, and ryanodine receptor [23,24]. They may be seen in patients with myasthenia gravis associated with thymoma [25] and paraneoplastic neurologic disorders with primary lung cancer but are usually not associated with gastrointestinal dysmotility [13]. Voltage-gated potassium channel autoantibodies have been reported in rare patients with slow transit constipation [26] and in a patient with diarrhea-predominant irritable bowel syndrome [27]. Their significance is currently unknown. Likewise, glutamic acid decarboxylase-65 (GAD) antibodies have been reported in a significant percentage of patients with achalasia, but, again, this finding is currently of uncertain pathogenetic significance [28]. CLINICAL PRESENTATION OF A PARANEOPLASTIC DYSMOTILITY SYNDROME Paraneoplastic dysmotility of the gastrointestinal tract is often manifested as esophageal dysmotility (pseudoachalasia), gastroparesis, intestinal pseudoobstruction, or constipation. Patients frequently present with a dominant symptom but often have pan-gut involvement. In patients who have had a full thickness biopsy, all had an inflammatory lymphocytic and plasma cell infiltrate of the myenteric plexus and loss of ganglion cells. The smooth muscle layers are often spared [29–31]. Pseudoachalasia Pseudoachalasia accounts for about 2% to 4% of all cases that have the manometric criteria of incomplete or absent relaxation of the lower esophageal sphincter seen in true achalasia. Most patients with pseudoachalasia have a primary tumor at the gastroesophageal junction [32,33]. This form of pseudoachalasia is not a form of paraneoplastic dysmotility, because it is usually due to obstruction of the lower esophageal sphincter by the tumor or direct involvement of myenteric plexus with the neoplastic cells. Nevertheless, depletion of ganglion cells in the dorsal nucleus of the vagus nerve as a consequence of neuronal degeneration distant to primary tumor can occur [34]. In a small proportion of patients, there is no evidence of neoplastic involvement of the gastroesophageal junction, but they demonstrate anti-neuronal antibodies,
ENTERIC AUTOANTIBODIES AND GUT MOTILITY DISORDERS
401
most often ANNA-1 [35,36]. Liu and colleagues [36] described a case series of 13 patients with pseudoachalasia in which eight patients had direct infiltration of the esophageal wall and involvement of the myenteric plexus. The total number of ganglion cells was normal, and it is unclear how the neoplastic cells altered ganglion cell function. In the same series, they described a patient with small cell lung cancer and lymph node metastasis with achalasia-like symptoms but no radiographic or histologic involvement of the esophagogastric junction. The patient had ANNA-1 antibodies, suggesting a paraneoplastic dysmotility. Histologically, there was complete absence of myenteric ganglion cells and both perineural and intraneural lymphocytic infiltration. Lee and colleagues [21] published a series of 12 cases in which esophageal dysmotility was seen in four patients with small cell lung cancer. Two patients had pseudoachalasia, one had a nonspecific esophageal motility disorder, and one had abnormal manometry but no esophageal symptoms. Paraneoplastic Gastroparesis Gastroparesis is characterized by reduced emptying of gastric content, often associated with decreased gastric motility. The presentation of gastroparesis often includes nausea, vomiting of food consumed several hours earlier, bloating, epigastric fullness, and the finding of retained gastric contents on endoscopy. Gastroparesis was reported to be the most common paraneoplastic syndrome associated with ANNA-1 antibodies by Lucchinetti and colleagues [5]. As is true for other paraneoplastic gastrointestinal dysmotilities, paraneoplastic gastroparesis is commonly associated with small cell lung cancer [37]; however, it has also been reported in association with other tumors, including pancreatic cancer with no other identifiable cause [38,39]. It has been described in a patient with untreated breast cancer in whom improvement occurred with cisapride and chemotherapy and resulting tumor remission [40]. It is not possible to conclude whether the patient’s symptoms improved with the prokinetic therapy or treatment of the underlying tumor, but it is more than likely that a paraneoplastic dysmotility responds to treatment of the underlying tumor. Gastroparesis has also been reported in a patient with retroperitoneal leiomyosarcoma with no evidence of local invasion or metastasis and resolved completely after resection of the tumor [41]. As described previously, Pardi and colleagues [9] reported a patient with small cell lung cancer with both ANNA-1 and P/Qtype calcium channel antibody with gastroparesis and a disrupted interstitial cell of Cajal network, suggesting that enteric neurons are not the only enteric neural target of the paraneoplastic autoimmunity. Gastroparesis has also been associated with other autoantibodies, including ganglionic acetylcholine receptor antibodies. This finding was observed by Vernino and colleagues [15] in a patient with bladder cancer and also in patients with idiopathic gastroparesis with no known antecedent risk factors and no underlying cancer. These observations strongly suggest that the histologic nature of the underlying malignancy does not always dictate a certain autoantibody formation or specific dysmotility syndrome.
402
KASHYAP & FARRUGIA
Paraneoplastic Chronic Intestinal Pseudo-Obstruction Chronic intestinal pseudo-obstruction is defined as recurrent episodes or persistent symptoms of bowel obstruction in the absence of mechanical obstruction. Most cases of chronic intestinal pseudo-obstruction are due to primary defects in the contractile apparatus (nerves, interstitial cells of Cajal, smooth muscle cells) of the gut or are secondary to an infiltrating disease such as amyloidosis or scleroderma [42–44]. Paraneoplastic intestinal pseudo-obstruction is most often reported in cases with small cell lung cancer and thymoma and is usually associated with the presence of circulating ANNA-1 antibodies [45–49]. There have also been several case reports of patients with pseudo-obstruction with other primary cancers. Chronic intestinal pseudo-obstruction along with achalasia, gastroparesis, and constipation was reported in a patient with metastasizing bronchial carcinoid [50]. Intestinal pseudo-obstruction associated with lymphoplasmacytic infiltration of the myenteric plexus and the presence of ANNA-1 antibodies was found in a patient about 1.5 years after removal of a paravertebral ganglioneuroblastoma [51]. Viallard and colleagues [52] reported a case of colonic dysmotility associated with antibodies against voltage-gated potassium channels in a patient with invasive thymoma and acquired neuromyotonia, which improved after plasmapheresis. As discussed previously, Lee and colleagues [21] published a case series of 12 patients in which chronic intestinal dysmotility and acute colonic pseudoobstruction were observed in patients with cancer and ANNA-1, PCA-1, or anti–N-type calcium channel antibodies. All of the patients with small cell lung cancer had ANNA-1 antibodies, a patient with ovarian cancer was positive for PCA-1 antibody, and a patient with lymphoma had N-type calcium channel antibody. Histologically, the ANNA-1 antibodies in these patients were reactive with both nucleus and cytoplasm, as opposed to the findings in cases of idiopathic colonic pseudo-obstruction, in which a predominant cytoplasmic staining has been observed. Interestingly, in this study, the investigators observed that the gastrointestinal dysmotility preceded the small cell lung cancer by a mean period of 8.7 months, whereas in the patients who had other malignancies, antibodies were found after the tumor diagnosis. These data suggest that the diagnosis of new onset gut dysmotility accompanied by the presence of ANNA-1 antibodies should prompt a search for occult small cell lung cancer. Even if the initial screen is negative, vigilance should be maintained in the subsequent years. It was also observed that the patients with colonic pseudoobstruction and small cell lung cancer often had additional disease, including peripheral, sensorimotor, or autonomic neuropathies, cerebellar degeneration, or encephalopathy. Chronic Constipation Chronic constipation without accompanying pseudo-obstruction is not a common presentation of a paraneoplastic syndrome. Vernino and colleagues [15] reported constipation in two patients with ganglionic acetylcholine receptor antibody associated with thymoma and small cell lung cancer.
ENTERIC AUTOANTIBODIES AND GUT MOTILITY DISORDERS
403
MANAGEMENT ALGORITHM FOR PARANEOPLASTIC DYSMOTILITY Currently, there is insufficient evidence to recommend a paraneoplastic antibody profile for every patient with new onset of a gut motility disorder; however, the presence of significant weight loss, a rapid onset of the disease, and a past or present smoking history should prompt the physician to consider testing for the presence of autoantibodies associated with paraneoplastic dysmotility. Several laboratories offer a paraneoplastic autoantibody profile. In patients testing positive for ANNA-1, together with the appropriate gastrointestinal motility work-up, including tests to exclude obstruction, investigations should be initiated to look for small cell lung cancer, because paraneoplastic dysmotility may precede the diagnosis of the primary malignancy in patients testing positive for ANNA-1 [21]. A reasonable strategy is to start with a CT of the chest. If it is negative, follow-up with a PET scan and directed biopsies of any suspicious lymph nodes or masses are indicated. In a subset of patients with a high suspicion of malignancy, if the work-up is negative, one may consider bronchoscopy followed by mediastinoscopy to increase the diagnostic yield. An alternative strategy is to repeat CT of the chest at 6-month intervals for at least 1 year. Importantly, finding an alternate primary malignancy should not stop the search for possible small cell lung cancer, because in nearly 13% patients, an unrelated primary malignancy coexists with small cell lung cancer, the most common being renal cell cancer [5]. The presence of other autoantibodies without concomitant ANNA-1 positivity is less likely to predict the presence of a malignancy. It is unclear what strategy to use in this situation. The authors’ current management algorithm, not evidence based, is to obtain a CT of the chest and, if negative, repeat the scan once in 6 months. Treatment of Paraneoplastic Dysmotility The diagnosis of a paraneoplastic dysmotility is unfortunately associated with a bad outcome. Death often occurs within 6 months of the diagnosis, usually from the underlying malignancy but accelerated by difficulties in maintaining nutrition. There are no effective treatments available for paraneoplastic dysmotility. Several treatments have been investigated, including immunosuppressive therapy with steroids and cyclophosphamide, plasmapheresis, and intravenous immunoglobulin, but none have been convincingly shown to alter outcome [5,53]. The mainstay of treatment is to address the underlying primary malignancy. It is important to provide supportive care including nutritional support either enterally or parenterally, adequate hydration, and the use of prokinetics to promote motility and the treatment of complications such as bacterial overgrowth. One additional management strategy is to use high-dose intravenous steroids for 3 days. If there is a clinical response, a switch is made to 6-mercatopurine or azathioprine; however, this approach often needs to be aborted due to the need for chemotherapy.
404
KASHYAP & FARRUGIA
CLINICAL PRESENTATION OF A NON-PARANEOPLASTIC DYSMOTILITY SYNDROME ASSOCIATED WITH CIRCULATING ANTIBODIES Achalasia Idiopathic achalasia is a relatively common esophageal motility disorder (1:100,000) characterized by ineffective peristalsis and an abnormal relaxation of the lower esophageal sphincter [54]. A consistent finding in idiopathic achalasia is the loss of nitrergic neurons with a relative preservation of cholinergic neurons [55–57]. As the disease progresses, the enteric neuronal loss becomes more generalized. Most studies that have reported on biopsy specimens taken from patients with relatively early disease have shown the presence of an inflammatory infiltrate in the myenteric plexus. This infiltrate is predominantly due to CD3-positive T cells, suggesting an underlying immune-mediated process [58–60]. The serum from patients with achalasia often contains antineuronal antibodies; however, studies on their role in causation of disease have been inconclusive because similar antibodies are present in patients with gastroesophageal reflux disease. Moses and colleagues [61] collected sera from 45 patients with achalasia and 16 with gastroesophageal reflux disease as well as from healthy controls and demonstrated that the sera from patients who had achalasia and that from patients with gastroesophageal reflux disease labeled myenteric neurons in the esophagus as well as the ileum and submucosal plexus neurons in guinea pig and mice with rare labeling of the spinal neurons. Based on these findings, they felt that the presence of antineuronal autoantibodies was likely secondary to tissue damage and not involved in the pathogenesis of achalasia. A recent study, at present reported only in abstract form, suggests that 60% of patients with primary idiopathic achalasia have circulating anti–GAD-65 antibodies. Like ANNA-1, GAD-65 is not a membrane protein; therefore, it is unclear what role the antibodies have in the pathophysiology of the disease [28]. Chronic Intestinal Pseudo-Obstruction Chronic intestinal pseudo-obstruction can be divided into primary chronic intestinal pseudo-obstruction and secondary chronic intestinal pseudo-obstruction, the latter a consequence of an underlying condition such as amyloidosis, scleroderma, and a variety of other systemic disorders. Primary chronic intestinal pseudo-obstruction may be due to genetic defects affecting the contractile apparatus but often is idiopathic with no identifiable cause detected. Antineuronal antibodies are detected in some patients with idiopathic chronic intestinal pseudo-obstruction, suggesting a possible cause. Antibodies such as ANNA-1 usually indicate occult malignancy such as small cell carcinoma, but some patients with autoantibodies do not have underlying malignancy. Intestinal dysmotility has been observed in patients with antibodies such as those for ganglionic acetylcholine receptor with no evidence of malignancy even on long-term follow-up [15]. They likely represent the idiopathic form of autoimmune autonomic neuropathy. Ganglionic acetylcholine receptor
ENTERIC AUTOANTIBODIES AND GUT MOTILITY DISORDERS
405
antibody has been associated with variable degrees of autonomic failure and is one of the more common antibodies detected in patients with intestinal dysmotility. As described previously in a series of 25 patients with ganglionic receptor antibodies published by Vernino and colleagues [15], at least four patients had constipation with no antecedent risk factors or malignancy. These antibodies are likely pathogenic because they cause disease when injected in animal models [16], and there was a positive correlation between antibody levels and the degree of autonomic failure [15]. Smith and colleagues [31] described two patients with intestinal pseudoobstruction characterized histologically with acquired aganglionosis and a T-cell infiltrate affecting the myenteric neurons. Their sera contained antineuronal IgG antibodies similar to anti-Hu but demonstrated strong cytoplasmic staining rather than nuclear staining. Neurogenic inflammatory chronic intestinal pseudo-obstruction is being increasingly recognized in a small but distinct subset of patients with chronic intestinal pseudo-obstruction who undergo a full-thickness biopsy. The typical finding is a dense plasma cell and lymphocytic infiltrate in and around the ganglia. The identity of the lymphocytic infiltrate (CD3-positive and both CD4- and CD8-positive cells) suggests a T-cell mediated injury to the ganglionic cells, although B cells have also been reported [42,43]. Myenteric ganglia are invariably involved; however, submucosal ganglia may also be involved. MANAGEMENT OF DYSMOTILITY ASSOCIATED WITH ANTIBODIES OF LIKELY PATHOGENIC IMPORTANCE Treatment needs to be directed toward managing the disorder as well as attempting to intervene in the immune-mediated process. Intestinal dysmotility is often associated with complications such as bacterial overgrowth and malabsorption. These conditions need to be treated appropriately. Treatment of the primary disorder is often difficult and requires the judicious use of available medications including prokinetics. Among patients with idiopathic myenteric ganglionitis with intestinal dysmotility, the greatest benefit has been shown with the use of immunosuppressive therapy such as steroids either alone or in conjunction with azathioprine or cyclophosphamide. Steroids that have been used in varying doses and tapers include prednisolone, methylprednisolone, and beclomethasone. A female patient with chronic intestinal pseudoobstruction with underlying lymphoid infiltrate in the myenteric plexus showed a mild improvement in symptoms with prednisone and cyclophosphamide [62]. A young male patient with idiopathic myenteric ganglionitis and intractable vomiting failed therapy with prokinetics, but he responded to a steroid taper starting at 60 mg/d of methylprednisolone, with a sustained response at 1 year [29]. Smith and colleagues [31] published a series of two patients with chronic pseudo-obstruction and IgG autoantibodies directed against enteric neurons. One patient improved with a trial of prednisolone used in a dose of 10 mg/d and relapsed on discontinuation of the steroid. The other patient was treated initially
406
KASHYAP & FARRUGIA
with prednisolone, 2 mg/kg for 4 weeks and reduced to 0.5 mg/kg every other day, allowing the introduction of enteral feeds. This patient did poorly in the long term and required intestinal transplantation. De Giorgio and colleagues [63] described a patient with intestinal subocclusion and the presence of anti-Hu (ANNA-1) antibodies who responded well to pulse dose steroids (100 mg of intravenous methylprednisolone for 3 days). Scha¨ppi and colleagues [64] described three patients with predominantly eosinophilic myenteric ganglionitis who had a good response to oral beclomethasone, prednisolone at 60 mg/d with azathioprine, or high-dose intravenous prednisolone. Treatment with steroids often has to be tailored to the individual. One approach is to attempt a short course of pulse steroids such as intravenous methylprednisolone or to start with oral steroids in a dose of 60 mg/d with a taper over the next 4 to 6 weeks. If symptoms relapse after discontinuation, steroids should be reinstituted and a slower taper attempted. At this time it would also be reasonable to consider immunomodulator therapy such as azathioprine starting at a dose of 2 mg/kg. Plasmapheresis may be beneficial in patients with ganglionic receptor antibodies, who often show improvement with a decrease in titers of the antibodies. Acetylcholinesterase inhibitors such as neostigmine [65] and pyridostigmine have been used successfully to treat intestinal dysmotility. Pasha and colleagues [66] described a patient with idiopathic gastroparesis associated with N-type calcium channel antibody and ganglionic acetylcholine receptor antibody who responded to treatment with pyridostigmine despite a 15-year history of gastrointestinal symptoms. Gastrointestinal motility tests such as gastroduodenal manometry and colonic motility testing combined with the administration of neostigmine may help in making the decision to pursue pyridostigmine therapy. A significant problem in the diagnosis and treatment of dysmotility attributed to an autoimmune cause is the lack of a well-defined way to assess the degree of inflammatory infiltrate and the response to therapy. We are currently limited to surgery, be it open or laparoscopic, to obtain full-thickness gut wall biopsies. This requirement significantly limits the ability to repeatedly obtain tissue and to make informed decisions on the need for immunosuppression as well as the duration of immunosuppression or other therapeutic modalities. A significant advance in the field will be the introduction of endoscopic methodology to obtain full-thickness biopsies. This tool would allow not only a determination of the utility of current approaches but would also permit studies to determine the potential role inflammation and autoimmunity may have in the pathophysiology of a wide variety of gastrointestinal disorders not currently thought to be immune mediated. SUMMARY Increasing evidence suggests an immune-mediated role in the pathogenesis of several gastrointestinal motility disorders. The role of autoantibodies in paraneoplastic dysmotility is now well established, and the role of autoantibodies
ENTERIC AUTOANTIBODIES AND GUT MOTILITY DISORDERS
407
in a subset of patients with non-paraneoplastic gastrointestinal dysmotility is being recognized and the antibodies characterized. It is hoped that a better understanding of the role of autoimmunity and the specific antibodies involved, coupled with the development of less invasive techniques to obtain tissue or the development of better biomarkers, will lead to earlier diagnosis and targeted treatment with appropriate immunosuppressant therapy. Acknowledgments The authors thank Kristy Zodrow for secretarial assistance. References [1] Dropcho EJ. Remote neurologic manifestations of cancer. Neurol Clin 2002;20(1): 85–122. [2] Lennon VA. Calcium channel and related paraneoplastic disease autoantibodies. Amsterdam: Elsevier B.V.; 1996. [3] Wakamatsu Y, Weston JA. Sequential expression and role of Hu RNA-binding proteins during neurogenesis. Development 1997;124(17):3449–60. [4] Kiers L, Altermatt HJ, Lennon VA. Paraneoplastic anti-neuronal nuclear IgG autoantibodies (type I) localize antigen in small cell lung carcinoma. Mayo Clin Proc 1991;66(12): 1209–16. [5] Lucchinetti CF, Kimmel DW, Lennon VA. Paraneoplastic and oncologic profiles of patients seropositive for type 1 antineuronal nuclear autoantibodies. Neurology 1998;50(3):652–7. [6] Caras SD, McCallum RW, Brashear HR, et al. The effect of human antineuronal antibodies on the ascending excitatory reflex and peristalsis in isolated guinea pig ileum. Gastroenterology 1996;110:A643. [7] De Giorgio R, Bovara M, Barbara G, et al. Anti-HuD-induced neuronal apoptosis underlying paraneoplastic gut dysmotility. Gastroenterology 2003;125(1):70–9. [8] Srinivasan S, Stevens MJ, Sheng H, et al. Serum from patients with type 2 diabetes with neuropathy induces complement-independent, calcium-dependent apoptosis in cultured neuronal cells. J Clin Invest 1998;102(7):1454–62. [9] Pardi DS, Miller SM, Miller DL, et al. Paraneoplastic dysmotility: loss of interstitial cells of Cajal. Am J Gastroenterol 2002;97(7):1828–33. [10] Lennon VA. The case for a descriptive generic nomenclature: clarification of immunostaining criteria for PCA-1, ANNA-1, and ANNA-2 autoantibodies. Neurology 1994;44(12): 2412–5. [11] Luque FA, Furneaux HM, Ferziger R, et al. Anti-Ri: an antibody associated with paraneoplastic opsoclonus and breast cancer. Ann Neurol 1991;29(3):241–51. [12] Lennon VA, Kryzer TJ, Griesmann GE, et al. Calcium-channel antibodies in the LambertEaton syndrome and other paraneoplastic syndromes. N Engl J Med 1995;332(22): 1467–74. [13] Griesmann GE, Lennon VA. Detection of autoantibodies in myasthenia gravis and Lambert-Eaton myasthenic syndrome. Manual of clinical laboratory immunology. 5th edition. Washington DC: ASM Press; 1997. [14] Vernino S, Adamski J, Kryzer TJ, et al. Neuronal nicotinic ACh receptor antibody in subacute autonomic neuropathy and cancer-related syndromes. Neurology 1998;50(6):1806–13. [15] Vernino S, Low PA, Fealey RD, et al. Autoantibodies to ganglionic acetylcholine receptors in autoimmune autonomic neuropathies. N Engl J Med 2000;343(12):847–55. [16] Vernino S, Ermilov LG, Sha L, et al. Passive transfer of autoimmune autonomic neuropathy to mice. J Neurosci 2004;24(32):7037–42. [17] Hetzel DJ, Stanhope CR, O’Neill BP, et al. Gynecologic cancer in patients with subacute cerebellar degeneration predicted by anti-Purkinje cell antibodies and limited in metastatic volume. Mayo Clin Proc 1990;65(12):1558–63.
408
KASHYAP & FARRUGIA
[18] Greenlee JE, Brashear HR. Antibodies to cerebellar Purkinje cells in patients with paraneoplastic cerebellar degeneration and ovarian carcinoma. Ann Neurol 1983;14(6): 609–13. [19] Altermatt HJ, Rodriguez M, Scheithauer BW, et al. Paraneoplastic anti-Purkinje and type I anti-neuronal nuclear autoantibodies bind selectively to central, peripheral, and autonomic nervous system cells. Lab Invest 1991;65(4):412–20. [20] Okano HJ, Park WY, Corradi JP, et al. The cytoplasmic Purkinje onconeural antigen cdr2 down-regulates c-Myc function: implications for neuronal and tumor cell survival. Genes Dev 1999;13(16):2087–97. [21] Lee HR, Lennon VA, Camilleri M, et al. Paraneoplastic gastrointestinal motor dysfunction: clinical and laboratory characteristics. Am J Gastroenterol 2001;96(2):373–9. [22] Folli F, Solimena M, Cofiell R, et al. Autoantibodies to a 128-kd synaptic protein in three women with the stiff-man syndrome and breast cancer. N Engl J Med 1993;328(8): 546–51. [23] Williams CL, Hay JE, Huiatt TW, et al. Paraneoplastic IgG striational autoantibodies produced by clonal thymic B cells and in serum of patients with myasthenia gravis and thymoma react with titin. Lab Invest 1992;66(3):331–6. [24] Williams CL, Lennon VA. Thymic B lymphocyte clones from patients with myasthenia gravis secrete monoclonal striational autoantibodies reacting with myosin, alpha actinin, or actin. J Exp Med 1986;164(4):1043–59. [25] Strauss AJ, van der Geld HW, Kemp PG Jr, et al. Immunological concomitants of myasthenia gravis. Ann N Y Acad Sci 1965;124(2):744–66. [26] Knowles CH, Lang B, Clover L, et al. A role for autoantibodies in some cases of acquired non-paraneoplastic gut dysmotility. Scand J Gastroenterol 2002;37(2):166–70. [27] Tornblom H, Lang B, Clover L, et al. Autoantibodies in patients with gut motility disorders and enteric neuropathy. Scand J Gastroenterol 2007;42(11):1289–93. [28] Kraichely RE, Lennon VA, Pittock SJ, et al. Autoantibodies in primary achalasia. Fifteenth Annual American Motility Society Scientific Meeting, Diabetes and the Gut. Grand Bahama Island, Freeport, March 1-4, 2007. Neurogastroenterol Motil 2007;19(5):425. [29] De Giorgio R, Barbara G, Stanghellini V, et al. Idiopathic myenteric ganglionitis underlying intractable vomiting in a young adult. Eur J Gastroenterol Hepatol 2000;12(6): 613–6. [30] De Giorgio R, Stanghellini V, Barbara G, et al. Primary enteric neuropathies underlying gastrointestinal motor dysfunction. Scand J Gastroenterol 2000;35(2):114–22. [31] Smith VV, Gregson N, Foggensteiner L, et al. Acquired intestinal aganglionosis and circulating autoantibodies without neoplasia or other neural involvement. Gastroenterology 1997;112(4):1366–71. [32] Moonka R, Patti MG, Feo CV, et al. Clinical presentation and evaluation of malignant pseudoachalasia. J Gastrointest Surg 1999;3(5):456–61. [33] Kahrilas PJ, Kishk SM, Helm JF, et al. Comparison of pseudoachalasia and achalasia. Am J Med 1987;82(3):439–46. [34] Ulla JL, Fernandez-Salgado E, Alvarez V, et al. Pseudoachalasia of the cardia secondary to nongastrointestinal neoplasia. Dysphagia 2007. [35] Gockel I, Eckardt VF, Schmitt T, et al. Pseudoachalasia: a case series and analysis of the literature. Scand J Gastroenterol 2005;40(4):378–85. [36] Liu W, Fackler W, Rice TW, et al. The pathogenesis of pseudoachalasia: a clinicopathologic study of 13 cases of a rare entity. Am J Surg Pathol 2002;26(6):784–8. [37] Moskovitz DN, Robb KV. Small cell lung cancer with positive anti-Hu antibodies presenting as gastroparesis. Can J Gastroenterol 2002;16(3):171–4. [38] Caras S, Laurie S, Cronk W, et al. Case report: pancreatic cancer presenting with paraneoplastic gastroparesis. Am J Med Sci 1996;312(1):34–6. [39] Barkin JS, Goldberg RI, Sfakianakis GN, et al. Pancreatic carcinoma is associated with delayed gastric emptying. Dig Dis Sci 1986;31(3):265–7.
ENTERIC AUTOANTIBODIES AND GUT MOTILITY DISORDERS
409
[40] Berghmans T, Musch W, Brenez D, et al. Paraneoplastic gastroparesis. Rev Med Brux 1993;14(9–10):275–8 [in French]. [41] Lautenbach E, Lichtenstein GR. Retroperitoneal leiomyosarcoma and gastroparesis: a new association and review of tumor-associated intestinal pseudo-obstruction. Am J Gastroenterol 1995;90(8):1338–41. [42] De Giorgio R, Sarnelli G, Corinaldesi R, et al. Advances in our understanding of the pathology of chronic intestinal pseudo-obstruction. Gut 2004;53(11):1549–52. [43] De Giorgio R, Camilleri M. Human enteric neuropathies: morphology and molecular pathology. Neurogastroenterol Motil 2004;16(5):515–31. [44] De Giorgio R, Guerrini S, Barbara G, et al. Inflammatory neuropathies of the enteric nervous system. Gastroenterology 2004;126(7):1872–83. [45] Anderson NE, Hutchinson DO, Nicholson GJ, et al. Intestinal pseudo-obstruction, myasthenia gravis, and thymoma. Neurology 1996;47(4):985–7. [46] Lennon VA, Sas DF, Busk MF, et al. Enteric neuronal autoantibodies in pseudoobstruction with small cell lung carcinoma. Gastroenterology 1991;100(1):137–42. [47] Kulling D, Reed CE, Verne GN, et al. Intestinal pseudo-obstruction as a paraneoplastic manifestation of malignant thymoma. Am J Gastroenterol 1997;92(9):1564–6. [48] Tabbaa MA, Leshner RT, Campbell WW. Malignant thymoma with dysautonomia and disordered neuromuscular transmission. Arch Neurol 1986;43(9):955–7. [49] Pande R, Leis AA. Myasthenia gravis, thymoma, intestinal pseudo-obstruction, and neuronal nicotinic acetylcholine receptor antibody. Muscle Nerve 1999;22(11):1600–2. [50] Gerl A, Storck M, Schalhorn A, et al. Paraneoplastic chronic intestinal pseudoobstruction as a rare complication of bronchial carcinoid. Gut 1992;33(7):1000–3. [51] Schobinger-Clement S, Gerber HA, Stallmach T. Autoaggressive inflammation of the myenteric plexus resulting in intestinal pseudoobstruction. Am J Surg Pathol 1999;23(5): 602–6. [52] Viallard JF, Vincent A, Moreau JF, et al. Thymoma-associated neuromyotonia with antibodies against voltage-gated potassium channels presenting as chronic intestinal pseudo-obstruction. Eur Neurol 2005;53(2):60–3. [53] Graus F, Vega F, Delattre JY, et al. Plasmapheresis and antineoplastic treatment in CNS paraneoplastic syndromes with antineuronal autoantibodies. Neurology 1992;42(3 Pt 1): 536–40. [54] Kraichely RE, Farrugia G. Achalasia: physiology and etiopathogenesis. Dis Esophagus 2006;19(4):213–23. [55] Holloway RH, Dodds WJ, Helm JF, et al. Integrity of cholinergic innervation to the lower esophageal sphincter in achalasia. Gastroenterology 1986;90(4):924–9. [56] De Giorgio R, Di Simone MP, Stanghellini V, et al. Esophageal and gastric nitric oxide synthesizing innervation in primary achalasia. Am J Gastroenterol 1999;94(9):2357–62. [57] Mearin F, Mourelle M, Guarner F, et al. Patients with achalasia lack nitric oxide synthase in the gastro-oesophageal junction. Eur J Clin Invest 1993;23(11):724–8. [58] Khelif K, De Laet MH, Chaouachi B, et al. Achalasia of the cardia in Allgrove’s (triple A) syndrome: histopathologic study of 10 cases. Am J Surg Pathol 2003;27(5):667–72. [59] Goldblum JR, Rice TW, Richter JE. Histopathologic features in esophagomyotomy specimens from patients with achalasia. Gastroenterology 1996;111(3):648–54. [60] Goldblum JR, Whyte RI, Orringer MB, et al. Achalasia: a morphologic study of 42 resected specimens. Am J Surg Pathol 1994;18(4):327–37. [61] Moses PL, Ellis LM, Anees MR, et al. Antineuronal antibodies in idiopathic achalasia and gastro-oesophageal reflux disease. Gut 2003;52(5):629–36. [62] McDonald GB, Schuffler MD, Kadin ME, et al. Intestinal pseudoobstruction caused by diffuse lymphoid infiltration of the small intestine. Gastroenterology 1985;89(4):882–9. [63] De Giorgio R, Barbara G, Stanghellini V, et al. Clinical and morphofunctional features of idiopathic myenteric ganglionitis underlying severe intestinal motor dysfunction: a study of three cases. Am J Gastroenterol 2002;97(9):2454–9.
410
KASHYAP & FARRUGIA
[64] Scha ¨ ppi MG, Smith VV, Milla PJ, et al. Eosinophilic myenteric ganglionitis is associated with functional intestinal obstruction. Gut 2003;52(5):752–5. [65] Calvet X, Martinez JM, Martinez M. Repeated neostigmine dosage as palliative treatment for chronic colonic pseudo-obstruction in a patient with autonomic paraneoplastic neuropathy. Am J Gastroenterol 2003;98(3):708–9. [66] Pasha SF, Lunsford TN, Lennon VA. Autoimmune gastrointestinal dysmotility treated successfully with pyridostigmine. Gastroenterology 2006;131(5):1592–6.
Gastroenterol Clin N Am 37 (2008) 411–428
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Celiac Disease and Autoimmunity in the Gut and Elsewhere Susan H. Barton, MD, Joseph A. Murray, MD* Division of Gastroenterology and Hepatology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA
C
eliac disease is a common immune-mediated enteropathy with a prevalence of approximately 1% within the US and European populations. There is a worldwide disease distribution including Mexico, South America, the Middle East, parts of India, and specific regions of Africa. Although the classic and usually obvious consequences of the enteropathy are malabsorption with diarrhea, weight loss, and nutritional deficiencies, the difficulty in diagnosis is due in large part to the silent form of the disease which affects the majority of patients. Overall mild clinical symptoms with nonspecific complaints such as fatigue, headaches, and arthralgias are common and can delay diagnosis. A large multicenter study in the United States showed an increased disease prevalence in high-risk groups, including patients with autoimmune insulin-dependent diabetes mellitus (AIDDM), Sjogren’s syndrome, osteoporosis, and first-degree relatives of patients with celiac disease. In addition to diarrhea, abdominal pain and constipation were among the most highly reported symptoms [1]. Because of the diagnostic difficulties inherent in this often mild or clinically silent presentation, the identification of high-risk groups for serologic screening may decrease the time to diagnosis and lessen the complications of untreated disease. Several studies have demonstrated the cost-effectiveness of screening the population with irritable bowel syndrome for celiac disease. The results from one recent study addressed the possibility of immunologically based mechanisms following gluten exposure contributing to irritable bowel syndrome symptoms that may represent a celiac-like disorder. This study showed decreases in stool frequency and improvement in the gastrointestinal symptoms score among 60% of patients with diarrhea-predominant irritable bowel
Work for this article was supported by NIH training grant T32 DK07198 (SHB) and NIH grants DK57892 and 071003 (JAM). Dr. Murray has been a consultant to Astra Zeneca, Alvine, and Novartis and an investigator for Alba Therapeutics and Dynagen.
*Corresponding author. E-mail address:
[email protected] (J.A. Murray). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.001
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
412
BARTON & MURRAY
syndrome who carried the celiac disease–associated HLA type (DQ2) but lacked fully developed celiac disease [2]. Various autoimmune diseases have been historically associated with celiac disease [3,4]. These findings raise interesting questions as to whether abnormal immune responses at the level of the gut mucosa, when exposed to environmental antigens, have a role in systemic autoimmune disease, or whether these associations reflect more an underlying genetic predisposition. Proposed mechanisms of association include abnormal regulation of intestinal permeability and increased autoantibody production in the setting of chronic gut inflammation. This review focuses on the autoimmune connective tissue diseases, endocrine, and dermatologic conditions associated with celiac disease, as well as the related gut inflammatory disorders of refractory celiac disease, autoimmune enteropathy, collagenous enteritis, and collagenous colitis. GUT IMMUNOGENESIS OF CELIAC DISEASE Celiac disease is the result of an unchecked immune reaction to gluten. This unchecked response results in inflammation of the proximal small intestine where the partially digested gluten proteins contact the gut immune system. This immune response extends beyond just a direct response to the exogenous substance, also including a potent and multifaceted immune response to autoantigens that results in substantial damage to the structure and function of the gut and other organs. The clinical manifestations are heterogeneous, including the complete absence of gastrointestinal complaints, a myriad of extraintestinal manifestations, signs of overt malabsorption, and, in rare cases, the development of ulcerative jejunitis and enteropathy associated T-cell lymphoma (EATL). In the small intestinal mucosa, CD4þ T cells are stimulated by gliadin peptides only when presented by DQ2 or DQ8 MHC molecules on the surface of antigen-presenting cells. This binding triggers a proliferation of pathogenic gluten-specific T cells, a potent Th1 inflammatory response characterized by high levels of interferon-c, and subsequent villous atrophy. In addition to this response to gluten, there is a potent and almost universal humoral response to autoantigen tissue transglutaminase (tTg) in the gut. The role and, indeed, the production of tTg autoantibodies are not fully understood. What is truly curious is the ability of cell surface tTg to bind and deamidate gliadin peptides [5]. These deamidated peptides subsequently bind tightly to DQ2/DQ8 molecules, potentiating CD4þ T-cell activation and proliferation. What the antibodies that are directed against the active site of the transglutaminase are doing is uncertain from a pathologic perspective. Both T- and B-cell responses diminish upon withdrawal of gluten, and normal villous architecture is restored. Although antibodies directed against wheat proteins have been pursued for many years, the recognition of more specific antibodies directed against extracellular antigens revolutionized the detection of celiac disease. Given the high specificity of positive anti-endomysial antibody (EMA) titers and the high
CELIAC DISEASE AND AUTOIMMUNITY
413
sensitivity of IgA tTg, both tests in conjunction have a high positive predictive value. Because EMA testing increases expense and is prone to interpretation errors, tTg antibody testing alone is recommended for screening. Subsequent tTg antibody titers can be followed to assess occult gluten exposure or dietary indiscretion. The time to normalization of IgA tTg titers in the setting of a strict gluten-free diet is variable, ranging from weeks to months. Although it has been widely hypothesized that there may be environmental triggers for many autoimmune diseases, celiac disease is unique in that we know the major and necessary precipitating factors. Additionally, celiac disease is different from most autoimmune diseases in that the environmental triggers need to be there all of the time or the disease and the autoantibodies regress. Most people with the known genetic predisposition (DQ2 or 8) who eat gluten do not get celiac disease; therefore, there must be other factors. These additional factors contributing to disease development are less well understood but likely include stimulation of an innate immune response to some environmental factor derived from gluten or some other factor present in the intestinal lumen. ENVIRONMENTAL FACTORS The timing or amount of gluten exposure may be important factors in disease development. The data to support this comes largely from observational studies. In the 1980s, the incidence of celiac disease among Swedish children increased significantly and remained high for approximately 10 years. Subsequently, with nationwide changes in infant feeding practices to reduce gluten exposure and to alter its timing, the incidence of celiac disease dropped back to pre-epidemic levels. This ‘‘Swedish epidemic’’ generated the hypothesis that the introduction of large quantities of gluten after the cessation of breastfeeding was responsible for the failure of development of immune tolerance to gluten in these young children. Breastfeeding as a single factor did not appear to be the major risk factor. Dietary gluten exposure earlier than 3 months or later than 6 months of age was found to be a risk factor in a longitudinal follow-up study of a cohort of children in Denver, Colorado [6]. These studies taken together suggest that there is a crucial window when tolerance to gluten occurs, and that overlapping the introduction of gluten with breastfeeding may provide the best protection against childhood celiac disease. Whether such a practice will ultimately prevent the lifetime development of celiac disease or another autoimmune disease is not known. Ivarsson and colleagues found that the risk of the development of celiac disease increases with the number of gastrointestinal infections before 6 months of age and among infants born between the months of March and July. This seasonal risk was consistently shown year to year throughout the 10-year epidemic [7,8]. Significant interest in rotavirus as a possible infectious trigger was generated following a study by Stene and colleagues [9] showing that frequent rotavirus infections increased the risk of celiac disease in 54 children under 4 years of age. Additional case reports of the development of celiac disease in children
414
BARTON & MURRAY
following rotavirus infection have been published. Of recent interest are the roles of interleukin-15 and gastrointestinal flora in the proximal small bowel in relation to precipitating disease [5,10]. The dysregulation of immune mechanisms in response to food and microbial antigens and how these mechanisms relate to gastrointestinal and systemic autoimmune diseases are the focus of ongoing research. RISK OF AUTOIMMUNE DISEASE A 1999 study by Ventura and colleagues [11] demonstrated an increased prevalence rate of autoimmune disease among 909 patients who had celiac disease when compared with controls, although no significant difference was found in a comparison with 163 patients with Crohn’s disease. Logistic regression analysis showed that the age at diagnosis of celiac disease was a significant predictor of autoimmune disease development later in life. Follow-up studies addressing the length of gluten exposure as a predictor of autoimmune disease by SategnaGuidetti and colleagues and Biagi and colleagues [12,13] did not confirm these findings. Cataldo and Marino [14] showed a higher prevalence of autoimmune disease (4.8%) among the first-degree relatives of celiac patients when compared with the first-degree relatives of healthy controls (0.86%). A subset of these first-degree relatives of patients with celiac disease was diagnosed with silent disease. A higher prevalence of autoimmune disease among this group was found in a comparison with healthy first-degree relatives of celiac disease patients (20% versus 3.8%, respectively). They suggested that this increased risk relates to the higher prevalence of silent celiac disease. Whether untreated celiac disease is responsible for an incremental incidence in autoimmune disease is unknown, although it is worthwhile for the clinician to be aware of these associations lest they occur. Generally, patients with newly diagnosed celiac disease can be reassured that their risk of developing another autoimmune disease is not greatly elevated after diagnosis and treatment. ASSOCIATIONS WITH AUTOIMMUNE DISEASE Sjogren’s Syndrome Several studies have revealed a high prevalence of celiac disease among patients with Sjogren’s syndrome. Iltanen and colleagues [15] showed that 14.7% of 34 patients with Sjogren’s syndrome had celiac disease confirmed by small bowel biopsy. A more recent study from Hungary of 111 patients with Sjogren’s syndrome found the prevalence rate of celiac disease to be 4.5 cases per 100 persons confirmed by small bowel biopsy [16]. Luft and colleagues [17] measured levels of tTg autoantibodies in the sera of patients who had systemic lupus erythematosus, Sjogren’s syndrome, and rheumatoid arthritis. Twelve percent of patients with Sjogren’s syndrome were positive for IgA tTg autoantibodies versus 4% of controls. Five of six patients with positive IgA tTg titers had villous atrophy on small bowel biopsy, confirming the diagnosis of celiac disease. Significantly lower levels of tTg autoantibodies were found in patients who had systemic lupus erythematosus, rheumatoid arthritis, and systemic sclerosis.
CELIAC DISEASE AND AUTOIMMUNITY
415
Inflammatory Arthritis Several studies have looked for celiac disease in cohorts with arthritis. Stagi and colleagues [18] found an increased prevalence of celiac disease among 151 children with juvenile idiopathic arthritis (6.7% versus 0.6% in controls). Prevalence rates reported by Lepore and colleagues [19] demonstrated an increased prevalence of biopsy-confirmed celiac disease (2.5%) among 119 children with juvenile chronic arthritis. This association does not extend to adult rheumatoid arthritis (Table 1) [20]. The prevalence of rheumatologic conditions is not convincingly elevated in the celiac disease population as a whole. Several small case series show that treatment with a gluten-free diet may be helpful in some patients with musculoskeletal symptoms. Lubrano and colleagues [21] demonstrated a 26% prevalence rate of arthritis among 200 adult patients with celiac disease. The prevalence of arthritis was significantly higher and more severe among the group on a regular diet as yet untreated when compared with patients already established on a gluten-free diet for an average of 58 months. This finding confirmed the observations in an earlier small case series that showed improvement of the joint symptoms in newly diagnosed celiac disease patients following treatment with a gluten-free diet [22]. The mechanisms underlying these arthritic complaints in celiac disease are unknown, although the alterations in intestinal permeability in untreated celiac disease may be a factor.
Table 1 Prevalence of celiac disease among arthritis patients
Study
Autoimmune disorder (number of patients)
Age (years)
Stagi, et al [18]
Juvenile idiopathic arthritis (151)
Lepore, et al [19] Francis, et al [20]
Juvenile chronic arthritis (119) Rheumatoid arthritis (60)
Study
Autoimmune disorder (number of patients)
Lubrano, et al [21]
Previously diagnosed celiac disease patients (200)
Abbreviation: SBBx, small bowel biopsy.
Prevalence of celiac disease
Celiac disease diagnosis confirmation
2.4 to 16.9 (median, 8.3)
6.7%
2 to 16 (mean, 11.5) 20 to 84 (mean, 61)
2.5%
Positive anti-EMA and IgA tTg serology, SBBx Positive anti-EMA serology, SBBx Positive anti-EMA serology; SBBx not reported
Age (years) 18 to 65 (mean, 32.8)
0.63%
Prevalence of nonerosive arthritis 26%
Effect of gluten-free diet Higher prevalence of arthritis on regular diet 41% versus 21% on gluten-free diet
416
BARTON & MURRAY
There have been a few case reports of clinically overt celiac disease in patients who have systemic lupus erythematosus [23]. The diagnosis of celiac disease consistently preceded that of systemic lupus erythematosus by an average 15 years in the pediatric cases, whereas in the adults, the diagnosis of celiac disease consistently followed the diagnosis of systemic lupus erythematosus by an average of 3.8 years [23]. The association of systemic lupus erythematosus with celiac disease is likely rare, because one study did not find hidden cases in 103 patients with systemic lupus erythematosus [24]. A more recent study from Northern Ireland did not reveal clinical evidence of systemic lupus erythematosus among a cohort of 60 patients with celiac disease [25]. Although gastrointestinal manifestations of systemic lupus erythematosus are common in both adults and children [26–28], celiac disease is relatively rare among this patient population but should be considered in the setting of malabsorption symptoms. False-positive antinuclear antibody tests may be found in patients with untreated celiac disease. ASSOCIATED AUTOIMMUNE ENDOCRINE DISEASES Addison’s Disease European studies estimate the prevalence of celiac disease among patients with autoimmune Addison’s disease to be between 1.2% and 8% [29]. More recent studies have demonstrated higher prevalence rates. A study from Ireland found five cases of celiac disease among 41 patients who had Addison’s disease [30]. A recent large Italian study showed a 5.4% prevalence of celiac disease among 109 patients with autoimmune Addison’s disease versus 0.06% among controls [31]. This association was confirmed by the analysis of more than 14,000 celiac disease patients from the Swedish national register, showing an increased risk for Addison’s disease among celiac patients [32]. Table 2 summarizes these recent studies. Autoimmune Insulin-Dependent Diabetes Mellitus AIDDM has a high prevalence rate among the patient population with celiac disease. Serologic screening for celiac disease in AIDDM populations showed a median prevalence of 4.1% among 40 studies; these studies included primarily European centers as well as six centers from the United States and United Kingdom [33]. Gastrointestinal symptoms attributed to celiac disease among AIDDM patients are generally mild. Active malabsorption may lead to unreliable carbohydrate absorption and unpredictable glucose responses to meals. This problem can lead to some degree of brittleness in diabetic controls. Folate and iron deficiency are also commonly seen. The diagnosis of celiac disease often follows within 1 year of the diagnosis of AIDDM [33]. It is has not been universally accepted that all patients who have AIDDM should be screened for celiac disease, although some advocate screening for celiac disease at intervals in children with AIDDM due primarily to the prevalence and the potential consequences of delayed diagnosis [34]. Although it is rational to expect that early treatment would prevent the complications of celiac disease in this population, few data document such benefit. There may be a slight increase in the insulin total dose with improved absorption that comes with correction of the
CELIAC DISEASE AND AUTOIMMUNITY
417
Table 2 Prevalence of celiac disease among patients with Addison’s disease
Study
Addison’s disease (number of patients)
Age (years)
Prevalence of celiac disease
Celiac disease diagnosis confirmation Positive IgA tTg and anti-EMA serology; SBBx confirmation Positive IgA tTg and anti-EMA serology; SBBx confirmation Positive anti-EMA serology; SBBx confirmation Positive tTg IgA, 3 of 4 patients confirmed by SBBx (1 silent celiac disease, 2 latent celiac disease)
O’Leary, et al [30]
41
7 to 66 (mean, 37)
12.2%
Myhre, et al [78]
76
6 to 85 (mean, 45.6)
Biagi, et al [79]
17
26 to 79 (mean, 53.9)
7.9% (6 of 76, 1 patient with known celiac disease) 5.9%
Betterle, et al [31]
109
11 to 87 (mean, 42.7)
Study Efstrom, et al [32]
Celiac disease (number of patients) Previously diagnosed celiac disease patients (14,366)
Age (years) 18 to 65 (mean, 32.8)
5.4% (6 of 109, 2 patients with known celiac disease)
Risk (hazard ratio) 11-fold increased risk of developing Addison’s disease
—
Abbreviation: SBBx, small bowel biopsy.
intestinal malabsorption. Certainly, physicians caring for patients with AIDDM should have a low threshold for celiac disease testing. Thyroid Disease A recent review of thyroid disease studies revealed a high prevalence of celiac disease among patients with autoimmune thyroid disease, Hashimoto’s disease, Graves’ disease, and pediatric autoimmune thyroid disease ranging from 2% to 7.8% (average, 4.1%) [35]. The occurrence of hyperthyroidism in adult patients with celiac disease ranges from 0.1% to 5.2% (average, 2.3%), whereas the occurrence of hypothyroidism ranges from 0% to 5.8% (average, 2.6%) [33]. Although there are insufficient data to advocate screening for celiac disease in patients with thyroid disease, a much greater awareness should lead to a low threshold for testing. An awareness of the converse relationship is also crucial, because thyroid disease may explain fatigue and additional constitutional symptoms in treated celiac disease.
418
BARTON & MURRAY
ASSOCIATED AUTOIMMUNE LIVER DISEASE Mild transaminitis in the absence of primary liver disease can occur in the setting of untreated celiac disease. Patients are often asymptomatic and develop mild periportal inflammation that resolves with a gluten-free diet. Multiple studies and case reports have demonstrated variable associations among primary biliary cirrhosis, autoimmune hepatitis, primary sclerosing cholangitis, and celiac disease. With the exception of two negative studies, a consistent association between primary biliary cirrhosis and celiac disease has been shown [36,37]. In a recent review of celiac and liver disease, the percentage of biopsyconfirmed cases of celiac disease among the primary biliary cirrhosis population ranged from 1.3% to 7% [38]. The largest study included greater than 13,800 patients with celiac disease from the Swedish national register and confirmed the association [39]. Prevalence rates of celiac disease among patients with autoimmune hepatitis range from 4% to 6.4% [40,41]. Several other autoimmune diseases are more commonly associated with autoimmune hepatitis, including rheumatoid arthritis, synovitis, Graves’ disease, autoimmune thyroiditis, and ulcerative colitis [42]. The association between autoimmune hepatitis and celiac disease appears to be less common in children than adults [43]. Initially reported in 1988, the association between celiac disease and primary sclerosing cholangitis has been investigated in a few large-scale studies. Ludvigsson and colleagues recently published one of the largest studies using the Swedish national register with the same cohort of greater than 13,800 patients with celiac disease, showing a fourfold to eightfold increased risk of primary sclerosing cholangitis. Testing for celiac disease should be included in strategies for investigating elevated transaminases. False-positive tTg autoantibodies may occur in the setting of liver disease; however, a positive anti-EMA titer is quite specific even in this setting and may be a preferable way to detect celiac disease in this regard. DERMATOLOGIC MANIFESTATIONS Abenavoli and colleagues [44] recently reviewed the dermatologic diseases associated with celiac disease, including dermatitis herpetiformis, psoriasis, vitiligo, alopecia areata, Behcet’s disease, oral lichen planus, dermatomyositis, pyoderma gangrenosum, and other more rare conditions. Dermatitis herpetiformis is strongly associated with celiac disease, tends to be more prevalent in males, and manifests as extremely pruritic, blistering lesions on the extensor surfaces of the extremities as well as back and buttocks. Data from Finnish centers have shown that one in four patients who have celiac disease is affected by dermatitis herpetiformis [45]. A recent large, retrospective study by Alonso-Llamazares and colleagues [46] found a 12.6% prevalence rate of celiac disease and a 22% prevalence rate of systemic autoimmune disorders among 263 patients with dermatitis herpetiformis. Diagnosis is made by immunofluorescent visualization of granular IgA deposits at the junction of
CELIAC DISEASE AND AUTOIMMUNITY
419
the epidermal/dermal layers in biopsy specimens taken close to but not in an affected area. Epidermal tissue transglutaminase (eTg) is considered one of the key target autoantigens. The intestinal inflammation occurring in celiac disease may be associated with increased production of antibodies directed against eTg [47,48]. Lesions regress with oral dapsone and compliance with a gluten-free diet, even in the absence of small bowel villous atrophy [45]. Recurrence of the lesions usually occurs with enteral or rectal challenge with gluten. Ojetti and colleagues recently demonstrated a high prevalence of celiac disease among 92 patients with psoriasis (4.34%). Larger scale studies have not yet been reported. Michaelsson and colleagues [49] showed improvement in psoriatic lesions following treatment with a gluten-free diet, most notably among patients with high anti-gliadin antibody titers and mild increases in duodenal epithelial lymphocytes or no enteropathy. These studies raise interesting questions in terms of the immunologic mechanisms connecting gastrointestinal and skin diseases and the opportunity for potential treatments. Multiple mechanisms have been proposed, including increased intestinal permeability, vitamin D deficiency, and gluten-induced T-cell activation [50]. NEUROLOGIC MANIFESTATIONS Approximately 10% to 12% of patients who have celiac disease show neurologic symptoms, including cerebellar ataxia, peripheral neuropathy, seizures, and myelopathy [51]. One small retrospective study found a broad spectrum of these neurologic disorders in 12% of 148 pediatric and adult patients with celiac disease [51]. Cognitive decline, dementia, myopathy, and a rarer clinical condition involving epilepsy and cerebral calcifications in relation to celiac disease have been described in case reports and series [52–55]. Several recent studies have evaluated the prevalence and clinical presentation of ataxia and peripheral neuropathy associated with celiac disease, as well as the potentially gluten-dependent neurologic conditions occurring in the absence of enteropathy. Gluten ataxia is considered a distinct disease process that can occur with or without enteropathy and is associated with high anti-gliadin antibody titers [56]. Hadjivassiliou and colleagues [56] demonstrated a 32% prevalence of anti-gliadin antibodies among patients with sporadic idiopathic ataxia. Of this gluten ataxia group, 24% had evidence of gluten-sensitive enteropathy and only 13% reported gastrointestinal symptoms. The mean age of neurologic symptom onset was 48 years, and the additional symptoms of dysarthria, upper and lower extremity ataxia, and ocular symptoms (eg, nystagmus) were reported most frequently. Anti-gliadin antibodies cross-reacting with epitopes on cerebellar Purkinje fibers was suggested as the likely mechanism. Whether gluten ataxia exists in the absence of celiac disease is controversial. Although antibodies to gliadin are common in patients with cerebellar ataxia, they are also common in other degenerative brain diseases such as Huntington’s chorea and the genetically precipitated spinocerebellar ataxias.
420
BARTON & MURRAY
Peripheral axonal neuropathy is a common neurologic manifestation of celiac disease. Prospective screening of 140 patients with idiopathic axonal neuropathy demonstrated a celiac disease prevalence of 9% [57]. Additionally in this study, among 100 patients with idiopathic neuropathy, the mean age of neurologic symptom onset was 55 years, 72% of patients had either DQ2 or DQ8 HLA typing, and significant variations in the types of neuropathy were reported. The term gluten neuropathy was proposed to describe patients with idiopathic neuropathy and positive anti-gliadin antibody titers with or without evidence of enteropathy. Often, the autoimmune markers typifying celiac disease are absent. Treatment with a gluten-free diet may improve neuropathic symptoms [58,59]; however, it has not been definitively proven that gluten exclusion alters the natural history of this neurologic state in the absence of true celiac disease. Overall, multiple neurologic manifestations of celiac disease can occur. The more common manifestations include ataxia and peripheral neuropathy; early onset dementia and cognitive decline may occasionally be due to celiac disease. The etiology of neurologic syndromes related to gluten exposure, although not directly associated with celiac disease, remains controversial. IGA DEFICIENCY The prevalence of celiac disease among patients with IgA deficiency is greater than tenfold higher when compared with that in controls [60]. Because IgA deficiency makes the detection of celiac disease based on tTg IgA titers unreliable, it is generally recommended to measure total levels of IgA to rule out a deficiency state. If IgA deficiency is present, tTg IgG is often, but not always, elevated, prompting the need for small bowel biopsies, and can be followed to measure occult gluten exposure or dietary indiscretion. Cataldo and colleagues [60] found that patients with IgA deficiency had the silent form of celiac disease more frequently and higher rates of allergic diseases, including asthma, atopic dermatitis, and gastrointestinal allergies. PERNICIOUS ANEMIA Dickey [61] showed a 12% prevalence of B12 deficiency among 159 patients with celiac disease but no increased rates of pernicious anemia. Assessments were based on gastric histology, intrinsic factor antibody testing, and serum gastrin levels. Only 2 of 19 patients displayed histologic evidence of corpus atrophy. This investigation is the largest study reported to date demonstrating no increased prevalence of pernicious anemia. Although B12 deficiency is common, it is likely due to other factors among patients with celiac disease. RELATED GASTROINTESTINAL AUTOIMMUNE DISEASES Refractory Celiac Disease Some cases of celiac disease do not respond to the removal of dietary gluten, resulting in persistent villous atrophy, ongoing clinical symptoms, and an increased risk for malignancy in the setting of continued immune activation.
CELIAC DISEASE AND AUTOIMMUNITY
421
Refractory celiac disease is a relatively rare complication occurring in approximately 2% to 5% of patients. It is classified as persistent or recurrent symptoms of malabsorption and enteropathy. Refractory celiac disease is divided into two key categories, type I and type II. This distinction is important in terms of predicted morbidity and mortality, strategies for immunosuppressive therapy, and implications for EATL screening. The clinical and histologic presentations of these two groups bear similarities in the initial stages, specifically, the persistence of gastrointestinal symptoms, villous atrophy despite strict compliance with a gluten-free diet for greater than 12 months, and normal tTg antibody titers. The distinction between type I and II refractory celiac disease and its implications on overall patient survival are based on the monoclonal transformation of intraepithelial T-cell populations from the polyclonal phenotype characteristic of type I refractory celiac disease. Diagnosis is made via immunohistochemical analysis of specific T-cell surface markers and PCR-based T-cell receptor gene rearrangement analysis. Referral to a specialized center for this testing and management are warranted. Type II refractory celiac disease increases the risk of developing ulcerative jejunitis and EATL. Capsule endoscopy is limited by the inability to obtain tissue biopsy, although double-balloon enteroscopy has recently been found to be effective in diagnosing or excluding EATL with or without ulcerative jejunitis in patients with refractory celiac disease [62]. One recent study found PET scanning to be of additional benefit in diagnosing early EATL and identifying extraluminal manifestations of lymphoma [63]. A combination PET-CT modality may offer even greater sensitivity for early detection. Additional risk factors associated with EATL include a significant delay in diagnosis, DQ2 homozygous HLA status, and male sex [64]. Given the overall high mortality of EATL, early diagnosis and intervention are areas of current interest. Refractory celiac disease type I is generally steroid responsive which is the first line therapy. Azathioprine may be used following induction of clinical remission. Corticosteroid treatment is used for refractory celiac disease type II, although steroid-resistant disease may warrant more aggressive immunosuppression regimens, including infliximab infusions [65]. Among patients who have refractory celiac disease type II, the risk of accelerating the development of a malignant T-cell clone with chronic immunosuppression makes this therapy controversial. Budesonide is effective in improving clinical symptoms in refractory celiac disease I and II groups, although histologic recovery of the small bowel architecture was not observed in a recent study [66]. Budesonide has significant first-pass metabolism, largely minimizing the systemic side effects associated with chronic steroid therapy. Additional etiologies should be excluded before diagnosing refractory celiac disease, including lymphocytic colitis, autoimmune enteropathy, pancreatic insufficiency, collagenous enterocolitis, bacterial overgrowth, and lactase deficiency. Screening and treatment for nutritional deficiencies including zinc, copper, B12, folate, iron indices, and albumin are generally recommended. In
422
BARTON & MURRAY
the event of severe hypoalbuminemia with associated edematous states and opportunistic infections, parenteral nutrition is frequently required. Autoimmune Enteropathy Autoimmune enteropathy is a rare disease predominantly occurring in children, although cases of adult onset have been reported. One of the initial reports of adult-onset autoimmune enteropathy was in 1997 by Corazza and colleagues [67]. They described two of four patients with DQ2 HLA typing who were clinically nonresponsive to gluten restriction and who were found to be positive for anti-enterocyte antibodies. Additional reported cases of adults with autoimmune enteropathy or autoimmune enteropathy variants include autoimmune enteropathy associated with atrophic gastritis and colitis, autoimmune enterocolitis, autoimmune enteropathy in association with systemic autoimmune disease (sicca syndrome, thyroiditis), and autoimmune enteropathy associated with rheumatoid arthritis. The pathophysiology of autoimmune enteropathy is unclear. Leon and colleagues propose that TNF-alpha production by activated T cells in the bowel mucosa and a unique subset of CD4þ T cells in the lamina propria may be important in the pathogenesis of disease. Akram and colleagues [68] recently published a case series describing the clinical and histologic findings in 15 adults diagnosed with autoimmune enteropathy and proposed criteria for diagnosis. This series included the frequency of associated autoimmune diseases and the occurrence of autoimmune enterocolitis, a variant of autoimmune enteropathy. This report is the largest case series to date on this rare disorder. The cohort included patients with a median age of 55 years who were predominantly Caucasian (87%) who had subtotal villous atrophy of the proximal small bowel, increased lymphocytosis within the lamina propria, positive anti-enterocyte antibody or anti-goblet cell antibody serology, and decreased numbers of intraepithelial lymphocytes. The clinical presentation included protracted diarrhea, weight loss, lack of responsiveness to a gluten-restricted diet, and exclusion of additional secondary causes of villous atrophy. Proposed criteria for the diagnosis of adult autoimmune enteropathy from the Mayo Clinic study include the following: (1) chronic diarrhea for more than 6 weeks in duration, (2) clinical signs and symptoms of malabsorption, (3) specific histology on small bowel biopsy, (4) the exclusion of additional secondary diagnoses causing villous atrophy, and (5) positive anti-enterocyte or goblet cell antibody serologies. The first four criteria are necessary for diagnosis, although negative serologies do not exclude the diagnosis of autoimmune enteropathy [68]. Anti-enterocyte antibodies have not been reported in patients with inflammatory bowel disease or celiac disease [68]. Sixty percent of patients in this study responded to high-dose steroids, although 66% of this group became steroid dependent or required adjunct immunosuppressive therapy. Two of the 15 patients were treated with infliximab with complete resolution of symptoms.
CELIAC DISEASE AND AUTOIMMUNITY
423
Given the rarity of autoimmune enteropathy, specific associations with autoimmune disease would be difficult to establish. Systemic autoimmune diseases reported in the previous study included rheumatoid arthritis, hypothyroidism, psoriatic arthritis, autoimmune gastroenterocolitis, common variable immunodeficiency, and myasthenia gravis, among others. One of the initial pediatric studies by Hill and colleagues [69] in 1991 reported varying degrees of colonic histopathologic changes in all eight patients tested in a comparison with controls. The age of symptom onset ranged from 2 to 12 months, with a consistent interval of several months before diagnosis. Among the eight patients and their first- or second-degree relatives, additional autoimmune diseases were present, including hypothyroidism, interstitial nephropathy, chronic hepatitis, AIDDM, and vitiligo. Collagenous Enteritis (Collagenous Sprue) Collagenous enteritis is closely related to celiac disease both histologically and clinically. Patients present with classic symptoms of malabsorption and histologic evidence of villous atrophy on small bowel biopsy. Whether collagenous enteropathy is a variant of celiac disease is controversial. Features that distinguish this disorder from celiac disease are the presence of a thickened subepithelial collagen band and nonresponsiveness to dietary gluten restriction. Positive anti-EMA serology and an increased risk for T- and B-cell lymphomas are additional shared features of both diseases [70–72]. The causes of collagenous enteritis are not known, and treatment usually requires prolonged steroid therapy. Collagenous enteritis may occur in conjunction with collagenous colitis [73,74]. Microscopic Colitis Recent case reports and series have shown simultaneous occurrence of celiac disease and collagenous colitis [75,76]. The series by Freeman showed a 22.9% prevalence of celiac disease among 35 patients with collagenous colitis. There is a clear association between celiac disease and microscopic colitis including both collagenous and lymphocytic colitis variants. Approximately 30% of patients with celiac disease may have evidence of microscopic colitis on colon histology, whereas between 2% and 10% of patients with microscopic colitis have villous atrophy suggestive of celiac disease [77]. A low threshold for celiac disease screening among these patient populations is warranted. SUMMARY The strongest associations between celiac disease and systemic autoimmune diseases include Addison’s disease, autoimmune thyroid disease, AIDDM, Sjogren’s syndrome, primary biliary cirrhosis, and autoimmune hepatitis (Fig. 1). Although more cost-benefit analyses need to be done to support celiac disease screening of these enriched populations and to facilitate the development of diagnostic algorithms, maintaining a low threshold for screening and high clinical suspicion among these groups is recommended. There is clear
424
BARTON & MURRAY
*Microscopic colitis Collagenous enteritis/colitis Transaminitis *Primary biliary cirrhosis *Autoimmune hepatitis *IgA deficiency
*Dermatitis herpetiformis Psoriasis Dermatomyositis Alopecia areata Vitiligo GI-related diseases
Endocrine-related diseases
Celiac Diseaseenteropathy Gut permeability Humoral immunity
Dermatological-related diseases
Rheumatological and Neurological-related diseases *Autoimmune diabetes mellitus *Addison’s disease *Autoimmune thyroid disease
*Sjogren’s syndrome Inflammatory arthritis SLE Gluten ataxia Peripheral neuropathies
Fig. 1. Autoimmune and inflammatory diseases in relation to celiac disease. *Strongest associations.
association between dermatitis herpetiformis and celiac disease such that a definitive diagnosis of dermatitis herpetiformis by direct immunofluorescence is sufficient evidence to justify initiation of a gluten-free diet, and the presence of positive tTg autoantibodies in the context of dermatitis herpetiformis implies substantial villous atrophy. Testing among arthritic patients is less clear given the limited large-scale studies, prevalence differences between adult and pediatric populations, and the varying types of arthritis studied to date. Maintaining a lower threshold for screening among pediatric versus adult populations is recommended. Screening for celiac disease among patients with systemic lupus erythematosus is not recommended, although it should be considered in the differential diagnosis of lupus patients with gastrointestinal symptoms. Neurologic patients with the broad spectrum of clinical symptoms described herein should be considered for celiac disease screening and also evaluated for gluten-dependent conditions occurring in the absence of enteropathy. Additional autoimmune diseases of the small bowel should be considered in the absence of a clinical response to dietary gluten restriction and in view of the lack of recommended diagnostic criteria for celiac disease. The connection between gut and systemic autoimmune diseases and the environmental factors affecting development is an evolving area of research. Further understanding of the underlying immune mechanisms will allow for the development of new therapies targeting intestinal permeability, mucosal regulatory T cells, and neutralization of key inflammatory cytokines. The immunologic basis for celiac disease provides a great opportunity to understand the interplay between genetic and environmental factors in autoimmune diseases in general.
CELIAC DISEASE AND AUTOIMMUNITY
425
Acknowledgments The authors thank Dr. Eric V. Marietta for his review of this manuscript. References [1] Fasano A, Berti I, Gerarduzzi T, et al. Prevalence of celiac disease in at-risk and not-at-risk groups in the United States: a large multicenter study. Arch Intern Med 2003;163(3): 286–92. [2] Wahnschaffe U, Schulzke JD, Zeitz M, et al. Predictors of clinical response to gluten-free diet in patients diagnosed with diarrhea-predominant irritable bowel syndrome. Clin Gastroenterol Hepatol 2007;5(7):844–50. [3] Farrell RJ, Kelly CP. Celiac sprue. N Engl J Med 2002;346(3):180–8. [4] Cooper BT, Holmes GK, Cooke WT. Coeliac disease and immunological disorders. Br Med J 1978;1(6112):537–9. [5] Jabri B, Sollid LM. Mechanisms of disease: immunopathogenesis of celiac disease. Nat Clin Pract Gastroenterol Hepatol 2006;3(9):516–25. [6] Norris JM, Barriga K, Hoffenberg EJ, et al. Risk of celiac disease autoimmunity and timing of gluten introduction in the diet of infants at increased risk of disease. JAMA 2005;293(19): 2343–51. [7] Ivarsson A. The Swedish epidemic of coeliac disease explored using an epidemiological approach–some lessons to be learnt. Best Pract Res Clin Gastroenterol 2005;19(3): 425–40. [8] Ivarsson A, Hernell O, Nystrom L, et al. Children born in the summer have increased risk for coeliac disease. J Epidemiol Community Health 2003;57(1):36–9. [9] Stene LC, Honeyman MC, Hoffenberg EJ, et al. Rotavirus infection frequency and risk of celiac disease autoimmunity in early childhood: a longitudinal study. Am J Gastroenterol 2006;101(10):2333–40. [10] Forsberg G, Fahlgren A, Horstedt P, et al. Presence of bacteria and innate immunity of intestinal epithelium in childhood celiac disease. Am J Gastroenterol 2004;99(5):894–904. [11] Ventura A, Magazzu G, Greco L. Duration of exposure to gluten and risk for autoimmune disorders in patients with celiac disease: SIGEP study group for autoimmune disorders in celiac disease. Gastroenterology 1999;117(2):297–303. [12] Sategna-Guidetti C, Solerio E, Scaglione N, et al. Duration of gluten exposure in adult coeliac disease does not correlate with the risk for autoimmune disorders. Gut 2001;49(4): 502–5. [13] Biagi F, Pezzimenti D, Campanella J, et al. Gluten exposure and risk of autoimmune disorders. Gut 2002;51(1):140–1. [14] Cataldo F, Marino V. Increased prevalence of autoimmune diseases in first-degree relatives of patients with celiac disease. J Pediatr Gastroenterol Nutr 2003;36(4):470–3. [15] Iltanen S, Collin P, Korpela M, et al. Celiac disease and markers of celiac disease latency in patients with primary Sjogren’s syndrome. Am J Gastroenterol 1999;94(4):1042–6. [16] Szodoray P, Barta Z, Lakos G, et al. Coeliac disease in Sjogren’s syndrome–a study of 111 Hungarian patients. Rheumatol Int 2004;24(5):278–82. [17] Luft LM, Barr SG, Martin LO, et al. Autoantibodies to tissue transglutaminase in Sjogren’s syndrome and related rheumatic diseases. J Rheumatol 2003;30(12):2613–9. [18] Stagi S, Giani T, Simonini G, et al. Thyroid function, autoimmune thyroiditis and coeliac disease in juvenile idiopathic arthritis. Rheumatology (Oxford) 2005;44(4):517–20. [19] Lepore L, Martelossi S, Pennesi M, et al. Prevalence of celiac disease in patients with juvenile chronic arthritis. J Pediatr 1996;129(2):311–3. [20] Francis J, Carty JE, Scott BB. The prevalence of coeliac disease in rheumatoid arthritis. Eur J Gastroenterol Hepatol 2002;14(12):1355–6. [21] Lubrano E, Ciacci C, Ames PR, et al. The arthritis of coeliac disease: prevalence and pattern in 200 adult patients. Br J Rheumatol 1996;35(12):1314–8.
426
BARTON & MURRAY
[22] Bourne JT, Kumar P, Huskisson EC, et al. Arthritis and coeliac disease. Ann Rheum Dis 1985;44(9):592–8. [23] Mirza N, Bonilla E, Phillips PE. Celiac disease in a patient with systemic lupus erythematosus: a case report and review of literature. Clin Rheumatol 2007;26(5):827–8. [24] Rensch MJ, Szyjkowski R, Shaffer RT, et al. The prevalence of celiac disease autoantibodies in patients with systemic lupus erythematosus. Am J Gastroenterol 2001;96(4):1113–5. [25] Courtney PA, Patterson RN, Lee RJ, et al. Systemic lupus erythematosus and coeliac disease. Lupus 2004;13(3):214. [26] Richer O, Ulinski T, Lemelle I, et al. Abdominal manifestations in childhood-onset systemic lupus erythematosus. Ann Rheum Dis 2007;66(2):174–8. [27] Hallegua DS, Wallace DJ. Gastrointestinal manifestations of systemic lupus erythematosus. Curr Opin Rheumatol 2000;12(5):379–85. [28] Sultan SM, Ioannou Y, Isenberg DA. A review of gastrointestinal manifestations of systemic lupus erythematosus. Rheumatology (Oxford) 1999;38(10):917–32. [29] Betterle C, Dal Pra C, Mantero F, et al. Autoimmune adrenal insufficiency and autoimmune polyendocrine syndromes: autoantibodies, autoantigens, and their applicability in diagnosis and disease prediction. Endocr Rev 2002;23(3):327–64. [30] O’Leary C, Walsh CH, Wieneke P, et al. Coeliac disease and autoimmune Addison’s disease: a clinical pitfall. QJM 2002;95(2):79–82. [31] Betterle C, Lazzarotto F, Spadaccino AC, et al. Celiac disease in North Italian patients with autoimmune Addison’s disease. Eur J Endocrinol 2006;154(2):275–9. [32] Elfstrom P, Montgomery SM, Kampe O, et al. Risk of primary adrenal insufficiency in patients with celiac disease. J Clin Endocrinol Metab 2007;92(9):3595–8. [33] Collin P, Kaukinen K, Valimaki M, et al. Endocrinological disorders and celiac disease. Endocr Rev 2002;23(4):464–83. [34] Hill ID, Dirks MH, Liptak GS, et al. Guideline for the diagnosis and treatment of celiac disease in children: recommendations of the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr 2005;40:1–19. [35] Ch’ng CL, Jones MK, Kingham JG. Celiac disease and autoimmune thyroid disease. Clin Med Res 2007;5(3):184–92. [36] Chatzicostas C, Roussomoustakaki M, Drygiannakis D, et al. Primary biliary cirrhosis and autoimmune cholangitis are not associated with coeliac disease in Crete. BMC Gastroenterol 2002;2:5. [37] Habior A, Lewartowska A, Orlowska J, et al. Association of coeliac disease with primary biliary cirrhosis in Poland. Eur J Gastroenterol Hepatol 2003;15(2):159–64. [38] Rubio-Tapia A, Murray JA. The liver in celiac disease. Hepatology 2007;46(5):1650–8. [39] Ludvigsson JF, Elfstrom P, Broome U, et al. Celiac disease and risk of liver disease: a general population-based study. Clin Gastroenterol Hepatol 2007;5(1):63–9, e61. [40] Volta U, De Franceschi L, Molinaro N, et al. Frequency and significance of anti-gliadin and anti-endomysial antibodies in autoimmune hepatitis. Dig Dis Sci 1998;43(10): 2190–5. [41] Villalta D, Girolami D, Bidoli E, et al. High prevalence of celiac disease in autoimmune hepatitis detected by anti-tissue transglutaminase autoantibodies. J Clin Lab Anal 2005;19(1):6–10. [42] Czaja AJ. Autoimmune liver disease. In: Zakim DBT, editor, Hepatology: a textbook of liver disease, vol. 2. Philadelphia: Saunders; 2003. p. 1163–202. [43] Bridoux-Henno L, Dabadie A, Briard D, et al. A case of celiac disease presenting with autoimmune hepatitis and erythroblastopenia. J Pediatr Gastroenterol Nutr 2001;33(5): 616–9. [44] Abenavoli L, Proietti I, Leggio L, et al. Cutaneous manifestations in celiac disease. World J Gastroenterol 2006;12(6):843–52. [45] Collin P, Reunala T. Recognition and management of the cutaneous manifestations of celiac disease: a guide for dermatologists. Am J Clin Dermatol 2003;4(1):13–20.
CELIAC DISEASE AND AUTOIMMUNITY
427
[46] Alonso-Llamazares J, Gibson LE, Rogers RS III. Clinical, pathologic, and immunopathologic features of dermatitis herpetiformis: review of the Mayo Clinic experience. Int J Dermatol 2007;46(9):910–9. [47] Sardy M, Karpati S, Merkl B, et al. Epidermal transglutaminase (TGase 3) is the autoantigen of dermatitis herpetiformis. J Exp Med 2002;195(6):747–57. [48] Marietta EV, Camilleri MJ, Castro LA, et al. Transglutaminase autoantibodies in dermatitis herpetiformis and celiac sprue. J Invest Dermatol 2008;128:332–5. [49] Michaelsson G, Gerden B, Hagforsen E, et al. Psoriasis patients with antibodies to gliadin can be improved by a gluten-free diet. Br J Dermatol 2000;142(1):44–51. [50] Abenavoli L, Leggio L, Gasbarrini G, et al. Celiac disease and skin: psoriasis association. World J Gastroenterol 2007;13(14):2138–9. [51] Vaknin A, Eliakim R, Ackerman Z, et al. Neurological abnormalities associated with celiac disease. J Neurol 2004;251(11):1393–7. [52] Hu WT, Murray JA, Greenaway MC, et al. Cognitive impairment and celiac disease. Arch Neurol 2006;63(10):1440–6. [53] Collin P, Pirttila T, Nurmikko T, et al. Celiac disease, brain atrophy, and dementia. Neurology 1991;41(3):372–5. [54] Gobbi G. Coeliac disease, epilepsy and cerebral calcifications. Brain Dev 2005;27(3): 189–200. [55] Hadjivassiliou M, Chattopadhyay AK, Grunewald RA, et al. Myopathy associated with gluten sensitivity. Muscle Nerve 2007;35(4):443–50. [56] Hadjivassiliou M, Grunewald R, Sharrack B, et al. Gluten ataxia in perspective: epidemiology, genetic susceptibility and clinical characteristics. Brain 2003;126(Pt 3):685–91. [57] Hadjivassiliou M, Grunewald RA, Kandler RH, et al. Neuropathy associated with gluten sensitivity. J Neurol Neurosurg Psychiatry 2006;77(11):1262–6. [58] Rigamonti A, Magi S, Venturini E, et al. Celiac disease presenting with motor neuropathy: effect of gluten free-diet. Muscle Nerve 2007;35(5):675–7. [59] Hadjivassiliou M, Kandler RH, Chattopadhyay AK, et al. Dietary treatment of gluten neuropathy. Muscle Nerve 2006;34(6):762–6. [60] Cataldo F, Marino V, Ventura A, et al. Prevalence and clinical features of selective immunoglobulin A deficiency in coeliac disease: an Italian multicentre study. Italian Society of Paediatric Gastroenterology and Hepatology (SIGEP) and ‘‘club del tenue’’ working groups on coeliac disease. Gut 1998;42(3):362–5. [61] Dickey W. Low serum vitamin B12 is common in coeliac disease and is not due to autoimmune gastritis. Eur J Gastroenterol Hepatol 2002;14(4):425–7. [62] Hadithi M, Mallant M, Oudejans J, et al. 18F-FDG PET versus CT for the detection of enteropathy-associated T-cell lymphoma in refractory celiac disease. J Nucl Med 2006;47(10): 1622–7. [63] Hadithi M, Al-toma A, Oudejans J, et al. The value of double-balloon enteroscopy in patients with refractory celiac disease. Am J Gastroenterol 2007;102(5):987–96. [64] Al-Toma A, Verbeek WH, Hadithi M, et al. Survival in refractory coeliac disease and enteropathy-associated T-cell lymphoma: retrospective evaluation of single-centre experience. Gut 2007;56(10):1373–8. [65] Daum S, Cellier C, Mulder CJ. Refractory coeliac disease. Best Pract Res Clin Gastroenterol 2005;19(3):413–24. [66] Brar P, Lee S, Lewis S, et al. Budesonide in the treatment of refractory celiac disease. Am J Gastroenterol 2007;102(10):2265–9. [67] Corazza GR, Biagi F, Volta U, et al. Autoimmune enteropathy and villous atrophy in adults. Lancet 1997;350(9071):106–9. [68] Akram S, Murray JA, Pardi DS, et al. Adult autoimmune enteropathy: Mayo Clinic Rochester experience. Clin Gastroenterol Hepatol 2007;5(11):1282–90. [69] Hill SM, Milla PJ, Bottazzo GF, et al. Autoimmune enteropathy and colitis: is there a generalised autoimmune gut disorder? Gut 1991;32(1):36–42.
428
BARTON & MURRAY
[70] Freeman HJ. Hyposplenism, antiendomysial antibodies and lymphocytic colitis in collagenous sprue. Can J Gastroenterol 1999;13(4):347–50. [71] Freeman HJ. Collagenous mucosal inflammatory diseases of the gastrointestinal tract. Gastroenterology 2005;129(1):338–50. [72] Freeman HJ. Collagenous sprue associated with an extensive T-cell lymphoma. J Clin Gastroenterol 2003;36(2):144–6. [73] Schreiber FS, Eidt S, Hidding M, et al. Collagenous duodenitis and collagenous colitis: a short clinical course as evidenced by sequential endoscopic and histologic findings. Endoscopy 2001;33(6):555. [74] McCashland TM, Donovan JP, Strobach RS, et al. Collagenous enterocolitis: a manifestation of gluten-sensitive enteropathy. J Clin Gastroenterol 1992;15(1):45–51. [75] Freeman HJ. Collagenous colitis as the presenting feature of biopsy-defined celiac disease. J Clin Gastroenterol 2004;38(8):664–8. [76] Smith P, Bishop P, Whorwell PJ. Collagenous colitis, ulcerative colitis, coeliac disease and hyperparathyroidism in one patient: implications for the management of collagenous colitis. Eur J Gastroenterol Hepatol 2005;17(11):1239–42. [77] Pardi DS. Microscopic colitis. Mayo Clin Proc 2003;78(5):614–6. [78] Myhre AG, Aarsetoy H, Undlien DE, et al. High frequency of coeliac disease among patients with autoimmune adrenocortical failure. Scand J Gastroenterol 2003;38(5): 511–5. [79] Biagi F, Campanella J, Soriani A, et al. Prevalence of coeliac disease in Italian patients affected by Addison’s disease. Scand J Gastroenterol 2006;41(3):302–5.
Gastroenterol Clin N Am 37 (2008) 439–460
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Autoimmune Pancreatitis Timothy B. Gardner, MD, Suresh T. Chari, MD* Miles and Shirley Fiterman Center for Digestive Diseases, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA
A
utoimmune pancreatitis (AIP) represents a unique subset of chronic inflammatory pancreatic disease with distinct clinical, morphologic, and histopathologic features that typically responds dramatically to steroid treatment [1–3]. Named in 1995 by Yoshida and colleagues [4,5], the clinical characteristics of the disease had been described as early as 1961. It took several decades for it to be accepted as a distinctive entity; in fact, as recently as 2003, the American Pancreatic Association still debated its existence [6]. In the last 5 years, however, AIP has generated significant interest from clinicians and researchers, resulting in a better understanding of its pathophysiology and clinical manifestations. Although still a rare disease, it is increasingly being recognized and its existence is no longer in doubt. This article details the full spectrum of AIP. It focuses on the history, definition, and pathophysiology of this disease. Clinical, radiographic, and histologic features are discussed as are the current diagnostic classification recommendations. Finally, this article outlines current therapeutic approaches and suggests future areas of study. HISTORICAL MILESTONES AIP is a relatively newly characterized disease entity, with much of our knowledge gained in only the last decade. Sarles and colleagues [5] in 1961 were the first to describe an autoimmune phenomenon in relation to ‘‘sclerosis of the pancreas’’. Although treatment with immune-modulating therapy was not suggested at that time, this spurred interest in this entity over the next several decades. Several terms were then used to subsequently describe the disease, including ‘‘chronic sclerosing pancreatitis,’’ ‘‘lymphoplasmacytic sclerosing pancreatitis,’’ ‘‘nonalcoholic duct-destructive pancreatitis,’’ ‘‘sclerosing pancreatitis,’’ ‘‘sclerosing pancreaticocholangitis,’’ and ‘‘autoimmune chronic pancreatitis’’ [7–11]. In 1991, Kawaguchi and colleagues [12] described a variant of cholangitis extensively involving the pancreas. This description was quickly followed by *Corresponding author. Division of Gastroenterology and Hepatology, College of Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. E-mail address: chari.
[email protected] (S.T. Chari). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.004
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
440
GARDNER & CHARI
several other reports describing other organ involvement in patients who had AIP [7,13–16]. Ito and colleagues [17] subsequently described the first three cases of AIP that were successfully treated with corticosteroids. In a seminal study published in 2001, Hamano and colleagues [10] reported that elevated serum IgG4 levels were associated with sclerosing pancreatitis and that treatment with corticosteroid therapy successfully decreased these levels. The first diagnostic criteria were proposed by the Japanese Pancreas Society in 2001 and subsequently modified in 2006 [18,19]. Chari and colleagues [1] in 2006 proposed new diagnostic criteria, which included histologic, imaging, and serologic characteristics and other organ involvement and response to corticosteroids. DEFINITION AIP has been defined as ‘‘the pancreatic manifestation of a systemic fibroinflammatory disease which affects not only the pancreas but also various other organs including bile duct, salivary glands, the retroperitoneum and lymph nodes. Organs affected by AIP have a lymphoplasmacytic infiltrate rich in IgG4 positive cells and the inflammatory process responds to steroid therapy’’ [1]. The systemic disease of which AIP is a manifestation has been called IgG4related systemic disease (ISD) in recognition that all organs afflicted show a dense lymphoplasmacytic infiltrate rich in IgG4-positive cells [20]. The detection of elevated serum IgG4 levels in 5% to 10% of subjects who do not have AIP and tissue infiltration with IgG4-positive cells in more than 20% of pancreatic cancer patients questions the pathogenic role of IgG4 in this systemic disorder, however [21]. EPIDEMIOLOGY Almost all data describing the epidemiology of AIP come from Japan. In 2002, Nishimori and colleagues [22] randomly surveyed Japanese hospitals on how many patients they had who had pancreatitis in 2002 who fulfilled the diagnostic criteria for AIP as proposed by the Japan Pancreas Society. Based on this survey, the prevalence of patients who had AIP in Japan was estimated to be 0.82 per 100,000. AIP was predominantly seen in men past middle age (older than 45 years). Other Japanese series have described the prevalence of disease between 5% and 6% of all patients who have chronic pancreatitis [23,24]. The prevalence of AIP in the United States is unknown, because no extensive population-based studies have been performed. Because AIP often mimics pancreatic cancer in its initial presentation, the best estimates of prevalence of AIP are in patients undergoing resection for presumed pancreatic cancer because of obstructive jaundice or a pancreatic mass. In three recent studies, 43 of 1808 (2.4%) pancreatic resections were reported to have lymphoplasmacytic sclerosing pancreatitis (LPSP) on histologic examination of the resected specimen [8,25,26]. Another retrospective evaluation of 245 pathology specimens at the Mayo Clinic from patients who underwent resection for benign
AUTOIMMUNE PANCREATITIS
441
pancreatic disease revealed that 27 (11%) represented AIP with a ‘‘tumefactive’’ presentation [27]. Males are nearly twice as likely as females to develop AIP and the average age of onset is in the fifth decade [22]. In fact, 85% of patients are older than 50 years of age. One study that has evaluated the clinical characteristics of younger (less than 40 years old) patients who had AIP found that compared with older patients, the young are more likely to present with features of abdominal pain and elevated amylase [28]. PATHOGENESIS Currently, the pathogenesis of AIP is unknown, although it almost certainly reflects an immune-mediated process. Genetic susceptibility to AIP has been linked to both the HLA-DRB1*0405-DQB1*0401 in the class II and the ABCF1 proximal to C3-2-11 telomeric of HLA-E in the class I regions [29]. In addition, a recent report described the genetic association of Fc receptorlike 3 polymorphisms with AIP in Japanese patients [30]. The trigger for AIP remains elusive. It is hypothesized that HLA-DR antigens on the pancreatic ductal and acinar cells may serve as antigens recognized by CD4þ producing interferon-c and CD8þ T lymphocytes, leading to subsequent inflammation. Polymorphisms of the cytotoxic T lymphocyte-associated antigen 4, a key negative regulator of the T-cell immune response, have been demonstrated in patients who have AIP [31]. An etiologic role for antigenic Helicobacter pylori infection by way of molecular mimicry has also been proposed [32]. Like many immune-mediated diseases, AIP has been linked to many other autoimmune conditions, such as Sjo¨gren syndrome, retroperitoneal fibrosis, and primary sclerosing cholangitis (PSC) [3,33–38]. The work by Kamisawa and colleagues, however [39], has shown that these associated conditions are in fact manifestations of a new clinicopathological entity, IgG4-related systemic disease (ISD) that is characterized by tissue infiltration with abundant IgG4positive plasma cells. These associated conditions often mimic other wellknown diseases. For example, unlike Sjo¨gren syndrome, the salivary and lacrimal gland pathology associated with AIP, called chronic sclerosing sialadenitis, is not associated with rheumatoid arthritis, is not usually associated with SSA and SSB antibodies, and responds to corticosteroids. In the past this condition has also been called Mikulicz disease [40] or Kuttner tumor [41]. Similarly, the biliary involvement in ISD has recently been termed IgG4-associated cholangitis (IAC) [42] and may resemble cholangiocarcinoma or PSC [43]. Other possible manifestations of ISD include Reidel thyroiditis and IgG4-associated nephritis [44]. Whether AIP is truly associated with any other distinct autoimmune disorder needs confirmation in case-control studies. CLINICAL FEATURES Most patients who have AIP are male and older than 50 years [22,45,46]. The male/female predominance is approximately 2:1 [47]. Female patients are more
442
GARDNER & CHARI
prevalent in the subset of AIP associated with parotid gland involvement, however [35]. Although the age of onset is typically in the sixth decade, AIP has been described in patients in their 30s [28]. Patients affected by AIP have been described in Eastern and Western first-world countries and in developing countries. The authors have seen histologically proven AIP in many patients from India and in Indian immigrants to the United States [23,48]. AIP has also been described as an incidental finding at autopsy [49]. AIP has diverse clinical presentations, which may be classified in relation to onset of disease as acute or late and by the organs affected as predominantly pancreatic or extrapancreatic presentations. The most common acute presentation is with painless obstructive jaundice [15,47,50]. Very uncommonly patients may present with typical acute pancreatitis. In the postacute phase AIP may present or be incidentally discovered because of a persistent pancreatic mass, atrophic pancreas with or without calcification, or pancreatic steatorrhea. Although patients who have AIP may have radiologic evidence of pancreatic calcification or atrophy, unlike in usual chronic pancreatitis it is painless. The pancreatic manifestations of AIP thus can mimic pancreatic cancer, acute pancreatitis, painless chronic pancreatitis, or unexplained pancreatic functional insufficiency. The extrapancreatic manifestations are equally diverse and may be seen simultaneously with its pancreatic manifestations, may precede them, or may occur many years later when the pancreatic disease may or may not be symptomatic. The most common extrapancreatic organ involved is the biliary tree, wherein distal biliary involvement mimics pancreatic cancer–related stricture. More proximal involvement may trigger suspicion of cholangiocarcinoma or PSC. Other well-described manifestations include salivary gland involvement resembling Sjo¨gren syndrome, mediastinal adenopathy resembling sarcoidosis, retroperitoneal fibrosis, and tubulointerstitial nephritis [16,44,51–53]. The acute presentation of AIP is with obstructive jaundice and it has received the most attention because it closely mimics the presentation of pancreatic cancer. The inflamed gland can often have the appearance of a pancreatic head mass, suggesting pancreatic adenocarcinoma. In fact, before recognition of AIP as a clinical entity, patients frequently underwent pancreatic resection for suspicion of pancreatic adenocarcinoma [6,27,46]. Current diagnostic criteria have significantly enhanced our ability to recognize AIP and the Japanese and Korean diagnostic criteria are designed exclusively to diagnose AIP in this setting and distinguish it from pancreatic cancer. Jaundice is usually secondary to entrapment of the intrapancreatic bile duct in an inflamed gland [54–56]; however, a true IgG4-associated distal cholangitis is also seen on histology in some patients [57]. Although jaundice is the most common clinical symptom at presentation occurring in up to 80% of patients, multiple other symptoms have been described, such as abdominal pain, back pain, recurrent vomiting, and weight loss [58,59]. Although described, acute or recurrent acute pancreatitis is a rare presenting complaint in AIP,
AUTOIMMUNE PANCREATITIS
443
and may be more common in younger patients [28]. Diabetes mellitus has been reported at presentation in AIP and up to 50% of patients may present with glucose intolerance [6,60]. About half of patients treated with corticosteroids have subsequent improvement in their glucose intolerance, and complete resolution of diabetes mellitus following treatment has been reported [61,62]. Virtually all patients who have AIP have narrowing of dorsal or ventral pancreatic ducts on pancreatography [63,64]. Idiopathic pancreatic steatorrhea should also prompt concern for AIP, especially in elderly males. Table 1 summarizes the common, atypical, and rare clinical features encountered in AIP.
HISTOPATHOLOGY On gross examination, the pancreas in AIP is often noted to be indurated and firm [6]. Classically, the predominant histologic feature of AIP has been dense infiltration of the periductal space with plasma cells and T lymphocytes. Associated with this infiltrate is acinar destruction, obliterative phlebitis involving the major and minor veins, and storiform or ‘‘whirling’’ fibrosis of the
Table 1 Clinical features of autoimmune pancreatitis Feature
Typical (>50%)
Presenting complaint
Obstructive jaundice
Histology
LPSP
Pancreatic imaging
Diffusely enlarged gland with delayed and rim enhancement, irregular narrowed pancreatic duct Elevated serum IgG4 levels Bile duct, kidneys, lymphadenopathy Complete
Serology Other organ involvement Response to steroids
Less common (10%–50%) Abdominal pain, weight loss, diabetes, steatorrhea IDCP
Parenchymal atrophy, intraductal calcification
Normal serum IgG4 levels Retroperitoneum, salivary gland Incomplete
Rare (<10%) Clinical acute pancreatitis, asymptomatic Extensive fibrosis, minimal inflammation Pseudocyst
Mesenteritis, inflammatory bowel disease Refractory to steroids
Abbreviation: IDCP, idiopathic duct-centric chronic pancreatitis. Data from Lara LP, Chari ST. Autoimmune pancreatitis. Curr Gastroenterol Rep 2005;7:101–6.
444
GARDNER & CHARI
pancreatic parenchyma, which can extend to contiguous peripancreatic soft tissue [47]. This constellation of histopathologic findings defines LPSP (Fig. 1) [12,65]. A less common histopathologic variant termed idiopathic duct-centric chronic pancreatitis is characterized by a neutrophilic infiltrate with occasional microabscesses and rare obliterative phlebitis [9,27,47,66]. Nearly a third of patients who have AIP develop pancreatic calcification and atrophy and the appearance can resemble usual chronic calcific pancreatitis. It is well accepted that AIP has a diagnostic pancreatic histology that distinguishes it from usual chronic pancreatitis and pancreatic cancer. Additionally, the pancreas and other organs involved in AIP show abundant infiltration with
Fig. 1. Three examples of the predominant histologic features of LPSP in AIP. (A) Trucut biopsy of the pancreas demonstrating the dense infiltration of the periductal space with plasma cells and T lymphocytes. (B) Obliterative phlebitis involving a major vein and storiform or ‘‘whirling’’ fibrosis of the pancreatic parenchyma. (C) Abundant parenchyma infiltration with IgG4-positive plasma cells.
AUTOIMMUNE PANCREATITIS
445
IgG4-positive cells. Immunostaining pancreatic tissue for IgG4 positivity has been shown to be a helpful adjunct in diagnosing AIP [67]. Pancreatic resection specimens show diagnostic histology in almost all patients who have AIP in whom they are available. Histologic diagnosis of AIP requires preservation of tissue architecture, however, and hence fine-needle aspirates used to diagnose pancreatic cancer are not suitable for diagnosing AIP. The use of endoscopic ultrasound (EUS)–guided Trucut biopsy has emerged as an effective and safe tool for obtaining pancreatic biopsies in AIP [68,69]. When core biopsies are obtained, the diagnostic sensitivity of pancreatic histology for AIP depends on the size of the tissue sample. For example, in a report from Mayo Clinic, only 7 of 16 (44%) subjects who underwent pancreatic core biopsy showed the full spectrum of diagnostic changes of LPSP, and 15 of 16 (94%) patients had diagnostic IgG4 immunostaining [67]. The pancreas is often not uniformly involved by classic AIP features, however, and thus sampling error can occur [70], especially if visibly uninvolved areas are biopsied. In addition, staining of involved extrapancreatic tissues, such as the biliary tree, retroperitoneum, and colon, has revealed IgG4 positivity in affected patients [71]. At this time, it is unclear whether pancreatic histology predicts disease severity or progression to endocrine or exocrine insufficiency.
IMAGING FEATURES Characteristic cross-sectional and ultrasound imaging features have been well described in AIP. Classically, the pancreatic parenchyma is diffusely enlarged, forming a sausage-shaped gland with featureless borders (Fig. 2) [1,58,72,73]. Other classic features that may be present include delayed and prolonged
Fig. 2. Pancreas-phase CT scan from a patient who had untreated AIP demonstrating a diffusely enlarged (sausage-shaped) heterogenous pancreatic parenchyma. Note the relatively smooth contour of the parenchymal border and the lack of an identifiable pancreatic duct.
446
GARDNER & CHARI
contrast enhancement, a rimlike capsule surrounding the gland on delayed enhancement sequences (the hypoattenuation halo), a nondilated, ectatic pancreatic duct, and the absence of peripancreatic fat hypoenhancement [47,74]. In a recent study, dual-phase CT scans of 74 patients (25 AIP, 33 pancreatic adenocarcinoma, and 16 normal pancreas) were independently evaluated by three radiologists blinded to clinical diagnosis [75]. Readers correctly identified 17 to 19 of 25 AIPs (sensitivity 68%–76%) and overall accuracy was 81% to 85%. The authors concluded that dual-phase CT of the pancreas was moderately accurate in the diagnosis of AIP and in differentiating it from pancreatic carcinoma. Findings that were relatively specific for AIP included a diffusely enlarged pancreas, diffusely decreased pancreatic enhancement, and presence of a capsule-like rim, bile duct wall enhancement, and solid renal lesions. MRI characteristically reveals enlargement of the pancreas with decreased signal intensity on T1-weighted MR images, increased signal intensity on T2-weighted MR images, and, occasionally, a hypointense capsule-like rim [76]. Magnetic resonance cholangiopancreatography (MRCP) is often helpful in characterizing the pancreatic and bile ducts, although the narrowed segment of the pancreatic duct is not well visualized. Narrowing of the anterior superior pancreaticoduodenal artery, posterior superior pancreaticoduodenal artery, and transpancreatic artery on angiographic evaluation has also been described in AIP patients [77]. Increasingly, endoscopic ultrasound has been used to evaluate patients for AIP [69,78,79]. Not only can the parenchyma and biliary and pancreatic ducts be visualized, EUS also provides an opportunity to obtain Trucut biopsy samples. Intraductal ultrasound can also be used to evaluate indeterminate biliary strictures. Levy and colleagues [68] have reported on the diagnostic usefulness of Trucut biopsy in three patients who had suspected pancreatic adenocarcinoma with planned surgical resection following indeterminate fine-needle aspiration. In two of the patients, AIP was diagnosed using Trucut biopsy; in the other chronic pancreatitis was diagnosed. In all patients, unnecessary surgery was avoided. Cholangiography has been shown to be an accurate method to differentiate AIP from primary sclerosing cholongitis (PSC). Nakazawa and colleagues [80] reported that bandlike stricture, beaded or pruned-tree appearance, and diverticulum-like formation were significantly more frequent in patients who had PSC. In contrast, segmental stricture, long stricture with prestenotic dilatation, and stricture of the distal common bile duct were significantly more common in sclerosing cholangitis with AIP. Increased uptake with whole-body (18)F-fluorodeoxyglucose positron emission tomography is seen in the pancreas and extrapancreatic lesions of patients who have AIP [81–83]. A characteristic feature of the acute presentation of AIP is that the pancreatic changes improve, if not completely resolve, with corticosteroid treatment [1,84,85]. In our experience, failure of the pancreatic imaging to significantly improve after a 2- to 4-week course of corticosteroids should cast doubt on the diagnosis of AIP (Fig. 3).
AUTOIMMUNE PANCREATITIS
447
Fig. 3. Representative pancreatograms in the same patient obtained at ERCP demonstrating the pancreas pre (A) and post (B) treatment with 6 weeks of corticosteroids. Note the narrowed, ectatic pancreatic duct seen before treatment has completely normalized in appearance.
SEROLOGY Increased numbers of circulating immunoglobulins, specifically immunoglobulin subclass 4, are a hallmark of the disease [10,21,86]. In a landmark study, Hamano and colleagues [10] reported that serum IgG4 levels were highly (95%) sensitive and highly (97%) specific for AIP. In a recent study of 510 patients from the United States [21] including 45 who had AIP, 135 who had pancreatic cancer, 62 who had no pancreatic disease, and 268 who had other pancreatic diseases, the sensitivity, specificity, and positive predictive values for elevated serum IgG4 (>140 mg/dL) for diagnosis of AIP were 76%, 93%, and 36%, respectively. When using a cutoff of twice the upper limit of normal for serum IgG4 (>280 mg/dL), the corresponding values were 53%, 99%, and 75%, respectively [21]. In this study, 5% to 10% of non-AIP patient groups, including 10% of patients who had ductal adenocarcinoma, had elevated IgG4 levels [21]. In addition, serum IgG4 levels, even in the presence of classic histologic findings of AIP, can be normal (Table 2) [58,87]. Elevated titers of many autoantibodies have been described in AIP. Autoantibodies against carbonic anhydrase II and IV and lactoferrin are detected in most patients who have AIP [88–91]. Involvement of antinuclear and anti– smooth muscle antibodies has also been described [89,92]. Autoantibodies to the pancreatic secretory trypsin inhibitor have been shown to be elevated in nearly 50% of AIP patients compared with controls [93]. None of these autoantibodies have prediction characteristics that equal that of IgG4. Levels of total IgG and gamma globulins are also increased in AIP. In our experience, however, it is unusual to have elevated serum levels of IgG or gamma globulins without elevation of serum IgG4 levels. Although a combination of serum IgG4 levels and autoantibody titers of antinuclear antibodies and rheumatoid
GARDNER & CHARI
448
Table 2 IgG4 level in patients who have different diseases of the pancreas
Numbera Mean IgG4 SM Range Proportion elevated >140 mg/dL
AIP
Normal Pancreatic Benign pancreatic Chronic pancreas cancer tumor pancreatitis
45 550 99 16–2890 76%
62 49 6 3–263 4.8%
135 68 9 3–1140 9.6%
64 47 5 3–195 4.7%
79 46 5 3–231 6.3%
a
Based on 510 patients referred to the Mayo Clinic for evaluation of pancreatic disease from January 2005 through June 2006. Data from Ghazale A, Chari ST, Smyrk TC, et al. Value of serum IgG4 in the diagnosis of autoimmune pancreatitis and in distinguishing it from pancreatic cancer. Am J Gastroenterol 2007;102(8):1646–53.
factor modestly increases sensitivity, it also significantly reduces specificity. The authors do not routinely use autoantibody titers to diagnose AIP. OTHER ORGAN INVOLVEMENT Other organs are often involved in AIP; their involvement may be diagnosed before, simultaneous with, or after the diagnosis of AIP. Biliary tract is involved in 60% to 100% of all patients presenting with AIP [1,51,55,59,63] and has recently been termed IAC [42]. IAC affects both intra- and extrahepatic bile ducts, with the distal common bile duct being the most common site of involvement [94]. Biliary imaging may not necessarily reveal involvement, even when present microscopically [55,76]. Histologically, a lymphoplasmacytic infiltrate surrounds the bile ducts in a pattern similar to that seen in the pancreas and IgG4-positive staining is often present [42,47,95]. AIP coexisting with PSC has been described [96], although this is likely not primary sclerosing cholangitis but IgG4-associated cholangitis. It has also been shown that in a small proportion of patients who have aggressive PSC, serum IgG4 levels are elevated suggesting a possible role for corticosteroid therapy [47,95]. One should be cautious in diagnosing IAC simply based on elevated serum IgG4 levels, however, because false-positive elevations may occur in true PSC. IAC differs from PSC in that there is generally less intrahepatic involvement, the strictures can be transient under observation, the strictures are usually segmental, and patients are typically pANCA negative [47,97]. Inflammatory bowel disease, which is present in 70% of PSC, is less common (6%) in IAC [98]. Analogous to the response seen in the inflammatory component of pancreatic involvement, inflammation of the biliary tree typically responds to corticosteroid treatment, although the specific response of each duct segment is still being evaluated [99]. In addition to the biliary system, multiple other organs may be involved in ISD [85]. Hamano and colleagues [51] reviewed the frequency, distribution, clinical characteristics, and pathology of five extrapancreatic lesions in 64
AUTOIMMUNE PANCREATITIS
449
patients who had AIP and found the most frequent extrapancreatic lesion was hilar lymphadenopathy (80.4%), followed by extrapancreatic bile duct lesions (73.9%), lacrimal and salivary gland lesions (39.1%), hypothyroidism (22.2%), and retroperitoneal fibrosis (12.5%). No patients had all five types of lesions. Patients who had hilar lymphadenopathy or lacrimal and salivary gland lesions were found to have significantly higher IgG4 levels than those who did not. Both intrinsic (tubulointestinal fibrosis) and extrinsic (hydronephrosis secondary to retroperitoneal fibrosis) renal disease have been associated with AIP, as has inflammatory pneumonitis and inflammatory pseudotumor of the liver [44,100–103]. Fig. 4 demonstrates some of the more common extrapancreatic manifestations associated with AIP. Based on the significant degree of extrapancreatic disease and the association with IgG4 staining of tissues, the concept of an IgG4-related autoimmune disease entity has been proposed by Kamisawa [39]. This distinct clinicopathologic disease process would encompass all IgG4 diseases, of which AIP would be only one manifestation. DIAGNOSTIC CRITERIA In 1995 Yoshida and colleagues [19] reported a list of 12 features suggestive of AIP, but stopped short of providing clinical criteria for its diagnosis. The first diagnostic criteria were proposed by the Japan Pancreas Society in 2002 and later modified in 2006. These guidelines were developed to distinguish between AIP and pancreatic adenocarcinoma. To make the diagnosis of AIP based on the Japanese guidelines, it is mandatory that findings on radiography be consistent with AIP. These findings include the presence of diffuse or segmental narrowing of the main pancreatic duct with irregular wall diagnosed by endoscopic retrograde pancreatogram and diffuse or localized enlargement of the pancreas on abdominal ultrasonography, CT, or MRI. In addition, one of serologic (high serum c-globulin, IgG, or IgG4, or the presence of autoantibodies,
Fig. 4. Extrapancreatic manifestations of AIP demonstrated in a patient who had untreated AIP. Retroperitoneal fibrosis (A, white arrow) and intrahepatic biliary dilatation secondary to diffuse inflammatory stricturing of the biliary system (B, white arrow) are shown.
450
GARDNER & CHARI
such as antinuclear antibodies and rheumatoid factor) or histologic (marked interlobular fibrosis and prominent infiltration of lymphocytes and plasma cells in the periductal area, occasionally with lymphoid follicles in the pancreas) criteria are required to satisfy Japanese criteria for diagnosis of AIP. The Japanese criteria do not take into account that AIP has unique histologic features, characteristic findings on IgG4 immunostaining of the organs involved, other organ involvement, or response to steroids. In 2006, Chari and colleagues [1] published an alternate set of guidelines based on the Mayo Clinic experience with AIP. These criteria, known by the mnemonic HISORt, recognize characteristic features of AIP on pancreatic histology and imaging, serology, other organ involvement, and response to corticosteroid therapy. Based on the HISORt criteria patients can be diagnosed as AIP if they fall into one of three groups: (A) diagnostic pancreatic histology or presence of 10 or more IgG4-positive cells per high-power field (HPF) on immunostain of lymphoplasmacytic infiltrate with storiform fibrosis; (B) typical pancreatic imaging with elevated serum IgG4 140 mg/dL, or (C) unexplained pancreatic disease with negative workup for other pancreatic diseases, especially malignancy, with elevated serum IgG4 levels 140 mg/dL or other organ involvement confirmed by presence of abundant IgG4-positive cells, and resolution/marked improvement in pancreatic or extrapancreatic manifestations with steroid therapy (Box 1). By including additional features, these criteria identify a wider spectrum of clinical presentations of AIP. When the imaging features are typical and there is confirmatory serologic evidence of elevated levels of serum IgG4, the diagnosis of AIP is relatively easy. It still requires a radiologist familiar with the characteristic imaging features of AIP, but that is a matter of education. In some patients, however, the radiologic features are simply not diagnostic. In such patients a pancreatic biopsy can be a helpful adjunct to the diagnosis. The diagnostic gold standard for AIP is the presence of LPSP with IgG4-positive immunostaining of pancreatic tissues [6,50]. The presence of the full spectrum of LPSP on histology requires core biopsy, however, because fine-needle aspiration is usually not sufficient to make the diagnosis. Core biopsies can be obtained percutaneously with ultrasound or CT guidance or by transmurally using EUS-guidance biopsy [68]. TREATMENT The cornerstone of treatment of AIP is the use of corticosteroids with multiple authors reporting dramatic response rates with prolonged therapy [1,10, 104–106]. A word of caution, however, is that it is imperative to thoroughly rule out other possible causes of pancreatic disease, most notably pancreatic malignancy, before initiating corticosteroid therapy. Although spontaneous remissions do occur in AIP, the use of corticosteroids seems to hasten recovery and may prevent recurrences. There are several reasons, therefore, to initiate treatment of AIP with corticosteroids. For one, if the diagnosis remains in doubt and malignancy has been excluded, response to corticosteroids can be a reasonable method of diagnosing AIP. The clinical
AUTOIMMUNE PANCREATITIS
451
Box 1: Mayo Clinic HISORt criteria for the diagnosis of autoimmune pancreatitis Diagnostic criteria Histology At least one of the following: Periductal lymphoplasmacytic infiltrate with obliterative phlebitis and storiform fibrosis Lymphoplasmacytic infiltrate with storiform fibrosis with abundant (10 IgG4 cells/HPF)
Imaging Typical: diffusely enlarged gland with delayed rim enhancement, diffusely irregular,
attenuated main pancreatic duct Other: focal pancreatic mass/enlargement, focal pancreatic ductal stricture, pancre-
atic atrophy, calcification, pancreatitis Serology Elevated serum IgG4 level (normal 8–140 mg/dL)
Other organ involvement Hilar/intrahepatic biliary strictures, persistent distal biliary stricture, parotid/lacrimal
gland involvement, mediastinal lymphadenopathy, retroperitoneal fibrosis Response to steroid therapy Resolution/marked improvement of pancreatic/extrapancreatic manifestation with
corticosteroid therapy Diagnostic groupsa Group A: diagnostic pancreatic histology Presence of one or more of the following criteria: Specimen demonstrating the full spectrum of LPSP 10 IgG4 cells/HPF on immunostain of pancreatic lymphoplasmacytic infiltrate
Group B: typical imaging + serology Presence of all of the following criteria: CT or MRI scan showing diffusely enlarged pancreas with delayed and rim
enhancement Pancreatogram showing diffusely irregular pancreatic duct Elevated serum IgG4 levels
Group C: response to corticosteroids Presence of all of the following criteria: Unexplained pancreatic disease after negative workup for other causes Elevated serum IgG4 or other organ involvement confirmed by presence of abundant
IgG4-positive cells Resolution/marked improvement in pancreatic or extrapancreatic manifestations with
corticosteroid therapy a
Patients meeting criteria for one or more of the groups have AIP. Data from Chari ST, Smyrk TC, Levy MJ, et al. Diagnosis of autoimmune pancreatitis: the Mayo Clinic experience. Clin Gastroenterol Hepatol 2006;4(8):1010–6.
452
GARDNER & CHARI
suspicion for AIP should be high before initiation of therapy, however, and patients should be followed closely for any symptoms (eg, profound weight loss, anorexia, night sweats) more consistent with malignancy than AIP. Treatment can also be a means of reducing clinical symptoms from acute pancreatic (pancreatic endocrine insufficiency, rarely acute pancreatitis) or extrapancreatic (jaundice from biliary strictures, sialadenitis) manifestations of disease [62,104,107–110]. In addition, there can sometimes be structural improvement, for example in the pancreatic duct, with corticosteroid therapy [56]. The degree of structural response depends on the extent of fibrosis versus inflammation; patients who have more inflammatory injury typically have a greater structural response and extensive fibrosis may not allow complete remission to occur [104]. The exact corticosteroid treatment protocol for patients who have AIP is not standardized; however, most practitioners initiate therapy with between 30 and 40 mg of prednisone daily. These doses are usually effective to induce remission; it is unclear if starting at lower doses would be equally effective. Resolution of symptoms is generally rapidly achieved within 2 to 3 weeks of corticosteroid initiation. It usually takes several weeks to months for evidence of serologic (normalization of IgG4) or radiologic remission. Occasionally, because of fibrotic involvement of tissue, radiologic remission is not seen, especially in proximal biliary strictures resembling cholangiocarcinoma, intrahepatic strictures resembling PSC, and retroperitoneal fibrosis. Progression toward normalization of serum IgG4 levels can be used to guide treatment in these instances. Because histologic specimens are often difficult to obtain, we generally do not use histology as a marker for remission. At the Mayo Clinic, we treat patients diagnosed with AIP with a prolonged steroid taper. Patients are started on 40 mg/d of prednisone for 4 weeks. After 4 weeks, their clinical response is gauged and repeat cross-sectional imaging and serologic evaluation are performed to check for response. If clinical, serologic, or radiographic response is documented, the prednisone dose is tapered 5 mg/wk until gone. In patients in whom a biliary stent has been placed, this usually can be removed at 6 to 8 weeks following initiation of therapy. IgG4 levels are also followed and in patients who have AIP, a decrease in IgG4 (although not necessarily normalization) should occur within 4 weeks of treatment initiation (Fig. 5). Recently, in patients who present with jaundice because of biliary stricturing disease who do not wish to undergo initial endoscopic retrograde cholongiopancreatography with stent placement, we have occasionally treated with a large initial bolus of intravenous corticosteroids. Although our experience is limited and at this time anecdotal, jaundice does seem to respond rapidly to this treatment, thus obviating the need for biliary stent placement. Further prospective studies are needed to determine if large initial doses of corticosteroids are an effective alternative to initial biliary stent placement in newly diagnosed AIP. Between 30% and 40% of patients have clinical or radiographic relapse following treatment with prolonged corticosteroids requiring retreatment with a second prolonged course [1,43,84,107,111]. These relapses generally occur
AUTOIMMUNE PANCREATITIS
453
Diagnosis of Autoimmune Pancreatitis
Start Prednisone 40 mg daily for 4 weeks
After four weeks, measure IgG4 and repeat imaging
Clinical response, decrease in IgG4 and/or improvement of inflammation on imaging
Taper Prednisone 5 mg/week until taper complete
No change in symptoms, IgG4 level or improvement of inflammation on imaging
Consider alternative diagnosis
Recheck IgG4 and imaging 4-6 weeks following completion of Prednisone taper
No symptoms, IgG4 and radiographic improvement
Serial clinical follow-up (6 months initially) and imaging
More than 1 relapse, initiate chronic suppressive therapy*
Clinical, biochemical or imaging relapse
Fig. 5. Algorithm for treatment of AIP. Recent data suggest that relapse following steroid withdrawal is likely in 70% of patients who have AIP who have proximal extrahepatic or intrahepatic biliary strictures. Chronic immunosuppression is therefore recommended after withdrawal of first course of steroid. (Data from Ghazale A, Chari ST, Zhang L. Immunoglobulin G4-associated cholangitis: clinical profile and response to therapy. Gastroenterology 2008;134:706–15.)
in the short term; data on long-term relapse are lacking. Relapse may be symptomatic, radiologic, serologic, or histologic. The presence of symptoms (recurrent abdominal pain, weight loss, and so forth) is often a clue to relapse within the pancreas; only in the presence of symptoms is cross-sectional imaging repeated. Serologic relapse alone can be seen in patients who do not have clinical symptoms or radiologic changes; whether or not this represents a subclinical disease relapse is unclear at this time. It is also unclear if certain types of organ involvement are more prone to relapse than others. For example, Ghazale and colleagues [43] found that in patients who had IAC treated with 11 weeks of
454
GARDNER & CHARI
corticosteroids, proximal biliary involvement (proximal extrahepatic and intrahepatic biliary strictures) relapsed with a rate of 65% compared with a 25% relapse rate in those who had intrapancreatic stricturing of the distal common bile duct. In a certain subset of patients who have relapse after a second prolonged course of steroids, either chronic prednisone therapy or use of another agent, such as azathioprine or 6-mercaptopurine, may be necessary. In Japan, it is often standard of care to continue patients on a chronic low dose (2.5–10 mg/d) of prednisone indefinitely [104,107]. Although there are only case reports evaluating the efficacy of long-term treatment of AIP with immunomodulating therapy, we have had favorable, although limited, experience with these medications [112]. As more experience is gained about the long-term pathogenesis of AIP, recommendations on the use of chronic suppressive therapy needs to be developed. In addition, because there is such a high frequency of shortterm relapse, whether maintenance therapy should be used in all patients, and what type of maintenance therapy should be used, remains to be established. It is unclear at this time whether corticosteroid therapy alters the long-term natural history of disease, prevents the development of future pancreatic or extrapancreatic involvement and organ dysfunction, or is adequate as a long-term suppressive strategy. MISDIAGNOSIS Increasingly, we are evaluating patients in whom AIP has been inaccurately diagnosed. The misdiagnosis occurs in the setting of patients who have other unrelated conditions, such as pancreatic adenocarcinoma, being treated inappropriately with corticosteroids. We have also seen several patients who had functional abdominal pain complaints treated with prolonged courses of high-dose corticosteroids without evidence of AIP. Conversely, multiple patients have undergone therapies, such as pancreatic head resections or partial hepatectomy, without consideration of an autoimmune cause. It is therefore important that AIP be considered in the differential diagnosis of patients who have chronic pancreatitis or biliary strictures. It is imperative, however, that a thorough evaluation be performed for other causes of disease, with histopathologic analysis if possible, before the initiation of corticosteroid therapy. Once corticosteroid therapy has been initiated, patients should be followed closely for signs of worsening or refractory disease symptoms; it is highly unusual for AIP not to quickly respond to appropriate corticosteroid therapy. In addition, clinicians must be cognizant that AIP is a rare disease, and in patients who do not meet the Japanese Pancreas Society or Mayo HISORt guidelines, corticosteroid therapy is likely not advisable. PROGNOSIS There are limited data about the long-term outcome of patients who have AIP. Hirano and colleagues [105] published the most comprehensive report on
AUTOIMMUNE PANCREATITIS
455
prognosis when they evaluated 42 patients who had AIP, 19 of whom received corticosteroid treatment at the time of diagnosis. In the 23 patients who did not have corticosteroid treatment initially, 16 developed unfavorable events, including obstructive jaundice attributable to distal bile duct stenosis in 4, growing pseudocyst in 1, and sclerogenic changes of extrapancreatic bile duct in 9, over an average observation period of 25 months. After an average observation period of 23 months in the initial treatment group, 6 patients developed unfavorable events consisting of interstitial pneumonia in 3 and recurrence of obstructive jaundice in 3. Their conclusions were that early introduction of corticosteroids is important to prevent subsequent disease complications. Most patients treated with corticosteroids develop ‘‘burn out’’ of disease, rendering the pancreas usually somewhat atrophic following treatment [10,47]. The degree of residual pancreatic exocrine or endocrine insufficiency is likely related to the degree of gland fibrosis at the time of treatment [104]. At this time, given the relatively recent description and active investigation of AIP, there are no data regarding the long-term mortality rate of patients who have this disease. In addition, it has not been investigated whether life expectancy is altered by the course of this disease. FUTURE DIRECTIONS There are several lines of investigation that need to be addressed in regard to AIP. Continued work to determine the cause of this disease and its relationship with IgG4 is imperative. Specifically, the antigenic trigger of CD4 and CD8 T-cell activation needs to be identified. Clinically, investigation should focus on the natural history of AIP with specific attention to the wide-ranging effects of IgG4-related systemic disease. It is not currently known whether different manifestations of IgG4-related disease have unique or alternate prognoses. The role of corticosteroids, specifically their role in changing the natural history of disease, needs to be investigated. In asymptomatic patients, it will be important to determine if treatment helps to prevent future organ dysfunction. Furthermore, the role of chronic suppressive therapy, either with corticosteroids or another immune-modulating drug, in preventing relapse and affecting longterm prognosis is yet to be determined. SUMMARY AIP is a unique subtype of recently identified chronic pancreatitis that is immune mediated and represents one manifestation of a systemic IgG4-related disease process. Although a rare condition, it is important to recognize because it responds often dramatically to immune system–modulating treatment. Diagnosing AIP can sometimes be challenging, however, and it is imperative that clinicians be cautious when considering this diagnosis in patients suspected of having a pancreatic malignancy. As clinical experience with AIP increases, refinement of diagnostic criteria and development of standardized therapeutic protocols should allow further optimization of care for our patients.
456
GARDNER & CHARI
References [1] Chari ST, Smyrk TC, Levy MJ, et al. Diagnosis of autoimmune pancreatitis: the Mayo Clinic experience. Clin Gastroenterol Hepatol 2006;4(8):1010–6. [2] Kamisawa T. IgG4-positive plasma cells specifically infiltrate various organs in autoimmune pancreatitis. Pancreas 2004;29(2):167–8. [3] Kamisawa T, Egawa N, Nakajima H. Autoimmune pancreatitis is a systemic autoimmune disease. Am J Gastroenterol 2003;98(12):2811–2. [4] Yoshida K, Toki F, Takeuchi T, et al. Chronic pancreatitis caused by an autoimmune abnormality. Proposal of the concept of autoimmune pancreatitis. Dig Dis Sci 1995;40(7): 1561–8. [5] Sarles H, Sarles JC, Muratore R, et al. Chronic inflammatory sclerosis of the pancreas—an autonomous pancreatic disease? Am J Dig Dis 1961;6:688–98. [6] Pearson RK, Longnecker DS, Chari ST, et al. Controversies in clinical pancreatology: autoimmune pancreatitis: does it exist? Pancreas 2003;27(1):1–13. [7] Sood S, Fossard DP, Shorrock K. Chronic sclerosing pancreatitis in Sjogren’s syndrome: a case report. Pancreas 1995;10(4):419–21. [8] Weber SM, Cubukcu-Dimopulo O, Palesty JA, et al. Lymphoplasmacytic sclerosing pancreatitis: inflammatory mimic of pancreatic carcinoma. J Gastrointest Surg 2003;7(1): 129–37. [9] Ectors N, Maillet B, Aerts R, et al. Non-alcoholic duct destructive chronic pancreatitis. Gut 1997;41(2):263–8. [10] Hamano H, Kawa S, Horiuchi A, et al. High serum IgG4 concentrations in patients with sclerosing pancreatitis. N Engl J Med 2001;344(10):732–8. [11] van Buuren HR, Meijssen MA, van der Werf SD. A patient with sclerosing autoimmune pancreaticocholangitis as the cause of recurrent cholangitis following a pylorus-sparing pancreaticoduodenectomy. Ned Tijdschr Geneeskd 2005;149(51):2888–9. [12] Kawaguchi K, Koike M, Tsuruta K, et al. Lymphoplasmacytic sclerosing pancreatitis with cholangitis: a variant of primary sclerosing cholangitis extensively involving pancreas. Hum Pathol 1991;22(4):387–95. [13] Ichimura T, Kondo S, Ambo Y, et al. Primary sclerosing cholangitis associated with autoimmune pancreatitis. Hepatogastroenterology 2002;49(47):1221–4. [14] Kamisawa T, Funata N, Hayashi Y, et al. Close relationship between autoimmune pancreatitis and multifocal fibrosclerosis. Gut 2003;52(5):683–7. [15] Chutaputti A, Burrell MI, Boyer JL. Pseudotumor of the pancreas associated with retroperitoneal fibrosis: a dramatic response to corticosteroid therapy. Am J Gastroenterol 1995;90(7):1155–8. [16] Fukumori K, Shakado S, Miyahara T, et al. Atypical manifestations of pancreatitis with autoimmune phenomenon in an adolescent female. Intern Med 2005;44(8):886–91. [17] Ito T, Nakano I, Koyanagi S, et al. Autoimmune pancreatitis as a new clinical entity. Three cases of autoimmune pancreatitis with effective steroid therapy. Dig Dis Sci 1997;42(7): 1458–68. [18] Kim KP, Kim MH, Kim JC, et al. Diagnostic criteria for autoimmune chronic pancreatitis revisited. World J Gastroenterol 2006;12(16):2487–96. [19] Choi EK, Kim MH, Kim JC, et al. The Japanese diagnostic criteria for autoimmune chronic pancreatitis: is it completely satisfactory? Pancreas 2006;33(1):13–9. [20] Kamisawa T. IgG4-related sclerosing disease. Intern Med 2006;45(3):125–6. [21] Ghazale A, Chari ST, Smyrk TC, et al. Value of serum IgG4 in the diagnosis of autoimmune pancreatitis and in distinguishing it from pancreatic cancer. Am J Gastroenterol 2007;102(8):1646–53. [22] Nishimori I, Tamakoshi A, Otsuki M. Prevalence of autoimmune pancreatitis in Japan from a nationwide survey in 2002. J Gastroenterol 2007;42(Suppl 18):6–8. [23] Kim KP, Kim MH, Lee SS, et al. Autoimmune pancreatitis: it may be a worldwide entity. Gastroenterology 2004;126(4):1214.
AUTOIMMUNE PANCREATITIS
457
[24] Okazaki K. Autoimmune pancreatitis is increasing in Japan. Gastroenterology 2003; 125(5):1557–8. [25] Abraham SC, Wilentz RE, Yeo CJ, et al. Pancreaticoduodenectomy (Whipple resections) in patients without malignancy: are they all ‘‘chronic pancreatitis’’? Am J Surg Pathol 2003;27(1):110–20. [26] Farnell MB, Pearson RK, Sarr MG, et al. A prospective randomized trial comparing standard pancreatoduodenectomy with pancreatoduodenectomy with extended lymphadenectomy in resectable pancreatic head adenocarcinoma. Surgery 2005;138(4): 618–28. [27] Yadav D, Notahara K, Smyrk TC, et al. Idiopathic tumefactive chronic pancreatitis: clinical profile, histology, and natural history after resection. Clin Gastroenterol Hepatol 2003;1(2):129–35. [28] Kamisawa T, Wakabayashi T, Sawabu N. Autoimmune pancreatitis in young patients. J Clin Gastroenterol 2006;40(9):847–50. [29] Ota M, Katsuyama Y, Hamano H, et al. Two critical genes (HLA-DRB1 and ABCF1) in the HLA region are associated with the susceptibility to autoimmune pancreatitis. Immunogenetics 2007;59(1):45–52. [30] Umemura T, Ota M, Hamano H, et al. Genetic association of Fc receptor-like 3 polymorphisms with autoimmune pancreatitis in Japanese patients. Gut 2006;55(9):1367–8. [31] Chang MC, Chang YT, Tien YW, et al. T-cell regulatory gene CTLA-4 polymorphism/ haplotype association with autoimmune pancreatitis. Clin Chem 2007;53(9):1700–5. [32] Kountouras J, Zavos C, Gavalas E, et al. Challenge in the pathogenesis of autoimmune pancreatitis: potential role of helicobacter pylori infection via molecular mimicry. Gastroenterology 2007;133(1):368–9. [33] Kamisawa T, Chen PY, Tu Y, et al. Autoimmune pancreatitis metachronously associated with retroperitoneal fibrosis with IgG4-positive plasma cell infiltration. World J Gastroenterol 2006;12(18):2955–7. [34] Kawa S, Hamano H. Autoimmune pancreatitis and bile duct lesions. J Gastroenterol 2003;38(12):1201–3. [35] Matsuda M, Hamano H, Yoshida T, et al. Seronegative Sjogren syndrome with asymptomatic autoimmune sclerosing pancreatitis. Clin Rheumatol 2007;26(1):117–9. [36] Okazaki K. Is IgG4-associated multifocal systemic fibrosis the same disease entity as autoimmune pancreatitis? Intern Med 2007;46(3):117–8. [37] Miyajima N, Koike H, Kawaguchi M, et al. Idiopathic retroperitoneal fibrosis associated with IgG4-positive-plasmacyte infiltrations and idiopathic chronic pancreatitis. Int J Urol 2006;13(11):1442–4. [38] Frulloni L, Morana G, Bovo P, et al. Salivary gland involvement in patients with chronic pancreatitis. Pancreas 1999;19(1):33–8. [39] Kamisawa T, Funata N, Hayashi Y, et al. A new clinicopathological entity of IgG4-related autoimmune disease. J Gastroenterol 2003;38(10):982–4. [40] Takagi T, Doi T, Sakamoto H, et al. A case report of autoimmune pancreatitis with Mikulicz’ s disease and diabetes mellitus. Nippon Shokakibyo Gakkai Zasshi 2006;103(2): 180–8. [41] Kitagawa S, Zen Y, Harada K, et al. Abundant IgG4-positive plasma cell infiltration characterizes chronic sclerosing sialadenitis (Kuttner’s tumor). Am J Surg Pathol 2005;29(6): 783–91. [42] Bjornsson E, Chari ST, Smyrk TC, et al. Immunoglobulin G4 associated cholangitis: description of an emerging clinical entity based on review of the literature. Hepatology 2007;45(6):1547–54. [43] Ghazale A, Chari ST, Zhang L, et al. IgG4-associated cholangitis: clinical profile and response to therapy. Gastroenterol 2008;134:706–15. [44] Saeki T, Nishi S, Ito T, et al. Renal lesions in IgG4-related systemic disease. Intern Med 2007;46(17):1365–71.
458
GARDNER & CHARI
[45] Krasinskas AM, Raina A, Khalid A, et al. Autoimmune pancreatitis. Gastroenterol Clin North Am 2007;36(2):239–57. [46] Okazaki K. Autoimmune-related pancreatitis. Curr Treat Options Gastroenterol 2001;4(5):369–75. [47] Lara LP, Chari ST. Autoimmune pancreatitis. Curr Gastroenterol Rep 2005;7(2): 101–6. [48] Varadarajulu S, Cotton PB. Autoimmune pancreatitis: is it relevant in the West? Gastroenterology 2003;125(5):1557. [49] Kitano Y, Matsumoto K, Chisaka K, et al. An autopsy case of autoimmune pancreatitis. JOP 2007;8(5):621–7. [50] Finkelberg DL, Sahani D, Deshpande V, et al. Autoimmune pancreatitis. N Engl J Med 2006;355(25):2670–6. [51] Hamano H, Arakura N, Muraki T, et al. Prevalence and distribution of extrapancreatic lesions complicating autoimmune pancreatitis. J Gastroenterol 2006;41(12):1197–205. [52] Hamano H, Kawa S. Are there any other organs in which autoimmune pancreatitis-associated lesions remain to be identified? Intern Med 2006;45(15):883–4. [53] Kamisawa T, Egawa N, Nakajima H, et al. Extrapancreatic lesions in autoimmune pancreatitis. J Clin Gastroenterol 2005;39(10):904–7. [54] Hirano K, Komatsu Y, Yamamoto N, et al. Pancreatic mass lesions associated with raised concentration of IgG4. Am J Gastroenterol 2004;99(10):2038–40. [55] Hirano K, Shiratori Y, Komatsu Y, et al. Involvement of the biliary system in autoimmune pancreatitis: a follow-up study. Clin Gastroenterol Hepatol 2003;1(6):453–64. [56] Horiuchi A, Kawa S, Hamano H, et al. ERCP features in 27 patients with autoimmune pancreatitis. Gastrointest Endosc 2002;55(4):494–9. [57] Notohara K, Burgart LJ, Yadav D, et al. Idiopathic chronic pancreatitis with periductal lymphoplasmacytic infiltration: clinicopathologic features of 35 cases. Am J Surg Pathol 2003;27(8):1119–27. [58] Chari ST. Diagnosis of autoimmune pancreatitis using its five cardinal features: introducing the Mayo Clinic’s HISORt criteria. J Gastroenterol 2007;42(Suppl 18):39–41. [59] Church NI, Pereira SP, Deheragoda MG, et al. Autoimmune pancreatitis: clinical and radiological features and objective response to steroid therapy in a UK series. Am J Gastroenterol 2007;102(11):2417–25. [60] Seicean A, Grigorescu M, Seicean R. Autoimmune chronic pancreatitis. Rom J Intern Med 2006;44(1):17–24. [61] Nishimori I, Tamakoshi A, Kawa S, et al. Influence of steroid therapy on the course of diabetes mellitus in patients with autoimmune pancreatitis: findings from a nationwide survey in Japan. Pancreas 2006;32(3):244–8. [62] Kamisawa T, Egawa N, Inokuma S, et al. Pancreatic endocrine and exocrine function and salivary gland function in autoimmune pancreatitis before and after steroid therapy. Pancreas 2003;27(3):235–8. [63] Kamisawa T, Tu Y, Egawa N, et al. Involvement of pancreatic and bile ducts in autoimmune pancreatitis. World J Gastroenterol 2006;12(4):612–4. [64] Kamisawa T, Egawa N, Shimizu M, et al. Autoimmune dorsal pancreatitis. Pancreas 2005;30(1):94–5. [65] Shimoda M, Kubota K, Sawada T, et al. Autoimmune pancreatitis diagnosed on the basis of immunohistology alone. A case report. JOP 2006;7(5):478–81. [66] Chari ST, Echelmeyer S. Can histopathology be the ‘‘Gold Standard’’ for diagnosing autoimmune pancreatitis? Gastroenterology 2005;129(6):2118–20. [67] Zhang L, Notohara K, Levy MJ, et al. IgG4-positive plasma cell infiltration in the diagnosis of autoimmune pancreatitis. Mod Pathol 2007;20(1):23–8. [68] Levy MJ, Reddy RP, Wiersema MJ, et al. EUS-guided Trucut biopsy in establishing autoimmune pancreatitis as the cause of obstructive jaundice. Gastrointest Endosc 2005;61(3): 467–72.
AUTOIMMUNE PANCREATITIS
459
[69] Levy MJ, Wiersema MJ, Chari ST. Chronic pancreatitis: focal pancreatitis or cancer? Is there a role for FNA/biopsy? Autoimmune pancreatitis. Endoscopy 2006;38(Suppl 1): S30–5. [70] Kloppel G, Sipos B, Zamboni G, et al. Autoimmune pancreatitis: histo- and immunopathological features. J Gastroenterol 2007;42(Suppl 18):28–31. [71] Aoki S, Nakazawa T, Ohara H, et al. Immunohistochemical study of autoimmune pancreatitis using anti-IgG4 antibody and patients’ sera. Histopathology 2005;47(2): 147–58. [72] Kamisawa T, Egawa N, Nakajima H, et al. Comparison of radiological and histological findings in autoimmune pancreatitis. Hepatogastroenterology 2006;53(72):953–6. [73] Sahani DV, Kalva SP, Farrell J, et al. Autoimmune pancreatitis: imaging features. Radiology 2004;233(2):345–52. [74] Okazaki K, Uchida K, Matsushita M, et al. How to diagnose autoimmune pancreatitis by the revised Japanese clinical criteria. J Gastroenterol 2007;42(Suppl 18):32–8. [75] Takahashi N, Fletcher JG, Fidler JL, et al. Dual-phase CT of autoimmune pancreatitis: multireader study. AJR Am J Roentgenol 2008;190:280–6. [76] Kamisawa T, Chen PY, Tu Y, et al. MRCP and MRI findings in 9 patients with autoimmune pancreatitis. World J Gastroenterol 2006;12(18):2919–22. [77] Kamisawa T. Angiographic findings in patients with autoimmune pancreatitis. Radiology 2005;236(1):371. [78] Farrell JJ, Garber J, Sahani D, et al. EUS findings in patients with autoimmune pancreatitis. Gastrointest Endosc 2004;60(6):927–36. [79] Salla C, Chatzipantelis P, Konstantinou P, et al. EUS-FNA contribution in the identification of autoimmune pancreatitis: a case report. JOP 2007;8(5):598–604. [80] Nakazawa T, Ohara H, Sano H, et al. Cholangiography can discriminate sclerosing cholangitis with autoimmune pancreatitis from primary sclerosing cholangitis. Gastrointest Endosc 2004;60(6):937–44. [81] Nakajo M, Jinnouchi S, Fukukura Y, et al. The efficacy of whole-body FDG-PET or PET/CT for autoimmune pancreatitis and associated extrapancreatic autoimmune lesions. Eur J Nucl Med Mol Imaging 2007;34:2088–95. [82] Nakajo M, Jinnouchi S, Noguchi M, et al. FDG PETand PET/CT monitoring of autoimmune pancreatitis associated with extrapancreatic autoimmune disease. Clin Nucl Med 2007;32(4):282–5. [83] Nakamoto Y, Saga T, Ishimori T, et al. FDG-PET of autoimmune-related pancreatitis: preliminary results. Eur J Nucl Med 2000;27(12):1835–8. [84] Wakabayashi T, Kawaura Y, Satomura Y, et al. Long-term prognosis of duct-narrowing chronic pancreatitis: strategy for steroid treatment. Pancreas 2005;30(1):31–9. [85] Sohn JH, Byun JH, Yoon SE, et al. Abdominal extrapancreatic lesions associated with autoimmune pancreatitis: radiological findings and changes after therapy. Eur J Radiol, in press. [86] Hirano K, Kawabe T, Yamamoto N, et al. Serum IgG4 concentrations in pancreatic and biliary diseases. Clin Chim Acta 2006;367(1–2):181–4. [87] Choi EK, Kim MH, Lee TY, et al. The sensitivity and specificity of serum immunoglobulin G and immunoglobulin G4 levels in the diagnosis of autoimmune chronic pancreatitis: Korean experience. Pancreas 2007;35(2):156–61. [88] Aparisi L, Farre A, Gomez-Cambronero L, et al. Antibodies to carbonic anhydrase and IgG4 levels in idiopathic chronic pancreatitis: relevance for diagnosis of autoimmune pancreatitis. Gut 2005;54(5):703–9. [89] Okazaki K, Uchida K, Ohana M, et al. Autoimmune-related pancreatitis is associated with autoantibodies and a Th1/Th2-type cellular immune response. Gastroenterology 2000; 118(3):573–81. [90] Nishi H, Tojo A, Onozato ML, et al. Anti-carbonic anhydrase II antibody in autoimmune pancreatitis and tubulointerstitial nephritis. Nephrol Dial Transplant 2007;22(4):1273–5.
460
GARDNER & CHARI
[91] Nishimori I, Miyaji E, Morimoto K, et al. Serum antibodies to carbonic anhydrase IV in patients with autoimmune pancreatitis. Gut 2005;54(2):274–81. [92] Kawa S, Hamano H. Clinical features of autoimmune pancreatitis. J Gastroenterol 2007;42(Suppl 18):9–14. [93] Asada M, Nishio A, Uchida K, et al. Identification of a novel autoantibody against pancreatic secretory trypsin inhibitor in patients with autoimmune pancreatitis. Pancreas 2006;33(1):20–6. [94] Nishino T, Toki F, Oyama H, et al. Biliary tract involvement in autoimmune pancreatitis. Pancreas 2005;30(1):76–82. [95] Mendes FD, Jorgensen R, Keach J, et al. Elevated serum IgG4 concentration in patients with primary sclerosing cholangitis. Am J Gastroenterol 2006;101(9):2070–5. [96] Takikawa H. Characteristics of primary sclerosing cholangitis in Japan. Hepatol Res 2007;37(Suppl 3):S470–3. [97] Ohara H, Nakazawa T, Ando T, et al. Systemic extrapancreatic lesions associated with autoimmune pancreatitis. J Gastroenterol 2007;42(Suppl 18):15–21. [98] Ghazale A, Chari ST, Takahashi N, et al. Biliary involvement in patients with autoimmune pancreatitis: clinical features and response to treatment. Gastroenterol 2007;132(4): S1216. [99] Horiuchi A, Kawa S, Hamano H, et al. Sclerosing pancreato-cholangitis responsive to corticosteroid therapy: report of 2 case reports and review. Gastrointest Endosc 2001;53(4): 518–22. [100] Hamano H, Kawa S, Ochi Y, et al. Hydronephrosis associated with retroperitoneal fibrosis and sclerosing pancreatitis. Lancet 2002;359(9315):1403–4. [101] Hirano K, Kawabe T, Komatsu Y, et al. High-rate pulmonary involvement in autoimmune pancreatitis. Intern Med J 2006;36(1):58–61. [102] Khalili K, Doyle DJ, Chawla TP, et al. Renal cortical lesions in patients with autoimmune pancreatitis: a clue to differentiation from pancreatic malignancy. Eur J Radiol, in press. [103] Sasahira N, Kawabe T, Nakamura A, et al. Inflammatory pseudotumor of the liver and peripheral eosinophilia in autoimmune pancreatitis. World J Gastroenterol 2005;11(6): 922–5. [104] Chari ST. Current concepts in the treatment of autoimmune pancreatitis. JOP 2007;8(1): 1–3. [105] Hirano K, Tada M, Isayama H, et al. Long-term prognosis of autoimmune pancreatitis without and with corticosteroid treatment. Gut;56:1719–24. [106] Ito T, Nishimori I, Inoue N, et al. Treatment for autoimmune pancreatitis: consensus on the treatment for patients with autoimmune pancreatitis in Japan. J Gastroenterol 2007;42(Suppl 18):50–8. [107] Kamisawa T, Yoshiike M, Egawa N, et al. Treating patients with autoimmune pancreatitis: results from a long-term follow-up study. Pancreatology 2005;5(2–3):234–8. [108] Matsushita M, Yamashina M, Ikeura T, et al. Effective steroid pulse therapy for the biliary stenosis caused by autoimmune pancreatitis. Am J Gastroenterol 2007;102(1): 220–1. [109] Ketikoglou IG, Elefsiniotis IS, Vezali EV, et al. Diabetes mellitus responsive to corticosteroids in autoimmune pancreatitis. J Clin Gastroenterol 2004;38(10):910. [110] Tanaka S, Kobayashi T, Nakanishi K, et al. Corticosteroid-responsive diabetes mellitus associated with autoimmune pancreatitis: pathological examinations of the endocrine and exocrine pancreas. Ann NY Acad Sci 2002;958:152–9. [111] Kamisawa T, Okamoto A. Prognosis of autoimmune pancreatitis. J Gastroenterol 2007;42(Suppl 18):59–62. [112] van Buuren HR, Vleggaar FP, Willemien Erkelens G, et al. Autoimmune pancreatocholangitis: a series of ten patients. Scand J Gastroenterol Suppl 2006;243:70–8.
Gastroenterol Clin N Am 37 (2008) 461–478
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Diagnosis and Treatment of Autoimmune Hepatitis Bruce A. Luxon, MD, PhD Division of Gastroenterology-Hepatology, University of Iowa, 200 Hawkins Drive, 4607 JCP, Iowa City, IA 52242, USA
A
utoimmune hepatitis (AIH) is an idiopathic hepatitis characterized by inflammation of the liver, presence of autoantibodies, and evidence of increased gamma globulins in the serum. It represents an enigmatic interaction between the immune system, autoantigens, and unknown triggering factors. AIH also has certain genetic predispositions, such as association with certain human leukocyte antigen (HLA) markers, such as DR3, D52, and DR4. The clinical manifestations of this disease were described in the 1950s and despite significant advances in our understanding of the immune system, the specific cause and pathogenesis of AIH remain unknown. Although medical science has made significant advances in the treatment of many chronic liver diseases, our treatment of AIH is essentially unchanged over the last 50 years. In the last 10 years, there have been more 1800 articles dealing with this disease state, with nearly half of them devoted to the treatment of AIH. This article provides a brief summary of the diagnosis of AIH, the natural history of AIH, an approach to the treatment and follow-up of AIH, and the role of liver transplantation in the treatment of AIH. EPIDEMIOLOGY The mean annual incidence of AIH among white, northern Europeans is 1.9 per 100,000 and its point prevalence is 16.9 per 100,000 [1]. Recent studies have documented that the incidence and prevalence of this disease have remained essentially unchanged over the last 2 decades. In Europe, AIH accounts for 2.6% of liver transplantations. In the United States, as published in 1998, AIH accounted for 5.9% of liver transplants [2]. Like many autoimmune diseases, AIH is a disease that affects women more frequently than men with a gender ratio of 3:1 [3,4]. AIH is typically a disease of younger patients but 23% of adults who have AIH develop the disease after the age of 60 [5]. Elderly patients are more likely to have cirrhosis at presentation (33% versus 10%). This finding suggests that older individuals may have a more aggressive disease that goes undetected until considerable liver damage is done. Elderly patients E-mail address:
[email protected]
0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.002
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
462
LUXON
also have a higher frequency of other autoimmune conditions at the time of their presentation with AIH [5]. NATURAL HISTORY The natural history of AIH is varied and depends on host issues that are still not clearly understood. Sentinel studies in the 1970s demonstrated that patients who have untreated severe disease are likely to die within 6 months of diagnosis [6]. These patients clearly need to be treated. Similar studies done in the latter part of the 1970s also demonstrated that certain patients who have AIH are likely to have the disease progress into cirrhosis with the typical development of esophageal varices, portal hypertension, and liver cancer [7]. These early studies suggested that patients who have a severe acute onset of AIH should be treated. Other studies, however, have suggested that there are patients who may not need immunosuppressant therapy. The risk-to-benefit ratio of immunosuppressive therapy with its potential toxicities is not always clear and treatment must be individualized. One subgroup of patients who might not need treatment is those patients who present with asymptomatic AIH. There are few evidence-based data available to guide treatment decisions in these patients whose AIH was diagnosed solely because they have elevated transaminases. Czaja [8] recommends that asymptomatic patients who have minimal enzyme elevations can be safely observed without the initiation of immunosuppressive therapy. Feld and colleagues [9] suggest that the rationale for using immunosuppressive therapy in a patient who has symptomatic disease would be to prevent the development of fibrosis. Feld found that the degree of liver enzyme elevation or the elevation of IgG in asymptomatic patients did not predict outcome, however. Others [10,11] suggest that the liver biopsy can be used to decide if treatment is needed in asymptomatic patients. Czaja [12] suggested that patients who had confluent necrosis on the initial liver biopsy should be treated. Others have suggested that many asymptomatic patients had only nonspecific changes on liver biopsy and that the liver biopsy was not helpful in deciding whether the patient should be treated. A significant proportion of patients who have AIH may have cirrhosis at the time of presentation. Approximately one third of patients have cirrhosis regardless of the age at which they present [9,14]. Most studies have found that patients who present with cirrhosis are more likely to die or develop complications of their liver disease during follow-up [15,16]. This conclusion, however, is not universal. Roberts and colleagues [17] found that the 10-year survival of patients who had cirrhosis was similar to that of patients who did not have cirrhosis at baseline. Patients who have cirrhosis are more likely to need to be treated for a longer time before achieving remission. These patients therefore may have a relatively poor outcome during follow-up [16,18]. Overall survival rates at 5 years (79% cirrhosis versus 97% no cirrhosis) and 10 years (67% cirrhosis versus 94% no cirrhosis) are good irrespective of the presence of cirrhosis [9,19]. Many centers treat patients who have cirrhosis with
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
463
immunosuppressive therapy in an attempt to improve their outcomes, although there are few data to suggest that long-term treatment is beneficial [20]. Some experts believe that patients who have ‘‘burned-out’’ AIH do not need to have immunosuppressant treatment. The rationale for this decision is that treatment of this inactive cirrhosis is fraught with potential side effects from the immunosuppression with little benefit. One additional caveat needs to be addressed regarding the treatment of AIH in patients who have documented cirrhosis. Several of the largest studies were done before viral infections could be excluded in patients who had AIH. These patients who had cirrhosis who were treated with steroids may have had a flare of their undiagnosed viral hepatitis (primarily hepatitis C) contributing to their poor outcome. In addition, many patients who had non alcoholic fatty liver disease may have positive autoantibodies and little or no fat on their biopsies despite the presence of NAFLD-induced cirrhosis. Treatment of these patients with steroids, although not deleterious, would not be expected to be beneficial. There is a growing body of literature that suggests that fibrosis and even cirrhosis attributable to AIH may be reversible with treatment of the AIH [17,21]. Conversely, there are multiple reports suggesting that fibrosis may progress despite immunotherapy [22]. It is not yet clear what differentiates these two groups of patients who respond differently to immunosuppressive therapy. DIAGNOSIS The diagnosis of AIH requires the presence of many typical features and, at the same time, the exclusion of other conditions that may cause chronic hepatitis. Like many autoimmune diseases, there currently are no pathognomonic features that clearly define AIH. Instead, an international panel has developed specific criteria to include or exclude the diagnosis of AIH. These recommendations are shown in Table 1 [13,23]. Other liver disease conditions that may be confused with the diagnosis of AIH are Wilson disease, chronic viral hepatitis (especially chronic hepatitis C), and drug-induced hepatitis. A separate scoring system has been established by the International Autoimmune Hepatitis Study Group. It is shown in Table 2 [13,23]. In using this scoring system, a patient is evaluated based on 11 biochemical, epidemiologic, and clinical markers before treatment and a pretreatment score is calculated. The pretreatment score can be modified to give a posttreatment score based on the response to treatment with corticosteroids. As designated by the committee, scores greater than 15 before corticosteroid treatment are consistent with a definite diagnosis of AIH. A posttreatment score greater than 17 constitutes a definite diagnosis. Similarly, patients who have relatively low scores (less than 10) are unlikely to have AIH. The scoring system was originally created to aid in the selection of homogeneous groups of patients who had AIH, and was primarily to be used for clinical research purposes. Currently the scoring system is most beneficial when trying to ascertain atypical or so-called ‘‘overlap’’ cases. The scoring system has been validated in multiple large studies with the scoring system showing
LUXON
464
Table 1 Features of autoimmune hepatitis needed or excluded for a definite diagnosis Diagnostic features Criteria
Clinical
Laboratory
Histologic
Inclusion
Female predominance Acute, fulminant, or indolent onset Concurrent immune diseases
Predominant aminotransferase abnormality c-globulin or immunoglobulin G level >1.5 normal ANA or SMA >1:80 Active infection with hepatitis A, B, or C viruses Epstein-Barr virus Cytomegalovirus a1-antitrypsin deficiency
Interface hepatitis with or without lobular hepatitis or bridging necrosis
Exclusion Blood transfusion or exposure Hepatotoxic medication Excessive alcohol use (>35 g/d in men and >25 g/d in women)
Bile duct lesions Granulomas Copper or iron accumulation Any lesions suggestive of another disease
Abbreviations: ANA, antinuclear antibody; SMA, smooth muscle antibody.
good sensitivity and specificity for excluding AIH. Because AIH has various presentations that are geographically diverse [24], the scoring system has been validated in the United States, South America, Europe, and Asia. Another factor affecting the usefulness of the scoring system is that the prevalence and types of autoimmune markers vary considerably between patient populations of different ethnic groups, ages, and genders (see later discussion on autoantibodies). Currently, an important use of the scoring system is to exclude AIH in patients who are already known to have hepatitis C. This situation occurs fairly commonly in the treatment of hepatitis C because of the relative frequency of a positive antinuclear antibody in patients who have hepatitis C. Despite its wide use, the scoring system is not successful in excluding the diagnosis of various cholestatic syndromes from AIH [13,23]. The liver biopsy remains essential to the diagnosis and evaluation of disease severity in patients who have AIH or in whom the diagnosis is being considered. The degree of elevation of the aminotransferases does not predict the histologic pattern of injury or the degree of fibrosis. A liver biopsy is also important for diagnosing overlap syndromes such as may occur between AIH and primary sclerosing cholangitis, AIH and primary biliary cirrhosis, or AIH and hepatitis C [25–28]. The liver biopsy typically has a portal cell infiltration that may extend from the portal tract into the lobule. A typical biopsy showing interface hepatitis in shown in Fig. 1 [4]. The presence of plasma cells is usually believed to be the hallmark of the disease, although large studies have demonstrated that one third of patients who have well-documented AIH may have few or no plasma cells [29]. A second classic pathologic finding in the biopsy of a patient who has AIH is a rosette of hepatocytes. A rosette is a collection of swollen hepatocytes
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
465
Table 2 Diagnostic scoring system for atypical autoimmune hepatitis in adults based on recommendations of the International Autoimmune Hepatitis Group Category
Factor
Score
Gender Alk phos: AST (or ALT) ratio
Female >3 <1.5 >2.0 1.5–2.0 1.0–1.5 <1.0 >1:80 1:80 1:40 <1:40 Positive Positive Negative Yes No <25 g/d >60 g/d Any nonhepatic disease of an immune nature Anti-SLA/LP, actin, LC1, pANCA Interface hepatitis Plasma cells Rosettes None of above Biliary changesb Atypical featuresc DR3 or DR4 Remission alone Remission with relapse
þ2 2 þ2 þ3 þ2 þ1 0 þ3 þ2 þ1 0 4 3 þ3 4 þ1 þ2 2 þ2
c-globulin or IgG (times above upper limit of normal) ANA, SMA, or anti-LKM1 titers
AMA Viral markers of active infection Hepatotoxic drugs Alcohol Concurrent immune disease Other autoantibodiesa Histologic features
HLA Treatment response Pretreatment score Definite diagnosis Probable diagnosis Posttreatment score Definite diagnosis Probable diagnosis
þ2 þ3 þ1 þ1 5 3 3 þ1 þ2 þ3 >15 10–15 >17 12–17
Abbreviations: Alk phos, serum alkaline phosphatase level; ALT, serum alanine aminotransferase level; AMA, antimitochondrial antibodies; ANA, antinuclear antibodies; AST, serum aspartate aminotransferase level; HLA, human leukocyte antigen; IgG, serum immunoglobulin G level; LKM1, antibodies to liver/kidney microsome type 1; SMA, smooth muscle antibodies. a Unconventional or generally unavailable antibodies associated with liver disease include perinuclear antineutrophil cytoplasmic antibodies (pANCA) and antibodies to actin, soluble liver antigen/liver pancreas (anti-SLA/LP), asialoglycoprotein receptor (ASGPR), and liver cytosol type 1 (LC1). b Includes destructive cholangitis, nondestructive cholangitis, or ductopenia. c Includes steatosis, iron overload consistent with genetic hemochromatosis, alcoholic hepatitis, viral features (ground-glass hepatocytes), or inclusions (cytomegalovirus, herpes simplex). Data from Ayata G, Gordon FD, Lewis WD, et al. Liver transplantation for autoimmune hepatitis: a longterm pathologic study. Hepatology 2000;32:185.
466
LUXON
Fig. 1. In this example, the hepatic cord architecture is abnormal because of extensive hepatitic rosetting (ovals). In addition, focal central necrosis is noted (arrow). (Courtesy of E. M. Brunt, MD, St. Louis, MO.)
separated from other hepatocytes that show clear signs of damage. Inflammatory cells and a collapsed stroma are also characteristic of the rosette (Fig. 2) [4]. TRADITIONAL AUTOANTIBODIES AIH is traditionally associated with varius autoantibodies. Three of them are found in most patients. These are an antinuclear antibody (ANA), a smooth muscle antibody (SMA), and an anti-liver/kidney microsomal antibody (antiLKM). Presence of these three antibodies should be routinely determined in
Fig. 2. This low-power view highlights the marked portal inflammation that characterizes most cases of AIH. Most cells in the infiltrate are plasma cells (single arrows); in addition, interface activity is severe (double arrows). (Courtesy of E. M. Brunt, MD, St. Louis, MO.)
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
467
all patients in whom the diagnosis of AIH is suspected. Although not all patients have all three antibodies (in fact this is the exception rather than the rule), these antibodies are an important component of the definition of AIH. In addition, the presence or absence of these antibodies may serve to subclassify AIH into three subtypes [13,23,30,31]. Antinuclear Antibodies Antinuclear antibodies are common markers of various immune-mediated diseases, including AIH. ANA may occur alone (10% to 15%) or with another antibody in patients who have AIH. Despite considerable research, the nuclear targets of ANA and AIH remain uncertain. In most laboratories the determination of an ANA is performed by indirect immunofluorescence. Literature from before 1990 put great emphasis on the pattern of immunofluorescence, describing either a homogenous or a speckled pattern. More recent studies have shown that the pattern of indirect immunofluorescence has no clinical significance, however [32]. For the diagnosis of AIH, it is not crucial to have a specific pattern of immunofluorescence. Knowledge of the specific molecular targets of the antibody does not increase the diagnostic precision or have prognostic importance. ANAs have been shown to react against various nuclear antigens, including the centromere, ribonucleoproteins, and ribonuclear protein complexes [10]. In patients determined to have ribonuclear protein complexes, ANAs usually do occur in high titers exceeding 1:160. The ANA titer does not accurately predict the stage of disease, the prognosis, or the activity at the time of diagnosis [33]. When interpreting laboratory tests that show the presence of antinuclear antibodies, it is important to remember that ANAs can be found in various other hepatitic and cholestatic liver diseases. For instance, other liver diseases that have been associated with a positive ANA include primary biliary cirrhosis, primary sclerosing cholangitis, chronic viral hepatitis (especially hepatitis C), drug-related hepatitis, and even nonalcoholic steatohepatitis. Evidence of a positive ANA especially in low titer does not assure that the patient has AIH. The presence of a low-titer ANA does complicate the diagnosis and potential management of patients, especially those who have documented hepatitis C [34]. Anti–Smooth Muscle Antibodies Another traditional group of antibodies found in AIH are SMAs. These are a heterogeneous group of antibodies directed against various cytoskeletal proteins, including actin, tubulin, vimentin, and desmin. The presence of an SMA is common in patients who have AIH, occurring either alone or in conjunction with an ANA in up to 87% of patients. Just as a positive ANA is not specific for a diagnosis of AIH, SMAs are not specific for AIH. They occur in other chronic liver diseases and various infectious, immunologic, and rheumatologic disorders. In most clinical laboratories, SMAs are determined by immunofluorescence using either murine stomach or kidney. The pattern of immunofluorescence again is not helpful in determining prognosis nor does
468
LUXON
it increase the diagnostic accuracy. The antibody titer of SMA does not correlate with disease course or prognosis. In contrast to ANA, it has been noted in multiple studies that SMA titers can change dramatically with time in individual patients [30]. The significance of this phenomenon is not yet known. Anti–Liver/Kidney Microsomal Antibodies The third traditional class of antibodies is antibodies directed against liver or kidney microsomes (anti-LKM). These are directed against specific cytochrome enzymes. In most laboratories the presence of these antibodies is detected by indirect immunofluorescence using activity against the proximal tubules of murine kidney or murine hepatocytes. Although the specific targets of ANA and SMA are largely unknown, there is a large body of literature regarding the targets of anti-LKM. For example, antibodies to LKM1 react with cytochrome monooxygenase CYP2D6. These antibodies also inhibit CYP2D2 activity in vitro [35,36]. This particular antigen has been studied extensively because homologies exist between CYP2D6 and the genome of the hepatitis C virus, suggesting that AIH is the result of misdirected immunologic recognition of the hepatitis C virus. This hypothesis is in contrast to the well-recognized fact that most patients who have chronic hepatitis C in the United States do not have an anti-LKM antibody. Other members of the cytochrome P450 system, including 2C9, 2A2, and 1A2, also interact with anti-LKM. Prevalence of antibodies to LKM depends on geography and age. Patients who have AIH in the United States rarely have an antibody to LKM (less than 4%) [37]. Adult European patients who have AIH are much more likely to have an anti-LKM antibody with an occurrence of up to 20% in some reports. In addition, in Europe anti-LKM is commonly found in pediatric patients. It is not clear why there are such dramatic geographic differences in the occurrence of anti-LKM. It has been suggested that genetic differences in the immune response to a particular target antigen may be responsible. Despite its relatively rare occurrence in the adult United States population, current guidelines suggest that adult patients suspected of having AIH should be tested for anti-LKM [10]. SUBTYPES OF AUTOIMMUNE HEPATITIS There has been an emphasis recently on defining subtypes of AIH. These subclassifications are based on the presence or absence of the three traditional antibodies described above. Type 1 (classical) AIH is the most common type and is the predominant form of AIH in the United States. By definition, patients who have type 1 AIH have either a positive ANA or SMA. In contrast, type 2 AIH is characterized by the presence of an anti-LKM antibody and the absence of ANA and SMA. In the United States type 2 AIH is found primarily in young children and is probably related to DRB1*0701. There was some suggestion that type 2 AIH may have a poorer prognosis than type 1 AIH. Current guidelines suggest, however, that both type 2 and type 1 respond well to corticosteroids [10]. Finally, the third type of AIH is characterized by antibodies
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
469
present to soluble liver antigen (anti-SLA/LP). This type is the least-established form of the disease and its clinical course is controversial. In practice, patients who have type 3 AIH are indistinguishable clinically from those who have type 1 AIH [38]. EXPERIMENTAL OR NONTRADITIONAL ANTIBODIES There are more than 300 articles published in the last decade describing new autoantibodies that occur in AIH. This plethora of publications is important because these new antibodies may help us in understanding the immunopathologic process in AIH. Unfortunately the laboratory tests for most of these nontraditional autoantibodies are not widely available. Only two of the several dozen nontraditional autoantibodies are described here. An antibody against the asialoglycoprotein receptor (anti-ASGPR) has been described since the 1980s [39]. Anti-ASGPR can exist in patients who have AIH who also express an ANA, SMA, or anti-LKM. In contrast to the traditional antibodies, anti-ASGPR may have prognostic importance. Its presence has been shown to correlate with histologic activity and response to treatment. If patients who have AIH continue to have anti-ASGPR, they are much more likely to relapse after corticosteroid withdrawal [40]. Antibodies directed against soluble liver antigen (anti-SLA/LP) are highly specific markers of AIH. The target of this autoantigen is a 50 kD cytosolic protein [41]. Current guidelines suggest that patients who have anti-SLA/LP do not define a valid subgroup of AIH but they are patients who have chronic hepatitis that is histologically compatible with AIH. Current laboratory tests use an immunoassay or a Western blot to detect anti-SLA/LP. From a treatment standpoint, patients who have anti-SLA/LP are more likely to relapse after corticosteroid withdrawal. TREATMENT Indications for Treatment The diagnosis of AIH does not mandate therapy in every case. The decision to treat a patient represents a careful balance between the risks and benefits of immunosuppressive therapy. As described previously, knowledge of the natural history of AIH helps guide the clinician in his decision about whether immunosuppression is needed. Steroid-based treatment regimens are the mainstay of therapy and have been documented to improve clinical and histologic features and survival in patients who have severe AIH. Three randomized controlled treatment trials were published between 1971 and 1974 and demonstrated that prednisone or prednisone in combination with azathioprine enhanced survival, improved symptoms, and showed a dramatic improvement in transaminitis. These controlled trials were done in patients who had severe type 1 AIH. Treatment trials have not been performed in patients who have less severe disease and there are no randomized controlled trials of patients who have type 2 or type 3
470
LUXON
disease. Most clinicians believe that patients who have severe type 2 or type 3 disease would also benefit from immunosuppressive therapy. Based on the clinical trials performed in the 1970s, clear indications for treatment of AIH have been published. The American Society for the Study of Liver Disease (AASLD) guidelines are summarized in Box 1 [10]. It is unlikely that treatment trials will be performed in patients who have less severe disease or type 2/type 3 AIH and hence the clinician must extrapolate from these previously published studies. The decision to treat patients who have AIH who do not satisfy the criteria in Box 1 must be individualized. Clinicians may use symptoms, the liver biopsy, and the potential for complications as decision points in deciding if immunosuppressive therapy is needed. For example, patients who have only a mild to moderate elevation in their transaminases and evidence of mild interface hepatitis on their liver biopsy have a low probability of cirrhosis and a normal 5-year life expectancy [11,42]. Even patients who have advanced fibrosis or cirrhosis may still benefit from a short-term (3–6 months) treatment trial. Those who have inflammation on their biopsy are most likely to respond. It is also these patients, however, who are much more likely to suffer drug-related complications. Several studies [43] have shown that patients who have cirrhosis have a higher frequency of complications (25% versus 8%) compared with patients who have less fibrosis. A potential explanation for this is that serum levels of unbound prednisolone that result from decreased protein binding may lead to untoward side effects. Nonetheless, patients who have significant fibrosis but have pronounced inflammation on their biopsy may benefit from treatment. As a final issue, however, it is unlikely that established fibrosis or cirrhosis will resolve completely with treatment. Current thinking is that treatment with immunosuppression may delay or obviate the need for liver transplantation in these patients. Patients who have
Box 1: Indications for treatment Absolute Serum AST>10-fold upper limit of normal Serum AST>5-fold upper limit of normal and c-globulin greater than twice normal Bridging necrosis or multiacinar necrosis on histologic examination Relative Symptoms (fatigue, arthralgias, jaundice) Serum AST or c-globulin less than absolute criteria Interface hepatitis Abbreviation: AST, serum aspartate aminotransferase. Data from Czaja AJ, Freese DK, American Association for the Study of Liver Disease. Diagnosis and treatment of autoimmune hepatitis. Hepatology 2002;36:479.
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
471
decompensated liver disease (Mayo Endstage Liver Disease Model [MELD] score greater than 15) attributable to AIH should not be treated with steroid therapy and instead should be considered for orthotopic liver transplantation [10,20]. Medical Therapy The treatment of AIH has remained unchanged for nearly 5 decades. Most experts agree that when possible combination therapy with prednisone and azathioprine is the preferred initial treatment of patients who have type 1 AIH [10,12,20]. The AASLD guidelines are shown in Table 3 as two recommended treatment regimens for adults. The combination regimen is associated with fewer side effects attributable to steroids than using the higher dose prednisone alone (10% versus 44%) [44]. There have been multiple modifications of the regimens shown in Table 3 but none of these have been shown to be superior to the standard regimen [10]. The proposed modifications have included giving patients who have advanced cirrhosis prednisolone rather than prednisone, administering prednisone on alternate days, or modifying the dose based on gender, age, or body weight [11]. Side Effects Patients who have AIH who are treated with steroids develop the typical cosmetic changes attributable to prednisone use. These include facial rounding, acne, truncal obesity, and potentially the disabling complication of osteoporosis. There are occasional reports in the literature of patients receiving high-dose monotherapy with prednisone developing hypertension and psychosis. These are exceedingly rare and only 13% of treated patients develop complications that require dose reduction of their steroid [10]. Remembering that patients who have AIH are often middle-aged women, the most common reasons for treatment withdrawal are cosmetic changes, osteopenia with vertebral compression, or worsening of diabetes [45]. To avoid the typical side effects of prednisone, most treating physicians use azathioprine as a steroid-sparing agent. Side effects from azathioprine are relatively rare and occur in less than 10% of patients. In patients who are receiving azathioprine, it has been postulated that it is important to measure the level of Table 3 Suggested doses for initial treatment of type I autoimmune hepatitis Combination therapy Time Week 1 Week 2 Week 3 Week 4 Until clinical endpoint reached
Prednisone 30 20 15 10 10
Azathioprine
Steroids alone Prednisone
50 50 50 50 50
60 40 30 30 20
Combination therapy using both prednisone and azathioprine is preferred initial therapy [37]. Steroids alone should be used in patients who have severe cytopenia, have a concurrent neoplasm, are pregnant or wish to become pregnant, or have known intolerance for azathioprine. All doses given in mg/d.
472
LUXON
thiopurine methyltransferase. Thiopurine methyltransferase is important in the elimination of 6-mecaptopurine and variations in its enzymatic activity affect of the therapeutic action of azathioprine. Low enzyme activity is rare and occurs in less than 1% of the population, however. Heterozygosity (which results in intermediate activity of thiopurine methyltransferase) is much more common and is found in up to 11% of the population. At the present time pretreatment testing for thiopurine methyltransferase activity is a reasonable clinical precaution but it has not been established that routine testing needs to be done in patients who have AIH. Patients who develop cytopenia or have a pre-existing cytopenia should be tested to avoid potential overdosing of azathioprine. In any case, patients treated with azathioprine should be monitored for the development of drug-related side effects, including pancytopenia. Evaluation of the Treatment Response The response to the initiation of immunosuppression therapy is determined on clinical and biochemical grounds [20]. It has been believed that failure to normalize the serum aminotransferases or to reduce gamma globulin levels means that the AIH is not being adequately treated. There is no prescribed minimum or maximum duration of treatment. Instead improvement in the biochemical markers of AIH is expected and treatment is continued until the AIH is judged to be in remission. An important caveat is that histologic improvement may not occur despite improvement in the aminotransferases. Clinical trials have demonstrated that AIH-induced fibrosis can progress despite normalization of aminotransferases. Prospective clinical trials have demonstrated that 5% to 10% of patients who have achieved persistently normal aminotransferases progress to significant fibrosis and even cirrhosis. It is also believed that this group of patients invariably will suffer relapse if immunosuppression is withdrawn [46]. In addition, it is generally believed that histologic improvement often lags behind biochemical improvement by as much as 6 months [20]. A repeat liver biopsy should therefore be considered following biochemical remission. Nearly two thirds of patients enter remission within 18 months and 80% achieve remission within 3 years. In these patients the successful treatment of AIH includes disappearance of the patient’s symptoms, return of their bilirubin and gamma globulin levels to normal, and a lowering of the serum aminotransferase levels to normal or less than twice normal. It is usually assumed but it is not necessarily the case that a biopsy at this point shows normal hepatic tissue or minimal inflammation with no interface hepatitis. In these patients who are believed to be in remission, it is prudent to begin gradual withdrawal of prednisone with the potential to discontinue azathioprine. It is equally important to continue regular monitoring to assess for the possibility of relapse. Using a second liver biopsy to determine treatment endpoints based on the patient’s histologic response remains controversial. Some authors have suggested that normal histology on repeat liver biopsy is mandatory before attempting to withdraw treatment. The phenomenon of a completely normal
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
473
liver biopsy occurs relatively infrequently, however, even in patients who undergo liver biopsy 6 months after normalizing their aminotransferases. Predictors of Relapse Considerable effort has been made to prospectively identify patients who may be withdrawn safely from immunosuppression without suffering a relapse. A study by Hegarty and colleagues [47] demonstrated that as many as 9 out of 10 patients would relapse despite being in remission for many years before treatment was withdrawn. Multiple studies have emphasized the need for histologic resolution before consideration of withdrawal of immunosuppressive therapy. Even histologic resolution of hepatitis is not sufficient to guarantee that the patient will remain in remission, however. An interesting report published by Montano-Loza and colleagues [48] demonstrated that patients who were in histologic remission often relapsed following withdrawal of therapy. In this article, 77% of patients relapsed despite being treated for 24 months before attempting withdrawal of treatment. All of these patients had serum AST within twofold of the upper limits of normal and were asymptomatic. In addition they all had the absence of interface hepatitis on biopsy. It was difficult to identify patients who might remain in remission. Al-Chalabi and Heneghan [49] looked at length of therapy, presence of a normal AST, and gamma globulins, in addition to biopsy findings. Overall most clinicians believe that it remains impossible to predict which patients will relapse. One potential variable that is still being investigated is the duration of therapy before withdrawal. Findings seem at least in some studies to suggest that longer duration of therapy before withdrawal is an important factor involved in reducing the relapse rate in AIH. SPECIAL CIRCUMSTANCES Pregnancy AIH is prevalent among young women in their childbearing years. Although women who have cirrhosis at the time of their AIH diagnosis are less likely to become pregnant often because of hypothalamic pituitary dysfunction, other patients who have less severe forms of AIH may become pregnant. The question remains whether pregnancy is a threat for the patient or the fetus. Several groups have done retrospective studies of pregnancy in patients who have AIH. Early studies suggested that there was a high rate of fetal and maternal complications. A fetal loss rate of 14% was reported in a large case series [50]. Other studies have suggested that pregnancy may not be contraindicated in well-controlled AIH, especially in those patients who have relatively mild liver disease. In a study by Heneghan and colleagues [50], the pregnancies of 162 women who had definite AIH were reviewed. Relatively few patients had flares in disease activity during pregnancy and most patients had normal rates of complication, including eclampsia and fetal loss. In patients who had cirrhosis, Heneghan found that complications were increased. Concern has been raised about using azathioprine during pregnancy because of animal models showing that high doses of azathioprine are teratogenic. In
474
LUXON
published series, physicians have advised patients who were on azathioprine to stop the medication if possible. An increase in the risk for a flare following cessation of azathioprine also poses risks to the pregnant patient who has AIH and to her fetus. Current suggestions are that patients who have stable AIH who are on azathioprine could continue with azathioprine at low doses [20]. Several authors have suggested that AIH may improve spontaneously during pregnancy. Buchel and colleagues [51] reported that in a series of 98 patients, 14 pregnancies occurred. In all the patients studied, the dose of corticosteroids could be reduced during pregnancy. There were no flares of AIH during pregnancy. These authors concluded, therefore, that pregnancy leads to an attenuation of the immune process, possibly the result of a transition of TH1 to TH2 predominance during pregnancy. Other authors have suggested that there is high likelihood of an AIH flare in the peripartum time period and that patients should have appropriate prophylaxis with increased immunosuppression following delivery [20,50–53]. Autoimmune Hepatitis after Liver Transplant In some patients, AIH progresses to end-stage liver disease, necessitating a liver transplant. Patients who have AIH who require transplant are usually young women who present with acute deterioration and have cirrhosis at the time of presentation with appropriately high MELD scores. Fortunately these patients are relatively rare. Various studies have estimated that less than 10% of patients who have AIH do not respond to medical therapy and may need a liver transplant. The issue of whether AIH recurs after transplant has been extensively studied. The estimated risk for recurrence of AIH after liver transplant is 8% for the first year and 25% after 2 years [54,55]. The timing of the recurrence and the diagnosis of the recurrence is still an issue of some debate. Unfortunately, there are currently no defined criteria for diagnosis of AIH recurrence after transplant. Most transplant centers use criteria similar to those used to diagnose AIH in the nontransplanted patient. Because many of the antibodies and other diagnostic tests persist after transplant even in a patient who does not have recurrence of AIH, the diagnosis of AIH in a patient posttransplant is difficult. In a 10-year follow-up of French patients transplanted for AIH, histologic recurrence occurred before clinical or biochemical recurrence [56]. In this study protocol biopsies were done documenting histologic recurrence. Histologic occurrence was documented a mean of 2.5 years after the transplant. This French study demonstrated that biologic and histologic recurrence of AIH may occur more than 10 years after transplant. Other studies have looked at risk factors that might predict AIH recurrence after a transplant. High-grade inflammation in the native liver has been reported as a strong predictor of recurrence of AIH [57]. Some evidence points to an increase frequency of HLA-DR3 haplotype in patients who have AIH occurrence although this is not universal [57,58]. HLA antigen mismatch between
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
475
the graft and the host was not found to differ significantly in patients who had AIH recurrence [57,59]. Studies have also looked at whether the choice of calcineurin inhibitors influences the likelihood of AIH recurrence. In a systematic review by Gautam and colleagues [60], 16 studies reporting AIH recurrence after transplant were summarized. There was no difference in AIH recurrence based on either the choice of the initial immunosuppression or the rate of corticosteroid withdrawal. No firm recommendations can be made about the choice of calcineurin inhibitors or the use of steroids posttransplantation in an attempt to limit recurrence of AIH posttransplant. SUMMARY AIH is a chronic inflammatory liver disease with an unknown cause that is associated with various autoantibodies. This immune disease affecting the liver responds well to prednisone or a combination of prednisone and azathioprine. The specific criteria for the diagnosis of AIH incorporate a wide range of biochemical, histologic, and immunologic features. With treatment, most patients can be brought into remission, although many patients require maintenance therapy with low-dose prednisone and azathioprine. In treating patients with immunosuppressive therapy, combination therapy is initially preferred because of a lower rate of adverse affects from corticosteroids. Not all patients need to be treated, even once the diagnosis of AIH is confirmed. Drug therapy should be considered in patients who have cirrhosis if biopsy demonstrates considerable inflammation. Finally, some patients, despite optimal medical management, require orthotopic liver transplantation. Liver transplantation should be considered in patients who have decompensated cirrhosis attributable to AIH or those who have severe fulminant hepatitis who fail to respond to initial therapy. References [1] Boberg KM, Aadland E, Jahnsen J, et al. Incidence and prevalence of primary biliary cirrhosis, primary sclerosing cholangitis, and autoimmune hepatitis in a Norwegian population. Scand J Gastroenterol 1998;33:99. [2] Wiesner RH, Demetris AJ, Belle SH, et al. Acute hepatic allograft rejection: incidence, risk factors, and impact on outcome. Hepatology 1998;28:638. [3] Luxon BA. Autoimmune hepatitis. Making sense of all those antibodies. Postgrad Med 2003;114:79. [4] Luxon BA. Autoimmune hepatitis: diagnosis, treatment, and prognosis. Curr Gastroenterol Rep 2006;8:83. [5] Czaja AJ, Carpenter HA. Distinctive clinical phenotype and treatment outcome of type 1 autoimmune hepatitis in the elderly. Hepatology 2006;43:532. [6] Soloway RD, Summerskill WH, Baggenstoss AH, et al. Clinical, biochemical, and histological remission of severe chronic active liver disease: a controlled study of treatments and early prognosis. Gastroenterology 1972;63:820. [7] Murray-Lyon IM, Stern RB, Williams R. Controlled trial of prednisone and azathioprine in active chronic hepatitis. Lancet 1973;1:735. [8] Czaja AJ. Natural history, clinical features, and treatment of autoimmune hepatitis. Semin Liver Dis 1984;4:1.
476
LUXON
[9] Feld JJ, Dinh H, Arenovich T, et al. Autoimmune hepatitis: effect of symptoms and cirrhosis on natural history and outcome. Hepatology 2005;42:53. [10] Czaja AJ, Freese DK. Diagnosis and treatment of autoimmune hepatitis. Hepatology 2002;36:479. [11] Schalm SW, Korman MG, Summerskill WH, et al. Severe chronic active liver disease. Prognostic significance of initial morphologic patterns. Am J Dig Dis 1977;22:973. [12] Czaja AJ. Drug therapy in the management of type 1 autoimmune hepatitis. Drugs 1999;57:49. [13] Alvarez F, Berg PA, Bianchi FB, et al. International autoimmune hepatitis group report: review of criteria for diagnosis of autoimmune hepatitis. J Hepatol 1999;31:929. [14] Kogan J, Safadi R, Ashur Y, et al. Prognosis of symptomatic versus asymptomatic autoimmune hepatitis: a study of 68 patients. J Clin Gastroenterol 2002;35:75. [15] Davis GL, Czaja AJ, Ludwig J. Development and prognosis of histologic cirrhosis in corticosteroid-treated hepatitis B surface antigen-negative chronic active hepatitis. Gastroenterology 1984;87:1222. [16] Verma S, Gunuwan B, Mendler M, et al. Factors predicting relapse and poor outcome in type I autoimmune hepatitis: role of cirrhosis development, patterns of transaminases during remission and plasma cell activity in the liver biopsy. Am J Gastroenterol 2004;99:1510. [17] Roberts SK, Therneau TM, Czaja AJ. Prognosis of histological cirrhosis in type 1 autoimmune hepatitis. Gastroenterology 1996;110:848. [18] Czaja AJ. Treatment strategies in autoimmune hepatitis. Clin Liver Dis 2002;6:799. [19] Verma S, Redeker A. In type 1 autoimmune hepatitis, is cirrhosis at presentation or follow-up associated with a poorer outcome? Hepatology 2005;42:1237 [author reply 1237]. [20] Heathcote J. Treatment strategies for autoimmune hepatitis. Am J Gastroenterol 2006;101: S630. [21] Dufour JF, DeLellis R, Kaplan MM. Reversibility of hepatic fibrosis in autoimmune hepatitis. Ann Intern Med 1997;127:981. [22] Czaja AJ, Carpenter HA. Progressive fibrosis during corticosteroid therapy of autoimmune hepatitis. Hepatology 2004;39:1631. [23] Johnson PJ, McFarlane IG. Meeting report: international autoimmune hepatitis group. Hepatology 1993;18:998. [24] Goldberg AC, Bittencourt PL, Oliveira LC, et al. Autoimmune hepatitis in Brazil: an overview. Scand J Immunol 2007;66:208. [25] Ben-Ari Z, Czaja AJ. Autoimmune hepatitis and its variant syndromes. Gut 2001;49:589. [26] Czaja AJ. Frequency and nature of the variant syndromes of autoimmune liver disease. Hepatology 1998;28:360. [27] Czaja AJ, Carpenter HA. Autoimmune hepatitis with incidental histologic features of bile duct injury. Hepatology 2001;34:659. [28] Czaja AJ, Carpenter HA, Santrach PJ, et al. Autoimmune cholangitis within the spectrum of autoimmune liver disease. Hepatology 2000;31:1231. [29] Czaja AJ, Carpenter HA. Sensitivity, specificity, and predictability of biopsy interpretations in chronic hepatitis. Gastroenterology 1993;105:1824. [30] Czaja AJ, Manns MP. The validity and importance of subtypes in autoimmune hepatitis: a point of view. Am J Gastroenterol 1995;90:1206. [31] Meyer O, Abuaf N, Cyna L, et al. Anti-mitochondrial type 5 antibodies and anti-cardiolipin antibodies in systemic lupus erythematosus and auto-immune diseases. Clin Exp Immunol 1987;69:485. [32] Czaja AJ, Cassani F, Cataleta M, et al. Antinuclear antibodies and patterns of nuclear immunofluorescence in type 1 autoimmune hepatitis. Dig Dis Sci 1997;42:1688. [33] Czaja AJ. Behavior and significance of autoantibodies in type 1 autoimmune hepatitis. J Hepatol 1999;30:394.
DIAGNOSIS AND TREATMENT OF AUTOIMMUNE HEPATITIS
477
[34] Kogure T, Ueno Y, Fukushima K, et al. Fulminant hepatic failure in a case of autoimmune hepatitis in hepatitis C during peg-interferon-alpha 2b plus ribavirin treatment. World J Gastroenterol 2007;13:4394. [35] Manns M, Zanger U, Gerken G, et al. Patients with type II autoimmune hepatitis express functionally intact cytochrome P-450 db1 that is inhibited by LKM-1 autoantibodies in vitro but not in vivo. Hepatology 1990;12:127. [36] Manns MP, Griffin KJ, Sullivan KF, et al. LKM-1 autoantibodies recognize a short linear sequence in P450IID6, a cytochrome P-450 monooxygenase. J Clin Invest 1991;88:1370. [37] Czaja AJ, Manns MP, Homburger HA. Frequency and significance of antibodies to liver/ kidney microsome type 1 in adults with chronic active hepatitis. Gastroenterology 1992;103:1290. [38] Czaja AJ, Freese DK. American Association for the Study of Liver Disease: diagnosis and treatment of autoimmune hepatitis. Hepatology 2002;36:479. [39] McFarlane IG, Hegarty JE, McSorley CG, et al. Antibodies to liver-specific protein predict outcome of treatment withdrawal in autoimmune chronic active hepatitis. Lancet 1984;2:954. [40] Treichel U, Gerken G, Rossol S, et al. Autoantibodies against the human asialoglycoprotein receptor: effects of therapy in autoimmune and virus-induced chronic active hepatitis. J Hepatol 1993;19:55. [41] Wies I, Brunner S, Henninger J, et al. Identification of target antigen for SLA/LP autoantibodies in autoimmune hepatitis. Lancet 2000;355:1510. [42] Baggenstoss AH, Soloway RD, Summerskill WH, et al. Chronic active liver disease. The range of histologic lesions, their response to treatment, and evolution. Hum Pathol 1972;3:183. [43] Uribe M, Go VL, Kluge D. Prednisone for chronic active hepatitis: pharmacokinetics and serum binding in patients with chronic active hepatitis and steroid major side effects. J Clin Gastroenterol 1984;6:331. [44] Summerskill WH, Korman MG, Ammon HV, et al. Prednisone for chronic active liver disease: dose titration, standard dose, and combination with azathioprine compared. Gut 1975;16:876. [45] Czaja AJ, Davis GL, Ludwig J, et al. Complete resolution of inflammatory activity following corticosteroid treatment of HBsAg-negative chronic active hepatitis. Hepatology 1984;4: 622. [46] Czaja AJ, Wolf AM, Baggenstoss AH. Laboratory assessment of severe chronic active liver disease during and after corticosteroid therapy: correlation of serum transaminase and gamma globulin levels with histologic features. Gastroenterology 1981;80:687. [47] Hegarty JE, Nouri Aria KT, Portmann B, et al. Relapse following treatment withdrawal in patients with autoimmune chronic active hepatitis. Hepatology 1983;3:685. [48] Montano-Loza AJ, Carpenter HA, Czaja AJ. Improving the end point of corticosteroid therapy in type 1 autoimmune hepatitis to reduce the frequency of relapse. Am J Gastroenterol 2007;102:1005. [49] Al-Chalabi T, Heneghan MA. Remission in autoimmune hepatitis: what is it, and can it ever be achieved? Am J Gastroenterol 2007;102:1013. [50] Heneghan MA, Norris SM, O’Grady JG, et al. Management and outcome of pregnancy in autoimmune hepatitis. Gut 2001;48:97. [51] Buchel E, Van Steenbergen W, Nevens F, et al. Improvement of autoimmune hepatitis during pregnancy followed by flare-up after delivery. Am J Gastroenterol 2002;97:3160. [52] Schramm C, Herkel J, Beuers U, et al. Pregnancy in autoimmune hepatitis: outcome and risk factors. Am J Gastroenterol 2006;101:556. [53] Steven MM, Buckley JD, Mackay IR. Pregnancy in chronic active hepatitis. Q J Med 1979;48:519. [54] Prados E, Cuervas-Mons V, de la Mata M, et al. Outcome of autoimmune hepatitis after liver transplantation. Transplantation 1998;66:1645.
478
LUXON
[55] Wright HL, Bou-Abboud CF, Hassanein T, et al. Disease recurrence and rejection following liver transplantation for autoimmune chronic active liver disease. Transplantation 1992;53: 136. [56] Duclos-Vallee JC, Sebagh M, Rifai K, et al. A 10 year follow up study of patients transplanted for autoimmune hepatitis: histological recurrence precedes clinical and biochemical recurrence. Gut 2003;52:893. [57] Ayata G, Gordon FD, Lewis WD, et al. Liver transplantation for autoimmune hepatitis: a long-term pathologic study. Hepatology 2000;32:185. [58] Gonzalez-Koch A, Czaja AJ, Carpenter HA, et al. Recurrent autoimmune hepatitis after orthotopic liver transplantation. Liver Transpl 2001;7:302. [59] Milkiewicz P, Hubscher SG, Skiba G, et al. Recurrence of autoimmune hepatitis after liver transplantation. Transplantation 1999;68:253. [60] Gautam M, Cheruvattath R, Balan V. Recurrence of autoimmune liver disease after liver transplantation: a systematic review. Liver Transpl 2006;12:1813.
Gastroenterol Clin N Am 37 (2008) 479–484
GASTROENTEROLOGY CLINICS OF NORTH AMERICA
Antimitochondrial Antibody–Negative Primary Biliary Cirrhosis Flavia Mendes, MDa, Keith D. Lindor, MDb,* a
Division of Hepatology, University of Miami Miller School of Medicine, 1500 NW 12 Avenue, Suite 1101, East Tower, Miami, FL 33136, USA b Division of Gastroenterology and Hepatology, Mayo Medical School, Mayo Clinic College of Medicine, 200 1st Street SW, Rochester, MN 55905, USA
P
rimary biliary cirrhosis (PBC) is a chronic cholestatic liver disease of unclear cause. It is characterized histologically by nonsuppurative destruction of the bile ducts and serologically by the presence of antimitochondrial antibodies (AMA). Several reports indicate that in approximately 5% of patients serologic evidence of AMA is absent, even though these patients share similar clinical, biochemical, and histologic features of PBC with those who are AMA-positive [1]. There is some controversy as to whether autoimmune cholangitis (AIC), AMA-negative PBC, and AMA-positive PBC represent the same entities clinically and immunologically. According to Czaja and colleagues [2] AIC cannot be assimilated into a single diagnostic category. It may represent variant forms of different autoimmune liver diseases, a transition stage between two autoimmune disorders, or a separate entity with varying manifestations. The advent of more sensitive and specific techniques other than indirect immunofluorescence, such as ELISA or immunoblotting, to determine the presence of AMA in patients suspected to have PBC have disclosed that as many as 79% of patients initially considered AMA-negative are in fact AMA-positive [3–7], strengthening the suspicion that the two are indeed one single condition. This article reviews the clinical, biochemical, serologic, and histopathologic features and treatment approach and outcomes in patients who have AMA-negative PBC. The term AMA-negative PBC is used in preference to AIC. CLINICAL AND BIOCHEMICAL FEATURES AMA-negative PBC, like AMA-positive PBC, occurs predominantly in middleaged women and is commonly asymptomatic. When symptoms are present, fatigue, pruritus, and symptoms of sicca syndrome are common initial features, whereas jaundice is usually a late finding. Elevation of the alkaline phosphatase is the typical biochemical feature, with transaminases being only mildly *Corresponding author. E-mail address:
[email protected] (K.D. Lindor). 0889-8553/08/$ – see front matter doi:10.1016/j.gtc.2008.02.006
ª 2008 Elsevier Inc. All rights reserved. gastro.theclinics.com
480
MENDES & LINDOR
elevated. Several studies comparing the clinical and laboratorial findings between patients who have AMA-negative PBC and classic PBC have been published and there is no disagreement that the two entities behave similarly [4–6,8,9]. Table 1 summarizes the most important studies comparing the various features of AMA-negative and -positive PBC. Michieletti and colleagues [5] reported a group of 20 patients who had AMAnegative PBC compared with 20 AMA-positive controls, matched for gender and serum bilirubin. There was no significant difference with respect to signs, symptoms, and associated autoimmune phenomena at the time of diagnosis. Also serum IgG, alkaline phosphatase, and alanine aminotransferase levels were not significantly different. Similar results were reported by Lacerda and colleagues [9] in 35 patients who had AMA-negative PBC compared with 180 patients who had classic PBC and later on by Invernizzi and colleagues [4], who compared 24 patients who were AMA-negative to 273 patients who were AMA-positive seen over a 2-decade period. There were no significant differences with regard to gender, age, incidence of complications, and development of liver failure resulting in death or referral for liver transplantation between the two populations. SEROLOGIC FEATURES Some serologic characteristics aside from the absence of the antimitochondrial antibody may be different between AMA-negative and -positive PBC. Several studies have reported a significantly higher rate of positivity for antinuclear antibodies (ANA) and smooth muscle antibodies (SMA) in patients who have AMA-negative PBC than that seen in AMA-positive PBC. In the different studies the prevalence of ANA ranged between 71% to 100%, and of SMA between 14% to 41% in patients who had AMA-negative PBC, compared with 18% to 33% and 6% to 14%, respectively, in patients who had classic PBC [4–6,9]. Serum concentration of IgM seems to be lower and gamma globulin higher in patients who have AMA-negative PBC [4–6,9]. Some authors have suggested that antibodies to carbonic anhydrase II and antibodies to lactoferrin may be seen more commonly in patients who have AMA-negative PBC than in classic PBC, but no disease specificity has been noted [3]. HISTOPATHOLOGIC FEATURES According to the current thinking classic PBC and AMA-negative PBC share the same histologic features, but minor differences might emerge as more experience is gained in the latter group. Some authors have identified differences in the inflammatory infiltrates of patients who have AMA-negative and AMApositive PBC [10,11]. Ludwig [12] proposes applying the same criteria for diagnosis and staging used for PBC to AIC or AMA-negative PBC. Stage 1 (portal stage) is characterized by infiltration of portal tracts by lymphocytes, neutrophils, and eosinophils.
Michieletti, et al [5] AMA No. of patients Age (y) Female gender (%) No. of asymptomatic (%) Complications of cirrhosis Histologic findings
ALP AST ALT TB Number with þANA Number with þSMA IgM Response to UDCA treatment
AMAþ
Lacerda, et al [9]
Invernizzi, et al [4]
Nakajima, et al [6]
AMAþ
AMA
AMA
AMAþ
AMA
35 52 11 All NA
180 53 9
8 57.9 7.2 80(100) NA
89 54.2 9.5 81(91)
NA
23 71 53.3 13.7 52.3 9.4 23(100) 63(89) Higher in AMA 21(91) 48(68) NA
Slightly less portal inflammation in AMA No difference Lower in AMA No difference Matched Higher in AMA
NA
NA
NA
No difference No difference NA No difference Higher in AMA
24 273 52 11 54 11 21(8) 244(89) No difference 17(71) 166(61) No difference 3(13) 22(8) No difference in proportion of patients who had stage II–IV No difference NA No difference No difference Higher in AMA
AMAþ
NA No difference NA No difference NA
Lower in AMA No difference No difference Lower in AMA Higher in AMA
Lower in AMA
No difference
Higher in AMA
NA
Higher in AMA
Lower in AMA NA
Lower in AMA No difference (biochemical and Mayo risk score)
Lower in AMA No difference (biochemical)
NA No difference (biochemical and clinical outcome—death or OLT)
Lower in AMA NA
Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; ANA, antinuclear antibodies; AST, aspartate aminotransferase; NA, not assessed; OLT, orthotopic liver transplantation; SMA, smooth muscle antibodies; TB, total bilirubin; UDCA, ursodeoxycholic acid.
481
AMA
17 17 51.1 55.0 Matched No difference 8(47) 6(35) NA
NA
AMAþ
Kim, et al [8]
NEGATIVE PRIMARY BILIARY CIRRHOSIS
Table 1 Studies comparing patients who have antimitochondrial antibody–negative and antimitochondrial antibody–positive primary biliary cirrhosis
482
MENDES & LINDOR
The typical duct destruction with granuloma formation, the ‘‘florid duct lesion,’’ is virtually diagnostic, but rarely found. In stage 2 (periportal stage) the inflammatory infiltrates described for stage 1 have increased and involve the periportal parenchyma, often accompanied by prominent portal and periportal ductular proliferation (Fig. 1). Stage 3 (septal stage) is described by the presence of bridging fibrosis, but the features of stage 2 may still be present. Ductopenia and features of cholestasis become noticeable in most instances. Finally stage 4 (cirrhotic stage) is characterized by biliary cirrhosis with nodular regeneration of the hepatic parenchyma and prominent ductopenia. RESPONSE TO TREATMENT The treatment of PBC involves two treatment modalities: ursodeoxycholic acid (UDCA) and orthotopic liver transplantation (OLT). UDCA has been shown to improve liver biochemistry and to slow disease progression, delaying time to death or liver transplantation in patients who have PBC, especially in those who have earlier-stage disease [13,14]. In patients who have end-stage liver disease from PBC, OLT has been shown to prolong survival and to improve quality of life [15,16]. Given the small number of patients who have AMAnegative PBC, and that positive AMA is used as a diagnostic criterion for enrollment in clinical trials, fewer data are available with respect to outcomes of treatment in patients who have AMA-negative PBC. A study by Kim and colleagues [8] evaluated the effect of AMA status in the outcome of patients treated with UDCA and OLT. In this study eight patients who had AMA-negative PBC who were excluded from enrollment in a randomized clinical trial because of the absence of AMA were treated with 13 to 15 mg/ kg/d UDCA and followed at regular intervals. The results of UDCA treatment were compared with the 89 patients who had AMA-positive PBC who were enrolled in the therapeutic arm of the clinical trial. Biochemical improvement
Fig. 1. Typical changes of stage 2 primary biliary cirrhosis with lymphocytic cholangitis and mild interface hepatitis.
NEGATIVE PRIMARY BILIARY CIRRHOSIS
483
and main clinical outcomes (death from hepatic failure or liver transplantation) were comparable between patients who were AMA-negative and those who were AMA-positive. In the same study they reported the outcomes of six patients who had AMA-negative PBC who had undergone OLT, compared with two matched AMA-positive controls for each case. After a median of 36 months of follow-up, graft and patient survival rates and subsequent histologic changes (disease recurrence and steroid-resistant or late rejections) were similar in the two groups. They concluded that AMA status does not affect the response in patients who have PBC to treatment with UDCA or OLT [8]. Gisbert and colleagues [17] reviewed the published studies on therapeutic regimens for AMA-negative PBC. In general the response to UDCA treatment did not seem to be affected by the AMA status. In 13 uncontrolled studies a total of 52 patients were treated with UDCA and serum biochemical improvement was seen in 83% of those patients. Also a favorable effect of immunosuppressive drugs occurred in 57% of 54 patients who had AMA-negative PBC in 17 uncontrolled trials. The problem with these trials is that they included very few patients and most evaluated the effects of treatment on surrogate markers of disease, as opposed to evaluating the effects on histology, survival, or need for liver transplantation. In general it is recommended that patients who have AMA-negative PBC be treated with UDCA 13 to 15 mg/kg/d, no different from what is recommended for patients who have classic PBC, because there is no evidence to support that outcomes from treatment are different. PROGNOSIS Several models to predict survival in patients who have PBC have been developed [18–20]; the Mayo risk score is the most widely validated. It takes into account age, total bilirubin, serum albumin, prothrombin time, and presence or absence of edema [19]. Although these models have not been validated specifically in the AMA-negative population, it is reasonable to use them in this setting, given that most of the evidence available supports that the prognosis and natural history of AMA-negative PBC is comparable to AMA-positive PBC. SUMMARY AMA-negative PBC, also known as AIC, shares similar clinical, biochemical, histologic, and prognostic features with classic PBC. Management of AMAnegative PBC should not differ from treatment of AMA-positive disease. The development of more sensitive and specific assays and increased understanding of the immunopathogenetic mechanisms of these diseases may, in the near future, prove them to be the same disease entity. References [1] Kaplan MM, Gershwin ME. Primary biliary cirrhosis. N Engl J Med 2005;353(12): 1261–73. [2] Czaja AJ, Carpenter HA, Santrach PJ, et al. Autoimmune cholangitis within the spectrum of autoimmune liver disease. Hepatology 2000;31(6):1231–8.
484
MENDES & LINDOR
[3] Vierling JM. Primary biliary cirrhosis and autoimmune cholangiopathy. Clin Liver Dis 2004;8(1):177–94. [4] Invernizzi P, Crosignani A, Battezzati PM, et al. Comparison of the clinical features and clinical course of antimitochondrial antibody-positive and -negative primary biliary cirrhosis. Hepatology 1997;25(5):1090–5. [5] Michieletti P, Wanless IR, Katz A, et al. Antimitochondrial antibody negative primary biliary cirrhosis: a distinct syndrome of autoimmune cholangitis. Gut 1994;35(2):260–5. [6] Nakajima M, Shimizu H, Miyazaki A, et al. Detection of IgA, IgM, and IgG subclasses of anti-M2 antibody by immunoblotting in autoimmune cholangitis: is autoimmune cholangitis an early stage of primary biliary cirrhosis? J Gastroenterol 1999;34(5):607–12. [7] Kitami N, Komada T, Ishii H, et al. Immunological study of anti-M2 in antimitochondrial antibody-negative primary biliary cirrhosis. Intern Med 1995;34(6):496–501. [8] Kim WR, Poterucha JJ, Jorgensen RA, et al. Does antimitochondrial antibody status affect response to treatment in patients with primary biliary cirrhosis? Outcomes of ursodeoxycholic acid therapy and liver transplantation. Hepatology 1997;26(1):22–6. [9] Lacerda MA, Ludwig J, Dickson ER, et al. Antimitochondrial antibody-negative primary biliary cirrhosis. Am J Gastroenterol 1995;90(2):247–9. [10] O’Donohue J, Wong T, Portmann B, et al. Immunohistochemical differences in the portal tract and acinar infiltrates between primary biliary cirrhosis and autoimmune cholangitis. Eur J Gastroenterol Hepatol 2002;14(10):1143–50. [11] Watanabe S, Deguchi A, Uchida N, et al. Histopathologic comparison of anti-mitochondrial antibody-positive primary biliary cirrhosis and autoimmune cholangiopathy. Hepatol Res 2001;19(1):41–51. [12] Ludwig J. The pathology of primary biliary cirrhosis and autoimmune cholangitis. Baillieres Best Pract Res Clin Gastroenterol 2000;14(4):601–13. [13] Lindor KD, Dickson ER, Baldus WP, et al. Ursodeoxycholic acid in the treatment of primary biliary cirrhosis. Gastroenterology 1994;106(5):1284–90. [14] Poupon RE, Poupon R, Balkau B. Ursodiol for the long-term treatment of primary biliary cirrhosis. The UDCA-PBC Study Group. N Engl J Med 1994;330(19):1342–7. [15] Markus BH, Dickson ER, Grambsch PM, et al. Efficiency of liver transplantation in patients with primary biliary cirrhosis. N Engl J Med 1989;320(26):1709–13. [16] Starzl TE, Demetris AJ, Van Thiel D. Liver transplantation (1). N Engl J Med 1989;321(15): 1014–22. [17] Gisbert JP, Jones EA, Pajares JM, et al. Review article: is there an optimal therapeutic regimen for antimitochondrial antibody-negative primary biliary cirrhosis (autoimmune cholangitis)? Aliment Pharmacol Ther 2003;17(1):17–27. [18] Christensen E, Altman DG, Neuberger J, et al. Updating prognosis in primary biliary cirrhosis using a time-dependent Cox regression model. PBC1 and PBC2 trial groups. Gastroenterology 1993;105(6):1865–76. [19] Dickson ER, Grambsch PM, Fleming TR, et al. Prognosis in primary biliary cirrhosis: model for decision making. Hepatology 1989;10(1):1–7. [20] Rydning A, Schrumpf E, Abdelnoor M, et al. Factors of prognostic importance in primary biliary cirrhosis. Scand J Gastroenterol 1990;25(2):119–26.