Background Due to the low number of individuals with HIV-2, no randomised trials of HIV-2 treatment have ever been done. We hypothesised that a non-comparative study describing the outcomes of several antiretroviral therapy (ART) regimens in parallel groups would improve understanding of how differences between HIV-1 and HIV-2 might lead to different therapeutic approaches. Methods This pilot, phase 2, non-comparative, open-label, randomised controlled trial was done in Burkina Faso, C & ocirc;te d'Ivoire, Senegal, and Togo. Adults with HIV-2 who were ART naive with CD4 counts of 200 cells per mu L or greater were randomly assigned 1:1:1 to one of three treatment groups. A computer-generated sequentially numbered block randomisation list stratified by country was used for online allocation to the next available treatment group. In all groups, tenofovir disoproxil fumarate (henceforth tenofovir) was dosed at 245 mg once daily with either emtricitabine at 200 mg once daily or lamivudine at 300 mg once daily. The triple nucleoside reverse transcriptase inhibitor (NRTI) group received zidovudine at 250 mg twice daily. The ritonavir-boosted lopinavir group received lopinavir at 400 mg twice daily boosted with ritonavir at 100 mg twice daily. The raltegravir group received raltegravir at 400 mg twice daily. The primary outcome was the rate of treatment success at week 96, defined as an absence of serious morbidity event during follow-up, plasma HIV-2 RNA less than 50 copies per mL at week 96, and a substantial increase in CD4 cells between baseline and week 96. This trial is registered at ClinicalTrials.gov, NCT02150993, and is closed to new participants. Findings Between Jan 26, 2016, and June 29, 2017, 210 participants were randomly assigned to treatment groups. Five participants died during the 96 weeks of follow-up (triple NRTI group, n=2; ritonavir-boosted lopinavir group, n=2; and raltegravir group, n=1), eight had a serious morbidity event (triple NRTI group, n=4; ritonavir-boosted lopinavir group, n=3; and raltegravir group, n=1), 17 had plasma HIV-2 RNA of 50 copies per mL or greater at least once (triple NRTI group, n=11; ritonavir-boosted lopinavir group, n=4; and raltegravir group, n=2), 32 (all in the triple NRTI group) switched to another ART regimen, and 18 permanently discontinued ART (triple NRTI group, n=5; ritonavir-boosted lopinavir group, n=7; and raltegravir group, n=6). The Data Safety Monitoring Board recommended premature termination of the triple NRTI regimen for safety reasons. The overall treatment success rate was 57% (95% CI 47-66) in the ritonavir-boosted lopinavir group and 59% (49-68) in the raltegravir group. Interpretation The raltegravir and ritonavir-boosted lopinavir regimens were efficient and safe in adults with HIV-2. Both regimens could be compared in future phase 3 trials. The results of this pilot study suggest a trend towards better virological and immunological efficacy in the raltegravir-based regimen.< Réduire