In a prospective study, 237 patients undergoing a primary coronary bypass operation were randomized to receive 2 units of fresh whole blood (study group) or stored (2-5 d) blood (control group) at the end of the extracorporeal circulation. Serious post-operative bleeding necessitating a re-thoracotomy occurred in 4.2% of all patients with an equal distribution over the two groups. Post-operative haemoglobin content and platelet counts were higher in the study group, but the differences were small and clinically not important. There were no differences in transfusion requirements, post-operative blood losses and haemostatic parameters between the trial groups. At low post-operative platelet counts (below 120 X 10(9) platelets/l) however, patients in the control group lost significantly more blood and had increased transfusion requirements compared with patients in the study group (7.1 versus 4.8 units). These differences must be attributed to qualitative platelet defects in the transfused units of stored blood. The small, clinically insignificant, differences in two laboratory parameters between the study and control groups, and the increased transfusion requirements of a subpopulation of patients with low platelet counts in the control group do not justify giving fresh blood or prophylactic platelet transfusions to coronary bypass patients.