To examine the association between 25-hydroxyvitamin D (25[OH]D) levels and the risk of multiple sclerosis (MS) in blood samples collected prospectively and during gestation.

In this nested case-control study, 2 population-based biobanks with 291,500 samples from 164,000 persons collected since 1975 in the northern half of Sweden were used. We identified prospectively collected blood samples from MS cases (n = 192, controls matched 2:1) and gestational samples from pregnant mothers where the offspring had later developed MS (n = 37, control mothers matched 5:1). 25(OH)D levels were measured using an ELISA, and the risk of MS was analyzed using matched logistic regression.

Levels of 25(OH)D ≥75 (vs <75) nmol/L in prospectively collected blood samples were associated with a decreased risk of MS (odds ratio [OR] 0.39, 95% confidence interval [CI] 0.16-0.98). No decrease in MS risk was found in the offspring exposed to gestational 25(OH)D levels ≥75 (vs <75) nmol/L (OR 1.8, 95% CI 0.53-5.8). The prevalence of 25(OH)D levels ≥75 nmol/L in female controls decreased gradually during 1976-2005 (p trend = 0.005).

This study supports the presence of an association between high 25(OH)D levels during the years preceding disease onset and a decreased risk of MS. In the very limited material with samples drawn in early pregnancy, where month-of-birth effects were controlled for, we found no association between gestational 25(OH)D levels and MS risk in the offspring. Decreasing 25(OH)D levels in the population may contribute to explain the increasing MS incidence that is suggested from epidemiologic studies.