Background: Low baseline plasma 25-hydroxyvitamin D (25(OH)D is associated with increased risk of acute respiratory infections, but its association with long-term risk of sepsis remains unclear. Methods: Case-cohort analysis of participants selected from the Reasons for Geographic and Racial Differences in Stroke study (REGARDS), a US cohort of 30,239 adults ≥ 45 years old. We measured baseline plasma 25(OH)D in 711 sepsis cases and in 992 participants randomly selected from the REGARDS cohort. We captured sepsis events by screening records with International Classification of Disease methods then adjudicating clinical charts for significant, suspected infection and severe inflammatory response syndrome criteria on presentation. Results: In the study sample, the median age of participants was 65.0 years, 41% self-identified as black, and 45% were male. Mean plasma 25-hydroxyvitamin D concentration was 25.8 ng/ml with 31% of subjects < 20 ng/ml. The adjusted risk of community-acquired sepsis was higher for each lower category of baseline 25(OH)D. Specifically, in a Cox proportional hazards model adjusting for multiple potential confounders, when compared to a baseline 25(OH)D >33.6 ng/ml, lower 25(OH)D groups were associated with higher hazards of sepsis (16.5-22.4 ng/ml hazard ratio 3.21, 95% confidence interval [CI] 1.98 to 5.21; < 16.5 ng/ml hazard ratio 6.81, 95%CI 3.95 to 11.73). Results did not materially differ in analyses stratified by race or age. Conclusions: In the REGARDS cohort of community-dwelling US adults, low plasma 25(OH)D measured at a time of relative health was independently associated with increased risk of sepsis.