Late survival and freedom from myocardial infarction were determined for 192 patients with coronary artery disease and depressed left ventricular ejection fraction at rest (less than or equal to 35%) determined by biplane angiography who were evaluated between 1970 and 1977. Seventy-seven patients had coronary artery bypass grafting and 115 patients were treated medically and were considered surgical candidates. The medical and surgical groups were comparable in all baseline characteristics examined except frequency of three vessel disease and angina pectoris, which occurred in a significantly greater percent of the surgically treated patients (p less than 0.01). Only three medically treated patients (2.6%) underwent coronary bypass grafting in the follow-up period. Seven year actuarial survival was 63% in the surgical and 34% in the medical group (p less than 0.001). Ninety-three percent of patients in the surgical group and 81% of those in the medical group were free of nonfatal myocardial infarction (p = 0.01), and 62 and 33%, respectively, were alive and free of myocardial infarction (p less than 0.001) at 7 years. Significant differences in survival favoring surgical treatment were observed for the subsets of patients with an ejection fraction of 25% or less (p = 0.0002) and 26 to 35% (p = 0.01), and for the subsets with three vessel coronary disease (p less than 0.001), normal left ventricular end-diastolic volume (less than or equal to 100 ml/m2) (p = 0.005) and elevated end-diastolic volume (greater than 100 ml/m2)(p = 0.001). After adjustment for other important prognostic variables, the type of treatment remained significant in predicting the relative risk (medical to surgical) of mortality at 5 and 7 years (2.58 and 2.12, respectively). These data corroborate the trends observed in several randomized trials of medical and surgical therapy in patients with abnormal left ventricular function. If hospital mortality for coronary artery bypass grafting is less than 5%, substantial benefit can be anticipated for the majority of patients with depressed ventricular function.