Pharmacologic evaluation of delayed long-acting cabotegravir administration among cisgender women in HPTN 084

Abstract

HPTN 084 demonstrated the superiority of long-acting injectable cabotegravir (CAB-LA) compared with daily oral tenofovir disoproxil fumarate-emtricitabine (F/TDF) for HIV prevention in women. CAB-LA (600 mg) or placebo injections were administered 4 weeks after an initial dose (loading dose) and every 2 months (Q2M) thereafter; this is the approved regimen. Participants experienced both loading dose and Q2M delays during the trial. CAB concentrations were evaluated before a delay, at the visit associated with the delay, and the visit after a delayed injection was administered. During the blinded phase of the trial, 194 participants randomized to CAB-LA experienced at least one injection delay. Plasma CAB concentrations were maintained above the 4× protein adjusted 90% inhibitory concentration (4× PA-IC₉₀; protocol-specific threshold) for all loading dose and 98% of Q2M delays when injections were administered up to 6 weeks late. The feasibility of shifting to an every 3-month (Q3M) regimen in females was interrogated via simulation studies using a population pharmacokinetic model. Q3M injections in both CAB-naïve (with a loading dose) and previously CAB-exposed females were predicted to yield higher steady-state exposures than in males on the approved Q2M regimen. Although there is observed forgiveness following an isolated delayed CAB-LA injection and simulations suggest acceptable CAB-LA exposures in women with a 600 mg CAB-LA Q3M regimen, empirical efficacy of this regimen has not been established, and transitioning to this dosing schema is not recommended. Future pharmacokinetic bridging studies are aimed at evaluating higher dose CAB-LA formulations administered less frequently.

Clinical trials: This study is registered with ClinicalTrials.gov as NCT03164564.

Keywords: HIV; cabotegravir; modeling; pharmacokinetics; prevention.

Fig 1

CAB concentrations at the delayed injection visit. CAB concentration for loading dose or Q2M injection delays are presented as a function of interval between injections. Red, loading dose delays. Blue, Q2M Dose delays. IQR, interquartile range.

Fig 2

CAB PA-IC₉₀ categorization following an injection delay. The frequency of (A) loading dose and (B) Q2M injection delays with CAB concentrations above target thresholds. The interval between a loading dose and a subsequent dose is stratified as 8–10 weeks (56–70 days), 10–12 weeks (70–84 days), and 12–14 weeks (84–98 days); interval between bimonthly administration is stratified as 12–14 weeks (84–98 days), 14–16 weeks (98–112 days), and 16–18 weeks (112–126 days). CAB concentrations were stratified into four groups relative to the PA-IC₉₀: <1 × PA-IC₉₀, 1 × to < 4× PA-IC₉₀ (0.166 to 0.664  µg/mL), 4 × to < 8× PA-IC₉₀ (0.664 to 1.33  µg/mL), and ≥8 × PA-IC₉₀ (≥1.33  µg/mL). PA-IC₉₀ is the in vitro protein-adjusted 90% maximal inhibitory concentration.

Fig 3

Comparison between observed (hollow circle) and model-predicted (solid line and grey band) CAB concentration–time profile for participants receiving delayed injections.

Fig 4

Simulated CAB plasma concentration–time profiles in females (black line median, gray-shaded 90% prediction interval) receiving Q3M 600 mg CAB-LA injections with single-loading (A and B) and zero-loading (C and D) for a CAB-naïve population (A and C) and a CAB-exposed population (B and D), where the transition from steady-state Q2M dosing occurs at week 8 with (B) or without (D) a loading at this transition. CAB-exposed reflects a population that reached steady state following Q2M injections (in females, the final Q2M injection was administered at time 0 and the first injection of the Q3M regimen was administered at week 8). Single-loading = first 2 injections 1 month apart, whereas all subsequent injections were 3 months apart; Zero-loading = all injections were 3 months apart. For simulated CAB profiles in males (red solid line median, red dashed line 90% prediction interval), all four panels indicate receiving Q2M 600 mg CAB-LA injections with a loading dose 4 weeks following the initial dose, which is apparent in both “naïve” panels (A and C) and before the displayed times in both “exposed” panels (B and D); the “zero-loading” designation is only for female simulations. The y axis is on log scale. PA-IC₉₀, in vitro protein-adjusted 90% maximal inhibitory concentration; PI, prediction interval; Q2M, once every 2 months; Q3M, once every 3 months; reference line, 0.664 µg/mL = 4× PA-IC₉₀.