Abstract Body

Adherence in pre-exposure prophylaxis (PrEP) trials can be inferred by measuring drug concentrations at the time of HIV diagnosis. However, results do not necessarily reflect pill taking behavior around the time of HIV infection. The Bangkok Tenofovir Study (BTS) evaluated the efficacy of daily oral PrEP with tenofovir (TDF) among persons who inject drugs. We used molecular tools and adherence diaries from BTS participants to define pill taking behavior at the estimated time of infection

HIV genetic diversity was studied in 11 participants who became infected with HIV during periods of daily directly observed therapy (DOT) with TDF and in 13 placebo recipients. HIV env (V1-V5) was amplified from plasma collected at the first nucleic acid test positive (NAT+) visit using single genome amplification. The diversity structure of the virus population seen in this sample was used to estimate time since infection (±95% confidence interval) using the Poisson-Fitter tool. The number of TDF doses taken within the confidence interval of infection were then computed from adherence diaries

The estimated time since infection in the TDF group ranged from 8 (95%CI=1-16) to 75 (49-100) days (median=41) and did not differ from that seen in the placebo group (median=15 days; p=0.433). Analysis of missed TDF doses within the estimated period when HIV infection occurred showed perfect adherence in 5/11 (45%) participants, nearly perfect adherence in two (67/69 and 95/100 possible doses taken), low adherence in one (52 out of 71 possible doses taken) and non-adherence (no TDF taken) in three. The 2-5 missed doses in the two participants with nearly perfect adherence were recorded immediately before (3-4 days) or after (1-2 days) the estimated day of infection. TDF exposure in adherent participants was associated with low plasma virus loads at the first NAT+ visit (13,465 RNA copies/ml vs. 294,000 in placebo infections; p=0.033) and infection with single variants (8/8 vs.8/13 in the placebo group; p=0.063) although the latter was not significant at the 0.05 level

We used molecular tools to characterize adherence to PrEP at the estimated time of HIV infection, and demonstrated breakthrough infections under conditions of perfect or near perfect adherence to TDF. These results highlight the utility of molecular tools to improve computation of the biological efficacy of PrEP. Pre-diagnosis drug activity by PrEP reduces acute viremia and may favor the transmission of single variants