Modern contraceptive use by HIV-infected women prevents maternal morbidity, economic disparity, and perinatal HIV infection. Despite the clear benefits of hormonal contraceptives, there continues to be significant unanswered questions about contraceptive safety in women living with HIV on anti-retroviral therapy (ART). Specifically, hormonal contraceptives are metabolized in the liver by cytochrome P450 (CYP) enzymes and some ART drugs can increase the function of these enzymes leading to lower systemic concentrations of the hormonal contraceptives and contraceptive failures. Currently, non-nucleoside reverse transcriptase inhibitors (NNRTIs), such as nevirapine (NVP) and efavirenz (EFV), are widely prescribed HIV medications in sub-Saharan Africa, and are known to be CYP inducers. It is currently recommended that all HIV-infected adults should be treated with ART, thus the approximately 13 million women living with HIV in SSA are candidates for NNRTI-based ART. Simultaneously, contraceptive implants are the fastest growing method of contraception in SSA with a 9-fold increase from 2008 to 2012 and 9 million implants shipped to SSA in 2015. Unfortunately, several pharmacokinetic (PK) studies of hormonal contraception with combined EFV use have shown a decrease in progestin exposure by ~60% with different progestins and routes of administration, including levonorgestrel (LNG) emergency contraception and implants, desogestrel combined oral contraceptive pill, and the ENG implant. Additionally, there have been pregnancies reported for women using both the LNG and etonogestrel (ENG) implant and EFV-based ART.
The purpose of the study was to evaluate the bi-directional PK interaction between the ENG implant and NNRTI-based ART (EFV and NVP). We enrolled 60 HIV-infected women desiring the implant for contraception into 3 groups: ART-naïve, EFV-based ART, and NVP-based ART (20 participants per group). The ENG implant was placed at enrollment and the participants were followed for 6 months. Women on EFV-based ART also had copper intrauterine devices or prior surgical sterilization in place to prevent pregnancy given the concern prompting this study. ENG concentrations were measured at 4, 12 and 24 weeks after implant placement; and the EFV and NVP concentrations were measured at screening, enrollment, then 4, 12, and 24 weeks after implant placement for participants on ART.
There were no participants lost to follow-up. One participant in the EFV group was excluded due to lack of ART adherence and one participant in the NVP group was excluded due to a sampling processing error. Geometric mean ENG concentrations at week 24 were 362, 66, and 341 pg/mL in the ART-naïve, EFV and NVP groups, respectively, indicating that ENG concentrations were 82% lower in the EFV group as compared to the ART-naïve group. In contrast, NVP did not significantly impact the concentrations of ENG from the implant. For women on EFV-based ART, the concentrations of EFV decreased after the placement of the implant. However, there was no significant decrease in NVP concentrations after implant placement. Notably, no participant in the EFV group and one participant in the NVP group had concentrations below the suggested threshold to suppress HIV, indicating that this decline is not likely to be clinically significant.
Over the 24 weeks of combined use, ENG exposure was significantly decreased for women on EFV-based ART, but not for women on NVP-based ART. HIV and reproductive health care providers should not use implantable contraception at the current dosage for women on EFV-based ART. Future studies should focus on dose escalation for implantable contraception for women on EFV-based ART. Additionally, ART concentrations were lower after ENG contraceptive implant placement, indicating a bi-directional interaction between hormonal contraceptives and NNRTI-based ART. Future studies should confirm impact of ENG on NNRTI metabolism and evaluate the clinical significance of this interaction with markers of HIV disease such as CD4 count and HIV viral load safety in women living with HIV on anti-retroviral therapy (ART).  Specifically, hormonal contraceptives are metabolized in the liver by cytochrome P450 (CYP) enzymes and some ART drugs can increase the function of these enzymes leading to lower systemic concentrations of the hormonal contraceptives and contraceptive failures.  Currently, non-nucleoside reverse transcriptase inhibitors (NNRTIs), such as nevirapine (NVP) and efavirenz (EFV), are the most widely prescribed HIV medications in sub-Saharan Africa, and are known to be CYP inducers.  EFV-based ART is the first-line WHO recommended regimen for all adults and is the most commonly used regimen in sub-Saharan Africa (SSA), including in women of child-bearing potential.  It is currently recommended that all HIV-infected adults should be treated with ART, thus the approximately 13 million women living with HIV in SSA are candidates for NNRTI-based ART.  Simultaneously, contraceptive implants are the fastest growing method of contraception in SSA with a 9-fold increase from 2008 to 2012 and 9 million implants shipped to SSA in 2015.  Unfortunately, several pharmacokinetic (PK) studies of hormonal contraception with combined EFV use have shown a decrease in progestin exposure by ~60% with different progestins and routes of administration, including levonorgestrel (LNG) emergency contraception and implants, desogestrel combined oral contraceptive pill, and the ENG implant.   Additionally, there have been pregnancies reported for women using both the LNG and etonogestrel (ENG) implant and EFV-based ART. 
The purpose of the study was to evaluate the bi-directional PK interaction between the ENG implant and NNRTI-based ART (EFV and NVP).  We enrolled 60 HIV-infected women desiring the implant for contraception into 3 groups:  ART-naïve, EFV-based ART, and NVP-based ART (20 participants per group).  The ENG implant was placed at enrollment and the participants were followed for 6 months.  Women on EFV-based ART also had copper intrauterine devices or prior surgical sterilization in place to prevent pregnancy given the concern prompting this study.  ENG concentrations were measured at 4, 12 and 24 weeks after implant placement; and the EFV and NVP concentrations were measured at screening, enrollment, then 4, 12, and 24 weeks after implant placement for participants on ART. 
There were no participants lost to follow-up. One participant in the EFV group was excluded due to lack of ART adherence and one participant in the NVP group was excluded due to a sampling processing error.  Geometric mean ENG concentrations at week 24 were 362, 66, and 341 pg/mL in the ART-naïve, EFV and NVP groups, respectively, indicating that ENG concentrations were 82% lower in the EFV group as compared to the ART-naïve group.  In contrast, NVP did not significantly impact the concentrations of ENG from the implant.  For women on EFV-based ART, the concentrations of EFV decreased after the placement of the implant.  However, there was no significant decrease in NVP concentrations after implant placement.  Notably, no participant in the EFV group and one participant in the NVP group had concentrations below the suggested threshold to suppress HIV, indicating that this decline is not likely to be clinically significant. 
Over the 24 weeks of combined use, ENG exposure was significantly decreased for women on EFV-based ART, but not for women on NVP-based ART.  HIV and reproductive health care providers should not use implantable contraception at the current dosage for women on EFV-based ART.  Future studies should focus on dose escalation for implantable contraception for women on EFV-based ART.  Additionally, ART concentrations were lower after ENG contraceptive implant placement, indicating a bi-directional interaction between hormonal contraceptives and NNRTI-based ART.  Future studies should confirm impact of ENG on NNRTI metabolism and evaluate the clinical significance of this interaction with markers of HIV disease such as CD4 count and HIV viral load.