Age-Dependent Determinants of Multiple Sclerosis Diagnosis: Insights from the UK Biobank
In this article, Nova and colleagues address a central limitation in multiple sclerosis (MS) epidemiology: most prior studies have relied on retrospective case-control designs, which estimate lifetime associations but do not adequately capture how risk unfolds across age. To overcome this limitation, the authors apply a time-to-event framework to UK Biobank data, asking not only which factors are associated with MS, but also whether those factors influence when diagnosis occurs. This distinction is scientifically important because MS is a disorder of early and mid-adulthood, and the biological relevance of genetic susceptibility or early-life exposures may vary markedly across the life course. By shifting from odds ratios to hazard ratios, the study provides a more dynamic account of disease emergence and offers a more nuanced view of MS susceptibility.
Study Design and Analytical Strategy
The investigators analyzed 345,027 unrelated White participants born and recruited in England, among whom 1,669 had an MS diagnosis. They treated age as the time scale and followed individuals from birth until MS diagnosis, death, loss to follow-up, or 31 December 2022. The principal exposure of interest was a multiple sclerosis polygenic risk score (MS-PRS), complemented by several early-life variables, including sex, birth year, season of birth, number of older siblings, maternal smoking around birth, breastfeeding, multiple birth status, and a body mass index polygenic risk score (BMI-PRS). Smoking and infectious mononucleosis were modeled as time-varying exposures to avoid immortal time bias. Methodologically, this is one of the paper’s strongest features: the use of inverse-probability weighting, multiple imputation, Cox proportional hazards modeling, and extensive diagnostics allows the authors to interrogate age-specific risk with considerably greater rigor than is typical in retrospective MS studies.
Genetic Liability Emerges as a Strong and Age-Dependent Determinant
The most striking result is that genetic susceptibility, as measured by MS-PRS, exerts a powerful but distinctly age-dependent effect. Higher polygenic burden was associated with substantially greater hazard of MS diagnosis at younger ages, and this effect attenuated progressively with age. For example, a two-standard-deviation increase in MS-PRS was associated with a hazard ratio of 6.40 at age 20, but only 2.23 at age 60. This finding has major implications for interpretation of prior genetic studies: conventional retrospective analyses may underestimate the role of genetic risk in early adulthood by averaging effects across the full lifespan. The study therefore suggests that polygenic risk is not merely a marker of overall susceptibility, but may also be a determinant of earlier disease timing, an insight that could reshape risk stratification models and motivate survival-based genome-wide analyses in future research.
Sex Differences and the Timing of Disease Susceptibility
The paper also demonstrates that sex is not simply a static covariate but an age-sensitive determinant of MS diagnosis risk. Females had a significantly higher hazard than males, but this disparity was strongest at younger ages and declined over time. The hazard ratio for females versus males decreased from approximately 3.88 at age 20 to 2.15 at age 60. This pattern is consistent with longstanding evidence that female predominance in MS may be linked to hormonal, immunological, and possibly X-chromosomal mechanisms that are especially relevant during reproductive years. Importantly, the authors also report a positive additive interaction between female sex and high MS-PRS, implying that their combined effect exceeds the sum of their separate contributions. In biological terms, this supports the view that sex and genetic background may jointly amplify vulnerability during the period of life when MS most commonly manifests.
Environmental and Early-Life Factors Beyond Genetics
Several non-genetic factors also showed meaningful associations with MS diagnosis hazard. Ever smoking on most or all days was associated with a 69% increase in hazard, while previous infectious mononucleosis doubled the hazard, reinforcing the now substantial body of evidence implicating smoking and Epstein-Barr virus-related pathways in MS pathogenesis. In addition, higher BMI-PRS values were linked to increased hazard, indirectly supporting the role of adiposity-related biology in disease development. The seasonal birth findings were particularly intriguing: compared with birth in autumn, being born in spring, summer, or winter was associated with higher MS hazard, a pattern the authors interpret in light of possible differences in prenatal sunlight exposure and maternal vitamin D status. By contrast, breastfeeding, maternal smoking around birth, multiple birth status, and number of older siblings were not statistically significant in the main model, although some weak trends and interaction patterns suggest that these exposures may still merit further investigation.
From Relative Risk to Predicted Lifetime Incidence
A particularly valuable contribution of the study is its move from statistical association to individualized risk prediction. Using the fitted Cox model, the authors estimated cumulative incidence curves for combinations of MS-PRS, smoking status, and infectious mononucleosis history, separately for females and males. The highest predicted risk was observed among individuals with MS-PRS at the 99th percentile who also had both smoking exposure and prior infectious mononucleosis. In this group, the predicted probability of MS diagnosis by age 80 approached 10% in females and roughly 4% in males. These estimates illustrate how genetic and environmental information can be integrated into clinically interpretable trajectories rather than abstract effect sizes alone. The declining cumulative incidence ratio with advancing age among those at very high polygenic risk further reinforces the central conclusion of the paper: some determinants of MS risk are especially potent early in life and should not be modeled as constant across the lifespan.
Scientific Significance, Limitations, and Future Directions
Overall, this study makes a significant methodological and conceptual contribution to MS research by showing that both genetic burden and sex exert age-dependent effects on the timing of diagnosis. It strengthens evidence for smoking, infectious mononucleosis, and elevated BMI-related liability as relevant contributors, while also proposing that risk prediction could be substantially improved through longitudinal survival modeling. At the same time, the authors appropriately acknowledge several limitations, including the use of diagnosis date as a proxy for disease onset, possible diagnostic misclassification, restriction to White individuals born in England, and the imperfect use of infectious mononucleosis as a surrogate for Epstein-Barr virus exposure. Even with these caveats, the study offers a compelling argument that MS etiology should be studied not only in terms of whether disease occurs, but when it emerges. That perspective may prove essential for future efforts in prevention, early detection, and biologically informed risk modeling.
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:
Nova, A., Di Caprio, G., Bernardinelli, L., & Fazia, T. (2024). Genetic and early life factors influence on time-to-multiple sclerosis diagnosis: A UK Biobank study. Multiple Sclerosis Journal, 30(8), 994-1003.
