BACKGROUND The standard treatment for patients with symptomatic multivessel coronary artery disease is coronary-artery bypass grafting (CABG). Percutaneous transluminal coronary angioplasty (PTCA) is widely used as an alternative approach to revascularization, but a systematic comparison of the two procedures is needed. We compared the outcomes in patients one year after complete revascularization with CABG or PTCA. METHODS A total of 8981 patients with multivessel coronary disease were screened at eight clinical sites, and 359 patients were randomly assigned to undergo CABG (177 patients) or PTCA (182 patients). Enrollment required that complete revascularization of at least two major vessels supplying different myocardial regions be deemed clinically necessary and technically feasible. RESULTS Among the patients in the CABG group, an average of 2.2 +/- 0.6 vessels were grafted, and among those in the PTCA group, 1.9 +/- 0.5 vessels were dilated. After CABG, hospitalization was longer (median, 19, as compared with 5 days for PTCA), and Q-wave myocardial infarction in relation to the procedure was more frequent (8.1 percent, as compared with 2.3 percent after PTCA; P = 0.022), whereas in-hospital mortality did not differ significantly between the two groups (2.5 percent in the CABG group and 1.1 percent in the PTCA group). At discharge 93 percent of the patients in the CABG group were free of angina, as compared with 82 percent of those in the PTCA group (P = 0.005). During the first year of follow-up, further interventions were necessary in 44 percent of the patients in the PTCA group (repeated PTCA in 23 percent, CABG in 18 percent, and both in 3 percent) but in only 6 percent of the patients in the CABG group (repeated CABG in 1 percent and PTCA in 5 percent; P < 0.001). Seventy-four percent of the patients in the CABG group and 71 percent of those in the PTCA group were free of angina one year after treatment. Exercise capacity improved similarly in both groups. However, 22 percent of the CABG group, as compared with only 12 percent of the PTCA group, did not require antianginal medication (P = 0.041). CONCLUSIONS In selected patients with multivessel coronary disease, PTCA and CABG as initial treatments resulted in equivalent improvement in angina after one year. However, in order to achieve similar clinical outcomes, the patients treated with PTCA were more likely to require further interventions and antianginal drugs, whereas the patients treated with CABG were more likely to sustain an acute myocardial infarction at the time of the procedure.