APACHE III score is superior to King's College Hospital criteria, MELD score and APACHE II score to predict outcomes after liver transplantation for acute liver failure

Transplant Proc. Jul-Aug 2013;45(6):2295-301. doi: 10.1016/j.transproceed.2013.02.125.

Abstract

Objectives: The Model for End-Stage Liver Disease score and King's College Hospital (KCH) criteria are accepted prognostic models acute liver failure (ALF), while the use of (APACHE) scores predict to outcomes of emergency liver transplantation is rare.

Materials and methods: The present study included 87 patients with ALF who underwent liver transplantation. We calculated (KCH) criteria, as well as MELD, APACHE II, and APACHE III scores at the listing date for comparison with 3-month outcomes.

Results: According to the Youden-Index, the best cut-off value for the APACHE II score was 8.5 with 100% sensitivity, 49% specificity, 24% positive predictive value (PPV), and 100% negative predictive value (NPV). Patients with <8.5 points had a significantly higher survival rate (P < .05). The proposed APACHE III cut-off was 80. The APACHE III score demonstrated the highest specificity and PPV (90% specificity, 50% PPV). The NPV was 92%. With a 90-point threshold the specificity increased to 98% with 75% PPV and 89% NPV. Only 1 of 4 patients with a score >90 survived transplantation (P = .001). MELD score and KCH criteria were not significant (P > .05). According to the Hosmer-Lemeshow test, only the APACHE III score adequately describe the data.

Conclusions: The APACHE III score was superior to KCH criteria, MELD score, and APACHE II score to predict outcomes after transplantation for ALF. It is a valuable parameter for pretransplantation patient selection.

Publication types

  • Comparative Study

MeSH terms

  • APACHE*
  • Adolescent
  • Adult
  • Chi-Square Distribution
  • Child
  • Decision Support Techniques*
  • Female
  • Humans
  • Kaplan-Meier Estimate
  • Liver Failure, Acute / diagnosis*
  • Liver Failure, Acute / surgery*
  • Liver Transplantation*
  • Logistic Models
  • Male
  • Middle Aged
  • Multivariate Analysis
  • Patient Selection
  • Predictive Value of Tests
  • Retrospective Studies
  • Risk Factors
  • Severity of Illness Index
  • Time Factors
  • Treatment Outcome
  • Young Adult