Serum 25-hydroxyvitamin D concentration is associated with device-estimated sleep metrics in healthy young and early middle-aged adults.

Study Objectives: We tested associations between serum 25-hydroxyvitamin D concentration ([25(OH)D]) and device-estimated sleep metrics, including sleep duration, sleep efficiency, sleep duration regularity, sleep timing regularity, and sleep regularity index (SRI), in young and early middle-aged adults (18-45 years). We also assessed the mediating effect of nighttime melatonin (urinary 6-sulfatoxymelatonin (aMT6s) excretion) on these associations. Methods: Participants (n = 79) completed 14 days of wrist actigraphy. Fasted blood sampling was performed to quantify serum [25(OH)D]. First morning void was used to quantify overnight urinary aMT6s excretion, normalized to creatinine clearance. Associations between [25(OH)D] and sleep metrics were evaluated using linear regression (model 1). Separate models adjusted for age, sex, race, and body fat % (model 2), season of testing, caffeine consumption, and education level (model 3), and device-estimated moderate-to-vigorous physical activity (model 4; n = 68). Results: Serum [25(OH)D] was positively associated with sleep duration, sleep efficiency, and SRI, and negatively associated with sleep duration regularity, sleep onset timing regularity, and sleep midpoint timing regularity in model 1 (all p < .03) and model 4 (all p < .02). In model 2, serum [25(OH)D] remained significantly associated with sleep duration only (p = .036). In model 3, serum [25(OH)D] remained significantly associated with all sleep metrics (p < .02) except sleep duration regularity and SRI. Serum [25(OH)D] was not associated with aMT6s:creatinine, indicating no grounds for performing mediation analyses. Conclusions: Serum [25(OH)D] is independently associated with several sleep metrics in healthy adults. However, nighttime melatonin concentration did not mediate these associations, thus other mechanistic pathways must be considered.