RESUMO
Results of 773 actual flow crossmatches (aFXMs) and virtual flow crossmatches (vFXMs) performed for living and deceased donor kidney transplantation in our center were analyzed retrospectively and evaluated for their concordance. Prediction of vFXMs was based on antibody identification using single antigen bead assay and locally established mean fluorescence intensity cutoff point compared with donor HLA antigens. The vast majority of aFXMs were in concordance with vFXMs with an overall concordance of 97%. Twenty-three predicted to be negative showed positive aFXMs; 12 of them had 0% calculated panel-reactive antibody, and 11 were found in patients with multiple non-donor-specific HLA antibodies. Three predicted positive vFXMs yielded negative aFXMs; 2 of them had allele-specific antibodies. CONCLUSIONS: vFXMs based on precise characterization of antibody specificities detected by single antigen bead assay using our cutoff point accurately predicted FXMs in the majority of patients and can be used safely to allocate kidney offers without performing physical crossmatches in selected patients.