Background Despite serious concerns about late thrombosis associated with drug-eluting stents (DES), data are limited regarding uniform evaluation of stent thrombosis in real-word practice. Methods Our study included 7221 patients (3160 DES and 4061 BMS patients) who had received coronary stents at two academic hospitals from 1998 to 2006. Since February, 2003, DES has been adopted as the default treatment. The primary end point was the cumulative incidence of stent thrombosis set by the Academic Research Consortium (ARC) definitions. Results After the propensity-score adjusting analyses, there was no significant difference of the overall cumulative incidence of stent thrombosis between DES and BMS groups over 3 years: definite (adjusted hazard ratio [HR], 1.20; 95% confidence interval [CI], 0.62-2.29), definite or probable (HR, 1.07; 95% CI, 0.60-1.91), and any ARC criteria (HR, 1.29; 95% CI, 0.84-1.98). However, after 1 year, patients with DES showed a significant higher rates of very late thrombosis: definite (HR, 4.16; 95% CI, 1.24-13.95), definite or probable (HR, 3.42; 95% CI, 1.20-9.74), and any ARC criteria (HR, 2.24; 95% CI, 1.14-4.37). The overall risk of cardiac death or myocardial infarction (MI) was similar between two groups (HR, 1.31; 95% CI, 0.90-1.90). Conclusions Stent thrombosis after 1 year was more common with DES treatment than with BMS in consecutive ���real-word��� patients. There were no significant differences in the cumulative rates of cardiac death or MI up to 3 years.