EDITORIAL
How much expansion to be diseased? Toward repeat size and myotonic dystrophy type 2
Benedikt Schoser, MD Tetsuo Ashizawa, MD
Address correspondence and reprint requests to Dr. Tetsuo Ashizawa, Department of Neurology, UTMB, 301 University Blvd., JSA9.128, Galveston, TX 77555-0539
[email protected]
Neurology® 2009;72:484–485
Myotonic dystrophies (DMs) encompass at least two genetically distinct disorders: the classic type 1 (DM1), also known as Steiner disease, and type 2 (DM2), also referred to as proximal myotonic myopathy or Moxley-Ricker disease. Both disorders are characterized by variably expressed multisystem phenotypes with core features of progressive skeletal muscle weakness and degeneration with myotonia. DMs are caused by unstable repeat expansions: in DM1, a CTG repeat expansion in the 3= untranslated region of DMPK in chromosome 19q13.3, and in DM2, a CCTG repeat expansion in intron 1 of ZNF9 in 3q21.3.1,2 In DM1, a reservoir of premutation alleles in the population provides the source of de novo DM1 alleles. The expanded progenitor allele has a propensity to increase the CTG repeat size in successive generations, causing anticipation. The congenital form of DM1, with a large allele and the inability to reproduce offspring, represents the endpoint of anticipation. The stoichiometric balance between de novo mutations and losses of the mutant allele may explain the relatively stable prevalence of DM1. The gender of the transmitting parent is crucial in the instability, as de novo mutations occur with paternal and congenital DM1 with maternal transmission. Expanded alleles also tend to further expand variably in soma and germline throughout the life of the patient, potentially contributing to the progressive phenotype.1,3 In DM2, an uninterrupted variant of the CCTG portion of the repeat tract is elongated and shows unprecedented somatic instability with significant increases in length over time (e.g., 2,000 bp/3 years).4 The extraordinary somatic instability complicates the analysis of genotype–phenotype correlations including those in the effect of the gender of transmitting parents and anticipation. There have
been no reports of premutation alleles for DM2. The copy number of DM2 CCTG repeats is below 30 in normal individuals and up to about 11,000 in patients.2,4 The smallest reported expansion in DM2 had an uninterrupted mosaic (CCTG)75 as estimated by Southern blot.2 Thus, one of the enigmatic questions is which expansion size predicts the phenotypic expression in DM2. Haplotype analyses of DM1 and DM2 families suggested that the respective expansions may have originated from one or few founder mutations.2,5,6 While the original DM1 mutation has been estimated to have occurred around the time of the human migration out of Africa,3 the age of the DM2 expansion mutation on the founding haplotype is estimated at approximately 200 to 540 generations.6 A family of Afghan/Tajik ancestry provided some evidence that the DM2 expansion occurred prior to Aryan migration of the Indo-Europeans in 2000 – 1000 BC.6,7 In this issue of Neurology®, Bachinski et al.8 looked for the existence of premutation alleles in DM2. They found in normal Caucasian chromosomes a unimodal distribution of alleles at 132 bp with rare expanded alleles in the tail of the distribution. In contrast, in African Americans alleles revealed a secondary peak at 174 bp. However, sequence analyses showed that large alleles in African Americans have interruptions with multiple short stretches of uninterrupted CCTG repeats. While some large Caucasian alleles are also interrupted, a long uninterrupted tract of CCTG repeat was found exclusively among Caucasian alleles on the DM2 haplotype. Based on the sequence data, Bachinski et al.9 postulated that the long uninterrupted alleles have derived from interrupted alleles by unequal crossover, a model which is supported by previous observations of CCTG repeats in E coli. Furthermore, long uninterrupted alleles found in Caucasians
See page 490 From Ludwig Maximilians University Munich (B.S.), Friedrich-Baur Institute, Department of Neurology, Germany; and Department of Neurology (T.A.), University of Texas Medical Branch, Galveston. Disclosure: B.S. is a member of the German Muscular Dystrophy Network (MD-NET; 01GM0601) funded by the German Ministry of Education and Research (BMBF, Bonn, Germany). MD-NET is a partner of TREAT-NMD (EC, 6th FP, proposal 036825; www.treat-nmd.eu). Studies on myotonic dystrophies are supported by grants from the Deutsche Gesellschaft fu¨r Muskelkranke, Freiburg, Germany (to B.S.), and T.A. is supported by NIH (NS 041547). 484
Copyright © 2009 by AAN Enterprises, Inc.
have approximately three times higher average mutation load (⬍40%) than large interrupted alleles in African Americans. Large uninterrupted CCTG repeats have a lower thermodynamic stability when compared to the DM1 CTG repeats, which could make them better targets for DNA repair events, thus explaining their expansion-prone behavior. 9 Finally, the unstable, uninterrupted (CCTG)23–33 alleles with lengths of 92–132 bp found by Bachinski et al. are consistent with the threshold for instability at 100 –200 bp of uninterrupted repeats at other repeat loci.8,10 The claim that these unstable alleles with a long uninterrupted CCTG repeat tract are true DM2 premutations must be confirmed by identifying families in which these alleles do expand to the DM2 mutation range in a later generation (i.e., de novo mutation). The DM2 family in which brothers have predominant blood alleles of 55 and 61 repeats reported in this article is interesting, but whether these alleles are results of de novo mutation remains unknown. It also needs to be confirmed that subjects who have a premutation do not have a subtle DM2 phenotype. Furthermore, there are no cases of congenital DM2, indicating that a reservoir of premutations would increase the prevalence of DM2 unless there are other unidentified selection biases against DM2. Thus, the functional and clinical consequences of these premutation alleles remain unclear.8 However, identification of unstable CCTG repeat alleles in the general population is an important step toward understanding the origin of the DM2 mutation
and provides potentially useful information for genetic counseling. REFERENCES 1. Schoser B, Schara U. Myotonic dystrophies type 1 and 2: a summary on current aspects. Semin Pediatr Neurol 2006; 13:71–79. 2. Liquori CL, Ricker K, Moseley ML, et al. Myotonic dystrophy type 2 caused by a CCTG expansion in intron 1 of ZNF9. Science 2001;293:864–867. 3. Ashizawa T, Harper PS. Myotonic dystrophies: an overview. In: Ashizawa T, Wells RD, eds. Genetic Instabilities and Neurological Disorders. 2nd edition. Burlington, MA: Elsevier; 2006:21–35. 4. Day LW, Ranum LP. RNA pathogenesis of the myotonic dystrophies. Neuromuscul Disord 2005;15:5–16. 5. Bachinski LL, Udd B, Meola G, et al. Confirmation of DM2 (CCTG)n expansion mutation in PROMM/PDM patients of different European origins: a single shared haplotype indicates ancestral founder effects. Am J Hum Genet 2003;73:835–848. 6. Liquori CL, Ikeda Y, Weatherspoon M, et al. Myotonic dystrophy type 2: human founder haplotype and evolutionary conservation of the repeat tract Am J Hum Genet 2003;73:849–862. 7. Schoser BGH, Kress W, Walter MC, et al. Homozygosity for CCTG mutation in myotonic dystrophy type 2. Brain 2004;127:1868–1877. 8. Bachinski LL, Czernuszewicz T, Ramagli LS, et al. Premutation allele pool in myotonic dystrophy type 2. Neurology 2009;72:490–497. 9. Dere R, Wells RD. DM2 CCTG*CAGG repeats are crossover hotspots that are more prone to expansions than the DM1 CTG*CAG repeats in Escherichia coli. J Mol Biol 2006;360:21–36. 10. Gatchel JR, Zoghbi HY. Diseases of unstable repeat expansion: mechanisms and common principles. Nat Rev Genet 2005;6:743–755.
Neurology 72
February 10, 2009
485
This week in Neurology® Highlights of the February 17 issue
Disability in optic neuritis correlates with diffusion tensor-derived directional diffusivities
Impact of cardiac complications on outcome after aneurysmal subarachnoid hemorrhage: A meta-analysis Patients with subarachnoid hemorrhage who suffer from cardiac complications may require different treatment as DCI and poor outcome are more prevalent in this group. This paper describes the prevalence and prognostic significance of cardiac complications after aneurysmal subarachnoid hemorrhage.
For patients with MS, clinicians still search for that imaging sequence that provides clinically relevant information. The authors report that quantitative MR diffusion tensor imaging in optic neuritis holds promise, with high correlations for clinical testing and a potential role for predicting outcome. See p. 589; Editorial, p. 584
Sample sizes for brain atrophy outcomes in trials for secondary progressive multiple sclerosis This paper identifies brain atrophy measures that are feasible for use as outcomes in clinical trials of neuroprotective therapies in patients with secondary progressive multiple sclerosis. It
See p. 635
draws attention to a number of relevant trial design issues. See p. 595; Editorial, p. 586
Demyelinating events in early multiple sclerosis have inherent severity and recovery
VIEWS & REVIEWS
Teaching the next generation of neurologists Patients with more severe early multiple sclerosis relapses or worse recovery from them tend to experience similar features in subsequent relapses. This paper demonstrates that early relapse severity and recovery in multiple sclerosis are, to some extent, predictable. See p. 602
B-type natriuretic peptide and cardiovalvulopathy in
All neurologists have a stake in training future neurologists. Improving education will ultimately benefit patient care. Overcoming challenges to neurology training will entail rewarding clinician-educators and incorporating trainees into curriculum redesign and use of new technologies. See p. 657
Parkinson disease with dopamine agonist Plasma B-type natriuretic peptide is expected to be a simple screening test for development of valvulopathy and myocardium damage in patients with Parkinson disease treated with ergot derivative dopamine agonists. See p. 621
Intracranial arterial wall imaging using high resolution 3-tesla contrast-enhanced MRI This paper investigates the feasibility of directly imaging intracranial blood vessel wall pathology using conventional pulse sequences on a clinical 3T MRI system. It demonstrates that the major vascular pathologies could be discriminated, thus
CLINICAL IMPLICATIONS OF NEUROSCIENCE RESEARCH
Potassium channels: Brief overview and implications in epilepsy Potassium (K⫹) channels include several families that are critical in regulation of neuronal excitability, axonal conduction, and neurotransmitter release. Over the past several years there has been increasing interest in developing drugs that modulate the activity of these channels, and some of them show promise for treatment of epilepsy. See p. 664
promising to increase diagnostic specificity compared to lumenographic techniques. See p. 627
Podcasts can be accessed at www.neurology.org
Copyright © 2009 by AAN Enterprises, Inc.
583
EDITORIAL
Picturing injury and recovery with diffusion tensor imaging The eyes have it
Robert A. Bermel, MD Robert J. Fox, MD
Address correspondence and reprint requests to Dr. Robert J. Fox, Mellen Center for MS, Cleveland Clinic, 9500 Euclid Ave., U-10, Cleveland, OH 44195
[email protected]
Neurology® 2009;72:584–585
Imaging plays an essential role in multiple sclerosis (MS) patient care and research, as lesions on T2 and post-gadolinium T1-weighted images identify areas of focal inflammation. However, lesion measures correlate weakly with overall disability, due to shortcomings in clinical scales and pathologic nonspecificity of lesions. Destructive pathology can be clinically silent in early disease, while progressive disability is usually seen without new lesions in late disease. New imaging metrics are needed to test novel experimental therapies which target tissue restoration and prevent degeneration. Brain atrophy is an early and clinically relevant feature of MS, but is nonspecific and insufficiently dynamic to guide short-term treatment decisions.1 Several advanced imaging modalities have emerged, including magnetization transfer imaging, diffusion tensor imaging (DTI), proton spectroscopy, and lesion evolution. To date, however, there has been little clinical application of these techniques. Are they clinically relevant? Do they predict future function? In this issue of Neurology®, Naismith et al.2 provide a preliminary clinical study of DTI applied to a specific functional pathway. DTI quantifies the threedimensional diffusion of water molecules within each imaging voxel. Summary measures of diffusion (mean diffusivity and fractional anisotropy) are abnormal in lesional and normal-appearing tissue in patients with MS.3 Diffusivity, directionally restricted by organized fiber bundles, is more precisely represented by its component eigenvectors (figure). Studies using animal models of CNS inflammation have found that axial diffusivity (兩兩) and radial diffusivity (⬜) reflect the integrity of axons and myelin, respectively, suggesting an improved pathologic specificity of DTI over conventional imaging.4,5 The functional relevance of DTI has been preliminarily studied in the ocular motility pathway.6 Naismith et al. evaluated the functional relevance of DTI in the anterior visual system. By using a pathwayspecific measure of disease activity, they minimized the
problems of integrating disparate neurologic functions into a single measure of neurologic injury. The visual system is a discrete model system with well-defined clinical outcomes.7 The authors assessed the predictive capacity of DTI from 12 optic nerves with acute optic neuritis from 10 patients. Markedly decreased 兩兩 was observed in the acute period, and this change predicted later visual outcome as measured by contrast sensitivity. Notably, in the acute phase, DTI was better than clinical measures in predicting visual outcome at 3 months. DTI was also performed on 28 patients with remote optic neuritis. DTI measures showed moderate to strong correlations with all four tests of visual outcome: visual acuity, contrast sensitivity, retinal nerve fiber layer thickness measured by optical coherence tomography, and P100 latency on visual evoked potentials. The strongest correlations were observed with ⬜. This study shows that axial diffusivity measured in the acute phase provides important prognostic information, and that radial diffusivity in the remote phase correlates best with coincident functional, structural, and physiologic tests of vision. Measurement of the component DTI eigenvectors was an essential step, adding predictive value not provided by clinical and summary DTI measures. This study provides criterion and predictive validity supporting DTI as a relevant imaging measure of tissue integrity. It is also a reminder that the quantification of DTI metrics is a technically demanding task. A previous study of optic nerve DTI (using earlier methodology) failed to identify significant correlations between DTI and clinical measures.8 Naismith et al. utilized custommade hardware and complex filtering to enhance sensitivity, factors which likely limit this specific method to academic research centers, although basic DTI sequences are available on most MRI scanners. Also, focusing solely on the optic nerve potentially ignores brain regions which, though remote to the immediate lesion, may have effects on clinical outcome.9 A fundamental question remains: What are the factors which determine the degree of recovery after an
See page 589 e-Pub ahead of print on December 10, 2008, at www.neurology.org. From Mellen Center for MS Treatment and Research, Neurological Institute (R.A.B., R.J.F.), and Cleveland Clinic Lerner College of Medicine (R.J.F.), Cleveland Clinic, Cleveland, OH. Disclosure: The authors report no disclosures. 584
Copyright © 2009 by AAN Enterprises, Inc.
Figure
Measuring diffusion tensor vectors in healthy and diseased tissue
(A) Artist’s depiction of diffusion tensor imaging (DTI) in a healthy fiber tract. Water diffusion is characterized as an ellipsoid, with three eigenvectors representing each of the individual directional diffusivities. 1 is the longest diffusivity eigenvector, called axial diffusivity (兩兩), and represents diffusion parallel to the longitudinal axis of the fiber tract. 2 and 3 represent diffusion across the fiber tract. They typically are similar in magnitude, so are averaged together and called radial, or transverse diffusivity (⬜). The degree to which water diffuses more in the axial than in the radial direction is expressed as fractional anisotropy (FA), which is roughly the ratio of longitudinal and radial diffusivities. (B) In diseased tissue marked by demyelination, partial remyelination, and transection of axons (green), overall diffusion (the size of the ellipsoid) is increased. In addition, the geometry is altered, whereby the ellipsoid becomes less elongated (lower FA). (Fox RJ. Multiple sclerosis: conventional and diffusion tensor imaging. Semin Neurol 2008;28:462. Reprinted with permission.)
inflammatory demyelinating event, and can treatments influence recovery? Patient characteristics such as age, genetic and pathologic heterogeneity, and other medical conditions may influence recovery. As we bring candidate therapies aimed at tissue protection and recovery into clinical trials, pathway-specific DTI provides one model through which we might effectively test new strategies and predict clinical outcomes.
4.
REFERENCES 1. Bermel RA, Bakshi R. The measurement and clinical relevance of brain atrophy in multiple sclerosis. Lancet Neurol 2006;5:158–170. 2. Naismith RT, Xu J, Tutlam NT, et al. Disability in optic neuritis correlates with diffusion tensor– derived directional diffusivities. Neurology 2009;72:589–594. 3. Filippi M, Cercignani M, Inglese M, Horsfield MA, Comi G. Diffusion tensor magnetic resonance imaging in multiple sclerosis. Neurology 2001;56:304–311.
7.
5.
6.
8.
9.
Song SK, Yoshino J, Le TQ, et al. Demyelination increases radial diffusivity in corpus callosum of mouse brain. Neuroimage 2005;26:132–140. Song SK, Sun SW, Ramsbottom MJ, Chang C, Russell J, Cross AH. Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. Neuroimage 2002;17:1429–1436. Fox RJ, McColl RJ, Lee JC, Frohman TC, Sakaie K, Frohman E. A preliminary validation study of diffusion tensor imaging as a measure of functional brain injury. Arch Neurol 2008;65:1179–1184. Frohman EM, Costello F, Stuve O, et al. Modeling axonal degeneration within the anterior visual system: implications for demonstrating neuroprotection in multiple sclerosis. Arch Neurol 2008;65:26–35. Trip SA, Wheeler-Kingshott C, Jones SJ, et al. Optic nerve diffusion tensor imaging in optic neuritis. Neuroimage 2006;30:498–505. Wu GF, Schwartz ED, Lei T, et al. Relation of vision to global and regional brain MRI in multiple sclerosis. Neurology 2007;69:2128–2135.
Neurology 72
February 17, 2009
585
EDITORIAL
Brain atrophy as an outcome measure for multiple sclerosis clinical trials A “no-brainer”?
Richard A. Rudick, MD Elizabeth Fisher, PhD
Address correspondence and reprint requests to Dr. Richard A. Rudick, Mellen Center, Area JJ3, Cleveland Clinic, Cleveland, OH 44195
[email protected]
Neurology® 2009;72:586–587
In recent years, it has become apparent that axonal degeneration and axonal loss are major determinants of neurologic disability in multiple sclerosis (MS).1 Axonal transection is observed in foci of brain inflammation, even in patients with very short disease durations.2 Later in the disease, axonal degeneration seems to proceed in the absence of detectable inflammation, and gray matter loss accelerates.3 During the later stages of MS, anti-inflammatory therapies do not significantly slow neurologic disability progression. For these reasons, attention has turned to therapeutic strategies that might protect neurons and axons, and thereby slow or prevent disability progression. Optimal clinical trial designs for this purpose have not been fully developed, however. Outcome measures specific for neurodegeneration are needed, as are trial designs that incorporate those measures. MRI-based measures of brain atrophy have been applied increasingly to characterize MS, to study the disease process, and to test the effectiveness of interventions.4 Accelerated rates of brain atrophy are observed from early in the disease, correlate with current and future clinical disability, and decelerate with anti-inflammatory therapy in patients with relapsing-remitting MS. Can MRI measures of brain atrophy be used successfully for early stage drug testing of putative neuroprotection drugs in patients with progressive MS? In this issue of Neurology®, Altmann et al.5 studied three image analysis techniques used to measure change in brain volumes. Serial MRI scans in 43 placebo patients from the European trial of interferon beta-1b in secondary progressive MS (SPMS) were evaluated.6 The longitudinal data generated by the three analysis methods were used to calculate sample sizes for future clinical trials; the suitability of each measure was judged by comparing the required sample sizes necessary to show therapeutic efficacy with reasonable statistical power. The results suggest that
phase II clinical trials targeting brain atrophy in SPMS would be feasible based on modest sample sizes using trial durations of 1 or 2 years, and further suggest that the SIENA method requires the fewest subjects. SIENA coregisters serial MRI scans and directly determines volume differences by measuring changes along the brain–CSF interface.7 Using SIENA, the authors found that 32 subjects per arm would provide an 80% power to detect a 50% treatment effect over 2 years, and suggested that SIENA or similar methods would be feasible for placebo-controlled trials in SPMS. This is an important result, since continuous decline in brain volume almost certainly reflects myelin and axon loss in MS brain. At present, MRI-based brain atrophy measures are suitable for multicenter trials, and the data provided by Altmann et al. suggest that required sample sizes will not be prohibitive. These results should be viewed as an important step toward effective clinical trial designs for trials of neuroprotective agents. Much work needs to be done, and there are caveats and pitfalls. First, the study population consisted of patients with early SPMS, who continued to experience relapses. The results may not be appropriately extrapolated to patients with more advanced, non-relapsing disease. Second, several issues will need to be considered to optimally design clinical trials using brain atrophy measures. The time course of brain atrophy following therapeutic intervention is only partly defined. Ongoing pathologic processes will drive accelerated axonal degeneration for some time after the start of effective therapy. Therefore, atrophy occurring during the initial stages of a clinical trial may reflect pathologic processes operative before treatment. If the intervention has anti-inflammatory effects, initially accelerated brain volume loss may result from resolution of tissue edema.8,9 Both of these effects—resolution of brain edema and ongoing axonal degeneration from
See page 595 From the Mellen Center for Multiple Sclerosis Treatment and Research (R.A.R., E.F.) and the Department of Biomedical Engineering (E.F.), Cleveland Clinic Foundation, OH. Disclosure: Richard A. Rudick, MD, conducts research on interferon and multiple sclerosis funded by the NIH (NINDS and NCRR) and National MS Society, and accepted consulting fees from Biogen Idec, Millennium Pharmaceuticals, Wyeth, and Novartis. Elizabeth Fisher, PhD, conducts research on multiple sclerosis and magnetic resonance imaging supported by the NIH (NINDS), and accepted consulting fees from Biogen Idec, Millennium Pharmaceuticals, and Wyeth. 586
Copyright © 2009 by AAN Enterprises, Inc.
prior injury—would mask the neuroprotective effects of intervention. This may result in the need for longer trial durations to establish stable baselines. Finally, no perfect technique exists, and there are various biologic and technical factors that may have a significant impact on atrophy measurements, such as hydration level, changes in lesions, the necessity for manual corrections, and scanner upgrades. Regional atrophy measures, such as gray matter fraction or structure-specific measures, may prove to be more informative than global measures of whole brain volume, particularly in SPMS. The authors appropriately comment on some of these issues in their article. Much work remains to create designs for neuroprotection trials in MS. Altmann et al. provide important information on an attractive outcome measure for these trials— brain atrophy. If used properly, measures of brain atrophy could lead to a new generation of MS clinical trials. However, at present, the use of MRI-based measures of brain atrophy for MS clinical trials is clearly not a “no-brainer.”
2.
3.
4.
5.
6.
7.
8.
9. REFERENCES 1. Bjartmar C, Kidd G, Mork S, Rudick R, Trapp BD. Neurological disability correlates with spinal cord axonal loss
and reduced N-acetyl aspartate in chronic multiple sclerosis patients. Ann Neurol 2000;48:893–901. Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mork S, Bo L. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998;338:278–285. Fisher E, Lee JC, Nakamura K, Rudick RA. Gray matter atrophy in multiple sclerosis: a longitudinal study. Ann Neurol 2008;64:255–265. Bermel RA, Bakshi R. The measurement and clinical relevance of brain atrophy in multiple sclerosis. Lancet Neurol 2006;5:158–170. Altmann DR, Jasperse B, Barkhof F, et al. Sample sizes for brain atrophy outcomes in trials for secondary progressive multiple sclerosis. Neurology 2009;72:595–601. Molyneux PD, Kappos L, Polman C, et al. The effect of interferon beta-1b treatment on MRI measures of cerebral atrophy in secondary progressive multiple sclerosis: European Study Group on Interferon beta-1b in secondary progressive multiple sclerosis. Brain 2000;123:2256–2263. Smith SM, De SN, Jenkinson M, Matthews PM. Normalized accurate measurement of longitudinal brain change. J Comput Assist Tomogr 2001;25:466–475. Rudick RA, Fisher E, Lee JC, Simon J, Jacobs L. Use of the brain parenchymal fraction to measure whole brain atrophy in relapsing-remitting MS: Multiple Sclerosis Collaborative Research Group. Neurology 1999;53:1698–1704. Hardmeier M, Wagenpfeil S, Freitag P, et al. Atrophy is detectable within a 3-month period in untreated patients with active relapsing remitting multiple sclerosis. Arch Neurol 2003;60:1736–1739.
Neurology 72
February 17, 2009
587
IN MEMORIAM
Wayne Alfred Hening, MD, PhD (1945–2008)
Richard P. Allen, PhD Arthur Walters, MD Sudhansu Chokroverty, MD, FRCP
Wayne Alfred Hening, MD, PhD
Wayne Alfred Hening, MD, PhD, whose storied life encompassed literary comments, published poetry, and most of all stellar published contributions to restless legs syndrome (RLS) and movement disorders, died September 15, 2008, in New York University (NYU) hospital of rapidly progressive pulmonary fibrosis. He was 63. Dr. Hening demonstrated a creative genius of seemingly unlimited knowledge matched by widely diverse interests. He was a member of that generation of gifted young men of the American Eastern establishment, educated in boarding school and then Yale, graduating magna cum laude in 1968 with majors in English and Culture and Behavior. He completed MD-PhD studies at NYU, internship at Stanford, neurology residency at ColumbiaPresbyterian Medical Center, PhD in neurobiology (1982), and movement disorders fellowship under Dr. Stanley Fahn. His research mentors included Dr. Claude Ghez and the Nobel laureate Dr. Eric Kandel. Early in his career, he and one of the authors (A.W.) recognized in one patient the peculiar but
588
Copyright © 2009 by AAN Enterprises, Inc.
neglected movement disorder, RLS. His subsequent interest in RLS produced the most productive period of his life, providing seminal contributions developing the field of RLS. He authored or coauthored 92 articles from 1986 to 2007, mostly on movement disorders and especially RLS. He developed a close relationship with the senior author (R.A.) and Dr. Christopher Earley at the Johns Hopkins RLS research program, and conducted there the most extensive family history study of RLS completed to date. He was one of the most respected experts in the modern field of RLS and sleep-related movement disorder studies. He leaves behind a large group of students and colleagues influenced and inspired by him. He had recently accepted the chair of the sleep section of the American Academy of Neurology and had announced ambitious plans to enhance the function and activities of this group. He was one of the founding members of the World Association of Sleep Medicine and associate editor of Sleep Medicine. To those who knew him well, Wayne stood larger than life. He expressed a remarkable gusto for life, enjoying food, wine, and all forms of great art. He was never seen without carrying around at least one well-marked-up book. He lived intensely for his 63 years, leaving behind not only published papers, but also wonderful poetry. One recently written poem praises commonality of great art in three very different Spanish artists: Salvador Dali, Antoni Gaudi, and the great chef Fernando Adria´. He saw depths of artistry and meaning in all aspects of life from hard science and history to the sensual pleasures of art and food. That poem reflects both the wide span of his immersion in life and in how many ways he informed and encouraged us to look deeper at the meaning in our own lives. In that poem, Wayne wrote: What is seen is only the beginning of what can be seen. Look deeper.
ARTICLES
Disability in optic neuritis correlates with diffusion tensor-derived directional diffusivities R.T. Naismith, MD* J. Xu, PhD* N.T. Tutlam, MPH A. Snyder, MD, PhD T. Benzinger, MD, PhD J. Shimony, MD, PhD J. Shepherd, MD K. Trinkaus, PhD A.H. Cross, MD* S.-K. Song, PhD*
Address correspondence and reprint requests to Dr. Robert T. Naismith, Neurology, Box 8111, 660 S. Euclid Ave., St. Louis, MO 63110
[email protected]
ABSTRACT
Objective: To determine the potential of directional diffusivities from diffusion tensor imaging (DTI) to predict clinical outcome of optic neuritis (ON), and correlate with vision, optical coherence tomography (OCT), and visual evoked potentials (VEP).
Methods: Twelve cases of acute and isolated ON were imaged within 30 days of onset and followed prospectively. Twenty-eight subjects with a remote clinical history of ON were studied cross-sectionally. Twelve healthy controls were imaged for comparison. DTI data were acquired at 3T with a surface coil and 1.3 ⫻ 1.3 ⫻ 1.3 mm3 isotropic voxels.
Results: Normal DTI parameters (mean ⫾ SD, m2/ms) were axial diffusivity ⫽ 1.66 ⫾ 0.18, radial diffusivity ⫽ 0.81 ⫾ 0.26, apparent diffusion coefficient (ADC) ⫽ 1.09 ⫾ 0.21, and fractional anisotropy (FA) ⫽ 0.43 ⫾ 0.15. Axial diffusivity decreased up to 2.5 SD in acute ON. The decrease in axial diffusivity at onset correlated with visual contrast sensitivity 1 month (r ⫽ 0.59) and 3 months later (r ⫽ 0.65). In three subjects followed from the acute through the remote stage, radial diffusivity subsequently increased to ⬎2.5 SD above normal, as did axial diffusivity and ADC. In remote ON, radial diffusivity correlated with OCT (r ⫽ 0.81), contrast sensitivity (r ⫽ 0.68), visual acuity (r ⫽ 0.56), and VEP (r ⫽ 0.54).
Conclusion: In acute and isolated demyelination, axial diffusivity merits further investigation as a predictor of future clinical outcome. Diffusion parameters are dynamic in acute and isolated optic neuritis, with an initial acute decrease in axial diffusivity. In remote disease, radial diffusivity correlates with functional, structural, and physiologic tests of vision. Neurology® 2009;72:589–594 GLOSSARY ADC ⫽ apparent diffusion coefficient; CI ⫽ confidence interval; CS ⫽ contrast sensitivity; DTI ⫽ diffusion tensor imaging; EAE ⫽ experimental autoimmune encephalitis; FA ⫽ fractional anisotropy; MD ⫽ mean diffusivity; MS ⫽ multiple sclerosis; NAWM ⫽ normal-appearing white matter; OCT ⫽ optical coherence tomography; ON ⫽ optic neuritis; RA ⫽ relative anisotropy; rFOV ⫽ reduced field of view; RNFL ⫽ retinal nerve fiber layer; ROI ⫽ region of interest; SD ⫽ standard deviation; SNR ⫽ signal-to-noise ratio; VA ⫽ visual acuity; VEP ⫽ visual evoked potentials.
Supplemental data at www.neurology.org
Pathologic heterogeneity is postulated as a key contributor to the MRI paradox in multiple sclerosis (MS). An imaging modality that better correlates with clinical disability and prognosis is needed. We hypothesize that directional diffusivity measurements from diffusion tensor imaging (DTI) can detect structural changes within white-matter tracts, and differentiate axonal injury from demyelination. DTI tissue measurements are based upon Brownian motion of water, influenced by tissue restrictions from cellular structures. Studies suggest that mean diffusivity (MD, also known as apparent diffusion coefficient [ADC]) and fractional or relative anisotropy (FA or RA) can detect both MS lesions and changes in normal-appearing white matter (NAWM).1 Unfortu-
Editorial, page 584 e-Pub ahead of print on December 10, 2008, at www.neurology.org. *These authors contributed equally. From the Departments of Neurology (R.T.N., J.X., N.T.T., A.S., A.H.C.), Radiology (A.S., T.B., J. Shimony, S.-K.S.), Ophthalmology (J. Shepherd), and Biostatistics (K.T.), and Hope Center for Neurological Disorders (A.H.C., S.-K.S.), Washington University, Saint Louis, MO. Funding included K23NS052430 (R.T.N.), K12RR02324902 (R.T.N.), K24 RR017100 (A.H.C.), P30 NS048056 (A.Z.S.), and UL1 RR024992 from the NIH, and CA1012 (A.H.C., S.K.S.), RG 3670 (S.K.S.), and FG1782A1 (J.X.) from the National MS Society USA. A.H.C. was supported in part by the Manny and Rosalyn Rosenthal-Dr. John L. Trotter Chair in Neuroimmunology. Disclosure: The authors report no disclosures. Copyright © 2009 by AAN Enterprises, Inc.
589
nately, summary parameters derived from DTI have not demonstrated sufficient specificity to define the underlying pathology of white matter injury. To improve DTI specificity, an analytical approach has been proposed by this group to assess the extent of axonal damage, demyelination, or both. The three eigenvalues or diffusivities (1, 2, and 3) are scalar indices that describe water diffusion through tissue microstructures in a local frame of reference. Within white matter tracts, the largest eigenvalue within a voxel, 1, is assumed to represent the water diffusivity parallel to the majority of axonal fibers. This is designated 兩兩, the “axial diffusivity.” The averaged diffusivity of 2 and 3 represents diffusion perpendicular to the axonal fibers, denoted as ⬜, the “radial diffusivity.” Previous work by this group using animal models of acute CNS injury has demonstrated that directional diffusivities within white matter tracts correlate with axon and myelin pathologies.2-9 Building upon the application in mouse models of axial and radial diffusivities as pathologic surrogates of axonal and myelin damage, respectively, the present studies undertook translation to humans. The human optic nerve was chosen as a simple and welldefined white matter tract emanating from single-ordered neurons. It is commonly affected in demyelinating disorders and has well-defined outcome measures. METHODS All subjects provided informed consent, after approval by the local Human Research Protection Office/Institutional Review Board.
Normal subjects. Twelve subjects with no symptoms or signs of neurologic or ocular pathology were recruited (6 men and 6 women, mean age 37 with range 21– 49). Control subjects had a normal funduscopic examination, visual acuity (VA) ⱖ20/25, and Pelli-Robson contrast sensitivity (CS) ⱖ1.65. Two individuals were imaged twice, and another two were imaged three times on different days to determine intersubject variability.
Subjects with ON. Twelve cases of acute optic neuritis (ON) in 10 subjects with no previous neurologic events were imaged within 30 days of onset. One subject had bilateral involvement, and another sequential involvement in the fellow eye, which was initially unaffected based upon all vision tests and VEP. Twelve nerves were prospectively followed for 3 months, with three followed for over 1 year. Nine of the 12 episodes received IV glucocorticoids, including all with VA worse than 20/40 (n ⫽ 7). For remote ON, 28 subjects with at least one clinical episode of ON at least 1 year prior were included. Visual measurements 590
Neurology 72
February 17, 2009
included Sloan 5% CS chart at 3 m in an illuminated cabinet (Precision Vision, IL) and corrected VA by 20 ft Snellen wall chart. Visual evoked potentials (VEP) P100 latency (normal mean 98.95 msec, upper limit 112.9 msec) was read blinded. If the waveform was unobtainable due to poor vision, the maximal obtained value of 170 msec was used. Optical coherence tomography (OCT) fast retinal nerve fiber layer (RNFL) thickness was obtained on a Zeiss Stratus OCT III with v4.0 software.
MR protocol. MR data were acquired using a custom fabricated transmitter coil and four-element phased array flexible surface receiver coil on a 3T MR scanner (Allegra, Siemens AG, Erlangen, Germany). A vacuum-molded pillow minimized head movement. Subjects were encouraged to sleep or close their eyes. Sleep was unlikely to cause excessive eye movement since the scan is shorter than the latency to the REM stage. A single shot spin-echo echoplanar imaging diffusion sequence was employed with fat suppression, and reduced field of view (rFOV) technique with twice refocused diffusion weighting.10 Diffusion-weighted images were acquired transaxially (field of view 168 mm ⫻ 84 mm, matrix 128 ⫻ 64, partial Fourier 6/8, echo time 65 msec) with two collated groups of 1.3-mm-thick slices. Each slice group comprised five interleaved slices and was cardiac gated (150 msec delay after the sphygmic wave by pulse oximetry), yielding a repetition time of 4 – 6 seconds. Eight to 12 image sets, each consisting of one with b ⫽ 0 (b0) and 12 diffusion-weighted images on 12 diffusion encoding directions (two sets of six icosahedral directions with opposite gradient polarity) with b ⫽ 600 s/mm2, were acquired for each slice group.11 Total scan time was 40 minutes.
DTI calculation. Each DTI data set was motion corrected using an iterative procedure.12 Diffusion images with excessive movement (ⱖ3 mm translation in either direction) were excluded from averaging. All transforms were rigid body affine and computed by vector gradient measure maximization.13 The b0 volumes of each DTI data set were aligned using intensity correlation maximization. The final motion-corrected result was obtained by algebraically composing all transforms (saved from the iterative procedure), and then averaging all data sets after application of the composed transforms using cubic spline interpolation. The final resampling step output 13 volumes with interpolated resolution of 0.65 ⫻ 0.65 ⫻ 0.65 mm3. The diffusion tensor matrix at each voxel was estimated by linear leastsquares fitting of the motion corrected and resampled DTI dataset.14
Region of interest analysis. The region of interest (ROI) was selected manually on the b0 image to include 15–20 voxels longitudinally (10 –13 mm in length). To avoid CSF partial volume artifacts, the ROI only included voxels at the nerve center. To avoid the region most prone to movement and confounded by lack of myelination, the ROI started 12–15 voxels (about 15.0 mm) posterior to the retina (figure 1). After ROI identification for controls, voxels having FA greater than 2 SD from the mean were discarded to avoid magnetic field inhomogeneity induced artifacts and incorrect diffusion calculations. Voxels with signal-to-noise ratio (SNR)15 lower than 32 were excluded, resulting in retention of 75% of the voxels. These same ROI criteria derived from the normal subjects were applied to the subjects with ON. Statistical analyses. Linear modeling for the normative data was estimated by taking 1,000 bootstrap samples of one measurement from each of the 12 normal subjects, such that the observations contributing to each mean were independent. In
Figure 1
Region of interest (ROI) selection on the b0 image
Table 2
Diffusion parameters in acute optic neuritis (ON) at onset and after 1 year ON 1
ON 2
ON 3
VA at onset
20/40
Motion
Motion
VA after 1 year
20/20
20/30
20/40
Axial diffusivity
1.66
1.48
1.18†
Radial diffusivity
0.72
0.76
0.67
Apparent diffusion coefficient
1.03
1.02
0.91
Fractional anisotropy
0.46
0.39
0.26
Axial diffusivity
2.28*
2.07*
2.04*
Radial diffusivity
1.44*
1.32*
1.61*
Apparent diffusion coefficient
1.72*
1.57*
1.76*
Fractional anisotropy
0.32
0.31
0.19
DTI at onset
DTI after 1 year
Isotropic voxels permitted visualizing the nerve in the transverse (A) and coronal (B) planes. The nerve center was selected to minimize CSF contamination. Fifteen to 20 voxels posterior to the unmyelinated anterior nerve section allowed adequate sampling for approximately 2.5 cm of the nerve. ROI was coregistered to transverse images of fractional anisotropy (C) and axial diffusion (D).
acute ON, Spearman correlation coefficients described the relationship between the diffusion and clinical parameters at onset, to the clinical parameters 1 and 3 months later. In remote ON, Spearman correlation described the relationship between the diffusion parameters to the clinical visual tests. RESULTS Normal subjects. Data from normal subjects revealed no age or gender effects on the diffusion parameters (table 1). Axial diffusivity (mean 1.66 m2/ms) was twice the magnitude of radial diffusivity (mean 0.81). The mean SNR for each voxel within the ROI was 37.8 ⫾ 9.3 (range 17.8 – 69.7). For voxels with low SNR, axial diffusion was dependent upon the SNR. Below SNR of 32, the Pearson correlation coefficient squared (r2) of voxel SNR vs axial diffusivity was 0.43. Above SNR of 32, r2 ⫽ 0.03 (both p values ⬍0.05). Voxels above the SNR cutoff demonstrated little dependency on other diffusion parameters (MD r2 ⫽ 0.05, FA r2 ⫽ 0.05,
Table 1
Diffusion parameters in optic nerve from healthy volunteers (n ⴝ 12) Mean
95% CI
SD
95% CI
Interscan SD
Intrascan SD
Axial diffusivity
1.66
(1.58, 1.69)
0.18
(0.16, 0.21)
0.12
0.13
Radial diffusivity
0.81
(0.73, 0.84)
0.26
(0.17, 0.28)
0.17
0.10
Apparent diffusion coefficient
1.09
(1.02, 1.12)
0.21
(0.15, 0.24)
0.15
0.08
Fractional anisotropy
0.43
(0.41, 0.46)
0.15
(0.12, 0.17)
0.12
0.12
The inter- and intrascan variability was based on two individuals imaged three times each, and another two individuals each imaged twice. Axial diffusivity, radial diffusivity, and apparent diffusion coefficient are given in m2/ms. Fractional anisotropy is without units. CI ⫽ confidence interval.
Axial, radial, and mean diffusivities (apparent diffusion coefficient) are given in m2/ms. Fractional anisotropy is without units. Pelli-Robson contrast sensitivity is logMAR units. *Value is ⱖ2 SD from the normative mean. VA ⫽ visual acuity; DTI ⫽ diffusion tensor imaging.
radial diffusivity r2 ⫽ 0.05). After the cutoff was applied, diffusion parameters displayed a Gaussian distribution. Only three percent of voxels had FA greater than 2 SD above the mean. The SD of the noise on b0 images was found to be consistent across all studies (mean 15.51, SD 1.83, range 11.40 – 18.50), as a measure of quality control. Acute ON. In 12 cases of ON imaged within 30 days
of clinical onset, FA was the only diffusion parameter correlated with the vision at onset (CS r ⫽ 0.66, p ⫽ 0.02; VA r ⫽ 0.66, p ⫽ 0.02). However, visual recovery 1 or 3 months later did not correlate with the initial alteration in FA. Axial diffusivity was the only diffusion parameter at onset to correlate with a visual function parameter at both 1 (CS r ⫽ ⫺0.59, n ⫽ 12, p ⫽ 0.04) and 3 months later (CS r ⫽ ⫺0.65, n ⫽ 11, p ⫽ 0.03) (figure e-1 on the Neurology® Web site at www.neurology.org). Of note, CS and VA at onset did not correlate with CS and VA 1 or 3 months later, suggesting that axial diffusivity is contributing prognostic information independent of that acquired clinically.16-18 Three cases of acute ON were available after 1 year to determine the evolution of DTI into the remote phase (table 2). Only ON 3 demonstrated a significant change in a diffusion parameter at onset, with axial diffusivity being decreased ⬍2 SD below normal. In the three nerves after 1 year of follow-up, axial and radial diffusivity, along with ADC, were significantly increased 2 SD above normal. ON 3, Neurology 72
February 17, 2009
591
Table 3
Correlation coefficients of diffusion parameters for clinical and physiologic measures in remote optic neuritis
DTI parameter Axial diffusivity
5% Contrast 0.47
Visual acuity
OCT RNFL
⫺0.51
⫺0.59
VEP P100 0.41
Radial diffusivity
0.68
⫺0.56
⫺0.81
0.54
Apparent diffusion coefficient
0.61
⫺0.58
⫺0.76
0.54
⫺0.64
0.49
0.74
⫺0.53
Fractional anisotropy
All subjects (n ⫽ 28) had a clinical episode of ON in at least one eye over 1 year ago. Recovery was variable. Values in the table represent r values using the Spearman test. All correlations were significant at p ⱕ 0.001, except axial diffusivity to VEP was p ⫽ 0.006. DTI ⫽ diffusion tensor imaging; OCT ⫽ optical coherence tomography; RNFL ⫽ retinal nerve fiber layer; VEP ⫽ visual evoked potentials.
with an initial 2.5 SD decrease in axial diffusivity at onset, had reversed after 1 year to become 2 SD increased above the mean. The DTI parameters from the three cases of acute ON followed prospectively over a year were not different from the mean DTI parameters for the 28 subjects with remote ON. Remote ON. The 28 individuals with a prior clinical
history of ON had a median age of 43 years (range 22–59), 11:3 female:male, median disease duration 8 years (range 1–27), median EDSS of 2.0 (range 0 –7), median Snellen VA 20/20 (range 20/13 to no light perception), and median CS 0.3 logMAR (range 0.1–1.0) (figure e-2). Spearman coefficients of all diffusion parameters demonstrated correlation with all clinical parameters (table 3). Of the four diffusion parameters, radial diffusivity had the highest correlations with concurrent visual system tests. Radial diffusivity was highly correlated with OCT (r ⫽ ⫺0.81, p ⬍ 0.001) and 5% CS (r ⫽ 0.68; p ⬍ 0.001), and moderately correlated with VA (r ⫽ ⫺0.56; p ⬍ 0.001) and VEP (r ⫽ 0.54, p ⬍ 0.001). Most episodes of ON have a favorable outcome, sometimes without regard for the severity at onset.17 However, visual impairment in ON remains a disabling problem, adversely affecting quality of life in many. The variability in recovery is likely due to the nature of the inflammatory response, the extent of tissue injury, along with the individual’s capacity to repair. The temporary vision loss in some severe cases may be due to reversible inflammation and edema. Lack of recovery for others may be due to extensive demyelination without adequate repair, along with variable injury to axons. In ON, as well as MS or NMO, a biomarker is needed to determine individual risk of poor recovery at onset of a demyelinating event. Directional diffusivities have potential to provide a window into the pathology of demyelinating disease.2-9 Because it is a single tract with quantifiable function, the optic nerve is an ideal system for invesDISCUSSION
592
Neurology 72
February 17, 2009
tigating directional diffusivities from DTI. Herein, the optic nerve was used to understand the dynamics of diffusion alterations as a prelude to studying more complex systems. In 12 acute episodes of ON, a decrease in axial diffusivity at presentation correlated with worse contrast sensitivity 1 and 3 months later. These time points are notable because recovery from ON begins to plateau 30 – 60 days from onset,17 and therefore recovery at 1–3 months may serve as a surrogate of long-term recovery. Contrast discrimination is a sensitive test that can reveal subtle and important deficits despite recovery of Snellen acuity.18 It is of interest that initial axial diffusivity did not correlate with initial clinical parameters, suggesting that axial diffusivity adds independent prognostic information. Initial Snellen VA and contrast sensitivity did not correlate with these same parameters 1 and 3 months later, suggesting in this cohort that recovery is not always predicted by the initial clinical severity.16-18 Additional subjects and longer follow-up are needed to confirm this finding. In these 12 acute cases, axial diffusivity was the first diffusion parameter to be altered (table e-1). Decline in axial diffusivity in the acute setting therefore has potential as a surrogate marker of tissue destruction and outcome. Radial diffusivity, FA, and ADC were less useful as predictors, becoming abnormal beyond the acute setting. Since tissue injury in demyelinating disease is presumably a dynamic process, one would expect diffusion parameters to also change over time. Thus, the timing of diffusion imaging, as well as the existence of prior injury within that tract, should be considered for proper interpretation of DTI results. In the group of 28 subjects with remote and stable ON, marked alterations in all diffusion parameters of the affected optic nerves were evident. Radial diffusivity, along with FA and ADC, had strong correlations to tests of vision (CS and VA), structure (OCT), and physiology (VEP). The correlation to RNFL thickness by OCT supports the underlying hypothesis, since the RNFL is thought to reflect tissue alterations within the optic nerve. Axial diffusivity was less strongly correlated to vision, structure, and physiology in remote disease. We hypothesize that axial diffusivity in the remote setting may be overwhelmed by increased diffusivity and decreased anisotropy from demyelination, tissue vacuolization, and gliosis.19 Axial diffusivity may be insensitive to axons if they are not tightly packed with myelin. To illustrate the effect of increased total diffusivity upon axial diffusivity, three cases of ON followed from the acute to the remote state demonstrate increased axial diffusivity with increasing ADC
and radial diffusivity over the year. The role of diffusion parameters with disease recurrence needs further investigation to better understand how to translate future DTI to those with established MS. Interpretation of axial diffusivity with acute but recurrent disease may be more challenging if there are preexisting alterations in overall diffusivity. Imaging of the optic nerve is difficult due to small size, mobility, and surrounding tissues. The diameter of a normal optic nerve is 3– 4 mm. Normal retrobulbar optic nerve has a fiber count of 1,141,000 ⫾ 212,000, a 20% range.20 In the present study, highresolution imaging with 1.3 mm3 isotropic voxels minimized CSF contamination and allowed 2–3 transverse slices of the nerve. Isotropic voxels permitted visualization in multiple planes for optimal ROI selection (figure 1). Central voxels were selected, sacrificing possible peripherally positioned plaques in order to exclude voxels bordering CSF. Distortion from high-resolution echoplanar sequences was lowered by a reduced echo train length by rFOV technique. A high SNR is necessary for reliable DTI quantification. This presents a challenge for high-resolution imaging, because signal strength decreases with increasing resolution. In this study, plots comparing diffusion parameters to voxel SNR in normal subjects found an optimal value of SNR to be 30 – 40. A custom-built optic nerve receiver coil and a 3T magnet with multiple averages increased SNR. An inversion pulse for CSF suppression was not used, since this reduces signal from the CNS tissue. Using these methods, the mean optic nerve SNR was 34.6. An SNR of 32 was selected as a criterion for inclusion in quantitative measurements (see Methods). This value is consistent with published theoretical analyses that show increasing measurement accuracy above 30.21-23 This investigational imaging protocol was optimized to obtain quality images. It required 40 minutes using a dedicated surface and transmitter coil. With proper instruction and positioning, only 2 of 38 scans were repeated due to motion. Thus, directional DTI of ON is feasible in the clinical setting, although ideally obtained in a dedicated imaging session. Published animal studies in different models of axonal injury and myelin loss support the concept of utilizing directional diffusivities as markers of pathology.2,3,5,24 This study in acute and isolated ON in humans demonstrates an initial decrease in axial diffusivity (figure 2). This early decrease in axial diffusivity has been demonstrated after corpus callosotomy within the brain.25 This decrease has also been observed in animal imaging, with histologic correlation to decreased axon count. Animal imaging is very useful for following diffusion changes longitudinally and corre-
Figure 2
Axial diffusivity is decreased in acute retrobulbar optic neuritis (ON)
This optic nerve was imaged 11 days after clinical onset of ON (ON 2). Vision was motion perception only, and funduscopic examination was normal. Decreased axial diffusivity was manifested by a warmer color. A moderate decrease in axial diffusivity within the retrobulbar portion of the right nerve was observed. Axial diffusion was 1 standard deviation below the normative range. Although this eye had a severe onset, it made a good recovery.
lating them with histopathology, but animal models have limitations. Experimental autoimmune encephalitis (EAE) is used as a model for MS, but there are important immunopathologic differences. For example, EAE is induced by immunization, whereas the inciting event for MS is unknown. Thus far, in the nonrelapsing C57BL/6 EAE model, a late elevation in ON axial diffusivity following the early decline in axial diffusivity has not been observed in studies up to 90 days.9 Other investigators have reported DWI/DTI studies of human optic nerves.26-29 Previous reports describe elevated mean diffusion in the remote state following ON, the ability to distinguish normal from abnormal optic nerves, and correlations of DTI with results of clinical testing.30,31 The present work expands upon the prior reports, and includes acute DTI in isolated ON with prospective determination of clinical outcome. The present report describes imaging and postprocessing procedures, a high SNR, sampling across a 10 –15 mm longitudinal segment of the optic nerve, and correlation of DTI measures to OCT. Our primary objective was to evaluate the use of the individual tensors of a human CNS tract to estimate the pathology within white matter. Highresolution diffusion imaging of the living human optic nerve can be reliably accomplished while maintaining high SNR. We report an early decrease in axial diffusivity in acute ON, and preliminary data Neurology 72
February 17, 2009
593
that the degree of decline in axial diffusivity may correlate with clinical outcome. DTI and directional diffusivities provide information about function, physiology, and structure of the optic nerve. This clinically important information is not obtained by conventional MRI. The present studies also showed that diffusion parameters are dynamic, and changing between the acute and remote stages. Future studies will determine whether these techniques have practical benefit in the care of patients. Received April 18, 2008. Accepted in final form August 4, 2008. REFERENCES 1. Rovaris M, Gass A, Bammer R, et al. Diffusion MRI in multiple sclerosis. Neurology 2005;65:1526–1532. 2. Sun SW, Liang HF, Trinkaus K, Cross AH, Armstrong RC, Song SK. Noninvasive detection of cuprizone induced axonal damage and demyelination in the mouse corpus callosum. Magn Reson Med 2006;55:302–308. 3. Kim JH, Budde MD, Liang HF, et al. Detecting axon damage in spinal cord from a mouse model of multiple sclerosis. Neurobiol Dis 2006;21:626–632. 4. Song SK, Yoshino J, Le TQ, et al. Demyelination increases radial diffusivity in corpus callosum of mouse brain. NeuroImage 2006;26:132–140. 5. Song SK, Sun SW, Ju WK, Lin SJ, Cross AH, Neufeld AH. Diffusion tensor imaging detects and differentiates axon and myelin degeneration in mouse optic nerve after retinal ischemia. NeuroImage 2003;20:1714–1722. 6. Song SK, Sun SW, Ramsbottom MJ, Chang C, Russell J, Cross AH. Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. NeuroImage 2002;17:1429–1436. 7. Harsan LA, Poulet P, Guignard B, et al. Brain dysmyelination and recovery assessment by noninvasive in vivo diffusion tensor magnetic resonance imaging. J Neurosci Res 2006;83:392–402. 8. Tyszka JM, Readhead C, Bearer EL, Pautler RG, Jacobs RE. Statistical diffusion tensor histology reveals regional dysmyelination effects in the shiverer mouse mutant. NeuroImage 2006;29:1058–1065. 9. Sun SW, Liang HF, Schmidt RE, Cross AH, Song SK. Selective vulnerability of cerebral white matter in a murine model of multiple sclerosis detected using diffusion tensor imaging. Neurobiol Dis 2007;28:30–38. 10. Jeong EK, Kim SE, Guo J, Kholmovski EG, Parker DL. High-resolution DTI with 2D interleaved multislice reduced FOV single-shot diffusion-weighted EPI (2D ss-rFOV-DWEPI). Magn Reson Med 2005;54:1575– 1579. 11. Hasan KM, Parker DL, Alexander AL. Comparison of gradient encoding schemes for diffusion-tensor MRI. J Magn Reson Imaging 2001;13:769–780. 12. Shimony JS, Burton J, Epstein AA, McLaren DG, Sun SW, Snyder AZ. Diffusion tensor imaging reveals white matter reorganization in early blind humans. Cereb Cortex 2006;16:1653–1661. 13. Rowland DJ, Garbow JR, Laforest RR, Snyder AZ. Registration of [18F]FDG microPET and small-animal MRI. Nucl Med Biol 2005;32:567–572. 14. Shimony JS, McKinstry RC, Akbudak E, et al. Quantitative diffusion-tensor anisotropy brain MR imaging: nor594
Neurology 72
February 17, 2009
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
mative human data and anatomic analysis. Radiology 1999;212:770–784. Edelstein WA, Bottomley PA, Pfeifer LM. A signal-tonoise calibration procedure for NMR imaging systems. Med Phys 1984;11:180–185. Beck RW, Ruchman MC, Savino PJ, Schatz NJ. Contrast sensitivity measurements in acute and resolved optic neuritis. Br J Ophthalmol 1984;68:756–759. Beck RW, Cleary PA, Anderson MM, et al. A randomized controlled trial of corticosteroids in the treatment of acute optic neuritis. The Optic Neuritis Study Group. N Engl J Med 1992;326:581–588. Trobe JD, Beck RW, Moke PS, Cleary PA. Contrast sensitivity and other vision tests in the optic neuritis treatment trial. Am J Ophthalmol 1996;121:547–53. Schierer K, Wheeler-Kingshott CAM, Boulby PA, et al. Diffusion tensor imaging of post mortem multiple sclerosis brain. NeuroImage 2007;35:467–477. Jonas JB, Schmidt AM, Mu¨ller-Bergh JA, Naumann GOH. Optic nerve fiber count and diameter of the retrobulbar optic nerve in normal and glaucomatous eyes. Graefes Arch Clin Exp Ophthalmol 1995;233:421–424. Basetin ME, Armitage PA, Marshall I. A theoretical study of the effect of experimental noise on the measurement of anisotropy in diffusion imaging. Magn Reson Imaging 1998;16:773–785. Pierpaoli C, Basser PJ. Toward a quantitative assessment of diffusion anisotropy. Magn Reson Med 1996;36:893– 906. Anderson AW. Theoretical analysis of the effects of noise on diffusion tensor imaging. Magn Reson Med 2001;46: 1174–1188. Deboy CA, Zhang J, Dike S, et al. High resolution diffusion tensor imaging of axonal damage in focal inflammatory and demyelinating lesions in rat spinal cord. Brain 2007;130:2199–2210. Concha Gross DW, Wheatley BM, Beaulieu C. Diffusion tensor imaging of time-dependent axonal and myelin degradation after corpus callosotomy in epilepsy patients. NeuroImage 2006;32:1090–1099. Iwasawa T, Matoba H, Ogi A, et al. Diffusion weighted imaging of the human optic nerve: a new approach to evaluate optic neuritis in multiple sclerosis. Magn Reson Med 1997;38:484–491. Vinogradov E, Degenhardt A, Smith D, et al. Highresolution anatomic, diffusion tensor, and magnetization transfer magnetic resonance imaging of the optic chiasm at 3T. J Magn Reson Imaging 2005;22:302–306. Chabert S, Molko N, Cointepas Y, Le Roux P, Le Bihan D. Diffusion tensor imaging of the human optic nerve using a non-CPMG fast spin echo sequence. J Magn Reson Imaging 2005;22:307–310. Wheeler-Kingshott C, Trip SA, Symms MR, Parker GJM, Barker GJ, Miller DH. In vivo diffusion tensor imaging of the human optic nerve: pilot study in normal controls. Magn Reson Med 2006;56:446–451. Trip SA, Wheeler-Kingshott C, Jones SJ, et al. Optic nerve diffusion tensor imaging in optic neuritis. NeuroImage 2006;30:498–505. Hickman SJ, Wheeler-Kingshott CA, Jones SJ, et al. Optic nerve diffusion measurement from diffusion-weighted imaging in optic neuritis. Am J Neuroradiol 2005;26:951– 956.
Sample sizes for brain atrophy outcomes in trials for secondary progressive multiple sclerosis
D.R. Altmann, DPhil B. Jasperse, MD F. Barkhof, PhD K. Beckmann, MSc M. Filippi, MD L.D. Kappos, MD P. Molyneux, MD C.H. Polman, PhD C. Pozzilli, MD A.J. Thompson, FRCP K. Wagner, MD T.A. Yousry, FRCR D.H. Miller, FRCP
Address correspondence and reprint requests to Dr. Dan R. Altmann, Medical Statistics Unit, London School of Hygiene and Tropical Medicine, Keppel Street, London, WC1E 7HT, UK
[email protected]
ABSTRACT
Background: Progressive brain atrophy in multiple sclerosis (MS) may reflect neuroaxonal and myelin loss and MRI measures of brain tissue loss are used as outcome measures in MS treatment trials. This study investigated sample sizes required to demonstrate reduction of brain atrophy using three outcome measures in a parallel group, placebo-controlled trial for secondary progressive MS (SPMS).
Methods: Data were taken from a cohort of 43 patients with SPMS who had been followed up with 6-monthly T1-weighted MRI for up to 3 years within the placebo arm of a therapeutic trial. Central cerebral volumes (CCVs) were measured using a semiautomated segmentation approach, and brain volume normalized for skull size (NBV) was measured using automated segmentation (SIENAX). Change in CCV and NBV was measured by subtraction of baseline from serial CCV and SIENAX images; in addition, percentage brain volume change relative to baseline was measured directly using a registration-based method (SIENA). Sample sizes for given treatment effects and power were calculated for standard analyses using parameters estimated from the sample.
Results: For a 2-year trial duration, minimum sample sizes per arm required to detect a 50% treatment effect at 80% power were 32 for SIENA, 69 for CCV, and 273 for SIENAX. Two-year minimum sample sizes were smaller than 1-year by 71% for SIENAX, 55% for CCV, and 44% for SIENA. Conclusion: SIENA and central cerebral volume are feasible outcome measures for inclusion in placebo-controlled trials in secondary progressive multiple sclerosis. Neurology® 2009;72:595–601 GLOSSARY ANCOVA ⫽ analysis of covariance; CCV ⫽ central cerebral volume; FSL ⫽ FMRIB Software Library; MNI ⫽ Montreal Neurological Institute; MS ⫽ multiple sclerosis; NBV ⫽ normalized brain volume; PBVC ⫽ percent brain volume change; RRMS ⫽ relapsing–remitting multiple sclerosis; SPMS ⫽ secondary progressive multiple sclerosis.
Supplemental data at www.neurology.org
Definitive clinical trials of potential new disease-modifying agents in multiple sclerosis (MS) often evaluate disability as the primary outcome measure. Because MS is characterized by a variable but generally slow clinical evolution, controlled studies with disability endpoints require large numbers of patients (several hundreds) to be studied over several years. Accordingly, there is considerable interest in developing surrogate laboratory markers of disease progression that, if more sensitive than disability, would enable trials to be performed more quickly and with fewer patients. Irreversible and progressive disability in MS is likely due to neuroaxonal loss and demyelination, which occur in focal white matter lesions1 and also in normal-appearing white2,3 and gray
Editorial, page 586 e-Pub ahead of print on November 12, 2008, at www.neurology.org. From the Medical Statistics Unit (D.R.A.), London School of Hygiene and Tropical Medicine, UK; Nuclear Magnetic Resonance Research Unit (D.R.A., D.H.M.), Department of Neuroinflammation, Institute of Neurology, University College London, UK; Multiple Sclerosis Centre Amsterdam and Image Analysis Centre (B.J., F.B., C.H.P.), VU University Medical Centre, The Netherlands; Bayer Schering Pharma AG (K.B., K.W.), Berlin, Germany; Neuroimaging Research Unit (M.F.), Department of Neurology, Scientific Institute and University Ospedale San Raffaele, Milan, Italy; Neurology and Department of Biomedicine (L.D.K.), University Hospital, Basel, Switzerland; Addenbrookes Hospital (P.M.), Cambridge, UK; Department of Neurological Science (C.P.), University of Rome “La Sapienza,” Italy; and Department of Brain Repair and Rehabilitation (A.J.T., T.A.Y.), Institute of Neurology, University College London, UK. The Nuclear Magnetic Resonance Research Unit is partly supported by The Multiple Sclerosis Society of Great Britain and Northern Ireland. The Multiple Sclerosis Centre Amsterdam is supported by the Dutch Foundation for MS Research (grant 05-538c). Disclosure: Bayer Schering Pharma AG supported the data collection for this study. F.B., M.F., P.M., C.H.P., and D.H.M. have received honoraria from Bayer Schering Pharma AG (less than $10,000). K.W. and K.B. are current employees of Bayer Schering Pharma AG. Copyright © 2009 by AAN Enterprises, Inc.
595
matter.4 MRI-measured brain atrophy has been proposed as a marker of progressive axonal and myelin loss,5 and it is now often acquired as an outcome measure in phase III trials.6-8 If brain atrophy is to be used as a reliable outcome measure in clinical trials, power calculations are required not only to determine the sample sizes needed to show therapeutic efficacy, but also to help identify the most suitable atrophy outcome measures, which is our primary aim here. In this report, based on data acquired in a multicenter sample of placebo-treated subjects with secondary progressive MS (SPMS), we calculate and compare sample sizes required in a parallelgroup, placebo-controlled trial for SPMS subjects, using three brain atrophy outcome measures: a semiautomated measure of a regional (central) cerebral volume that has previously been used in MS cohorts9-11 and two whole-brain automated measures—SIENA and SIENAX—also used extensively.7,12,13 Two secondary aims are to contrast the sample sizes required for different trial durations and analyses and to examine the relationships between the three atrophy outcomes. METHODS Patients. A substudy10 from five centers in a placebo-controlled trial of interferon beta-1b in SPMS acquired 6-monthly T1-weighted brain MRI over 3 years. There were 46 placebo-treated patients from the five centers (20 women, 26 men), 43 of which provided usable data. The mean age at entry was 40.9 years (SD 7.9 years), the mean disease duration was 13.4 years (SD 7.5 years), the mean time since evidence of progression was 3.8 years (SD 3.4 years), and the mean Expanded Disability Status Scale score was 5.2 (SD 1.1, range 3– 6.5). These patients underwent 6-monthly T1-weighted spin echo MRI (repetition time 500 –700 msec, echo time 5–25 msec, 256 ⫻ 256 matrix, 24-cm field of view) for 3 years with 5-mm-thick contiguous axial slices acquired through the brain on each occasion.
Brain atrophy measures. Central cerebral volume (CCV) was measured using an automated technique that segments cerebral tissue from surrounding scalp and other extracerebral tissue using a four-step algorithm. The details of the methodology are described elsewhere.9,10 The slices were chosen with the most caudal being at the level of the velum interpositum cerebri. Four contiguous, axial, 5-mm-thick slices were studied. This region of the cerebral hemispheres was chosen because in a previous study 1) there had been substantial atrophy seen over an 18-month period in subjects with SPMS9 and 2) the measure–reposition– rescan–remeasure coefficient of variability of the method was 0.56%.9 SIENAX was used to measure normalized brain volume.14 SIENAX automatically segments brain from nonbrain matter, calculates the brain volume, and applies a normalization factor to 596
Neurology 72
February 17, 2009
correct for skull size. The normalization factor is obtained by registering the subject’s scan to the Montreal Neurological Institute (MNI) 152 standard image using the skull to normalize spatially. Percentage brain volume change (PBVC) for each time point relative to baseline was measured using SIENA.14 SIENA registers the baseline and follow-up magnetic resonance image using the skull as scale and skew constraint, and then estimates the displacement of the brain edge for each point of the brain edge between these two scans. The brain edge displacements of all edge points are used to calculate the “overall” PBVC, which is expressed as a single value. Because not all scans included the full brain, the SIENAX and SIENA analyses were restricted to a prespecified interval along the z-axis, ranging from ⫺52 to ⫹60 mm in standard MNI152 space. When necessary, errors in brain extraction were corrected manually by a single experienced observer; this has been shown previously13 to reduce unwanted variability in SIENA and SIENAX results without materially introducing interobserver/intercenter variability; all scans required manual correction to a varying extent. SIENAX and SIENA are part of the FMRIB Software Library (FSL).15 All SIENAX and SIENA analyses were performed using FSL version 3.1.
Statistical methods and issues. Sample size estimates were calculated for trial durations of 12, 24, and 36 months to detect treatment effects of 30%, 40%, 50%, and 60% at 80% and 90% power, all with a two-tailed ␣ (significance level) of 5%. Treatment is assumed to have an immediate and constant effect, and in the absence of a healthy control group treatment effects assume zero atrophy in healthy subjects, 100% equating with zero volume loss. For each duration, three standard statistical analysis methods were considered for the comparisons between active and placebo trial groups: 1) comparison of the mean change from baseline, using a t test; 2) comparison of baseline adjusted mean change from baseline, using analysis of covariance (ANCOVA)16; and 3) comparison of mean rates of change estimated from longitudinal linear mixed models,17 using either 6-monthly or annual time points. Relative efficiencies are used to summarize comparisons: the relative efficiency of procedure A vs B is the inverse of the ratio of the corresponding sample sizes required to achieve the same power. These methods are discussed further below, but technical details of the statistical models and calculations are given in appendix e-1 on the Neurology® Web site at www.neurology.org. A number of issues are relevant to the comparisons we present and to their potential impact on trial design. Chiefly, these relate to the choice of sample required to obtain valid comparisons between outcomes or between different trial durations or statistical analyses, and issues regarding outcome type. Choice of samples for comparison. For the primary comparison, between atrophy measures, best estimates come from subjects with all three measures available at a given time point, “all-three” samples. This ensures that differences between measures are not due to different subjects. For these comparisons, at different time points, sample sizes were calculated just for a 50% treatment effect (because the relative efficiency of the volume measures is approximately constant over different treatment effects for a given analysis method and duration). For any given trial duration and analysis method, this gives a valid comparison across the atrophy measures. For the simplest analysis method, the t test of changes, the nonparametric bias-corrected bootstrap18 (1,000 replicates), was used to assess the statistical significance of sample size differences between the measures: standard errors for the differences in sample size estimates are not theoret-
Table 1
Volumes and changes from baseline for the three measures, by month, with numbers of patients contributing (maximum n ⴝ 43)
Measure
CCV, cm3 n; Mean (SD)
0
Volumes
43; 307.87 (269.77)
41; 1,460.57 (63.15)
—
41†
6
Volumes
34; 302.97 (26.11)
40; 1,446.88 (55.04)
—
32
Changes
34; ⫺2.73 (4.17)
40; ⫺11.23 (31.97)
—
[As %‡]
[⫺0.89 (1.35)]
[⫺0.73 (2.14)]
40; ⫺0.63 (0.78)
Volumes
38; 304.95 (27.08)
40; 1,452.79 (58.82)
—
Changes
38; ⫺4.37 (5.26)
39; ⫺7.68 (38.25)
—
[As %‡]
[⫺1.40 (1.69)]
[⫺0.48 (2.59)]
39; ⫺1.24 (1.05)
Volumes
38; 301.27 (25.13)
37; 1,443.88 (66.52)
—
Changes
38; ⫺6.44 (9.11)
35; ⫺12.19 (44.86)
—
[As %‡]
[⫺2.02 (2.83)]
[⫺0.79 (3.09)]
35; ⫺1.73 (1.19)
Volumes
33; 300.36 (24.24)
36; 1,431.92 (61.82)
—
Changes
33; ⫺8.82 (10.66)
34; ⫺25.55 (50.36)
—
[⫺2.76 (3.30)]
[⫺1.68 (3.40)]
34; ⫺2.47 (1.57) —
12
18
24
[As %‡] §
30
36
SIENAX, cm3 n; Mean (SD)
No. with all three measures
Month
SIENA, % n; Mean (SD)
Correlation
0.91
0.65
Volumes
29; 298.99 (24.59)
29; 1,433.86 (68.52)
—
Changes
29; ⫺8.76 (10.64)
27; ⫺22.39 (42.46)
—
[As %‡]
[⫺2.74 (3.38)]
[⫺1.49 (2.93)]
27; ⫺2.43 (1.64)
Volumes
31; 293.73 (27.18)
30; 1,431.57 (75.41)
—
Changes
31; ⫺11.96 (10.96)
29; ⫺26.28 (50.03)
—
[As %‡]
[⫺3.86 (3.53)]
[⫺1.77 (3.50)]
29; ⫺3.24 (2.23)
Correlation§
0.92
0.76
—
37
34
30
24
28
*Unless otherwise indicated. †Baseline implicit for SIENA in subjects with at least one later observation. ‡As percentage of baseline: mean (SD). § Pearson correlation coefficients between absolute volume at time point, and baseline (all p ⬍ 0.0001). CCV ⫽ central cerebral volume.
ically available, but in this context the bootstrap method gives a valid test, estimating confidence intervals for the differences empirically by multiple resampling (replicates) of the data. (p value ranges are given because of the computationally intensive nature of the bootstrap). For best results within each individual measure and also for the secondary comparison between analysis methods and trial durations using a given measure, optimal estimates are given for each volume measure separately by fitting a longitudinal model using an “all-data” sample: the 36-month duration 6-monthly longitudinal model, which uses every available time point for that measure. Because the “all-three” samples have to drop a subject at a given time point if one of the three measures is missing, the “all-data” sample gives additional information on the robustness of the “all-three” comparisons to missing data. The estimated slope and variance parameters for the “all-data” model were then used to deduce the parameters relevant to the different statistical analyses and time points and thus generate the appropriate sample sizes. Thus, from the single set of “master” 36-month parameters, we obtain a valid comparison of the different analysis methods and durations in each measure, assuming constant atrophy over the period. Under this assumption, these parameters also allow estimation of the effect of altering observation times. It has been shown19 that the timing of observations is relevant to gains in power, e.g., adding a third observation midway between baseline and final follow-up provides no additional information with which to estimate linear
change. Though our primary aim is to compare the volume measures rather than establish optimal design, for interest we report some efficiency gains from a theoretically more efficient concentration of observations toward the trial period extremes. Volume measures. The methodology of SIENA, calculating the percentage brain volume change (PBVC), is a “direct”20 measure of change, with theoretically less measurement error compared to indirect measures of change obtained by numerical subtraction between volumes calculated at separate time points, as is required for CCV and SIENAX. The superior precision of SIENA compared with indirect volume measures has been noted previously in cohorts with relapsing–remitting MS (RRMS).21-23 However, direct difference methods have a different error structure than absolute measures, and this was taken account of in constructing the longitudinal models to estimate SIENA parameters.20 To examine the concordance between the three measures, the “all-three” sample was used, with CCV and SIENAX converted into PBVC units using 100 ⫻ (volume at time point ⫺ baseline volume)/baseline volume. Pearson correlation coefficients and Bland–Altman plots24 were obtained, and the standard deviations of the measures were statistically compared using the Pitman test25 for paired variances.
Of the 46 patients available, a maximum of 43 patients were used in the analyses: 2 subjects were excluded having only SIENAX baseline and no other valid measurements (both dropped out at 6 months), and 1 subject with only baseline measures in CCV and SIENAX (6-month scan electronic data rejected and then dropped out at 12 months) was also excluded. The patients provided a maximum of 246 data points for the analyses. From a theoretical maximum of 43 ⫻ 7 ⫽ 301 observations, 55 were missing: 25 because of patient dropout, 3 because of scan nonacquisition, 17 because of electronic data rejection, 1 because of hard copy (and therefore electronic data) rejection, and 9 because of unavailable electronic data. Table 1 shows the number of patients with all three measures available at any one time point, along with summary statistics of changes in volume from baseline and, for CCV and SIENAX only, absolute volumes and correlations between baseline and later volumes.
RESULTS
Concordance between the volume measures. There
was in general much better agreement between SIENA and CCV percentage changes than with SIENAX (table 1; figure). Concordance between the three measures is further detailed in appendix e-2; figure e-1, A–C; and figure e-2, A–C. Comparison of sample size estimates between the measures. Table e-1 gives the parameter estimates on
which the sample size calculations for the “all-three” comparisons are based. (Details of the longitudinal parameters are given in appendix e-1.) Longitudinal model residuals did not show any serious nonnormality. Table 2 shows sample size estimates for 50% Neurology 72
February 17, 2009
597
Figure
(36-month longitudinal) and 15.2 (12-month t test) compared with SIENAX. Bootstrap inference, for the pairwise differences in t test sample sizes between measures, showed that all sample size differences were p ⬍ 0.05: in particular, SIENA vs SIENAX gave p ⬍ 0.001 at all three durations; SIENA vs CCV gave 0.03 ⬍ p ⬍ 0.04 at 12 months, 0.004 ⬍ p ⬍ 0.005 at 24 months, and 0.01 ⬍ p ⬍ 0.02 at 36 months; and CCV vs SIENAX gave 0.001 ⬍ p ⬍ 0.002 at 12 months, 0.02 ⬍ p ⬍ 0.03 at 24 months, and 0.01 ⬍ p ⬍ 0.02 at 36 months.
Mean percentage changes from baseline by month for the three measures— central cerebral volume, SIENAX, and SIENA— calculated in the “all-three” sample
Comparison of sample size estimates between analysis methods and trial durations. Table e-2 gives the pa-
rameter estimates underlying these sample size calculations. Table e-3 shows the sample size estimates across the different analysis methods and trial durations, for each volume measure separately, allowing valid comparisons within the columns. For all measures, the most influential factor in determining sample sizes is trial duration. Minimum 2-year sample sizes per arm for 50% treatment effect at 80% power were 32 for SIENA, 69 for CCV, and 273 for SIENAX and were 71%, 55%, and 44% lower than corresponding 1-year sizes. Detailed comparisons between analysis methods and trial durations are presented in appendix e-3. Key points are that adding an observation at the midpoint of the follow-up period does not add relevant information to the baseline and
CCV ⫽ central cerebral volume.
treatment effect across the three measures, but the sample size ratios (relative efficiencies) within any single row would be the same for other treatment effects. SIENA has relative efficiencies between 2 (36-month t test) and 2.5 (24-month t test) compared with CCV and between 6.8 (36-month longitudinal) and 31.8 (12-month t test) compared with SIENAX. CCV has relative efficiency between 3.2 Table 2
Comparison* of the three measures for 50% treatment effect: n per trial arm Volume measure
Statistical analysis
Patients† contributing to all three measures
CCV power
SIENAX power
SIENA power
80%
90%
80%
90%
80%
90%
t Test of change 12 m–baseline
37
98
130
1,494
2,000
47
63
24 m–baseline
30
70
93
250
335
28
37
36 m–baseline
28
60
80
235
315
30
40
12 m
37
97
129
1,318
1,764
—
—
24 m
30
68
91
199
267
—
—
36 m
28
57
76
207
277
—
—
ANCOVA
Longitudinal, with time points 0, 6, 12
41:226
139
186
1,137
1,521
56
74
0, 6, 12, 18, 24
41:226
69
92
316
423
31
42
0, 6, 12, 18, 24, 30, 36
41:226
58
77
184
246
27
36
*Because different numbers of patients contributed to the estimates for different analyses and time points, comparisons are validhorizontally across the measures but not generally across analyses or time points. †For change and analysis of covariance (ANCOVA), each patient contributed two data points toward estimating the parameters from which the sample sizes are calculated. For the longitudinal models, parameters were estimated from the largest sample, with 41 patients contributing a total of 226 data points. CCV ⫽ central cerebral volume. 598
Neurology 72
February 17, 2009
final scans, while the effect of additional informative (noncentral) time points for a given duration is greater the more variable the measure. Thus, additional informative time points have an impact for SIENAX, with its greater variability and lower correlation between times; but for CCV, and particularly for SIENA, adding time points between baseline and last follow-up gives little theoretical gain, even if the scans are clustered at the period extremes, provided there is negligible patient dropout. DISCUSSION Sample sizes based on four volume measures including SIENA21 and SIENA precision23 have been estimated previously in RRMS cohorts, reporting the superior precision of SIENA compared with indirect measures of volume change. Our results show generally better agreement between CCV and SIENA than between either of these and SIENAX. Differences between CCV and SIENA may be because the latter is a registration-based method directly measuring brain volume changes, whereas the former involves numerical subtraction. Additionally, these differences may be due to using a greater portion of the brain for SIENA. Nevertheless, there was good agreement between these two measures, particularly regarding longitudinal trajectory. Comparing the three measures for the same analyses/durations gives highest sample sizes for SIENAX, followed by CCV and then SIENA, with the advantage of SIENA more pronounced at shorter durations. These results are explained by the comparative standard deviations of the three measures, relative to treatment effects. Although the variability of SIENAX absolute volumes, as a percentage of the volume, is actually lower than for CCV, the SIENAX changes have much higher variability than the other two measures, leading to higher SIENAX sample sizes for the analyses of changes. For the longitudinal models, sample sizes over shorter durations are dominated by the within-subject standard deviation, which was highest relative to treatment effect for SIENAX and lowest for SIENA. Over longer durations, sample sizes are influenced more by the between-subject atrophy rate standard deviation, which was again highest for SIENAX and lowest for SIENA. Although some patients were lost to the “allthree” sample underlying direct between-measure comparisons, the general similarity in sample sizes from the “all-three” and the “all-data” samples suggest the between-measure comparison is robust to patient loss. Although in theory analyzing CCV with adjustment for baseline intracranial volume would only reduce the variability between subjects at baseline rather than of atrophy rates and, therefore, may not
greatly enhance power in longitudinal studies, further work is required to assess the potential gains from such adjustment. Further work is also required to assess any change in power from calculating SIENA direct changes between consecutive time points, rather than from baseline as in these data; or from using ANCOVA to adjust SIENA for baseline SIENAX, though our data suggest little gain from this because ANCOVA results tend to approach but not improve on the corresponding longitudinal analysis with annual time points. Detecting smaller treatment effects, or increasing test power, naturally increased the required sample sizes. Comparing analyses and durations, for all three measures, increasing the duration or the number of informative (i.e., not midway) time points reduced the required sample size, with increased duration generally having greater impact than number of time points. In general, “noisier” measures gain more than precise measures from an increase in the number of informative data points: thus, SIENAX gains the most from increasing the intrinsic power of the analysis by extending duration or adding points (particularly points toward the period extremes), followed by CCV, with the least gains for SIENA. SIENA sample sizes for different trial durations have previously been estimated21 as 69 (1 year), 44 (2 years), and 40 (3 years), based on an RRMS cohort to be analyzed with t tests of change at 90% power and 50% treatment effect, close to our corresponding 77, 45, and 39 in an SPMS cohort (table e-3). This might suggest that— despite of the use of different T1-weighted sequences on which atrophy was measured (three-dimensional in the RRMS group, twodimensional in the SPMS group)—the average rate of brain atrophy and its variance between subjects may be similar in RRMS and SPMS cohorts.26 The SPMS cohort in our European trial of interferon beta-1b had more ongoing relapses and a shorter disease duration than the SPMS cohort that took part in a North American trial of interferon beta-1b,27 and further work might investigate sample sizes in a longer-disease-duration nonrelapsing SPMS cohort. One assumption that may exaggerate the study power is that 100% treatment effect equates to zero volume loss. However, healthy controls experience some brain volume loss (0.1%– 0.3% per year), and if disease-specific treatment effects do not affect the “normal” atrophy associated with aging, a larger sample size will be required to show the same diseasespecific effect. If 0.1% “healthy” annual loss is assumed, the SIENA sample size of 28 required for a 50% treatment effect, 80% power 3-year longitudinal analysis increases to 33; if 0.3% is assumed, the new sample size is 50. This effect might be allowed Neurology 72
February 17, 2009
599
for in analysis models where healthy controls are scanned using the same protocol. Determining optimal trial design has to take careful consideration of issues such as dropout rate and scanning burden on patients, and is outside the scope of this article; we can here only highlight relevant factors. It is important to note that the relatively small gain in power for SIENA and CCV shown by multi–time point longitudinal analyses compared with t tests and ANCOVA conceals an important advantage of the more sophisticated models: missing one data point at either baseline or final follow-up will remove a subject from the simpler analyses, whereas the longitudinal models can use all available data points efficiently and thus minimize the impact of missing data, in terms of both power and potential bias from differential dropout. Possible dropout toward the end of follow-up may also limit the power gains from timing scans near the trial end rather than spacing them regularly.19 We assumed a linear volume change over time. Testing for nonlinearity, we found weak evidence of trajectories leveling off over time, consistent with a proportionate change, which is linear on a logarithmic volume scale. As a precaution, we repeated the sample size calculations on the log outcomes, but obtained sizes almost identical to those we report for SIENAX and SIENA and around 10% greater for CCV (probably because the changes tend to be larger as a proportion of absolute volumes for CCV than for the other measures). Further work on larger data sets would be required to assess possible nonlinearity satisfactorily. For CCV and particularly for SIENA, extending the trial duration from 2 to 3 years reduces sample sizes relatively modestly. In contrast, extending the duration from 1 to 2 years can roughly halve the sample sizes required for these outcomes. A further disadvantage of 1-year duration is the possible shortterm effect of biologic confounds tending to undermine sample size calculations, which, as here, assume immediate onset and constancy of treatment effect. First, any wallerian degeneration from axonal injury before the commencement of treatment may continue to evolve, and thus cause atrophy, for several months after the start of treatment, possibly delaying any treatment benefit from manifesting as reduced atrophy rate. Second, if the therapy has an antiinflammatory as well as a neuroprotective effect, it may cause an initial decrease in brain volume due to resolution of inflammation. Such an effect has been proposed to contribute to decreases in brain volume seen after treatment with IV methylprednisolone,28 beta interferon,6,29 and natalizumab.8 To avoid these confounds, baseline for analysis could be taken after 600
Neurology 72
February 17, 2009
an initial treatment “burn in” period. The appropriate interval is uncertain, but 3 or 6 months might be considered reasonable.29 AUTHOR CONTRIBUTIONS Statistical analysis was conducted by D.R.A.
ACKNOWLEDGMENT The authors thank Stenmar van Steenbrugge for assisting in the SIENA and SIENAX analyses and Chris Frost and Jonathan Bartlett for their statistical advice.
Received May 19, 2008. Accepted in final form August 20, 2008. REFERENCES 1. Trapp BD, Peterson J, Ransohoff RM, et al. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998;338:278–285. 2. Evangelou N, Esiri MM, Smith S, Palace J, Matthews PM. Quantitative pathological evidence for axonal loss in normal appearing white matter in multiple sclerosis. Ann Neurol 2000;47:391–395. 3. Kutzelnigg A, Lucchinetti CF, Stadelmann C, et al. Cortical demyelination and diffuse white matter injury in multiple sclerosis. Brain 2005;128(pt 11):2705–2712. 4. Peterson JW, Bo¨ L, Mo¨rk S, Chang A, Trapp BD. Transected neuritis, apoptotic neurons, and reduced inflammation in cortical multiple sclerosis lesions. Ann Neurol 2001;50:389–400. 5. Miller DH, Barkhof F, Frank JA, Parker GJM, Thompson AJ. Measurement of atrophy in multiple sclerosis: pathological basis, methodological aspects and clinical relevance. Brain 2002;125:1676–1695. 6. Rudick RA, Fisher E, Lee JC, Simon J, Jacobs L. Use of the brain parenchymal fraction to measure whole brain atrophy in relapsing-remitting MS. Multiple Sclerosis Collaborative Research Group. Neurology 1999;53:1698–1704. 7. Filippi M, Rovaris M, Inglese M, et al. Interferon beta-1a for brain tissue loss in patients at presentation with syndromes suggestive of multiple sclerosis: a randomised, double-blind, placebo-controlled trial. Lancet 2004;364: 1489–1496. 8. Miller DH, Soon D, Fernando KT, et al. MRI outcomes in a placebo-controlled trial of natalizumab in relapsing MS. Neurology 2007;68:1390–1401. 9. Losseff NA, Wang L, Lai HM, et al. Progressive cerebral atrophy in multiple sclerosis: a serial MRI study. Brain 1996;119:2009–2019. 10. Molyneux PD, Kappos L, Polman C, et al. The effect of interferon beta-1b treatment on MRI measures of cerebral atrophy in secondary progressive multiple sclerosis. Brain 2000;123:2256–2263. 11. Stevenson VL, Smith SM, Matthews PM, Miller DH, Thompson AJ. Monitoring disease activity and progression in primary progressive multiple sclerosis using MRI: sub-voxel registration to identify lesion changes and to detect cerebral atrophy. J Neurol 2002;249:171–177. 12. Smith SM, De Stefano N, Jenkinson M, Matthews PM. Normalized accurate measurement of longitudinal brain change. J Comput Assist Tomogr 2001;25:466–475. 13. Jasperse B, Valsasina P, Neacsu V, et al. Intercenter agreement of brain atrophy measurement in multiple sclerosis patients using manually-edited SIENA and SIENAX. J Magn Reson Imaging 2007;26:881–885.
14.
15.
16.
17. 18.
19.
20.
21.
22.
Smith SM, Zhang YY, Jenkinson M, et al. Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 2002;17:479–489. Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 2004;23(suppl 1):208– 219. Frison L, Pocock SJ. Repeated measures in clinical trials: analysis using mean summary statistics and its implications for design. Stat Med 1992;11:1685–1704. Goldstein H. Multilevel Statistical Models. Kendall’s Library of Statistics Series 3. London: Hodder Arnold; 1995. Carpenter JR, Bithall JF. Bootstrap confidence intervals: when, which, what? A practical guide for medical statisticians. Stat Med 2000;19:1141–1164. Schott JM, Frost C, Whitwell JL, et al. Combining short interval MRI in Alzheimer’s disease: implications for therapeutic trials. J Neurol 2006;253:1147–1153. Frost C, Kenward MG, Fox NC. The analysis of repeated “direct” measures of change illustrated with an application in longitudinal imaging. Stat Med 2004;23:3275–3286. Anderson VM, Bartlett JW, Fox NC, Fisniku L, Miller DH. Detecting treatment effects on brain atrophy in relapsing remitting multiple sclerosis: sample size estimates. J Neurol 2007;254:1588–1594. Anderson VM, Fernando KT, Davies GR, et al. Cerebral atrophy measurement in clinically isolated syn-
23.
24.
25.
26.
27.
28.
29.
dromes and relapsing remitting multiple sclerosis: a comparison of registration-based methods. J Neuroimaging 2007;17:61–68. Sormani MP, Rovaris M, Valsasina P, Wolinsky JS, Comi G, Filippi M. Measurement error of two different techniques for brain atrophy assessment in multiple sclerosis. Neurology 2004;62:1432–1434. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310. Armitage P, Berry G, Matthews J. Statistical Methods in Medical Research, 4th ed. Oxford: Blackwell Science; 2002. Kalkers NF, Ameziane N, Bot JC, Minneboo A, Polman CH, Barkhof F. Longitudinal brain volume measurement in multiple sclerosis: rate of brain atrophy is independent of the disease subtype. Arch Neurol 2002;59:1572–1576. Panitch H, Miller A, Paty D, et al. Interferon beta-1b in secondary progressive MS: results from a 3-year controlled study. Neurology 2004;63:1788–1795. Rao AB, Richert N, Howard T, et al. Methylprednisolone effect on brain volume and enhancing lesions in MS before and during IFNbeta-1b. Neurology 2002;59:688–694. Hardmeier M, Wagenpfeil S, Freitag P, et al. Rate of brain atrophy in relapsing MS decreases during treatment with IFNbeta-1a. Neurology 2005;64:236–240.
Learn. Earn. Network. 2009 AAN Annual Meeting: An Excellent Value • Learn about the latest scientific advances in neurology • Earn valuable CME credit and fulfill Maintenance of Certification requirements • Network with your peers at exciting social events all week long • Enjoy the convenience and value of all this and more—in just one meeting Early registration and hotel deadline is March 20, 2009. Register today at www.am.com/AM2009.
Support Research at the AAN Foundation Wine Tasting & Auction You are cordially invited to sample some of the finest wines in the Pacific Northwest. Enjoy an array of food, music, and fun. This exclusive 2009 Annual Meeting event takes place Wednesday, April 29, from 7:00 p.m. to 9:30 p.m. in the Grand Ballroom at the Sheraton Hotel in Seattle. Bid on a variety of silent auction prizes including exquisite wine, entertainment, gifts, and more. The evening culminates with a live auction of exciting prizes. Bid high and often to support research. Proceeds benefit the AAN Foundation Research Program. Tickets are $100 per guest. Tickets are limited, so go to www.aan.com/wine and register now!
Neurology 72
February 17, 2009
601
Demyelinating events in early multiple sclerosis have inherent severity and recovery E.M. Mowry, MD M. Pesic, MD B. Grimes, PhD S. Deen, MD P. Bacchetti, PhD E. Waubant, MD, PhD
Address correspondence and reprint requests to Dr. Ellen M. Mowry, Department of Neurology, Multiple Sclerosis Center, University of California, San Francisco, 350 Parnassus Avenue, Suite 908, San Francisco, CA 94117
[email protected]
ABSTRACT
Background: It is unclear whether the severity of and recovery from the initial demyelinating event (IDE) are recapitulated in subsequent multiple sclerosis (MS) relapses. We sought to identify the factors associated with relapse severity and recovery and to evaluate whether events have inherent severity or recovery.
Methods: Patients seen at the UCSF MS Clinic within 1 year of disease onset were identified from a prospective database. Ordinal logistic regression was used to analyze predictors of three-level categorizations of event severity and recovery.
Results: We identified 330 patients with MS or clinically isolated syndrome; 152 had a second event and 63 had a third event. Nonwhite and younger patients were at an increased risk of more severe demyelinating events. A severe prior event predicted a substantial increase in the odds of being above any given severity cutoff for a severe subsequent event (for second event severity, odds ratio [OR] ⫽ 5.62, 95% confidence interval [CI] [2.39, 13.26], p ⬍ 0.0001; for third event severity, OR ⫽ 6.74, 95% CI [1.67, 27.18], p ⫽ 0.007). Similarly, poor recovery of the IDE predicted poor second event recovery (OR ⫽ 5.28, 95% CI [1.95, 14.25], p ⫽ 0.001), while fair or poor second event recovery predicted about a 5- or 13-fold increase in the odds of poor third event recovery. A more severe event also predicted a substantial increase in the odds of poor recovery.
Conclusions: Patients with severe presentation and poor recovery at disease onset continue on a similar trajectory with subsequent demyelinating events. Whether genetic or other biologic factors are responsible for this pattern remains to be determined. Neurology® 2009;72:602–608 GLOSSARY BS/CE ⫽ brainstem/cerebellum; CI ⫽ confidence interval; CIS ⫽ clinically isolated syndrome; DMT ⫽ disease-modifying therapy; EDSS ⫽ Expanded Disability Status Scale; FS ⫽ Functional Systems; IDE ⫽ initial demyelinating event; MS ⫽ multiple sclerosis; N/A ⫽ not applicable; OR ⫽ odds ratio; RRMS ⫽ relapsing-remitting multiple sclerosis; VA ⫽ visual acuity.
The significant variability in the severity of and recovery from demyelinating events in relapsing-remitting multiple sclerosis (RRMS) reflects patients’ heterogeneity.1,2 The clinical determinants of event severity and recovery are poorly understood, and accumulation of residual disability due to incomplete relapse recovery remains largely unpredictable. Initial demyelinating event (IDE) severity and recovery may be important predictors for both short-and long-term disability (i.e., long-term prognosis may be determined very early in the disease process).1,3-9 A few authors reported that IDE severity is the most significant predictor of IDE recovery, but polyregional or polysymptomatic onset, onset location, and age may also be important.7,10,11 Predictors of event severity in early MS, other predictors of IDE recovery, and predictors of subsequent event recovery have not been studied. Furthermore, it is unknown if a
From the MS Center, Department of Neurology (E.M.M., M.P., S.D., E.W.), and Department of Epidemiology and Biostatistics (B.G., P.B.), University of California, San Francisco. Supported by the NMSS (RG-3692A), the Nancy Davis Foundation, a National Multiple Sclerosis Society Sylvia Lawry Fellowship Award, and the Partners MS Center Clinical Fellowship Program. Statistical analysis for this publication was made possible by NIH/NCRR UCSF-CTSI Grant Number UL1 RR024131. Disclosure: Dr. Waubant has received research support from Biogen Idec, Genentech Inc, Pfizer, and Sanofi-Aventis, and honorarium for one educational presentation from Teva and Biogen. 602
Copyright © 2009 by AAN Enterprises, Inc.
given patient has an inherent tendency to develop clinical demyelinating events of similar severity and recovery. In this prospective cohort study, we sought to identify clinical factors associated with the severity and recovery of early clinical events in MS. In addition, we evaluated whether an event’s severity and recovery are associated with the severity and recovery of subsequent events. METHODS This project was approved by the UCSF Committee on Human Research. Clinical and demographic information for all patients seen at the UCSF MS Center is entered into a Microsoft SQL server database (retrospectively at the first visit and prospectively for subsequent visits).10,12 Routine follow-up visits typically occur every 6 months; unscheduled visits occur if a patient has an exacerbation. We queried the database for all patients with clinically isolated syndrome (CIS) and RRMS seen within the first year of disease onset between January 2000 and June 2007.13 We included only patients seen early after disease onset since accumulated disability and frequent symptom fluctuations may preclude accurate relapse characterization later in the disease course. Patients were divided into two groups to analyze the effects of race/ethnicity: white, non-Hispanic (hereafter referred to as white) and nonwhite (all others). Demyelinating events were new or recurring neurologic symptoms referable to the CNS lasting for at least 48 hours after a remission of 30 days or more since the previous event. Pseudoexacerbations were excluded. Based on clinical history and examination, each patient’s relapses were coded as occurring in the spinal cord, brainstem/cerebellum, optic nerve, or cerebrum. An event was considered polyregional if it involved at least two of these locations. A patient was considered as being on disease-modifying therapy (DMT) if he or she had at least 90 days of continuous treatment, since it is thought that there is a lag between initiation of therapy and the onset of therapeutic effectiveness.14,15 The severity of and recovery from the IDE was determined by trained individuals (S.D., E.M.) based on definitions derived from previous publications.10,11,16,17 Mild IDE severity was defined as Functional Systems (FS) scores of 0 to 1 in one to three FSs or visual acuity (VA) better than or equal to 20/40, Expanded Disability Status Scale (EDSS) score range of 0 to 1.5 inclusive; moderate severity was defined as a score of 2 in one or two FSs or four or more scores of 1 or VA of 20/50 to 20/190, EDSS score of 2.0 to 2.5 inclusive; relapses exceeding prior criteria were considered severe. Recovery was scored using the lowest EDSS and FS scores reported between 2 and 12 months after the attack. IDE recovery was considered complete (no residual complaint, normal follow-up examination, all FS scores ⫽ 0, follow-up EDSS score ⫽ 0), fair (residual subjective complaint that does not impair activity, or at least one FS score of 1 at most or VA better or equal to 20/40, follow-up EDSS score ⫽ 1.0 to 1.5), or poor (at least one FS score of 2 or more or VA of 20/50 or worse, follow-up EDSS score 2.0 or greater). For second and third events, severity was scored the same way as for the IDE if the pre-event EDSS was 0; if the pre-event EDSS was ⬎0, the severity was defined as mild (EDSS increase by 0.5 point, or 1 point change in up to three FS scores), moderate (EDSS increase by 1 or 2 points, or 2 points change in up to
two FS, or 1 point change in four or more FS), or severe (exceeding prior criteria). Recovery from the second or third event was defined as complete if no residual signs or symptoms remained above those present before the attack, fair if EDSS increased by up to 1 point or if there was an increase of 1 point on one or two FS (residual subjective complaint or new residual finding compared to baseline that does not impair activity), or poor if exceeding prior criteria. Potential predictors of severity and recovery of all events included sex, race/ethnicity, age at onset, event location, and abnormal (at least one T2-weighted hyperintensity) vs normal baseline brain MRI. There were not enough cerebral second or third events to provide useful analyses. Monoregional vs polyregional event status was evaluated for an association with event recovery. The severity of the preceding event was used to predict that event’s recovery as well as the subsequent event’s severity. Preceding event recovery was used to predict subsequent event recovery. DMT use was added as a predictor of severity and recovery of second and third events. Multivariate models were generated to evaluate potential confounding. Because event severity is to some extent collinear by definition with monoregional vs polyregional status, the multivariate models for recovery included only severity as a predictor.
Statistics. Statistical analyses were performed by Barbara Grimes, PhD, and Peter Bacchetti, PhD, University of California, San Francisco. Appropriate summary statistics were used to describe categorical and continuous variables. Since severity and recovery were measured on an ordered, three-level scale, ordinal logistic regression was used. This method assumes a common odds ratio (OR) for each predictor’s association with both severe vs mild or moderate severity and severe or moderate vs mild severity; similar assumptions were made for recovery (the proportional odds assumption). When there was evidence against this assumption, we dichotomized the outcome and performed logistic regression. For the severity analyses, the dichotomized outcome was severe/moderate vs mild events. Violations occurred for univariate predictors of IDE (brainstem/cerebellar or optic nerve involvement), second event (spinal cord symptoms), and third event (nonwhite race/ethnicity, brainstem/cerebellar or optic nerve symptoms) severity as well as for the IDE and third event multivariate models. For the recovery analyses, the dichotomized outcome was poor/fair vs complete recovery. It was required for univariate analyses for IDE (gender, IDE severity, brainstem/cerebellar, or spinal cord symptoms) and second event (disease-modifying therapy) recovery as well as for the multivariate model of IDE recovery. RESULTS Patient and event characteristics. We identified 330 patients (224 women) seen at the UCSF MS Center within a year of the first MS symptoms; mean follow-up was 759 ⫾ 575 days (median 633 days, range 23–2,692 days). The mean age at IDE onset was 34 ⫾ 12 years. A total of 267 (81%) patients were white; the remaining patients were African American (21), Asian (15), Hispanic or Latino (14), Native American (1), or unknown (12). At onset, 301 (93%) of the 323 patients who had available imaging had an abnormal brain MRI. A total of 153 (46%) patients received high dose IV steroids for the IDE, and DMT was initiated in 54% of patients (n ⫽ 178) during the entire Neurology 72
February 17, 2009
603
Table 1
polyregional in 48 (15%) of first, 20 (13%) of second, and 13 (21%) of third relapses.
Locations, severity, and recovery of first, second, and third demyelinating events First event (n ⴝ 330)
Second event (n ⴝ 152)
Third event (n ⴝ 63)
Spinal cord
175 (53)
98 (65)
45 (71)
Brainstem/cerebellum
110 (33)
50 (33)
21 (33)
Optic nerve
86 (26)
25 (16)
10 (16)
Cerebrum
10 (3)
Event characteristic
Factors associated with event severity. Univariate anal-
Location, n (%)*
0
0
Severity, n (%) Mild
134 (41)
73 (48)
36 (57)
Moderate
139 (42)
58 (38)
21 (33)
57 (17)
21 (14)
6 (10)
Complete
150 (46)
84 (55)
26 (41)
Fair
118 (37)
52 (34)
24 (38)
Severe Recovery, n (%)†
Poor
54 (17)
11 (7)
6 (10)
*Total does not add up to 100% since some patients had multiple sites affected. †Could not be calculated for a few individuals.
follow-up period (Avonex 29%, Rebif 10%, Copaxone 8%, Betaseron 6%, other 1%). Only 36 patients (24%) who had a second event initiated DMT before it occurred; 35 patients (56%) who had a third event initiated DMT beforehand. A second clinical event was experienced by 46% (n ⫽ 152); 19% (n ⫽ 63) experienced a third. The severity, recovery, and locations of events are presented in table 1. Events were
Table 2
yses. Nonwhite and younger patients had a higher odds of experiencing more severe first, second, and third events (table 2). The associations of race and age with third event severity had wide confidence intervals, reflecting the smaller number of subjects. Treatment with DMT prior to the second or third event was not meaningfully associated with severity (table 2). A more severe preceding event was associated with a substantial increase in the odds of a more severe second or third event, with a more than threefold increase in the odds if the first event was moderate compared to mild and a greater than fiveto sixfold increase in the odds if the preceding event was severe compared to mild (table 2). Poor recovery of the prior event predicted substantially increased odds of a more severe subsequent event (table 2). Multivariate analyses. In the multivariate analysis evaluating predictors of IDE severity, which included age, location, and race, there were not substantial changes from the univariate analyses except that optic neuritis was more likely to be associated with IDE severity, whereas spinal cord onset did not seem to meaningfully predict severity (table 3). The multivariate model for second event severity included age, race, location, IDE severity and recovery, and DMT. The results were not meaningfully dif-
Predictors (univariate) of increased first, second, and third event severity First event
Second event p
Third event
Predictor
OR
95% CI
OR
95% CI
OR
95% CI
p
Nonwhite race/ethnicity
1.79
1.07, 3.00
0.027
2.85
1.37, 5.93
p 0.005
1.75
0.57, 5.37
0.33
Age (10-year decrease)
1.32
1.10, 1.56
0.003
1.54
1.15, 2.04
0.004
1.25
0.78, 1.96
0.36
Spinal cord
0.46
0.30, 0.69
0.0002
0.46
0.24, 0.89
0.021
0.68
0.24, 1.98
0.48
BS/CE
4.48
2.60, 7.71
⬍0.0001
2.01
1.05, 3.83
0.034
1.42
0.51, 3.92
0.50
Optic nerve
0.99
0.60, 1.64
0.98
1.22
0.54, 2.74
0.63
0.80
0.23, 2.77
0.72
Cerebrum*
7.67
2.17, 27.13
0.002
N/A
N/A
N/A
N/A
N/A
N/A
Prior event severity (moderate vs mild)
N/A
N/A
N/A
3.04
1.48, 6.21
0.002
3.87
1.18, 12.68
0.026
Prior event severity (severe vs mild)
N/A
N/A
N/A
5.62
2.39, 13.26
⬍0.0001
6.74
1.67, 27.18
0.007
Prior event recovery (fair vs complete)
N/A
N/A
N/A
0.65
0.33, 1.28
0.22
1.31
0.44, 3.92
0.63
Prior event recovery (poor vs complete)
N/A
N/A
N/A
2.29
0.91, 5.81
0.080
15.38
2.19, 108.20
0.006
DMT before event
N/A
N/A
N/A
0.91
0.44, 1.85
0.79
0.62
0.23, 1.64
0.33
*There were not enough second or third events in the cerebrum to generate an odds ratio. OR ⫽ odds ratio; CI ⫽ confidence interval; BS/CE ⫽ brainstem/cerebellum; N/A ⫽ not applicable; DMT ⫽ disease-modifying therapy. 604
Neurology 72
February 17, 2009
Table 3
Multivariate predictors of increased first, second, and third event severity First event
Second event p
Third event
Predictor
OR
95% CI
OR
95% CI
p
OR
95% CI
p
Nonwhite race/ethnicity
1.59
0.85, 2.99
0.15
2.05
0.92, 4.59
0.081
1.18
0.23, 6.00
0.84
Age (10-year decrease)
1.27
1.02, 1.56
0.034
1.14
0.83, 1.56
0.43
1.30
0.63, 2.70
0.47
Spinal cord
1.67
0.71, 3.91
0.24
1.40
0.50, 3.95
0.52
0.82
0.11, 6.08
0.85
BS/CE
7.13
2.96, 17.18
Optic nerve
2.05
0.89, 4.68
⬍0.0001
2.46
0.97, 6.22
0.057
1.17
0.19, 7.16
0.87
0.090
1.41
0.43, 4.66
0.57
1.56
0.14, 17.80
0.72
Cerebrum*
12.99
1.33, 126.76
0.027
N/A
N/A
N/A
N/A
N/A
N/A
Prior event severity (moderate vs mild)
N/A
N/A
N/A
2.75
1.24, 6.11
0.013
3.24
0.72, 14.67
0.13
Prior event severity (severe vs mild)
N/A
N/A
N/A
4.74
1.79, 12.56
0.002
2.89
0.42, 19.92
0.28
Prior event recovery (fair vs complete)
N/A
N/A
N/A
0.56
0.26, 1.19
0.13
0.94
0.24, 3.63
0.93
Prior event recovery (poor vs complete)
N/A
N/A
N/A
1.35
0.49, 3.75
0.56
N/A
N/A
N/A
DMT before event
N/A
N/A
N/A
0.97
0.43, 2.18
0.93
0.69
0.19, 2.49
0.57
*There were not enough second or third events in the cerebrum to generate an odds ratio. OR ⫽ odds ratio; CI ⫽ confidence interval; BS/CE ⫽ brainstem/cerebellum; N/A ⫽ not applicable; DMT ⫽ disease-modifying therapy.
ferent than in the univariate analyses for nonwhite race, optic nerve or brainstem/cerebellar involvement, IDE severity, fair vs complete IDE recovery, and DMT (table 3), but there was an attenuation of the association of age and of poor vs complete IDE recovery with second event severity. Spinal cord involvement of the second event did not appear to be meaningfully associated with the event’s severity (table 3). The multivariate model for third event severity included age, race, location, second event severity and recovery, and DMT. The measures of association for race, age, location, and DMT were similar to those in the univariate analyses. There was still a large OR for a more severe third event if the prior event was moderate (OR ⫽ 3.24, 95% confidence interval [CI] [0.72, 14.67], p ⫽ 0.13) or severe (OR ⫽ 2.89, 95% CI [0.42, 19.92], p ⫽ 0.28) vs mild. The OR associated with fair vs complete second event recovery was not meaningfully different than in the univariate analysis but due to small numbers, an estimate could not be obtained for poor vs complete second event recovery. Factors associated with event recovery. Univariate analyses.
The most important predictors of relapse recovery were the concurrent event severity and the degree of recovery from the prior event (table 4). A moderate vs mild event was associated with a two- to threefold increase in the odds of poorer recovery of the first, second, and third events, while a severe vs mild event was associated with a 4- to 17-fold increase in the
odds of worse recovery, although the CIs surrounding the ORs for the second and third events are wide due to the lower number of patients who had them (table 1). Multivariate analyses. In the multivariate model for IDE recovery, which included age, race, event location, and IDE severity, IDE severity remained a strong predictor of worse recovery (OR for moderate vs mild severity ⫽ 2.38, 95% CI [1.38, 4.07], p ⫽ 0.002; OR for severe vs mild IDE ⫽ 5.08, 95% CI [2.42, 10.69], p ⬍ 0.0001) (table 5). Increased age was weakly associated with a worse recovery; nonwhite race remained an apparently unimportant predictor. IDE onset in the spinal cord, but not other locations, was associated with poorer recovery (OR ⫽ 2.85, 95% CI [1.30, 6.28], p ⫽ 0.009). The multivariate model for second event recovery included age, race/ethnicity, location, DMT, IDE recovery, and second event severity (table 5). The severity of the second relapse was an even stronger predictor of poor recovery than in the univariate analyses (OR for moderate vs mild event ⫽ 5.44, 95% CI [2.35, 12.59], p ⬍ 0.0001; OR for severe vs mild event ⫽ 9.21, 95% CI [2.99, 28.40], p ⫽ 0.0001). The ORs associated with age, IDE recovery, location, DMT, and race/ethnicity did not substantially differ from those obtained in the univariate models. The multivariate model for third event recovery included age, race/ethnicity, location, DMT, second event recovery, and third event severity. Only the Neurology 72
February 17, 2009
605
Table 4
Predictors (univariate) of poorer first, second, and third event recovery First event
Second event p
Third event
Predictor
OR
95% CI
OR
95% CI
OR
95% CI
p
Nonwhite race/ethnicity
1.17
0.69, 1.98
0.56
1.53
0.72, 3.28
p 0.27
2.35
0.75, 7.40
0.144
Age (10-year increase)
1.11
0.93, 1.34
0.25
1.12
0.83, 1.50
0.46
1.15
0.73, 1.82
0.55
Polyregional vs monoregional event
2.09
1.17, 3.75
0.013
0.85
0.33, 2.18
0.73
0.55
0.13, 2.28
0.41
Spinal cord
1.50
0.97, 2.34
0.069
1.34
0.68, 2.63
0.39
0.58
0.20, 1.71
0.32
BS/CE
0.97
0.61, 1.54
0.88
0.61
0.30, 1.23
0.17
1.11
0.38, 3.25
0.85
Optic nerve
0.97
0.61, 1.55
0.90
1.26
0.54, 2.91
0.59
1.09
0.28, 4.25
0.90
Cerebrum*
2.79
0.87, 8.99
0.086
N/A
N/A
N/A
N/A
N/A
N/A
Event severity (moderate vs mild)
2.08
1.27, 3.38
0.003
3.51
1.68, 7.36
0.0009
2.92
0.93, 9.17
0.067
Event severity (severe vs mild)
4.07
2.05, 8.09
⬍0.0001
7.61
2.78, 20.82
17.56
2.37, 130.18
0.005
Prior event recovery (fair vs complete)
N/A
N/A
N/A
0.94
0.46, 1.93
0.87
4.92
1.53, 15.82
0.008
Prior event recovery (poor vs complete)
N/A
N/A
N/A
5.28
1.95, 14.25
0.001
12.93
1.51, 110.65
0.019
DMT before event
N/A
N/A
N/A
1.15
0.53, 2.50
0.73
1.53
0.55, 4.24
0.41
⬍0.0001
*There were not enough second or third events in the cerebrum to generate an odds ratio. OR ⫽ odds ratio; CI ⫽ confidence interval; BS/CE ⫽ brainstem/cerebellum; N/A ⫽ not applicable; DMT ⫽ disease-modifying therapy.
latter two predictors appeared to be meaningful, but the CIs surrounding the ORs were large (table 5). Age, location, and nonwhite race did not appear to Table 5
substantially impact recovery. There was a trend for DMT to be associated with a worse recovery (OR ⫽ 2.54, 95% CI [0.65, 9.88], p ⫽ 0.18).
Multivariate predictors of poorer first, second, and third event recovery First event
Second event p
Third event
Predictor
OR
95% CI
OR
95% CI
OR
95% CI
p
Nonwhite race/ethnicity
1.03
0.55, 1.91
0.93
1.08
0.44, 2.68
p 0.87
1.09
0.20, 5.91
0.92
Age (10-year increase)
1.18
0.95, 1.46
0.13
1.40
0.99, 1.97
0.058
1.31
0.66, 2.63
0.44
Spinal cord
2.85
1.30, 6.28
0.009
1.18
0.37, 3.73
0.78
0.75
0.10, 5.68
0.78
BS/CE
1.51
0.68, 3.32
0.31
0.49
0.17, 1.38
0.18
1.70
0.22, 13.33
0.61
Optic nerve
1.50
0.70, 3.19
0.30
1.26
0.35, 4.58
0.72
0.49
0.04, 5.83
0.57
Cerebrum*
2.11
0.45, 10.01
0.35
N/A
N/A
Event severity (moderate vs mild)
2.38
1.38, 4.07
0.002
5.44
2.35, 12.59
⬍0.0001
N/A
N/A
N/A
N/A
3.23
0.78, 13.40
0.11
Event severity (severe vs mild)
5.08
2.42, 10.69
⬍0.0001
9.21
2.99, 28.40
0.0001
60.58
3.73, 984.97
0.004
Prior event recovery (fair vs complete)
N/A
N/A
N/A
1.20
0.54, 2.69
0.66
7.60
1.85, 31.22
0.005
Prior event recovery (poor vs complete)
N/A
N/A
N/A
4.94
1.70, 14.35
0.003
6.74
0.59, 76.47
0.12
DMT before event
N/A
N/A
N/A
1.08
0.44, 2.68
0.87
2.54
0.65, 9.88
0.18
*There were not enough second or third events in the cerebrum to generate an odds ratio. OR ⫽ odds ratio; CI ⫽ confidence interval; BS/CE ⫽ brainstem/cerebellum; N/A ⫽ not applicable; DMT ⫽ disease-modifying therapy. 606
Neurology 72
February 17, 2009
Our finding that individuals with MS inherently experience relatively similar severity of and recovery from relapses over time early in the disease course substantiates the concept that at the individual level, patients with MS may have predetermined disease features. This notion is supported by our previous reports that a given patient with RRMS is likely to have consecutive relapses in the same location within the nervous system and that there may be pathologic homogeneity within, but not between, individuals with MS.12,16,18 Whether stereotyped severity, recovery, and location of exacerbations is explained by genetic polymorphisms or other underlying biologic processes remains to be determined. Since poor recovery from early events predicts a worse long-term MS prognosis, interventions that modify this tendency toward more severe relapses with poor recovery are needed.1,3-5,7,9 Incomplete recovery may result from a more severe initial injury, such as axonal damage,19 or from limited repair processes. Studying relapse severity and recovery as secondary outcome measures in trials of neuroprotective agents may give insight into the responsible mechanisms. Whether current DMTs modify relapse severity or recovery is uncertain. DMT reduced the annual rate of moderate and severe exacerbations in one study.20 Here, DMT did not seem to attenuate relapse severity or recovery, although many patients did not receive treatment prior to the second or third attack. Also, this was not a randomized study, so patients who received DMT may have been different from those who did not. In previous studies, African Americans had more rapid disease progression or were more likely to be disabled than were Caucasians, suggesting that the long-term course of MS may be more aggressive in the former group, although predictors of such outcomes have not been fully characterized.21-24 Here, we demonstrate that nonwhite race/ethnicity was associated with a higher odds of more severe demyelinating events. We report elsewhere that nonwhites have a twofold increase of the risk of an early second demyelinating event.25 It is less clear if nonwhite race is predictive of poor recovery from events since the CIs are so wide. Extended follow-up of the cohort will help further define the long-term outcomes of this heterogeneous group of nonwhite patients. This study confirms previous reports that poor recovery of a first MS event is associated with severe presentation and polyregional onset.7,10,11 We expand these findings by showing that the severity of the second and third events also predicts their recovery. Younger patients appeared more likely to have a DISCUSSION
more severe IDE, and there was a trend for younger patients to experience better recovery of the first and second events. These results should be confirmed in a larger cohort. If the differential effect of age exists, perhaps younger patients have attacks associated with more edema such that symptoms are worse at their peak but also resolve more completely as the edema subsides. Younger patients may also have more plasticity and therefore better repair. The effect of location, as evidenced in the multivariate models, was complex and requires further study in a larger cohort. Onset in the spinal cord did not appear to meaningfully influence relapse severity but was predictive of poor recovery of the IDE. Conversely, onset in the other three locations was associated with increased severity of the IDE (with a trend for the same when the second event involved the brainstem/cerebellum) but did not seem to meaningfully predict recovery. These data suggest that inflammation and repair processes are different in some CNS locations. There are some limitations to our study. Those who were determining severity and recovery of attacks were not blinded to the assignations for previous attacks. While this could raise questions of whether misclassification bias was introduced, the fact that rigid definitions based on objective examination findings were used makes it unlikely. We used 2 to 12 months as the postexacerbation time interval in which we measured recovery, which may generate concern that some patients whose recovery was documented early in this period may have experienced continued recovery later in the interval that was not captured in this analysis. However, one study demonstrated that the proportion of individuals who had incomplete recovery after a relapse did not change when recovery was measured 30 to 59, 60 to 89, or 90 or more days after the exacerbation.26 These data suggest that there is little meaningful change in recovery over time when it is measured beyond 30 days postrelapse. Our definitions of severity imply that milder events are less likely to be followed by poor recovery. While the EDSS scales are nonlinear, potentially limiting the scoring of severity and recovery, this cannot explain the within-patient tendency to experience similar severity or recovery for their first three MS exacerbations. The sample sizes for the second and particularly the third event are small, resulting in widened CIs for predictors of these events’ severity and recovery. However, for the main predictors of interest, the point estimates were generally in the same direction as for the IDE model, and most of the wider CIs were similar to or encompassed those of the IDE models, indicating biologic consistency that bolsters the credibility of the findings. Finally, long-term outcomes and CNS imaging with quantiNeurology 72
February 17, 2009
607
tative measures of injury and repair are not available for this cohort but could add to our understanding of the pathophysiologic processes at play. Patients with a more severe presentation and poor recovery at MS onset have an inherent tendency to continue on a similar trajectory for subsequent events. It is unclear if earlier or more aggressive treatment in patients with such poor short-term prognosis limits the long-term accrual of disability. We are currently investigating if specific genetic polymorphisms are associated with severity of and recovery from demyelinating events.
12.
13.
14.
15. ACKNOWLEDGMENT The authors thank the neurologists of the UCSF MS center for their participation in data collection.
16.
Received August 25, 2008. Accepted in final form November 10, 2008. REFERENCES 1. Confavreux C, Vukusic S, Adeleine P. Early clinical predictors and progression of irreversible disability in multiple sclerosis: an amnesic process. Brain 2003;126: 770–782. 2. Pittock SJ, Mayr WT, McClelland RL, et al. Change in MS-related disability in a population based cohort: a 10-year follow-up study. Neurology 2004;62:51–59. 3. Phadke JG. Clinical aspects of multiple sclerosis in northeast Scotland with particular reference to its course and prognosis. Brain 1990;113:1597–1628. 4. Weinshenker BG, Rice GP, Noseworthy JH, Carriere W, Baskerville J, Ebers GC. The natural history of multiple sclerosis: a geographically based study: 3: multivariate analysis of predictive factors and models of outcome. Brain 1991;114:1045–1056. 5. Amato MP, Ponziani G, Bartolozzi ML, Siracusa G. A prospective study on the natural history of multiple sclerosis: clues to the conduct and interpretation of clinical trials. J Neurol Sci 1999;168:96–106. 6. Scott TF, Schramke CJ, Novero J, Chieffe C. Short-term prognosis in early relapsing remitting multiple sclerosis. Neurology 2000;55:689–693. 7. Runmarker B, Andersen O. Prognostic factors in a multiple sclerosis incidence cohort with twenty-five years of follow-up. Brain 1993;116:117–134. 8. Kurtzke JF, Beebe GW, Nagler B, Kurland LT, Auth TL. Studies on the natural history of multiple sclerosis: 8: early prognostic features of the later course of the illness J Chron Dis 1977;30:819–830. 9. Amato MP, Ponziani G. A prospective study on the prognosis of multiple sclerosis. Neurol Sci 2000;21:S831– S838. 10. West T, Wyatt M, High A, Bostrom A, Waubant E. Are initial demyelinating event recovery and time to second event under differential control? Neurology 2006;67:809– 813. 11. Beck RW, Cleary PA, Anderson MM, et al. A randomized, controlled trial of corticosteroids in the treatment
608
Neurology 72
February 17, 2009
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
of acute optic neuritis. N Engl J Med 1992;326:581– 588. Deen S, Bacchetti P, High A, Waubant E. Predictors of the location of multiple sclerosis relapse. J Neurol Neurosurg Psychiatry 2008;79:1190–1193. Polman CH, Reingold SG, Edan G, et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald criteria.” Ann Neurol 2005;58:840–846. Comi G, Filippi M, Wolinsky J. European/Canadian multicenter, double-blind, randomized, placebo-controlled study of the effects of glatiramer acetate on magnetic resonance imaging–measured disease activity and burden in patients with relapsing multiple sclerosis. Ann Neurol 2001;49:290–297. Waubant E, Goodkin DE, Sloan R, Andersson PB. A pilot study of MRI activity before and during interferon beta-1a therapy. Neurology 1999;53:874–876. Mowry EM, Deen S, Malikova I, Pelletier J, Bacchetti P, Waubant E. The onset location of multiple sclerosis predicts the location of subsequent relapses. J Neurol Neurosurg Psychiatry Epub 2008 December 9. Panitch H, Goodin DS, Francis G, et al. EVIDENCE Study Group: randomized, comparative study of interferon beta-1a treatment regimens in MS: the EVIDENCE trial. Neurology 2002;59:1496–1506. Lucchinetti C, Bruck W, Parisi J, Scheithauer B, Rodriguez M, Lassman H. Heterogeneity of multiple sclerosis lesions: implications for the pathogenesis of demyelination. Ann Neurol 2000;47:707–717. Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mork S, Bo L. Axonal transection in the lesions of multiple sclerosis. N Engl J Med 1998;338:278–285. IFNB Multiple Sclerosis Study Group. Interferon beta-1b is effective in relapsing-remitting multiple sclerosis: I: clinical results of a multicenter, randomized, double-blind, placebo-controlled trial. Neurology 1993; 43:655–661. Cree BA, Khan O, Bourdette D, et al. Clinical characteristics of African Americans vs Caucasian Americans with multiple sclerosis. Neurology 2004;63:2039–2045. Kaufman MD, Johnson SK, Moyer D, Bivens J, Norton HJ. Multiple sclerosis: severity and progression rate in African Americans compared to whites. Am J Phys Med Rehab 2003;82:582–590. Marrie RA, Cutter G, Tyry T, Vollmer T, Campagnolo D. Does multiple sclerosis-associated disability differ between races? Neurology 2006;66:1235–1240. Weinstock-Guttman B, Jacobs LD, Brownscheidle CM, et al. Multiple sclerosis characteristics in African American patients in the New York State Multiple Sclerosis Consortium. Mult Scler 2003;9:293–298. Mowry EM, Pesic M, Deen SR, Grimes B, Bacchetti P, Waubant E. Age, race, and initial demyelinating events with fewer affected functional systems predict an increased hazard of early relapse in multiple sclerosis. Mult Scler 2008;14:P563. Abstract. Lublin FD, Baier M, Cutter G. Effect of relapses on development of residual deficits in multiple sclerosis. Neurology 2003;61: 1528–1532.
NGF, DCX, and NSE upregulation correlates with severity and outcome of head trauma in children A. Chiaretti, MD G. Barone, MD R. Riccardi, MD A. Antonelli, PhD P. Pezzotti, DStat O. Genovese, MD L. Tortorolo, MD G. Conti, MD
ABSTRACT
Background: Secondary brain damage after traumatic brain injury (TBI) involves neuroinflammatory mechanisms, mainly dependent on the intracerebral production of specific biomarkers, such as cytokines, neurotrophic factors, and neuron-specific enolase (NSE). NSE is associated with neuronal damage, while neurotrophic factors play a neuroprotective role due to their ability to modulate neuronal precursor biosynthesis, such as doublecortin (DCX). However, the relationships between the expression of these factors and the severity and outcome of TBI are not understood.
Methods: To determine whether the concentrations of neurotrophic factors (nerve growth factor Address correspondence and reprint requests to Dr. Antonio Chiaretti, Terapia Intensiva Pediatrica, Policlinico Gemelli, Largo Gemelli, 1-00168 Rome, Italy
[email protected]
[NGF], brain-derived neurotrophic factor [BDNF], glial-derived neurotrophic factor [GDNF]), DCX, and NSE in the CSF of children with TBI correlate with the severity of brain damage and neurologic outcome, we prospectively collected CSF samples from 32 children at 2 and 48 hours after admission for severe TBI and from 32 matched controls. Severity of TBI was evaluated by Glasgow Coma Scale and neurologic outcome by Glasgow Outcome Score.
Results: Early NGF, DCX, and NSE concentrations correlated significantly with the severity of head injury, whereas no correlation was found for BDNF and GDNF. Furthermore, NGF and DCX upregulation and lower NSE expression were associated with better neurologic outcomes. No significant association was found between BDNF and GDNF expression and outcome. Conclusions: Nerve growth factor (NGF), doublecortin (DCX), and neuron-specific enolase concentrations in the CSF are useful markers of brain damage following severe traumatic brain injury (TBI). NGF and DCX upregulation correlates also with better neurologic outcome and could be useful to obtain clinical and prognostic information in children with severe TBI. Neurology® 2009; 72:609–616 GLOSSARY BDNF ⫽ brain-derived neurotrophic factor; DCX ⫽ doublecortin; FIM ⫽ Functional Independence Measures; GCS ⫽ Glasgow Coma Scale; GDNF ⫽ glial-derived neurotrophic factor; GOS ⫽ Glasgow Outcome Score; NGF ⫽ nerve growth factor; NSE ⫽ neuron-specific enolase; PICU ⫽ pediatric intensive care unit; SGZ ⫽ subgranular zone; SVZ ⫽ subventricular zone; TBI ⫽ traumatic brain injury.
Severe traumatic brain injury (TBI) commonly causes impairment of brain functions and extensive neuronal loss. Neuronal cells require the action of some neurotrophic factors, such as nerve growth factor (NGF), brain derived neurotrophic factor (BDNF), and glial derived neurotrophic factor (GDNF), to stimulate growth and differentiation, and to promote recovery of injured brain neurons.1 TBI induces intracerebral neurotrophic factor biosynthesis, mainly of NGF, by activated neurons, microglia, and astrocytes.2 Recently, it has been proposed that changes in NGF levels in the CNS might be a neuroprotective response after TBI.3,4 In experimental brain injury, exogenous NGF administration prevents or significantly reduces severe neurologic deficits, apoptosis, and brain cell death.5 Genes and pathways involved in this neuroprotective role remain largely unknown, although recent studies have shown that NGF can influence neurogenesis and neuronal repair.6 NGF upregulation and intraventricular NGF Supplemental data at www.neurology.org From the Pediatric Intensive Care Unit (A.C., A.A., O.G., L.T., G.C.) and Pediatric Oncology (G.B., R.R.), Catholic University Medical School; and Agency for Public Health of Lazio Region (P.P.), Rome, Italy. Disclosure: The authors report no disclosures. Copyright © 2009 by AAN Enterprises, Inc.
609
Figure 1
Scatter plots and estimated linear regression of neurotrophic factors, doublecortin (DCX), and neuron-specific enolase (NSE) concentrations measured in the CSF at T1 with Glasgow Coma Scale (GCS) on admission
Nerve growth factor, DCX, and NSE levels correlated significantly with GCS scores. No significant association was found between brain-derived neurotrophic factor and glial-derived neurotrophic factor expression and severity of traumatic brain injury.
administration increase the number of new neurons generated in several forebrain structures.7,8 Neurogenesis is restricted to limited regions of the hippocampus, such as the subventricular zone (SVZ) and the subgranular zone (SGZ), where stem cells divide to form neuronal precursors, and then migrate into the damaged regions of the brain.9 Hypoxic610
Neurology 72
February 17, 2009
ischemic brain injury and TBI increase the production of new striatum neurons that express doublecortin (DCX).10 DCX contributes to neuronal repair by stabilizing microtubules in neuronal cells and represents a marker for tracking the migration of new neurons into the injured sites of the brain.11 An increase of DCX expression was recently demonstrated
Figure 2
Box plots of neurotrophic factors, doublecortin (DCX), and neuronspecific enolase (NSE) concentrations measured in the CSF at T1 sorted by category of Glasgow Outcome Score
Only nerve growth factor, DCX, and NSE expression was significantly associated with neurologic outcome of children.
following intraventricular BDNF and NGF administration.12-14 Nevertheless, the interactions between neurotrophins and DCX in children with TBI and their relationships with the severity of brain injury and clinical outcome are not understood. Conversely, neuron-specific enolase (NSE) is a specific biomarker of neuronal damage.15 Although raised NSE levels indicate neuronal destruction, its ability to quantify brain injury severity in children with TBI remains controversial.16 In this study we investigated the expression of NGF, BDNF, GDNF, DCX, and NSE in the CSF of children with TBI, in order to clarify
whether these factors can be useful in assessing clinical severity and neurologic outcome and whether these neurotrophins can modulate DCX biosynthesis in the injured brain. METHODS Population studied. We conducted a prospective observational clinical study in all children with severe TBI (Glasgow Coma Scale [GCS] scores ⱕ8) consecutively admitted to our pediatric intensive care unit (PICU) and referred within 4 hours from injury. No child with severe TBI was excluded from the study. All patients were treated according to the international guidelines for the management of severe TBI (appendix e-1 on the Neurology® Web site at www.neurology.org). To measure NGF, BDNF, GDNF, DCX, and NSE levels at 2 hours (T1) and 48 hours (T2) after admission to the PICU, we collected CSF samples in a closed, sterile system using the same intraventricular catheter set for ICP monitoring. All CSF samples were centrifuged for 10 minutes at 5,000 rpm, and the supernatant was immediately stored at ⫺70°C until analysis. As controls, we used CSF samples collected from 32 children subjected to lumbar puncture in order to rule out meningitis suspected from clinical and laboratory findings (fever, mental state alteration, and acute phase reactant elevation). None of these patients showed any evidence of meningitis (as confirmed by negative cultures, chemical analysis, and lack of pleocytosis). TBI patients and control subjects were matched for age, sex, and weight. In children with TBI outcomes were assessed 6 months after injury using the Glasgow Outcome Score (GOS), which assigns a score of 1 to children who died, 2 to persistent vegetative state, 3 to severe neurologic deficits, 4 to mild neurologic deficits, and 5 to completely healthy children. Using the GOS, we decided to dichotomize the clinical outcome into two categories: poor outcome when the GOS was between 1 and 3, and good outcome when the GOS was 4 or 5. Although GOS is a global and nonspecific clinical score in infants and children with severe TBI in respect to other more fine disability scales, such as Functional Independence Measures (FIM/WeeFIM),17,18 it is widely used in the assessment of late neurologic outcome in childhood morbidity after severe TBI, demonstrating a high sensitivity and specificity from acute injury predictors in this subset of patients.18,19 This study was approved by our University Ethical Committee. Parents of all children provided written informed consent. See appendix e-1 for neurotrophic factors, DCX, and NSE assays.
Statistical analysis. Patrizio Pezzotti, Doctor in Statistics at the Agency for Public Health of Lazio Region, Rome, Italy, conducted the statistical analysis. Mean concentrations of neurotrophic factors, DCX, and NSE measured in the CSF of patients were compared with controls using the Mann-Whitney twotailed test. T1 mean concentrations of neurotrophic factors, DCX, and NSE measured in the CSF of patients were compared with T2 concentrations using the Wilcoxon test. To evaluate the association of neurotrophic factors, DCX, and NSE concentrations with the GCS scores assessed upon admission, we used bidimensional scatter plots on which the estimated linear regression line was superimposed. Spearman correlation was determined and the p value was calculated to evaluate if this was significantly different from zero. To evaluate the association of neurotrophic factors, DCX, and NSE concentrations with GOS assessed at 6 months after TBI, we constructed box plots stratified by class of GOS (GOS ⬍3 ⫽ poor outcome; GOS ⬎3 ⫽ Neurology 72
February 17, 2009
611
Figure 3
Box plots of neurotrophic factors, doublecortin (DCX), and neuronspecific enolase (NSE) concentrations measured in the CSF at T2 sorted by category of Glasgow Outcome Score
Thirtytwo children with severe TBI (17 boys and 15 girls) and with a median age of 7.6 years (range: 1.3–15.6 years) were admitted to our PICU (table e-1). All children had suffered an isolated head injury caused by a motor vehicle accident (n ⫽ 26), fall (n ⫽ 4), or domestic accident (n ⫽ 2). GOS was 1 in 7 patients, 2 in 1 patient, 3 in 6 patients, 4 in 6 patients, and 5 in 12 patients. RESULTS Demographic and clinical data.
Neurotrophic factors, DCX, and NSE expression in patients and in controls. CSF concentrations of NGF,
NSE, and DCX measured 2 hours after admission (figure e-1) were dramatically increased in patients with TBI compared to controls, by 22-, 42- and 9-fold (p ⬍ 0.01) (NGF: 406.44 ⫾ 23.12 compared with 18.13 ⫾ 2.14; NSE: 90.84 ⫾ 7.12 compared with 2.12 ⫾ 0.3; DCX: 0.87 ⫾ 0.1 compared with 0.1). Significant, albeit much smaller, differences were measured also for BDNF, which increased fivefold in patients with TBI compared to controls (p ⬍ 0.01) (BDNF: 31.16 ⫾ 3.32 compared with 6.44 ⫾ 1.55). Conversely, no significant differences were found in GDNF concentration between patients and controls (p ⫽ 0.29). A further increase was noted in the concentrations of NGF, NSE, and DCX in TBI patients at 48 hours postadmission, whereas BDNF and GDNF concentrations decreased from T1 to T2 (figure e-2).
NSE overexpression was significantly associated with poor neurologic outcome, while nerve growth factor and DCX upregulation was significantly associated with a better clinical outcome of patients.
Correlation with severity of TBI. The scatter plots in figure 1 show the linear regression analysis of neurotrophic factors, DCX, and NSE concentrations measured at T1 and GCS scores recorded upon admission to the PICU. NGF, DCX, and NSE concentrations correlated significantly with GCS scores (p ⬍ 0.01) (Spearman correlation ⫽ ⫺0.83, ⫺0.70, and ⫺0.90). In contrast, no significant correlation was found between BDNF and GDNF (Spearman correlation ⫽ ⫺0.17; p ⫽ 0.36, and ⫺0.14; p ⫽ 0.29) concentrations and GCS scores. Correlation with outcome. The box plots in figures
good outcome). The choice of dichotomizing GOS was done by convenience given that for some scores there were very few observations (i.e., the score GOS 2 was reported only for one patient). To evaluate if there was a significant difference between the two outcome groups for any of the considered variables, we used the Wilcoxon test. Furthermore, several bivariate ordinal logistic regression analyses were performed to test the association of GOS with each parameter and simultaneously adjusting for GCS. A multiple ordinal logistic regression model adjusting simultaneously for GCS and all CSF concentrations at time T1 and T2 and its changes was also performed. However, because of the small sample size available, this model did not reach the convergence criterion. A p value ⬍0.05 was considered significant. Statistical analysis of the data was performed using Stata 8.0 software (Stata Corporation, College Station, TX). 612
Neurology 72
February 17, 2009
2, 3, and 4 illustrate the concentrations of neurotrophic factors, DCX, and NSE in the CSF measured at T1 and T2, as well as their differential (⌬T2 ⫺ T1), sorted by GOS category. NGF, DCX, and NSE concentrations measured at T1 were significantly lower in patients with better neurologic outcomes (p ⬍ 0.01; figure 2). However, these patients also had significantly higher NGF (p ⬍ 0.01) and DCX (p ⬍ 0.01) together with lower NSE (p ⬍ 0.01) concentrations measured at T2 (figure 3). When evaluating the differentials between T1 and T2 (figure 4), we found that patients with better neurologic outcomes had lower NSE levels (p ⬍ 0.01), but also significantly stronger NGF and DCX upregulation (p ⬍ 0.01). No significant correlation was found between
Figure 4
Box plots of neurotrophic factors, doublecortin (DCX), and neuronspecific enolase (NSE) changes between T1 and T2 sorted by category of Glasgow Outcome Score (GOS)
expression in one patient with NGF upregulation from time T1 to time T2. We observed the same DCX behavior in all patients with NGF upregulation, while in other patients and in controls the DCX expression remained at the basal levels (data not shown). To determine the interrelationship between the neurotrophic factor expression and DCX biosynthesis we also evaluated the correlation of their levels in the CSF of patients. Figure e-3 shows the correlation between NGF and BDNF expression and DCX biosynthesis at different time points after TBI. While NGF is significantly associated with DCX biosynthesis both at time T1 and at time T2, the BDNF correlated significantly with DCX only at 48 hours after TBI (time T2). No significant association was found between GDNF expression and DCX biosynthesis (data not shown). This study shows that children with severe TBI develop a marked and early increase in CSF concentrations of NGF, DCX, and NSE with a smaller but significant increase also in BDNF, but not in GDNF levels. Whereas BDNF and GDNF tend to decline, NGF, DCX, and NSE increase further by 48 hours post-trauma, confirming that these molecular markers are involved in the neurobiological response of the brain to injury. The increased levels of NGF, DCX, and NSE soon after head trauma correlated significantly with the severity of neurologic compromise assessed upon admission, while no significant correlation was found between BDNF and GDNF concentrations and GCS scores. Also, and more meaningfully, concurrent strong upregulation of NGF and DCX, relative to low NSE expression soon after injury, seems to predict a better neurologic outcome. The better outcome in this subset of patients might be due to the action of both NGF and BDNF in modulating DCX biosynthesis into the injured brain in the early phases after traumatic insult. Some specific biomarkers of brain injury, such as S100B, myelin basic protein, and NSE have been associated with poorer outcome in children with severe TBI.20 NSE is an intracellular enzyme detected both in the serum and in the CSF following structural damage of neuronal cells, which appears to be a good marker for the severity of intracranial injury.21 In our study, the increase in NSE levels was observed soon after brain injury (time T1) and NSE expression was followed by a sustained and delayed peak after 48 hours postinjury (time T2), as previously reported in the literature.22 This time course of NSE expression in the CSF of TBI patients is characterized by both an early and a late peak, presumably representing two waves of neuronal death, the second of which may represent neuronal apoptoDISCUSSION
NSE upregulation was significantly associated with poor neurologic outcome (GOS ⱕ3), whereas nerve growth factor and DCX upregulation was significantly associated with favorable neurologic outcome (GOS ⬎3) of children with severe traumatic brain injury.
BDNF and GDNF expression and GOS scores. In bivariate ordinal logistic regressions where the outcome was the GOS, the association of CSF concentrations was evaluated adjusting by GCS (table e-2); at time T1, NGF and NSE concentrations were significantly associated with GOS; at time T2, only NSE level was significantly associated with GOS; when considering the change of CSF concentrations between time T1 and T2, NGF, DCX, and NSE changes were significantly associated after having been adjusted by GCS. Correlation between neurotrophic factor expression and DCX biosynthesis. Figure 5 reports an example of
Western blot panels showing the increase of DCX
Neurology 72
February 17, 2009
613
Figure 5
Western blot analysis for doublecortin (DCX) in the CSF of patient 1 in the first 48 hours after admission
The figure shows a significant increase in DCX expression in this patient concomitantly to NGF upregulation from time T1 to time T2. The figure also shows the negative and positive control (Cn, Cp) run in the same blot with patients’ samples. All bands are normalized with a housekeeping gene as GAPDH.
sis.22 The association between increased NSE expression and poor outcome suggests that the intracerebral production of this biomarker may indicate the extent of neuronal damage in patients with severe TBI.22 Our results are in keeping with findings showing that NSE upregulation is significantly associated both with the severity of TBI and poorer outcomes of brain injured patients.23-25 TBI determines also an increased production of neurotrophic factors in the injured brain soon after traumatic insult.4,26 Our results are consistent with the timing of neurotrophic factor expression in the CSF of patients with head trauma, suggesting that this process lies upstream from the complex cascade of neuroreactive mechanisms after TBI.2,26,27 In our patients the expression of NGF, BDNF, and GDNF was differentially modulated, suggesting that these factors might act on different target cells and serve different functions in post-traumatic injury. In particular, NGF upregulation was significantly stronger than the changes in BDNF and GDNF expression and correlated significantly with both clinical severity and neurologic outcomes of patients. However, due to the lack of studies on NGF expression in children with TBI, the specific role of this neurotrophin on the outcome of patients often remains undetected, although NGF exerts a neuroprotective action in brain injured patients.2-4,26 In our study, children with severe TBI showed an early NGF upregulation that reveals a significant correlation both with the severity of head injury and with patient’s outcome. Interestingly, patients with a poor outcome (GOS ⱕ3), mainly children with brain death, showed a flattening of the NGF curve in the CSF and decrease of NGF expression from time T1 to time T2. In contrast, children with a higher GOS had an NGF increase at time T2 and this upregulation shows a significant correlation with a good patient outcome. In these patients, the NGF levels increase strongly and this upregulation appears to be enhanced in the early phases of traumatic injury, in which NGF biosynthesis can be 614
Neurology 72
February 17, 2009
stimulated by activated neurons, microglia, and astrocytes, involved in the intracerebral inflammatory host response after TBI.28 Cellular brain responses to NGF are elicited through binding and activation of its receptors, TrkA (high affinity receptor) and p75 (low affinity receptor), located in the hippocampus, the SVZ, the SGZ, the nucleus basalis of Meynert, and the cerebral cortex.29 NGF upregulation in neurotrauma patients may have a beneficial impact on the regenerative capacity of the injured brain via stimulation of TrkA, which influences cellular processes such as proliferation, growth, differentiation, and other cell-specific functions, as well as regeneration.30 On the other hand, NGF does lead to positive trophic effects, including the nicotinic receptor upregulation in the brain, significantly improving cortical blood flow and counteracting post-traumatic cerebral ischemia and excitotoxicity process, the two primary mechanisms of neuronal death following TBI.31,32 Recently, it has also been reported that intraventricular NGF administration improves neuronal survival in infants with hypoxic-ischemic brain injury and the neuroprotective role of NGFtransfected stem cells after experimental TBI.13,33,34 However, little is known about the role of NGF in determining the mechanisms of neuronal repair in the damaged brain. New striatum neurons can generate when neurotrophic factors are infused in the lateral ventricle of the brain and increased levels of intrastriatal neurotrophins promote the initial phase of induced neurogenesis with a considerable increase in the number of DCX-immunoreactive migrating neuroblasts.12,13,35 Our study shows that there is a significant correlation between NGF and BDNF upregulation and DCX expression in the CSF of children with severe TBI and that increased levels of both NGF and DCX correlate significantly with a good outcome for these patients. DCX appears in the CSF soon after traumatic insult, raising the possibility that the release of this marker may have a pivotal role in neuronal connection reorganization after brain injury. From this point of view, DCX can be a useful marker for the assessment of clinical severity and a possible screening tool for the outcome of patients with TBI and other CNS dysfunctions.14,36 On the other hand, previous experimental studies have shown that intracerebral neurotrophin administration and transplantation of marrow stromal cells cultured with BDNF and NGF in rats subjected to TBI promote functional recovery by stimulating the biosynthesis of different neuronal markers, such as Neu N, MAP-2, and DCX.12,13,34,37,38 DCX-positive cells were present near and among the glial scars early after brain injury migrating from SVZ toward cerebral cortex lesions changing from immature to mature
neurons.39,40 According to those experimental studies, we observed that the interrelationship between NGF expression and DCX biosynthesis in the CSF of children with TBI increased significantly in the early phase after injury. The correlation between NGF and DCX suggests that both these markers are involved in the neurobiological response of the brain to injury and that NGF, together with BDNF, probably acts by modulating DCX biosynthesis in the brain, reflecting an endogenous attempt against the biochemical cascades triggered by traumatic insult.14 Limitations of the study. Because of the low sample
size and the high correlation with GCS, it remains difficult to understand whether these molecular mechanisms can provide an independent contribution to predict the outcome of children with severe TBI. As a consequence, no multiple regression models including simultaneously all CSF biomarker concentrations together with GCS and their interaction terms were performed to better disentangle the independent effect of these biomarkers with GOS. Moreover, our data may not be generalizable to all children with severe TBI, and to minimize the selection bias of patients clear guidelines for the assessment and treatment must be used. Also, the difficulties with handling and stocking CSF samples and the limitations and difficulties of the techniques used to determine CSF biomarkers may limit the development of this type of study, but defining the relationships between NGF and DCX expression in the injured brain might allow individualization of new strategies for the treatment of TBI patients. Received June 17, 2008. Accepted in final form November 6, 2008.
REFERENCES 1. Barde YA. Neurotrophins: a family of proteins supporting the survival of neurons. Prog Clin Biol Res 1994;390: 1855–1859. 2. DeKosky ST, Goss JR, Miller PD, et al. Up-regulation of nerve growth factor following cortical trauma. Exp Neurol 1994;130:173–177. 3. Chiaretti A, Antonelli A, Riccardi R, et al. Nerve growth factor expression correlates with severity and outcome of traumatic brain injury in children. Eur J Paediatr Neurol 2008;12:195–204. 4. Chiaretti A, Antonelli A, Mastrangelo A, et al. Interleukin-6 and nerve growth factor upregulation correlates with improved outcome in children with severe traumatic brain injury. J Neurotrauma 2008;25:225–234. 5. Kromer LF. Nerve growth factor treatment after brain injury prevents neuronal death. Science 1987;235:214–216. 6. Sofroniew MV, Howe CL, Mobley W. Nerve growth factor signalling, neuroprotection and neural repair. Ann Rev Neurosci 2001;24:1217–1281. 7. Frielingsdorf H, Simpson DR, Thal LJ, Pizzo DP. Nerve growth factor promotes survival of new neurons in the adult hippocampus. Neurobiol Dis 2007;26:47–55.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Chiaretti A, Genovese O, Riccardi R, et al. Intraventricular nerve growth factor infusion: a possible treatment for neurological deficits following hypoxic-ischemic brain injury in infants. Neurol Res 2005;27:741–746. Collin T, Arvidsson A, Kokaia Z, et al. Quantitative analysis of the generation of different striatal neuronal subtypes in the adult brain following excitotoxic injury. Exp Neurol 2005;195:1–80. Ong J, Plane JM, Parent JM, et al. Hypoxic-ischemic injury stimulates subventricular zone proliferation and neurogenesis in the neonatal rat. Pediatr Res 2005;58:600–606. Horesh D, Sapir T, Francis F, et al. Doublecortin, a stabilizer of microtubules. Hum Mol Genet 1999;8:1599–1610. Zigova T, Pencea V, Wiegand SJ, et al. Intraventricular administration of BDNF increases the number of newly generated neurons in the adult olfactory bulb. Mol Cell Neurosci 1998;11:234–245. Chiaretti A, Antonelli A, Genovese O, et al. Intraventricular nerve growth factor infusion improves cerebral blood flow and stimulates doublecortin expression in two infants with hypoxic-ischemic brain injury. Neurol Res 2008;30: 223–228. Chiaretti A, Antonelli A, Genovese O, et al. Nerve growth factor and doublecortin expression correlates with improved outcome in children with severe traumatic brain injury. J Trauma 2008;65:80–85. Marangos PJ, Schmechel DE. Neuron specific enolase, a clinically useful marker for neurons and neuroendocrine cells. Annu Rev Neurosci 1987;10:269–295. Vos PE, Lamers KJ, Hendriks JC, et al. Glial and neuronal proteins in serum predict outcome after severe traumatic brain injury. Neurology 2004;62:1303–1310. Wong V, Wong S, Chan K, Wong W. Functional Independence Measure (WeeFIM) for Chinese children: Hong Kong Cohort. Pediatrics 2002;109:E36. Robertson CM, Watt JM, Joffe AR, et al. Childhood morbidity after severe traumatic brain injury: Increased detection with the Multiattribute Health Status Classification. Pediatr Crit Care Med 2001;2:145–150. Ducrocq SC, Meyer PG, Orliaguet GA, et al. Epidemiology and early predictive factors of mortality and outcome in children with traumatic severe brain injury: experience of a French pediatric trauma center. Pediatr Crit Care Med 2006;7:461–467. Berger RP, Adelson PD, Pierce MC, et al. Serum neuronspecific enolase, S100B, and myelin basic protein concentrations after inflicted and noninflicted traumatic brain injury in children. J Neurosurg 2005;103:61–68. Bandyopadhyay S, Hennes H, Gorelick MH, et al. Serum neuron-specific enolase as a predictor of short-term outcome in children with closed traumatic brain injury. Acad Emerg Med 2005;12:732–738. Berger RP, Pierce MC, Wisniewski SR, et al. Neuronspecific enolase and S100B in cerebrospinal fluid after severe traumatic brain injury in infants and children. Pediatrics 2002;109:E31. Fridriksson T, Kini N, Walsh-Kelly C, et al. Serum neuron-specific enolase as a predictor of intracranial lesions in children with head trauma: a pilot study. Acad Emerg Med 2000;7:816–820. Varma S, Janesko KL, Wisniewski SR, et al. F2isoprostane and neuron-specific enolase in cerebrospinal fluid after severe traumatic brain injury in infants and children. J Neurotrauma 2003;20:781–786.
Neurology 72
February 17, 2009
615
25.
Celtik C, Acunas¸ B, Oner N, et al. Neuron-specific enolase as a marker of the severity and outcome of hypoxic ischemic encephalopathy. Brain Dev 2004;26:398–402. 26. Chiaretti A, Piastra M, Polidori G, et al. Correlation between neurotrophic factor expression and outcome of children with severe traumatic brain injury. Intens Care Med 2003;29:1329–1338. 27. Longo FM, Selak I, Zovickian J, et al. Neuronotrophic activities in cerebrospinal fluid of head trauma patients. Exp Neurol 1984;84:207–218. 28. Kossmann T, Hans V, Imhof HG, et al. Interleukin-6 released in human cerebrospinal fluid following traumatic brain injury may trigger nerve growth factor production in astrocytes. Brain Res 1996;25:143–152. 29. Bothwell M. Functional interactions of neurotrophins and neurotrophin receptors. Annu Rev Neurosci 1995;18:223–253. 30. Cui Q. Actions of neurotrophic factors and their signaling pathways in neuronal survival and axonal regeneration. Mol Neurobiol 2006;33:155–179. 31. Cheng B, Mattson MP. NT-3 and NGF protect CNS neurons against metabolic/excitotoxic insults. Brain Res 1994; 640:56–67. 32. Holtzman DM, Sheldon RA, Jaffe W, Cheng Y, Ferriero DM. Nerve growth factor protects the neonatal brain against hypoxic-ischemic injury. Ann Neurol 1996;39: 114–122. 33. Calza L, Giuliani A, Fernandez M, et al. Neural stem cells and cholinergic neurons: regulation by immunolesion and treatment with mitogens, retinoic acid, and nerve growth factor. Proc Natl Acad Sci 2003;100:7325–7330.
616
Neurology 72
February 17, 2009
34.
35.
36.
37.
38.
39.
40.
Mahmood A, Lu D, Wang L, et al. Intracerebral transplantation of marrow stromal cells cultured with neurotrophic factors promotes functional recovery in adult rats subjected to traumatic brain injury. J Neurotrauma 2002;19:1609– 1617. Gustafsson E, Andsberg G, Darsalia V, et al. Anterograde delivery of brain-derived neurotrophic factor to striatum via nigral transduction of recombinant adeno-associated virus increases neuronal death but promotes neurogenic response following stroke. Eur J Neurosci 2003;17:2667– 2678. Winner B, Couillard-Despres S, Geyer M, et al. Dopaminergic lesion enhances growth factor-induced striatal neuroblast migration. J Neuropathol Exp Neurol 2008;67: 105–116. Mahmood A, Lu D, Chopp M. Marrow stromal cell transplantation after traumatic brain injury promotes cellular proliferation within the brain. Neurosurgery 2004;55: 1185–1193. Collin T, Arvidsson A, Kokaia Z, Lindvall O. Quantitative analysis of the generation of different striatal neuronal subtypes in the adult brain following excitotoxic injury. Exp Neurol 2005;195:71–80. Sundholm-Peters NL, Yang HK, Goings GE, et al. Subventricular zone neuroblasts emigrate toward cortical lesions. J Neuropathol Exp Neurol 2005;64:1089–1100. Itoh T, Satou T, Nishida S, et al. Immature and mature neurons coexist among glial scars after rat traumatic brain injury. Neurol Res 2007;29:734–742.
A novel Frabin (FGD4) nonsense mutation p.R275X associated with phenotypic variability in CMT4H Henry Houlden, MD, PhD Simon Hammans, MD Haider Katifi, MD Mary M. Reilly, MD
Address correspondence and reprint requests to Dr. Henry Houlden, Institute of Neurology, Queen Square, London, UK WC1N 3BG
[email protected]
ABSTRACT
Background: Charcot Marie Tooth (CMT) disease is a heterogeneous group of inherited peripheral motor and sensory neuropathies. CMT4H is an early onset autosomal recessive demyelinating neuropathy. The locus responsible for CMT4H was assigned to chromosome 12p11.21-q13.11 by homozygosity mapping and mutations in the Frabin gene (FGD4 Rho GDP/GTP exchange factor) were subsequently identified in six families.
Methods: We sequenced the Frabin gene in a cohort of 12 UK CMT families with clinically defined autosomal recessive demyelinating neuropathy.
Results: We identified a novel homozygous Frabin p.R275X mutation in a family from Northern Ireland. The two affected cases in this family had a very slowly progressive neuropathy with both cases remaining ambulant into middle age. Examination of mRNA from lymphoblasts showed that this stop mutation caused very little nonsense mediated mRNA decay and the predominant mRNA species was the mutant form that is likely to be translated into a truncated protein.
Conclusions: This work extends the understanding of the pathogenesis of Frabin mutationassociated Charcot Marie Tooth (CMT) 4H and suggests that mutations in Frabin should also be considered in ambulant adults with CMT1. Neurology® 2009;72:617–620 GLOSSARY AR ⫽ autosomal recessive; CMT ⫽ Charcot Marie Tooth; MCV ⫽ motor conduction velocity; MRC ⫽ Medical Research Council; NMD ⫽ nonsense mediated mRNA decay.
Charcot Marie Tooth (CMT) is clinically and genetically heterogeneous and forms the most common group of inherited neuromuscular disorders, with an estimated overall prevalence of 1 in 2,500 individuals.1 The autosomal recessive (AR) CMT phenotype is usually more severe and has an earlier onset than dominant CMT.2,3 The AR form may have additional clinical features such as scoliosis and cranial neuropathies that can give clinical clues to the genetic cause.4 There are at least 13 causative genes for ARCMT1 and this group is classified as CMT4.5,6 The CMT4H gene was recently identified as the Frabin gene on chromosome 127 in six families with early onset demyelinating CMT.8,9 Frabin is a GDP/GTP nucleotide exchange factor, a member of the Rho family of small GTP binding proteins, and plays a role in Cdc42-mediated cell shape changes.8,9 The expression of mutant Frabin (M298R)8 induced fewer microspikes in rat primary motoneurons and Schwann cells.10 These data along with mRNA expression studies suggest a loss of function disease mechanism for this particular mutation. The reported Frabin mutations cause a childhood onset moderately severe, progressive AR demyelinating CMT. There is associated scoliosis in two families and focally folded myelin on sural nerve biopsy.
From MRC Centre for Neuromuscular Disease and Department of Molecular Neurosciences (H.H., M.M.R.), The National Hospital for Neurology and Neurosurgery and The Institute of Neurology, Queen Square, London; and Wessex Neurological Centre (S.H., H.K.), Southampton General Hospital, UK. Supported by Medical Research Council’s clinician Scientist Fellowship (H.H.) as well as other grants, the Muscular Dystrophy Campaign, and the Brain Research Trust. This work was undertaken at University College London Hospitals/University College London, which received a proportion of funding from the Department of Health’s National Institute for Health Research Biomedical Research Centers funding scheme. Disclosure: The authors report no disclosures. Copyright © 2009 by AAN Enterprises, Inc.
617
Figure 1
A homozygous p.R275X, c.723C>T defect in the Frabin gene
(A) Chromatogram of the Frabin homozygous affected patient, unaffected heterozygous mother, and control. The R275X (CGA to TGA) mutation is indicated by an arrow. (B) Family tree. (C) Sequencing of Frabin exon 6 to 8 RT-PCR of the heterozygous case at 26, 28, and 30 cycles. Mutation is indicated by an arrow. The mutation is clearly present and not subject to significant nonsense decay and there is very little difference among 26, 28, and 30 cycles.
To assess the frequency of Frabin mutations in our CMT cohort, we analyzed 12 families, identifying one mildly affected Belfast kindred with a homozygous mutation. The effect of this mutation on lymphoblast mRNA was analyzed to assess the mechanism and possible genotype-phenotype associations.
turer’s instructions. Primers for Frabin were designed to span the introns around the mutation (introns 6 and 7); RT-PCR was amplified for 26, 28, and 30 cycles. Products were sequenced to determine the mutation and the ratio of the mutant base to the wild type base in the heterozygous carrier. RT-PCR reactions were carried out using GAPDH-primers as a housekeeper control gene. Equal volumes of these RT-PCR products were resolved on metaphor agarose gels and stained with SYBR green-1 and densitometry analysis of RT-PCR products was carried out.
We identified a homozygous p.R275X, c.723C⬎T defect in the Frabin gene (figure 1, A and B). This stop mutation was present in the two affected individuals in the family (figure 1A) and heterozygous in the unaffected mother. The arginine at codon 275 was highly conserved (figure 2C). Analysis of lymphoblast mRNA from one affected patient and the unaffected mother showed that the mutation was present in mRNA and not decayed. In the heterozygous mother, the mutant
RESULTS Genetic analyses. METHODS All patients gave informed consent and ethics approval was obtained from the ethics committee at The National Hospital for Neurology and Neurosurgery.
Molecular genetics. The Frabin gene was amplified and sequenced by standard methods, primers available on request. Mutations were confirmed by repeat sequencing where the p.R275X mutation was not identified in 190 control chromosomes. Total RNA was isolated from blood lymphoblasts from a homozygous affected and a heterozygous unaffected individual by the Trizol procedure (Invitrogen) according to the manufac618
Neurology 72
February 17, 2009
Figure 2
Genetic analysis of the Frabin gene
(A) Frabin RT-PCR 6F to 8R shows that there is no loss of density of the homozygous affected vs the unaffected heterozygous RNA expression of the Frabin gene. Paradoxically there is probably slightly greater total expression in the affected case. The GAPDH expression is not significantly different between the two cases. (B) Frabin 6F to 8R RT-PCR in affected, carrier, and controls. The top band is the Frabin transcript with exon 6, 7, and 8 and the lower band is exon 6 (5= region spliced out), exon 7, and exon 8. This splicing was confirmed with sequencing and the same in affected and controls. C ⫽ control; ⫹/⫹ ⫽ affected; ⫹/⫺ unaffected carrier; E ⫽ empty well; B ⫽ Brain cDNA. (C) Conservation of the mutated amino acid within other members of the Rho gene family as indicated in red.
T and the wild type C were both clearly seen at 26, 28, and 30 RT-PCR cycles (figure 1C). The T was not significantly below the C as compared to genomic sequencing. RT-PCR between exons 6 and 8 showed that the 5= region of exon 6 and the first base of exon 7 (89 bp) was spliced in or out in controls and the family and this was confirmed by sequencing (figure 2A). Products were analyzed on an agarose gel and by peak areas ratios with GAPDH and no difference was observed between the PCR product of the affected patient, unaffected mother, and controls (figure 2A). Clinical features and electrophysiology. The proband
in the family (figure 1) was assessed at age 58. She first had problems obtaining comfortable shoes as a young child and she found it difficult to run and had poor balance. She worked in a public house and her CMT progressed very slowly with occasional cramps and paraesthesia in her feet in her 30s. In her 50s she developed greater weakness distally and worsening sensory symptoms in her limbs and she remains ambulant with one stick. Examination at the age of 58 shows pupil size asymmetry, pes cavus, and wasted small hand and foot muscles. Power was decreased in
the hands (Medical Research Council [MRC] grade 4/5) and feet (ankle dorsiflexion and plantar flexion, MRC 4/5). She was areflexic with a glove and stocking sensory loss to the elbows and knees for pinprick and touch. Joint position sense was absent in the toes. Vibration sense was absent to the hips. Motor conduction velocities (MCVs) (age 49) were slow in the upper limbs with left ulnar MCV 13.0 m/s (distal motor latency 9.0 m/s, amplitude 0.4 mV) and left median MCV 8 m/s (latency 8.4 m/s, amplitude 2.3 mV). There were no recordable sensory responses. The patient’s brother had a similar phenotype. As a child he was clumsy and tended to drop things. He stopped work as a television technician at age 36 years due to lack of fine coordination in his hands. By age 50 he used two crutches or a wheelchair to mobilize. At 40 years, his gait was clumsy and he was unable to stand on his heels. He had pes cavus, claw feet, wasting, and weakness distally in the limbs. He was areflexic with distal reduction in all modalities of sensation. Limited nerve conduction study carried out at age 22 of the right median nerve showed a velocity of 6 m/s and latency 12.7 m/s (compound muscle action potential not documented). EMG of tibialis anterior showed a striking neurogenic pattern. DISCUSSION We report a further mutation in the Frabin gene that is associated with CMT. The mutation is a stop codon change (p.R275X, c.723C⬎T) segregating with the disease. The arginine 275 is highly conserved in species and other members of the Rho family. The previous two reports on mutations in the Frabin gene8,9 identified families with missense, premature stop as well as splice site mutations. The Frabin M298R mutation was shown to have an interesting mechanism; it is a splicing rather than a missense mutation and is predicted to result in a frameshift mutation (p.Met298fsX8), which explains the nonsense mediated mRNA decay (NMD) of the Frabin gene with this change.8 Interestingly, there was still some residual mutant mRNA leftover (40% of control), suggesting some abnormal mutated protein was still present. Analysis of mRNA extracted from lymphoblasts in our family showed the mutation was clearly present in the cDNA on sequencing of the proband and the unaffected heterozygous mother; affected mRNA was not degraded at 26, 28, and 30 cycles as would be expected with NMD. These data indicate that a truncated protein may be translated; however, a rapid protein degradation and resulting haploinsufficiency cannot be excluded at this stage of the investigation. Neurology 72
February 17, 2009
619
The phenotype in our family was less severe than the previously published families.8,9 This may be due to a genotype-phenotype effect although the position and type of the mutation is similar and in close proximity to three other reported stop mutations. Two other mutations causing a premature stop site (E543fs and G586X) are located toward the C-terminal end of the Frabin gene and these have a more severe phenotype.9 This indicates that the length of the Frabin protein truncation has little effect on predicting the phenotype and suggesting that multiple domains of this protein are important for Cdc42-mediated cell shape changes. This report extends the understanding of the pathogenesis of Frabin mutation-associated CMT4H and suggests that mutations in Frabin should also be considered in ambulant adults with CMT1. ACKNOWLEDGMENT The authors thank the families for their help.
Received August 11, 2008. Accepted in final form November 12, 2008. REFERENCES 1. Skre H. Genetic and Clinical Aspects of Charcot-MarieTooth’s Disease: Proceedings of the Third International Congress on Muscle Diseases: Excerpta Med Int Cong Series, No 334. Amsterdam: Excerpta Medica; 1974. 2. Harding AE, Thomas PK. Autosomal recessive forms of hereditary motor and sensory neuropathy. J Neurol Neurosurg Psychiatry 1980;43:669–678.
3. Thomas PK. Autosomal recessive hereditary motor and sensory neuropathy. Curr Opin Neurol 2000;13:565– 568. 4. Houlden H, King RH, Wood NW, Thomas PK, Reilly MM. Mutations in the 5= region of the myotubularinrelated protein 2 (MTMR2) gene in autosomal recessive hereditary neuropathy with focally folded myelin. Brain 2001;124:907–915. 5. Kabzinska D, Hausmanowa-Petrusewicz I, Kochanski A. Charcot-Marie-Tooth disorders with an autosomal recessive mode of inheritance. Clin Neuropathol 2008; 27:1–12. 6. Vallat JM, Grid D, Magdelaine C, Sturtz F, Levy N, Tazir M. [Autosomal recessive forms of Charcot-Marie-Tooth disease.] Bull Acad Natl Med 2005;189:55–68; discussion 68 – 69. 7. De Sandre-Giovannoli A, Delague V, Hamadouche T, et al. Homozygosity mapping of autosomal recessive demyelinating Charcot-Marie-Tooth neuropathy (CMT4H) to a novel locus on chromosome 12p11.21-q13.11. J Med Genet 2005;42:260–265. 8. Delague V, Jacquier A, Hamadouche T, et al. Mutations in FGD4 encoding the Rho GDP/GTP exchange factor FRABIN cause autosomal recessive Charcot-Marie-Tooth type 4H. Am J Hum Genet 2007;81:1–16. 9. Stendel C, Roos A, Deconinck T, et al. Peripheral nerve demyelination caused by a mutant Rho GTPase guanine nucleotide exchange factor, frabin/FGD4. Am J Hum Genet 2007;81:158–164. 10. Nakanishi H, Takai Y. Frabin and other related Cdc42specific guanine nucleotide exchange factors couple the actin cytoskeleton with the plasma membrane. J Cell Mol Med 2008;12:1169–1176.
Calling All Artists! Submit Your Art to Help Raise Money for Neurologic Research Are you an artist? The AAN Foundation invites you to donate your work to the Art for Research: An AAN Gallery Show. Pieces will be displayed at the Annual Meeting in Seattle and put on sale with proceeds going to support clinical research training in neuroscience. Academy members and/or their families may donate pieces for the show. The show accepts paintings, sculptures, textiles, ceramics, and more. Choose how to make your donations: ● Donate a piece of art for the Academy to sell at the meeting ● Sell a piece of art with 20% of the proceeds going to support research ● Submit your art for showcase only for a $50.00 fee For additional details on this event and to learn how to contribute, visit www.aan.com/art.
620
Neurology 72
February 17, 2009
B-type natriuretic peptide and cardiovalvulopathy in Parkinson disease with dopamine agonist H. Watanabe, MD M. Hirayama, MD A. Noda, PhD M. Ito, MD N. Atsuta, MD J. Senda, MD T. Kaga, MD A. Yamada, MD M. Katsuno, MD T. Niwa, MD F. Tanaka, MD G. Sobue, MD
Address correspondence and reprint requests to Dr. Gen Sobue, Department of Neurology, Nagoya University Graduate School of Medicine, Nagoya 466-8550, Japan
[email protected]
ABSTRACT
Objective: To elucidate the usefulness of plasma B-type natriuretic peptide (BNP) values for evaluating adverse effects of pergolide or cabergoline on cardiovalvulopathy in patients with Parkinson disease.
Methods: Twenty-five patients treated with pergolide or cabergoline (ergot group) and 25 patients never treated with ergot derivatives (non-ergot group) were enrolled. Plasma BNP values and detailed echocardiography were evaluated. Thirty age- and gender-matched controls were similarly evaluated.
Results: Patients with regurgitation more than grade 3 were more frequent in the ergot group than in the non-ergot group as well as control groups (24%, 0%, 3%, p ⫽ 0.001). Both composite regurgitation scores and plasma BNP values were significantly higher in the ergot group than in controls. In the ergot group, the cumulative dose correlated to both tenting area (r ⫽ 0.57, p ⫽ 0.004) and tenting distance (r ⫽ 0.62, p ⫽ 0.001). Furthermore, plasma BNP values were higher in patients with severe or multiple regurgitation groups (p ⬍ 0.001), and were correlated with composite regurgitation score (r ⫽ 0.70, p ⬍ 0.001). Multiple regression analyses revealed that BNP values were independently correlated with both composite regurgitation and left ventricular ejection fraction.
Conclusion: The combination of comprehensive echocardiography and plasma B-type natriuretic peptide levels elucidates the presence of cardiac damage in patients with Parkinson disease using ergot derivative dopamine agonists. Neurology® 2009;72:621–626 GLOSSARY AR ⫽ aortic regurgitation; BNP ⫽ B-type natriuretic peptide; MR ⫽ mitral regurgitation; PD ⫽ Parkinson disease; TR ⫽ tricuspid regurgitation; UPDRS ⫽ Unified Parkinson’s Disease Rating Scale.
Ergot derivative dopamine agonists including pergolide and cabergoline are some of the most effective drugs to treat parkinsonian symptoms and have the potential to reduce the motor complications observed in patients with Parkinson disease (PD) treated with L-dopa.1 However, several reports have shown an association of ergot derivative dopamine agonists and cardiac multivalvular regurgitation.2-18 In particular, high cumulative doses and long-term treatment with pergolide and cabergoline have been considered to be risk factors for increased valvulopathy in patients with PD. The US Food and Drug Administration public health advisory of March 29, 2007, cautions against abruptly stopping pergolide and is looking for ways to provide the drug to those people who cannot successfully switch to alternative treatments. On the contrary, over 60% of patients did not show valvulopathy, despite several years’ exposure.2 In Japan, similar to some European countries, pergolide and cabergoline are still used, and moderate doses of pergolide are associated with a low incidence of restrictive valvulopathy.4,9
From the Departments of Neurology (H.W., M.H., M.I., N.A., J.S., T.K., M.K., F.T., G.S.) and Cardiology (A.Y.), Nagoya University Graduate School of Medicine; and Departments of Medical Technology (A.N.) and Clinical Preventive Medicine (T.N.), Nagoya University School of Health Sciences, Japan. Supported by Health and Labor Sciences Research grants for research on measures for intractable diseases and comprehensive research on Aging and Health of the Ministry of Health, Labor and Welfare, Japan. Disclosure: The authors report no disclosures. Copyright © 2009 by AAN Enterprises, Inc.
621
Although echocardiography is an essential tool to evaluate valvulopathy, a simple screening test that does not require specialized techniques would be beneficial for management of patients under various conditions in particular institutes without a department of cardiovascular medicine. B-type natriuretic peptide (BNP), which is secreted mainly from the heart and belongs to the natriuretic peptide family, is indicative of cardiac dysfunction in patients with not only heart failure and coronary artery disease but also valvular disorders.19-21 Since plasma BNP values can be measured in serum by fully automated and commercially available assays with excellent test precision, it would be beneficial for monitoring the cardiac findings in patients with PD. In this study, we investigated the usefulness of plasma BNP values for identifying and monitoring cardiac involvement in patients with PD treated with ergot derivative dopamine agonists. METHODS The records of 121 patients with PD who attended the Department of Neurology, Nagoya University Hospital, and were nominated to our clinical cohort study at Nagoya area22 during January to December 2006 were investigated. Of these, 34 patients with PD who fulfilled probable PD criteria according to the established diagnostic criteria23 and were continuously taking ergot agonists (pergolide or cabergoline) but not non-ergot ones for a minimum of 1 year with or without levodopa were enrolled. Switching dopamine agonists between pergolide and cabergoline or combined use of pergolide and cabergoline often occurs in clinical practice. Although there are no controlled data for the comparison of two or more dopamine agonists to define equivalent dosages, several reports have been published on the clinical experience of experts.24 In addition, the stimulus strength on 5-hydroxytryptamine 2B (5-HT2B) receptors is similar between cabergoline and pergolide, in parallel with their molecular weights.25,26 Thus, according to a previous report,24 we calculated that 2 mg of pergolide is equal to 3 mg of cabergoline. Age- and sex-matched patients with PD who were never treated with ergot derivative dopamine agonists were also included. Six patients taking nonpermitted medication (anorectic or ergot alkaloid agents, Chinese herbs, anticancer or immune-suppressive drugs before enrollment), having a history of significant coronary heart disease, impaired function/dilatation of left/right ventricle, history of peripheral artery occlusive disease, and any clinically significant illnesses that may interfere with their capability to participate in the study were excluded. We also excluded three patients who were not treated in our institute because their history of taking dopamine agonists was found to be inaccurate. As a result, 25 patients treated with pergolide or cabergoline were enrolled in the ergot group. Thirteen patients were treated with pergolide or cabergoline. Five patients were only treated with pergolide and seven patients only with cabergoline. In addition, 25 patients never treated with ergot derivatives were also enrolled in the non-ergot group. Disease 622
Neurology 72
February 17, 2009
severity was assessed with the Unified Parkinson’s Disease Rating Scale (UPDRS) and the Hoehn & Yahr stages. All patients showed normal renal function. Two of the ergot group and three of the non-ergot group patients had mild hypertension. Patients were interviewed with a structured questionnaire about the frequency of dyspnea, fatigue, leg edema, and palpitation, and were scored from 0 (no disability) to 4 (maximum). As for the controls, 30 age- and sex-matched normal volunteers who have no history of cardiac disorders and related conditions requiring medication were examined (age at examination: 67 ⫾ 11 years; 16 women, 14 men). This study was approved by the ethics committee of Nagoya University Graduate School of Medicine. We obtained written informed consent from each participant before data collection. All patients were assessed by an echocardiography GE VIVID 7 machine (GE Medical Systems, Milwaukee, WI) with two independent observers (A.N. and A.Y.) who were blinded to the clinical information. Mitral, aortic, and tricuspid valves were recorded from all possible views with the zoom function. In addition, a stethoscope examination was performed before echocardiography by A.N. and A.Y. Semiquantitative and quantitative measurements for quantification of regurgitant valvular diseases from the continuous wave, pulsed wave, and color Doppler examinations were assessed. Tenting distance and tenting area of the mitral valve were also evaluated as quantitative data.7,8,10,15 We quantified regurgitant lesions by integration of all semiquantitative and quantitative measurements, and a final score was given as follows: absent, 0; trace, 1; mild, 2; moderate, 3; severe, 4.27 A composite regurgitation score was calculated by adding the scores for aortic regurgitation (AR), mitral regurgitation (MR), and tricuspid regurgitation (TR).7,13 The proportion of patients with any regurgitation grade from 3 to 4 was also assessed.7 We derived the systolic pulmonary artery pressures from the TR jet, adding 10 mm Hg to the maximum gradient of the TR jet or 5 mm Hg if the vena cava inferior diameter was less than 10 mm with complete respiratory collapse and 15 mm Hg if the vena cava inferior was greater than 20 mm without respiratory variation. Left ventricular end-diastolic and end-systolic dimensions were measured, and the left ventricular ejection fraction was calculated by the Teichholz method. All patients with PD were investigated for both the diameter and flow of the hepatic vein and inferior vena cava using ultrasonography. In addition, a chest X-ray was also performed if necessary. Blood for BNP quantification was collected in the fasting state in EDTA acid-treated tubes and placed on ice. After centrifugation at 2,500 rpm and 3°C, the plasma was stored at ⫺80°C. BNP levels were measured directly with a specific immunoradiometric assay kit TOSOH AIA-PACK BNP (TOSOH Corp., Tokyo, Japan) including 30 age- and gender-matched controls. Statistical analyses were performed using SPSS 15.0 for Windows (SPSS Inc.). Comparisons of age and disease duration between groups were performed using one-way analysis of variance followed by post hoc Bonferroni correction. Group comparisons of frequencies of valvular regurgitation were restricted to grades 3 and 4 and were performed using the Fisher exact test. The statistical threshold for post hoc comparisons between each treatment group vs the control group was set at p ⬍ 0.017 (0.05/3). The relationships between the cumulative dose of ergot derivative dopamine agonists and tenting area, tenting distance, composite regurgitation score, and BNP were analyzed using Pearson correlation test. Statistical significance was considered as p ⬍ 0.05.
Patient characteristics were as follows: 22 men and 28 women; age at
RESULTS Patient characteristics.
examination, 66 ⫾ 9 years; duration, 11.8 ⫾ 9.6 years; mean Hoehn & Yahr stage 2.8. There were no significant differences between the ergot group and the non-ergot group in terms of Hoehn and Yahr staging (3.0 ⫾ 0.9 vs 2.6 ⫾ 0.9), duration of illness (12.9 ⫾ 9.1 vs 10.6 ⫾ 10.2 years), dosages of levodopa (494 ⫾ 216 mg vs 481 ⫾ 257 mg), age at examination (64.9 ⫾ 9.2 vs 66.4 ⫾ 7.7 years), and gender. Mean daily/cumulative dosage of pergolide was 1.1 ⫾ 0.4/1,752 ⫾ 1,512 mg and that of cabergoline was 3.1 ⫾ 1.0/14,230 ⫾ 2,566 mg. There were no patients who required a surgical operation during the course of this study. The frequency of leg edema, dyspnea, palpitation, and fatigue were not significantly different between the ergot group and the non-ergot group. All patients demonstrated more than 50% ejection fraction and did not show heart failure fulfilling the criteria of the Framingham study.28,29 Echocardiographic findings and plasma BNP levels.
With respect to regurgitation, frequencies of equal to or greater than grade 3 regurgitation were only observed in the ergot group (12%, aortic valve; 12%, mitral valve; and 8%, tricuspid valve), except for one subject in the control group. The frequency of any grade 3 to 4 regurgitation was significant between the ergot group and non-ergot group as well as the ergot group and control group (ergot group vs non-ergot group; p ⬍ 0.001, ergot group vs control group; p ⬍ 0.001). Composite regurgitation scores in the ergot group were higher than in the control group (p ⬍ 0.001), but there were no differences between the ergot group and non-ergot group as well as between the non-ergot group and control group. Differences of tenting area and tenting distance between the ergot group and non-ergot group were slight (tenting area; ergot group, 1.26 ⫾ 0.42, non-ergot group, 1.05 ⫾ 0.21, p ⫽ 0.04, tenting distance; ergot group, 7.53 ⫾ 2.57, non-ergot group, 6.10 ⫾ 1.65, p ⫽ 0.09). The plasma BNP levels as well as the composite regurgitation score were elevated in the ergot group vs the control group (p ⫽ 0.004, p ⬍ 0.001) (table). The BNP levels and composite regurgitation score in the ergot group showed a tendency to be increased as compared to those in the non-ergot group but did not show a significant difference. The BNP levels and composite regurgitation score in the non-ergot group were slightly elevated compared to control, although there was no significant difference. Relationship between cumulative dose of ergot derivative dopamine agonists and echocardiographic findings. The
cumulative dose of ergot derivative dopamine agonists was related to tenting distance (r ⫽ 0.62, p ⫽ 0.001) as well as to tenting area (r ⫽ 0.57, p ⫽ 0.004) but did not show any relationship with the
composite regurgitation score (r ⫽ 0.36, p ⫽ 0.08). Within the high dose group (more than 4,000 mg), 33.3% of patients showed grade 3 to 4 regurgitation, while 15.4% of the low dose group (less than 4,000 mg) exhibited similar grades. Plasma BNP levels in patients with severe valvulopathy. Plasma BNP values were higher in the ergot pa-
tient group with grade 3 to 4 regurgitation, which were seen only in the ergot group, than in those without such a high grade of regurgitation in the ergot group as well those in the non-ergot group (65.3 ⫾ 47.8 pg/mL vs 24.7 ⫾ 17.1 pg/mL vs 21.1 ⫾ 15.4 pg/mL, p ⬍ 0.001). Patients with multiple regurgitation equal to or greater than grade 2 also had higher BNP values than those without (57.8 ⫾ 46.1 vs 22.5 ⫾ 11.9 vs 21.1 ⫾ 15.4 pg/mL, p ⬍ 0.001). According to receiver operating characteristic curve analyses to determine the adequate values for discriminating patients with severe regurgitation from those without, the most appropriate cutoff level of plasma BNP was 39.6 pg/mL, which showed 67.4% sensitivity and 84.4% specificity. In the ergot group, the positive predictive value was 66.7% and the negative predictive value was 89.4% if the plasma BNP level of 39.6 pg/mL was determined as the cutoff value. Relationship between BNP and echocardiographic findings. The BNP levels showed a correlation to the
composite regurgitation scores (r ⫽ 0.70, p ⬍ 0.001, figure) and a correlation to the left ventricular ejection fraction (r ⫽ ⫺0.42, p ⬍ 0.04) but not age at examination, motor examination section (part III) of the UPDRS, and disease duration. Multiple regression analyses demonstrated that BNP values were independently correlated with composite regurgitation scores (t ⫽ 4.08, p ⫽ 0.001) and the left ventricular ejection fraction (t ⫽ ⫺2.07, p ⫽ 0.045, R 2 ⫽ 0.60). We demonstrated that a significant elevation of plasma BNP values was observed in the ergot group vs in control groups. In particular, the BNP values were significantly elevated in the ergot group with more severe or multiple regurgitation than those with no to mild regurgitation and those in the non-ergot group. Furthermore, composite scores of regurgitation were well correlated with BNP values. Serum BNP is elevated in patients with valvular disorders due to ventricular pressure and volume load,20,21 and this may be one reason why plasma BNP values increased in the ergot group. More recently, animal models have demonstrated that left ventricular cardiomyocytes were hypertrophic in both serotonin and pergolide-treated animals com-
DISCUSSION
Neurology 72
February 17, 2009
623
Table
Valvular abnormalities and plasma BNP values in the ergot group and non-ergot group
Grade of regurgitation, no. (%) of patients
Ergot group (n ⴝ 25)
Non-ergot group (n ⴝ 25)
Figure
Correlation between composite regurgitation scores and serum Btype natriuretic peptide (BNP) values
Control (n ⴝ 30)
Aortic regurgitation 0 to 1
13 (60)
22 (88)
27 (90)
2
7 (28)
3 (12)
3 (10)
3
2 (8)
0 (0)
0 (0)
4
1 (4)
0 (0)
0 (0)
13 (60)
18 (72)
26 (87)
2
7 (28)
7 (28)
3 (10)
3
3 (12)
0 (0)
1 (3)
4
0 (0)
0 (0)
0 (0)
19 (76)
21 (84)
25 (83)
2
4 (16)
4 (16)
5 (17)
3
1 (4)
0 (0)
0 (0)
4
1 (4)
0 (0)
0 (0)
6 (24)*
0 (0)
1 (3)
Mitral regurgitation 0 to 1
Tricuspid regurgitation 0 to 1
Any grade from 3 to 4 regurgitation, mean (SD) Composite regurgitation score
3.30 (2.31)†
2.39 (1.29)
1.73 (1.83)
BNP (pg/mL)
33.6 (31.8)‡
21.1 (15.4)
14.2 (8.3)
Composite regurgitation score is the sum of mitral, aortic, and tricuspid regurgitation scores. * The frequency of any grade 3 to 4 regurgitation was statistically significant between the ergot group and the non-ergot group (p ⬍ 0.0001) as well the ergot group and the control group (p ⬍ 0.001). † Composite regurgitation score in the ergot group was significantly higher than that in the control group (p ⬍ 0.001). Composite regurgitation score of the ergot group tended to be increased when compared to the non-ergot group, but was not significantly different. Patients with grade 3 to 4 regurgitation were seen only in the ergot group. ‡ The plasma BNP level was significantly elevated in the ergot group vs in the control group (p ⫽ 0.004). The plasma BNP level of the ergot group tended to be increased when compared to the non-ergot group, but was not significantly different. The plasma BNP of patients with grade 3 to 4 regurgitation seen only in the ergot group was significantly elevated. BNP ⫽ B-type natriuretic peptide.
pared with placebo-treated animals, and macroscopically, left ventricular cavities were more dilated in both the serotonin and pergolide groups.30 Thus, the second possible explanation is that direct toxic effects on the cardiomyocytes may have an influence on increased plasma BNP values. Since ventricular involvement in patients with PD using ergot derivative dopamine agonists has not been fully assessed, further prospective and pathologic studies will be needed to clarify this issue. BNP is expected to detect preclinical structural and functional myocardial alterations not detectable by current techniques. Thus, BNP testing for structural heart disease screening in community-based populations is useful for cohorts with a high prevalence of heart disease.31,32 However, age, renal dysfunction, and fluid overload can also contribute to 624
Neurology 72
February 17, 2009
Composite regurgitation scores were correlated with BNP values (r ⫽ 0.70, p ⬍ 0.001). Composite regurgitation score was calculated by adding the score of aortic regurgitation, mitral regurgitation, and tricuspid regurgitation.
elevated BNP concentrations.33 Furthermore, elderly hypertensive subjects with orthostatic systolic blood pressure decrease also show significantly higher BNP values than in the control group, suggesting greater cardiac burden although the influence of orthostatic hypotension on BNP is not well known.34 In this study, no patients exhibited symptomatic orthostatic hypotension but sympathetic dysfunction in PD might result in a slight elevation of plasma BNP levels in the non-ergot group compared with controls. Since ergot derivative dopamine agonists can exacerbate not only cardiac fibrosis but also renal dysfunction and orthostatic hypotension, measurement of plasma BNP values may be beneficial to detect and prevent the worsening of these clinical conditions by means of administration of such dopamine agonists. This study showed that plasma BNP values were significantly higher in the ergot group than in controls, while the plasma BNP values showed a tendency to be elevated in the ergot group vs the non-ergot group, but did not show a significant difference. In this study, over three-quarters of patients in the ergot group did not develop significant valvular regurgitation. Such a low occurrence of severe valvulopathy may be a consequence of the lower dose of pergolide and cabergoline prescribed to our Japanese patients when compared to Western countries.2,9 However, six patients of the ergot group with grade 3 to 4 regurgitation clearly showed a significant elevation of plasma BNP values compared to those of the non-ergot group. These results demonstrate that plasma BNP values can be used as a marker for patients who have reached a significant degree of heart valve involvement prior to heart failure.
This study also showed that the composite regurgitation score in the non-ergot group showed a slight elevation when compared to controls. We cannot rule out the possibility that other drugs including L-dopa or sympathetic dysfunction have an influence on the increase of regurgitation in the non-ergot group, but no patients with grade 3 to 4 regurgitation were observed in this group. According to a recent review, a considerably large proportion of patients do not develop valvulopathy, despite several years’ exposure to high doses of pergolide, suggesting the presence of patients with a low susceptibility to pergolide.2 Furthermore, the low dose of pergolide used in Japan can be associated with the low frequency of severe valvulopathy in our patients treated with ergot as mentioned above.2,9 The striking point here is, as mentioned above, patients with a grade 3 to 4 composite score were present in the ergot group but not in the non-ergot group. Both pergolide and cabergoline are potent agonists of not only dopamine but also the 5-HT2B receptor. It is supposed that stimulation of 5-HT2B, which is expressed in heart valves, induces prolonged activation of fibroblast mitogenesis resulting in valvular fibroplasia.35,36 Thus, high dose and long-term ergot derivative administration is thought to be a risk factor for valvulopathy in patients with PD.2-18 The significant association of the cumulative dose of ergot derivatives and mitral valve tenting area/distance, which have been proposed as restrictive changes due to valvular fibroplasia, was observed.7,8,10,15 However, these significant adverse events did not occur in all ergot patients, including those administered high cumulative doses as previously reported,2 suggesting that patients who receive benefit from ergot-derived dopamine agonists without valvulopathy will exist at a constant rate under careful follow-up. In Germany, if any abnormalities are seen on echocardiography, non-ergot dopamine agonists are recommended.37 Although there has been no report concerning plasma BNP values in patients with PD, our results support the view that plasma BNP levels will be a beneficial marker for monitoring cardiac fibrosis due to ergot derivative dopamine agonists. Measurement of plasma BNP levels is quicker, more accessible, and cheaper than echocardiography and may contribute to the assessment of not only the development of valvulopathy and myocardium damage but also several other important factors deteriorated by ergot derivative dopamine agonists in patients with PD. In addition, plasma BNP values can predict the prognosis of patients with chronic heart failure38 and mitral regurgitation.39 Echocardiography is effective and able to identify valvulopathy as a cause of incipient or present right heart failure, and it is a
satisfactory screen for valvulopathy itself, but BNP is a suitable marker for the relevant forms of cardiac dysfunction. The combination of comprehensive echocardiography and plasma BNP levels will complementarily elucidate the presence of cardiac damage in patients with PD using ergot derivative dopamine agonists. Received May 23, 2008. Accepted in final form November 17, 2008.
REFERENCES 1. Nutt JG, Wooten GF. Clinical practice: diagnosis and initial management of Parkinson’s disease. N Engl J Med 2005;353:1021–1027. 2. Antonini A, Poewe W. Fibrotic heart-valve reactions to dopamine-agonist treatment in Parkinson’s disease. Lancet Neurol 2007;6:826–829. 3. Corvol JC, Anzouan-Kacou JB, Fauveau E, et al. Heart valve regurgitation, pergolide use, and Parkinson disease: an observational study and meta-analysis. Arch Neurol 2007;64:1721–1726. 4. Ru˚zicka E, Lı´nkova´ H, Penicka M, Ulmanova´ O, Nova´kova´ L, Roth J. Low incidence of restrictive valvulopathy in patients with Parkinson’s disease on moderate dose of pergolide. J Neurol 2007;254:1575–1578. 5. Dewey RB 2nd, Reimold SC, O’Suilleabhain PE. Cardiac valve regurgitation with pergolide compared with nonergot agonists in Parkinson disease. Arch Neurol 2007;64: 377–380. 6. Schade R, Andersohn F, Suissa S, Haverkamp W, Garbe E. Dopamine agonists and the risk of cardiac-valve regurgitation. N Engl J Med 2007;356:29–38. 7. Zanettini R, Antonini A, Gatto G, Gentile R, Tesei S, Pezzoli G. Valvular heart disease and the use of dopamine agonists for Parkinson’s disease. N Engl J Med 2007;356: 39–46. 8. Junghanns S, Fuhrmann JT, Simonis G, et al. Valvular heart disease in Parkinson’s disease patients treated with dopamine agonists: a reader-blinded monocenter echocardiography study. Mov Disord 2007;22:234–238. 9. Yamamoto M, Uesugi T, Nakayama T. Dopamine agonists and cardiac valvulopathy in Parkinson disease: a casecontrol study. Neurology 2006;67:1225–1229. 10. Kim JY, Chung EJ, Park SW, Lee WY. Valvular heart disease in Parkinson’s disease treated with ergot derivative dopamine agonists. Mov Disord 2006;21:1261–1264. 11. Peralta C, Wolf E, Alber H, et al. Valvular heart disease in Parkinson’s disease vs. controls: An echocardiographic study. Mov Disord 2006;21:1109–1113. 12. Waller EA, Kaplan J, Heckman MG. Valvular heart disease in patients taking pergolide. Mayo Clin Proc 2005;80: 1016–1020. 13. Baseman DG, O’Suilleabhain PE, Reimold SC, Laskar SR, Baseman JG, Dewey RB Jr. Pergolide use in Parkinson disease is associated with cardiac valve regurgitation. Neurology 2004;63:301–304. 14. Horvath J, Fross RD, Kleiner-Fisman G, et al. Severe multivalvular heart disease: a new complication of the ergot derivative dopamine agonists. Mov Disord 2004;19:656– 662. Neurology 72
February 17, 2009
625
15.
Van Camp G, Flamez A, Cosyns B, et al. Treatment of Parkinson’s disease with pergolide and relation to restrictive valvular heart disease. Lancet 2004;363:1179–1183. 16. Van Camp G, Flamez A, Cosyns B, Goldstein J, Perdaens C, Schoors D. Heart valvular disease in patients with Parkinson’s disease treated with high-dose pergolide. Neurology 2003;61:859–861. 17. Pritchett AM, Morrison JF, Edwards WD, Schaff HV, Connolly HM, Espinosa RE. Valvular heart disease in patients taking pergolide. Mayo Clin Proc 2002;77:1280–1286. 18. Flowers CM, Racoosin JA, Lu SL, Beitz JG. The US Food and Drug Administration’s registry of patients with pergolide-associated valvular heart disease. Mayo Clin Proc 2003;78:730–731. 19. Weber M, Hamm C. Role of B-type natriuretic peptide (BNP) and NT-proBNP in clinical routine. Heart 2006; 92:843–849. 20. Detaint D, Messika-Zeitoun D, Chen HH, et al. Association of B-type natriuretic peptide activation to left ventricular end-systolic remodeling in organic and functional mitral regurgitation. Am J Cardiol 2006;97:1029–1034. 21. Eimer MJ, Ekery DL, Rigolin VH, Bonow RO, Carnethon MR, Cotts WG. Elevated B-type natriuretic peptide in asymptomatic men with chronic aortic regurgitation and preserved left ventricular systolic function. Am J Cardiol 2004;94:676–678. 22. Watanabe H, Atsuta N, Ito M, et al. Relationship between non-motor function and quality of life in Parkinson’s disease; longitudinal study. Rinsho Shinkeigaku 2007;47: 1015. Abstract. 23. Calne DB, Snow BJ, Lee C. Criteria for diagnosing Parkinson’s disease. Ann Neurol 1992;32:S125–S127. 24. Junghanns S, Glo¨ckler T, Reichmann H. Switching and combining of dopamine agonists. J Neurol 2004;251 suppl 6:VI/19–23. 25. Millan MJ, Maiofiss L, Cussac D, et al. Differential actions of antiparkinson agents at multiple classes of monoaminergic receptor: I: a multivariate analysis of the binding profiles of 14 drugs at 21 native and cloned human receptor subtypes. J Pharmacol Exp Ther 2002;303:791–804. 26. Jahnichen S, Horowski R, Pertz HH. Agonism at 5-HT2b receptor is not a class effect of the ergolines. Eur J Pharmacol 2005;513:225–229. 27. Zoghbi WA, Enriquez-Sarano M, Foster E, et al. Recommendations for evaluation of the severity of native valvular regurgitation with two-dimensional and Doppler echocardiography. J Am Soc Echocardiogr 2003;16:777–802.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
McKee PA, Castelli WP, McNamara PM, et al. The natural history of congestive heart failure: the Framingham study. N Engl J Med 1971;285:1441–1446. Ho KK, Anderson KM, Kannel WB, et al. Survival after the onset of congestive heart failure in Framingham Heart Study subjects. Circulation 1993;88:107–115. Droogmans S, Franken PR, Garbar C, et al. In vivo model of drug-induced valvular heart disease in rats: pergolideinduced valvular heart disease demonstrated with echocardiography and correlation with pathology. Eur Heart J 2007;28:2156–2162. Niinuma H, Nakamura M, Hiramori K. Plasma B-type natriuretic peptide measurement in a multiphasic health screening program. Cardiology 1998;90:89–94. Nakamura M, Endo H, Nasu M, Arakawa N, Segawa T, Hiramori K. Value of plasma B type natriuretic peptide measurement for heart disease screening in a Japanese population. Heart 2002;87:131–135. Redfield MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett JC Jr. Plasma brain natriuretic peptide concentration: impact of age and gender. J Am Coll Cardiol 2002;40:976–982. Eguchi K, Kario K, Hoshide S, et al. Greater change of orthostatic blood pressure is related to silent cerebral infarct and cardiac overload in hypertensive subjects. Hypertens Res 2004;27:235–241. Setola V, Hufeisen SJ, Grande-Allen KJ, et al. 3,4Methylenedioxymethamphetamine (MDMA, “Ecstasy”) induces fenfluramine-like proliferative actions on human cardiac valvular interstitial cells in vitro. Mol Pharmacol 2003;63:1223–1229. Newman-Tancredi A, Cussac D, Quen-tric Y, et al. Differential actions of antiparkinson agents at multiple classes of monoaminergic receptor. III. Agonist and antagonist properties at serotonin, 5-HT(1) and 5-HT(2), receptor subtypes. J Pharmacol Exp Ther 2002;303:815–822. Reichmann H, Bilsing A, Ehret R, et al. Ergoline and nonergoline derivatives in the treatment of Parkinson’s disease. J Neurol 2006;253 suppl 4:iv36–iv38. Miller WL, Hartman KA, Burritt MF, et al. Serial biomarker measurements in ambulatory patients with chronic heart failure: the importance of change over time. Circulation 2007;116:249–257. Detaint D, Messika-Zeitoun D, Avierinos JF, et al. B-type natriuretic peptide in organic mitral regurgitation: determinants and impact on outcome. Circulation 2005;111: 2391–2397.
No Charge for Color Figures Neurology® is committed to presenting data in the most descriptive way for the benefit of our readers. To make possible the publication of a greater number of color figures, we have elimated our color figure charges to authors.
626
Neurology 72
February 17, 2009
Intracranial arterial wall imaging using high-resolution 3-tesla contrast-enhanced MRI R.H. Swartz, MD, PhD S.S. Bhuta, MD R.I. Farb, MD R. Agid, MD R.A. Willinsky, MD K.G. terBrugge, MD J. Butany, MD B.A. Wasserman, MD D.M. Johnstone, RTR, RTMR F.L. Silver, MD D.J. Mikulis, MD
Address correspondence and reprint requests to Dr. David Mikulis, Division of Neuroradiology, Department of Medical Imaging, New East Wing, Toronto Western Hospital, University Health Network, 399 Bathurst St., Toronto, Ontario, Canada M5T 2S8
[email protected]
ABSTRACT
Background: Conventional arterial imaging focuses on the vessel lumen but lacks specificity because different pathologies produce similar luminal defects. Wall imaging can characterize extracranial arterial pathology, but imaging intracranial walls has been limited by resolution and signal constraints. Higher-field scanners may improve visualization of these smaller vessels.
Methods: Three-tesla contrast-enhanced MRI was used to study the intracranial arteries from a consecutive series of patients at a tertiary stroke center. Results: Multiplanar T2-weighted fast spin echo and multiplanar T1 fluid-attenuated inversion recovery precontrast and postcontrast images were acquired in 37 patients with focal neurologic deficits. Clinical diagnoses included atherosclerotic disease (13), CNS inflammatory disease (3), dissections (3), aneurysms (3), moyamoya syndrome (2), cavernous angioma (1), extracranial source of stroke (5), and no definitive clinical diagnosis (7). Twelve of 13 with atherosclerotic disease had focal, eccentric vessel wall enhancement, 10 of whom had enhancement only in the vessel supplying the area of ischemic injury. Two of 3 with inflammatory diseases had diffuse, concentric vessel wall enhancement. Three of 3 with dissection showed bright signal on T1, and 2 had irregular wall enhancement with a flap and dual lumen. Conclusions: Three-tesla contrast-enhanced MRI can be used to study the wall of intracranial blood vessels. T2 and precontrast and postcontrast T1 fluid-attenuated inversion recovery images at 3 tesla may be able to differentiate enhancement patterns of intracranial atherosclerotic plaques (eccentric), inflammation (concentric), and other wall pathologies. Prospective studies are required to determine the sensitivity and specificity of arterial wall imaging for distinguishing the range of pathologic conditions affecting cerebral vasculature. Neurology® 2009;72:627–634 GLOSSARY CTA ⫽ computed tomographic angiography; DSA ⫽ digital subtraction angiography; FIESTA ⫽ fast imaging employing steady state acquisition; FLAIR ⫽ fluid-attenuated inversion recovery; FRFSE ⫽ fast recovery fast spin echo; ICA ⫽ internal carotid artery; MCA ⫽ middle cerebral artery; MR ⫽ magnetic resonance; MRA ⫽ magnetic resonance angiography; TE ⫽ echo time; TI ⫽ inversion time; TR ⫽ recovery time.
Many modalities exist for imaging the lumina of blood vessels, including conventional CT and magnetic resonance (MR) angiography. However, these approaches have a limited ability to differentiate vascular pathologies, because different pathologies can produce the same luminal defects. Direct imaging of the blood vessel wall offers the potential to discriminate between these pathologies. Characterization of atherosclerotic plaque using MRI is already well established.1-3 High-resolution MR studies of the extracranial carotid arteries have identified features of atherosclerotic plaque that may convey increased risk of thromboembolic events1 even in cases with less than 50% luminal stenosis.2 Extracranial large artery plaques have been reported to cause asymmetric wall thickening with enhancement.4 In contrast, extracranial vasculitis has been reported to show circumferential thickening and enhancement on MRI.5 Supplemental data at www.neurology.org From the Departments of Neurology (R.H.S.), Neuroradiology (S.S.B., R.I.F., R.A., R.A.W., K.G.t.B., F.L.S., D.J.M.), Medical Imaging (D.M.J.), and Pathology (J.B.), Toronto Western Hospital, University Health Network, University of Toronto, Ontario, Canada; and Russell H. Morgan Department of Radiology and Radiological Sciences (B.A.W.), Johns Hopkins Hospital, Baltimore, MD. Disclosure: The authors report no disclosures. Copyright © 2009 by AAN Enterprises, Inc.
627
Intracranial arterial wall imaging poses a greater challenge given the smaller size and relatively deep location of the target vessels. Threetesla (T) MRI has been used to characterize carotid atherosclerosis1,6 and dissection,7 as well as extracranial vessel wall imaging to assess giant cell arteritis.8,9 Although the need to look “beyond the lumen” has been increasingly recognized in assessing the extracranial vasculature, it has not yet been systemically applied to multiple intracranial vascular pathologies using highresolution 3T MRI. To make this assessment, we applied high-resolution 3T MRI in a consecutive case series from a tertiary stroke center with the goal of identifying the unique imaging appearance of intracranial vessel pathologies. METHODS Hypotheses. Based on our understanding of the morphology and pathophysiology of the three major disorders affecting the wall blood of vessels, we developed the following imaging hypotheses: • Atherosclerotic plaques show eccentric irregular wall thickening; gadolinium enhancement of the plaque correlates with plaque instability. • Vasculitis produces smooth circumferential concentric wall thickening with diffuse gadolinium enhancement of the inflamed wall. • Dissection shows eccentric wall thickening with T1 bright wall components representing intramural hematoma.
Subjects. This study was a retrospective analysis of consecutive patients presenting to neurology or neurosurgery at a tertiary care center (Toronto Western Hospital) with focal neurologic symptoms for which the treating clinicians requested additional information about the intracranial vasculature, beyond conventional angiographic techniques (computed tomographic angiography [CTA], magnetic resonance angiography [MRA], or digital subtraction angiography [DSA]). All patients had to have no contraindications to MRI and be medically stable to be in the MR scanner for 30 to 90 minutes without general sedation. The sequences were applied, when requested, beginning in August 2005, and scans were collected until the beginning of the analysis period in September 2007. Where patients presented under acute stroke protocols, plain CT brain scans had been performed. All participants also had routine MRI pulse sequences (fluid-attenuated inversion recovery [FLAIR], T2-weighted, and diffusion-weighted images) and at least one method of vessel lumen assessment (MRA, CTA, or DSA) before the wall imaging protocol. When possible (in 34 of 37 people), wall imaging was performed during the acute hospitalization, within days to weeks of the presenting symptoms. The three patients who did not have acute wall imaging returned for follow-up at 2 months (table e-1 on the Neurology® Web site at www.neurology.org, case 11), 3 months (case 32), and 5 months (case 37) after symptom onset. Institutional ethics approval. All MR images were acquired using vendor-supplied sequences approved for clinical use. To correlate the MR findings with patient demographics and clinical diagnoses, a retrospective chart review and retrospective de628
Neurology 72
February 17, 2009
tailed imaging review was undertaken. Hospital research ethics board approval was obtained for a systematic, retrospective chart and imaging review of all patients who had the 3T wall imaging protocol performed (Toronto Western Hospital Research Ethics Board).
Imaging protocol and analysis. Patients were scanned on a 3T MRI system (HDX platform, GE Healthcare, Milwaukee) using an eight-channel head coil. All sequences applied were standard, approved, vendor-supplied pulse sequences. No new experimental sequences were pioneered in this study. The “protocol” described in this study refers to the implementation of a standardized set of sequences and the disciplined application of those sequences (e.g., with specific axial, coronal, and sagittal sampling of the site of intracranial stenosis). The number of pulse sequences used for 3T wall imaging was reduced over the 2 years that the protocol was performed. Initially, multiple sequences were applied: T1 spin echo (repetition time [TR]/echo time [TE] 450/21, slice thickness 3 mm, 384 ⫻ 384 matrix), T1 FLAIR (TR/TE/inversion time [TI] 2108/12/860, slice thickness 3 mm, 384 ⫻ 384 matrix), three-dimensional fast imaging employing steady state acquisition (FIESTA) (TR/TE 6/2.5, slice thickness 0.8 mm, 256 ⫻ 256 matrix), and T2 fast recovery fast spin echo (FRFSE) (TR/TE 3450/92, echo train length 17, slice thickness 3 mm, 512 ⫻ 512 matrix). All T1 FLAIR and T2 FRFSE sequences used a parallel factor of 2. Field of view was 16 to 22 cm. The ability to distinguish the vessel wall from the lumen, as well as the presence of flow artifacts, was assessed for each of the sequences. It was observed that T1 FLAIR improved visualization of blood vessel walls via its black blood characteristics including the absence of blood signal and decreased sensitivity to flow artifacts. The T1 spin echo sequences, which had these effects, were replaced with multiplanar T1 FLAIR, either with or without fat saturation, but always with and without gadolinium. The MRI technologist, in consultation with the neuroradiologist, selected the best combination of acquisition orientations and target vessels. Three-dimensional FIESTA was found to be more susceptible to artifacts than T2, without providing additional information. These were thus eliminated from the protocol. The final common “wall protocol” consisted of a minimum of seven sequences on the 3T MRI: 1) axial T2-weighted images; precontrast 2) axial, 3) sagittal, and 4) coronal T1 FLAIR images; and postcontrast 5) axial, 6) sagittal, and 7) coronal T1 FLAIR images. All sequences had to be monitored for quality to ensure that the orientation was successful to capture the affected artery at the site of stenosis on at least one slice both parallel and perpendicular to a site of abnormality. Where no clear focal abnormality was noted on baseline angiography (MRA, CTA, or DSA), the acquisitions were targeted to ensure sampling of the clinically suspected vessel (e.g., the left middle cerebral artery [MCA] in a patient presenting with an aphasic TIA). Imaging analysis was performed on a radiology information system–picture archiving and communication system and, if necessary, multiplanar reformatting was performed. At least one method of conventional luminal imaging (MRA, CTA, or DSA) was available for all patients. Retrospective visual analysis of all imaging information from each case was performed to describe any cerebral pathology, to identify intracranial and extracranial sites of stenosis, to evaluate focal wall thickening, and to assess postcontrast enhancement, as well as to categorize any enhancement as either concentric or eccentric. All imaging analysis was performed by a neuroradiologist (D.J.M.) blinded to the final clinical diagnosis.
Table
Patient demographics by final clinical diagnoses
Diagnosis
No. of patients
Sex, F:M
Average age, (min–max), y
Intracranial atherosclerotic stroke
13
3:10
64 (44–80)
Stroke with extracranial source
5
2:3
60 (32–84)
Vasculitic/Inflammatory diseases
3
3:0
43 (18–66)
Dissection
3
1:2
56 (49–60)
Other vascular lesions (3 aneurysms, 1 cavernoma)
4
4:0
48 (23–65)
Moyamoya syndrome
2
0:2
38 (30–46)
No final clinical diagnosis
7
2:5
50 (30–79)
15:22
55 (18–84)
Total
37
Wall thickening and enhancement patterns were categorized by visual inspection. Enhancement was considered concentric if it was circumferential and uniform (defined specifically as width of the thinnest wall segment at least 50% or more of the thickest segment). In contrast, eccentric enhancement was defined as either clearly limited to one side of the vessel wall (e.g., not 360degree circumferential enhancement) or, where circumferential enhancement was noted, the thinnest part of the wall enhancement was estimated to be less than 50% of the thickest point (figure e-4). More precise measurements are not feasible based on the limits of resolution of the imaging.
Case histories. Charts were reviewed for demographic data, medical history (including cerebrovascular risk factors), presenting symptoms and physical findings (blood pressure, height, weight, neurologic deficits), laboratory results, and imaging findings.
Diagnostic categorization. Subjects were classified as “atherosclerotic” if they had two or more vascular risk factors: men older than 50 years, women older than 60 years, hypertension, hypercholesterolemia, diabetes, overweight or obesity, smoking, or calcified intracranial arterial plaques on plain CT head. Subjects were classified as “inflammatory” if they had one or fewer vascular risk factors with blood tests or prior diagnoses consistent with inflammatory diseases (lupus, erythrocyte sedimentation rate or C-reactive protein elevations, antinuclear antigen, or anti–phospholipid antibody positivity). Subjects whose workup discovered a carotid or cardiac source for their presenting symptoms were classified as “extracranial source.” Those subjects who had structural pathology diagnosed on conventional angiographic techniques (MRA, CTA, or DSA) were categorized as either “dissection,” “structural (aneurysm/cavernoma),” or “moyamoya.” Subjects were classified as “no final clinical diagnosis” if they met none, or more than one, of the above criteria.
Thirty-seven patients underwent the vessel wall protocol. Demographics by final clinical diagnoses are shown in table. Individual case data are available in table e-1.
RESULTS
Wall imaging patterns. Stroke with intracranial athero-
All 13 of the patients with stroke/TIA and intracranial atherosclerotic disease showed focal luminal narrowing of intracranial vessels on MRA. Twelve had diffusion-weighted abnormalities suggestive of recent cerebral infarction, whereas 1 patient sclerosis.
with transient focal symptoms had no changes on diffusion-weighted imaging. The narrowed areas on conventional luminal studies corresponded to focal areas of thickened wall with narrowed lumen seen on T1 FLAIR (precontrast) and T2-weighted images (figure 1). Of these 13 patients, 12 had focal areas of eccentric wall enhancement in the intracranial vessel supplying the territory of acute infarct (figures 1, 2, and e-1). Enhancing plaques were visualized in the relevant major branches of the circle of Willis, including the MCA (figure 1), anterior cerebral artery (figure e-1), and basilar artery (figure 2). Most patients with enhancement (10 of 12) had enhancement only in the vessel supplying the area of acute infarction, even with multiple areas of wall thickening and luminal stenosis indicative of plaques elsewhere (figure e-2). Two patients had multiple vessels with focal eccentric enhancement, not all of which had evidence of ischemic injury in the territory they supplied (figure e-3). Atherosclerotic enhancement was typically irregular and eccentric (figures 1 and 2). Circumferential, but still eccentric, enhancement was occasionally seen (figure e-4). One patient had a thickened wall with luminal stenosis but no enhancement; that patient had wall imaging only at followup, 5 months postinfarct (figure e-5). Extracranial source of stroke. In the five patients with an extracranial source of stroke, none showed intracranial wall enhancement. Inflammatory disease. In contrast to the pattern seen in patients with atherosclerotic risk factors, one patient with biopsy-proven giant cell arteritis (figure 3) showed a smooth, diffuse, concentric pattern of enhancement. The same pattern of smooth, concentric enhancement was seen in another patient with systemic vasculitis, as well as one patient with druginduced vasculopathy. These patterns are difficult to discern on conventional 1.5T sequences. One patient with systemic lupus and amaurosis fugax showed no intracranial wall enhancement of the circle of Willis vessels. Notably, the ophthalmic artery was not well visualized; vessels of this caliber are likely at the limit of the resolving power of the 3T system. Intracranial dissection. Intracranial dissection had a similar pattern to atherosclerosis, including eccentric wall thickening with enhancement. Distinguishing features included T1 bright wall elements on nonenhanced T1 FLAIR sequences (indicating methemoglobin in the arterial wall), as well as visualization of a false lumen (figure 4). Other vascular lesions. In three patients with intracranial aneurysms, a thin wall with smooth, circumferential, concentric enhancement was noted (figure e-6). The prognostic significance of aneurysmal wall Neurology 72
February 17, 2009
629
Figure 1
Focal, eccentric enhancement of the left middle cerebral artery wall
A 57-year-old man with untreated hypertension presented with recurrent right hemiplegia and dysarthria (table e-1, case 4). Axial diffusion-weighted image (A) shows an acute infarct (arrow) in the left corona radiata. Time-of-flight magnetic resonance angiogram maximum-intensity projection image (B) demonstrates focal stenosis (arrow) in the left middle cerebral artery (MCA). Sagittal T1 fluid-attenuated inversion recovery (FLAIR) image (C) shows eccentric thickening (arrow) of the anterosuperior wall of the left MCA. Axial T2 fast recovery fast spin echo image (D) shows focal eccentric wall thickening (arrow) of the left M1 segment of the MCA. Axial T1 FLAIR images before (E) and after (F) gadolinium show an eccentrically enhancing plaque (arrow) on the anterior wall of the left MCA. This is likely representing an inflamed active plaque (“hot plaque”).
enhancement patterns is uncertain. One patient who was initially thought to have an aneurysm underwent wall imaging and was found to have a nonenhancing cavernous angioma. Moyamoya syndrome. Two patients with moyamoya syndrome had severely narrowed or occluded MCA branches and extensive collateralization. One presented with hemorrhage, and the other presented with hemodynamic symptoms. Neither patient showed wall thickening or enhancement of the circle of Willis vessels (including the distal internal carotid artery). No final clinical diagnosis. In the seven people with no final clinical diagnosis, three had acute infarcts on diffusion-weighted imaging. All seven had complex case histories with either risk factors for multiple types of disease or insufficient risk factors for any 630
Neurology 72
February 17, 2009
specific disease (table e-1). Two showed eccentric wall enhancement, two showed concentric enhancement, one had a mixed pattern, and two had thickened walls without enhancement; both of these had been imaged months after acute infarctions. The assessment of the wall of intracranial vessels is limited by several factors. Conventional imaging sequences (T1, T2, FLAIR), even at 3 T, do not render these vessels with sufficient detail to fully assess the vessel wall. Sequence optimization using higher spatial resolution is required. Increasing spatial resolution by decreasing slice thickness and decreasing in-plane voxel size is limited by the signalto-noise ratio, which is already deficient for small vessels using a 1.5T scanner. One prior study used MRI to examine the wall in atherosclerosis10; how-
DISCUSSION
Figure 2
Focal eccentric enhancement of the basilar artery wall
A 61-year-old man presented with recurrent vertebrobasilar TIAs and small strokes in the posterior circulation on diffusion imaging (not shown) (table e-1, case 5). Sagittal T1 fluid-attenuated inversion recovery (FLAIR) image (A) shows eccentric thickening (arrow) and high signal in the wall of a short segment of the tortuous basilar artery. Axial T2-weighted image (B) shows thickening of the wall (arrow) with iso to low signal. On a pregadolinium T1 axial image with fat saturation (C), minimal high signal intensity is noted in the arterial plaque (arrow). A coronal T1 FLAIR postgadolinium image (D) shows a larger area of eccentric enhancement (arrow) in the mid basilar artery. Axial T1 FLAIR precontrast (E) shows eccentric wall thickening of the basilar artery. The eccentric enhancement of the basilar artery wall is best appreciated on the corresponding axial T1 FLAIR postcontrast image (F). Gadolinium dynamic contrast enhanced magnetic resonance angiogram maximum-intensity projection image (G) and digital subtraction angiogram (H) demonstrate a corresponding area of severe stenosis (arrow) in the mid basilar artery (compare G and H with D).
ever, this study used 1.5T MRI, which cannot achieve sufficient spatial resolution to assess the intracranial wall. The improved signal-to-noise provided by 3T MRI is therefore used to achieve higher spatial resolution. However, examinations still require targeting the vessel of interest, because the intracranial vessels are smaller, are more variable in their distribution than the extracranial carotid arteries, and require longer acquisition times. This can be time-consuming for the radiologist, technologist, and patient. Finally, it is difficult to validate radiologic findings of intracranial arteries because, unlike the case of imaging the carotid bifurcation where endarterectomy is often performed, pathology is rarely obtained from the intracranial vessels. In the current study, we used the demonstrated advantages
of 3T MRI11,12 to overcome these limitations and demonstrate pathology within the wall of the smaller caliber intracranial vessels. The use of T1 FLAIR sequences, acquired in multiple slice dimensions precontrast and postcontrast, as well as high-resolution T2 images, has provided adequate signal to resolve the smaller vessels. It was chosen for its superior block blood characteristics, where flowing blood returns no signal. The eccentric wall enhancement observed in intracranial atherosclerosis is consistent with findings using 3T MRI of atherosclerotic plaques in carotid arteries,6 femoral arteries,3 and the abdominal aorta,3 all of which have been validated with histologic correlations. Although most of our atherosclerotic plaque cases show homogeneous enhancement within the plaque, case 4 may be an exNeurology 72
February 17, 2009
631
Figure 3
Biopsy-confirmed giant cell arteritis with smooth concentric middle cerebral artery enhancement
A 67-year-old woman presented with multiple TIAs followed by headache and multiple focal neurologic deficits, progressing to decreased level of consciousness (table e-1, case 9). Diffusion images (not shown) revealed dozens of small cortical and subcortical areas of restriction. Reformatted CT angiogram (A) shows bilateral narrowing (arrows) of the cavernous and supraclinoid internal carotid arteries (ICAs). Catheter digital subtraction angiogram images (B, combined right and left injections) confirm these findings (black arrows). Axial fast imaging employing steady state acquisition images (C) at the level of the cavernous sinuses show wall thickening (arrow) of bilateral cavernous ICAs. Axial T1 fluid-attenuated inversion recovery (FLAIR) (D and F) demonstrates mural thickening (arrow) of both cavernous ICAs with luminal narrowing. After gadolinium administration, corresponding axial T1 FLAIR images (E and G) show smooth, concentric wall enhancement (arrows) of the cavernous ICAs. Similar enhancement (arrow) is also noted in both vertebral arteries (H). Note the normal-appearing, nonenhancing basilar artery (starred, D–G). Temporal artery biopsy (I and J) shows destruction of the internal elastic lamina with inflammatory cells (predominately lymphocytic with occasional multinucleated giant cells) in the media and adventitia. There were also luminal occlusion and fibrin platelet plugs in the vasa vasorum confirming the diagnosis of giant cell arteritis.
ample of an atherosclerotic plaque with fibrous cap enhancement superficial to a lipid core. In studies of extracranial vessel imaging with pathologic confirmation, inflammatory conditions are associated with concentric, circumferential wall thickening and enhancement,5 whereas atherosclerotic disease is frequently eccentric.4 In our study, we found similar patterns in the intracranial vessels of patients with atherosclerotic and inflammatory disease (biopsy confirmed in one case) confirming our hypotheses. The enhancement patterns were consistent and largely confined to the vessels supplying the area of acute infarction. In some cases, remote vessels also enhanced, raising concern of active or “hot” plaque that might be at risk of future thromboembolic events. It remains to be determined whether enhancement within intracranial atherosclerotic plaques can be used to select patients at high risk of recurrent ischemic stroke. None of the three individuals who had wall imaging months after their acute events showed wall enhancement, and none of the patients with extracranial sources for their strokes showed intracra632
Neurology 72
February 17, 2009
nial wall enhancement. In addition, all 12 patients with atherosclerotic disease imaged acutely showed enhancement, and some had enhancement in vessels that had not yet shown signs of infarction. This suggests that enhancement may reflect active plaques or “plaques at risk.” Studies with carotid wall imaging have shown similar correlations between plaque rupture and symptomatic presentations.13 The concept of using imaging to identify vulnerable plaques is well established.14 Inflammation in chronic atherosclerotic plaques plays a role in destabilizing the plaque.15 Denuded vascular endothelium with inflammatory components or neovascularization is most likely to enhance and also likely serve as a focus for clot formation and embolization.14 It is therefore not surprising that similar enhancement patterns can be seen in intracranial vessels using higher-resolution scanners and optimized acquisitions. It remains to be seen whether current treatments (e.g., statin loading, anti-inflammatory therapies) can alter the enhancement pattern and affect clinical outcome. Our dissection cases also confirmed the hypothesis that eccentric bright T1 elements (higher signal
Figure 4
Basilar artery dissection
A 60-year-old man with neck pain after a rugby injury presented with ataxia and left-sided weakness (table e-1, case 13). Diffusion-weighted imaging (A) showed restricted diffusion in the pons, and conventional angiography was interpreted as a dissection of the basilar artery (B). Contrast-enhanced magnetic resonance angiogram (C, axial slice) also shows a severely narrowed lumen of the basilar artery at the level of the midpons (level shown by arrow on slice G). At the same level, blooming artifact can be seen on the gradient echo T2* images (D, arrow). T1 fluid-attenuated inversion recovery (FLAIR) image without contrast (E) shows bright signal consistent with hematoma in the wall (arrow), with narrowed lumen. Postcontrast T1 FLAIR (F) shows additional enhancement and thickening of the wall throughout the length of the basilar artery (arrows) and into the vertebral arteries bilaterally (slices not shown). Sagittal T1 FLAIR precontrast images also show clot extending the length of the basilar artery (G). High-resolution T2-weighted images show a widened basilar artery with two lumens (H, black and white arrows) and an intervening flap. Full characterization of intracranial wall pathology may be able to reduce the need for diagnostic angiography for the diagnosis of intracranial dissections, as has been the case for extracranial dissections.
intensity than brain) in the vessel wall indicate methemoglobin. It remains to be seen whether T1 bright wall elements alone are specific for dissection or whether they may also be seen in vulnerable atherosclerotic plaque with hemorrhage. Using 3T MRI to image intracranial vessel walls addresses several technological concerns, but it also has its own limitations. First, the acquisitions are time-consuming and require cooperative patients with minimal movement. Second, to identify vessels of interest (and aid in the diagnostic workup of patients), baseline imaging (e.g., acute stroke protocols, CTA, MRA, or DSA) had already been performed. Wall imaging thus requires additional imaging sessions, which add expense and imaging time. Third,
proper vessel targeting requires interaction between an experienced and motivated MR technologist and a neuroradiologist to monitor the sequence acquisitions and ensure adequate coverage of the vessels. Given limited 3T MRI availability, these laborintensive studies are not routinely available at most centers. Fourth, this study was entirely retrospective and reviewed consecutive patients. Patients were selected to receive these additional imaging sequences based entirely on clinical requests by physicians for more information about the intracranial process. The population is thus biased toward the more complicated patients in whom clinicians or radiologists wanted further clarification about their disease. This selection bias means that our cohort may not be repNeurology 72
February 17, 2009
633
resentative of the full population of patients presenting to a neurologist or neurosurgeon with focal neurologic symptoms. The results presented here are hypothesis generating and must be interpreted with caution. Large, longitudinal, prospective studies of a broader cohort of patients using wall imaging are required to determine the sensitivity, specificity, and predictive values of this technique. AUTHOR CONTRIBUTIONS Data acquisition, statistical analysis, and manuscript writing were performed by R.H.S.
Received February 15, 2008. Accepted in final form November 14, 2008. REFERENCES 1. Yuan C, Mitsumori LM, Beach KW, Maravilla KR. Carotid atherosclerotic plaque: noninvasive MR characterization and identification of vulnerable lesions. Radiology 2001;221:285–299. 2. Wasserman BA, Wityk RJ, Trout HH III, Virmani R. Low-grade carotid stenosis: looking beyond the lumen with MRI. Stroke 2005;36:2504–2513. 3. Koops A, Ittrich H, Petri S, et al. Multicontrast-weighted magnetic resonance imaging of atherosclerotic plaques at 3.0 and 1.5 tesla: ex-vivo comparison with histopathologic correlation. Eur Radiol 2007;17:279–286. 4. Adams GJ, Greene J, Vick GW III, et al. Tracking regression and progression of atherosclerosis in human carotid arteries using high-resolution magnetic resonance imaging. Magn Reson Imaging 2004;22:1249–1258. 5. Bley TA, Uhl M, Venhoff N, Thoden J, Langer M, Markl M. 3-T MRI reveals cranial and thoracic inflammatory changes in giant cell arteritis. Clin Rheumatol 2007;26: 448–450.
6.
Cury RC, Houser SL, Furie KL, et al. Vulnerable plaque detection by 3.0 tesla magnetic resonance imaging. Invest Radiol 2006;41:112–115. 7. Bachmann RF, Nassenstein IF, Kooijman HF, et al. Spontaneous acute dissection of the internal carotid artery: high-resolution magnetic resonance imaging at 3.0 tesla with a dedicated surface coil. Invest Radiol 2006;41:105– 111. 8. Markl M, Uhl M, Wieben O, et al. High resolution 3T MRI for the assessment of cervical and superficial cranial arteries in giant cell arteritis. J Magn Reson Imaging 2006; 24:423–427. 9. Bley TA, Weiben O, Uhl M, et al. Assessment of the cranial involvement pattern of giant cell arteritis with 3T magnetic resonance imaging. Arthritis Rheum 2005;52: 2470–2477. 10. Klein IF, Lavallee PC, Touboul PJ, Schouman-Claeys E, Amarenco P. In vivo middle cerebral artery plaque imaging by high-resolution MRI. Neurology 2006;67:327–329. 11. Al-Kwifi O, Emery DJ, Wilman AH. Vessel contrast at three tesla in time-of-flight magnetic resonance angiography of the intracranial and carotid arteries. Magn Reson Imaging 2002;20:181–187. 12. Anumula S, Song HK, Wright AC, Wehrli FW. Highresolution black-blood MRI of the carotid vessel wall using phased-array coils at 1.5 and 3 tesla. Acad Radiol 2005;12: 1521–1526. 13. Yuan C, Zhang SX, Polissar NL, et al. Identification of fibrous cap rupture with magnetic resonance imaging is highly associated with recent transient ischemic attack or stroke. Circulation 2002;105:181–185. 14. Chen JW, Wasserman BA. Vulnerable plaque imaging. Neuroimaging Clin North Am 2005;15:609–21, xi. 15. Stoll G, Bendszus M. Inflammation and atherosclerosis: novel insights into plaque formation and destabilization. Stroke 2006;37:1923–1932.
Get Moving for the 2009 Run/Walk for Brain Research! Join your colleagues at the 2009 AAN Annual Meeting in Seattle for this exciting event benefiting neuroscience research. The (shoe) rubber meets the road on Tuesday, April 28, starting at 6:30 a.m. for a 5k run or mile long walk along the beautiful Seattle waterfront. Proceeds support Clinical Research Training Fellowships in neurology. Run up even more support by letting friends and family sponsor your run/walk for a flat donation. The more sponsors you recruit, the more raised for research. The runner with the most donations raised will receive FREE registration to the 2010 AAN Annual Meeting. Special prizes given for Best Male and Female Runners. Take your mark, get set, and GO now to www.aan.com/run to register!
634
Neurology 72
February 17, 2009
Impact of cardiac complications on outcome after aneurysmal subarachnoid hemorrhage A meta-analysis
I.A.C. van der Bilt, MD D. Hasan, MD, PhD W.P. Vandertop, MD, PhD A.A.M. Wilde, MD, PhD A. Algra, MD, PhD F.C. Visser, MD, PhD G.J.E. Rinkel, MD, PhD
Address correspondence and reprint requests to Dr. Ivo A.C. van der Bilt, Academic Medical Centre, Department of Cardiology, PO Box 22660, 1100 DD Amsterdam, The Netherlands
[email protected]
ABSTRACT
Impact of cardiac complications after aneurysmal subarachnoid hemorrhage (SAH) remains controversial. We performed a meta-analysis to assess whether EKG changes, myocardial damage, or echocardiographic wall motion abnormalities (WMAs) are related to death, poor outcome (death or dependency), or delayed cerebral ischemia (DCI) after SAH.
Methods: Articles on cardiac abnormalities after aneurysmal SAH that met predefined criteria and were published between 1960 and 2007 were retrieved. We assessed the quality of reports and extracted data on patient characteristics, cardiac abnormalities, and outcome measurements. Poor outcome was defined as death or dependence by the Glasgow Outcome Scale (dichotomized at ⱕ3) or the modified Rankin scale (dichotomized at ⬎3). If studies used another dichotomy or another outcome scale, we used the numbers of patients with poor outcome provided by the authors. We calculated pooled relative risks (RRs) with corresponding 95% confidence intervals for the relation between cardiac abnormalities and outcome measurements.
Results: We included 25 studies (16 prospective) with a total of 2,690 patients (mean age 53 years; 35% men). Mortality was associated with WMAs (RR 1.9), elevated troponin (RR 2.0) and brain natriuretic peptide (BNP) levels (RR 11.1), tachycardia (RR 3.9), Q waves (RR 2.9), STsegment depression (RR 2.1), T-wave abnormalities (RR 1.8), and bradycardia (RR 0.6). Poor outcome was associated with elevated troponin (RR 2.3) and creatine kinase MB (CK-MB) levels (RR 2.3) and ST-segment depression (RR 2.4). Occurrence of DCI was associated with WMAs (RR 2.1), elevated troponin (RR 3.2), CK-MB (RR 2.9), and BNP levels (RR 4.5), and ST-segment depression (RR 2.4). All RRs were significant.
Conclusion: Markers for cardiac damage and dysfunction are associated with an increased risk of death, poor outcome, and delayed cerebral ischemia after subarachnoid hemorrhage. Future research should establish whether these cardiac abnormalities are independent prognosticators and should be directed toward pathophysiologic mechanisms and potential treatment options. Neurology® 2009;72:635–642 GLOSSARY AF ⫽ atrial fibrillation; BBB ⫽ bundle branch block; BNP ⫽ brain natriuretic peptide; CI ⫽ confidence interval; CK-MB ⫽ creatine kinase MB; DCI ⫽ delayed cerebral ischemia; GCS ⫽ Glasgow Coma Scale; HR ⫽ hazard ratio; LVH ⫽ left ventricular hypertrophy; NT-proBNP ⫽ N-terminal prohormone brain natriuretic peptide; pt ⫽ patient; RR ⫽ relative risk; SAH ⫽ subarachnoid hemorrhage; STROBE ⫽ Strengthening the Reporting of Observational Studies in Epidemiology; WFNS ⫽ World Federation of Neurosurgical Societies; WMA ⫽ echocardiographic wall motion abnormality.
Case fatality after aneurysmal subarachnoid hemorrhage (SAH) is reported from 30% to 50%.1 The main causes of death are the impact of the initial bleeding and neurologic complications, such as rebleeding and delayed cerebral ischemia (DCI). Apart from the neurologic complications, cardiac abnormalities occur frequently after SAH. These abnormalities include EKG changes, elevated biochemical markers of myocardial damage and heart failure, and decreased left ventricular function. Supplemental data at www.neurology.org From the Departments of Cardiology (I.A.C.v.d.B., A.A.M.W.) and Neurosurgery (W.P.V.), Academic Medical Center, Amsterdam; Department of Intensive Care (D.H.), Viecuri, Venlo; Department of Neurology (A.A., G.J.E.R.) and Julius Centre for Patient Oriented Research (A.A.), University Medical Center Utrecht; and Department of Cardiology (F.C.V.), Erasmus Medical Center Rotterdam, The Netherlands. Disclosure: The authors report no disclosures. Copyright © 2009 by AAN Enterprises, Inc.
635
We hypothesized that cardiac abnormalities after SAH are related to the occurrence of death, poor outcome, and DCI. Therefore, we performed a meta-analysis on observational studies to assess the associations of cardiac complications with death, poor outcome, and DCI after aneurysmal SAH. METHODS Search strategy. Three investigators (I.A.C.v. d.B., D.H., and F.C.V.) independently performed a systematic search for studies regarding SAH accompanied by cardiac abnormalities, published between January 1960 and January 2007, using the electronic search engine PubMed. The following key words were used: SAH* OR SAB* OR subarachnoid hemorrhage* OR subarachnoid hemorrhage* OR subarachnoid bleed* OR subarachnoid blood* OR intracranial aneurysm* OR intracranial bleed*. All these key words were combined with the key words ECG*, electrocardiographic*, electrocardiography*, echocardiography*, echocardiographic, stunning*, myocardial damage*, myocardial necrosis*, left ventricular dysfunction*, LV dysfunction*, takotsubo*, apical ballooning*, CK*, CPK*, CKMB*, MB*, troponin, BNP*, brain natriuretic peptide*, and NT-pro-BNP* in different combinations. Reference lists were manually cross-checked for additional publications. This procedure was followed until no additional studies were found. Because many publications were more than 10 years old, no efforts were made to contact authors in case of missing data.
Eligibility. Two investigators (I.A.C.v.d.B. and D.H.) assessed eligibility of studies independently. Criteria for inclusion of studies in this review were publication after 1960 in the English, French, or German language. All studies had to report on cardiac abnormalities and outcome after aneurysmal SAH. Cardiac abnormalities were defined as echocardiographic wall motion abnormalities (WMAs), diastolic dysfunction, biochemical evidence of myocardial damage (defined as elevated troponin levels or elevated creatine kinase MB [CK-MB] levels), elevated brain natriuretic peptide levels (brain natriuretic peptide [BNP] and N-terminal prohormone BNP [NT-proBNP]), or EKG changes. SAH had to be documented by either CT scanning or CSF examination. Studies with fewer than 10 patients, case reports, and reviews were excluded. To avoid selection bias, only studies that included consecutive patients were eligible. When a study group published more articles on the same data set, only the report with the largest number of patients was eligible for data extraction.
Quality assessment of studies. To systematically assess the quality of the studies, we modified the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist (https://www.strobe-statement.org). Although this checklist is designed to improve the reporting of observational studies for optimal data extraction and interpretation, we used the list as a quality assessment tool. For every article, each of the 22 STROBE items was assigned a 0 or 1 by two independent observers (I.A.C.v.d.B. and F.C.V.) and summed as the STROBE score. Several STROBE items consist of subitems. These subitems were also scored as 0 or 1 and averaged. Disagreement was solved by direct communication between the two observers. The following items were assessed: cohort identification in title or abstract possible (item 1a); abstract informative and addressing key items (item 1b); background/rationale re636
Neurology 72
February 17, 2009
ported (item 2); objectives given with hypothesis (item 3); study design given (item 4); setting, locations, and data collection period given (item 5); inclusion and exclusion criteria given and sources and methods of selection of patients given (item 6a); period and methods of follow-up given (item 6b); variables of interest defined (cardiac data and outcome measures) (item 7); method of assessment of variable of interest given (item 8a); comparability of assessment methods across groups given (item 8b); sources of bias given (item 9); rationale for sample size given (item 10); statistical methods given (item 11a); missing data described (item 11b); subgroup and sensitivity analysis given (item 11c); analysis of quantitative variables given (item 12a); continuous and grouped analysis presented (item 12b); source of funding and role of funders presented (item 13); number of patients at each stage of the study given (item 14a); reasons for nonparticipation given (item 14b); period of recruitment defined (item 14c); baseline characteristics given (item 15a); completeness of data for each baseline variable given (item 15b); average and total amount of follow-up given (item 15c); number of outcome events presented (item 16); association between determinants and outcomes given (item 17a); categories of quantitative variables compared (item 17b); absolute outcome data given (item 17c); subgroup analysis performed (item 18); key results summarized (item 19); limitations discussed (item 20); external validity of study findings discussed (item 21); overall interpretation of the results given (item 22).
Data extraction. The three investigators who assessed quality and eligibility reviewed the publications independently. In case of disagreement, the authors reviewed the article in question together until consensus was reached. The following data were extracted: author, year of publication, study design (prospective or retrospective), definition of inclusion and exclusion criteria, number of included patients, sex, mean age, and follow-up period. The neurologic condition on admission was dichotomized as good or poor according to the scoring system used in the particular article: Hunt–Hess,2 World Federation of Neurosurgical Societies (WFNS),3 Glasgow Coma Scale (GCS),4 or Botterell.5 A poor condition on admission was considered when Hunt–Hess was ⱖ3, WFNS was ⱖ3, GCS was ⬍12, or Botterell was ⱖ3. Additionally, if an article did not use any of these scoring systems or used different criteria for poor outcome, the clinical condition reported by the authors was recorded. As determinants, we extracted the incidence of WMAs, diastolic dysfunction, elevated troponin levels, elevated NTproBNP levels, elevated CK-MB levels, atrial fibrillation, tachycardia, bradycardia, wandering P, extrasystoles, peaked P, P mitrale, short PR interval, long PR interval, bundle branch block, pathologic Q waves, ST-segment depression, ST-segment elevation, T-wave changes (inverted or flat), U-wave changes (present, inverted, or prominent), left ventricular hypertrophy, and prolonged QT time. Combined EKG criteria were disregarded, because the separate EKG abnormalities may harbor different prognostic values. For outcome measurements, we recorded the number of deaths from any cause, the number of patients with poor outcome, and the number of patients with DCI. Poor outcome was defined as death or dependence on activities of daily living, preferably defined by means of a handicap scale such as the Glasgow Outcome Scale (dichotomized at ⱕ3) or the modified Rankin scale (dichotomized at ⬎3). If studies used another dichotomy or another outcome scale, we used the numbers of patients with poor outcome provided by the authors, without adjusting to our preferred definition of poor outcome.
Figure 1
Search strategy explained
SAH ⫽ subarachnoid hemorrhage; pt ⫽ patient; WMA ⫽ echocardiographic wall motion abnormality; DCI ⫽ delayed cerebral ischemia; CK-MB ⫽ creatine kinase MB; BNP ⫽ brain natriuretic peptide.
Several studies used different definitions of DCI. Given this heterogeneity of definitions, we recorded the number of patients with DCI as given by the authors, without adjusting these numbers to a predefined definition of DCI.
Data analysis. The relation between the three outcome measures and the 22 determinants was analyzed. The crude proportions of the extracted variables were calculated. Cross-tables were made to calculate relative risks (RRs) for each determinant and outcome in each article. The pooled RRs with their corresponding 95% confidence intervals (CIs) were calculated by means of Cochrane’s Review Manager 4.2. Using the same program, statistical heterogeneity of the effects was tested. This is indicated as I2. A value greater than 50% may be considered substantial heterogeneity.6
with a median of 17. In 15 studies, data on the neurologic condition on admission were available. The proportion of patients with a poor neurologic condition on admission varied from 23% to 68%, with a mean of 49% (15 studies). The follow-up duration (for outcome assessment) varied from follow-up during hospital stay (14 studies) to 6 months’ follow-up. The proportion of patients who died during the observation period varied from 8% to 38%, with a mean of 22% (18 studies). The percentage of patients with poor outcome varied from 16% to 63%, with a mean of 46% (7 studies). Finally, the percentage of patients with DCI varied from 8% to 44%, with a mean of 23% (10 studies). As stated earlier, the definition of DCI varied7: scored temporary focal neurologic signs as DCI,8,9 equated imaging signs of vasospasm with DCI,10 used neurologic deterioration with imaging evidence of spasms,11-14 used neurologic deterioration with exclusion of other causes (using CT),15 used neurologic deterioration and imaging evidence of spasms or CT evidence of infarction,16 did not use a clear definition (clinical vasospasm). Table 2 shows the proportions of patients with cardiac abnormalities: WMAs varied from 13% to 31% of the patients, with a mean of 22% (9 studies). Diastolic dysfunction was reported in 1 study, in 71% of patients. Elevated troponin levels were reported in 21% to 50% of patients, with a mean of 34% (6 studies). CK-MB presence varied from 13% to 60%, with a mean of 33% (6 studies). Three studies reported elevated BNP levels, but each used different criteria for BNP elevation. Using these different criteria, elevated BNP levels were found between 9% and 100%. Table 3 summarizes prevalence of any EKG abnormality that was reported in the studies. T-wave changes were the most frequently observed EKG abnormality (22%), whereas pathologic Q waves were present in 1%. Relation of determinants with outcome. Figure 2A
In the initial PubMed search, 2,169 studies were found that reported on cardiac abnormalities after aneurysmal SAH (figure 1). After title screening and detailed abstract evaluation, 86 studies were selected. After detailed full article evaluation, 25 studies were included in this meta-analysis. RESULTS Study characteristics.
Baseline characteristics. Table 1 presents the baseline
characteristics of patients included in the analysis. A total of 2,690 patients were included. In the 21 studies where sex distribution was reported, the mean percentage of men was 35%. Mean age varied from 36 to 59 years, with a weighted mean of 53 years (21 studies). Sixteen studies were prospective, and 9 were retrospective. The STROBE score varied from 11 to 20,
shows the pooled RRs and corresponding 95% CI of the defined cardiac abnormalities for death. Additionally, heterogeneity of the data (I2) is presented. WMAs, elevated troponin and NT-proBNP levels, tachycardia, Q waves, ST-segment depression, and T-wave abnormalities were significantly associated with an increased risk of death. Bradycardia was significantly associated with a higher chance of survival. However, Q waves and ST-segment depression had significantly heterogeneity. Figure 2B presents the pooled RRs for cardiac abnormalities on DCI. WMAs; elevated troponin, CK-MB, and NT-proBNP levels; and ST-segment depression were significantly associated with an increased risk of the development of DCI. HeterogeneNeurology 72
February 17, 2009
637
Table 1
Baseline characteristics of the included studies
Men, %
Mean age, y
Prospective study
Strobe score
Patients with poor condition on admission (%)
Follow-up period
No. of deaths (%)
Patients with poor outcome (%)
No. of patients with DCI (%)
⫹
17
—
6 mo
35 (33)
—
—
—
Reference
Year
No. of patients
31
1965
106
46
—
32
1969
20
55
36
⫹
11
33
1976
100
—
—
⫺
13
7
1977
16
—
—
⫺
13
34
1983
17
47
53
⫺
11
—
2 (10)
—
—
—
In hospital
38 (38)
—
—
—
In hospital
—
—
—
—
48 (48)
35
1983
76
—
—
⫹
15
20
1986
23
48
46
⫹
16
36
1988
13
39
52
⫹
14
3 (23)
In hospital
8
1989
50
52
46
⫺
13
26 (52)
In hospital
In hospital 6 (26)
6 mo
—
—
14 (18)
—
—
7 (30)
—
—
3 (23) —
10
1997
13
—
56
⫹
11
3 (23)
In hospital
11
1999
72
35
51
⫺
20
26 (36)
In hospital
—
28
1999
313
31
55
⫺
17
9
2000
39
39
54
⫹
18
27
2002
122
46
59
⫺
17
16
2003
43
33
55
⫹
14
— 14 (36) — 25 (58)
3 (19)
5 (29)
12 (52) —
1 (8)
— 20 (40)
3 (23) —
4 (31) 26 (36)
In hospital
56 (18)
—
—
In hospital
5 (13)
—
16 (41)
In hospital
42 (34)
—
In hospital
7 (16)
27 (63)
— 19 (44)
37
2003
97
62
37
⫺
18
—
21 (22)
—
—
12
2003
38
45
49
⫹
17
14 (39)*
6 mo
—
—
14 (37)
29
2004
159
30
50
⫺
19
64 (40)
3 mo
43 (27)
38
2005
223
37
54
⫹
17
98 (44)
In hospital
—
—
13
2005
253
28
55
⫹
17
171 (68)
117 (46)
30
2005
68
22
—
⫹
17
—
32 (47)†
3 mo
70 (28)
3 mo
—
54 (34)
40 (59)
— — 33 (13) —
39
2006
173
32
54
⫹
18
—
8 days
31 (18)
—
—
40
2006
300
32
55
⫹
18
145 (48)
—
39 (13)
—
—
14
2006
121
26
55
⫹
18
15
2006
235
40
55
⫹
18
—‡
2,690
35
53
16/25
Median 17
733/1,511 (49)
Total
58 (48)
3 mo
—
In hospital
38 (16)
—
58 (48)
52 (43) 18 (8)
457/2,102 (22)
311/670 (46)
205/880 (23)
*Two patients missing in article, 36 patients used. †Poor outcome was defined as World Federation of Neurosurgical Societies ⱖ4. ‡Baseline characteristics in study given for 300 patients, but 150 are analyzed. DCI ⫽ delayed cerebral ischemia.
ity for the association of DCI and troponin and BNP levels was high. Figure e-1 on the Neurology® Web site at www. neurology.org shows the poor outcome data. Elevated troponin and CK-MB levels and STsegment depression were significantly associated with poor outcome. This meta-analysis patently indicates that cardiac abnormalities after SAH are related to death, poor outcome, and DCI. Although the main causes of death and poor outcome after SAH are the initial hemorrhage and neurologic complications, we found that also WMAs, troponin, CKMB, BNP, Q waves, ST-segment depression, and T-wave abnormalities are associated with death and poor outcome. These markers for cardiac
DISCUSSION
638
Neurology 72
February 17, 2009
damage and dysfunction are often present after SAH and are similar to those observed in ischemic heart disease, but the underlying pathophysiologic mechanism is probably different. Several mechanisms for the occurrence of cardiac complications after SAH have been suggested, but none is proven. However, a generally accepted hypothesis is that sympathetic stimulation induces catecholamine release in the myocardium, which may lead to impaired systolic and diastolic function, repolarization abnormalities, and myocardial damage. In ischemic heart disease, QT prolongation and STsegment elevation are associated with death. Against our expectations, these EKG abnormalities were not associated with death or poor outcome in patients with SAH. Possible explanations for the lack of association are that criteria used for QT prolongation in the articles were
Table 2
Prevalence of cardiac abnormalities
Reference
WMA
Diastolic dysfunction
1 Troponin
1 CK-MB
1 BNP
31
—
—
—
—
—
32
—
—
—
—
—
33
—
—
—
—
—
7
—
—
—
7 (44)
—
34
—
—
—
7 (41)
—
35
—
—
—
20
—
—
—
36 8
4 (31) —
— 7 (30)
— —
—
—
—
—
—
—
30 (60)
—
10
2 (15)
—
—
—
13 (100)
11
9 (13)
—
—
15 (21)
—
—
—
—
—
28 9 27 16
— 5 (13) — 7 (16)
— — —
8 (21) — 12 (28)
5 (13)
—
—
—
—
—
37
—
—
—
—
—
12
—
—
—
—
17 (45)*
29
—
—
—
—
—
38
—
146 (71)
—
—
—
13
55 (22)
—
126 (50)
—
—
30
—
—
35 (51)
—
—
39
48 (28)
—
41 (24)
—
—
40
79 (26)
—
—
—
—
14
—
—
—
—
—
15
45 (19)
—
—
14 (9)†
Total
254/1,141 (22)
146/223 (71)
71/217 (33)
44/286 (15)
52 (22) 274/811 (34)
Data are presented as number (percentage). *Brain natriuretic peptide (BNP) ratios were used. †BNP level ⬎600 pg/ml was used as cutoff value. WMA ⫽ echocardiographic wall motion abnormality; CK-MB ⫽ creatine kinase MB.
heterogeneous and arrhythmias as a cause of death were not reported in the studies. Bradycardia was associated with decreased risk of death. This corroborates with the results of several publications that suggest a beneficial effect of -blockade on outcome after SAH.17-20 However, it is unclear whether this beneficial effect was due to decrease of heart rate or due to systemic or neuroprotective effects of the -blockade. In contrast, tachycardia was associated with higher risk of death. Tachycardia may be a sign of poor hemodynamic condition or inotropic stimulation. Additionally, the presence of a P mitrale, a sign of left atrial dilatation, was associated with death. This abnormality may also represent a poor hemodynamic condition. We also found a relation between cardiac abnormalities and DCI after SAH. DCI is a complication that occurs in around 30% of patients with SAH,
usually between 4 and 12 days after the SAH,21 and is an important contributor to poor outcome. In contrast to thromboembolic stroke, which has a sudden onset, is unifocal, and usually does not affect consciousness, DCI usually has a gradual onset, often with waxing and waning focal deficits, a decreasing level of consciousness, or both, and often is multifocal. The pathogenesis of DCI has not been elucidated yet but is often attributed to vasospasm of the intracranial arteries. However, vasospasm cannot be the only initiator of DCI, because one-third of patients with severe vasospasm do not develop DCI, and onethird of patients with DCI do not have vasospasm.22 Powerful and independent predictors are the duration of loss of consciousness at time of the ictus and the total amount of extravasated blood,23 and the occurrence of hypovolemia and hypotension.24 Because many patients with SAH have narrowed arteries and hypovolemia and also because autoregulation of cerebral perfusion is disturbed after SAH,25,26 left ventricular dysfunction may directly affect cerebral perfusion. This is a potential explanation for the finding that cardiac abnormalities are also related to DCI. With this respect, not only the presence but also the degree of the cardiac abnormality may influence outcome. This notion is supported by the finding of a linear relation between the BNP levels and vasospasm severity.12 Additionally, the degree of troponin elevation has been associated with poor outcome.13 Although an association has been established, it remains unclear whether a definite causal relation exists between cardiac abnormalities and outcome after SAH. Several studies have found an independent effect of cardiac abnormalities on outcome, adjusted for clinical variables,11,15,27 whereas others did not.14,28-30 Caution should be perceived when interpreting these results because of shortcomings of the included studies. First, the included studies were published over a period of more than 40 years. During this period, diagnosis and treatment of SAH have improved with decreased case fatality rates,1 possibly affecting prevalence and consequence of cardiac complications on outcome. Second, reference values of cardiac markers were not given in all studies. Abnormal BNP levels were differently defined in the three studies that used BNP as prognosticator. One study defined QT prolongation as more than 410 msec, whereas others have used a cutoff value of more than 460 msec. T-wave abnormalities included both T-wave inversion and T-wave flattening. U waves were defined as present, greater than 1 mm, or negative. Moreover, some studies did not provide criteria for abnormalities at all. This obviously might influence reported prevaNeurology 72
February 17, 2009
639
640
Neurology 72
February 17, 2009
—
—
—
1 (6)
32
33
—
—
20
36
—
—
—
—
2 (5)
—
—
—
—
—
—
—
—
—
—
14/185 (8)
11
28
9
27
16
37
12
29
38
13
30
39
40
14
15
Total
33/378 (9)
—
—
—
—
—
—
—
9 (6)
—
—
7 (16)
—
—
—
—
—
—
—
—
7 (9)
—
—
108/808 (13)
—
—
—
—
—
—
—
17 (11)
—
26 (27)
2 (5)
—
—
4 (1)
—
—
—
—
—
21 (28)
—
—
32 (32)
6 (30)
—
Bradycardia
2/20 (10)
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
2 (10)
—
Wandering P wave
13/176 (7)
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
8 (11)
—
—
5 (5)
—
—
Extrasystole
6/335 (2)
—
—
—
—
—
—
—
2 (1)
—
—
—
—
—
—
—
—
—
—
—
2 (3)
—
—
2 (2)
—
—
Peaked P
8/248 (3)
—
—
—
—
—
—
—
4 (3)
—
—
—
—
—
—
—
—
—
1 (8)
—
3 (4)
—
—
—
—
—
P mitrale
21/226 (9)
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
16 (32)
—
—
4 (5)
—
—
1 (1)
—
—
Short PR
7/335 (2)
—
—
—
—
—
—
—
2 (1)
—
—
—
—
—
—
—
—
—
—
—
2 (3)
—
—
3 (3)
—
—
Long PR
37/391 (9)
—
—
—
—
—
—
—
6 (4)
—
—
9 (21)
—
—
—
—
—
—
1 (8)
—
9 (12)
—
—
12 (12)
—
—
BBB
47/846 (1)
—
—
—
—
—
—
—
16 (10)
—
—
5 (12)
8 (7)
—
8 (3)
—
—
—
1 (8)
—
6 (8)
—
—
3 (3)
0 (0)
—
Pathologic Q wave
126/1,081 (12)
—
17 (14)
—
—
—
—
—
9 (6)
—
5 (5)
5 (12)
35 (29)
—
13 (4)
—
—
18 (36)
—
—
15 (20)
—
—
9 (9)
—
—
ST-segment depression
65/872 (7)
—
10 (8)
—
—
—
—
—
21 (13)
—
7 (7)
1 (2)
10 (12)
—
9 (3)
—
—
—
—
—
—
—
—
7 (7)
—
—
ST-segment elevation
234/1,077 (22)
—
38 (31)
—
—
—
—
—
32 (20)
—
15 (16)
11 (26)
21 (17)
—
48 (15)
—
7 (54)
—
5 (39)
—
13 (17)
—
—
34 (34)
10 (