Background
HIV Prevention Trials Network (HPTN) 083 Trial was a randomized controlled trial that demonstrated the superiority of long-acting injectable cabotegravir (CAB-LA) over daily oral TDF/FTC for HIV PrEP among cisgender men and transgender women who have sex with men. The threshold of plasma CAB concentrations ([CAB]) needed for HIV protection in humans is difficult to define due to limited numbers of incident infections, unknown HIV-exposure timing, and unknown [CAB] at HIV exposure. A nested case-control study was conducted among HPTN 083 participants randomized to the CAB arm to investigate the association between plasma [CAB] and HIV protection.
Methods
For each of the 25 CAB-arm participants with confirmed HIV during the randomized blinded study and the first year post-unblinding, four matched, HIV-negative controls were randomly selected; matching was based on region, gender identity, race, and duration of follow-up. Plasma [CAB] was assessed throughout follow-up, and a multicompartment PK model was used to estimate [CAB] concentration-time profiles. The analysis window for each case and their matched controls was defined as the time between the last-negative HIV test and the first-positive HIV test visits for HIV-positive cases, i.e., the period when HIV acquisition likely occurred. All participants were classified according to whether their minimum concentrations during the analysis window fell within [CAB] strata bounded by the protein adjusted inhibitory concentration 1x PA-IC90 and 4x PA-IC90. Conditional logistic regression was used to estimate the association between [CAB] and risk of HIV infection.
Results
During the analysis window, minimum plasma [CAB] was below 1x PA-IC90 for 64% of confirmed HIV cases, and above 4x PA-IC90 in 76% of matched controls (Table). Minimum [CAB] above the 4x PA-IC90 was associated with a 93% (95%CI: 77%, 98%) reduction in risk of HIV acquisition compared to participants with a minimum [CAB] < 1x PA-IC90. A minimum [CAB] between 1x and 4x had an estimated risk reduction of 79% (95% CI: -20%, 96%).
Conclusions
Plasma [CAB] above 4x PA-IC90 is estimated to provide >90% protection against HIV acquisition, consistent with the target Cτ for the selected dose derived from non-human primate models. Data are insufficient for a precise estimate for plasma [CAB] between 1x and 4x PA-IC90. Q2M dosing has been shown to achieve [CAB] above 4x PA-IC90. Novel formulations, dosing, and time-to-protection estimates may be informed by these results.