ObjectiveThe role of osteocalcin (OC) in cardiovascular disease (CVD) is unresolved. We aimed to study the association between plasma OC concentrations and the risk of non-fatal and fatal CVDs. We also aimed to investigate whether such an association, if present, would be mediated by established metabolic risk factors.DesignA population-based longitudinal cohort study.MethodsIn 1995/1996, OC was determined in blood samples drawn from 1319 subjects aged 65–88 years participating in the Longitudinal Aging Study Amsterdam in 1995/1996. The self-reported CVD events were collected every 3 years until 2005/2006, and CVD deaths until 1st January 2007. Cox proportional hazards regression was performed, considering potential confounders (smoking, physical activity, and BMI) and mediators (blood pressure, plasma triglycerides, total and HDL cholesterol, fructosamine, and aortic calcification).ResultsDuring the median 4.1 years follow-up, 709 subjects (53.8%) suffered a CVD event. There was no overall association between OC and CVD: hazard ratio (HR) was 0.97 (95% CI 0.90–1.04) per nmol/l higher plasma OC, adjusted for age and sex. There was a statistical interaction between plasma OC, age, and sex on CVD (P=0.014). In those subjects aged ≥75 years, age-adjusted HRs (95% CI) were 0.86 (0.75–0.99) in men and 1.16 (1.03–1.31) in women per nmol/l higher plasma OC. Adjustment for covariates only slightly attenuated the association in older-old men, but did not affect the association in older-old women.ConclusionA higher plasma OC concentration was associated with a reduced risk of CVD in older-old men and with an increased risk of CVD in older-old women. We found no evidence that this was mediated by arterial calcification or metabolic risk factors.