When Genes Shape the First Years of Multiple Sclerosis: What a Prospective Study Reveals About Relapses and Disability
Genome-wide association studies have robustly mapped susceptibility loci for multiple sclerosis (MS), particularly within immune-regulatory pathways, yet the genetic determinants of clinical course—conversion after a first demyelinating event, relapse propensity, and early disability accumulation—have remained comparatively elusive. A central challenge is methodological: clinical course is inherently longitudinal, while many large genetic studies are cross-sectional or case–control, limiting the ability to model time-to-event outcomes and early disease dynamics. Pan and colleagues address this gap by testing whether established MS-risk variants also predict near-term disease activity and disability progression when measured prospectively from the earliest clinical presentation.
Cohort framework: following participants from a first demyelinating event
The investigators leveraged the Ausimmune/AusLong programme, recruiting individuals at their first clinical presentation suggestive of central nervous system inflammatory demyelination (first demyelinating event, FDE) and following them for approximately five years with repeated neurological review. The present analyses focus on 127 participants with a “classic” FDE and available genotype data (125 for disability analyses). By five-year review, 68 participants (53.5%) had converted to clinically definite MS and experienced 151 neurologist-verified relapses; the cohort was predominantly female (77.2%) with a mean entry age of 37.8 years, and the median Expanded Disability Status Scale (EDSS) at five years was 1 (IQR 0–2).
Genotyping strategy and rationale for a targeted susceptibility-variant approach
Rather than pursuing a de novo genome-wide search for progression loci, the study adopted an efficient hypothesis-driven design: evaluate known MS susceptibility single nucleotide polymorphisms (SNPs) as predictors of subsequent clinical outcomes. DNA was genotyped using an Illumina Human Exome BeadChip with an MS-relevant custom component, and stringent quality control was applied (high call-rate thresholds, Hardy–Weinberg filtering, and population-structure assessment indicating a Caucasian sample without evident stratification outliers). The analysis extracted 110 previously published non-HLA MS-associated SNPs plus six HLA-tagging variants, yielding 116 candidate predictors in total, with proxy selection applied when necessary.
Endpoints and statistical modelling aligned to longitudinal disease biology
Three clinically meaningful outcomes were modelled: (i) time to conversion to definite MS (principally by dissemination in space and time using 2005 McDonald criteria), (ii) relapse risk using a recurrent-event survival framework, and (iii) annualised disability progression from FDE to five-year review (ΔEDSS), derived by assuming EDSS = 0 immediately before FDE and scaling the five-year EDSS by follow-up duration. Cox proportional hazards models (including a gap-time approach for relapses) and adjusted linear regression for ΔEDSS (with log transformation to satisfy model assumptions) were used, controlling for key confounders such as age, sex, study site, and relapse status at disability assessment. Bonferroni correction for 116 tests was applied, and the authors complemented p values with a permutation simulation to interrogate type I error.
Genetic predictors of conversion and relapse: signal consistency and a cumulative risk gradient
The study identified several susceptibility variants associated with conversion to MS and/or relapse risk over follow-up, including two non-HLA SNPs (rs12599600 near RMI2/PRM1 and rs1021156 near ZC2HC1A/PKIA) that predicted both outcomes (table on page 3). While individual SNP associations did not retain statistical significance after multiple-comparison correction, the directionality and internal consistency between related inflammatory outcomes (conversion and relapse) supported a polygenic interpretation. When the seven top conversion/relapse-associated SNPs were aggregated into a cumulative genetic risk score (CGRS), a pronounced risk gradient emerged: participants with ≥5 risk genotypes had approximately sixfold higher hazards of converting to MS (HR 5.98) and of relapse (HR 6.07) compared with those carrying ≤2 risk genotypes (table on page 4).
Genetic predictors of disability progression: a distinct polygenic architecture
A separate set of seven non-HLA variants predicted annualised disability progression (ΔEDSS), with no meaningful overlap with the conversion/relapse SNP set (table on page 5). This separation is biologically provocative because it suggests that early inflammatory activity (relapses and conversion) and disability accumulation may be influenced by partially distinct genetic pathways. The disability CGRS displayed a strong dose–response relationship (visualised in the line plot on page 6): compared with those carrying ≤2 disability-risk genotypes, individuals with ≥6 progressed by an additional 0.48 EDSS points per year on average, an effect size that—if sustained—would be clinically material across several years. Notably, the disability CGRS model explained 32.7% of the variance in disability progression within this cohort (R² = 0.327; ptrend ≈ 1.5×10⁻⁹), underscoring the potential informativeness of multi-variant models even when single-variant effects are modest.
Interpretation, limitations, and translational implications
The authors interpret their findings as evidence that MS susceptibility loci can influence early clinical course in a polygenic manner, with a plausible mechanistic split between pathways driving inflammatory disease activity (conversion/relapse) and pathways contributing to disability progression, potentially reflecting neurodegenerative components that are not tightly coupled to relapse biology. The work is strengthened by prospective inception-cohort design, neurologist-verified relapses, and disability assessment at a timepoint where EDSS stability is more likely; however, the modest genotyped sample size limits power for stringent multiple-testing correction and risks overestimation of effect sizes, making replication in independent longitudinal cohorts essential. If validated, cumulative genetic risk models could support earlier prognostication after FDE, refine risk stratification for monitoring intensity, and motivate functional studies of implicated loci as candidate targets for therapies tailored to distinct components of MS progression.
Disclaimer: This blog post is based on the provided research article and is intended for informational purposes only. It is not intended to provide medical advice. Please consult with a healthcare professional for any health concerns.
References:
Pan, G., Simpson, S., Van der Mei, I., Charlesworth, J. C., Lucas, R., Ponsonby, A. L., ... & Taylor, B. V. (2016). Role of genetic susceptibility variants in predicting clinical course in multiple sclerosis: a cohort study. Journal of Neurology, Neurosurgery & Psychiatry, 87(11), 1204-1211.
