Background Recent reports have indicated that there may be an increased risk of mortality with the use of drug-eluting stents (DES) in the high-risk subgroup of patients with diabetes mellitus, as compared with bare-metal stents (BMS). Methods We evaluated 1700 consecutive diabetic patients treated with DES (n=865) and BMS (n=835) between Jan, 1997 and Mar, 2006, in which DES have been used as the default treatment for coronary intervention from Feb, 2003. The clinical outcomes up to 3 years regarding the safety and efficacy were compared after the adjustment for differences in baseline characteristics. Results During 3 years of follow-up, a significant difference of the all-cause mortality was observed in favor of the DES over the BMS (adjusted hazard ratio [HR], 0.56; 95% confidence interval [CI], 0.35-0.90; P=0.02), mainly due to higher non-cardiac mortality of the BMS (HR 0.48; 95% CI, 0.26-0.89; P=0.02). The rates of target-vessel revascularization (TVR) and the composite of death, myocardial infarction, or TVR were markedly reduced in the DES group, as compared with the BMS group (HR 0.48; 95% CI, 0.33-0.69 and HR 0.48; 95% CI, 0.36-0.65, respectively). There was no significant difference in the overall risk of stent thrombosis with DES versus BMS by the Academic Research Consortium (ARC) definitions. However, after 1 year, there was a significant increase of stent thrombosis with DES (any ARC criteria; HR 6.85; 95% CI, 1.56-30.05; P=0.011, and definite or probable; HR 11.5; 95% CI, 0.95-138.9; P=0.055). Conclusions In contrast to results of clinical trials, all-cause mortality was significantly lower in ���real-world��� diabetic patients after DES treatment, as compared with BMS, due to the differences of the non-cardiac death. Despite the substantial reduction of TVR, there was the increased risk of stent thrombosis with DES after 1 year.