Polygenic Risk Scores and Lifetime Multiple Sclerosis Risk: Translating Genetic Susceptibility Into Absolute Probability
Genome-wide association studies have identified more than 200 common genetic variants associated with susceptibility to multiple sclerosis (MS), yet a persistent translational gap remains: odds ratios from association studies do not directly inform a person’s absolute probability of developing MS across the life course. Loonstra and colleagues addressed this gap by quantifying how a polygenic risk score (PRS)—a weighted aggregation of many risk variants—maps onto lifetime risk of MS in a population-based setting, rather than merely differentiating cases from controls. Their central premise is that if a PRS captures a meaningful fraction of inherited liability, then individuals in the extreme tails of the PRS distribution should show markedly different absolute risks, even when the overall disease incidence is relatively low.
A methodological advantage: an almost complete national birth-year cohort
A major strength of this work is its unusually rigorous ascertainment strategy. The investigators leveraged “Project Y,” which attempted to identify nearly all persons with MS born in 1966 in the Netherlands (239,611 live births) by combining nationwide outreach via MS-specialist neurologists, hospitals, patient associations, and care facilities, thereby reducing the selection biases common in clinic-based cohorts. From this birth-year cohort, they identified 452 persons with MS (approximately 73% female), yielding sex-specific lifetime risk estimates of 282 per 100,000 women and 100 per 100,000 men in that cohort. For genetic analyses, the MS-PRS was computed in 285 persons with MS and compared with 267 non-MS controls assembled from the Project Y cohort (non-MS) and age-matched non-MS participants from the Amsterdam Dementia Cohort restricted to similar birth years.
Genotyping, quality control, and construction of an MS-specific PRS
All participants were genotyped on the same array platform, followed by standard quality control, ancestry restriction to European genetic background, relatedness filtering, genome build harmonization, and imputation using a large external reference panel. The PRS itself was based on 215 genome-wide significant MS susceptibility variants drawn from the largest MS GWAS available to the authors, including variants in (and outside) the major histocompatibility complex (MHC). Each individual’s PRS was computed as the sum of risk alleles weighted by the log-odds effect sizes from GWAS summary statistics, then standardized (Z-scored) to facilitate interpretation. Conceptually, this approach treats MS susceptibility as highly polygenic, with many loci of small effect jointly shifting liability.
From relative genetic liability to absolute lifetime risk: the decile simulation framework
To translate PRS differences into lifetime risk, the authors assigned individuals with MS to PRS deciles defined by the PRS distribution in controls, then estimated sex-stratified lifetime risk within each decile. They implemented a simulation of 100,000 individuals per decile, calibrated to the sex-specific MS lifetime risk observed in the Dutch 1966 birth cohort, and assumed that the PRS distribution in controls approximates that of the general population. In effect, the procedure propagates the observed enrichment of MS cases across PRS deciles into an absolute-risk scale, enabling statements such as “1 in X” lifetime risk for people in the lowest versus highest genetic-risk strata. While this approach cannot eliminate all assumptions (notably about representativeness and competing risks), it provides an interpretable bridge between polygenic stratification and population incidence.
Core result: extreme PRS stratification yields large absolute-risk gradients, especially in women
The findings demonstrate pronounced risk stratification across the PRS distribution. Women in the lowest 30% of genetic risk had an estimated lifetime MS risk of approximately 1 in 2,739, whereas women in the highest 10% had an estimated lifetime risk of approximately 1 in 92 (≈1.1%). Men showed the same qualitative pattern but lower absolute risk at comparable PRS ranks: about 1 in 7,900 in the lowest 30% versus 1 in 293 (≈0.34%) in the highest 10%. Importantly, the relative contrast between the top and bottom strata was similar by sex (relative risk on the order of ~26–29), yet the absolute probabilities differed substantially, consistent with the higher baseline incidence of MS in women. These estimates are the most clinically legible outputs of the study, because they convert polygenic ranking into absolute lifetime probability within a defined population context.
Notably absent: PRS does not explain age at onset or progression to secondary progressive MS
Beyond susceptibility, the authors evaluated whether the same PRS associates with clinically relevant disease features: age at first symptom onset, age at secondary progression, and time from onset to secondary progression, using Cox proportional hazards models adjusted for genetic principal components (and sex where appropriate). After multiple-testing correction, the MS-PRS was not significantly associated with these outcomes in either sex, and stratified analyses across higher PRS deciles similarly failed to show robust effects. This negative result supports an emerging view in MS genetics: variants that influence susceptibility may be partly distinct from those shaping disease trajectory, progression, or central nervous system resilience. In other words, the genetic architecture of “getting MS” may differ from that of “how MS behaves after onset,” limiting the PRS’s utility for prognostication even if it is informative for risk stratification.
Clinical and research implications: decision support, diagnostic caution, and generalizability limits
The most immediate translational implication proposed by the authors concerns diagnostic support rather than population screening. A high PRS could incrementally strengthen diagnostic confidence in an individual with suspected MS, but the study emphasizes a particularly practical asymmetry: a PRS in the lowest tail of the distribution may function as a “red flag” against MS and could help reduce misdiagnosis in patients with MS-mimicking conditions. At the same time, several constraints must temper interpretation: the cohort is Dutch, birth-year specific (1966), and limited to individuals of European ancestry, and the PRS is derived primarily from European-ancestry GWAS signals—factors that may reduce portability to other ancestries, geographies, or birth cohorts with different environmental exposures and healthcare pathways. Finally, the analysis did not fully model competing risks such as mortality, though the authors argue this is unlikely to materially distort onset-focused estimates in their setting. Taken together, the paper provides a rigorous demonstration that polygenic burden can strongly modulate absolute lifetime MS risk, while also delineating the boundaries of what this particular PRS can—and cannot—predict in clinical neurology.
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:
Loonstra, F. C., Álvarez Sirvent, D., Tesi, N., Holstege, H., Strijbis, E. M., Salazar, A. N., ... & Uitdehaag, B. (2024). Association of Polygenic Risk Score With Lifetime Risk of Developing Multiple Sclerosis in a Population-Based Birth-Year Cohort. Neurology, 103(7), e209663.
