Effect of efavirenz on levonorgestrel concentrations among Malawian levonorgestrel implant users for up to 30 months of concomitant use: a subanalysis of a randomized clinical trial

Abstract

Objectives: Our primary objective was to compare geometric mean levonorgestrel concentrations between levonorgestrel implant users who were or were not taking the antiretroviral efavirenz, for up to 30 months after implant initiation. Our secondary objective was to evaluate the pregnancy rate among levonorgestrel implant users on efavirenz.

Study design: We performed a subanalysis of 42 Malawian women randomized to initiate the levonorgestrel implant as part of a parent randomized clinical trial. Our subset included 30 HIV-infected women taking efavirenz and 12 HIV-uninfected women not taking efavirenz. They underwent urine pregnancy testing every 3 months and serum levonorgestrel testing at day 3 and months 1, 3, 6, 12, 18, 24, 27 and 30 after implant initiation. Geometric mean levonorgestrel concentrations were calculated for efavirenz users and non-efavirenz users at each time point.

Results: The geometric mean levonorgestrel concentrations were lower for efavirenz users than non-efavirenz users at every time point; the geometric mean ratio for efavirenz users:non-efavirenz users ranged from 0.60 [90% confidence interval (CI) 0.46-0.79] at 1 month to 0.27 (90% CI 0.12-0.61) at 30 months after implant insertion. No pregnancies occurred over 60 woman-years of concomitant levonorgestrel implant and efavirenz use, although 11 women had levonorgestrel concentrations < 180 pg/mL (the previously suggested minimum threshold concentration for efficacy).

Conclusions: Efavirenz users had lower levonorgestrel concentrations than non-efavirenz users, and one third of our concomitant efavirenz and levonorgestrel implant users had concentrations < 180 pg/mL. Continued evaluation of the contraceptive efficacy of the levonorgestrel implant may be needed for efavirenz users.

Implications: Among 42 Malawian women using the levonorgestrel implant for contraception, women who were taking the antiretroviral efavirenz had lower serum levonorgestrel concentrations than women who were not taking efavirenz. However, none of the women who were taking efavirenz became pregnant over 60 women-years of follow-up.

Keywords: Contraceptive implant; Drug concentrations; Drug interactions; Efavirenz; Levonorgestrel.

Fig. 1

Graphical comparison of geometric mean serum levonorgestrel concentrations by time since levonorgestrel implant initiation between efavirenz users and efavirenz nonusers. Legend: EFV = efavirenz.

Fig. 2

Levonorgestrel concentrations rebound after stopping efavirenz among two HIV-infected levonorgestrel implant users who switched to a nonefavirenz-based antiretroviral therapy regimen. Legend: Levonorgestrel concentrations for two users undergoing switching to a nonefavirenz-based antiretroviral therapy regimen after having received the implant are plotted over time. Color-coded arrows correspond to the first study visit occurring after the switch from efavirenz. In both instances, serum levonorgestrel concentrations rebound above baseline concentrations.