An analysis of patients undergoing coronary artery bypass for unstable postinfarction angina (less than or equal to 30 days of infarct) during two time periods was undertaken: Group I, January, 1982, through December, 1982; Group II, September, 1983, through August, 1984. Clinical, angiographic, and operative data were coded, and statistical analysis was used to compare the two patient groups, evaluate operative results, and identify risk factors. The incidence of unstable postinfarction angina as an indication for bypass grafting increased significantly (p less than 0.01) from the first to the second time frame, 8.7% (24/276) to 18% (51/283). A greater proportion of Group II patients were operated upon within 7 days of infarct (37% versus 21%, p less than 0.01). All other variables examined were similar in the two patient groups. Analysis of the combined Group I and II patients (N = 75) indicates the following: The ratio of transmural to nontransmural infarction was 39%/61%, and 39% of patients had a previous infarction. Three-vessel disease was present in 76%, two-vessel in 21%, one-vessel in 3%, and left main disease in 20%. Left ventricular ejection fraction was greater than or equal to 40% in 27% of patients, less than 40% in 32%, and not obtained in 41%. Mean left ventricular end-diastolic pressure was 19.5 mm Hg. Intra-aortic balloon pumping was necessary preoperatively in 39%. The mean interval from infarction to revascularization was 12 days, and the mean number of grafts was 3.1 (range one to six). The overall in-hospital mortality was 8% (6/75). Statistical analysis demonstrated that decreased ejection fraction was associated with an increased risk of mortality. No other variables were correlated with mortality. Mean follow-up for the combined Group I and II patients is 13 months (range 4 to 32). Ninety percent of survivors remain in Canadian Heart Association Functional Class I and 6% in Class II. No late deaths have occurred. Patients with unstable postinfarction angina constitute an ever-increasing subset of the coronary bypass population of the 1980s. Operation can be performed with a satisfactory mortality and excellent long-term outlook compared to less acceptable published results with medical management alone. Preoperative left ventricular function constitutes the major indicator of operative risk.