Safety, adherence, and HIV-1 seroconversion among women using the dapivirine vaginal ring (DREAM): an open-label, extension study.

Background: The Ring Study, a phase 3 trial in 1959 sexually active women (randomised 2:1), showed a favourable safety profile and a 31% HIV-1 infection risk reduction for a vaginal ring containing 25 mg of dapivirine, compared with a placebo ring. We report here the DREAM study, which aimed to evaluate safety, adherence, and HIV-1 incidence in those using the dapivirine vaginal ring (DVR) in open-label use.

Methods: The DREAM study is an open-label extension of The Ring Study, done at five research centres in South Africa and one research centre in Uganda. Former participants from The Ring Study, who remained HIV-negative and who did not discontinue the study due to an adverse event or safety concern that was considered to be related to the investigational product, were eligible. Women who were pregnant, planning to become pregnant, or breastfeeding at screening for DREAM were excluded. All participants received the DVR for insertion at the enrolment visit. Participants attended a 1-month follow-up visit and could either proceed with visits once every 3 months or attend monthly visits up to month 3 and then continue with visits once every 3 months. At each visit, HIV testing and safety evaluations were done, and residual dapivirine measured in used rings (approximately 4 mg is released from the DVR over 28 days of consistent use). HIV-1 incidence was compared descriptively with the simulated incidence rate obtained from bootstrap sampling of participants in the placebo group of The Ring Study, matched for research centre, age, and presence of sexually transmitted infections at enrolment. This study is registered with ClinicalTrials.gov, NCT02862171.

Findings: Between July 12, 2016, and Jan 11, 2019, 1034 former participants from The Ring Study were screened, 941 were enrolled and 848 completed the trial. 616 (65·5%) of 941 participants reported treatment-emergent adverse events. Of these, six (0·6%) had events considered to be treatment-related. No treatment-related serious adverse events were reported. Measurements of monthly ring residual amounts in participants enrolled in both trials showed consistently lower mean values in DREAM than in The Ring Study. Arithmetic mean ring residual amounts of participants in The Ring Study DVR group who enrolled in DREAM were 0·25 mg lower (95% CI 0·03-0·47; p=0·027) than the mean ring residual amounts of these participants in The Ring Study. 18 (1·9%) HIV-1 infections were confirmed during DVR use, resulting in an incidence of 1·8 (95% CI 1·1-2·6) per 100 person-years, 62% lower than the simulated placebo rate.

Interpretation: Although efficacy estimation is limited by the absence of a placebo group, the observed low HIV-1 incidence and improved adherence observed in DREAM support the hypothesis that increased efficacy due to improved adherence occurs when women know the demonstrated safety and efficacy of the DVR. The feasibility of a visit schedule of once every 3 months was shown, indicating that the DVR can be used in a real-world situation in usual clinical practice.

Funding: The Ministry of Foreign Affairs (MFA) Denmark, Flanders MFA, Irish Aid, Dutch MFA, UK Aid from the UK Government's Foreign, Commonwealth and Development Office, and the US President's Emergency Plan for AIDS Relief through the US Agency for International Development.