Transfusion of red blood cells (RBC) and other blood products in patients undergoing coronary artery bypass grafting (CABG) is associated with increased mortality and morbidity. We retrospectively analyzed data of patients who underwent an isolated coronary bypass graft operation between January 1998 and December 2007. Mean follow-up was 1696+/-1026 days, with exclusion of 122 patients lost to follow-up and 80 patients who received 10 units of RBC. Of the remaining patients, 8001 (76.7%) received no RBC, 1621 (15.2%) received 1-2 units of RBC, 593 (5.7%) received 3-5 units and 220 (2.1%) received 6-10 units. The number of transfused RBC was a predictor for early but not for late mortality. When compared to expected survival, survival of patients not receiving any blood product was better, while survival of patients receiving >3 units of RBC was worse. Transfusion of RBC is an independent, dose-dependent risk factor for early mortality after revascularization. Compared to expected survival, receiving no RBC improves patient long-term survival, whereas receiving three or more units of RBC significantly decreases patient survival.