Exploring new noninvasive parameters to predict oesophageal varices in patients with NAFLD-associated compensated liver cirrhosis.

Introduction: Nonalcoholic fatty liver disease is a leading cause of liver cirrhosis and portal hypertension, which can be complicated by oesophageal varices and variceal bleeding. Screening for oesophageal varices is essential for initiating bleeding prophylaxis. Several noninvasive parameters for predicting oesophageal varices have been suggested (e.g., Baveno VI, expanded Baveno VI, and platelet count/spleen diameter ratio), although with variable efficacy in patients with nonalcoholic fatty liver disease.

Aim: This study aimed to compare the non-invasive predictors of oesophageal varices.

Material and methods: We retrospectively analyzed the medical records of patients with nonalcoholic fatty liver disease-related compensated liver disease who underwent screening endoscopy at a tertiary care unit in United Arab Emirates. The accuracy of the established (Baveno VI and expanded Baveno VI) and newly devised (platelet count)/(spleen diameter) + liver stiffness × (40 - albumin) noninvasive parameters in predicting oesophageal varices and the presence of large oesophageal varices was assessed in our population.

Results: We found that the (platelet count)/(spleen diameter) + liver stiffness × (40 - albumin) formula had a higher accuracy than both Baveno VI (p = 0.030) and expanded Baveno VI criteria (p = 0.050) in predicting the presence of oesophageal varices. The (platelet count)/(spleen diameter) + liver stiffness × (40 - albumin) formula was associated with a higher number of spared endoscopies than Baveno VI (n = 16, 21.9%) and expanded Baveno VI (n = 9, 12.3%) criteria.

Conclusions: The new formula could provide superior predictive value than the currently practiced noninvasive predictors of oesophageal varices. However, large-scale studies are warranted to confirm its predictive performance in patients with nonalcoholic fatty liver and other etiologies of chronic liver disease.