Some evidence has shown that platelet crossmatching is useful in multitransfused patients with hypoplastic bone marrows who are refractory to platelet therapy through alloimmunization. Several immunoglobulin binding assays other than enzyme-linked immunospecific assay (ELISA) have been studied previously. We performed 51 ELISA crossmatches on six patients receiving single donor platelets. One bone marrow transplant patient receiving 33 single donor HLA matched (related and unrelated) was also studied. Effectiveness of transfusion was closely monitored by patient evaluation and corrected platelet count increment (CCI) at 1-2 and 18-24 hours posttransfusion. We found the ELISA method very sensitive, specific, and predictive, 85, 96, and 95.6% respectively in the 51 crossmatches studied in six patients with either leukemia, solid tumors, or aplastic anemia. However, variation existed among individual recipients, with sensitivity ranging from 70-100%. The distribution of true positives and negatives and false positives and negatives in the 33 crossmatches performed in the bone marrow transplant patient differed significantly (chi 2 = 101.2; P less than 0.001) from single donor recipients. The specificity in the 51 crossmatches on the six patients was also significantly different from the 33 crossmatches performed in the bone marrow transplant (96 vs 74%). This suggests individual variation occurs as well as differences in diseases and bone marrow suppressive agents affecting platelet crossmatching.