The clinical significance of perioperative myocardial infarction (MI) after coronary artery bypass surgery is not known. Therefore, strategies for the risk stratification of these patients do not exist. This study was undertaken to define the effect of perioperative MI on prognosis after discharge from the hospital and to develop an approach to the risk stratification of these patients. Fifty-nine patients with and 115 patients without perioperative MI were observed for 30 months for the development of cardiac events (death, nonfatal MI, and admission to hospital for unstable angina or congestive heart failure). Patients with perioperative MI were significantly more likely than patients without to have a cardiac event (31% versus 12%, p less than 0.01) and multiple events (19% versus 1%, p less than 0.001). Cox regression analysis identified two independent predictors of cardiac events other than perioperative MI (relative risk, 2.7): inadequate revascularization (relative risk, 3.5) and depressed (less than 40%) postoperative ejection fraction (EF) (relative risk, 2.1). Event-free survival rate of patients with perioperative MI varied markedly depending on the number of other negative prognostic variables present. Patients with perioperative MI who were adequately revascularized and had a postoperative EF greater than 40% had an event-free survival rate similar to patients without a perioperative MI (92% versus 87%, p = NS). Patients with perioperative MI who were inadequately revascularized and had depressed postoperative EF had an event-free survival rate of 13% (p less than 0.001 versus all other subsets). Event-free survival rate was intermediate (68%) in patients with perioperative MI and with only one of the other two variables (p less than 0.001 versus other subsets). In conclusion, perioperative MI adversely affects prognosis. Patients can be stratified into low, high, and intermediate risk subsets based on a simple assessment of the adequacy of revascularization and a determination of residual left ventricular function.