Background: Antiretroviral therapy (ART) regimens without tenofovir have risks related to hepatitis B virus (HBV). As these regimens are being used more frequently, clinicians and policy makers would benefit from an accurate estimate of the risk of HBV outcomes following tenofovir cessation.
Methods: We conducted a systematic review of HBV reactivation and infection in people with HIV who discontinue tenofovir. We searched Medline, Embase, Google Scholar, and the Cochrane database (1 January 2006-15 February 2026) for studies related to tenofovir cessation in people with HIV and (1) active HBV: positive HBV surface antigen (HBsAg); (2) resolved HBV: positive HBV core antibody (HBcAb) without HBsAg; and (3) susceptibility to HBV: negative HBsAg, HBcAb, and HBV surface antibody. We used a random effects model to calculate the pooled incidence rate of HBV reactivation (in resolved HBV) and new HBV infection (in those susceptible to HBV) after tenofovir cessation. We used multivariable meta-regression to compare pooled outcomes across subgroups. We used I2 to assess inter-study heterogeneity, and we tested for publication bias using Egger's test. This review is registered with Prospero (#2025-CRD420251080460).
Findings: Of 144 abstracts and articles, 53 underwent full-text review, and 22 unique observational studies were included. In eight studies of active HBV infection, only descriptive data could be retrieved without a meta-analysis: HBV DNA increase or hepatitis flares commonly occurred in most studies after tenofovir cessation or interruption. In six observational studies of resolved HBV infection with active surveillance of HBV serological markers after tenofovir cessation, pooled incidence of HBV reactivation was 3.79/100 person-years (95% CI 1.98-7.23; I2 = 57.6%; moderate certainty). This was higher than the pooled incidence in studies without active monitoring for HBV reactivation (0.35/100 person-years, 95% CI 0.12-1.03; p < 0.001). Six observational studies reported new HBV infections in susceptible people following transition to tenofovir-sparing ART (pooled incidence = 1.63/100 person-years, 95%CI = 0.27-9.92; I2 = 78.6%; very low certainty). Egger's test showed no statistically significant evidence of publication bias (p = 0.216).
Interpretation: In people with HIV, tenofovir-sparing ART is associated with risks related to HBV reactivation and infection. Higher quality prospective studies are needed in people with HIV who stop tenofovir.
Funding: National Institutes of Health and Harvard University Center for AIDS Research.
Keywords: HIV; Hepatitis B; Long-acting injectable ART; Tenofovir.
