The transfusion records of 189 patients with acute leukemia were analyzed to correlate lymphocytotoxic antibody (LCTAb) levels with response to a series of random-donor platelet transfusions (Tx). Twenty-one patients were studied twice at times of different LCTAb levels. All transfusions were given when patients were clinically stable without disseminated intravascular coagulation, bleeding, temperature greater than 101 degrees F, or splenomegaly. The mean 1-hr and 24-hr corrected count increments (CCI) for all patients with negative LCTAb were 16,100 and 12,000, and values for patients with positive LCTAb were 5,600 and 2,600 (P less than 0.0005). Thirteen patients had intermediate LCTAb (10-20%) and a variable response to Tx. Of the 137 patients with negative LCTAb levels 106 (77%) had good mean CCI of greater than 10,000 at 1 hr and greater than 7,500 at 24 hr following transfusion. In contrast of 60 patients with positive LCTAb (greater than 20% cytotoxicity), 53 (88%) had poor CCI of less than 10,000 at 1 hr and less than 7,500 at 24 hr after transfusion. Only 4 patients with positive LCTAb had a good response to random donor platelets at both 1 and 24 hrs. Eighteen patients had negative LCTAb with a high 1-hr and low 24-hr CCI. Thirteen of these had a history of positive LCTAb and in 9 there was an anamnestic rise following transfusion. Nine of 137 patients had negative LCTAb with low 1-hr and 24-hr CCI. LCTAb is highly predictive of response to random donor platelets. Cytotoxicity to greater than 20% of tested lymphocytes virtually precludes a good CCI at 1 and 24 hr.