One hundred consecutive patients undergoing coronary artery bypass surgery were randomly allocated to a preoperative (24 h) intravenous infusion of isosorbide dinitrate (1.5-15 mg/hr) (50 patients) or to placebo (50 patients). The characteristics of the two groups were similar. Evidence of acute myocardial ischaemia was sought by continuous electrocardiographic Holter recordings and acute myocardial infarction by the appearance of new Q waves and increased activity of the creatine kinase MB isoenzyme. Episodes of acute myocardial ischaemia were found in 18% of patients in the control group and in none of those who received isosorbide dinitrate. None the less, the frequency of perioperative myocardial infarction was similar (22% and 18% respectively) in the two groups. Perioperative infarction was significantly more common in women, in patients with unstable angina or poor left ventricular function, in those who had coronary endarterectomy, and in those in whom the aortic clamping time was greater than 50 minutes. These factors may have obscured any effect that prevention of preoperative ischaemia had on perioperative infarction. Preoperative infusion of isosorbide dinitrate eliminated preoperative ischaemia but did not influence the occurrence of perioperative infarction. The probable benefits of prevention of preoperative ischaemia on postoperative left ventricular function, which is a determinant of long term survival, remain to be established.