Long-term outcomes of patients with type 1 or 2 autoimmune hepatitis presenting in childhood

Background & Aims: In children with autoimmune hepatitis, uncertainties include outcomes associated with type 2 hepatitis, the possibility of and criteria for attempting withdrawal of treatment, and long-term outcomes. We report our experience on these issues. Methods: From 1973 to 2002, 117 children with type 1 (n = 65) or type 2 (n = 52) hepatitis, excluding fulminant hepatitis, were treated, primarily with prednisone and azathioprine. Median follow-up was 20 years in survivors. Results: Normalisation of aminotransferases and prothrombin ratio were observed in 93% and 84% of children, respectively; sustained remission after treatment withdrawal was recorded in 24% of the entire population, with a median follow-up of 7 years. Sustained treatment-free remission was obtained in 11 of 24 children with follow-ups of 4-22 years based on durable normalisation of aminotransferases (without histological assessment). Gastrointestinal bleeding from varices and the emergence of extrahepatic autoimmune disorders occurred in 10 and 22 patients, respectively. Liver transplantation was performed in 23 patients at a median age of 21 years. The 30-year probabilities of overall and native liver survival were 81% and 61%, respectively. No differences were observed between type 1 and 2 hepatitis for any of the component parts of outcome. In the multivariate analysis, a persistent abnormal prothrombin ratio was associated with worse probabilities of overall and native liver survival. Conclusions: In terms of liver outcome, type 2 hepatitis is not different from type 1. Withdrawal of treatment is possible without prior liver histology. A persistent abnormal prothrombin ratio identifies patients who will require liver transplantation in adolescence or early adulthood. Impact and implications: In children with autoimmune hepatitis, there are conflicting reports on the differences in outcome between type 1 and type 2 hepatitis, and on the possibility of treatment withdrawal, before which liver histology is required; data concerning >10-year overall and native liver survival rates are limited. In this study, we found no differences in outcomes between type 1 and 2 hepatitis; a durable treatment-free state was achieved in 19% of all patients throughout childhood and early adulthood, and in 45% of children for whom treatment withdrawal was attempted without prior liver histology; prothrombin was found to be predictive of 30-year overall and native liver survival. The results allow for a less-strict approach to treatment withdrawal in children, avoiding the risks of a liver biopsy, and they provide a tool to help anticipate the need for liver transplantation before complications occur.