BACKGROUND & AIMS: Adult studies of autoimmune hepatitis (AIH) have shown that the model of end-stage liver disease is associated with resistance to first-line treatment. Using a multicentre retrospective database, we sought to determine if the paediatric end-stage liver disease (PELD) score would similarly predict treatment resistance in paediatric AIH.

METHODS: One hundred and seventy-one children from 13 Canadian centres who fulfilled the International Autoimmune Hepatitis Group (IAIHG) criteria were included and assessed for change to second-line therapy within 24 months of primary treatment onset. Those with PSC overlap at presentation, or missing data on the PELD variables were excluded. PELD was calculated for all remaining patients. Univariate analysis and receiver-operator characteristic (ROC) curves were performed to determine the predictive ability of the PELD score to change to second-line therapy.

RESULTS: A total of 103 children were included with median age of 11 years (range 2-17). Mean PELD was -2.51±8.58. Second-line therapy was used within 24 months of diagnosis in 13 patients. Univariate analysis revealed that change to second-line therapy was associated with higher PELD (P=.028) and internal normalized ratio (INR) (P=.011). ROC curves for PELD and its individual components were performed. The strength of association was strongest with INR (AUC 0.72; CI: 0.58-0.86) although the composite PELD score also showed some predictive ability (AUC 0.67; CI: 0.52-0.81).

CONCLUSION: In this paediatric AIH cohort, higher PELD at presentation predicted change to second-line therapy within the first 2 years of follow-up. INR appeared to be the main contributor to that association.