We read with interest the article on a large case-control study that a hepatitis B surface antigen (HBsAg) level <200 IU/mL is predictive of HBsAg seroclearance within 3 years.1 Their results confirmed our earlier observation that serum HBsAg level ≤200 IU/mL has a negative predictive value (NPV) of 100% and 92% for HBsAg seroclearance at 1 and 3 years, respectively, with a positive predictive value (PPV) of 97% and 100% if combined with a ≥1 log10 IU/mL reduction in the preceding 2 years.2 Both studies have shown that HBsAg level <200 IU/mL is an optimal level for the prediction of HBsAg seroclearance at 1 and 3 years. Interestingly, Seto et al.1 further showed that a 0.5 log reduction in HBsAg during the next year in those with serum HBsAg >200 IU/mL may predict HBsAg seroclearance in 3 years with a sensitivity of 74% and a specificity of 89.4%.1 Reanalyzing our data also showed that an HBsAg decline of 1 log10 IU/mL in the following 2 years (equivalent to 0.5 log/year) had an NPV of 98% and a PPV of 67% for HBsAg seroclearance. Using the receiver operating characteristic and the area under curve (0.966; 95% confidence interval [CI] 0.915-1.000; P < 0.001), 2-year HBsAg decline of 1 log (0.5 log/year) is a good predictor for HBsAg seroclearance in those with HBsAg >200 IU/mL.
Prediction of HBsAg seroclearance using HBsAg levels has attracted much attention recently. Two Asian studies used an HBsAg level <100 IU/mL as a remote predictor of HBsAg seroclearance within 6 to 10 years.3, 4 Obviously, prediction of spontaneous HBsAg seroclearance within a much shorter period of 1-3 years, using an HBsAg level of 200 IU/mL,1, 2 is more useful in daily clinical practice.