A descriptive analysis of international transfusion practice and bleeding outcomes in patients with acute leukemia.

BACKGROUND Recently, bleeding has been used in platelet (PLT) trials rather than surrogate outcomes. The purpose of this study was to provide a descriptive summary of data from PLT studies conducted in four countries and exploratory analyses to determine the relationship between bleeding and PLT count. STUDY DESIGN AND METHODS A descriptive analysis was performed on original data from the Italian trigger study, the US TRAP study, the Canadian febrile reaction study, and a clinical chart review from one hospital site in the United Kingdom. The relationship between bleeding and PLT count was explored with the Italian data. RESULTS A total of 897 patients with acute leukemia received 10,506 PLT transfusions. Grade 3 or Grade 4 WHO bleeding frequency was 28.1 percent (252/897) but varied by country: Italy, 10.8 percent (27/250); United States, 36.4 percent (217/598); Canada, 18.9 percent (7/37); and the United Kingdom, 8.3 percent (1/12). Grade 1 or Grade 2 bleeding was reported only in the Italian study (46.4%[116/250] and 11.6%[29/250], respectively). The relative rates of WHO Grade 2, 3, or 4 bleeding for PLT counts in the ranges of 0 to 4, 5 to 9, 10 to 14, and 15 to 19 (x10(9)/L) were 8.8, 1.9, 1.8, and 1.2, respectively, compared to those counts within the range of 20 to 29 (x10(9)/L). CONCLUSION The study provides descriptive data on PLT use and frequency of bleeding. When PLT counts were 0x10(9) to 4x10(9) per L there was an eightfold increase in bleeding and a twofold risk increase when counts were 5x10(9) to 14x10(9) per L compared to the 20x10(9) to 29x10(9) per L reference range. The increased rate of bleeding at low counts occurred despite PLT therapy.