There is almost general agreement that removal of leukocytes from blood components reduces the incidence of HLA-antibody formation and refractoriness to random platelet transfusions. Recently filters have become available, which are able to reduce leukocyte contamination in platelet suspensions with acceptable platelet loss. We evaluated a cellulose acetate (CA) and a polyester (PE) filter, and stored buffy coat-derived platelet suspensions after filtration. Both filters are effective for the removal of leukocytes to levels below 5 x 10(6) per transfusate. For the CA filter, platelet recovery was 73 +/- 13% yielding 256 +/- 53 x 10(9) platelets per transfusate from 6 donors. For the PE filter, platelet recovery was 90 +/- 9% and 327 +/- 51 x 10(9) platelets per transfusate. When a loading dose of less than 5 x 10(8) leukocytes was applied, 98% of the CA-filtered suspensions and 100% of the PE-filtered suspensions contained less than 5 x 10(6) residual leucocytes. In 123 patients transfusion results of CA-filtered platelet suspensions stored for 72 h, were compared with those obtained by non-stored, non filtered, random platelet suspensions which had been leukocyte depleted by differential centrifugation. Platelet increments 1 and 20 h after transfusion showed no statistical difference between CA-filtered platelet transfusions stored for 72 h and non-stored, non-filtered platelet transfusions. In a new cohort of 117 patients, two filters and various postfiltration storage times were compared. Using both filters, the 1-hour posttransfusion increments decreased to approximately 60% after 96 h of storage compared to results of storage periods of 72 h or less.(ABSTRACT TRUNCATED AT 250 WORDS)