We evaluated the prevalence, determinants, and misclassification of different types of myocardial infarction in 3,272 men and women age 55 years or older. We defined self-reported myocardial infarction with electro-cardiographic evidence as "typical myocardial infarction." We defined self-reported myocardial infarction without electrocardiographic evidence, but verified with additional clinical information, as "non-Q-wave myocardial infarction." Finally, we defined myocardial infarction detected by electrocardiogram that was not self-reported as "silent myocardial infarction," after verification of absence of symptoms. Overall, the prevalence of typical myocardial infarction was 4.1% [95% confidence interval (CI) = 3.5-4.9], of non-Q-wave myocardial infarction 2.8% (95% CI = 2.2-3.4), and of silent myocardial infarction 3.9% (95% CI = 3.2-4.5). Silent myocardial infarction was more prevalent in women, hypertensives, cigarette smokers, and those with higher post-load blood glucose. Self-reported myocardial infarction without electrocardiographic characteristics could be verified as myocardial infarction by means of additional clinical information in 56% of the cases. We conclude that myocardial infarction occurs frequently in the elderly without typical symptoms or electrocardiographic changes. As all these manifestations of myocardial infarction convey an increased risk of symptomatic heart disease or death, they require further attention. Misclassification due to limited sources of information can be considerable and should be taken into account in the design and interpretation of epidemiologic studies.