Non-Invasive Predictors of Moderate–Large Esophageal Varices in Compensated Cirrhotic Patients

Aim: to identify a non-invasive and low-cost tool that could predict the presence of moderate-to-large EV in compensated cirrhotic patients to lower the number of unnecessary endoscopies. Materials and Methods: A cross-sectional study was conducted in newly diagnosed patients with non-Child-Pugh C liver cirrhosis, without a history of variceal bleeding. The presence and degree of esophageal varices were determined during endoscopy. Univariate and multivariate logistic regression were used to identify independent predictors for the presence of esophageal varices. Results: Of the 180 patients enrolled in this study, EV was found in 99 patients (55%). Moderate-large varices were detected in 23.3% and 51.7% of patients with CP-A and CP-B class, respectively. In univariate analysis, CP class B, low platelet count (≤100 000/μL), increased bipolar diameter (≥135 mm), and ratio of platelet count to bipolar spleen diameter (≤847) were associated with the presence of moderate-to-large EV. In multivariate analysis, CP class and ratio of platelet count to bipolar spleen diameter were associated with the presence of moderate-to-large EV. The AUC, sensitivity, and specificity of combined predictors were 0.76, 88% and 55.4%, respectively. Conclusion: Platelet count/bipolar spleen diameter ratio value of ≤847 could be used as a non-invasive predictor of the presence of moderate-to-large EV in compensated cirrhotic patients. Future validation in a larger cohort is needed before applying this parameter in clinical practice.