PURPOSE: Patients with epilepsy have increased risk of myocardial infarction (MI). Valproate can exert anti-atherosclerotic effects. We therefore examined the risk of MI in patients with epilepsy receiving valproate. METHODS: Two cohorts of patients with valproate-treated epilepsy and sex- and age-matched individuals (controls) from the general Danish population were identified by individual-level-linkage of nationwide registries and followed for 10 years. The two cohorts comprised patients treated with valproate at baseline and valproate-naïve patients initiating treatment in the study period, respectively. The hazard ratios (HR) of MI and all-cause death were estimated by two different Cox proportional-hazard models; valproate treatment was analysed as a baseline categorical covariate in the first cohort and as a time-dependent exposure covariate in the second cohort. RESULTS: The two cohorts comprised 53¿086 and 102¿003 individuals, respectively. In the first cohort, the risk of MI was decreased (HR 0.75, 95% confidence interval 0.59-0.97) while the risk of all-cause death was increased (HR 2.11, 95% confidence interval 1.95-2.28), compared to the controls. In the second cohort, the risk of MI was decreased (HR 0.62, 95% confidence interval 0.53-0.73) while the risk of all-cause death was similar to the controls (HR 1.02, 95% confidence interval 0.97-1.07). CONCLUSIONS: In this nationwide pharmacoepidemiological study, we found a consistent association between valproate treatment and a reduced risk of MI in patients with epilepsy. Copyright © 2010 John Wiley & Sons, Ltd.