This study evaluates whether patients with coronary artery disease and severely depressed left ventricular ejection fraction benefit from coronary artery bypass grafting. From 1981 to 1991, 118 consecutive patients with ejection fraction less than or equal to 0.25 underwent isolated coronary artery bypass grafting at Duke University Medical Center. Operative mortality was 11%. Ventricular arrhythmia requiring treatment was the most common postoperative complication (27%), followed by low cardiac output state (22%). Median length of postoperative hospitalization was 9 days. Kaplan-Meier estimate of survival at 1 year and 5 years was 77.2% and 57.5%, and was better than estimated survival with medical therapy alone. Survivors experienced significant improvement in angina class (p < 0.0001), congestive failure class (p < 0.0001), and follow-up ejection fraction (p < 0.005). Of 22 preoperative factors evaluated by univariate survival analysis, five were associated with significantly greater mortality: other vascular disease (p < 0.005), female sex (p < 0.005), hypertension (p < 0.005), elevated left ventricular end-diastolic pressure (p < 0.05), and depressed cardiac index (p < 0.05). Considering length of hospitalization, three factors showed significant adverse effect in a multivariate Cox model: time on cardiopulmonary bypass (p < 0.005), acute presentation (p < 0.005), and female sex (p < 0.01). These data and review of the literature suggest that patients with coronary artery disease and severely depressed ejection fraction benefit from coronary artery bypass grafting, and specific preoperative factors may help determine optimal treatment.