This study was designed to determine whether left ventricular performance measured noninvasively from the systolic time intervals could identify patients in whom coronary bypass surgery may improve survival. 71 patients with two- or three-vessel disease undergoing coronary bypass surgery were compared with 78 matched medically treated patients. All patients had recuperated from myocardial infarction by a mean of 17.6 months when systolic time intervals were performed. Surgical and medical patients were classified preoperatively into those with normal and those with abnormal left ventricular performance by preejection period/left ventricular ejection time (PEP/LVET ≤ 0.42 and > 0.42, respectively). Survival was analyzed by life table and log-rank test. Cumulative 5-year survival in patients with normal left ventricular performance was not statistically different in surgical and medical groups (96 vs. 93%, respectively). In contrast, cumulative survival in patients with abnormal left ventricular performance was significantly greater in the surgical group when compared to the medical group (84 vs. 62, p < 0.01). Among the patients with abnormal left ventricular function, the mean PEP/LVET and the average vessel disease were not different in the medical and surgical groups. Multivariate analysis of 17 other clinical and laboratory risk variables were not different between these two groups. It is concluded that coronary bypass surgery my improve survival in patients with two- or three-vessel disease and left ventricular dysfunction.