Reply to Letter to the Editor: Regarding the “Hepatitis B surface antigen glycan isomer is a predictor of the development of hepatocellular carcinoma during nucleoside/nucleotide analog therapy”

We appreciate Huang et al. for their interest in our recent publication.1 In this study, we measured serum hepatitis B surface antigen glycan isomer (HBsAgGi) levels at baseline and 48 weeks after nucleoside/nucleotide analog (NA) therapy in chronic hepatitis B virus (HBV) genotype C-infected patients. Our findings suggest that the on-treatment HBsAgGi/HBsAg ratio is a valuable predictor of hepatocellular carcinoma (HCC) development during NA therapy. A recombinant monoclonal antibody against the HBsAgGi has been developed using the O-glycosylated PreS2 peptide in M-HBsAg of HBV genotype C.2, 3 HBsAgGi specifically recognizes O-glycosylated Pre-S2 on M-HBsAg, which is predominantly found on infectious particles, whereas conventional HBsAg recognizes all HBV particles, including non-infectious ones. Therefore, HBsAgGi, as well as HBV DNA, HBsAg, and immunoassay for total antigen including complex via pretreatment for hepatitis B core-related antigen, may be a potential biomarker for predicting therapeutic response to NA or pegylated interferon-α therapy and HCC development. Okumura et al. demonstrated a significantly higher HBsAgGi/HBsAg ratio in patients with an HCC history compared with those without.4 Furthermore, Inoue et al. emphasized the clinical utility of HBsAgGi for predicting HBsAg seroclearance.5 The letter of Huang et al. pointed out statistical issues that could affect the predictive results of this study. According to standard logistic regression principles, there should be at least 10 outcome events per predictor variable. Considering that only 10 patients developed HCC during NA therapy in this study, a multivariate analysis should be performed using one variable. In contrast, there were 20 variables in Table 2, which were analyzed to predict the HCC development during NA therapy in HBV genotype C-infected patients. We analyzed the prediction of HCC development during NA therapy separately at baseline and 48 weeks after NA therapy. Therefore, this study did not perform a multivariate analysis using all 20 variables (shown in Table 2 of that paper). In short, we reported that a combination of on-treatment HBsAgGi and HBsAg predicted the development of HCC during NA therapy using only variables at 48 weeks after NA therapy. As they pointed out, we agree that it should be analyzed using one or two variables, because 10 patients developed HCC. As described in the statistical analysis section, a multivariate analysis was performed using two variables (HBsAg and HBsAgGi/HBsAg ratio at 48 weeks after NA therapy) that were selected based on their significance in univariate analysis among the five variables (HBsAg, immunoassay for total antigen including complex via pretreatment for hepatitis B core-related antigen, HBV DNA, HBsAgGi, and HBsAgGi/HBsAg ratio) assessed at 48 weeks after NA therapy, and showed that the HBsAgGi/HBsAg ratio at 48 weeks after NA therapy was independently and significantly associated with the development of HCC. In conclusion, the number of variables used in this analysis is appropriate. Considering the relatively small sample size and limited number of HCC cases in our study, the predictive power of the HBsAgGi/HBsAg ratio requires validation in larger cohorts. Further large-scale studies are needed to confirm its utility as a predictor of HCC development during NA therapy. The authors declare that they have no financial conflicts of interest.