Percutaneous coronary intervention (PCI) of the unprotected left main coronary artery (LMCA) is controversial. In 143 patients who underwent PCI of the unprotected LMCA, 30-day mortality was compared with predicted cumulative risk-adjusted perioperative surgical mortality based on logistic European System for Cardiac Operative Risk Evaluation. One-year clinical follow-up was completed in all patients. The overall major adverse cardiac event rate at 1 year was 34.3%, reflecting the high-risk profile of the patient population. Twelve patients (8%) experienced an acute myocardial infarction and 16 (11%) underwent target lesion revascularization. In 31 patients (22%) who died during the first year, median logistic European System for Cardiac Operative Risk Evaluation was 30%. Calculated RRs showed significantly lower 30-day mortality using PCI compared with predicted surgical mortality (RR 0.54, 95% confidence interval 0.31 to 0.86). Angiographic follow-up in 90 of the 118 patients alive at 6 months showed binary restenosis of 6% in patients treated with drug-eluting stents versus 29% in patients receiving bare-metal stents (p < or =0.01). In conclusion, PCI for unprotected LMCA disease was associated with acceptable short- and medium-term outcomes in patients at low to intermediate risk of bypass surgery. Mortality remains high in very high-risk patients unsuitable for surgery. However, in selected indications, PCI of the LMCA can offer an alternative to surgery, especially when using drug-eluting stents.