AUTHOR=Tao Xuemei , Chen Lin , Zhao Youfei , Liu Yonggang , Shi Ruifang , Jiang Bei , Mi Yuqiang , Xu Liang 

TITLE=A Novel Noninvasive Diagnostic Model of HBV-Related Inflammation in Chronic Hepatitis B Virus Infection Patients With Concurrent Nonalcoholic Fatty Liver Disease

JOURNAL=Frontiers in Medicine

VOLUME=Volume 9 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2022.862879

DOI=10.3389/fmed.2022.862879

ISSN=2296-858X

ABSTRACT=BACKGROUND AND AIMS: Chronic Hepatitis B Virus Infection (CBI) patient with concurrent non-alcoholic fatty liver disease (NAFLD) is becoming increasingly common in clinical practice, and it is quite important to identify the etiology when hepatitis occurs. A non-invasive diagnostic model was constructed to identify patients who need anti-hepatitis B virus (HBV) therapies (histologic activity index, HAI≥4) in CBI Patients with concurrent NAFLD by analyzing clinical routine parameters.
APPROACH AND RESULTS: 303 out of 502 CBI patients with concurrent NAFLD proven by liver biopsy from January 2017 to December 2020 in Tianjin Second People’s Hospital were enrolled, and they were divided into HBV-related inflammation group (HAI≥4, 176 cases) and non-HBV-related inflammation group (HAI<4, 127 cases) according to hepatic pathology. Univariate analysis and multivariate logistic regression analysis were performed on the two groups of patients, and then the HBV-related inflammation model (HBV-I) of CBI patients with concurrent NAFLD was constructed. The areas under ROCs (AUROCs) were used to evaluate the parameters of the regression formula. Another 115 CBI patients with concurrent NAFLD proven by liver biopsy from January 2021 to January 2022 were enrolled as the validation group. There were some statistical differences in demographic data, biochemical indicators, immune function, thyroid function, virology indicator, blood routine indicators between the two groups(P<0.05), and liver stiffness measurement (LSM) in the HBV-related inflammation group were significantly higher than those in the non-HBV-related inflammation group (P<0.05), While controlled attenuation parameter (CAP) were lower than those in the non-HBV-related inflammation group (P<0.05); (2) We developed a novel model by logistic regression analysis: HBV-I = -0.020×CAP+0.424×LSM+0.376×lg(HBV DNA)+0.049×AST, and the accuracy rate was 82.5%. The AUROC is 0.907, the cut-off value is 0.671, the sensitivity is 89.30%, the specificity is 77.80%, the positive predictive value is 90.34%, and the negative predictive value is 81.89%; (3) The AUROC of HBV-I in the validation group was 0.871, and the overall accuracy rate is 86.96%.
CONCLUSIONS: Our novel model HBV-I [combining CAP, LSM, lg(HBV DNA) and AST] shows promising utility for predicting HBV-related inflammation in CBI patients with concurrent NAFLD with high sensitivity, accuracy and repeatability, which may contribute to clinical application.