Fifty consecutive patients undergoing coronary artery bypass grafting for chronic stable angina were assessed by serial electrocardiography, preoperative and postoperative myocardial scanning with technetium 99m pyrophosphate, gated radionuclide ventriculography, and serial measurement of creatine kinase, aspartate aminotransferase, urea stable lactic dehydrogenase, and creatine kinase isoenzyme (MB) to assess the incidence of perioperative myocardial infarction and identify the most appropriate diagnostic techniques. The correlation between myocardial scanning and the measurement of peak enzyme and isoenzyme activity was excellent in the diagnosis of perioperative infarction, although electrocardiography proved less helpful. There appeared to be no advantage in measuring creatine kinase MB rather than the more routinely measured enzymes. There were two deaths and evidence of myocardial infarction in five other patients, an incidence of 14%. Perioperative infarction was associated with a significant reduction in resting ejection fraction in two cases. In those patients without evidence of perioperative infarction the mean increase in ejection fraction of 7.8% was statistically significant.