Abstract
Introduction. Upper digestive tract endoscopy remains the gold-standard for detecting esophageal or gastric varices and assessment of bleeding risk, but this method is invasive. The aim of the study was to identify non-invasive factors that could be incorporated into an algorithm for estimating the risk of variceal bleeding.
Methods. A prospective study was performed on 130 cirrhotic patients. Tests were performed on all patients which included liver enzymes, complete blood count and coagulation parameters, abdominal ultrasound, elastography of both the liver and the spleen. Upper gastrointestinal endoscopy was performed in all patients included in the study and the results were classified, in accordance with Baveno VI into 2 outcome groups: Group 1 – patients with low bleeding risk and Group 2 – patients with varices needing treatment.
Results. The study lot (130 patients) was divided into: Group I (low bleeding risk – 102 patients), and Group II (high bleeding risk – 28 patients). Parameters found to have significant differences in univariate analysis were transaminases, platelet count, spleen size, INR, portal vein diameter and both liver and spleen elastography. Calculating AUROC for each parameter identifies spleen elastography as having the best result, followed by INR, AST and platelet count. Liver elastography had the worst AUROC. Independent variables identified by logistic regression included spleen elastography, INR, platelet count, spleen diameter, ALT, age, and gender.
Conclusions. Spleen stiffness is the best single parameter predicting the presence of high-risk esophageal varices.