ESPN 54th Annual Meeting

ESPN 2022

Assessment of HLA incompatibility at the molecular compared to antigenic HLA level enables better prediction of graft function deterioration in paediatric kidney transplantation.
JON JIN KIM 1 Alexander Fichtner 2 Hannah Copley 1 Caner Susal 2 Lars Pape 3 Jun Oh 4 Kai Krupka 2 Burkhard Toenshoff 2 Vasilis Kosmoliaptsis 1

2- University of Heidelberg
3- University of Hannover
4- University of Hamburg

Introduction: HLA mismatching has a detrimental effect on graft survival after paediatric kidney transplantation. Assessment of HLA incompatibility at the molecular level (molecular HLA mismatch; molMM) has emerged as a promising method for predicting primary alloimmunity risk. In this study, we aimed to assess whether molMM compared to conventional antigenic mismatching (antMM) enables better prediction of graft function deterioration in paediatric kidney transplantation.

Material and methods:

Methods: We performed a retrospective analysis of 177 paediatric patients from the ABMR study of the CERTAIN registry (median follow-up 4.5 (IQR 3-5) years). Only five patients experienced graft loss. Therefore, we used the time to 50% reduction in eGFR, from month-3 post-transplant baseline, as a surrogate endpoint for long-term graft loss (eGFR-50). HLA molMM was assessed using the Cambridge amino acid mismatch score (AAMS) which on a separate study was predictive of primary alloimmunity risk in this cohort. Survival analysis was performed using Cox models, adjusted for donor and recipient baseline characteristics.


Results: 27 (15%) patients met the primary outcome. In multivariable analysis, recipient and donor age, baseline eGFR, and re-transplant status had a significant association with eGFR-50. Importantly, only mismatches at HLA-DQα1β1, and not at other loci, were associated with the primary outcome (adjusted HR (aHR) 10.2; 95% CI, 10.1-10.4 per 10 AAMS increase, and aHR 1.8; 95% CI, 1.02-3.4 per antigen increase). There was a wide range of AAMS values (0-49) within each HLA-DQ antMM (0-2). We used a predetermined molMM score (AAMS=16 derived from analyses of donor-specific antibody risk) to classify patients according to HLA-DQα1β1 mismatch risk into “low/low”, “low/high” and “high/high” risk groups. Patient risk for eGFR-50 was associated with molMM stratification, regardless of their antMM. eGFR-50 risk was equivalent in patients with “low/low” risk mismatches (aHR 2.2, 0.2-23, p=0.5 for 2 v 0 antMM). Patients with “low/high” and “high/high” HLA-DQα1β1 mismatches had worse allograft outcomes (“low/high”: aHR 4.7, 1.9-11.4, p<0.05; “high/high” aHR 5.7, 1.4-22.7, p<0.05 versus “low/low”, Figure 1). Compared to antMM, molMM showed better stratification of outcomes whilst increasing the number of patients in the low risk group (“low/low” n=100, v 0 antMM n=65).


Conclusion: Assessment of HLA incompatibility at the molecular level enables better stratification of graft deterioration risk compared to conventional serology-based HLA mismatching. Further validation of molMM in independent cohorts is required before clinical implementation.