Background: VOICE was a phase IIB trial of daily oral tenofovir disoproxil fumarate (TDF), oral TDF-emtricitabine (FTC), and 1% vaginal tenofovir (TFV) gel for HIV chemoprevention, in Uganda, Zimbabwe and South Africa (RSA). Plasma TFV levels objectively measure product use and correlate with efficacy. We measured plasma TFV levels (≥0.3 ng/mL) in a subset of VOICE participants on active products at their first quarterly visit, and assessed factors associated with early product use.
Methods: Of 5029 enrollees, we analyzed available TFV levels in plasma samples taken at 3 months post-randomization from women assigned to active product and not on product hold. We examined a pre-determined set of demographic, socio-behavioral and clinical factors assessed at baseline or 3 months and their association (at p<0.05 significance level) with TFV detection in the oral and gel groups. All analyses were weighted to adjust for the sampling design. Weighted generalized linear models with a log-link and robust standard errors were fitted to estimate prevalence ratio (PR) of detectable TFV in plasma, adjusting for country.
Results: Of 1146 participants assessed at 3-month follow-up, 33% had TFV detectable in the oral group (N=637) and 27% in the gel group (N=509), respectively. TFV detection by group and country is summarized in the table below. Controlling for country, factors associated with TFV detection in the oral group were: receiving material support from partner (PR=0.64; p=0.02) and frequent alcohol use (PR=1.93; p<0.01). In the gel group, factors associated with TFV detection were: age ≥25 (PR=1.6; p=0.03), partner disapproval of product use (PR=0.34, p=0.02) and social harm (PR=0.23 p=0.04). In both groups, moderate perceived HIV risk was associated with detecting TFV.
Conclusions: In these VOICE participants, use of oral tablets and gel at initial PK assessment varied by country. Further, different factors correlated with product use in the oral and gel groups, suggesting geographical, demographic and relationship difference in end-users characteristics for different routes of administration. This study adds to the growing body of evidence that choice of biomedical prevention options will be critical to address the varying prevention needs of women at risk of HIV.