You are here
Does cerebrospinal fluid analysis add predictive value to magnetic resonance imaging for long term irreversible disability in patients with early multiple sclerosis?
Journal of the Neurological Sciences, 1-2, 354, pages 51 - 55
The independent prognostic value of cerebrospinal fluid analysis in multiple sclerosis is not established.
To determine the prognostic value of intrathecal synthesis in a cohort of patients with relapsing-onset MS taking into consideration demographic and imaging parameters.
In this prospective cohort study conducted from 1993 to 2013, we analyzed the time to confirmed disability (persistent above 6 months) and irreversible disability (persistent for the entire disease course) of two disability milestones, Expanded Disability Status Scale score ≥ 4 or 6, and the time to secondary progressive onset in 579 patients with relapsing-onset multiple sclerosis. Demographic parameters (age at onset, gender) and imaging parameters (periventricular lesions) were included in the Cox models.
447 patients (77.2%) had intrathecal synthesis (oligoclonal bands and/or increased immunoglobulin G index value). No statistically significant relation was found between intrathecal synthesis and the time to reach each disability milestone or secondary progressive onset. An age older than 40 years and more than 3 periventricular lesions predicted a worse prognosis.
Cerebrospinal fluid analysis did not predict the time to disability milestones in relapsing-onset multiple sclerosis independently of age and imaging data.
- Data from a prospective cohort study conducted from 1993 to 2013 were analyzed.
- Age at onset, gender and imaging parameters were included in the Cox models.
- End-points were time to reach confirmed and irreversible disability (EDSS ≥ 4 or 6).
- No relation was found between intrathecal synthesis and the disability milestones.
- Older age and more than 3 periventricular lesions predicted a worse prognosis.
Keywords: Multiple sclerosis, Cerebrospinal fluid, Magnetic resonance imaging, Prognosis, Cohort, Disability.
Multiple sclerosis (MS) is the most common disabling neurological disease that occurs in young adults. The course of the disease and the risk for developing permanent disability are very different from one patient to another and predicting long term disability is not possible in a patient presenting with a new diagnosis of MS. Several long term clinical studies have been conducted to determine the clinical predictors of disability accumulation and some consistent results have been found. A review of natural history studies performed before 2009 concluded that negative prognostic factors include progressive disease type and disability at 2 and 5 years when considering all clinical phenotypes  . In the same study, the negative prognostic factors for relapsing–remitting MS (RRMS) and secondary progressive MS (SPMS) were identified as the onset of progression, a higher early relapse rate, greater disability in the first 5 years, shorter interval to the second relapse and the involvement of more functional systems  . None of these predictors are available when a patient is presenting with a first neurological episode that is suggestive of MS, in so-called clinically isolated syndromes (CIS). The significance of other potential predictors, such as an older age of onset or male gender, are equivocal but some studies have found that they are significantly associated with a worse outcome, , , , , and .
Among the parameters available at the CIS stage that have a potential prognostic value, brain magnetic resonance imaging (MRI) is of particular interest. MRI data collected at the onset of the disease have been shown in two longitudinal studies to be prognostic markers of interest, , and . The London CIS study, which followed a group of patients with CIS over 20 years, yielded important results in this regard and . T2 lesion volume at baseline and at several other time points during follow-up moderately predicted an increase in two disability scores: the Expanded Disability Status Scale (EDSS) and the Multiple Sclerosis Functional Composite (MSFC) and . Another cohort study consisting of patients presenting with CIS followed for a median of 7 years, identified the baseline number of Barkhof criteria and the baseline number of brain lesions as predictors of the EDSS score at year 5  . A further potentially predictive parameter available to CIS stage patients is cerebrospinal fluid (CSF) analysis. The presence of intrathecal synthesis of immunoglobulin G (IgG) is frequent in MS patients and is a strong predictor of conversion from CIS to MS  . Intrathecal synthesis can be demonstrated by identification of oligoclonal bands (OB) detected in the CSF but not in the serum or an increased IgG index  . Data concerning the prognostic value of CSF analysis are more equivocal. A recent meta-analysis suggested that OB could predict disability outcomes  . However, the results did not take into account several negative studies. It is not currently known whether CSF data add any prognostic information to MRI parameters.
The aim of this study was to determine the prognostic value of the presence of intrathecal synthesis (OB or increased IgG index in the CSF) in a cohort of patients with relapsing-onset MS followed in one center since 1993, taking into account early MRI results and demographic data as covariates.
2. Patients and methods
2.1. Patients and data collection
Patients were included in this study after being identified through our local EDMUS (European Database for Multiple Sclerosis) database. This database software has been used since 1993 to follow individual patients affected by MS, regardless of the clinical course, starting from the first neurological examination in the MS clinic at Bordeaux University Hospital. The database is reported to the French Commission Nationale Informatique et Libertés, and patients are informed of the use of this database.
All data have been entered into the database prospectively from 1993 to 2013. However, some historical data, such as relapses that occurred prior to the first visit to the MS clinic, have been added retrospectively based on patients' interviews and referring physician information at the first visit. These data cannot be considered with the same level of confidence. Therefore, relapse dates and numbers were not included in the analyses. We not only considered all clinical, biological and imaging data obtained in our center but also considered those reported by other neurologists and local hospitals when available and clearly documented. Demographic and clinical variables, such as disease course, disability, biological parameters and MRI data have been analyzed retrospectively. At each medical consultation with a senior neurologist of the MS clinic, data entered in the database previously was checked.
To evaluate the representativeness of our cohort, the demographic characteristics and the proportion of patients with CSF results available were compared with the national French MS cohort (data provided by the Observatoire Français de la Sclérose en plaques, OFSEP).
Before the present analysis, all data used in the study were checked in the patients' medical files, which were considered source files.
The inclusion criteria for this study were as follow: relapsing–remitting (RR) or secondary progressive (SP) clinically definite multiple sclerosis (CDMS) according to the standard diagnostic criteria and , that was diagnosed between 1993 and January 2013, available CSF analysis results, and a detailed results of a brain MRI performed within two years after their first episode available.
Subjects were excluded from the analysis if they were suffering from CIS, primary progressive MS (PPMS), an unconfirmed diagnosis or an alternative diagnosis.
2.2. Patients' assessment
The primary endpoint was the time to disability onset (confirmed and irreversible).
The disability assessment was performed during each visit by senior neurologists, according to a French-adapted version of the Kurtzke EDSS  . Two milestones, reflecting the main steps of disease evolution that are easy to determine by the clinician, were taken into account: confirmed EDSS 4 (limited walking ability, but without aid or rest for more than 500 m), and confirmed EDSS 6 (ability to walk with unilateral support for no more than 100 m without rest). All patients were treated according to current recommendations. No patient received disease modifying therapy before the lumbar puncture.
Confirmed disability scores were only taken into account if they persisted for at least a six month period and were checked by a neurologist. This information was not retained if the disability score was recorded during a relapse. A relapse was defined as the occurrence, recurrence or worsening of neurological dysfunction symptoms lasting more than 24 h and usually ending with a remission (partial or complete). Symptoms occurring within 1 month were considered part of the same relapse. However, if EDSS decreased after a six month delay, disability was considered confirmed for at least 6 months but was not considered as irreversible. Irreversible disability is a sustained and confirmed disability and was defined as a steady EDSS after at least a six month delay and for the duration of the follow-up period.
Secondary progression was defined as the initial RR disease course followed by progression for at least 6 months, with or without occasional relapses, minor remissions, and plateaus  . The date of SPMS onset was determined by senior neurologists during individual consultations with patients. In a limited number of cases, missing EDSS scores were calculated based on detailed neurological evaluations of medical files when available, according to a published algorithm  .
2.3. Potential prognosis factors taken into account
To determine the role of CSF analysis and consequently the predictive value of intrathecal synthesis of immunoglobulin on multiple sclerosis-related disability, the presence of at least two IgG oligoclonal bands (OB) in the CSF but not in the serum, or an IgG index above 0.7 was considered pathological. The parameter taken into account was, then, the presence or not of an intrathecal synthesis. From 1993 to 2013, different techniques have been used to detect intrathecal synthesis. All of them met standards found in hospital laboratories and are concordant with actual literature  . Before 2000, high resolution agarose gel electrophoresis was used. In 2000 sensitized immunofixation electrophoresis with paired CSF and serum was introduced. Isoelectrofocusing on agarose gel performed on semi-automatic HYDRASYS system, followed by immunofixation with anti-IgG serum was generalized after 2004. In the latter analysis, serum and CSF IgG are detected and identified using HYDRAGEL 9 CSF ISOFOCUSING SEBIA kit. Since these techniques have different sensitivity, we analyzed separately these three periods and did not take into account the absolute number of supplementary OB as a prognostic factor.
To study CDMS prognosis, we also studied baseline parameters known or suggested to have a prognostic value in early MS such as age at disease onset, gender and periventricular MRI lesions (≥ 3 or < 3) as covariates. MRI was performed according to technical standards. Neuroimaging study protocol included at least a whole-brain T1-weighted scan with intravenous injection of gadolinium-based contrast agent and a fast fluid-attenuated inversion-recovery (FLAIR) axial images or T2-weighted scan. Since the follow-up period lasted over twenty years, MRI acquisitions have changed over time, according to current recommendations. Image analysis was confirmed by senior neurologists.
2.4. Statistical analysis
The clinical characteristics of the Bordeaux EDMUS database and the national OFSEP database were compared using the chi square test, Fisher test and t test as required.
The predictive value of intrathecal synthesis and covariates on the time to occurrence of confirmed disability or irreversible disability was assessed using a Cox model for censored data. Disability milestones could be determined by clinical evaluation in 91% of cases. In the remaining patients the EDSS scores increased directly from one evaluation to the next one from a score inferior to the milestone to a score superior to the milestone. Considering the rarity of this event, interval censoring was not used.
Independent variables were the presence or absence of intrathecal synthesis, age at onset > or ≤ 40 years, female or male gender and periventricular lesions (≥ 3 or < 3), parameters previously shown to be predictive of CIS conversion to MS16. The dependent variables used in the analysis were the following: time to confirmed and irreversible disability onset defined by an EDSS ≥ 4 or 6. The secondary progressive course was studied as a disability stage as well. Statview ® (version 5.0 for Windows ®) software was used for the analysis.
On January 1, 2013, 579 patients met the inclusion criteria. Among these patients 84 had SPMS.
3.1. Subject characteristics
Among the 579 patients extracted from the EDMUS database and included in analysis, the sex-ratio was 3.4. The mean age at disease onset was 32.6 years, and the mean follow-up time was 88.0 months (7.3 years). The number of patients reaching each disability end-point and the delays of occurrence are summarized in Table 1.1 . The comparison of demographic characteristics of our cohort with the national French MS cohort (OFSEP) is presented in Table 1.2 . No statistical difference was observed between the two cohorts.
|N = 579||Mean delay from onset to disability endpoint (years)|
|Gender: women||449 (77.5%)|
|Age at onset ≤ 40 years||452 (78,1%)|
|Intrathecal synthesis||447 (77.2%)|
|≥ 3 periventricular lesions||347 (59.9%)|
|EDSS ≥ 4||145 (25.1%)||5.83|
|Irreversible EDSS ≥ 4||120 (20.7%)||6.61|
|EDSS ≥ 6||66 (11.4%)||7.49|
|Irreversible EDSS ≥ 6||57 (9.8%)||8.12|
EDSS = expanded disability scale; SPMS = secondary progressive multiple sclerosis.
N = 579
|Bordeaux's total cohort
N = 2403
|Total EDMUS cohort
N = 31 486
|Mean age at onset||32.60 years||32.38 years||33.68 years|
|Disease duration||7.33 years||13.44 years||13.21 years|
|Availability of CSF results||100%||54.47%||53.10%|
3.2. Prognostic value of CSF analysis on time to disability
Table 2 and Table 3show the results of the Cox models concerning the predictive value of CSF, MRI and demographics on disability progression. Concerning the whole population, the mean times from MS onset to EDSS 4 or 6 (confirmed or irreversible) or to secondary progression onset were not predicted by intrathecal synthesis.
P < 0.05.
P < 0.01.
EDSS = expanded disability status scale; SPMS = secondary progressive multiple sclerosis; CSF = cerebrospinal fluid; MRI = magnetic resonance imaging; HR = hazard ratio; ITS: intrathecal synthesis of IgG; PV ≥ 3: presence of 3 or more periventricular lesions on MRI.
P < 0.05.
P < 0.01.
P < 0.001.
EDSS = expanded disability scale; SPMS = secondary progressive multiple sclerosis; CSF = cerebrospinal fluid; MRI = magnetic resonance imaging; HR = hazard ratio. ITS: intrathecal synthesis of IgG; PV ≥ 3: presence of 3 or more periventricular lesions on MRI.
Gender did not predict these disability end-points either, except for secondary progressive outcome predicted by male gender.
However, an age at onset older than 40 years or having more than 3 periventricular lesions on an early MRI predicted long term reversible or irreversible disability but did not predict the onset of secondary progression or irreversible EDSS 6 disability when considering MRI data.
3.3. Subgroup analysis according to CSF technique analysis
All results were confirmed regardless CSF analysis (Table 2 and Table 3). 249 patients had CSF analysis after 2004, corresponding to isoelectric focusing which is considered most sensitive. Among them, intrathecal synthesis did not predict neither confirmed nor irreversible disability. Secondary progression was not influenced neither. These results coincide with those obtained with techniques used before 2000 and between 2000 and 2004.
Concerning, gender and age at onset results are also confirmed in the subgroup analysis. However, imaging study did not predict disability or progression in patients studied after 2000.
In the present dataset, neither the time to confirmed or irreversible disability at several EDSS milestones, nor the time of secondary progression onset was predicted by the presence of intrathecal synthesis in patients with relapsing onset MS when analyzed with other covariates such as MRI periventricular lesions, age and gender.
The potential prognostic value of oligoclonal bands on disability outcomes in MS have been examined in a few studies. A recent meta-analysis concluded that the presence of OB was predictive of reaching disability outcomes earlier  . An odds-ratio of 1.96 was reported based on 4 studies using isoelectrofocusing, , , and including one study concerning IgM OB, which is a different issue and . The results were mainly driven by a study performed retrospectively in 1488 Swedish registry patients and used unadjusted Cox-models to predict the age at attainment of an EDSS of 6.0  . None of these studies used other potential predictors of disability as covariates, except gender.
Several negative and important studies have been published. In a study of 415 CIS patients followed up to 50 (SD 31) months prospectively in Barcelona  , the time to EDSS 3.0 was not predicted by the presence of OB. Using the MS severity score (MSSS) and the progression index as outcome measures, a retrospective study of 451 patients did not show any significant predictive value of OB  . In a retrospective study of 143 MS patients followed for more than 5 years, the occurrence of secondary progression and increased EDSS during the follow-up period did not differ between OB positive and OB negative patients  . Recently, a study performed in 957 patients between 1982 and 2010 showed no significant differences between OB positive and negative patients in terms of the progression index and the proportion of patients reaching EDSS 6.0 within 10 years, in both PPMS and RRMS patients  . On the contrary, a worse prognosis in OB negative patients was reported in a study of 210 Turkish MS patients  . One study took MRI into account as a covariate, but it only concerned a sample of patients treated with interferon beta  . No differences were found.
The discrepancies between these studies could most likely be due to methodological differences and a potential bias because the use of non-confirmed disability as an outcome is of limited value.
In the present study we considered several covariates as potential predictors. Two of them significantly and independently predicted disability outcomes: age at onset older than 40 years or having more than 3 periventricular lesions on an early MRI predicted long term reversible or irreversible disability. Several natural history studies have been performed but not all found that an older age of onset was a negative prognostic factor  . We chose a cut-off of 40 years because it was found to be useful in the prediction of conversion to MS after CIS and . The explanation of the effect of age is not clear. It has been suggested that disability milestones are associated with age in general rather than with disease duration and . Another possible explanation is based on the higher frequency of comorbidity in older patients.
To analyze the potential prognostic value of early MRI, we used a very simple parameter that was easy to assess and is available in most of the files in our database, the presence of at least 3 periventricular lesions, which is one of the Barkhof criteria. This parameter has been previously shown to be predictive of CIS conversion to MS and . In the Barcelona CIS cohort, the number of Barkhof–Tintoré criteria at baseline was predictive of disability  . Interestingly, the frequency of abnormal MRI in that cohort was higher in patients with OB (85% of patients with 3 to 4 Barkhof criteria had OB versus only 31% of patients without any Barkhof criteria) suggesting that both parameters are not independent  .
This study has some limitations and strengths. First of all, some data were obtained retrospectively. However, the main end-points used (confirmed and irreversible disability) were acquired prospectively. Follow-up was done in real life conditions meaning that scheduled variations are to be considered as a potential bias, regarding the risk that people with more active disease might be more likely to present for ongoing clinical assessment. Although we included a large number of patients, the study took place in a single MS center, and some patients were lost to follow up. However, the characteristics of the sample are close to that of other studies and the national French MS cohort (OFSEP) regarding clinical features, disease course, CSF data, and mean disease duration, and the follow-up period is one of the longest among studies regarding the prognostic value of CSF analysis. Table 1.2 illustrates the comparison between our patients, Bordeaux's EDMUS and the whole French OFSEP cohort.
One possible other limitation is that between 1993 and 2013, CSF analysis techniques have changed, and this could lead to an underestimation of intrathecal synthesis in some patients. These techniques were standardized according to current recommendations and progress and are congruent with literature and . However, the subgroup analysis did not show significant results during the different studied periods whatever the technique used. Hazard ratios were higher in the most recent period, but only for EDSS 4 and the results were still not significant. Treatments were not included in our multivariate analysis, but the majority of patients have also been treated according to current recommendations.
The main strengths of the study were, first, the use of consolidated end-points like confirmed and irreversible disability, which were used for the first time in this type of research, and second, the analysis of other potential predictors of disability as covariates, in particular, early MRI data. This design allowed us to conclude that the demonstration of intrathecal synthesis does not seem to add prognostic information beyond MRI and age at the earliest stages of MS.
Conflict of interest statement
The authors did not report any conflict of interest involving the work under consideration for publication. Other relevant financial activities outside the submitted work were as follows: BB has received honoraria for speaking at scientific meetings and serving as member of scientific advisory boards for Bayer Pharma, Biogen-Idec, Merck Serono, Genzyme, Novartis and Teva. JCO has received consultancy fees, speaker fees, research grants (non-personal), and honoraria from Novartis, Biogen-Idec, Merck-Serono, Bayer Schering, Roche, Almirall, Teva and Genzyme Sanofi-Aventis.
AM and BB had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. It was performed as part of the cluster of excellence program, LabEx (ANR-LabEx 57 TRAIL).
-  A. Degenhardt, S.V. Ramagopalan, A. Scalfari, G.C. Ebers. Clinical prognostic factors in multiple sclerosis: a natural history review. Nat Rev Neurol. 2009;5:672-682 Crossref
-  C. Confavreux. Natural history of multiple sclerosis: a unifying concept. Brain. 2006;129:606-616 Crossref
-  M.P. Amato, G. Ponziani, M.L. Bartolozzi, G. Siracusa. 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 Crossref
-  C. Confavreux. Early clinical predictors and progression of irreversible disability in multiple sclerosis: an amnesic process. Brain. 2003;126:770-782 Crossref
-  B.G. Weinshenker, G.P. Rice, J.H. Noseworthy, W. Carriere, J. Baskerville, G.C. Ebers. 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
-  B.G. Weinshenker, M. Issa, J. Baskerville. Long-term and short-term outcome of multiple sclerosis: a 3-year follow-up study. Arch Neurol. 1996;53:353-358 Crossref
-  A. Scalfari, A. Neuhaus, A. Degenhardt, et al. The natural history of multiple sclerosis, a geographically based study 10: relapses and long-term disability. Brain. 2010;133:1914-1929 Crossref
-  L.K. Fisniku, P.A. Brex, D.R. Altmann, et al. Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain. 2008;131:808-817 Crossref
-  P.A. Brex, O. Ciccarelli, J.I. O'Riordan, M. Sailer, A.J. Thompson, D.H. Miller. A longitudinal study of abnormalities on MRI and disability from multiple sclerosis. N Engl J Med. 2002;346:158-164 Crossref
-  M. Tintoré, A. Rovira, J. Río, et al. Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology. 2006;67:968-972
-  R. Dobson, S. Ramagopalan, A. Davis, G. Giovannoni. Cerebrospinal fluid oligoclonal bands in multiple sclerosis and clinically isolated syndromes: a meta-analysis of prevalence, prognosis and effect of latitude. J Neurol Neurosurg Psychiatry. 2013;84:909-914 Crossref
-  H. Link, Y.-M. Huang. Oligoclonal bands in multiple sclerosis cerebrospinal fluid: an update on methodology and clinical usefulness. J Neuroimmunol. 2006;180:17-28 Crossref
-  C.M. Poser, D.W. Paty, L. Scheinberg, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol. 1983;13:227-231 Crossref
-  F.D. Lublin, S.C. Reingold. Defining the clinical course of multiple sclerosis: results of an international survey. National Multiple Sclerosis Society (USA) Advisory Committee on Clinical Trials of New Agents in Multiple Sclerosis. Neurology. 1996;46:907-911 Crossref
-  B. Brochet. Assessing incapacity at early stages of multiple sclerosis using the EDSS. Rev Neurol (Paris). 2009;165:S173-S179 Crossref
-  J. Mandrioli, P. Sola, R. Bedin, M. Gambini, E. Merelli. A multifactorial prognostic index in multiple sclerosis: cerebrospinal fluid IgM oligoclonal bands and clinical features to predict the evolution of the disease. J Neurol. 2008;255:1023-1031 Crossref
-  J.I. Rojas, S. Tizio, L. Patrucco, E. Cristiano. Oligoclonal bands in multiple sclerosis patients: worse prognosis?. Neurol Res. 2012;34:889-892 Crossref
-  M. Koch, D. Heersema, J. Mostert, A. Teeljen, J. De Keyser. Cerebrospinal fluid oligoclonal bands and progression of disability in multiple sclerosis. Eur J Neurol. 2007;14:797-800 Crossref
-  K. Imrell, E. Greiner, J. Hillert, T. Masterman. HLA-DRB1-15 and cerebrospinal-fluid-specific oligoclonal immunoglobulin G bands lower age at attainment of important disease milestones in multiple sclerosis. J Neuroimmunol. 2009;210:128-130 Crossref
-  P. Sola, J. Mandrioli, A.M. Simone, D. Ferraro, et al. Primary progressive versus relapsing-onset multiple sclerosis: presence and prognostic value of cerebrospinal fluid oligoclonal IgM. Mult Scler. 2010;17:303-311
-  M. Tintoré, A. Rovira, J. Río, et al. Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis?. Neurology. 2008;70:1079-1083
-  S. Siritho, M.S. Freedman. The prognostic significance of cerebrospinal fluid in multiple sclerosis. J Neurol Sci. 2009;279:21-25 Crossref
-  P. Lourenco, A. Shirani, J. Saeedi, J. Oger, W.E. Schreiber, H. Tremlett. Oligoclonal bands and cerebrospinal fluid markers in multiple sclerosis: associations with disease course and progression. Mult Scler. 2013;19:577-584 Crossref
-  E. Idiman, S. Ozakbas, Y. Dogan, G. Kosehasanogullari. The significance of oligoclonal bands in multiple sclerosis: relevance of demographic and clinical features, and immunogenetic backgrounds. J Neuroimmunol. 2009;212:121-124 Crossref
-  P. Annunziata, A. Giorgio, L. De Santi, et al. Absence of cerebrospinal fluid oligoclonal bands is associated with delayed disability progression in relapsing–remitting MS patients treated with interferon-β. J Neurol Sci. 2006;244:97-102 Crossref
-  A. Ruet, M.S.A. Deloire, J.-C. Ouallet, S. Molinier, B. Brochet. Predictive factors for multiple sclerosis in patients with clinically isolated spinal cord syndrome. Mult Scler. 2011;17:312-318 Crossref
-  A. Ruet, G. Arrambide, B. Brochet, et al. Early predictors of multiple sclerosis after a typical clinically isolated syndrome. Mult Scler. 2014;13:1721-1726
-  C. Confavreux. Age at disability milestones in multiple sclerosis. Brain. 2006;129:595-605 Crossref
a CHU de Bordeaux, Service de Neurologie, Bordeaux, France
b INSERM U 862, Université de Bordeaux, France
c OFSEP - Université de Lyon, France - Hospices Civils de Lyon, France
© 2015 Elsevier B.V., All rights reserved.