AIMS: Drug-eluting stents (DES) have never been sufficiently studied in patients with atrial fibrillation (AF). The latter are considered as a high-risk population with uncertainty over the optimal antithrombotic therapy strategy to prevent stroke, stent thrombosis, and recurrent cardiac ischaemia, balanced against the high risk of haemorrhage. The aim of this study was to evaluate the safety and efficacy of the use of DES vs. bare-metal stents (BMS) in a cohort of patients with AF. METHODS AND RESULTS: We reviewed 604 patients with AF who had undergone percutaneous coronary intervention with stent over a period of 7 years (January 2001-January 2008). After a propensity score selection, we identified two matched cohorts who received DES (n = 207) or BMS (n = 207). Clinical follow-up was performed, and all bleeding episodes, thrombo-embolism, and major adverse cardiac events (MACE; i.e. death, acute myocardial infarction, target vessel failure) were recorded. Complete follow-up was achieved in 95.9% of the cohort (mean: 693 +/- 427 days, median: 564). The incidence density of MACE as well as the incidence of all-cause mortality in both groups was similar. There was a higher incidence of major bleeding in DES group (2.26 vs. 1.19 per 10 000 days of exposure; P = 0.03). In a multivariate analysis, age, chronic AF, chronic renal failure, and non-use of dicoumarin were predictors of MACE and of all-cause mortality. The use of DES was not a predictor of reduced events. CONCLUSION: On the basis of this study, the routine use of DES in patients with AF does not seem to be justified. A higher risk of major bleeding with DES in comparison with BMS raises the possibility that DES should be limited to lesions or patients with a high risk of restenosis.