AIMS: To identify possible underuse by analysing initiation and persistence with clopidogrel treatment in an unselected population of patients admitted with myocardial infarction (MI) with or without subsequent percutaneous coronary intervention (PCI). METHODS: Patients admitted with first-time MI from 2000 to 2005 and subsequent prescription claims of clopidogrel were identified by individual-level linkage of nationwide administrative registries in Denmark. Independent factors affecting initiation and persistence with treatment were analysed by multivariable logistic regression models and Cox proportional hazard models. RESULTS: A total of 46,190 MI patients were included in the study, of whom 14,939 were treated with PCI. From 2000 to 2005 initiation of clopidogrel increased from 80.4 to 93.7% among MI patients with PCI and from 2.8 to 39.3% among MI patients without PCI. MI patients with concomitant heart failure received less treatment [odds ratio (OR) 0.49, confidence interval (CI) 0.43, 0.56 among patients with PCI and OR 0.90, CI 0.81, 0.99 among patients without PCI in 2002-2003, and OR 0.89, CI 0.80, 1.00 in 2004-2005, respectively]. Of MI patients with PCI, 77.5% completed 9 months' clopidogrel treatment in 2004-2005, the corresponding figures for MI patients without PCI being 53.9%. CONCLUSIONS: Initiation and persistence with clopidogrel treatment is high in MI patients with PCI. However, we found substantial underuse among MI patients without PCI and in MI patients with heart failure.