Multiple sclerosis (MS) affects an estimated 2.9 million people worldwide, with Canada among the most affected countries, reporting prevalence exceeding 200 per 100,000 people [1], [2], [3]. Understanding long-term disability progression and factors associated with disease course remains crucial for patient counseling, treatment decisions, and healthcare planning. However, robust long-term data from well-characterized populations remain limited, particularly from Canadian cohorts where demographic, genetic, and healthcare system factors may influence disease outcomes differently. The MS registry at the Centre hospitalier de l’Université de Montréal (CHUM) represents one of Canada's earliest comprehensive MS databases, established in 1976 and now encompassing over 5,000 individuals across five decades. This extensive temporal span and large sample size provide a unique opportunity to examine disability progression patterns in the Canadian context while addressing methodological challenges that complicate interpretation of observational MS research.

Previous large-scale studies have consistently identified male sex, older age at onset, and frequent early relapses as predictors of faster disability accumulation [4], [5], [6], [7], [8], [9]. Disease-modifying therapies have demonstrated efficacy in reducing relapse rates and slowing disability progression, but observational treatment effect estimates are complicated by indication bias, where therapeutic decisions are influenced by unmeasured disease severity markers [10], [11], [12], [13], [14], [15].

The substantial proportion of women with pregnancy experience in MS cohorts highlights the clinical importance of understanding reproductive effects on disease progression. However, pregnancy studies show variable findings [16], [17], [18], potentially reflecting both genuine population differences and methodological challenges including healthy mother bias and selection effects.

Several methodological factors further complicate cross-study comparisons. The choice of time origin, whether disease onset or first clinical assessment, can substantially influence progression estimates, particularly in cohorts with heterogeneous diagnostic delays. While extensive covariate adjustment and propensity score methods can address some confounding, these approaches may introduce interpretational complexity that limits clinical translation.

The extensive follow-up period and large sample size of our Canadian cohort provides an opportunity to explore the consistency of prognostic associations across different analytical approaches. Given the inherent complexities of observational cohort studies, we performed sensitivity analyses to assess whether our findings were robust to methodological choices. This study leverages one of the longest-running MS registries in Canada to address two important objectives: (1) provide comprehensive Canadian data on long-term disability progression to complement international literature and inform local clinical practice; and (2) examine the consistency of findings across different analytical approaches in observational MS research. This approach allows us to contribute valuable Canadian natural history data while acknowledging the methodological challenges inherent in observational research.