The role of humoral immunity in causing antibody-mediated rejection (AMR) of organ allografts has been extensively documented. For this reason, negative complement-dependent cytotoxicity (CDC) cross-matches between recipient sera and donor T and B lymphocytes have become a mandatory requirement for cadaveric kidney transplantation. However, the significance of donor-specific antibodies (DSAs) detectable only by flow cytometry (FC) or solid phase assays (SPA) but not CDC is still controversial. We have performed a retrospective analysis of FC cross-matching results in 80 consecutive cadaver kidney allograft recipients. Antibodies against HLA class I and class II antigens were measured by CDC and SPA in sequential samples of sera obtained prior to transplantation. The preoperative cross-match was performed by CDC using magnetically sorted T and B cells from donor spleen. Sera obtained from each patient before and at the time of transplantation were included in the final cross-match. The sample of serum obtained at the time of transplantation was cross-matched retrospectively by FC and analyzed for anti-HLA antibody specificity on high resolution SPA. The actuarial kidney allograft survival at one year was 98%. Two of these eighty patients lost the graft, one due to AMR, the other for reasons unrelated to DSAs. Donor-specific antibodies were detected by FC in 17 of 80 patients, yet only 6 of 17 had an early episode of AMR. This episode was successfully reversed by desensitization therapy using intravenous immunoglobin (IVIG) and plasmapheresis. Flow cytomery cross-matching showed 95% specificity but only 35% sensitivity for prediction of AMR (p = 0.002). There was a significant correlation between high panel reactive antibodies (PRA) and positive FC cross-matching (p = 0 .0001), as well as high PRA and AMR (p = 0.0004 by CDC and 0.0011 by Luminex). Reversible AMR occurred 12-30 days post-transplantation in 8 patients. Of these 8 patients, 3 had no detectable DSAs in spite of C4d positivity, 4 had C4d deposition in conjunction with anti-HLA antibodies, and 1 patient had DSAs (anti-MICA) yet no C4d deposition. We conclude that early initiation of desensitization protocols can prevent transplant failure and that retrospective FC cross-matches may facilitate the diagnosis of AMR. Extensive analysis of patients' sera using a comprehensive set of tests may contribute to early treatment and better understanding of the mechanism underlying humoral rejection.