We studied the transfusion response from random donor platelet concentrates in 15 stable multitransfused, thrombocytopenic patients by comparing the platelet counts measured before and 20 hours after transfusion. The observed platelet increments were corrected (corrected increment, C.I.) for the number of units of platelet concentrate transfused and the patient's body surface area in square meters (platelets/microliter per unit/m2). Using platelet concentrates stored for less than 24 hours, the patients achieved a median C.I. of 9500 (range: 5000-18,000). When platelet concentrates stored for 24 to 48 hours or 48 to 72 hours were given, the median C.I. markedly decreased to 1000 (range: 0-4800) and 0 (range: 0-5100), respectively (p less than 0.001). These differences could not be explained by further recipient alloimmunization. Transfusion with platelet concentrates less than 24 hours old on a second occasion, bracketing the transfusions of older platelet concentrates, resulted in a median C.I. of 7200 (range: 5400-14,500). Similar results were obtained in three patients when HLA-identical sibling platelet concentrates were employed. In vitro tests, including pH, morphology, and aggregation, demonstrated no statistically significant differences among the platelet concentrates stored for less than 24 hours, 24 to 48 hours, and 48 to 72 hours. These studies suggest that, although platelet concentrates can be stored for 72 hours without loss of in vitro function, the in vivo recovery is significantly diminished after 24 hours of storage, and preferably patients should not be transfused prophylactically with platelet concentrates greater than 24 hours old.