Background: HIV infections are ongoing globally despite efficacious biomedical prevention options. We sought to determine whether an HIV prevention package providing choice of daily pills or long-acting injectable cabotegravir and opportunities to change prevention options could increase biomedical prevention coverage and reduce new HIV infections.
Methods: This study was an extension of three randomised trials that used SEARCH dynamic choice HIV prevention to recruit adults (aged ≥15 years) at risk for HIV from antenatal, outpatient, and community settings in rural Uganda and Kenya. In this 48-week open-label extension, participants maintained their original (1:1) randomisation group; the option to choose cabotegravir long-acting injectable was added for intervention participants. Inclusion criteria for the extension were previous enrolment in a SEARCH dynamic choice HIV prevention trial, negative HIV rapid test, and residence in study region. The intervention provided person-centred choice of oral pre-exposure prophylaxis (PrEP) or post-exposure HIV prophylaxis (PEP) or cabotegravir long-acting injectable, with the option to switch according to participant preference. The control provided standard-of-care access to oral PrEP and PEP, but not cabotegravir long-acting injectable. Biomedical prevention coverage (proportion of follow-up covered by oral PrEP, PEP, or cabotegravir long-acting injectable; primary outcome) and HIV incidence (secondary outcome) were compared between groups using targeted minimum loss-based estimation. The trial (NCT05549726) is closed to recruitment.
Findings: Of 1534 participants initially randomly assigned (from April 15, 2021 to Sept 29, 2022), 984 (487 in the intervention group and 497 in the standard-of-care group) reconsented to the extension (from Jan 2 to March 3, 2023). The mean proportion of follow-up covered by biomedical HIV prevention was 69·7% (95% CI 64·9–74·5) in the intervention group versus 13·3% (10·2–16·3) in the standard-of-care group, corresponding to an absolute difference of 56·4 percentage points (95% CI 50·8–62·1; p<0·0001). The intervention significantly improved coverage across prespecified subgroups (sex and age groups). During the study, 274 (56%) of 485 intervention participants used cabotegravir long-acting injectable, 255 (53%) used oral PrEP, and ten (2%) used PEP. Among cabotegravir longacting injectable initiators, 118 (43%) of 274 were not previously using oral PrEP or PEP. There were seven incident HIV infections in 390 person-years of follow-up in the standard-of-care group and no infections in 400 person-years of follow-up in the intervention group (incidence rate difference per 100 person-years 1·8, 95% CI 0·4–3·2; p=0·014).
Interpretation: Offering people the choice of HIV biomedical prevention options including cabotegravir long-acting injectable in a flexible model can increase prevention coverage and reduce incident HIV infections. HIV programmes should support dynamic choice HIV prevention programmes that include effective oral and injectable long-acting products.