Some of the patients had liver stiffness GSK458 mw measurement assessed by transient elastography in our previous study and the last visit was taken as the day of transient elastography.14 Otherwise, the last visit was taken as the last follow-up visit with serum sample available. Detailed analyses were also performed to investigate the changes of HBsAg levels at the time of HBeAg seroconversion and hepatitis flare. For Group 3 patients, the HBsAg levels immediately before and after HBeAg seroconversion were measured and compared. Hepatitis flare was defined as an abrupt elevation of ALT to > 200 IU/L or >3 times the baseline level,

whichever was higher.15 HBsAg levels were determined in the available serum samples at the visits immediately before the flare, at the flare, and immediately after the flare. HBsAg was quantified by Architect HBsAg QT (Abbott Diagnostic, Germany) according to the manufacturer instructions.16 The sensitivity of Architect assay ranged from 0.05 to 250 IU/mL. Samples with HBsAg titer higher than 250 IU/mL were diluted to 1:500 to 1:1000 to bring the reading within the range

of the calibration curve. HBV DNA was quantified by TaqMan real-time polymerase chain reaction assay as described.17 This assay was standardized by serial dilution of EUROHEP genotype D HBV standard, which contained 2.7 × 109 viral copies per mL and was validated by the World Health Organization HBV standard. The range of HBV DNA detection was from 102 to 109 copies/mL with correlation coefficient of the standard curve routinely greater than 0.990. For samples with HBV DNA > 109 copies/mL, BVD-523 clinical trial HBV DNA assay would be repeated after dilution to 1:100. In this assay, 4.86 copies/mL equaled 1 IU/mL. HBV genotyping was determined by restriction fragment length polymorphism and was confirmed by direct sequencing in case of doubt in the residual serum sample at initial visit, as described.18 Statistical analysis was performed by SPSS (version 15.0; SPSS, Inc., Chicago, IL). Continuous variables were expressed as mean ± standard deviation Adenosine or median (range) as appropriate. HBV DNA (IU/mL) and HBsAg

(IU/mL) were logarithmically transformed for analysis. For patients with undetectable HBV DNA and negative HBsAg, the results were taken as the lower limit of detection (20.6 IU/mL for HBV DNA and 0.05 IU/mL for HBsAg) for calculation. Ratio of HBsAg (log IU/mL) to HBV DNA (log IU/mL) was determined to reflect the proportion of subviral particles to virions. Continuous variables including HBV DNA and HBsAg were compared by Student t test or Mann-Whitney U test as appropriate. Wilcoxon signed rank test was used to compare the reduction in log HBV DNA and log HBsAg levels during the longitudinal follow-up. The annual decline of HBsAg was computed by dividing the HBsAg decline from the first to the last follow-up visit by the total duration of follow-up.