Platelet concentrates transfused for correction of thrombocytopenia or reduced platelet function do not consistently improve primary haemostasis in the recipient. Insufficient therapeutic effects may be caused by impaired donor platelet function and by unfavourable donation and storage conditions, as well as by immunological interactions with the recipient blood. The present study was designed to investigate whether the effect of platelet transfusion on recipient platelet function can be predicted by in vitro methods. METHODS. Blood samples were taken from 12 thrombocytopenic patients before (20 ml, P0) and after (10 ml, P(vivo)) transfusion of one unit of platelets previously stored for 24-120 h in acid citrate dextrose. An additional sample was taken from the platelet concentrate (TK) immediately before transfusion. P0 was divided into two specimens and TK platelets were added to one of them (P(vitro) in order to obtain a platelet count similar to that in P(vivo). Bleeding time (BT) and bleeding volume (BV) of the samples P0, P(vivo) and P(vitro) were measured using the method of Kratzer and Born (Fig. 2); mean values were calculated for each sample from six measurements. Aggregability of TK platelets was determined in addition by aggregometry. In contrast to previous studies, physiological Ca2+ concentrations were restored and secondary haemostasis was inhibited by low-molecular-weight heparin (Fragmin P, Pfrimmer Kabi GmbH und Co. KG, Erlangen) in the platelet-rich plasma used for aggregometry. RESULTS. Platelet counts increased in all patients after transfusion (P(vivo) vs P0, Table 1) and were nearly identical in P(vitro) and P(vivo) (r = 0.94, P < 0.001; Fig. 3). Parameters of primary haemostasis were significantly improved by addition of platelets to P0 in vitro (BT P < 0.05, BV P < 0.01) as well as by platelet transfusion (BT P < 0.05, BV P < 0.01). Direct comparison of P(vitro) and P(vivo) yielded a very close correlation of BT (r = 0.88, P < 0.001) and BV (r = 0.89, P < 0.01) in both samples. Although aggregometry revealed decreasing platelet function with increased storage time, aggregability was considerably higher compared to previous studies of platelet concentrates stored for 2-5 days. CONCLUSION. A new technique has been developed which allows reliable prediction of the effect of platelet concentrates on primary haemostasis of the recipient by in vitro measurement of bleeding time and bleeding volume prior to transfusion using the method of Kratzer and Born.