Background: The reduction of white cells in blood components before transfusion by filters with at least 3 log(10) depletion may prevent adverse transfusion reactions such as HLA alloimmunization, febrile reactions, transmitted infections, and immunomodulation. A new generation of filters with 4 log(10) depletion is now available. Study Design and Methods: The aim of this study is to compare the efficiency of white cell reduction by six commercial filters for packed red cells with 3 and 4 log(10) depletion (claimed by manufacturers). The analysis of white cell concentration In the white cell-reduced units was performed by flow cytometry and with a Nageotte chamber. Results: The last generation of filters (BPF4, RC400, R01 Plus) show mean residual white cell numbers of 0.18 +/- 0.14, 0.26 +/- 0.21, and 0.25 +/- 0.15 x 10(6), respectively, by flow cytometric analysis and 0.05 +/- 0.04, 0.18 +/- 0.16, and 0.38 +/- 0.23 x 10(6), respectively, by Nageotte chamber evaluation. The 3 log(10) depletion filters (R01, Leucostop-4LT-mono, R200) have mean residual white cell numbers of 1.41 +/- 0.92, 2.4 +/-. 1.99, and 1.05 +/- 0.64 x 10(6), respectively, by flow cytometric analysis and 3.56 +/- 1.7, 1.67 +/- 1.3, and 3.21 +/- 4.1 x 10(6), respectively, by Nageotte chamber evaluation. The data show that the BPF4, RC400, and R01 Plus filters are likely to be more efficient by 1 log(10) reduction than the R01, Leucostop-4LT-mono, and R200 filters. Conclusion: The most recent generation of filters is able to deplete white cells from packed red cells by 4 log(10) in particular, with one of the filters, the residual WBC content was less than 0.5 x 10(6) per unit in all experiments, while two other filters reached that level in 9 of 10 experiments.