A, Neurological systems affected during the index attack leading to PLEX/IA treatment, stratified by immunopathological patterns. Data are presented as the percentage of patients with the specified functional system affected in relation to all patients with the indicated disease pattern. Consciousness includes somnolence, sopor, and coma; cerebral function includes aphasia, apraxia, and other impairments; cognitive function includes memory dysfunction and disorientation. Bowel/bladder involvement differs significantly between patients in patterns 1 and 2 (4 of 16, or 25%, affected vs 2 of 40, or 5%; P = .04). B, Functional response of different systems, stratified per disease patterns. Bars show the percentage of patients with moderate or marked improvement after apheresis treatment in relation to all patients with the respective pattern and functional system affected. For some patients, clinical data were not sufficient to judge therapy response in single functional systems. Differences in functional responses of the motor system between patterns 1 and 2 are significant (improved: 1 of 11, or 9%, in pattern 1, vs 13 of 27, or 48%, in pattern 2; P = .002).
A, Functional response to apheresis therapy is presented as the percentage of patients successfully treated with apheresis treatment, stratified according to pathological patterns; differences between patterns 1 and 3 and between patterns 2 and 3 are significant (P = .03 and P < .001, respectively). B, The percentage of patients with lesion improvement, as established by magnetic resonance imaging is shown; the difference between patterns 2 and 3 is significant (P = .03). C, The percentage of patients with an Expanded Disability Status Scale (EDSS) score in response to apheresis therapy is shown; the difference between patterns 2 and 3 is significant (P = .003). D, The median reductions in EDSS score within 1 month after apheresis therapy in therapy responders with pattern 1 and 2 lesions are shown. Bars and boxes define the minimum, 25th quartile, 75th quartile, and maximum changes in scores. the dots represent individual patients. Data to evaluate MRI and/or EDSS response were omitted when patient data were incomplete.
This figure presents estimates on functional response to apheresis therapy. Estimates were obtained by multivariate logistic regression using Firth correction. The estimated log odds ratio of relevant covariates, including penalized likelihood profiles-based 95% confidence intervals, are given per patient experience of moderate/marked functional improvement after apheresis. Covariates with a negative log odds ratio predict no response, while covariates with a positive log odds ratio predict therapy success. Estimates are significant when the 95% CI bars do not cross the log odd ratio 0. Multivariate adjustment included immunopattern; neurological system, brainstem, or cognitive involvement; therapy with immunoadsorption, disease duration, and apheresis treatment delay.
eTable. Antibody list
Customize your JAMA Network experience by selecting one or more topics from the list below.
Stork L, Ellenberger D, Beißbarth T, et al. Differences in the Reponses to Apheresis Therapy of Patients With 3 Histopathologically Classified Immunopathological Patterns of Multiple Sclerosis. JAMA Neurol. 2018;75(4):428–435. doi:10.1001/jamaneurol.2017.4842
Are there any differences in the response to apheresis therapy for steroid refractory relapses among patients with histologically defined immunopathological patterns of multiple sclerosis?
This cohort study of 69 patients defined 3 patterns of multiple sclerosis and observed functional improvement after apheresis therapy in patients with histological pattern 1 (31%) and pattern 2 (55%), but not in patients with pattern 3. Secondary outcome parameters (magnetic resonance imaging and expanded disability status scale improvement) strongly supported the primary outcome.
Response to apheresis therapy may be associated with immunopathological patterns and thus with pathological mechanisms of lesion development.
Plasma exchange and immunoadsorption are second-line apheresis therapies for patients experiencing multiple sclerosis relapses. Early active multiple sclerosis lesions can be classified into different histopathological patterns of demyelination. Pattern 1 and 2 lesions show T-cell– and macrophage–associated demyelination, and pattern 2 is selectively associated with immunoglobulin and complement deposits, suggesting a humoral immune response. Pattern 3 lesions show signs of oligodendrocyte degeneration. Thus it is possible that pathogenic heterogeneity might predict therapy response.
To evaluate the apheresis response in relation to histopathologically defined immunopathological patterns of multiple sclerosis.
Design, Setting and Participants
This single-center cohort study recruited 69 patients nationwide between 2005 and 2016. All included patients had a diagnosis of early active inflammatory demyelination consistent with multiple sclerosis; were classified into patterns 1, 2, or 3 based on brain biopsy analysis; and underwent apheresis treatments. Patients who had concomitant severe disease, neuromyelitis optica, or acute disseminated encephalomyelitis were excluded.
Main Outcomes and Measures
The primary therapy outcome was a functionally relevant improvement of the relapse-related neurological deficit. Radiological and Expanded Disability Status Scale changes were secondary outcome parameters.
The mean (SD) age of patients was 36.6 (13.3) years; 46 of the 69 participants (67%) were female. Overall, 16 patients (23%) exhibited pattern 1 lesions, 40 (58%) had pattern 2 lesions, and 13 (19%) had pattern 3 lesions. A functional therapy response was observed in 5 of the 16 patients with pattern 1 disease (31%) and 22 of the 40 patients with pattern 2 disease (55%), but none of the 13 patients with pattern 3 disease exhibited improvement (pattern 2 vs 3 P < .001). Radiological improvements were found in 4 (25%), 22 (56%), and 1 (11%) of patients with patterns 1, 2, and 3, respectively. The respective rates of response measured by changes in Expanded Disability Status Scale scores were 25%, 40%, and 0%. Brainstem involvement was a negative predictive factor for the functional therapy response (logarithmic odds ratio [logOR], −1.43; 95% CI, −3.21 to 0.17; P = .03), while immunoadsorption (as compared with plasma exchange) might be a positive predictive factor (logOR, 3.26; 95% CI, 0.75 to 8.13; P = .01).
Conclusions and Relevance
This cohort study provides evidence that the response to apheresis treatment is associated with immunopathological patterns. Patients with both patterns 1 and 2 improved clinically after apheresis treatment, but pattern 2 patients who showed signs of a humoral immune response benefited most. Apheresis appears unlikely to benefit patients with pattern 3 lesions.
Quiz Ref IDApheresis therapies, including plasma exchange (PLEX) and immunoadsorption (IA), are used to treat steroid-unresponsive multiple sclerosis (MS) relapses.1 Although the percentage of steroid-unresponsive cases is fairly low at 5%, PLEX and IA are important rescue therapies with response rates of 40% to 90%.2-6 The mechanism of action is assumed to be the removal of disease-causing agents such as antibodies and autoantibodies, immune complexes, and cytokines.7-9 As apheresis therapies are associated with potentially severe risks such as infections,10 predictors of treatment response are needed.
Quiz Ref IDHistopathologically early active demyelinating MS lesions can be classified into 3 main, intraindividually stable immunopathological patterns of demyelination (patterns 1, 2, and 3), which suggests pathogenic heterogeneity between individuals and may also predict therapy response to PLEX and IA.11-13 Patterns 1 and 2 share similar features of demyelination. However, only pattern 2 lesions are selectively associated with immunoglobulins and complement deposited along myelin sheaths and present within macrophages, which suggests an antibody and complement-mediated mechanism of action. Pattern 3 lesions show a more degenerative character and possibly reflect primary oligodendrocytic damage.14
Previously, an analysis of 19 patients showed that only patients with pattern 2 pathology responded to PLEX treatment, but none of the patients with patterns 1 or 3 did so.12 These results suggest that apheresis therapies may be beneficial for patients with pattern 2 lesions, which are characterized by a humoral immune response. This is in line with the known efficacy in antibody-mediated diseases such as neuromyelitis optica.15 We aimed to retrospectively analyze the PLEX and IA responses in a cohort of 69 patients with MS with histopathologically defined patterns of demyelination who were treated for steroid-unresponsive MS relapses.
This study was approved by the ethics committee of the University Medical Center Goettingen. Written informed consent was obtained from participants. All brain biopsies included in the analysis had been performed for differential diagnostic procedures in a clinical setting.
The study cohort was recruited from our German brain biopsy databank, which includes 774 patients nationwide with histologically proven inflammatory demyelination consistent with MS. Not all biopsied patients can be classified into patterns 1, 2, or 3, because the presence of early active demyelination, which represents the earliest stages of lesions, is a prerequisite for classification. Artifacts or insufficient material may also hinder classification. A total of 386 cases in the databank were classifiable into immunopathological patterns, and of these, 69 patients met the inclusion criteria, which were defined as (1) histopathological diagnosis of inflammatory demyelination consistent with early active demyelinating MS, classified into immunopathological patterns 1, 2, or 3 per published criteria11;Quiz Ref ID (2) at least 2 treatments with PLEX and/or IA for steroid-unresponsive acute attacks of MS; (3) clinical documentation available for the analysis of the index attack and the therapeutic outcomes of PLEX and/or IA treatments; and (4) no concomitant severe diseases (for example, mental retardation). The index attack was defined as the relapse leading to apheresis therapy. In addition to persons with severe comorbidities, patients with a histopathological diagnosis of neuromyelitis optica or acute disseminated encephalomyelitis, defined according to published criteria,16,17 were excluded. This cohort does not overlap with the cohort on which Keegan et al12 have reported previously.
Histological and immunohistochemical stainings were performed as indicated in the eTable in the Supplement. For the classification of lesions, the research team first determined the demyelinating activity based on published criteria describing early active demyelinating lesions containing myelin-laden macrophages immunoreactive for minor and major myelin proteins.18 Those lesions were subsequently classified into 1 of the immunopathological patterns 1, 2, and 3 by 2 board-certified neuropathologists (W.B. and I.M.), who were blinded to PLEX and IA treatment response.11
Data analysis was performed in a single center using a standardized protocol. Treatment response, magnetic resonance imaging (MRI) changes, and Expanded Disability Status Scale (EDSS) scores were assessed retrospectively by 1 study investigator (L.S.) and approved by a board-certified neurologist (I.M.), both of whom were blinded to immunopatterns. Clinical information was obtained from medical record review for all 69 participants. Radiological information before PLEX/IA treatment was available for 62 patients (90%) and after apheresis treatment for 46 patients (67%). Multiple sclerosis diagnosis and clinical course at the time of index PLEX/IA treatment were evaluated based on published criteria.19,20 Different neurological systems were evaluated to assess which deficits occurred with the index attack (Figure 1). Only new or worsening symptoms occurring with the index attack that influenced EDSS or significantly impacted function were considered.
The functional outcome of the PLEX/IA treatments was evaluated using an established grading system2,12 focused on changes in the affected (targeted) neurological systems. This system defines the response as (1) none (ie, no gain in neurological function); (2) mild gain (ie, subjective or minimal change, without changes in function); (3) moderate gain (ie, gain in neurological status that effects function); and (4) marked gain (ie, substantial difference from baseline with major gain in function). Changes in each neurological system were evaluated separately within 30 days after apheresis treatment.
Secondary outcomes included magnetic resonance imaging (MRI) and EDSS responses. Magnetic resonance imaging changes were investigated using T2-weighted images as well as gadolinium-enhanced T1-weighted images from primary centers that used different protocols. Scans by MRI at the time of index attack and before PLEX/IA treatment (a maximum of 3 weeks between the MRI and initiation of apheresis treatment) were analyzed for new or enlarging hyperintense T2-weighted lesions and gadolinium enhancement (1 or more index lesions). Within 30 days after PLEX/IA, MRI scans were evaluated for index lesion development. The EDSS scores were collected from charts or retrospectively estimated for 2 points: first, the highest EDSS score associated with the index attack before PLEX/IA treatment, and second, the EDSS score within 1 month after apheresis therapy. The EDSS treatment response was defined as a reduction in the EDSS score of 0.5 points or more for patients who had had an EDSS score of 6.0 or more at the time of the index attack, and a reduction of 1.0 or more for patients with an EDSS score of 5.5 or less before PLEX/IA treatment was started.21
Descriptive statistics are given for the cohort as a whole and for patients in each immunopattern strata. Comparisons for global group differences were made with the Fisher exact test and the Kruskal-Wallis test or 1-way analysis of variance. To adjust for relevant covariates, the association of the immunopatterns with treatment response in the primary and secondary outcomes were analyzed using logistic regression models. The Firth correction was used to avoid model fitting problems owing to very low response rates in some subgroups. Univariate and multivariate effect measures along with penalized likelihood profiles–based 95% confidence intervals as well as predicted probabilities for various subgroups are given. Selection of relevant covariates was initially done using least absolute shrinkage and selection operator (lasso) and subsequently by backward variable selection, eliminating less informative statistical variables to avoid overfitting. Only variables that significantly differed between the pattern strata were always kept in the model, unless major collinearities appeared. When addressing longitudinal measurements of serial PLEX/IA sessions within single patients, generalized estimation equations with a compound symmetry covariance structure were used to estimate whether a response is predictive for future responses while adjusting for relevant covariates. Statistical analyses were carried out with SAS, version 9.4 (SAS Institute) and R, version 3.1.2 (R Foundation for Statistical Computing). Two-tailed P values smaller than or equal to 0.05 were regarded as statistically significant.
Demographic data as well as clinical baseline characteristics of 69 eligible patients, stratified by immunopathological patterns, are summarized in Table 1. The groups showed no statistically significant differences in most demographical and clinical parameters listed. However, patients with pattern 3 lesions more frequently showed a clinically isolated syndrome (9 of 13 patients; 69%; in a comparison for all 3 groups, P = .02), whereas patients with pattern 2 lesions more often had a relapsing-remitting disease course (26 of 40 patients; 65%; P = .02). There was a tendency to longer time intervals between the start of the index attack and the apheresis therapy (PLEX/IA delay) in patients with pattern 3 lesions compared with patients exhibiting patterns 1 and 2 (P = .07). The disease duration (from new-symptom onset to apheresis therapy) tended to be longer for patients with pattern 2 lesions (P = .07). Subsequent analyses were corrected for these variables to exclude any possible influence on outcome measures.
At the time of apheresis therapy, more than two-thirds of the patients (51 of 69; 74%) had clinically definite MS according to the 2011 McDonald criteria.19 Index attack symptoms leading to PLEX/IA treatment are shown in Figure 1A. Most patients (58 of 69; 84%) presented with multifocal neurological deficits involving more than 1 functional system. A similar distribution of affected functional systems was observed when comparing the different pathological patterns.
Functioning improved after apheresis therapy in a subgroup of patients exhibiting patterns 1 and 2, but not in patients with pattern 3 disease. Quiz Ref IDTreatment success, defined as moderate or marked therapy response with a functional improvement of deficits, was found in 27 of 69 patients (39%) within 1 month of apheresis. This percentage is consistent with published data.22,23 Treatment success differed when comparing patients across the different immunopathological patterns (Figure 2A). The highest response rate to PLEX/IA was in patients with pattern 2 disease, 22 (55%) of whom experienced improvement compared with 0 of 13 patients with pattern 3 disease who experienced improvement (P < .001). One-third of the patients with pattern 1 disease also benefited from therapy (5 of 16 patients; pattern 1 vs 3 P = .03). Clinical improvement could be observed in most functional systems (Figure 1B).
Most pronounced lesion improvements on MRI were in patients with pattern 2 lesions, followed by patients with pattern 1 lesions. Changes to MRI results were analyzed as a secondary outcome. Lesion improvement, defined as lesions becoming smaller and/or showing less gadolinium enhancement, was found in 18 of 46 patients (39%). Patients with pattern 2 disease showed lesion regression more often (14 of 25; 56%) compared with patients with pattern 3 disease (1 of 9 patients; 11%; P = .03; Figure 2B). In patients with pattern 1 lesions, lesion improvement was observed in 3 of 12 patients (25%).
The overall rate of response as measured by EDSS was 28% (n = 19/67) and thus lower than the functional response observed in 39% of patients as described above. The EDSS focuses on the function of the lower extremities and has a poor sensitivity for disabilities that occur because of involvement of upper extremities or because of cognitive changes.24 Again, the highest EDSS response rate was found for patients exhibiting pattern 2 (15 of 38; 40%), followed by patients exhibiting pattern 1 (4 of 16; 25%). None of the patients with pattern 3 lesions showed an EDSS improvement (0 of 13; 0%; Figure 2C). Focusing on therapy responders, responders with pattern 2 lesions showed a median EDSS reduction of 0.5 points, while responders with pattern 1 lesions showed a median reduction of 1.5 points (Figure 2D). Thus, EDSS evaluation supports the conclusion that patients with patterns 1 and 2 lesions have a more favorable response to PLEX and IA than do patients exhibiting pattern 3, who showed no EDSS response.
To analyze which factors predict PLEX/IA functional response, various parameters were tested in univariate and multivariate logistic regressions (Figure 3). Immunopattern 1 (logarithmic odds ratio [logOR], 3.35; 95% CI, 0.57-8.59; P = .01) and pattern 2 (logOR, 5.61; 95% CI, 2.49-11.32; P < .001) were found to be positive predictive factors for therapy response. In addition, therapy with IA, compared with PLEX, turned out to be a positive predictive factor for patients in patterns 1 and 2 (logOR, 3.26; 95% CI, 0.75-8.1; P = .01). Clinical involvement of the cognitive system also positively influences therapy response, although with a lower effect size (logOR, 1.56; 95% CI, 0.03-4.37; P = .046). In contrast, involvement of the brainstem emerged as a negative predictive factor (logOR, −1.43; 95% CI, −3.21 to −0.17; P = .03). Thus, we found a high predicted probability that patients would benefit from apheresis, for example, if they had pattern 2 disease without brainstem involvement and were treated with IA (99%; 90% CI, 0.83-1.0; Table 2). It should be noted that in peripheral subgroups, the predicted therapy response rate might be overfitted.
Prior publications3,4,22,23 have found male sex, a shorter delay before PLEX initiation, a shorter disease duration, an EDSS of 5.0 or less at the time of PLEX initiation, and the presence of ringlike contrast enhancement on MRI to be associated with a better PLEX response, factors that were not predictive in our study. However, ringlike enhancement was significantly more often present in patients with pattern 2 lesions (13 of 39; 33%) compared with patients with pattern 3 lesions (0 of 13; 0%; P = .02). The absence of reflexes, which in prior studies was associated with a negative therapy response,20,22 was only observed in 1 patient with pattern 1 disease, and this patient showed no therapy response.
Fourteen patients received more than 1 PLEX/IA session. Longitudinal measurements of therapy responses to consecutive PLEX/IA sessions within 1 patient were not positively correlated (ρ = −0.27). Patients with pattern 3 disease did not respond to either the first or any following PLEX/IA sessions.
Multiple sclerosis is a heterogeneous disease with respect to its clinical, genetic, radiographic and pathological features. Three main immunopathological patterns of early active MS lesions have been described, with lesions showing an intraindividual pathological homogeneity and interindividual heterogeneity that persists over time.11,13 However, this concept has been a matter of debate, and a time-dependent heterogeneity of lesions has been suggested.25 The clinical and therapeutic relevance of these histopathologically defined patterns is still poorly understood.12 We analyzed the efficiency of apheresis therapy for steroid-resistant relapses in patients with MS, stratified according to their pattern of early demyelination and therefore in relation to their assumed pathophysiological mechanisms of lesion development. In a large cohort of 69 patients, we confirmed the successful apheresis treatment in more than 50% of patients with pattern 2 lesions. Importantly, we showed that, in addition, every third patient with pattern 1 pathology also responded to the PLEX/IA therapy. This is in contrast to a prior report,12 which had found no therapy response in patients with pattern 1 lesions, most likely because only 3 patients with pattern 1 lesions were analyzed. Patients with pattern 3 lesions did not display any treatment response.
Several studies support the heterogeneity of early MS lesions by observing distinct chemokine receptor profiles in pattern 2 compared with pattern 3 lesions, and by noting mitochondrial defects only in patients with pattern 3 pathology.26,27Quiz Ref ID The histopathology of pattern 3 lesions resembles white matter stroke, and the mitochondrial changes described in these lesions suggest a hypoxia-like tissue injury rather than an inflammation-driven pathogenesis.27 This might explain the nonresponse to PLEX/IA treatment.
In contrast, the pathological features of pattern 1 and 2 lesions are similar. They are only distinguishable from each other by the presence of immunoglobulin and complement deposits in patients exhibiting pattern 2, suggesting antibody/complement-mediated demyelination. Antigen microarrays analyzing central nervous system proteins, lipid autoantigens, and heat shock proteins identified unique serum antibody signatures in patients exhibiting pattern 1 vs pattern 2.28 However, specific pathogenic autoantibodies in patients with MS have not been identified yet, although myelin oligodendrocyte glycoprotein–IgG antibodies may be pathogenic in a low percentage of adult patients with pattern 2 lesions.29-32 Removal of antibodies and circulating immune complexes is suggested as a mechanism of action of apheresis therapies, reducing serum antibodies by 85% from preapheresis levels.15,33
In pattern 1 lesions, proinflammatory mediators such as cytokines and chemokines produced by activated microglia or macrophages and T cells were suggested to cause myelin damage.34 Elimination of these substances may be beneficial in patients with pattern 1 lesions, but data on the removal with PLEX are controversial.35 Cytokine levels were not lowered after PLEX in septic patients.36 In contrast, a reduction of interleukin 8 and tumor necrosis factor α cytokine levels was observed after PLEX therapy for thrombotic thrombocytopenic purpura.37 Fibrinogen and complement 3 were reduced in plasma after PLEX for MS relapses.38 Thus, elimination of factors other than antibodies may be relevant for the treatment outcomes observed in about one-third of patients with pattern 1 pathology.
In addition to the removal of pathological agents, changes in immune cell numbers, composition and activation after apheresis treatment can also be observed.39 Changes may occur either because of alterations in concentrations of soluble plasma factors, or because of the apheresis procedure itself.40
Both the immunopathological patterns 1 and 2, as well as the application of IA (compared with PLEX), were strongly associated with a beneficial outcome after apheresis treatment. Prior studies have shown significant clinical improvement after IA in 73% to 85% of patients with MS compared with 40% to 70% after PLEX.41 A combination of both therapies may even be more effective.42
Our study was limited by the low number of patients treated with IA (a total of 10, of whom 3 had pattern 1 lesions, 3 had pattern 2 lesions, and 4 had pattern 3 lesions), so that further studies are necessary to explore whether IA may even have superior treatment outcomes. Interestingly, IA not only removes antibodies but also other proteins that may be involved in MS pathogenesis.6
Previous studies3 reported that lesions with ringlike enhancement on MRI were associated with a beneficial therapy response to PLEX, and they correlate with a macrophage rim at the lesion border in pattern 1 and 2 lesions.43 Although we found a ringlike contrast enhancement significantly more often in patients with pattern 2 than in patients with pattern 3 (none of the pattern 3 patients showed ring enhancement), it was not independently associated with a favorable outcome. This may be related to the limited number of patients (16 of 69; 23%) with available MRI scans and ringlike contrast enhancement in our study.
Importantly, the immunopathological patterns can only be determined by histology, which is not routinely available. As there are inherent risks to biopsy, this should not be performed solely for apheresis treatment decisions. Further studies need to identify serological and MRI biomarkers to diagnose MS patterns without taking biopsies.
Another limitation of our study is that the longer delay to PLEX initation in patients with pattern 3 lesions could influence therapy outcome. However, the observed therapy responses in patients with pattern 1 and 2 up to 92 days after symptom onset argues against this possibility. Analyses were corrected for the PLEX delay.
A third limitation is that the biopsied cohort with MS is characterized by atypical clinical presentations, such as tumefactive lesions, which led to the neurosurgical diagnostic procedure. However, comparing patients with MS who undergo biopsy with a nonbiopsied cohort showed no significant differences in demographic parameters and subsequent clinical course.44 This suggests that, although biopsied patients are atypical in their first clinical presentation, studying them yields important pathological and clinical observations.
Finally, this study is based on brain biopsies and the question arises as to whether our results could be valid for other lesion locations, eg, in the spinal cord or optic nerve. As most of the patients in this study presented with multifocal symptoms involving these regions, it seems likely that these results are valid for other parts of the central nervous system.
In conclusion, this study shows that pathological differences in early MS lesions could not only explain different pathophysiological mechanisms and targets of tissue injury, but also variable responses to apheresis treatment. Therapy response is influenced by multiple factors, including histopathological patterns and possibly also factors associated with IA, brainstem involvement, and disturbed cognitive functions, all of which may be important for therapy planning.
Corresponding Author: Imke Metz, MD, Institute of Neuropathology, University Medical Center Goettingen, Robert-Koch-Strasse 40, 37075 Göttingen, Germany (firstname.lastname@example.org).
Accepted for Publication: September 15, 2017.
Published Online: February 5, 2018. doi:10.1001/jamaneurol.2017.4842
Author Contributions: Dr Metz and Ms Stork had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Metz, Brück and Lucchinetti.
Acquisition, analysis, or interpretation of data: Stork, Ellenberger, Metz, Brück.
Drafting of the manuscript: Stork, Ellenberger, Metz.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Ellenberger, Beißbarth, Friede.
Obtained funding: Brück, Metz.
Administrative, technical, or material support: Lucchinetti, Brück, Metz.
Study supervision: Metz.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by the German Federal Ministry of Education and Research (Bunderministerium fur Bildung und Forschung) and the German Competence Network Multiple Sclerosis (Kompetenznetz Multiple Sklerose), Pattern Multiple Sclerosis/Neuromyelitis Optica (Drs Metz and Brück).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We acknowledge Sven Müller, Institute of Neuropathology, University Medical Center, Goettingen, Germany, for his outstanding administrative support. No compensation from a funding sponsor was involved in his contribution.
Create a personal account or sign in to: