A comparison of two validated scores for estimating risk of mortality of children with intestinal failure associated liver disease and those with liver disease awaiting transplantation

Background and aims: To evaluate risk of mortality in children with intestinal failure associated liver disease (IFALD) compared with other liver disease using two validated scores.

Methods: Sixty-seven children listed for transplant were studied: cholestatic liver disease (CLDn23); liver disease secondary to other processes (LDsec n11); (IFALDn22), acute liver failure (ALFn11). Paediatric Hepatology Score (PHD) score and Pediatric end-stage liver disease score (PELD) were evaluated by Receiver Operating Curves (ROC), proportional hazards regression.

Results: The highest PHD and PELD scores were found in ALF; the lowest in LDsec. Both scores correlated well in identifying waiting list (WL) mortality in patients with CLD and ALF, but not in those with IFALD where PELD scores were lower. Cox proportional hazard regression of time spent on the waiting list prior to death or transplant/delisting showed significant associations with PHD (P=0.006) and PELD (P=0.008). WL mortality was strongly predicted by disease group (6/8 deaths in IFALD). ROC analysis of all data showed that a PHD score greater than 15.5 had sensitivity of 87.5% and specificity of 81% for waiting list mortality (P<0.001); PELD greater than 8 had a sensitivity of 87.5% and specificity of 40%. Neither PHD nor PELD scores correlated with post-transplant mortality.

Conclusion: PHD and PELD scores had the same sensitivity for identifying risk of WL mortality in all patients, but PELD failed to identify the sickest children with IFALD, lowering its specificity. The PHD score has the added advantage for European centres of being in SI units, not requiring a computer application to calculate and was simpler to use at bedside.