Fifty-three patients receiving long-term platelet transfusions were regularly screened for platelet-associated antibodies by a platelet suspension immunofluorescence test (PSIFT) and a lymphocytotoxicity test (LCT). Subsequently, 24 patients became alloimmunized; all of their antibodies were of HLA specificity. Eighty-two single-donor platelet transfusions were given, and the clinical responses were considered satisfactory if the 18-hour corrected count increment was 7.5 x 10(3) per microL or higher. In the meantime, 82 pairs of patient sera and donor lymphocytes were crossmatched. Among 63 crossmatched transfusions, 53 (84%) resulted in a satisfactory increment, with a mean (+/- SEM) of 17.71 +/- 1.96 (x 10(3)/microL), and 10 did not result in a satisfactory increment. The increments after 19 unmatched transfusions and 25 random-donor (uncrossmatched) transfusions were 0.7 +/- 0.3 and 2.39 +/- 0.66, respectively. The difference was not significant (p greater than 0.05). The agreement between the LCT results and clinical response was 88 percent. Retrospectively, the corrected count increments showed no significant differences (p greater than 0.05) among three groups of HLA grading: the increments for A/BU/BX, C/D, and random HLA matches were 22.97 +/- 4.07, 15.1 +/- 1.97, and 14.85 +/- 2.04, respectively. These results suggest that platelet crossmatching by LCT is an effective method for use in alloimmunized patients, especially Chinese patients.