OBJECTIVE: To examine whether specific platelet pheresis (minimal plasma harvested) would contribute toward reduced blood loss and allogenic blood requirements after cardiac surgery. DESIGN: A prospective randomized trial. SETTING: A large cardiothoracic surgical center. PARTICIPANTS: Consenting patients undergoing routine coronary artery or valve surgery (n = 54). INTERVENTIONS: Patients in the pheresis group underwent platelet pheresis in the anesthetic preparation room before general anesthesia. Pheresed platelets were stored during cardiopulmonary bypass and were returned to the patients after reversal of heparin with protamine toward the end of surgery. Control patients underwent their operations without this intervention. MEASUREMENTS AND MAIN RESULTS: Primary endpoints were blood loss and transfusion requirements. There were no differences between the 2 groups (pheresis v control: median loss, 960 mL v 1100 mL, p = 0.15; median blood transfused, 896 mL v 635 mL, p = 0.71). Secondary endpoints included analysis of platelet counts, platelet function, and surface markers. Counts remained the same after retransfusion of platelets up to 2 hours after surgery. Platelet aggregation to ristocetin was well preserved, but adenosine diphosphate caused almost no aggregation of the harvested platelets. Flow cytometry revealed the platelets to have a reduced surface density of the glycoprotein 1b receptor, and 13% of them were irreversibly activated. CONCLUSION: Platelet pheresis activates a proportion of the harvested platelets and impairs the function of the remainder; this may explain its failure to reduce postoperative blood loss and transfusion requirements.