Background: In West Africa, 10% of HIV-infected adults are co-infected with hepatitis B virus (HBV). The impact of HBV co-infection on mortality is unknown. We analyzed the association between HBV replication and mortality during long-term follow-up in the Temprano trial.
Methods: Between March 2008 and July 2012, HIV-1 infected adults with CD4< 800/mm3 and no criteria for starting ART according to most recent WHO guidelines were randomized to deferred ART or early ART , and to receive or not Isoniazid Preventive Therapy. At inclusion, hepatitis B surface antigen (HBsAg) was tested for all those included and plasma HBV DNA was quantified for HBsAg-positive samples using an in-house PCR technique (detection limit=2 copies/ml). All first-line ART regimens contained Tenofovir/Emtricitabine. 30-month mortality and severe morbidity were previously described in the paper reporting the final results of the trial. After their 30-month visits, all participants continued to be followed up until the last participants reached 30 months. Here we present mortality during and after the trial in all participants in Temprano. We used Cox regression to assess the risk of mortality in patients with high levels of HBV DNA at baseline, compared to other patients, adjusting for early/deferred ART and IPT.
Results: Of the 2056 participants in Temprano (78% women, median age 35 years, median CD4 count 465/mm3), 193 (9%) were HBsAg-positive. Of the 173 co-infected participants with available plasma HBV DNA, 119 (69%) had detectable HBV DNA (median 3880 copies/ml, IQR 660-2120000), including 73 (42%) with HBV DNA >2000 copies/ml. Median follow-up time was 58 months (IQR 40-69), totaling 9322 person-years (PY). During follow-up, 1814 (89%) patients started ART, 85 (4%) died, and 187 (9%) were lost to follow-up. The incidence of mortality was 0.9 /100 PY overall, 2.1/100 PY in HBsAg-positive patients with baseline HBV DNA >2000 copies/ml and 0.9/100 PY in other patients (p=0.02). In multivariate analysis, the risk of mortality was independently higher in patients with HBV DNA >2000 copies/ml (adjusted Hazard Ratio 2.23, 95%CI 1.02-4.85, p=0.04).
Conclusions: In these West African HIV-infected adults with high baseline CD4 count, mortality was 2.2 times higher in patients with high levels of HBV replication.