BACKGROUND: Centrifugal pumps are being used increasingly for short-term extracorporeal circulation purposes such as during heart operations. Whether the centrifugal pump improves the cardiopulmonary bypass biocompatibility has not been fully documented. METHODS: A roller pump (n = 20) was compared in vivo with a centrifugal pump (n = 20) in groups of patients in which cardiopulmonary bypass circuits that were either totally heparin coated (Carmeda BioActive Surface; n = 20) or uncoated (n = 20) were used. We expected the heparin coating to attenuate blood activation, thus possibly making the comparison of the two pumps easier with respect to their different blood activation potentials. Samples of blood plasma, obtained during cardiopulmonary bypass from low-risk coronary artery bypass grafting patients, were analyzed for hemolysis (plasma haemoglobin), complement activation (C3bc and the terminal complement complex), a complement lytic inhibitor (vitronectin), coagulation activation (fibrinopeptide A), granulocyte activation (lactoferrin), and platelet activation (beta-thromboglobulin). RESULTS: The concentrations of terminal complement complex, lactoferrin, and beta-thromboglobulin were significantly lower in association with heparin-coated surfaces. The concentration of plasma hemoglobin was significantly lower in association with the centrifugal pump. In uncoated circuits, the beta-thromboglobulin level was significantly higher in association with the roller pump than with the centrifugal pump, but this significant reduction in the beta-thromboglobulin level did not hold true for the heparin-coated circuit group. CONCLUSIONS: A heparin-coated cardiopulmonary bypass surface reduces the blood activation potential during cardiopulmonary bypass, and the centrifugal pump causes less hemolysis than the roller pump.