To the Editor:-Capdevila et al. noted that the mechanism underlying the beneficial effect of aprotinin was not fully elucidated. We wish to make some comments in this regard. The blood- sparing effect of aprotinin has long been attributed to an early protective effect of aprotinin on platelet membrane glycoproteins, mainly Gp1b, altered by the contact of blood components with the foreign, non-biological surface of the cardiopulmonary bypass circuit (CPB). [2,3]Until recently the blood-sparing effect of aprotinin was only demonstrated in cardiac surgery conducted with CPB. During CPB, inhibition of fibrinolysis by aprotinin [3,4]could explain the protective effect on platelet membrane glycoproteins. Plasmin and d-dimer may degrade Gp1b receptors and the von Willebrand factor, which is responsible for platelet adhesion. [3,4]However, the inhibition of fibrinolysis is unlikely to be the mechanism by which aprotinin reduces intraoperative blood loss with CPB: the activation of fibrinolysis is apparent only at the end of CPB, whereas the Gp1b receptor consistently decreases within 5 min after start of CPB. Also, as noted by the authors, the reduction of intraoperative cytokine release by aprotinin has been assessed when a CPB is used. However, this effect has never been associated with any hemostatic changes. 
The intraoperative blood-sparing effect of aprotinin has been demonstrated in situations in which CPB is not used: elective liver resection and orthopedic surgery. [7,8]Consequently, the blood-sparing effect of aprotinin cannot be attributed primarily to a protective effect on platelet function previously altered by the contact of blood cells with the non-biological surface of a CPB. However, available data are controversial. Inhibition of intraoperative hyperfibrinolysis has been demonstrated in elective liver resection. Conversely, in two prospective, randomized, placebo-controlled studies of hip surgery, [7,8]a reduction in blood loss with high-dose aprotinin occurred but without any associated antifibrinolytic effect. [7,8]In these two studies, fibrinolysis was assessed by two reliable tests: postoperative fibrinogen and d-dimer concentration. The effect of aprotinin on platelet function was investigated in two randomized, placebo-controlled studies of hip replacement surgery. [7,9]In the former study, in the aprotinin and placebo groups:(1) the intraoperative platelet count and the bleeding time did not change significantly;(2) no evidence of in vivo intraoperative platelet activation was found (b thromboglobulin remained normal); and (3) no significant change in in vitro intraoperative ADP and collagen-induced platelet aggregability was observed. In the latter study, platelet function was altered intraoperatively in the control group, whereas (1) aprotinin preserved spontaneous platelet aggregation and ADP induced platelet aggregation and platelet adhesivity; and (2) postoperative platelet aggregates concentration decreased significantly with aprotinin when compared with placebo. We have not discussed liver transplantation because aprotinin has failed to significantly decrease blood loss in such cases. 
That aprotinin can achieve by different mechanisms an identical therapeutic goal in different surgical fields would be puzzling. Consequently, in regard to the mechanism of the blood-sparing effect of aprotinin, it is our opinion that this hypothesis should be revisited.
Claude Lentschener, M.D.
Dan Benhamou, M.D.
Professor Department of Anesthesiology; Hopital Antoine-Beclere; Universite Paris-Sud; 157, rue de la Porte de Trivaux; 92141 Clamart Cedex, Paris; France;dBenhamou.firstname.lastname@example.org
(Accepted for publication July 20, 1998.)