Between March 1978 and July 1981, 217 symptomatic patients underwent coronary angioplasty as an alternative to coronary bypass surgery. Angioplasty was successful in 143 patients (66%), unsuccessful but uncomplicated in 65 (30%) and complicated in 9 (4%) by one or more of the following criteria: Q wave myocardial infarction (2%), emergency surgery (4%) or death (0.5%). Late follow-up evaluation was obtained in 213 patients at a mean of 9 +/- 1 years. Of patients in whom angioplasty was successful, 59 (42%) of 140 required another revascularization procedure (repeat angioplasty in 26% and bypass surgery in 16%). The actuarial survival rate at 5, 9 and 10 years after successful angioplasty was 98%, 93% and 92%, respectively. Of the 65 patients with unsuccessful and uncomplicated angioplasty (usually as a result of technical factors), 58 underwent elective bypass surgery within 2 months and 56 survived. These 56 surgical patients were compared with the 140 patients with successful angioplasty. Univariate analysis of prognostic factors did not reveal significant differences between these two groups. At late follow-up study, the successful angioplasty and the successful surgical groups had similar rates of survival (93% versus 95%, p = NS) and of death or infarction, or both (11% versus 12.5%, p = NS). Repeat revascularization was required more frequently after successful angioplasty than after surgery (42% versus 18%, p less than 0.001). Crossover from angioplasty to surgery occurred slightly more often than from surgery to angioplasty (16% versus 12.5%, p = NS). The time to crossover from angioplasty to surgery occurred earlier than from surgery to angioplasty (mean 21 versus 76 months, p less than 0.001).