Background: During the past decade, an increasing number of prospective studies have focused on the association between vitamin D and cardiovascular disease (CVD). However, the evidence on the relation between serum 25-hydroxyvitamin D [25(OH)D] and the risk of overt CVD is inconclusive.

Objective: We performed a dose-response meta-analysis to summarize and prospectively quantify the RR of low serum 25(OH)D concentration and total CVD (events and mortality).Design: We identified relevant studies by searching PubMed and EMBASE up to December 2015 and by hand-searching reference lists. Prospective studies based on the general population and reported RRs and 95% CIs were included. A random-effects model was used to calculate the pooled RRs. Nonlinear association was assessed by using restricted cubic spline analyses.

Results: A total of 34 publications with 180,667 participants were eligible for the meta-analysis. We included 32 publications (27 independent studies) for total CVD events and 17 publications (17 independent studies) for CVD mortality. We observed an inverse association between serum 25(OH)D and total CVD events and CVD mortality, and the pooled RRs per 10-ng/mL increment were 0.90 (95% CI: 0.86, 0.94) for total CVD events and 0.88 (95% CI: 0.80, 0.96) for CVD mortality. A nonlinear association was detected for total CVD events (P-nonlinear < 0.001) and CVD mortality (P-nonlinear = 0.022).

Conclusion: Serum 25(OH)D concentration was inversely associated with total CVD events and CVD mortality from the observed studies.