Abstract
BACKGROUND AND AIM: A low hepatitis B surface antigen (HBsAg) level is reported to be predictive of future HBsAg seroclearance. A hospital-based cohort study was conducted to clarify the clinical features of patients with low HBsAg levels and to demonstrate the usefulness of low HBsAg levels for predicting HBsAg seroclearance. METHODS: A total of 1459 patients with chronic hepatitis B were included in the study. Of these, 587 had repeated measurements for HBsAg levels and two or more records of HBsAg-positive results. HBsAg levels were measured with a commercially available HBsAg assay. Based on a cut-off index (COI) of 2000, a high HBsAg level was defined as HBsAg ≥2000 COI, and a low HBsAg level was defined as HBsAg <2000 COI. RESULTS: The proportion of patients with low HBsAg levels at baseline tended to increase with age. Patients with low HBsAg levels at baseline had significantly older age, lower transaminase levels, and lower hepatitis B virus (HBV) DNA levels than those with high HBsAg levels. The annual HBsAg seroclearance rate was 1.30%/year. The cumulative incidences of HBsAg seroclearance differed significantly by HBsAg level at baseline (<2000 vs ≥2000 COI), age (≥50 vs <50 years), and HBV DNA level (<4.0 vs ≥4.0 log copies/mL). Cox proportional hazards regression analyses showed that low HBsAg level (<2000 COI) and low HBV DNA level (<4.0 log copies/mL) were significantly associated with HBsAg seroclearance. CONCLUSION: Aging was one of the factors affecting HBsAg level. HBsAg seroclearance was significantly associated with low HBsAg level and low HBV DNA level at baseline.