Factors that predict mortality in children with Wilson disease associated acute liver failure and comparison of Wilson disease specific prognostic indices

BACKGROUND AND AIMS

Wilson disease (WD) associated acute liver failure (ALF) affects children more than adults. The predictors of mortality and outcome in patients without encephalopathy are not clear. We investigated the ability of prognostic factors and various models including model for end-stage liver disease (MELD) to predict mortality among children with WD and ALF.

METHODS

We analyzed the admission characteristics in 61 children <18 years with WD and ALF. Factors associated with mortality on univariate Cox regression analysis were analyzed by forward stepwise Cox hazards regression. The prognostic models such as Nazer's model, revised Kings College Model, and pediatric end-stage liver disease/model for end-stage liver disease (PELD/MELD) score were compared.

RESULTS

Of the 145 children < 18 years with WD, 61 experienced ALF of whom 33 (54%) died, including 22/27 (81.5%) with encephalopathy and 11/34 (32.4%) without encephalopathy. The mean age of children with ALF was 9.7 years, 38(62.3%) were boys. Prognostic factors significant for mortality included encephalopathy, international normalized ratio, total proteins, total and direct bilirubin, alkaline phosphatase, serum creatinine, and white blood cell count. Forward stepwise Cox proportional hazards regression identified encephalopathy (hazard ratio 2.88; CI 1.1-7.4) and total bilirubin (hazard ratio 1.05; CI: 1.02-1.09) as predictors of outcome. The area under the receiver operating curve (AUC) of the Nazer index, revised King's College Criteria, and PELD/MELD were 0.74, 0.76, and 0.75, respectively.

CONCLUSIONS

Mortality in children with WD and ALF is 54% including 81.5% with encephalopathy and 32.4% without encephalopathy. The prognostic models, MELD/PELD score, Nazer index and Kings College Criteria are comparable with a AUC between 0.74-0.76.