Background In a randomized study of unstable angina, medically treated patients with impaired left ventricular (LV) ejection fractions (EF=0.3 to 0.58) were at significantly higher risk of mortality than patients treated by coronary artery bypass graft surgery (CABG). Because the duration of this surgical advantage is unknown, 10-year cumulative mortality rates of patients with impaired LVEF were determined and compared with the previously observed rates at 2, 5, and 8 years.Methods and Results Of 468 patients with unstable angina, 237 were randomized to receive medical treatment alone and 231 patients to have CABG.Baseline characteristics, which were equally distributed between the two treatment groups, included age, LVEF, number of diseased coronary arteries, diabetes, clinical presentation (type I or type II), prior myocardial infarction, and smoking. Mortality was determined by life-Table analysisand risk factors by logistic regression analysis. Patients were divided into terciles according to LVEF, and the mortality rates of medical and surgical patients in the lowest tercile were compared. The 10-year mortality rate for all medical patients was 38% and for all surgical patients, 39%. When LVEF was treated as a continuous variable, there was a significant relation between mortality and LVEF for medically treated patients but not for surgical patients. The cumulative mortality rate for the lowest-tercile (EF 0.3 to 0.58) medical patients was 49%; for the lowest-tercile surgical patients, 41% (P=.15).Conclusions The surgical advantage for patients with impaired LVEF that was significant at 5 years (P=.03) and 8 years (P=.05) appears to have diminished at 10 years (P=.15). (Circulation. 1994;90(part 2):II-120-II-123.)