HIV-HBV coinfection in Southern Africa and the effect of lamivudine- versus tenofovir-containing cART on HBV outcomes.

This study assessed HIV-hepatitis B virus (HBV) coinfection in southern Africa in terms of prevalence, viral characteristics, occult HBV, and the effect of lamivudine- versus tenofovir-containing first-line combination antiretroviral treatment (cART) on HBV-related outcomes. A multicenter prospective cohort of HIV-infected adults in Zambia and South Africa who initiated cART. Outcomes by month 12 on cART were immunological recovery, hepatitis B surface antigen (HBsAg) loss, viral suppression, and drug resistance. We used descriptive statistics, logistic regression, and linear mixed models. Of the 1087 participants, 92 were HBsAg seropositive, yielding a sample-weighted prevalence of 7.4% (95% confidence interval: 5.6 to 9.2), with 76% genotype HBV-A1. The estimated CD4 recovery on cART was similar between HIV monoinfection and HIV-HBV coinfection groups and between lamivudine- and tenofovir-treated participants. HBsAg loss was documented in 20% (4/20) of lamivudine-treated and 18% (3/17) of tenofovir-treated participants (P = 0.305). Viral suppression (HBV-DNA < 20 IU/mL) was achieved in 61.5% (16/26) of lamivudine-treated and 71.4% (15/21) of tenofovir-treated participants (P = 0.477). HBV pol sequencing demonstrated M204I (n = 3) and N236T (n = 1) resistance-associated mutations in 4 of 8 (50%) lamivudine-treated participants and none in tenofovir-treated participants. Occult HBV infection was present in 13.3% before cART, but by month 12, HBV-DNA was below the limit of detection (<15 IU/mL) in 90.5% (19/21) of lamivudine-treated and 100% (18/18) of tenofovir-treated participants (P = 0.179). Tenofovir-containing first-line cART is preferred for HIV-HBV coinfection in Africa because of a superior resistance profile relative to lamivudine monotherapy. Extended follow-up will be needed to determine long-term complications of occult HBV coinfection. Improved access to HBsAg screening and tenofovir is needed.