Familial factors influence disability in MS multiplex families

Patients.

Families were identified by the French Multiple Sclerosis Genetics Group. Since 1991, neurologists from university and community hospitals, as well as private practitioners, have been contacted all over France to recruit sibpairs with MS. Informed written consent was given by each individual participating in the study in accordance with the Helsinki Convention (1964) and the French law relating to biomedical research. All patients were seen by at least one of us. Diagnostic criteria were those of Goodkin et al.,27 which were designed to minimize the risk of diagnostic errors in research on familial MS. Geographic and ethnic origin, sex, age at onset (defined by the first well-documented signs), duration of the disease, and disease course from onset were recorded. Disability was assessed with the Expanded Disability Status Scale (EDSS).28 Disease course was classified as relapsing-remitting (RR) (relapses with full recovery or with a residual deficit, and lack of progression between relapses), secondary progressive (SP) (initial RRMS followed by progression), or primary progressive (PP) (disease progression from onset).

A total of 132 families were referred to our group. Twenty-five families were excluded because one sib was unavailable or because there were no sibpairs. Eighteen families were excluded because at least one affected sib did not meet the research diagnostic criteria for familial MS.27 Two families were excluded after genotyping for a false paternity. A total of 87 families were available: 84 families with two affected sibs and 3 families with three affected sibs, for a total of 177 patients. There were 61 pairs with a disease duration of at least 5 years in both sibs. All families were of European origin and had lived in France for as long as family members could remember, except for two families who had emigrated from North Africa at least two generations before. HLA haplotypes were determined as previously described.16

Progression index.

Evaluation of disease severity in an MS population is difficult because patients are heterogeneous for disability and disease duration at first examination. The ideal method to evaluate disease severity would be a longitudinal follow-up study. In studies of the natural history of MS with a follow-up of large numbers of patients over a period of 10 to 20 years, disease severity was evaluated by the time taken to reach specific levels on the EDSS.29-31 We could not use this method because it would have introduced a bias toward concordance for disability because only individuals who reached a specific level would be taken into account. Pairs with one of the siblings not reaching a specific level would be excluded.

We therefore used the progression index (PI), defined as the ratio of the EDSS score to disease duration.32,33 The PI is relevant to the evaluation of disease severity if a linear relationship between the EDSS score and disease duration exists.32,33 However, this is not the case at the beginning of the disease. To minimize the denominator effect, Poser et al.30 excluded patients with a disease duration of less than 1 year. Between 1970 and 1980, 170 patients were studied and the PI was stable, demonstrating that the PI could be a marker of disability independent of the disease duration.30 More recently, Weinshenker et al.32,33 applied the PI to patients with a disease duration of more than 5 years and found a wide and approximately normal distribution of ranked severity scores. We used the same condition in our study, and MS pairs with a patient having a disease duration less than 5 years were excluded for the evaluation of the PI.

Statistical methods.

Patients were defined as individuals with MS whether they were probands or secondary cases. We randomly chose either member of each pair (i.e., a group of 87 unrelated patients) for the analysis of clinical features and of relationships between HLA status and clinical features to allow comparison with sporadic MS. Groups defined according to their HLA status were compared using Pearson’s chi-square or an unpaired t-test for each of the relevant criteria. The familial concordance analysis was carried out with a group of 61 pairs. The concordance of categorical criteria was assessed and tested through Cohen’s kappa coefficient.34 The concordance between siblings for quantitative criteria was assessed through an intra-class correlation coefficient. It was tested with the analysis of variance technique to establish that the variability within sibpairs was lower than the residual variability.34 The obvious positively skewed distribution of PI did not fit the theory, which requires a Gaussian distribution. However, the logarithm transformation of the variable yielded a more symmetric and fairly Gaussian histogram. As stressed by Robertson et al.,26 the presence of young sibpairs among patients can generate a bias toward concordance when looking for a pairwise concordance of any criterion linked to age at onset or disease duration. To avoid such a bias, we stratified the sample according to the mean age at onset of sibling pairs based on 10-year intervals.

Clinical features.

There were 119 females and 58 males (sex ratio: 2.05). The mean age at onset was 28.7 years, the mean disease duration was 15.3 years, and the mean EDSS score was 3.73 (table 1). The disease course was RR for 109 patients (61.3%), SP for 59 (33.3%), and PP for 9 (5.4%) (table 2). Similar clinical features were observed in all patients and in the group of 87 unrelated patients (see tables 1 and 2⇓).

HLA status and clinical features.

Of the 177 patients, 105 were HLA DR15 negative, 65 were heterozygous, and 7 were homozygous for HLA DR15. Clinical features (sex, age at onset, EDSS score, and PI) were similar in groups of patients defined according to their HLA status (data not shown) as well as the disease course (see table 2). Similar results were obtained when considering all patients or 87 unrelated patients (see table 2).

Concordance for age at onset, EDSS score, and PI in sibling pairs.

In the sibpairs, no significant concordance was found for sex or disease course (data not shown). Interestingly, a positive concordance was found between the age at onset (r = 0.204, p = 0.024) and the EDSS score (r = 0.221, p = 0.020). However, after stratification by mean age at onset for each pair as explained above, significance was not reached. The major observation of our study was a significant concordance for PI of the 61 sibpairs with a disease duration of at least 5 years (r = 0.234, p = 0.030).