Diagnostic model of esophageal varices in alcoholic liver disease.
Esophageal varices may cause life-threatening bleeding with attendant high hospital cost. Since effective preventive modalities for variceal hemorrhage have been established, early detection of esophageal varices is critical for prevention of bleeding. Currently, endoscopic screening is widely recommended to patients who have the diagnosis of cirrhosis. However, the diagnosis of cirrhosis relies on histological evaluations, which is costly and invasive, and endoscopic screening also burdens medical resources. Recent cost-effectiveness studies suggested that empiric prophylaxis with beta-blockers may be viable in comparison with endoscopic screening in patients with cirrhosis. However, this issue need to be also evaluated taking account of societal and patient perspectives and is not yet decided. An accurate non-invasive diagnostic model to predict the presence of esophageal varices may reduce unnecessary endoscopic procedures and prophylactic medication and improve cost-benefit of these approaches. Use of an accurate serum marker for severe hepatic fibrosis may also improve accuracy of non-invasive diagnostic models. Hyaluronic acid, a serum marker for severe hepatic fibrosis, has been reported to have a high diagnostic performance in assessing the severity of hepatic fibrosis in patients with alcoholic liver disease. In this issue, a non-invasive diagnostic model including hyaluronic acid was shown to have excellent performance in excluding the presence of medium to large esophageal varices in severe alcohol abusers. Based on current evidence, the strategy of using a non-invasive diagnostic model together with a serum marker for severe hepatic fibrosis may improve cost-benefit in the prevention of variceal hemorrhage among patients with alcoholic liver disease. The independent verification of such diagnostic models is needed.
Suzuki, A; Mendes, F; Lindor, K
Volume / Issue
Start / End Page
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)