The ability of Hepascore to predict liver fibrosis in chronic liver disease: a meta-analysis.

Abstract

Hepascore is a serum model that was developed to assess the severity of liver fibrosis. It has been well validated in common causes of chronic liver disease. This study performed a meta-analysis to evaluate the pooled diagnostic performance of Hepascore and to compare it for different aetiologies of chronic liver disease. Two reviewers searched electronic databases from October 2005 to September 2015 for studies that evaluated the diagnostic performance of Hepascore for liver fibrosis in chronic liver disease. 21 studies were included. The AUROC was adjusted according to the distribution of fibrosis stages. The mean adjusted AUROC was 0.83 (95% CI, 0.81-0.85) for significant fibrosis, 0.89 (95% CI, 0.85-0.92) for advance fibrosis and 0.93 (95% CI, 0.91-0.95) for cirrhosis. A cut point of 0.50-0.55 achieved a summary sensitivity of 70% and a summary specificity of 79% to predict significant fibrosis. A cut point of 0.50-0.61 had a summary sensitivity of 81% and a summary specificity of 74% to predict advanced fibrosis. A cut point of 0.80-0.84 had a summary sensitivity of 72% and a summary specificity of 0.88% to predict cirrhosis. The accuracy of Hepascore was similar among all disease aetiologies for the prediction of cirrhosis. However, Hepascore had better diagnostic ability for significant and advanced fibrosis in chronic hepatitis C, chronic hepatitis B and alcoholic liver disease than for non-alcoholic fatty liver disease and HIV co-infected viral hepatitis. Hepascore is a clinically useful measure of liver fibrosis in patients with common causes of chronic liver disease.