Reports of successful use of excimer laser coronary angioplasty for complex coronary artery disease abound, yet firm indications for its use have not been defined. We attempted to treat 858 coronary stenoses in 764 consecutive patients (mean age 61 years; range 32 - 91 years; 75% men; 76% with Class III or IV angina) with excimer laser angioplasty at 308 nm. Successful treatment was achieved in 86% of patients, as indicated by <EQ 50% residual stenosis and no in-hospital complication. To define the potential indications for excimer laser angioplasty, we used relative risk analysis. This showed that certain angiographic features, such as lesions at a vessel bifurcation (odds ratio, OR equals 0.46; 95% confidence interval 0.23, 0.88; P equals 0.017;) or in a tortuous segment (OR equals 0.54; 95% CI equals 0.34, 0.88; P equals 0.041), have decreased likelihood of clinical success. On the other hand, ostial stenoses (OR equals 1.06; 95% CI equals 0.44, 2.56, P equals 0.903) and saphenous vein graft lesions (OR equals 2.17; 95% CI equals 0.98, 4.82; P equals 0.051) have acceptable success rates. Diffuse disease (> 20 mm), total occlusions and calcified lesions were treated as successfully as all other lesion types. Successful treatment with excimer laser coronary angioplasty was also achieved in almost all patients (15/16) who had a prior unsuccessful attempt at balloon angioplasty in the lesion was crossed with a guidewire yet resists either balloon catheter passage or full dilatation. Follow-up angiography was obtained in 70% of eligible patients. Angiographic restenosis, defined by > 50% stenosis, was seen in 60% of patients. Relative risk analysis showed an increased risk of restenosis when adjunctive balloon angioplasty was not used (OR equals 1.68; 95% CI equals 1.02, 2.28; P equals 0.039). Other variables known to affect the outcome of balloon angioplasty, such as lesion length or stenosis in degenerated saphenous vein bypass graft, did not influence the restenosis rates. This analysis defines the profile of risk for excimer laser angioplasty and now provides a sound basis for future, rigorous comparison of excimer laser and balloon angioplasty to balloon angioplasty for a group of lesions that appear to have acceptable success rates with excimer laser coronary angioplasty, which we have denoted the `alpha class' -- saphenous vein lesions, long lesions, ostial stenoses, lesions in calcified vessels, total occlusions, and balloon dilatation failures. Although adjunctive balloon angioplasty may be beneficial, restenosis remains a significant limitation of the procedure and will restrict the usefulness of excimer laser coronary angioplasty for may other types of lesions.