Abstract
The clinical utility of quantifying hepatitis B surface antigen (qHBsAg) levels in African subjects with chronic hepatitis B virus (HBV) infection has been poorly documented. From a multicentre cohort of 944 HBV-infected African patients, we aimed to assess whether qHBsAg alone can accurately identify i) those in a HBeAg-negative chronic HBV infection phase at low risk of liver disease progression and ii) those in need of antiviral therapy according to the 2017 EASL guidelines. We analysed 770 HBV mono-infected treatment-naïve patients, mainly males (61%) from West Africa (92%), median age 35 years (IQR: 30-44), median HBV DNA: 95.6 IU/ml (10.0-1,300.0), median qHBsAg 5,498 IU/ml (1,171-13,000) and HBeAg-pos 38 (5%). A total of 464/770 (60.2%) patients were classified as HBeAg-negative chronic infection (median age 36 years (31-46), median ALT 23 IU/l (18-28), median HBV-DNA 33.5 IU/ml (3.8-154.1), median LSM 4.8 kPa (4.1-5.8)) and qHBsAg levels had poor accuracy to identify these subjects with an AUROC at 0.58 (95%CI: 0.54-0.62), sensitivity 55.0% and specificity 55.6%; 118/770 (15.3%) patients were eligible for treatment according to the 2017 EASL criteria. qHBsAg correlated poorly with HBV DNA and had poor accuracy to select patients for antiviral therapy with an AUROC at 0.54 (0.49-0.60), sensitivity 46.6% and specificity 46.9%. In African treatment-naïve HBV-infected subjects, the clinical utility of qHBsAg to identify subjects in HBeAg-negative infection phase or subjects eligible for antiviral therapy seems futile. Whether qHBsAg levels can be used as a predictor of long-term liver complications in Africa needs to be further investigated.