Successful clinical ABO-incompatible renal transplantation has been achieved by the removal of anti-A or anti-B antibodies using double filtration plasmapheresis (DFPP). We have compared changes in the levels of anti-donor antibodies and the histopathology of the renal grafts following human ABO-incompatible allotransplantation and pig-to-baboon xenotransplantation using pretransplant DFPP. DFPP was performed on days 6, -4, -2 and -1 before the ABO-incompatible transplants (n=25) and on days -2 and 0 (immediately before reperfusion) in the xenotransplants (n=4). In two baboons (XenoTx Group I) the extent of antibody removal was comparable to that in the ABO-incompatible patients, and an even greater level of removal was achieved in another two baboons (XenoTx Group II). Anti-A and anti-B and anti-pig IgM and IgG antibodies were measured by flow cytometry. All clinical ABO-incompatible renal grafts are functioning, except two which were lost from recurrence of the original disease or from chronic rejection. Three other grafts underwent humoral rejection episodes, which were successfully treated. DFPP reduced the mean anti-A/B IgM and IgG antibody levels to 8% and 13% of pretreatment levels, respectively. After kidney transplantation, they were maintained at 37% and 46% of pre-DFPP level. No antibody binding to the transplanted kidney was detected at any time (1 h to 2 yr) after ABO-incompatible allotransplantation. In contrast, in XenoTx Group I, the same extent of antibody removal (90%) prevented hyperacute rejection, but the two grafts were rejected on days 6 and 7, respectively, from acute vascular rejection. In XenoTx Group II, the additional DFPP that was required to deplete the remaining 10% of anti-pig antibody was poorly tolerated and the two baboons died 4 h and 2 days, respectively, after renal transplantation. Although anti-pig IgM antibodies were reduced to 2% of pre-treatment level, IgM and C3 binding were detected in the graft as early as 1 h posttransplantation. These data suggest that the concentration of xeno-antigen epitopes expressed on pig organs may need to be reduced by genetic engineering to the much lower level of blood group A/B antigens on human kidneys if discordant xenotransplantation is to be successful.