An optimal platelet-count threshold for prophylactic platelet transfusion in hematopoietic stem cell transplant (HSCT) recipients has yet to be determined. Between July 1997 and December 1999, we performed the first prospective randomized clinical trial addressing this issue in 159 HSCT recipients who received a prophylactic platelet transfusion when the morning platelet count fell below a 10,000/microL (10K) or 20,000/microL (20K) threshold. Subsequent prophylactic transfusions were administered according to a predetermined algorithm. The number of prophylactic and therapeutic transfusions and the incidence of minor and major bleeding were compared between the 2 groups. The groups were matched according to patient and transplantation characteristics. There were no significant differences in bleeding incidence or severity. Fourteen percent of patients in the 10K arm compared to 17% in the 20K arm had major bleeding events. Only 3 central nervous system bleeds occurred, 2 in the 10K group and 1 in the 20K group. No deaths were attributed to bleeding. An average of 11.4 days of bleeding occurred in both groups. An average of 10.4 platelet transfusions per patient were administered in the 10K group compared to 10.2 in the 20K group (P = .94). More transfusions were given above the assigned transfusion threshold in the 10K group than in the 20K group (4.3/patient versus 1.9/patient, respectively, P = .05). Safety measures incorporated into our study may have precluded demonstration of significant differences in platelet use between the groups. In conclusion, a platelet transfusion trigger of 10K was found to be safe; however, a decrease in platelet use was not achieved because of safety measures incorporated into our study design.