We conducted a multicenter, randomized, prospective study comparing medical therapy alone with coronary-artery bypass surgery plus medical therapy in 468 men with unstable angina pectoris. Patients were entered in the study from June 1, 1976, to June 30, 1982. Among those assigned to surgery who received bypass grafts, operative mortality was 4.1 percent. Arteriography performed after one year of follow-up revealed that 74.8 percent of the grafts studied were patent. The cumulative rate of crossover from medical to surgical therapy after two years was 34 percent; the operative mortality among patients crossed over was 10.3 percent. Nonfatal myocardial infarction occurred in 11.7 percent of the patients treated surgically and 12.2 percent of those treated medically (no significant difference). Most of the nonfatal myocardial infarctions in the surgical group occurred in the perioperative period. Overall, the two-year survival rate computed by life-table analysis did not differ between the two groups. However, the curves reflecting mortality as a function of left ventricular ejection fraction were significantly different (P = 0.03); surgery was associated with a significantly reduced mortality among patients with lower ejection fractions. We conclude that patients with unstable angina pectoris have a similar outcome after two years whether they receive medical therapy alone or coronary bypass surgery plus medical therapy. However, patients with reduced left ventricular ejection fractions may have a better two-year survival rate after coronary bypass surgery.