We examined our practice of invasive therapy for one- and two-vessel coronary disease to assess the impact of the randomized trials of coronary surgery and the current use of angioplasty. We first reviewed our results with coronary artery bypass graft in equivalent patients in the Coronary Artery Surgery Study with one- and two-vessel disease between 1976 and 1981. Among 1376 patients, hospital mortality was 0.07%, and 5-year survival was 95.2% +/- 0.8%. To define trends in invasive therapy, which have since occurred, we compared 100 patients with one- and two-vessel disease in each of three groups: 1979 coronary artery bypass graft, 1984 coronary artery bypass graft, and 1984 percutaneous transluminal coronary angioplasty. Preoperative characteristics in the average 1979 and 1984 patients were similar; however, in 1984, patients who had a coronary artery bypass graft were older than patients who had percutaneous transluminal coronary angioplasty (61.5 versus 56.7 years, p less than 0.01), they required more heart medications (2.1 versus 1.5, p less than 0.01), had more previous infarctions (0.8 versus 0.5, p less than 0.01), and more patients had an ejection fraction of less than 50% (34% versus 7%, p less than 0.01). Patients who had angioplasty had a shorter postoperative stay (median number of days 7, 7, 2, p less than 0.01). Freedom from major complications was similar among the groups (91%, 87%, 85%). Unstable symptoms were the most frequent indication for invasive therapy (approximately 80%), whereas long-term symptoms, those considered in the randomized trials, occurred in relatively few patients. The number of patients without at least one definite indication for invasive therapy was 13%, 3%, and 11%, p less than 0.05, suggesting that the indications for the 1984 coronary artery bypass graft group have become more restrictive since the 1979 coronary artery bypass graft group. Indications for the 1984 percutaneous transluminal coronary angioplasty group remained less restrictive, being similar to those for the 1979 coronary artery bypass graft group. A continuing trend toward the use of percutaneous transluminal coronary angioplasty was evident, as 56% of the 1979 coronary artery bypass graft group of patients and 32% of the 1984 coronary bypass group of patients would be offered percutaneous transluminal coronary angioplasty rather than coronary artery bypass graft on the basis of 1986 percutaneous transluminal coronary angioplasty criteria. The p values were obtained with analysis of variance or chi 2 test.