Abstract
BACKGROUND: Assessing the potential increased risk of viral rebound (VR) in migrants requires adequate control for sex and acquisition risk groups. METHODS: People living with HIV1, enrolled in the ANRS CO4-French Hospital Database on HIV, who achieved virological suppression with antiretroviral therapy (ART) initiated between 2006 and 2016 were included. We first compared the risk of VR, with loss to follow-up and death considered as competing events, across origin among the HIV acquisition groups, then across acquisition groups among the different origins, and finally across modality of a variable combining sex, acquisition group, and origin. Models were adjusted for clinical and biological confounding factors. RESULTS: We included 21 571 French natives (FRA), 10 148 migrants from sub-Saharan Africa (SSA), 1137 migrants from the non-French West Indies (NFWI), and 4205 other migrants (OTHER). The 5-year probability of VR was 19% (95% confidence interval [CI] 19-20) overall, 15% in FRA, 21% in OTHER, 26% in SSA, and 34% in NFWI (p < 0.0001). It was 14% in men who have sex with men (MSM), 23% in heterosexual men, and 23% in women (p < 0.0001). After adjustment, all acquisition groups had a higher risk of VR than MSM from FRA, with men and women from NFWI having the highest risk (adjusted hazard ratio [aHR] 2.46; 95% CI 2.12-2.86 and aHR 2.59; 95% CI 2.20-3.04, respectively). Within each acquisition group, all groups of origin had a higher risk of VR than FRA. Within each region of origin, except the NFWI, heterosexual men had a higher risk of VR than MSM. CONCLUSIONS: After accounting for sex and acquisition group, migration, especially from NFWI, remains prognostic of VR.