To determine the relative importance of clinical factors on the efficacy of platelet transfusions, 941 pooled platelet transfusions from HLA-unmatched donors were studied prospectively in 133 patients with bone marrow failure. Multiple linear regression analyses identified the major factors influencing one-hour-corrected increments (CI) as prior splenectomy, bone marrow transplantation, disseminated intravascular coagulation, concurrent intravenous amphotericin B, splenomegaly, and HLA antibody grade. The relative impact of these factors on CI has been quantitated by using a formula developed from these data. A linear relationship was demonstrated between increasing percentage of HLA antibody grade and decreasing CI. A number of other factors were less important in the linear regression model than the aforementioned major factors. These included platelet-specific antibodies, concurrent antibacterial antibiotics, clinical bleeding grade, and temperature. Factors that did not influence CI included the number of prior platelet transfusions, prior granulocyte transfusions, prior red cell transfusions, infection, age, blood group, diagnosis, sex, pretransfusion platelet count, prior pregnancies, and concurrent antineoplastic drugs. This study identified major clinical factors that significantly influenced CI and were major causes of refractoriness to pooled platelet transfusions.