Transient elastography as a predictor of oesophageal varices in patients with liver cirrhosis

Abstract

One of the most serious complications of liver cirrhosis is variceal bleeding. Early recognition of the oesophageal varices is of primary importance in the prevention of variceal bleeding. Endoscopy is the only means to directly visualize varices and measure their size, as one of the most important predictor of the risk of bleeding. During the course of cirrhosis repeated oesophago-gastro-bulboscopic examinations are recommended. As these interventions are expensive and often poorly accepted by patients who may refuse further follow-up, there is a need for non-invasive methods to predict the progression of portal hypertension as well as the presence and the size of oesophageal varices. After several combinations of biological and ultrasonographical parameters proposed for the detection of advanced fibrosis, it was suggested that liver stiffness measured by transient elastography, a novel non-invasive technology may reflect not only fibrosis and portal pressure but it may even predict the presence or absence of large oesophageal varices in patients with cirrhosis. The aim of the authors was to study the diagnostic accuracy of transient elastography using FibroScan for selecting patients who are at risk of bearing large (Paquet-grade ≥ II) oesophageal varices and high risk of bleeding. The authors performed upper tract endoscopy and transient elastography in 74 patients with chronic liver disease (27 patients with chronic hepatitis and 47 patients with liver cirrhosis). The relationships between the presence of oesophageal varices (Paquet-grade 0-IV) and liver stiffness (kPa), as well as the hematological and biochemical laboratory parameters (prothrombine international normalized ratio, platelet count, aspartate aminotransferase, alanine aminotransferase, albumin, and aspartate aminotransferase/platelet ratio index) were investigated. The predictive role of liver stiffness for screening patients with varices and those who are at high risk of variceal bleeding was also analysed. Liver stiffness values significantly correlated with the grade of oesophageal varices (Paquet-grade) (r = 0.67, p<0.0001). The liver stiffness value of 19.2 kPa was highly predictive for the presence of oesophageal varices (AUROC: 0.885, 95% CI: 0.81-0.96) and for the presence of high grade varices (P≥II) (AUROC: 0.850, 95% CI: 0.754-0.94). Using the cut-off value of 19.2 kPa, the sensitivity of transient elastography was 85%, specificity was 87%, positive predictive value was 85%, negative predictive value was 87% and validity was 86% for the detection of varices. Liver stiffness values less than 19.2 kPa were highly predicitive for the absence of large (P≥II) varices (sensitivity, 95%; specificity, 70%; positive predictive value, 54%; negative predictive value, 97%). Transient elastography may help to screen patients who are at high risk of bearing large (P≥II) oesophageal varices which predict variceal bleeding and, therefore, need endoscopic screening. Lives stiffness values higher than 19.2 kPa indicate the need for oesophageal-gastro-bulboscopy, while liver stiffness values lower than 19.2 kPa make the presence of large oesophageal varices unlikely.