Early and late results of primary nonemergency coronary artery bypass grafting in 1,000 consecutive patients and primary nonemergency percutaneous transluminal coronary angioplasty performed concurrently in 389 patients were retrospectively compared. The coronary bypass population was significantly older and more symptomatic and had more prior myocardial infarctions, more left main and multiple-vessel coronary artery disease, and poorer ventricular function. Hospital mortality rates for coronary bypass grafting and angioplasty were 0.4% and 0.5%, respectively, and infarction rates were 1.7% and 5.1%, respectively (p less than 0.01). Including hospital events for the coronary bypass and angioplasty populations, actuarial survival at 5 years was 92.3% versus 96.3% (p = 0.04), freedom from myocardial infarction was 94.6% versus 88.1% (p less than 0.001), freedom from subsequent angioplasty was 99.5% versus 75.2% (p less than 0.001), freedom from subsequent coronary bypass grafting was 98.8% versus 84.9% (p less than 0.001), and freedom from all morbidity and mortality was 87.1% versus 66.0% (p less than 0.001), respectively. By Cox regression analysis for all 1,389 patients, only diminished ejection fraction and advanced age predicted poor long-term survival (p less than 0.001). The only significant predictor of nonfatal late events was having had coronary angioplasty.