Refractoriness to infused platelets becomes a major clinical problem for many patients with acute myeloid leukemia (AML). Inadequate post-transfusion platelet count increments can be due to a number of host-related factors such as: spleno-megaly, severe infection with high fever, disseminated intrava-scular coagulation, drug-mediated antibodies and/or alloim-munization, and occasionally by administration of platelets damaged or activated during collection or storage. Lymphocy-totoxic antibody directed against HLA-A orB antigens is an excellent serologic marker for alloimmunization. In one large study, the presence of anti-HLA antibodies was correlated with a poor post-transfusion platelet increase in over 90% of such patients, whereas patients lacking antibodies had satisfactory increments after transfusion 85% of the time. 1 Similar predictive information can be obtained using a variety of anti-platelet antibody tests. Post-transfusion counts obtained at the conclusion of transfusion, at a time that the patient must be seen by the physician or nurse to switch infusion bags, are identical to counts obtained 1 h later and it is therefore relatively simple to assess the results of every platelet transfusion to determine if the expected count increment was achieved. 2 The incidence of alloimmunization appears to be unrelated to the number of transfusions that patients receive. In a series of 114 newly diagnosed AML patients receiving 2-10+ transfusions of pooled non-leukocyte-depleted platelet concentrates (PC) from random donors, there was no relationship between the number of transfusions or donor exposures, and the development of new anti-HLA antibody. 3 At the end of induction therapy, 35-40% of patients had developed anti-body. The majority had persistence of the antibody, but 20% had disappearance of antibody over time. After marrow recovery from the induction therapy, 60% of patients remained anti-body-negative and, indeed, never developed antibody or refractoriness to platelet transfusion despite multiple subsequent red blood cell and platelet transfusions, suggesting that they had become immune tolerant to histocompatibility antigens. These patients continued to be easily supported with random donor transfusions and therefore measurement of lym-phocytotoxic antibody at this time in a patient's treatment course is helpful in predicting and arranging for their future transfusion needs. The traditional management of alloimmune-mediated plate-let refractoriness is by HLA typing of donors and recipients. A significant fraction of HLA 'matched' transfusions do not produce satisfactory increments, however, while some 'mismat-ched' transfusions are successful, and a number of studies have evaluated subtleties of HLA serology to attempt to explain these discrepancies. As one example, there is a high