Background: High vitamin K intake is associated with a reduced risk of coronary heart disease (CHD). This is thought to be mediated by increased activation of the vitamin K-dependent matrix γ-carboxyglutamate protein (MGP). Desphospho-uncarboxylated MGP (dp-ucMGP) is associated with both vitamin K status and vascular calcification. However, the association of dp-ucMGP with CHD and stroke in the general population has not been investigated to date. Objective: To investigate the association of dp-ucMGP with incident CHD or stroke. Methods: A prospective case-cohort study with a representative baseline sample of 1406 participants and 1154 and 380 incident cases of CHD and stroke, respectively, was nested within the EPIC-NL study. Circulating dp-ucMGP levels were measured with ELISA in baseline plasma samples. The incidence rates of fatal and non-fatal CHD and stroke were obtained by linkage to national registers. Cox proportional hazard models were used to calculate hazard ratios (HRs) per standard deviation (SD) and per quartile of circulating dp-ucMGP levels. Results and conclusion: The average follow-up was 11.5 years. Levels of dp-ucMGP were not associated with CHD risk, with an HR per SD of 1.00 (95% confidence interval [CI] 0.93-1.07) and an HRQ4 vs. Q1 of 0.94 (95% CI 0.79-1.13) after adjustment for cardiovascular risk factors. There was no association of dp-ucMGP stroke risk (HRSD 0.98, 95% CI 0.90-1.08; and HRQ4 vs. Q1 1.09, 95% CI 0.78-1.51). This study could not confirm that high dp-ucMGP levels, reflecting poor vitamin K status, are associated with increased CHD or stroke risk in the general population.