BACKGROUND & AIMS: Loss of hepatitis B surface antigen (HBsAg) usually indicates that hepatitis B virus (HBV) infection has been cured. However, little is known about factors predicting HBsAg loss in patients who spontaneously clear hepatitis B e antigen (HBeAg). METHODS: We studied 390 Taiwanese HBeAg-positive patients with chronic hepatitis who had spontaneously cleared HBeAg (seroconversion) during follow-up. Serum levels of HBV DNA and HBsAg were determined 1 year after HBeAg seroconversion, and their relationships with subsequent HBsAg loss were investigated. RESULTS: In a mean follow-up of 7.4 years, the average annual rate of HBsAg loss was 0.62%. Serum levels of HBsAg and HBV DNA were inversely associated with HBsAg loss in a dose-response manner. Compared with patients with HBsAg levels ≥1000 IU/mL, the HBsAg loss rate was higher for those with HBsAg levels of 100 to 999 and <100 IU/mL, with hazard ratios of 4.4 (95% confidence interval, 1.1-17.0) and 24.3 (8.7-67.5), respectively. Among those who underwent HBsAg loss within 6 years of follow-up, serum HBsAg levels were a better predictor than HBV DNA levels by receiver operating characteristic curve analysis (area under the receiver operating characteristic curve, 0.90 vs 0.69; P = .012); an HBsAg level <100 IU/mL predicted HBsAg loss within 6 years with a diagnostic accuracy of 91.5%, sensitivity of 83.3%, specificity of 92.1%, positive predictive value of 45.5%, and negative predictive value of 98.6% in patients with an HBV DNA level <200 IU/mL. CONCLUSIONS: Low serum levels of HBsAg, alone or in combination with HBV DNA levels, 1 year after HBeAg seroconversion can predict HBsAg loss in patients with HBV genotype B or C infection.
BACKGROUND & AIMS: Loss of hepatitis B surface antigen (HBsAg) usually indicates that hepatitis B virus (HBV) infection has been cured. However, little is known about factors predicting HBsAg loss in patients who spontaneously clear hepatitis B e antigen (HBeAg). METHODS: We studied 390 Taiwanese HBeAg-positive patients with chronic hepatitis who had spontaneously cleared HBeAg (seroconversion) during follow-up. Serum levels of HBV DNA and HBsAg were determined 1 year after HBeAg seroconversion, and their relationships with subsequent HBsAg loss were investigated. RESULTS: In a mean follow-up of 7.4 years, the average annual rate of HBsAg loss was 0.62%. Serum levels of HBsAg and HBV DNA were inversely associated with HBsAg loss in a dose-response manner. Compared with patients with HBsAg levels ≥1000 IU/mL, the HBsAg loss rate was higher for those with HBsAg levels of 100 to 999 and <100 IU/mL, with hazard ratios of 4.4 (95% confidence interval, 1.1-17.0) and 24.3 (8.7-67.5), respectively. Among those who underwent HBsAg loss within 6 years of follow-up, serum HBsAg levels were a better predictor than HBV DNA levels by receiver operating characteristic curve analysis (area under the receiver operating characteristic curve, 0.90 vs 0.69; P = .012); an HBsAg level <100 IU/mL predicted HBsAg loss within 6 years with a diagnostic accuracy of 91.5%, sensitivity of 83.3%, specificity of 92.1%, positive predictive value of 45.5%, and negative predictive value of 98.6% in patients with an HBV DNA level <200 IU/mL. CONCLUSIONS: Low serum levels of HBsAg, alone or in combination with HBV DNA levels, 1 year after HBeAg seroconversion can predict HBsAg loss in patients with HBV genotype B or C infection.