BACKGROUND Recent longitudinal analyses suggested that low levels of serum 25-hydroxyvitamin D (25-OH-D) predict incident cardiovascular disease in initially healthy populations. Because the prognostic value of vitamin D for the occurrence of secondary cardiovascular events remains unclear, we examined the association of baseline 25-OH-D levels with prognosis in patients with stable coronary heart disease (CHD). METHODS Serum 25-OH-D levels from 1,125 CHD patients of 2 German clinics undergoing a 3-week rehabilitation program after an acute cardiovascular event were measured, and participants were followed for up to 8 years. We used multivariate Cox regression analysis to model cardiovascular event incidence (fatal and nonfatal, including myocardial infarction, stroke, and death due to cardiovascular diseases) and all-cause mortality according to 25-OH-D quartiles, categories based on cut points of 15 and 30 ng/mL, or continuous vitamin D concentrations. RESULTS During follow-up, 148 cardiovascular events and 121 deaths were recorded. Elevation of risk for the lowest quartile or category in comparison to the highest category was weak and nonsignificant for both incidence (hazard ratio [HR](quartile1) = 1.15 [0.72-1.84], HR(<15 ng/mL) = 1.17 [0.61-2.23]) and mortality (HR(quartile1) = 1.29 [0.77-2.14], HR(<15 ng/mL) = 1.87 [0.91-3.82]) in unadjusted Cox regression analysis and disappeared entirely after adjustment for potential confounders (cardiovascular events: HR(quartile1) = 0.84 [0.47-1.50], HR(<15 ng/mL) = 0.90 [0.41-1.96]; mortality: HR(quartile1) = 0.63 [0.33-1.21], HR(<15 ng/mL) = 0.93 [0.39-2.21]). Models treating vitamin D as a continuous variable likewise suggested no significant associations. CONCLUSIONS Unlike previous population-based studies, our analysis in high-risk patients with stable CHD does not support a prognostic value of baseline-25-OH-D levels for secondary cardiovascular event incidence or all-cause mortality.