The application of a centrifugal pump might lead to a reduced release of tissue factor (TF) due to less blood cell damage. This could result in a decrease in activation of the extrinsic pathway of coagulation and embolus formation. In the present study, 60 patients undergoing coronary artery bypass grafting were randomly assigned to a centrifugal or a roller pump. Plasma concentrations of TF, thrombin-antithrombin complex (TAT), and prothrombin fragments F1 + 2 were investigated before, during, and after cardiopulmonary bypass (CPB). Embolus detection was performed at the arterial line of CPB and transcranially by Doppler ultrasound. The centrifugal pump group revealed a lower TF release (area under the curve during CPB) when compared with the roller pump group [5661 (696-10359) vs 12681 (6383-17538) microg x min/l; median (lower - upper quartiles); P = 0.009]. In contrast, TAT and F1 + 2 formation did not differ between the groups, and neither did the total embolus count of both Doppler systems. Embolus counts did not correlate with TAT or F1 + 2 formation. In conclusion, the reduction in TF release by the application of a centrifugal pump seems to have little consequence on total thrombin formation. Since the applied Doppler systems seem to detect mainly microbubbles, conclusions regarding differences between the two pumps in the formation of thrombofibrinous clots cannot be drawn.