Eighty-three patients underwent coronary artery bypass during acute evolving myocardial infarction 6.8 +/- 2.8 hours after the onset of symptoms. Linear discriminant analysis of preoperative variables identified predictors of mortality with an accuracy of 84%. Significant predictors in decreasing order of importance were cardiogenic shock, age over 65 years, left ventricular ejection fraction less than or equal to 0.30, cardiac index less than or equal to 2.0 L/min/m2, and absent collateral flow. Time to reperfusion did not influence outcome nor did the infarct-related artery. Hospital mortality was 15.6% (13/83). Among 51 low-risk patients under 65 years of age without cardiogenic shock, there were three deaths (5.9%). Follow-up angiography was performed in 21 patients. The graft patency rate was 94%. Left ventricular ejection fraction improved from 0.39 +/- 0.10 to 0.49 +/- 0.11 (p less than 0.05). Left ventricular end-systolic volume decreased from 53.2 +/- 19.3 ml/m2 to 41.4 +/- 16.8 ml/m2 (p less than 0.05), and end-diastolic volume remained unchanged: 86.2 +/- 21.2 ml/m2 before operation and 78.7 +/- 24.0 ml/m2 after operation (no significant difference). Regional ejection fraction of the infarct area, determined by the centerline method, increased 0.23 +/- 0.15. In contrast, among 215 patients treated by nonsurgical reperfusion (intracoronary thrombolysis or angioplasty, or both), mortality was 13.5%. In this group, reperfusion was successful in 144 patients (67%) and 89 underwent follow-up angiography. Persistent patency of the infarct artery was demonstrated in 73 (82%). Ejection fraction increased from 0.45 +/- 0.10 to 0.50 +/- 0.15 (p less than 0.05). We conclude that preoperative variables enable identification of patients with evolving acute myocardial infarction in whom coronary artery bypass is associated with low operative mortality and improved ventricular performance.