Predictive value of HLAMatchmaker and PIRCHE-II scores for de novo donor-specific antibody formation after adult and pediatric liver transplantation.

Production of de novo DSA (dnDSA) is associated with an increased risk of antibody mediated rejection after liver transplantation. Antibodies not only recognize the entire antigen but are able to bind specific functional epitopes present on the HLA molecule surface. The HLAMatchmaker and the PIRCHE-II (predicted indirectly recognizable HLA epitopes) algorithms are able to determine predictive epitope mismatches scores and de novo DSA (dnDSA) synthesis based on alloreactive eplets' identification. The aim of the present study was to assess, for the first time in liver transplantation, the complementarity between these two algorithms. We retrospectively analyzed a cohort of 407 adult and 133 pediatric liver transplant patients without preformed DSA, transplanted between 1991 and 2019 in Lyon and Montpellier. HLA antibodies were detected by single antigen bead assay. HLA typing of the donor-recipient pair was achieved by serological and/or DNA-based techniques. PIRCHE-II and HLAMatchmaker algorithms were then applied on both groups. During follow-up, 27.3% of adults and 38.3% of children developed dnDSA. HLA-DRB1 and DQB1-PIRCHE-II and HLAMatchmaker scores were significantly higher in dnDSA group compared to no DSA group for both pediatric and adult patients (except for PIRCHE-II HLA-DRB1 locus score in pediatrics). ROC curves allowed determining score thresholds classifying patients in low- and high-risk of dnDSA synthesis. The two algorithms' Kaplan-Meier curves showed a predicted incidence of dnDSA 20 years after transplantation significantly lower in the low-risk group compare with the high-risk group (log rank <0.05), in both cohorts, with a good negative predictive value. In conclusion, HLAMatchmaker and PIRCHE-II algorithms both are effective tools to identify anti-HLA immunization risk and to predict dnDSA formation after liver transplantation.

Informations de copyright

Copyright © 2020 Elsevier B.V. All rights reserved.