OBJECTIVE First-generation drug-eluting stents have been proven to be very effective for the treatment of bare-metal stent in-stent restenosis (BMS ISR). Efficacy of second-generation drug-eluting stents in this setting remains less well defined. This study compared 3-year clinical outcomes after treatment of BMS ISR using second-generation everolimus-eluting stents (EES) to first-generation paclitaxel-eluting stents (PES) or sirolimus-eluting stents (SES). METHODS This was a retrospective observational study. A total of 264 consecutive patients with BMS ISR underwent percutaneous coronary intervention using EES (75 patients), PES (95 patients), or SES (94 patients) from 2003 to 2009. The primary endpoint of the study was survival free of major adverse cardiac events (MACE) at 3 years. Secondary endpoints were survival free of need for revascularization of the target lesion and definite stent thrombosis. Clinical follow-up could be obtained in 99% of patients. RESULTS Baseline clinical and angiographic parameters were comparable between the three groups. MACE at the 3-year follow-up were 27, 30, and 27%, for the EES, PES, and SES groups, respectively (P=0.874). The target lesion revascularization rates for EES, PES, and SES groups were 15, 20, and 23%, respectively (P=0.429). Rates of definite stent thrombosis at the 3-year follow-up were comparable between the three groups at 0, 2.1, and 1.0%, respectively (P=0.437). Rates of myocardial infarction and death were also similar between the three groups. Diabetes mellitus was the only independent predictor of MACE at the 3-year follow-up (odds ratio=1.14, 95% confidence interval 1.00-1.30; P=0.038), whereas renal insufficiency was the only independent predictor for death (odds ratio=1.10, 95% confidence interval 0.850-1.274; P=0.011). CONCLUSIONS Second-generation EES is as effective and safe as the first-generation PES or SES in the treatment of BMS ISR. Diabetes mellitus is the only independent predictor for MACE at the long-term follow-up.