Background. We have previously presented a method for performing coronary artery bypass graft operation on the beating heart without cardiopulmonary bypass (CPB). This method has now been explored. Method. Thirty-two patients were prospectively randomized. The study group was operated on using an axial blood now pump (Hemopump; HP) as circulatory support. Operations were performed on the beating heart. The control group was operated on using CPB, aortic cross-clamping, and cardioplegic arrest. Results. All patients went through the procedure without major complications, and were discharged from the hospital. No statistical differences were observed between the groups for time on support (HP, 60.5 minutes; CPB, 70.5 minutes) or total operating time (HP, 178 minutes; CPB, 162 minutes). The number of grafts was greater in the CPB group (HF, 1.8; range, 1 to 3; CPB, 2.5; range, 1 to 4; p = 0.03). Statistical differences were found for intraoperative bleeding (HP mean, 312 mt; CPB mean, 582 mt; p = 0.0003) and myocardial trauma as measured by postoperative troponin-Tr values (HP, 0.23 mu g/L; CPB, 1.17 mu g/L; p = 0.004). Conclusions. Hemopump-supported coronary artery bypass graft operation has been shown to be a safe and feasible procedure with the potential benefits of reduced operative bleeding and myocardial damage without prolonging intraoperative support or total operating time. (C) 1999 by The Society of Thoracic Surgeons.