Strict electrocardiographic, enzymatic, scintigraphic, and hemodynamic criteria for perioperative myocardial infarction (MI) were defined and related to serial assessments of left ventricular performance during rest and exercise in patients seen early and late after coronary artery bypass graft operation. Global left ventricular performance was determined by radionuclide ventriculography from which changes in the pattern of serial postoperative ejection fractions (EF) were obtained. Patients were divided into two groups based on the presence or absence of perioperative MI, and were matched in pairs on the basis of preoperative EF and extent as well as location of coronary artery obstructions. The results indicate that neither short- nor long-term depression in resting EF occurred subsequent to perioperative MI. However, an exercise-related increase in EF eight months postoperatively was depressed in patients who had perioperative MI compared with those who did not. Patients with new Q waves and abnormal postoperative elevation in serum levels of the myocardial isoenzyme of creatine kinase (CK-MB) had a greater early decrease in EF compared with patients without evidence of perioperative MI. However, seven days after operation, the EF in both groups returned to preoperative levels. Patients with abnormal technetium 99m-pyrophosphate scintigrams had changes in perioperative EF similar to those in patients without MI. The presence of low cardiac output syndrome immediately after operation was associated with immediate and short-term decreases in EF, which were not seen in any of the other patient subgroups.