The usefulness of measuring serum MB creatine kinase activity (CK-MB) for the diagnosis of per- and postoperative myocardial infarction (MI) was assessed in 104 patients undergoing coronary artery bypass grafts. In each patient, 15 samples were taken during the week which followed the surgical procedure. New Q waves were considered to be a criteria of MI. 19 patients developed new Q waves (MI group), whereas 57 had no significant ECG changes (control group); 13 showed only ST changes, whilst 15 had unassessable recordings. In the MI group, CK-MB was greater than in the control group, both at the first peak (8 to 10 h after induction of anaesthesia) and at the greater peak (13 to 21 h after induction) (p less than 0.05). Significant differences were also seen between both groups between 8 and 32 h after induction, but there was also a large overlap. An area under the curve (AUC) greater than 50,000 IU.l-1.min-1 had a positive predictive value of 0.64, and an AUC less than 50,000 IU.l-1.min-1 a negative predictive value of 0.89 if all the groups of patients were taken into account. An AUC greater than 65,000 IU.l-1.min-1 was always seen in MI patients, but only 25% of MI patients had a value greater than this threshold. There were no significant differences between the patient groups in the first peak time, nor in the CK-MB/total CK ratio. CK-MB appeared therefore as a less reliable criterium of per- and postoperative MI during coronary artery bypass operations than previously reported, especially when intermediate values are found.