The effectiveness of different methods of leukoreduction in preventing alloimmune platelet refractoriness was evaluated in a canine model. Platelets from a random donor dog were administered for up to 8 weeks or until platelet refractoriness. Standard (STD; unmodified) platelets were accepted by 14% of recipients (n = 7) compared with 14% for centrifuge leukoreduced (C-LR) platelets (n = 21) and 31% for filter leukoreduced (F-LR) platelets (n = 13; no significant differences). Surprisingly, using both F-LR and C-LR platelets was highly effective (87% acceptance, n = 15). Transfusing F-LR/C-LR red blood cells (n = 4) or F-LR/C-LR plasma (n = 4), along with F-LR/C-LR platelets, did not affect platelet acceptance (100% acceptance). Overall acceptance of F-LR/C-LR platelets was 91% (n = 23; P < or = .05 versus STD, C-LR, or F-LR platelets). F-LR/C-LR transfusions also induced tolerance to subsequent STD platelet transfusions from the same donor (82% acceptance, n = 19) as well as to donor skin grafts without recipient immunosuppression (57% acceptance, n = 7). To evaluate mechanisms of tolerance induction, F-LR/C-LR platelets were gamma-irradiated. Although the gamma-irradiated F-LR/C-LR platelets were uniformly accepted (n = 6), tolerance to STD platelets was lost. These data suggest that some allostimulatory white cells are filter adherent, whereas others escape filtration but can be removed by centrifugation and tolerance requires a residual functioning white cell.