Evaluation of pharmacokinetic interactions between long-acting cabotegravir or emtricitabine/tenofovir disoproxil fumarate and hormonal contraceptive agents: a tertiary analysis of South African participants in HPTN 084

Abstract

Introduction: HPTN 084 found that long-acting cabotegravir (CAB-LA) was well-tolerated and significantly reduced the risk of HIV acquisition in women compared to tenofovir disoproxil fumarate/emtricitabine (F/TDF). During the blinded phase of the trial, participants were required to use an effective method of contraception, including an injectable or implantable hormonal contraceptive (HC) agent. A contraceptive sub-study assessed the pharmacokinetic interactions between pre-exposure prophylaxis agents (CAB-LA or F/TDF) and etonogestrel (ENG), medroxyprogesterone acetate (MPA) or norethindrone enanthate (NET-EN). Methods: Participants were enrolled in a nested sub-study between 24 February 2020 and 26 October 2020. Via a convenience sampling strategy, plasma concentrations of ENG, MPA and NET-EN were evaluated at enrolment and weeks 25, 49 and 73; plasma tenofovir (TFV) and CAB concentrations were determined at contemporaneous visits. Participants were allowed to switch contraceptives, and HC assessments were adjusted accordingly. Geometric mean concentrations were calculated and compared using t-tests or Fisher's exact tests. Results: One hundred and seventy participants were included in this analysis. Hormone concentrations at all study visits were comparable between the CAB-LA and F/TDF study arms. Among participants randomized to the CAB-LA arm, geometric mean concentrations declined from enrolment to the follow-up period for ENG (335 to 202 pg/ml), MPA (1520 to 1138 pg/ml) and NET-EN (3715 to 1888 pg/ml); similar findings were observed among participants randomized to the F/TDF arm. Observed HC declines are likely attributed to the timing of contraceptive administration relative to sampling; the percentage of participants with hormone concentrations above thresholds associated with ovulation suppression was high (73-100%) and did not differ between arms. CAB concentrations were comparable across contraceptive types, with 97.8-98.1% of participants yielding trough CAB concentrations above the protocol-specified target threshold. TFV concentrations were unquantifiable for most participants, irrespective of contraceptive agent, rendering comparisons largely uninformative. Conclusions: Given the comparable hormone concentrations between arms and the likely influence of the timing of sample collection on observed measurements, clinically significant interactions between CAB-LA and HC are not expected. Associations between F/TDF and hormone concentrations could not be effectively evaluated due to low adherence to F/TDF.