Evaluation of long-acting cabotegravir safety and pharmacokinetics in pregnant women in eastern and southern Africa: a secondary analysis of HPTN 084

Abstract

Introduction: Long-acting injectable cabotegravir (CAB-LA) for pre-exposure prophylaxis significantly reduced HIV acquisition in HPTN 084. We report on the safety and CAB-LA pharmacokinetics in pregnant women during the blinded period of HPTN 084.

Methods: Participants were randomized 1:1 to either active cabotegravir (CAB) plus tenofovir disoproxil fumarate/emtricitabine (TDF/FTC) placebo or active TDF/FTC plus CAB placebo. Pregnancy testing was performed at each visit; participants with a positive test had study product withheld and were offered open-label TDF/FTC. Pregnancies were confirmed on two tests at least 4 weeks apart. All participants with a positive pregnancy test prior to November 5, 2020 are included in this analysis. Pregnancy incidence, maternal adverse event (AE) incidence, pregnancy outcomes (including composite outcome of spontaneous abortion <20 weeks, intrauterine foetal death or stillbirth ≥20 weeks, premature birth <37 weeks, or small for gestational age) were assessed. The apparent terminal phase half-life (t_1/2app) of CAB-LA in pregnant women in HPTN 084 was compared to non-pregnant women from the phase 2a HPTN 077 trial. Multivariable models assessed associations with t_1/2app. RESULTS: Fifty-seven pregnancies (30 CAB-LA, 27 TDF/FTC) were confirmed over 3845 person-years [py] (incidence 1.5/100 py, 95% CI 1.1-1.9). CAB-LA group participants had a median 342 days (IQR 192, 497) of CAB-LA exposure prior to pregnancy detection. Grade 2 or higher maternal AE incidence did not differ by study arm (CAB 157, 95% CI 91-271 per 100 py vs. TDF/FTC 217, 95% CI 124-380 per 100 py; p = 0.256). Most pregnancies (81%) resulted in live births (25 CAB-LA, 22 TDF/FTC). Composite poor pregnancy outcomes did not differ significantly by group (CAB 6/30 vs. TDF/FTC 4/27; p = 0.476). No congenital anomalies were observed. The CAB t_1/2app geometric mean was 52.8 days (95% CI 40.7-68.4) in pregnant women compared to 60.3 days (95% CI 47.7-76.3; p = 0.66) in non-pregnant women; neither pregnancy nor body mass index were significantly associated with t_1/2app.

Conclusions: CAB-LA concentrations post-cessation of injections were generally well tolerated in pregnant women. The t_1/2app was comparable between pregnant and non-pregnant women. Ongoing studies will examine the safety and pharmacology of CAB-LA in women who choose to continue CAB-LA through pregnancy and lactation.

Keywords: HIV prevention; PrEP; cabotegravir; long‐acting; pregnancy; women.

Figure 1

Box plots of standardized infant growth measures at birth, by study arm.

Figure 2

Geometric mean CAB‐LA concentrations following first positive pregnancy date in study participants who received at least one injection of CAB‐LA (n = 27). The in vitro protein‐adjusted 90% CAB‐LA inhibitory concentration (PA‐IC90) is 166 ng/ml [28]. The assay LLOQ is 25.0 ng/ml.

Figure 3

Individual participant log‐linear regression curves of plasma CAB‐LA concentrations from the visit after last injection to the last available quantifiable sample in (a) pregnant women from HPTN 084 (n = 17) and (b) non‐pregnant women from HPTN 077 (n = 35). Curves were fitted for each individual extrapolated to the intersection with LLOQ and all extended beyond the observed concentrations.