• Kao-Chi Chang, Chih-Yi Lee, Te-Sheng Chang, Chao-Hung Hung, Wei-Ming Chen, Mei-Yen Chen, Tung-Jung Huang, Wen-Nan Chiu, Jing-Hong Hu, Yu-Chih Lin, Wei-Cheng Huang, Nien-Tzu Hsu, and Sheng-Nan Lu.
• Division of Hepatology and Gastroenterology, Department of Internal Medicine, Chang Gung Memorial Hospital, Chiayi, Taiwan.
• J Formos Med Assoc. 2021 Feb 1; 120 (2): 847-853.

Background/PurposeLow viral load (LVL) of hepatitis B virus (HBV) is a predictor of chronic HBV infection. However, the usefulness of quantitative hepatitis B surface antigen (qHBsAg) in predicting LVL in community-based screening has not been well studied. We aimed to measure the prevalence of LVL in HBV carriers and validate the efficacy of qHBsAg in predicting LVL.MethodsThis community-based screening study was conducted in Taiwan. HBV DNA was assayed in HBsAg carriers. Participants were randomized to training and validation sets to determine the ability of qHBsAg to predict LVL. Receiver operating characteristic curves were used to identify the best cutoff values in the training set.ResultsAmong the 2919 participants, 359 (12.2%) were HBsAg carriers. There were 132 and 137 carriers in the training and validation sets, respectively. Significant correlations were found between qHBsAg and HBV DNA in both training and validation sets. Thirty and 29 participants with qHBsAg <8 IU/mL in the training and validation sets, respectively, had LVL. Using 8 IU/mL as the cutoff, negative predictive value (NPV) of qHBsAg for HBV DNA levels >2000 IU/mL was 100%. The best cutoff level of qHBsAg to predict HBV LVL was 200 IU/mL, with a sensitivity, specificity, and accuracy of 75.0%, 76.1%, and 75.8%, respectively, in the training set. The positive predictive value and NPV were 70.0% and 77.9%, respectively, in the validation set.ConclusionApproximately 60% of HBsAg carriers had HBV LVL, and qHBsAg <8 IU/mL accurately predicts LVL. This quantitative test provides additional information for community-based screening.Copyright © 2020 Formosan Medical Association. Published by Elsevier B.V. All rights reserved.

15 keywords...

hide…