A prospective controlled trial was performed to determine whether the use of ABO-identical platelets from the start of treatment might provide higher post-transfusion platelet increments, reduce the number of platelet transfusions and ultimately delay the onset of refractoriness. Forty newly diagnosed patients with haematological diseases were randomized to receive either pooled ABO-identical platelets or pooled platelets unmatched for ABO group throughout their course. The corrected platelet count increments (CCI) were calculated for the first 25 transfusions of each patient and non-immune factors present at the time of each platelet transfusion were documented. The mean CCI for the first 25 transfusions in the ABO-identical group was significantly higher (6600 +/- 7900 SD) than that achieved with ABO unmatched platelets (5200 +/- 7900; p < 0.01). The effect was most marked for the first 10 transfusions for each patient where the CCI was 64% higher in the ABO-identical group (8200 +/- 7500 vs 5000 +/- 8100; p < 0.0002). Patients given ABO-identical platelets required only about half as many transfusions in the first 30 days (10 versus 17, p < 0.05) or during the first admission (11 versus 21 p < 0.01) as patients in the ABO-unmatched group. A smaller percentage of patients in the ABO-identical group became refractory (36% vs 75% p < 0.03). The data suggest that patients requiring long-term platelet support should be transfused with ABO-identical platelets.