Background: The population impact of universal test and treat (UTT) on HIV transmission has not yet been evaluated.
Methods: A cluster-randomized trial was implemented in 2x11 rural communities in KwaZulu-Natal, South Africa. All residents ≥16 years were offered rapid HIV testing and provided dried blood spots (DBS) during 6-monthly home-based survey rounds. HIV-positive participants were referred to cluster-based trial clinics to receive ART regardless of CD4 count (intervention arm) or according to national guidelines (control arm). Standard of care ART was also available in the Department of Health clinics. HIV incidence was estimated on repeat DBS using cluster-adjusted Poisson regression.
Results: Between 03/2012 and 04/2016, 13,239 and 14,916 individuals (63% women, median age 30 years) were registered in the intervention and control arms. Contact frequency per round among registered individuals ranged from 64% to 83%, HIV ascertainment from 74% to 85%. Baseline HIV prevalence was 29.4%(95%CI 28.8-30.0), with 7,578 individuals identified as HIV-positive. 1,513(36%) of 4,172 HIV-positive individuals not previously in care linked to trial clinics within 6 months of referral. ART initiation in trial clinics at 3 months was 90.9%(576/634) and 52.3%(332/635) in the intervention and control arms; viral suppression (< 400 copies/mL) 12 months after ART initiation was 94.9%(300/316) and 94.2%(194/206), respectively. Overall ART coverage at entry was 31% and 36% in the intervention and control arms, reaching 41% in both arms by closing date. Repeat DBS tests were available for 13,693 individuals HIV-negative at baseline, yielding 461 seroconversions in 20,833 person-years (PY). HIV incidence was 2.16 per 100 PY (1.88-2.45) in the intervention arm and 2.26 (1.98-2.54) in the control arm (adjusted relative risk: 0.95 [0.82-1.10]). Severe adverse events rates were 3.4%(45/1,323) and 3.5%(57/1,604) in the intervention and control arms. Follow-up will be completed by 06/2016.
Conclusions: Our trial shows high acceptance of home-based HIV testing and high levels of viral suppression among individuals on ART. However overall linkage to care remains poor. No reduction in HIV incidence was demonstrated. Several factors are being investigated, including determinants of poor linkage, change in national ART guidelines, migration and geography of sexual networks.
(Funded by ANRS, GiZ and 3ie; Clinical Trials registration NCT00332878).