P70 Evaluating strategies for patient selection in NSBB treatment for chronic advanced compensated liver disease: a comparative study of varices and non-invasive tests

<h3></h3> There is currently no consensus on the best method to select patients with chronic advanced compensated liver disease (cACLD) for treatment with non-selective beta-blockers (NSBB). Traditionally, the presence of varices was used as the primary criterion, focusing on preventing variceal hemorrhage. However, recent guidelines suggest using noninvasive surrogates to predict clinically significant portal hypertension for initiating NSBB therapy. This study aims to evaluate the effectiveness of different strategies for selecting cACLD patients for NSBB treatment. Data from patient cohorts at two large UK hospitals were analyzed. The study compared the prediction of decompensation using varices and noninvasive tests (NITs) and their impact on NSBB treatment decisions. Five predictors were evaluated: presence of varices, liver stiffness measurement (LSM) &gt;25 kPa, AASLD rule-in criteria, Score 3P, and Score 3P + LSM. Outcomes included model performance (discrimination and calibration), net benefit for treatment decisions, and potential treatment effects. The cohort included 2117 patients with a median age of 56 years, and 59% were male. The primary etiologies were MASLD, ARLD, and viral hepatitis. Patients identified to have varices had a decompensation rate of 32.7% (hazard ratio (HR) of 4.68 (95% CI: 3.41–6.42)) with an AUC of 0.595 (95% CI: 0.566–0.626) for the prediction of decompensation. However, relatively few patients underwent upper GI endoscopy in the study period. In contrast, while the incidence of decompensation was lower among those test positive with NITs, the predictive performance of NIT based approaches was superior to the identification of varices. For instance, the AASLD rule-in criteria had a decompensation rate of 18.0% among test positive patients, and an AUC of 0.669 (95% CI: 0.633–0.702). Overall, the Score 3P + LSM predictor had the best discriminative performance (AUC of 0.682) with good calibration to outcomes. Decision curve analysis indicated the greatest net benefit for Score 3P + LSM and AASLD rule-in criteria and in simulation analyses to understand the impact of treatment selection on outcomes the Score 3P + LSM and AASLD rule-in criteria gained the most life years free of decompensation with NSBB treatment. Limited data on treatment efficacy in NIT-identified persons necessitate further research. While the presence of oesophageal varices is strongly associated with future decompensation the effectiveness of this approach is reduced by low uptake of upper GI endoscopy. Consequently, the NIT strategies had greater performance and were associated with increased net benefit and predicted gains in life years free of decompensation with NSBB treatment.