The use of effective contraception, including hormonal, has been widely promoted in HIV-positive women in the last years.1,2 There is limited available data on the pharmacokinetic (PK) interactions between combined hormonal contraceptives and certain common classes of antiretrovirals (ARVs), such as nonnucleoside reverse transcriptase inhibitors (NNRTIs) and protease inhibitors.1 Because of common metabolic pathways in the liver of sex steroid hormones and the above-mentioned ARVs, the blood levels of both combined hormonal contraceptives and ARVs could be affected—decreased with possible compromise in effectiveness or increased with possible enhanced toxicity. In this article, we evaluate interactions between combined hormonal contraception, in particular combined oral contraceptives (COCs) and NNRTIs.
Few studies in this field have been reported and they showed conflicting results. These studies generally had small sample sizes, were conducted outside real-life conditions, and without reaching a steady state of hormones and/or ARVs.3–7 The interaction between nevirapine (NVP) and COC in HIV-positive women has been reported in 2 studies,3,4 1 showed a decrease in sex steroid hormones,3 and the other showed the contrary.4 The interaction between efavirenz (EFV) and sex steroid hormones, administered either as COC or as ethinyl estradiol (EE) only, or progestin only, has been assessed in HIV-negative women.5–7 These studies reported comparable outcomes—no change or slight increases in EE levels, significant reduction in the progestin levels, and no change in the EFV levels. These studies concluded that there is a need to use reliable barrier contraception when taking COC with EFV,7 or higher dose of progestin, in case of progestin-only contraceptive, such as emergency contraception.6 Two studies assessed the interaction between depot medroxyprogesterone acetate and EFV8,9 or NVP8 in HIV-positive women. They found no changes in the progestin level,8,9 low progesterone levels, consistent with anovulation,8,9 and no clinically significant changes in the EFV and NVP levels.8 The authors concluded that depot medroxyprogesterone acetate can be used safely with EFV and NVP.8 Studies on new NNRTIs, etravirine, and rilpivirine, subject to similar limitations, did not find significant changes in either the hormonal or the ARV levels.10,11 Nucleoside/nucleotide analogue reverse transcriptase inhibitors (NRTIs/NtRTIs), a common backbone regime of combined ARV therapy, have a different metabolic pathway and are not expected to go into PK interactions with sex steroid hormones.12
This article reports on a marker of the likelihood of ovulation suppression (endogenous progesterone level) and the safety of a steady-state COC formula administered in HIV-positive women on stable NVP- or EFV-based ARV therapy in real-life conditions. We hypothesized that there is no risk for contraceptive or ARV therapeutic failure, if we administer COC and NVP/EFV, and that neither the contraceptive nor the ARV will cause increased toxicity. The results will add to the scarce research in this field and will help to optimize combined hormonal contraceptive use in HIV-positive women.
This was a prospective, open-label, nonrandomized steady-state clinical trial assessing the interaction between a combined oral contraceptive formula of 0.150 mg desogestrel (DSG)/0.030 mg EE and either NVP or EFV. As a combined contraceptive pill, we used Marvelon 150/30 tablets, manufactured by Organon (Oss, The Netherlands). Women were recruited from the HIV–Netherlands, Australia, Thailand (HIV-NAT) Research Collaboration Clinic and The Thai Red Cross Anonymous Clinic in Bangkok, Thailand. All subjects gave informed consent, and the study was approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University in Bangkok.
Eligible subjects had to be Thai women between 18 and 45 years of age, with regular menstrual periods, and willing to use COC pills for a period of 8 weeks. Subjects were excluded if they were pregnant or lactating, smokers, using injectable progestins, or implants within 6 months of study entry. Subjects were also excluded if they had history of thrombosis, hypertension, diabetes mellitus with vascular involvement, hyperlipidemia grade 2 and higher, presence or history of hepatic disease, malignant disease of genital organs or breast, or an undiagnosed uterine bleeding. The women had to be on a stable ARV therapy for at least 30 days before the study entry, containing either NVP or EFV combined with 2 NRTI/NtRTI. We targeted a minimum sample of 20 subjects per group.
Between the first and the third day of a regular spontaneous menstrual cycle, all eligible subjects started 1 tablet of Marvelon per day for 21 days, followed by 7 days of no COC tablets (first cycle, study days 1–28); on the 29th day the subjects restarted Marvelon for another cycle (second cycle, days 29–56). Subjects were asked not to use any other medication, including vitamins, herbs, or food supplements, apart from ARV therapy and Marvelon. Steady-state concentrations for EE were reached after 3–4 days of administration, when serum drug levels are higher by 30%–40% as compared with single dose; steady-state concentrations for etonogestrel (the active metabolite of DSG) are reached during the second half of a treatment cycle.13 Contraceptive pills were taken in the morning by all subjects. Subjects from the NVP group took the COC tablet together with the ARV drugs; subjects from the EFV group took the EFV in the evening and the COC tablet in the morning, 12 hours after the EFV tablet. During the screening process, before initiation of COC tablets, we collected blood at a single time point to measure concentrations of NVP or EFV 12 hours after administration (C12). We collected blood a second time at a single time point around study day 44 ± 2 days (second half of the second hormonal cycle) to measure C12 concentrations of NVP/EFV in the presence of DSG/EE, and for progesterone levels. Adherence to the COC and ARV tablets was assessed by self-reporting. We gathered data on all medications used from enrollment in the study until the second blood collection, and on the appearance of any adverse events. We gathered available data from records for CD4, HIV RNA, complete blood count , fasting blood sugar, lipids, and liver enzymes.
The main outcomes of the study concerned the proportion of women with sufficient contraceptive effect based on progesterone levels (levels >3.0 ng/mL might be indicative of ovulation14), the changes in NVP/EFV blood level after introduction of DSG/EE (desired therapeutic levels were > 3.1 mg/L for NVP and 1.0–4.0 mg/L for EFV), and the proportion of women exhibiting adverse effects after introduction of DSG/EE. Laboratory tests for the quantitative determination of progesterone in serum were performed at Bangkok RIA Laboratory (BRIA, Bangkok, Thailand), using the chemiluminescent microparticle immunoassay method (US Food and Drug Administration–approved reagents). Laboratory tests for EFV and NVP plasma concentrations were performed at HIV-NAT Research Laboratory, The Thai Red Cross AIDS Research Centre (Bangkok, Thailand), by using a validated reversed-phase high-performance liquid chromatography method. Both methods were originally developed at the Clinical Pharmacy, Laboratory, Radboud University (Nijmegen, The Netherlands).14,15 HIV-NAT laboratory participates in an international interlaboratory quality control program for therapeutic drug monitoring in HIV infection (Association for Quality Assessment in Therapeutic Drug Monitoring and Clinical Toxicology, The Hague, The Netherlands) for external quality control.16 For statistical analysis of the data, STATA/IC version 11.2 for Windows (Statacorp LP, College Station, TX) was used. Baseline characteristics of subjects were summarized by calculating median and interquartile range for quantitative data, and number and percentage for categorical data. For assessing the significance in the change between ARV levels with and without DSG/EE, we applied the nonparametric Wilcoxon signed-rank test. At day 44, we assessed whether the ARV levels were increased or decreased from baseline. Logistic regression was then applied to test whether any other factors under study influenced the changes in the ARV levels. Comparisons of baseline characteristics between NVP and EFV groups were assessed using the Wilcoxon rank-sum test. All hypothesis tests were 2-sided, with 5% statistical significance.
Between August 2011 and April 2012, we enrolled 49 women, 24 were in the NVP group and 25 were in the EFV group. Thirty-four subjects completed the study and were included in the analysis (18 in the NVP group and 16 in the EFV group). Fifteen subjects did not complete the study because of the inability to come for blood collection at the second time point (NVP group, n = 3; EFV group, n = 4), did not take COC tablets according to instructions (NVP group, n = 3; EFV group, n = 3), irregular menstrual period before enrollment (EFV group, n = 1), adverse effects, nausea, and dizziness (EFV group, n = 1).
Table 1 shows the characteristics of the subjects who completed the study. The median age was 36 years, and body mass index and blood pressure were normal. Baseline characteristics were comparable between groups except for systolic blood pressure and alanine aminotransferase. The backbone regime for all subjects contained 2 NRTIs/NtRTIs. The median length of the current ARV regime was 5.5 years for the NVP group (200 mg NVP twice daily) and 3 years for the EFV group (600 mg EFV once daily). No subject had an AIDS-related illness. All subjects, included in the analysis, had undetectable plasma HIV-1 RNA and reported 100% adherence rate to the ARVs and COC during the study period.
Endogenous Progesterone Levels
All subjects (n = 18) in the NVP group had serum progesterone <1.0 ng/mL (Table 2). In the EFV group, 12 subjects of 16 (75%) had serum progesterone levels <1.0 ng/mL; 4 of 16 subjects (25%) had serum progesterone >1.0 ng/mL, including 3 subjects who had >3.0 ng/mL (which might indicate ovulation). The difference in progesterone levels between the 2 groups was statistically significant (P = 0.04).
NVP and EFV Levels With and Without DSG/EE
The median C12 of NVP without DSG/EE was 6.8 (5.5–8.7) mg/L, and it increased insignificantly by 17%–7.2% (6.5–9.3 mg/L) with DSG/EE (Table 2). Six of 18 subjects (33%) showed decreased NVP levels, including 1 subject (6%) with NVP C12 below a target therapeutic level of 3.1 mg/L (2.96 mg/L). The median C12 of EFV without DSG/EE was 3.3 (2.7–4.6) mg/L, and it decreased significantly (P = 0.03) by 22%–2.7% (1.5–4.2 mg/L) with DSG/EE. In 3 of the 16 subjects (19%), C12 of EFV dropped below 1.0 mg/L, a target therapeutic threshold, after DSG/EE. In addition, 5 of the 16 subjects (31%) experienced a slight increase in EFV C12. By logistic regression analysis, in both groups the change of C12 of NVP/EFV was not associated with age, body mass index, absolute CD4 cell count, ARV therapy duration, or reported adverse effects (results not shown). The 3 subjects, mentioned in the previous paragraph, with progesterone >3.0 ng/mL were not the same subjects as the 3 subjects mentioned here whose EFV level dropped below 1.0 mg/L.
Adverse Effects After Administration of DSG/EE
Significantly more women in the EFV group reported adverse effects attributed to DSG/EE in comparison to the NVP group (P = 0.04). In the NVP group, 4 of the 18 women (22%) reported one or more of the following: headache (n = 1), dizziness (n = 1), breast tenderness (n = 3), or mood change (n = 1). In the EFV group, 9 of the 16 women (56%) reported adverse effects, including irregular uterine bleeding (n = 4), nausea (n = 5), mood change (n = 1), and dizziness (n = 1).
We found that the coadministration of DSG/EE COC in HIV-positive women on NVP-based ARV therapy did not affect the effectiveness of the contraceptive formula, based on the measurement of progesterone, or of the NVP, based on C12 of NVP. On the other hand, in the EFV group, we found that the coadministration of DSG/EE COC in HIV-positive women led to possible compromise in both, that is, the contraceptive effectiveness of DSG/EE (serum progesterone >1.0 ng/mL, including 3 subjects with >3.0 ng/mL), and ARV activity of EFV (C12 of EFV <1.0 mg/L in 19% of subjects). COCs achieve their contraceptive effect mainly by inhibiting ovulation, and serum progesterone levels >3.0 ng/mL might indicate ovulation.17 EFV levels below the therapeutic threshold of 1 mg/mL might be insufficient for effective viral suppression possibly leading to therapeutic failure.18 Furthermore, 56% of the subjects in the EFV group reported adverse effects attributed to COC such as irregular uterine bleeding and nausea.
There is no clear cutoff point in the literature, which confirms ovulation, based on a single midluteal level of progesterone. Levels >10 ng/mL are indicative of ovulation,19 although it cannot be excluded with progesterone above 3.0 ng/mL.17 According to the historical reference range for progesterone of the National Institute of Health in Bethesda, progesterone levels should be <1.0 ng/mL with oral contraceptives.20 Our findings regarding contraceptive efficacy for the NVP group in Thai women are in line with the findings of the recently published small study (n = 3) in Malawian women.4 When assessments are done in the steady state of hormones and ARVs, NVP does not seem to affect the contraceptive effectiveness of COC pills assessed by progesterone levels. In our study, single-point progesterone measurement was <1.0 ng/mL in all subjects. In the EFV group, 25% of the subjects had progesterone >1.0 ng/mL, including 3 women (19%) with progesterone >3.0 ng/mL indicative of possible COC failure. This finding is different from previous reports of administering COC pills with EFV,7 where all subjects (n = 19) had serum progesterone <1.5 ng/mL. Women on EFV might be in need of reliable contraception because of its potential teratogenicity.21,22 Until more data are available, we caution the use of combined hormonal contraceptives in HIV-positive women on EFV for pregnancy prevention. Furthermore, several case studies were reported on contraceptive failure of Implanon, an implantable contraceptive releasing etonogestrel (the active metabolite of DSG), in HIV-positive women on EFV-based therapy.23,24
The coadministration of COC in HIV-positive women on NVP-based ARV therapy led to an insignificant median increase in the NVP plasma concentration. The median C12 of NVP (6.8 mg/L), which we found without DSG/EE, is comparable to earlier reports of trough NVP levels in Thai HIV-positive people.25 In 1 subject, NVP levels dropped slightly below the 3.1 mg/L therapeutic threshold.18 In our study, women did not report adverse effects, which we could attribute to NVP. Therefore, our results suggest, that despite slight changes in NVP plasma concentration when coadministering with DSG/EE, the drug combination was well tolerated and did not lead to increased NVP toxicity/failure. These results are comparable to the results of other studies assessing the interaction between sex steroid hormones and NVP.3,4,8 However, in contrast to other studies, which did not find any significant changes in the EFV levels,5,7,8 we found a significant decrease by 22% (P = 0.03). Trough plasma concentrations of EFV between 1.0 and 4.0 mg/L are considered sufficient for preserved efficacy, and levels <1.0 mg/L in a single measurement are considered as predictive of viral failure.18,26 Three subjects (19%) had values of EFV below 1.0 mg/L after introducing DSG/EE. Therefore, our results suggest that the administration of DSG/EE in HIV-positive women on EFV-based ARV therapy could lead to decrease in EFV levels below targeted therapeutic threshold with possible therapeutic failure. Our results may have been influenced by several factors, including the different hormonal combination we used, the timing of laboratory testing during the steady state of EFV and of the hormones, and the fact that we conducted the research exclusively with Thai women, whose metabolic rate may be different. Some authors, however, question the use of trough plasma concentrations for NVP and EFV for the prediction of virological failure.18
We attributed all reported adverse effects to the COC tablets, as they were common for combined hormonal contraceptives, and subjects were stable for a median of 3 (EFV group) to 5 years (NVP group) on their designated ARV regime. COCs adverse effects are usually related to the estrogen or progestin component of the pill, plasma concentrations of which might vary significantly from individual to individual, and from one population to another.27 The EFV group experienced significantly more adverse effects than the NVP group. This may be an additional reason why the use of combined hormonal contraceptives should be cautioned in HIV-positive women on EFV.
The metabolic pathways involved in the processing of sex steroid hormones and NNRTIs are complex and involve mainly the cytochrome P450 microsomal enzyme system. Depending on their specific structure, COC and NNRTIs might suppress or induce specific isoforms of cytochrome P450.21,28–31 The metabolism of the above-discussed molecules also depends on other factors, such as transport systems, with further possibilities for wide genetic variations between individuals.32–34 Therefore, it is not possible to attribute the outcome of a PK study of 2 specific molecules to the whole class, neither for the ARVs nor for the combined hormonal contraceptives.
The study has certain limitations—the sample size is relatively small and progesterone levels were measured at a single time point, which might affect the actual result due to pulsatile secretion of the hormone with large variations in the levels.35,36 We do not have results of blood levels of the sex steroid hormones in the study (DSG and EE), which could contribute to a better understanding of the findings. In the study we used one specific type of COC tablets, containing 0.15 mg DSG/0.03 mg EE, and the results may not be directly relevant to COC with different components. Both EE and the progestin contribute to the contraceptive effect of the COC tablets, with the progestin being primarily responsible for preventing ovulation.27 DSG is a third generation 19-nortestosterone derivate. Even if the structural differences between the different progestins are minor, it is considered inappropriate to relate the various effects of the different molecules as class effects.37 This could explain different findings of studies in which HC with different progestin component had been used.6–8 Therefore, our findings are relevant to the COC tablets under study—DSG/EE. Nevertheless, our study has some significant advantages in that it was conducted in real-life conditions with HIV-positive women who were on stable treatment providing steady-state levels for both ARVs and COCs.
In conclusion, our results suggest that the coadministration of COC, containing DSG/EE with EFV is associated with risks for contraceptive and EFV failures. On the other hand, the study results suggest that the use of such COC with NVP resulted in more favorable ARV and progesterone levels. As EFV-based ARV is becoming the preferred first-line option worldwide, it is crucial to study in detail the PK interaction between COC and NNRTI and its clinical impact on contraceptive and HIV treatment effects in larger and more diverse populations.
The authors are grateful to the research and clinical staff and clients at the HIV-NAT Clinic and at The Thai Red Cross Anonymous Clinic for their contribution to this study. The team appreciated the support of Theeradej Boonmangum for helping with the data entry and Piraporn June Ohata for helping with the submission of this article.
2. Centers for Disease Control and Prevention (CDC). Update to CDC’s U.S. Medical Eligibility Criteria for Contraceptive Use, 2010. MMWR Morb Mortal Wkly Rep. 2012;61:449–452.
3. Mildvan D, Yarrish R, Marshak A, et al.. Pharmacokinetic interaction between nevirapine and ethinyl estradiol/norethindrone when administered concurrently to HIV-infected women. J Acquir Immune Defic Syndr. 2002;29:471–477.
4. Stuart GS, Moses A, Corbett A, et al.. Combined oral contraceptives and antiretroviral PK/PD in Malawian women: pharmacokinetics and pharmacodynamics of a combined oral contraceptive and a generic combined formulation antiretroviral in Malawi. J Acquir Immune Defic Syndr. 2011;58:e40–e43.
5. Joshi AS, Fiske WD, Benedek IH, et al.. Lack of a pharmacokinetic interaction between efavirenz (DMP 266) and ethinyl estradiol in healthy female volunteers. Conf Retrovir Oppor Infect. 1998;5:abstract no. 348.
6. Carten ML, Kiser JJ, Kwara A, et al.. Pharmacokinetic interactions between the hormonal emergency contraception, levonorgestrel (Plan B), and Efavirenz. Infect Dis Obstet Gynecol. 2012;2012:Article 137192.
7. Sevinsky H, Eley T, Persson A, et al.. The effect of efavirenz on the pharmacokinetics of an oral contraceptive containing ethinyl estradiol and norgestimate in healthy HIV-negative women. Antivir Ther. 2011;16:149–156.
8. Cohn SE, Park JG, Watts DH, et al.. Depo-medroxyprogesterone in women on antiretroviral therapy: effective contraception and lack of clinically significant interactions. Clin Pharmacol Ther. 2007;81:222–227.
9. Nanda K, Amaral E, Hays M, et al.. Pharmacokinetic interactions between depot medroxyprogesterone acetate and combination antiretroviral therapy. Fertil Steril. 2008;90:965–971.
10. Schöller-Gyüre M, Kakuda TN, Woodfall B, et al.. Effect of steady-state etravirine on the pharmacokinetics and pharmacodynamics of ethinylestradiol and norethindrone. Contraception. 2009;80:44–52.
11. Crauwels H, van Heeswijk R, Cornelis L, et al.. Pharmacokinetic Interaction Study between TMC278, an NNRTI, and the Contraceptives Norethindrone Plus Ethinylestradiol. Presented at the 12th European AIDS Conference, November 11–14, 2009, Cologne, Germany.
12. Kearney BP, Mathias A. Lack of effect of tenofovir disoproxilfumarate on pharmacokinetics of hormonal contraceptives. Pharmacotherapy. 2009;29:924–929.
14. Droste JA, Verweij-Van Wissen CP, Burger DM. Simultaneous determination of the HIV drugs indinavir, amprenavir, saquinavir, ritonavir, lopinavir, nelfinavir, the nelfinavir hydroxymetabolite M8, and nevirapine in human plasma by reversed-phase high-performance liquid chromatography. Ther Drug Monit. 2003;25:393–399.
15. Aarnoutse RE, Grintjes KJ, Telgt DS, et al.. The influence of efavirenz on the pharmacokinetics of a twice-daily combination of indinavir and low-dose ritonavir in healthy volunteers. Clin Pharmacol Ther. 2002;71:57–67.
16. Burger D, Teulen M, Eerland J, et al.. The international interlaboratory quality control program for measurement of antiretroviral drugs in plasma: a global proficiency testing program. Ther Drug Monit. 2011;33:239–243.
17. Schwartz JL, Creinin MD, Pymar HC, et al.. Predicting risk of ovulation in new start oral contraceptive users. Obstet Gynecol. 2002;99:177–182.
18. Leth FV, Kappelhoff BS, Johnson D, et al.. Pharmacokinetic parameters of nevirapine and efavirenz in relation to antiretroviral efficacy. AIDS Res Hum Retroviruses. 2006;22:232–239.
19. Warne DW, Tredway D, Schertz JC, et al.. Midluteal serum progesterone levels and pregnancy following ovulation induction with human follicle-stimulating hormone: results of a combined-data analysis. J Reprod Med. 2011;56:31–38.
22. Ford N, Mofenson L, Kranzer K, et al.. Safety of efavirenz in first-trimester of pregnancy: a systematic review and meta-analysis of outcomes from observational cohorts. AIDS. 2010;24:1461–1470.
23. Leticee N, Viard JP, Yamgnane A, et al.. Contraceptive failure of etonogestrel implant in patients treated with antiretrovirals including efavirenz. Contraception. 2012;85:425–427.
24. Matiluko AA, Soundararjan L, Hogston P. Early contraceptive failure of Implanon in an HIV-seropositive patient on triple antiretroviral therapy with zidovudine, lamivudine and efavirenz. J Fam Plann Reprod Health Care. 2007;33:277–278.
25. Manosuthi W, Kiertiburanakul S, Chaovavanich A, et al.. Plasma nevirapine levels and 24-week efficacy of a fixed-dose combination of stavudine, lamivudine and nevirapine (GPO-VIR) among Thai HIV-infected patients. J Med Assoc Thai. 2007;90:244–250.
26. Langmann P, Weissbrich B, Desch S, et al.. Efavirenz plasma levels for the prediction of treatment failure in heavily pretreated HIV-1 infected patients. Eur J Med Res. 2002;7:309–314.
27. Speroff L, Fritz MA. Clinical Gynecologic Endocrinology and Infertility. Philadelphia, PA: Lippincott Williams & Wilkins; 2005:862–940.
31. Wang B, Sanchez RI, Franklin RB, et al.. The involvement of CYP3A4 and CYP2C9 in the metabolism of 17 alpha-ethinylestradiol. Drug Metab Dispos. 2004;32:1209–1212.
32. Liptrott NJ, Pushpakom S, Wyen C, et al.. Association of ABCC10 polymorphisms with nevirapine plasma concentrations in the German Competence Network for HIV/AIDS. Pharmacogenet Genomics. 2012;22:10–19.
33. Weiss J, Herzog M, König S, et al.. Induction of multiple drug transporters by efavirenz. J Pharmacol Sci. 2009;109:242–250.
34. Heil SG, van der Ende ME, Schenk PW, et al.. Associations between ABCB1, CYP2A6, CYP2B6, CYP2D6, and CYP3A5 alleles in relation to efavirenz and nevirapine pharmacokinetics in HIV-infected individuals. Ther Drug Monit. 2012;34:153–159.
35. Beitins IZ, Dufau ML. Pulsatile secretion of progesterone from the human corpus luteum: poor correlation with bioactive LH pulses. Acta Endocrinol (Copenh). 1986;111:553–557.
36. Rossmanith WG, Laughlin GA, Mortola JF, et al.. Secretory dynamics of oestradiol (E2) and progesterone (P4) during periods of relative pituitary LH quiescence in the midluteal phase of the menstrual cycle. Clin Endocrinol (Oxf). 1990;32:13–23.
37. Sitruk-Ware R. Pharmacological profile of progestins. Maturitas. 2004;47:277–283.