The influence of four different kinds of intravascular volume replacement on platelet function was investigated in 60 patients undergoing elective aortocoronary bypass grafting using cardiopulmonary bypass (CPB). In a randomized sequence, high-molecular weight hydroxyethyl starch solution (HMW-HES, mean molecular weight [Mw] 450,000 d), low-molecular weight HES (LMW-HES, Mw 200,000 d), 3.5% gelatin or 5% albumin were infused preoperatively to double reduced filling pressure (pulmonary capillary wedge pressure [PCWP] <5 mm Hg). Fifteen untreated patients served as a control. Platelet function was assessed by aggregometry using turbidometric technique (inductors: ADP, epinephrine, collagen). Maximum aggregation, maximum gradient of aggregation, and platelet volume were measured before, during, and after CPB until the first postoperative day. HMW-HES 840 ± 90 mL, LMW 850 ± 100 mL, gelatin 950 ± 110 mL, and albumin 810 ± 100 mL were given preoperatively. Maximum platelet aggregation (ranging from −23% to −44% relative from baseline value) and maximum gradient of platelet aggregation (ranging from −26% to −45% relative from baseline values) were reduced only in the HMW-HES patients. After CPB, aggregometry also was impaired most markedly in these patients. The other volume groups showed less reduction in platelet aggregation and were similar to the untreated control. On the first postoperative day, aggregation variables had returned almost to baseline in all patients. Platelet volume was the same among the groups within the investigation period. Postbypass blood loss was highest in the HMW-HES group (890 ± 180 mL). There was significant (P < 0.04) correlation in this group between blood loss and change in platelet aggregation. The need for homologous blood was comparable among all groups. It is concluded that volume replacement with HMW-HES resulted in the most pronounced impairment of platelet aggregation associated with the highest postoperative blood loss. Thus HMW-HES should be avoided particularly in patients at risk of enhanced and prolonged postoperative bleeding, whereas the other solutions were similar with regard to platelet function. (Anesth Analg 1993;76:1185-90)