Liver stiffness measurement, Doppler ultrasound, and serum non-invasive markers as predictors of varices and variceal bleeding in cirrhotic patients

Abstract Background and aims Variceal bleeding is a life-threatening complication of liver cirrhosis, necessitating reliable non-invasive predictors to minimize the use of the more invasive esophagogastroduodenoscopy (EGD). This study evaluates the diagnostic accuracy of liver stiffness measurement (LSM), Doppler ultrasound, and serum biomarkers (FIB-4, APRI, PC/SD ratio) for predicting varices and variceal bleeding. Patient and methods In this cross-sectional study, we enrolled 144 cirrhotic patients who were divided into 3 groups: I, no varices ( n = 37); II, non-bleeding varices ( n = 41); and III, variceal bleeding ( n = 66). All participants underwent LSM, Doppler ultrasound, and laboratory tests. The diagnostic performance of these non-invasive methods was determined using ROC curves and multivariate logistic regression. Results LSM (cutoff > 30 kPa) showed high accuracy for predicting variceal bleeding (90% sensitivity, AUC 0.91), outperforming all Doppler parameters (AUC 0.41–0.55). PC/SD ratio (cutoff ≤ 650) also demonstrated high predictive value for variceal bleeding. LSM (cutoff > 20 kPa) and PC/SD ratio (cutoff < 900) showed high accuracy for predicting the presence of varices (AUC 0.90 and 0.84, respectively). Multivariate analysis identified LSM (OR 6.350, p < 0.001), APRI score (OR 2.256, p = 0.047), and PC/SD ratio (OR 2.314, p = 0.048) as the key independent predictors for variceal bleeding. Doppler ultrasound showed low predictive value ( p > 0.05). Conclusions LSM, APRI, FIB-4, and PC/SD ratio are accurate non-invasive tools for prediction of varices and variceal bleeding, supporting their role in risk stratification of cirrhotic patients and reducing unnecessary EGD. Doppler ultrasound has limited value in this context.