The role of excimer laser angioplasty in treating complex coronary artery disease remains uncertain. A randomized trial comparing this new technology with balloon angioplasty cannot be designed until systematic analysis identifies the lesion types that are likely to benefit from treatment with excimer laser angioplasty. In a cohort of 764 patients who had 858 coronary stenoses treated with excimer laser-facilitated angioplasty, relative risk analysis was used to examine acute success, complications and restenosis rates, and the results were compared with those of balloon angioplasty to identify the lesion types that show the greatest benefit with the new treatment. Clinical success was achieved in 657 patients (86%), as indicated by < or = 50% residual stenosis and no in-hospital complication. A major in-hospital complication (death, bypass surgery, or Q-wave or non-Q-wave myocardial infarction) occurred in 58 patients (7.6%). Follow-up angiography was obtained in 70% of eligible patients. Combining angiographic and noninvasive restenosis rates yielded an overall restenosis rate of 46%. Relative risk analysis showed that major complications occurred frequently in lesions at an arterial bifurcation (odds ratio [OR] 5.96 [2.76, 12.6]; p = 0.001). However, certain complex lesions that are difficult to treat with balloon angioplasty (saphenous vein graft lesions, lesions > 10 mm, ostial lesions, calcified stenoses, total occlusions and unsuccessful balloon dilatations), analyzed together as a group, had lower complication rates by univariate (OR 0.59 [0.35, 1.00]; p = 0.051) and multivariate logistic regression (p = 0.006) analyses.(ABSTRACT TRUNCATED AT 250 WORDS)