A prospective clinical study was conducted to ascertain if a patient's postoperative elevation in serum creatine kinase MB isoenzyme coupled with determintion of the lactate dehydrogenase1/lactate dehydrogenase2 ratio could differentiate whether atrial or ventricular myocardium was the source of these changes. Animal studies have shown that atrial myocardium is as rich a source of creatine kinase MB as is ventricular myocardium. Atrial myocardium has a lactate dehydrogenase1/lactate dehydrogenase2 ratio less than 1.00, whereas in ventricular myocardium the ratio is greater than 1.00. Sixty-four patients were assigned to six groups on the basis of serial electrocardiograms and vectorcardiograms by a cardiologist who was unaware of their clinical courses. The control group (Group 1) consisted of 16 patients admitted to the coronary care unit who had no electrocardiographic changes. Three surgical groups without electrocardiographic or vectorcardiographic evidence of perioperative myocardial infarction were studied: 10 patients undergoing routine coronary artery bypass procedures (Group 2), six adults undergoing repair of secundum atrial septal defect (Group 3), and 13 patients having mitral valve replacement (Group 4). Two groups of surgical patients who had acute perioperative transmural myocardial infarctions confirmed by serial electrocardiograms and vectorcardiograms were studied: 15 patients (Group 5) who had elective coronary artery bypass procedures and four (Group 6) who had mitral valve replacement. This study suggests that serum creatine kinase MB levels in excess of 50 IU/L on the postoperative day 1 and day 2 samples coupled with serum lactate dehydrogenase1/lactate dehydrogenase2 ratios greater than 1.00 on the postoperative day 2 and day 3 samples support the diagnosis of acute myocardial infarction. Patient groups undergoing procedures necessitating atriotomies had average elevations in serum creatine kinase MB and in the lactate dehydrogenase1/lactate dehydrogenase2 ratio, but these were significantly less than those seen when acute perioperative myocardial infarction had occurred.