Kinetics of hepatitis B surface and envelope antigen and prediction of treatment response to tenofovir in antiretroviral-experienced HIV–hepatitis B virus-infected patients

Abstract

Hepatitis B surface (HBsAg) and envelope (HBeAg) antigen loss are the primary goals of treating chronic hepatitis B virus (HBV). Although their quantification is useful for other antivirals, such has not been the case with tenofovir disoproxil fumarate (TDF), particularly in HIV infection. Prospective, multicenter, cohort study in 143 antiretroviral-experienced HIV-HBV-co-infected patients initiating TDF. HBsAg (IU/ml) and HBeAg levels (S/CO) were measured every 6 months. HBsAg and HBeAg decline (Δ) were assessed by mixed-effect linear models. Quantification criteria were used to assess predictability of antigen loss with time-dependent receiver operating characteristic curves. After a median follow-up of 30.3 months, cumulative incidence rate of HBsAg loss was 4.0% (n = 4) in the entire study population and HBeAg loss was 21.0% (n = 17) in the 96 HBeAg-positive patients. ΔHBsAg was steady during follow-up (HBeAg-positive: -0.027; HBeAg-negative: -0.017 log(10) IU/ml per month), whereas ΔHBeAg ratio was strongly biphasic (-27.1 S/CO per month before and -6.5 S/CO per month after 18 months). Baseline HBeAg and ΔHBeAg were significantly different in patients harboring precore mutations (P < 0.01), whereas both ΔHBsAg and ΔHBeAg were significantly slower among HBeAg-positive patients with CD4(+) T-cell count less than 350 cells/μl (P < 0.05). HBeAg-ratio of 10 S/CO or less at 12 months of therapy was the optimal marker of HBeAg loss, with high sensitivity (0.82) and specificity (0.84) at 36 months. In patients with HBsAg loss, three of four (75.0%) patients had a baseline level of HBsAg of 400 IU/ml or less. During TDF treatment, HIV-induced immunosuppression and HBV genetic variability are associated with differences in HBsAg and HBeAg decline among antiretroviral-experienced, co-infected patients. Considering the decline of HBsAg level is slow, further evaluation is needed to determine its role as a marker of therapeutic efficacy.