Application of Noninvasive Tools to Decide the Need for Beta-Blockers for Variceal Bleeding Prophylaxis in Compensated Advanced Liver Disease: A Decision Curve Analysis

Abstract

Noninvasive tools (NITs) reliably categorise patients with compensated advanced chronic liver disease (cACLD) into high-risk and low-risk group for harbouring varices needing treatment. Here, we assess the ability of these NITs to predict the need for nonselective beta-blockers at baseline based on risk of variceal bleeding (VB) on follow-up. This was a retrospective multicentre analysis of patients with cACLD categorised at baseline into different risk groups by NITs (Baveno-VI, expanded Baveno-VI, platelet-albumin, platelet-model for end-stage liver disease (MELD) and anticipate study platelet criteria) and by endoscopy (high risk vs low risk/no varices). VB event rates on follow-up were estimated in different risk strata. Decision curve analysis (DCA) was used to estimate the benefit of administering nonselective beta-blockers (NSBB) using NITs over endoscopic classification at different threshold probabilities of VB event rates and estimating the number needed to treat (NNT) to identify one additional bleeder over endoscopy. A total of 1284 patients (mean age: 44.7±13.5 years, 72.4% males) of hepatitis B (29.2%), nonalcoholic fatty liver disease (24.9%), hepatitis C (20.1%), and alcohol (17.5%)-related cACLD were included with 323 (25.2%) having high-risk varices. Ninety-eight (7.6%) patients developed VB over a median follow-up of 20 (9-35) months. The 1-year and 3-year rate of VB with all NITs was 5.7-7.4% and 13.2-16.4% among high-risk and 0-2.3% and 0-5% among low-risk subgroups, respectively (P<0.001) in both viral and nonviral aetiologies. Among patients classified as low risk on Baveno-VI criteria, none developed VB on follow-up. At thresholds of <3% event rate of VB, Baveno-VI (NNT-176), platelet-albumin (NNT-576) and anticipate platelet (NNT-233) criteria were superior, whereas endoscopic stratification was superior above this event rate on DCA. The use of both elastography and blood-based NITs at baseline can accurately identify the need for NSBB for VB prophylaxis in patients of cACLD on follow-up.