We prospectively compared, according to their preoperative clinical profiles, the in-hospital outcomes of patients operated on consecutively (but without randomization) for isolated coronary artery disease with on-pump or off-pump techniques. During 2001, 324 patients underwent coronary artery bypass grafting: 216 patients (mean age, 66.7 +/- 8.9 years; range, 41-85 years) underwent on-pump revascularization, and 108 patients (mean age, 676 +/- 10 years; range, 37-90 years) underwent full-sternotomy off-pump revascularization. The 2 groups were homogeneous with regard to female sex (22.6% vs 26.8%), previous cardiac operation (2.8% vs 4.6%), cardiogenic shock (1.3% vs 1.9%), diabetes (30% vs 33%), and chronic renal failure that required hemodialysis (3% vs 3.5%). Postoperative complications, including bleeding, myocardial infarction, acute renal failure, mediastinitis with sternal dehiscence, cerebrovascular events, and prolonged respiratory assistance were more frequent in on-pump patients (P = 0.004). The total number of grafts and the grafts per patient ratio were significantly higher in on-pump patients (P = 0.0001), whereas the total number of full arterial revascularizations was higher in off-pump patients (P = 0.0001). Off-pump patients showed a significantly shorter intensive care unit stay (P = 0.02), and less need for intra-aortic balloon pump insertion (P = 0.04). In-hospital mortality was 2.8% in on-pump patients and 2.7% in off-pump patients (P = NS). Although the hospital mortality rate was comparable for the 2 techniques, the in-hospital comparison between the 2 groups showed how the avoidance of cardiopulmonary bypass can significantly reduce the cumulative postoperative incidence of complications in patients undergoing coronary artery bypass grafting.