Prognostication in multiple sclerosis (MS) remains challenging. Biomarkers capable of providing this information at diagnosis would be valuable in shaping therapeutic decisions. Measurement of neurofilament light (NfL) has shown promise in predicting clinical outcomes in established MS, but its ability to predict outcomes in real-world cohorts at diagnosis requires further validation.
We used linear regression to evaluate the relationship between serum NfL (sNfL), measured at the time of diagnosis with short-term (1-year) and medium-term (5-year) clinical outcomes in 164 people with MS from a real-world, population-based cohort. Cox proportional hazards regression was used to analyse the association between sNfL and subsequent hazard of relapse or sustained accumulation of disability (SAD). Analyses were adjusted for age and disease-modifying treatment (DMT).
sNfL concentration at diagnosis was modestly associated with baseline EDSS score (β = 0.272, 95% CI 0.051 to 0.494, p = 0.016). However, no significant associations were found between baseline sNfL and odds of relapse at 12-months, 5-year EDSS change, or the hazard of relapse or SAD over 5 years follow-up. Dichotomising baseline sNfL according to the median sNfL did not change these findings.
sNfL appears to be of limited clinical utility in predicting future irreversible neurological disability in a largely untreated real-world population, and remains insufficiently validated to shape treatment decisions at the time of diagnosis. Further studies may be needed for sNfL to be considered as a prognostic marker in the MS clinic. However the masking effect of DMTs on the natural disease trajectory will continue to pose challenges.

Copyright © 2020 Elsevier B.V. All rights reserved.