Greater dapivirine release from the dapivirine vaginal ring is correlated with lower risk of HIV-1 acquisition: a secondary analysis from a randomized, placebo-controlled trial

Abstract

Introduction: A vaginal ring containing 25 mg of the antiretroviral dapivirine has demonstrated efficacy in reducing women's risk of sexually acquiring HIV-1; however, imperfect ring use likely diluted efficacy estimates in clinical trials. The amount of dapivirine remaining in returned rings may reflect the extent of product use, permitting estimation of HIV protection in the context of consistent use.

Methods: We measured the amount of dapivirine in returned rings from a placebo-controlled trial of the dapivirine vaginal ring conducted between August 2012 and June 2015 among 2629 African women. Phase I/II studies established that greater than 4 mg of dapivirine on average is released from the ring when used consistently over 28 days and ≤0.9 mg released suggested non-use. We assessed the relative risk reduction associated with levels of ring use using residual dapivirine in returned rings as a time-dependent covariate for HIV-1 infection in multivariable Cox models, including multiple exploratory analyses designed to estimate upper limits of efficacy given uncertainty in timing of HIV-1 acquisition. All models were adjusted for baseline covariates associated with HIV risk and adherence.

Results: Residual dapivirine levels indicating at least some use (>0.9 mg released over a month) were associated with a 48% relative reduction in HIV-1 acquisition risk (95% confidence interval (CI): 21% to 66%; p = 0.002) compared to the placebo. Exploratory analyses accounting for potential misclassification in timing of HIV-1 acquisition estimated 75% to 91% HIV-1 risk reduction with> 4 mg released when compared to placebo. Results limited to the subgroup of women <25 years of age, who tended to have lower adherence, were generally consistent to those overall.

Conclusions: Residual dapivirine levels, an objective measure of adherence, were correlated with HIV-1 protection in a secondary analysis of a randomized trial. Periods of ring use were associated with approximately 50% protection, with exploratory analyses suggesting higher protection with more consistent use. The dapivirine vaginal ring is the first method to fulfil the promise of a fully reversible, long-acting, woman-initiated approach for discreet HIV-1 prevention.

Keywords: Africa; HIV prevention; HIV prevention trials; adherence; clinical trials; women.

Figure 1

Distributions of remaining dapivirine (DPV) in returned rings (A), DPV released from returned rings (B) and rate of DPV release (C). The range of DPV release has a lower limit below 0 due to measurement error. Remaining DPV in returned rings is the measure returned from the laboratory. Released DPV subtracts the lab value from the average load corresponding to the manufacturing batch for that ring. Average rate of release is calculated as the released DPV divided by the time that the participant had the ring. Panel D illustrates the relationship between rate of release and amount released. The red line is the expected relationship (intercept = 0; slope = 1). The green line is the expected rate of release (intercept = 4 mg/28 days; slope = 0) if fully adherent over 28 days. This plot shows how it is important to account for time the ring was available so as to not to potentially mis‐assign participants as fully adherent because the ring was available for longer than the usual 28 days