Association between liver function at hospital admission and mortality in pediatric patients receiving postcardiotomy venoarterial extracorporeal membrane oxygenation support.

BACKGROUND: Preoperative liver function in children with congenital heart disease is often compromised to varying degrees because of the unique pathophysiology. We aimed to investigate the relationships between liver function indicators at hospital admission and mortality in children receiving veno-arterial extracorporeal membrane oxygenation (VA-ECMO) support following cardiac surgery. METHODS: We retrospectively analysed the clinical data of pediatric patients who received postcardiotomy VA-ECMO support at Fuwai Hospital between January 2010 and June 2020. Univariable and multivariable-adjusted Cox proportional hazard models were constructed to evaluate the risk factors associated with 30-day and 180-day mortality. The cut-off values for the liver function variables measured at hospital admission were categorized into high and low groups and then compared using Kaplan-Meier survival curves and log-rank tests. RESULTS: Our study included 96 pediatric patients who received VA-ECMO support after cardiotomy. Among the patients receiving VA-ECMO, the 30-day and 180-day mortality rates were 37.5% and 52.1%, respectively. The level of aspartate aminotransferase (AST) at admission was associated with 30-day mortality (hazard ratios [HRs]=1.852, 95%CI 1.010-3.398, P=0.046). The AST and alkaline phosphatase (ALP) levels were predictors of 180-day mortality, with adjusted HRs of 1.799 (95%CI 1.074-3.014; P=0.025) and 1.384 (95%CI 1.050-1.825; P=0.021), respectively. The cut-off value for AST to predict mortality at 30 d was 77 U/L, and that for ALP to predict mortality at 180 d was 269 U/L. CONCLUSION: Liver function indicators, including AST and ALP, at hospital admission are associated with mortality risk in children with congenital heart disease receiving VA-ECMO after cardiac surgery.