Pediatric discard risk index for predicting pediatric liver allograft discard.

ABSTRACT

BACKGROUND: Of the 600 pediatric candidates added to the liver waiting list annually, 100 will remain waiting while over 100 liver allografts are discarded, often for subjective reasons. METHODS: We created a risk index to predict discard to better optimize donor supply. We used the UNOS database to retrospectively analyze 17 367 deceased donors (≤18 years old) through univariate and multivariate logistic regression models. Deceased donor clinical characteristics and laboratory values were independent variables with discard being the dependent variable in the analysis. Significant univariate factors (P-value < .05) comprised the multivariate analysis. Significant variables from the multivariate analysis were incorporated into the pDSRI, producing a risk score for discard. RESULTS: From 17 potential factors, 11 were identified as significant predictors (P < .05) of pediatric liver allograft discard. The most significant risk factors were as follows DCD; total bilirubin >10 mg/dL, and alanine transaminase (ALT) ≥500 IU/L. The pDSRI has a C-statistic of 0.846 for the training set and 0.840 for the validation set. CONCLUSION: The pDSRI uses 11 significant risk factors, including elevated liver function tests, donor demographics, and donor risk/type to accurately predict risk of pediatric liver allograft discard and serve as a tool that may maximize donor yield.