Introduction: Tacrolimus dose adjustment most commonly based on its trough concentrations (FKC0) is fundamental to prevent graft rejection. Risk factors including HLA histocompatibility are documented in several solid organ transplants except in liver transplantation. Moreover, molecular HLA eplet mismatch (eMM) has demonstrated its superiority as a biomarker for the immunological risk stratification but scarce data is available in pediatric liver transplantation. Our aim was to evaluate the correlation between HLA eMM, FKC0 variability and the interaction of both variables as risk factors for biopsy-proven acute rejection (BPAR).

Methods: Pediatric patients transplanted between 2018 and 2021 at our hospital were included and prospectively followed. Liver-graft recipients and donor pairs were HLA-typed by Next Generation Sequencing, and the number eMM for every loci was quantified using the HLAMatchmaker version of HLA Fusion software 4.6. Tacrolimus variability was calculated using the percent coefficient of variation (CV%) and tortuosity. The percent of C0 above the minimum threshold at each window of time after transplantation was also calculated (CaT). Demographic, clinical, and pharmacological covariares were included in univariate (p<0.2) and Cox multivariate models. ROC analysis was performed to identify specific thresholds for BPAR development in the variables retained in the multivariate model. Kaplan-meier curves for the BPAR-free survival according to the thresholds obtained in the ROC analysis were constructed. For variables retained in the multivariate model, Fisher’s exact test was performed.

Results: Fifty-two of the 117 enrolled patients with a mean (range) age of 1.5 years (0.5-17.3), had full available data and were on tacrolimus as primary immunosuppression. Patients were followed for a mean of 350 days (range: 36-1025).  BPAR-free survival was 75.2% (CI95%, 63.7-88.7) and 62.7% (CI95%, 49.1-80.0) at 1 and 2 years post-transplant, respectively. Tacrolimus tortuosity (HR 14.04, 95%CI, 4.37-45.12; p<0.001) and HLA-DQ eMM (HR 1.14, 95%CI, 1.03-1.30; p=0.013) were significant risk factors for BPAR. Having a tacrolimus tortuosity≥1.1 (sensitivity 50%, specificity 79.4%; AUC=0.676) and HLA-DQ eMM load≥5 (sensitivity 72%, specificity 44.1%; AUC=0.590) was considered high tortuosity and high HLA-DQ eMM load according to ROC curves, respectively. Patients with low HLA-DQ eMM load and low tacrolimus variability were as likely to develop BPAR as those with high variability. However, patients with high HLA-DQ eMM load and high tortuosity (12/32) showed a higher proportion of BPAR (n=8) compared to those with high HLA-DQ eMM and low tortuosity (20/32) that developed BPAR (n=5), p=0.03.

Conclusion: Tacrolimus variability and HLA-DQ eMM were identified as risk factors for BPAR. Patients with high HLA-DQ eMM load are less likely to tolerate high tacrolimus variability without developing BPAR compared to those with low HLA-DQ eMM load.