Background: Limited data exists regarding use of drug-eluting stents (DES) outside of approved indications in real-world settings. Recent reports have reported that there may be an increased risk of stent thrombosis up to 1 year with off-label use of DES, as compared with on label use. However, little data are available regarding long-term (3-years) comparison of DES and bare-metal stent (BMS) implantation for acute myocardial infarction. Methods: We evaluated 1.368 patients treated with at least 1 DES (n=437) or BMS (n=931) from January 2003 to March 2006 at Asan Medical Center. We applied a classification of stent thrombosis set by the Academic Research Consortium (ARC). The outcome analysis was based on cardiac death, myocardial infarctions (MI), and stent thrombosis, and adjusted for differences in baseline characteristics. Results: Stent length (34.2짹16.9 mm vs. 21.2짹7.7 mm, p<0.001) was longer in DES group versus BMS group. Mean number of stent per lesion (1.3짹0.6 vs. 1.0짹0.2, p<0.001) and primary stenting (61.6% vs. 35.0%, p<0.001) were higher in DES group versus BMS group. Adjusted risk of death or MI (adjusted relative risk, 1.16; 95% CI, 0.51 to 2.64) was not statistically different with DES versus BMS up to 3 years of follow-up. The risk of MI (adjusted relative risk, 1.63; 95% CI, 0.63 to 4.18) and death (adjusted relative risk, 1.01; 95% CI, 0.26 to 3.89) was not significantly higher in patients treated with DES. Overall stent thrombosis was statistically higher in DES versus BMS up to 3 years of follow-up (adjusted relative risk, 3.34; 95% CI, 1.21 to 9.25). The risk of definite or probable stent thrombosis as defined by the ARC was significantly higher in patents treated with DES (adjusted relative risk, 3.52; 95% CI, 1.09 to 11.38). Conclusions; In the current practice, DES implantation showed higher incidence of stent thrombosis assessed by ARC criteria than BMS implantation for acute myocardial infarction. However, overall incidence of death or MI was similar between two groups.