Previous studies have shown impressive short- and medium-term outcomes from drug-eluting stent (DES) implantation in coronary artery disease. We assessed long-term outcomes from DES versus bare metal stent (BMS) implantation in standard and off-label lesions. In 2,345 patients who survived event-free for > or = 30 days after stent implantation for standard (n = 1,540, 66%) or off-label (805, 34%) lesions, we assessed time to occurrence of death, myocardial infarction (MI), death or MI, stent thrombosis, target vessel revascularization (TVR), and major adverse cardiovascular events (defined as composite of all study outcomes). Comparisons were made between standard and off-label lesion subsets and between DES and BMS in lesion subsets. Multivariable differences in outcomes between DES versus BMS were assessed using propensity-adjusted proportional hazards regression. Median follow-up duration was 3.4 years. Stenting of off-label lesions was associated with uniformly worse outcomes than stenting of standard lesions. After adjustment for lesion classification, propensity to receive DES, and baseline differences, DES implantation was associated with statistically significant decreases in death (adjusted hazard ratio 0.71, 95% confidence interval 0.51 to 0.98), TVR (hazard ratio 0.58, 95% confidence interval 0.39 to 0.85 for off-label subset; hazard ratio 0.33, 95% confidence interval 0.24 to 0.46 for standard subset), and major adverse cardiovascular events (hazard ratio 0.51, 95% confidence interval 0.42 to 0.61), without increase in MI, death/MI, or stent thrombosis. Elective TVR occurred in 272 patients and resulted in only 1 early death. In conclusion, compared with BMS, use of DES is associated with clinical benefit in standard and off-label lesions at late follow-up. Decrease in elective TVR does not explain the apparent mortality benefit from DES implantation.