Objectives: The primary objective of this study was to characterize the pharmacokinetics of etonogestrel (ENG) released from a contraceptive implant in Ugandan women living with HIV who were receiving efavirenz (EFV) or nevirapine (NVP)-based antiretroviral therapy (ART), compared with ART-naive women over 24 weeks. Design: Nonrandomized, parallel-group study with three arms: ART-naive, NVP, or EFV-based ART (N==20/group). Methods: Sparse pharmacokinetic sampling of ENG, NVP, or EFV were performed at screening, entry, and then 1, 4, 12, and 24-week postimplant insertion. The primary endpoint was ENG concentrations at week 24, compared between the ART-naive group and each ART group, using geometric mean ratio (GMR) with 90% confidence intervals. Results: Sixty participants competed the 24-week study and data from 58 participants are included; one participant each was excluded from the NVP group and EFV group because of a sample processing error and ART nonadherence, respectively. At week 24, geometric mean ENG was 362, 341, and 66=pg/ml in the ART-naive, NVP, and EFV groups, respectively [GMR: NVP=:ART-naive 0.94 (0.90-1.01); EFV=:ART-naive 0.18 (0.17-0.20)]. NVP and EFV concentrations were lower at week 24 compared to preimplant [NVP: geometric mean 5.7 versus 6.8=mg/l, respectively, GMR 0.84 (0.83-0.85); EFV: geometric mean 3.6 versus 4.9=mg/l, respectively, GMR 0.73 (0.69-0.80)]. Conclusion: After 24 weeks of combined use, ENG exposure was 82% lower in women using EFV-based ART compared with ART-naive women. In contrast, NVP did not significantly impact ENG exposure. These results raise concerns about reduced effectiveness of implantable contraception for women taking EFV-based ART.
