We compared 1-year outcomes in patients treated with paclitaxel-eluting stents (PESs) or sirolimus-eluting stents (SESs) in "real-world" clinical practice. Clinical outcomes were evaluated in 1,558 consecutive, unselected, retrospectively collected patients treated with drug-eluting stents (DESs; PES = 816, SES = 742) at 19 United States centers. The primary end point was 1-year target vessel revascularization (TVR). The study included a prespecified diabetic cohort (PES = 289, SES = 247), for which efficacy comparisons between DESs were analyzed according to vessel diameter and presence of chronic kidney disease. Baseline demographic, angiographic, and procedural characteristics were similar between patients treated with PESs and those treated with SESs. At 1 year, there were no overall statistical differences in death, myocardial infarction, TVR, or stent thrombosis. In the diabetic cohort, however, the cumulative incidence of TVR was significantly lower for patients treated with PESs (3%) compared with SESs (9%, p <0.01), which persisted after adjustment for baseline differences (hazard ratio 0.29, 95% confidence interval 0.12 to 0.67). This decrease in TVR with PES was similar in insulin- and noninsulin-requiring diabetic patients. In multivariate analysis, independent predictors of TVR included diabetes, bifurcation stenting, and overlapping stents; in the diabetic cohort, treatment with SESs was also a multivariate predictor of TVR. In conclusion, in this observational, retrospective analysis of DES-treated patients, PESs and SESs demonstrated similar overall safety and efficacy, but PESs were associated with a significant decrease in 1-year TVR rates in diabetic patients.