Prediction and Validation of Severe Early Allograft Dysfunction Following Pediatric Living Donor Liver...

Severe early allograft dysfunction (EAD) markedly compromises transplant outcomes for both the graft and recipient. There remains a scarcity of research focused on severe EAD following pediatric living donor liver transplantation (LDLT). The present study aims to identify risk factors associated with severe EAD after pediatric LDLT and to develop and validate a corresponding predictive model.

This retrospective study enrolled patients (training cohort, n = 218; validation cohort, n = 101) who underwent pediatric LDLT between May 2018 and January 2025. The severity of EAD was graded using the liver graft assessment following transplantation (L-GrAFT) risk score. Univariate and multivariate logistic regression analyses were employed to identify independent risk factors associated with severe EAD. A predictive nomogram was developed based on risk factors to estimate the probability of severe EAD. The area under the receiver operating characteristic curve (AUC) was used to assess predictive performance.

Severe EAD occurred in 26 (11.9%) patients in the training cohort and 5 (5%) patients in the validation cohort. A multivariate analysis confirmed that leukocyte count (odds ratio (95% confidence interval), P-value; 0.74 (0.62-0.88), 0.001), serum creatinine (0.88 (0.78-0.98), 0.02), albumin (0.85 (0.77-0.93), 0.001), graft steatosis (>30%) (132.88 (6.78-2604.32), 0.001), cold ischemia time (1.01 (1.00-1.02), 0.04), and duration of surgery (3.54 (1.87-6.69), <0.001) were independent risk factors for severe EAD. A nomogram was developed using independent risk factors. The calibration curve showed excellent consistency between nomogram-predicted probabilities and actual observations in the training cohort. The nomogram demonstrated an AUC of 0.92 (95%CI: 0.87-0.97) in the training cohort and 0.76 (0.52-1) in the validation cohort for predicting severe EAD.

This study is the first to identify risk factors for severe EAD following pediatric LDLT and to develop and validate a predictive nomogram. The application of this model may assist clinicians in evaluating the risk of severe EAD in an objective and efficient manner.