Degree of HLA class II eplet mismatch load improves prediction of antibody-mediated rejection in living donor kidney transplantation

Abstract Background HLA mismatching is a well known risk factor for worst outcomes in kidney transplantation. Methods In the present study, HLA antigen and eplet mismatches were determined in 151 living donor-recipient pairs transplanted between 2007 and 2014 and rejection episodes and graft survival were evaluated. Results We found that high HLA-II eplet mismatch load (EpMM ≥ 13, versus low EpMM ≤ 5), was an independent predictor of AMR (adjusted HR = 14.839; P = 0.011), while HLA-II AgMM was not. We also showed that HLA-II EpMM load was a significant better predictor of AMR than AgMM (c-statistic = 0.064; P = 0.023). After discriminating HLA-II into HLA-DR and HLA-DQ loci we demonstrated that high versus low eplet mismatch load for HLA-DR (T3 ≥ 6 versus T = 0–1, p = 0.013) and HLA-DQ (T3 ≥ 7 versus T = 0–1, p = 0.009) are independent predictors for AMR. HLA-II EpMM increased discrimination performance of the classical HLA-II AgMM risk model (IDI, 0.061, 95%CI: 0.005–0.195) for AMR. Compared with AgMM, HLA-II eplet model adequately reclassified 13 of 17 patients (76.5%) with AMR and 92 of 134 patients (68.7%) without AMR (cfNRI, 0.785, 95%CI: 0.300–1.426). Conclusions Our study evidences that eplet-based matching is a refinement of the classical HLA antigen mismatch analysis in LDKT and is a potential biomarker for personalized assessment of alloimmune risk.