We analyzed the validity of the diagnosis of acute myocardial infarction (MI) in national registries by individual record linkage with the Danish MONICA registry. MI events were identified by record linkage between The National Hospital Registry and The National Death Registry. The sensitivity and specificity were estimated based on three different definitions of an MI event. For events with MI as the primary diagnosis or underlying cause of death the predictive value was 93.6% and sensitivity 77.6% when compared to DANMONICA definite or possible MI. Compared to DANMONICA, definite MI the predictive value was 75.0% and the sensitivity 92.8%. When secondary diagnosis and contributory cause of death were included, the sensitivity increased and the predictive value decreased only slightly. The combination of the national registries was found to be a valid and powerful tool for monitoring the population incidence of MI.