Albumin-bilirubin grade versus MELD score for predicting survival after transjugular intrahepatic portosystemic shunt (TIPS) creation.

Abstract

The purpose of this study was to compare the albumin-bilirubin (ALBI) grade and model for end-stage liver disease (MELD) scores for predicting survival after transjugular intrahepatic portosystemic shunt (TIPS) creation. A retrospective study of pre-procedure ALBI and MELD scores was performed in 197 patients who underwent TIPS from 2005 to 2012. There were 140 men and 57 women, with a mean age of 56±11 (SD) (range: 19-90years). The prognostic capability of ALBI and MELD scores were evaluated using competing risks survival analysis. Discriminatory ability was compared between models using the C-index derived from cause specific Cox proportional hazards models. TIPS were created for ascites or hydrothorax (128 patients), variceal hemorrhage (61 patients), or both (8 patients). Prior to TIPS, 5 patients were ALBI grade 1, 76 were grade 2, and 116 were grade 3. The average pre-TIPS MELD score was 14. Pre-TIPS ALBI score, ALBI grade, and MELD were each significant predictors of 30-day mortality from hepatic failure and overall survival (all P<0.05). Based on the C-index, the MELD score was a better predictor of both 30-day and overall survival (C-index=0.74 and 0.63) than either ALBI score (0.70 and 0.59) or ALBI grade (0.64 and 0.56). In multivariate models, after accounting for MELD score ALBI score provided no additional short- or long-term survival information. Although ALBI score and grade were statistically significantly associated with risk of death after TIPS, MELD remains the superior predictor.