BACKGROUND Patients with diabetes experience a more aggressive form of atherosclerosis than patients without diabetes. Whether or not the long-term results of coronary stenting and coronary-artery bypass grafting (CABG) will be same for diabetic and non-diabetic patients remains to be controversial. METHODS We performed a pooled analysis of 5775 patients in 3 observational cohort studies evaluating the relative safety and efficacy of PCI with stenting and CABG for the treatment of multivessel and/or left main disease. We compared adverse outcomes (death; a composite outcome of death, Q-wave myocardial infarction [MI], or stroke; and repeat revascularization) according to diabetic status. RESULTS Over a median follow-up of 5.5 years, risk-adjusted mortality after stenting relative to CABG were similar in patients with diabetes (hazard ratio [HR], 1.114; 95% confidence interval [CI], 0.874-1.420, P=0.383) and without diabetes (HR, 1.112; 95% CI, 0.874-1.417, P=0.387) (P=0.328 for interaction). Also, the adjusted risks for a composite of death, Q-wave MI, or stroke were similar in patients with diabetes (HR, 0.975; 95% CI, 0.787-1.209, P=0.821) and without diabetes (HR, 0.975; 95% CI, 0.787-1.207, P=0.814) (P=0.661 for interaction). However, patients undergoing stenting had significantly higher rates of repeat revascularization (diabetic patients: HR, 3.146; 95% CI, 2.375-4.167, P<0.001, and non-diabetic patients: HR, 3.142; 95% CI, 2.375-4.158, P<0.001) (P=0.061 for interaction). CONCLUSIONS In this pooled analysis of 3 observational studies, PCI with stenting was associated with a long-term mortality and safety profile similar to that of CABG, which was not modified by the diabetic status. However, the rate of repeat revascularization was persistently lower in the CABG patients. Key words: coronary disease, revascularization, stents, surgery.