A Pharmacokinetic and Pharmacogenetic Evaluation Of

Home , Norgestrel

1995 , d Global g Center for i , , b M , a,b,c h 0.001] < P 0.73]. Pos- ¼ P , Christina J. Kim Department of Pharmaceutics, f d , Jared M. Baeten Department of Pharmacy Practice and Science, j 0.006) compared with the control , Athena P. Kourtis , Katherine K. Thomas b,i Division of Reproductive Health, National Center for g ¼ c h P 1995–2004 : 33 2019 Wolters Kluwer Health, Inc. All rights reserved. Department of Epidemiology, c ß , David W. Erikson f 2019, 0.64; etonogestrel 1.07 (0.77, 1.50); ¼ P Copyright Department of Clinical Pharmacy, University of California, San Francisco, San , Nelly Mugo AIDS e , for the Partners PrEP Study Team 2019 Wolters Kluwer Health, Inc. All rights reserved. j , Kavita Nanda a,b Q d , Randy M. Stalter a,b Concomitant use of EFV significantly reduces levonorgestrel or etonoges- To evaluate pharmacokinetics and pharmacogenetics of contraceptive Plasma samples collected every 6 months were analyzed for levonorgestrel Department of Global Health, Our analysis included 11 women who initiated EFV, 13 who initiated NVP, We analyzed stored samples from women self-reporting implant use in the b , Steven W. Blue e Endocrine Technologies Core, Oregon National Primate Research Center, Beaverton, Oregon, f among women in Kenya and Uganda Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. and the etonogestrel aGMR was 0.51 (0.34, 0.76; session of four alleleconcentrations variants among were those receiving found EFV. toConclusion: result in furthertrel reductions concentrations in by 61 progestin andreport 49%, allelic respectively, compared variants with indrug-interaction no hepatic between ART use. enzymes progestins We that and also EFV. influenced the extent of the observed and 36 who remained ART-naive.(aGMR) In the of EFV levonorgestrel group, was the 0.39 adjusted [90% geometric confidence mean intervals ratio (0.31,group. 0.49); No difference wasnorgestrel observed 0.93 in (0.74, the 1.18); NVP group compared with controls [levo- containing or nevirapine (NVP)-containing antiretroviralDesign: therapy (ART). Partners PrEP Study. Methods: and etonogestrel concentrations. Progestin concentrations fromART samples initiation collected after were comparedparisons. with We used pre-ART adjusted concentrations linearbetween for mixed individuals intraindividual models on to EFV com- compare andpossessing hormone NVP certain concentrations alleles to with a known nometabolism or ART possible were group. influences associated We on with then EFV, changesassociation NVP, evaluated in or whether between progestin progestin ART concentrations or use modified andResults: the progestin concentrations. Objectives: implant progestin concentrations in HIV-positive women initiating efavirenz (EFV)- Edward J. Kelly Kimberly K. Scarsi ISSN 0269-9370 Copyright Jairam R. Lingappa Rena C. Patel of contraceptive implants and antiretroviral therapy Members of the study team listed after the Acknowledgements. A pharmacokinetic and pharmacogenetic evaluation Department of Medicine, Bani Tamraz Received: 27 March 2019; revised: 28 MayDOI:10.1097/QAD.0000000000002308 2019; accepted: 31 May 2019. a Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Georgia, USA, Francisco, California, University of Nebraska MedicalCorrespondence Center, to Omaha, Rena Nebraska, C. USA. Patel,Tel: UW +1 Box 206 520 359927, 3800; 325 Ninth e-mail: [email protected] Avenue, Seattle, WA 98104, USA. Health, Population, and Nutrition, FHI 360, Durham, North Carolina, University of Washington, Seattle, Washington, Clinical Research, Kenya Medical Research Institute, Nairobi, Kenya, and

Downloaded from https://journals.lww.com/aidsonline by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3Dkys9+3wZnLBAQwV9m967KYSyNLNJCeWjq0jy3w0N9e13Ci0yNfjvg== on 05/06/2020 Downloaded from https://journals.lww.com/aidsonline by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3Dkys9+3wZnLBAQwV9m967KYSyNLNJCeWjq0jy3w0N9e13Ci0yNfjvg== on 05/06/2020 1996 AIDS 2019, Vol 33 No 13

Keywords: antiretroviral therapy, efavirenz, etonogestrel, hormonal contraception, implants, levonorgestrel, pharmacogenetic, pharmacokinetic

Introduction within each woman (intrasubject comparisons) as well as comparing the changes in concentrations across The majority of people living with HIVare women, with groups by ART category (intersubject comparisons). an estimated 13 million HIV-positive women of repro- Our second objective was to characterize pharmacoge- ductive age living in sub-Saharan Africa [1]. Nearly two- netic polymorphisms associated with progestin concen- thirds of pregnancies among HIV-positive women in sub- tration changes among these women, and whether any Saharan Africa are unintended [2], and effective contra- pharmacogenetic polymorphisms modify the effect of ception can prevent unintended pregnancies, reducing ARTon hormone concentrations. maternal mortality and perinatal HIV transmission [3,4]. Contraceptive implants, containing the progestins levonorgestrel or etonogestrel, are the most effective contraceptives available and are increasingly being used by HIV- Methods positive women in sub-Saharan Africa [5]. Study population The WHO recommends initiation of antiretroviral therapy We used data from HIV-positive women less than 50 years (ART) for HIV-positive individuals, and dolutegravir- of age enrolled in the Partners PrEP Study in Kenya and containing ART is recommended as a first-line regimen Uganda from 2008 to 2010 who self-reported implant [6]. Enthusiasm for dolutegravir’s wide-spread rollout in use, some of whom went on to initiate ART later. resource-limited settings has been hampered by its possible Detailed enrollment and follow-up procedures for link with neural tube defects [7]. Therefore, efavirenz participants are described elsewhere [15–17]. Women (EFV)-containing ART remains the primary regimen for not eligible for ART initiation at the time of Partners women of reproductive age in resource-limited settings, PrEP Study enrollment had their clinical and immuno- including those on contraceptive implants. However, due logical status monitored at study visits. If participants to drug–drug interactions resulting from induction of became eligible to initiate ART according to national hepatic cytochrome P450 (CYP450) enzymes, the guidelines during follow-up, they were referred to HIV concomitant use of EFV reduces the effectiveness of facilities for ART initiation. Contraception was not contraceptive implants. In a prior study among Kenyan provided by the study; instead, women were referred to women, those using implants and EFV-containing ART routine clinical care if desiring contraception. Women faced three-fold higher rates of pregnancies than those self-reported implant use and the implant insertion dates using nevirapine (NVP)-containing ART [8], a result were not available. Women self-reporting the use of other supported by pharmacokinetic data showing reduced hormonal methods, intrauterine device, or surgical implant progestin concentrations with concomitant EFV methods, or those without documented ART status use [9–12]. The existing pharmacokinetic evaluations are were excluded. limited by large intersubject variability. Furthermore, though both levonorgestrel and etonogestrel are believed We analyzed 6-monthly plasma samples, which were to be metabolized primarily by CYP3A4, the mechanisms collected regardless of menstrual cycles, for each woman. that underlie metabolism of progestins are poorly For women initiating ART, we included women with at understood. Genetic polymorphisms in hepatic enzymes least one sample available for analysis pre or post-ART. and transporters, which often vary by ethnic group, may be Five women in the EFV group had one sample either pre an important factor influencing progestin metabolism, or post-ART initiation and the remainder of the women extent of drug–drug interactions, and subsequent con- had at least two samples either pre or post-ART. To be traceptive efficacy [13,14]. included in this analysis, we required that the plasma samples were within 6 months of each other and the date Pharmacokinetic studies that address intersubject vari- of ART initiation, when applicable. To help account for ability by including within-subject sampling and the expected decreases in progestin concentrations over pharmacogenetic analyses with ethnic groups tradition- time [14,18], we included samples from women who did ally less represented in such analyses are needed to fill not initiate ART but had samples available from at least six knowledge gaps regarding concomitant implant and study visits to represent levonorgestrel or etonogestrel EFV use. Our first objective was to conduct a exposure over time in the absence of ART. pharmacokinetic evaluation studying the effect of initiating EFV-containing or NVP-containing ART on One woman in the Partners PrEP Study became pregnant levonorgestrel and etonogestrel plasma concentrations, while concomitantly using an implant and EFV-contain- comparing concentrations pre and post-ART initiation ing ART. Because she initiated the implant after EFV, we

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. Contraceptive implants and antiretroviral therapy Patel et al. 1997 could not include her in the pharmacokinetic or Primary outcomes: progestin concentrations in pharmacogenetic evaluations. Nonetheless, we analyzed plasma all progestin plasma samples collected for this woman and We simultaneously quantified plasma concentrations of descriptively present these findings. levonorgestrel, etonogestrel, and endogenous progesterone using a validated, high-performance liquid chroma- Antiretroviral therapy classification tography-heated electrospray ionization–tandem triple We classified ART regimen use at the time of sample quadrupole mass spectrometry assay [21]. The lower limit collection into three categories: EFV-containing ART, of quantification (LLQ) for levonorgestrel and etonogestrel NVP-containing ART, or no ART. The no ART group were both 0.02 ng/ml and endogenous progesterone was included all samples from ART-naive controls as well as 0.01 ng/ml. When both levonorgestrel and etonogestrel pre-ART initiation samples for women who initiated concentrations above the LLQ were detected for a sample, ART. The EFV-containing and NVP-containing ART and the levonorgestrel concentrations were markedly groups consisted of samples collected after initiation of higher than concentrations consistent with implant use the respective ART regimens. We defined an ART (possibly due to oral levonorgestrel-containing contracep- regimen as at least a three-drug combination of tive use) [14], the levonorgestrel values were excluded (16 antiretrovirals, and monotherapy, for instance, for samples from 10 women). One woman had several prevention of mother-to-child transmission was not consecutive levonorgestrel concentrations quantified and considered ‘initiation of ART.’ ART initiation dates were then a single etonogestrel concentration detected; the self-reported and not verified against their clinical care etonogestrel observation was excluded from analysis. One records. Date of ART initiation was imputed as the woman had both levonorgestrel and etonogestrel con- midpoint of the preceding interval before ART use was centrations detected in the same samples consistent with reported or, when available, using the self-reported either implant use; all observations were excluded from number of days of ART use. For the purposes of analysis as it was not possible to determine which implant characterizing intrasubject comparisons in the no ART was being used. For endogenous progesterone concentra- group, the time of the median sample collection for each tions, we quantified the number of samples with ART-unexposed woman was used to split her observa- concentrations above both a lower cutoff of 3 ng/ml and tions into ‘pre-ART’ and ‘post-ART’ periods a higher cutoff of 5 ng/ml; endogenous progesterone for comparison. concentrations above 3 ng/ml are associated with recent ovulation but controversy remains regarding the exact Candidate gene selection cutoff most predictive of ovulation [22]. For the pharmacogenetic evaluation, the following 18 alleles were chosen for analysis based on known or Statistical analysis possible influences on EFV, NVP, or implant progestin We used frequencies (and percentages) and median [and metabolism [19,20]: CYP2B6: rs4803419 (15582C>T), interquartile ranges (IQR)] to describe categorical and rs3745274 (516G>T); ABCB1: rs3842 (4046A>G), continuous variables, respectively. Etonogestrel and rs1045642 (3435C>T), rs2032582 (2677G>A/T); levonorgestrel concentrations were log-transformed for CYP2B6: rs28399499 (983T>C); CYP3A5: rs776746 analyses. For the pharmacokinetic evaluations, we used (3 allele, 6986A>G), rs10264272 (6 allele, 14690G> the observed geometric means and ratios (GMR) to A), rs41303343 (7 27131insT); CYP3A4: rs2740574 summarize progestin concentrations pre vs. post-ART (392A>G), rs35599367 (22 allele, 15497C>T); (intrasubject comparisons). We used linear mixed models CYP2A6: rs8192726 (9B allele, 1836G>T), to compare hormone concentrations for observations rs28399433 (48A>C); NR1I2: rs1523130 (44477A> post-ART initiation in the EFV and NVP groups to the G), rs2472677 (63396C>T); NR1I3 rs2307424 no ART group (intersubject comparisons). Beta coeffi- (540C>T), and rs3003596 (1089T>C). cients were exponentiated to obtain GMRs of hormone levels and 90% confidence intervals for the post-ART vs. Single nucleotide polymorphism genotyping no ARTobservations. All models were adjusted for days We extracted genomic DNA from 100 ml of whole blood from ART initiation, and multivariate models were also using a Qiagen (Germantown, Maryland, USA) adjusted for a priori selected covariates of age, BMI, DNAeasy blood and tissue kit. We then genotyped the nationality (Kenyan or Uganda) and HIV viral load DNA samples using Applied Biosystems (Foster City, measured closest to analyzed sample, as a marker for California, USA) Taqman genotyping assays according to antiretroviral adherence. the manufacturer’s instructions. PCR amplification was done using a BioRad (Hercules, California, USA) CFX For the pharmacogenetic evaluations, of our 18 candidate Connect Real-Time PCR Detection System and allelic alleles, two alleles were excluded from the analysis. One discrimination determined by postrun analysis using CFX allele (s35599367) was monomorphic with all participants Manager 3.1 software. For quality control, we ran samples with C/C expression. Two alleles in CYP2A6 from the UW School of Pharmacy human liver bank of (rs28399433 and rs8192726) were in perfect linkage known genotypes for each assay. disequilibrium (r2 ¼ 1), so rs8192726 was excluded from

analysis. One allele was not in Hardy–Weinberg main effects pharmacogenetic models, we report the equilibrium at a threshold of P less than 0.0001 GMRs of progestin concentrations among women with (rs3745274, P ¼ 0.0001); however, since this allele is each additional variant copy of the allele, obtained by one of the best characterized alleles regarding progestin exponentiating the beta coefficient for the allele variant and ART drug–drug interactions, we chose to report the variable, and the corresponding P values. For the findings. We calculated minor allele frequencies (MAF) interaction pharmacogenetic models, we report the fold by nationality (Kenya vs. Uganda), and used a Chi-square change in progestin GMRs between women on ARTand test to assess for differences in the MAFs. All P values were women not on ART with each additional copy of the greater than 0.05, indicating that any potential bias arising variant allele, obtained by exponentiating the beta from population structure for these variants by nationality coefficient of the interaction term, and the corresponding would not be significant, so we did not conduct stratified P values. analysis by nationality (data not shown). We ran separate, additive linear mixed models for each of the 16 alleles in For pharmacokinetic and pharmacogenetic linear mixed our pharmacogenetic evaluations. Due to reduced sample models, we selected a spatial autoregression correlation sizes when restricting analyzes by implant type and allelic structure based on Akaike information criterion. The variants, we combined women using either implant by spatial correlation structure accounts for correlation conducting a z-score transformation of their hormone between women’s repeated measures and is ideal for data concentrations, so that the values were ‘normalized’ for like ours where spacing of repeated measures differs the distribution of concentrations for each hormone type. between individuals. We conducted all analyses with SAS We then tested for an association of number of allele version 9.4 statistical software (SAS Institute Inc., Cary, variant copies (0,1,2) with changes in hormone North Carolina, USA). concentrations, adjusting for ART category (main effects model). Next, we tested whether each allelic variant modified the association between ARTuse and hormone concentration changes by adding an interaction term for Results number of variant allele copies and ART category (interaction model). In addition to the adjustment This analysis included 11 women who initiated EFV, 13 variables described for the pharmacokinetic models, who initiated NVP, and 36 who remained ART-naı¨ve we also adjusted pharmacogenetic models for implant during follow-up (Table 1). In the EFV group, two type. We used a Bonferroni adjustment to determine the women had only one pre-ART initiation sample, one significance threshold of 0.05/16 ¼ 0.003 for the multiple woman had only one post-ART initiation sample, and comparisons in the pharmacogenetic analysis. For the one woman had only one pre and one post-ART

Table 1. Baseline characteristics of HIV-infected women on contraceptive implants by antiretroviral therapy category. Characteristic Efavirenz, n ¼ 11 Nevirapine, n ¼ 13 ART-naive, n ¼ 36

n (%) or median (IQR) Age (years) 34 (31, 36) 32 (31, 35) 33 (30, 37) Parity 3 (2, 3) 4 (1, 4) 4 (3, 5) Married 11 (100%) 12 (92%) 36 (100%) Educational status (years in school) 7 (3, 12) 7 (4, 8) 4 (3, 7) Ethnicity Kikuyo 3 (27%) 1 (8%) 5 (14%) Iteso 2 (18%) 5 (38%) 14 (39%) Luo 1 (9%) 3 (23%) 4 (11%) Other 5 (45%) 4 (31%) 13 (36%) Country Kenya 7 (63%) 7 (56%) 13 (37%) Uganda 4 (37%) 6 (44%) 23 (63%) BMI (kg/m2) 22.8 (21.8, 25.7) 22.5 (21.5, 24.0) 22.2 (21.5, 24.5)

Samples Pre-ART Post-ART Pre-ART Post-ART Pre-ARTa Post-ARTa

Total number of samples 27 22 34 37 121 104 contributing to analysis Number of days since ART 239 (630, 36) 209 (40, 658) 210 (630, 42) 236 (41, 799) 252 (732, 38) 252 (0, 566) initiation, median (min, max) Number of samples with viral 11/27 (41%) 17/22 (77%) 8/34 (24%) 36/37 (97%) 30/121 (25%) 34/104 (33%) suppression (<1000 copies/ml)

ART, antiretroviral therapy; IQR, interquartile range. aThe median time for each woman in the no ART group was used to split the observations into ‘pre-ART’ and ‘post-ART’ so as to facilitate characterizing changes over time (i.e., intrasubject comparisons) in this group.

Table 2. Pharmacokinetic results by antiretroviral therapy category and implant type.

Unadjusteda Adjustedb

No. of Pre-ART GM hormone Post-ART Observed Predicted ART: Predicted ART: ART women concentrations GM hormone pre-ART: Post- no ART GMRe no ART GMRe category (samples) (pg/ml)c concentrations (pg/ml)c ART GMRd (90% CI) P value (90% CI) P value

Levonorgestrel ART-naive 14 (85) 441.95 402.40 0.91 Ref – Ref – EFV 8 (33) 333.40 89.91 0.27 0.33 (0.27, 0.41) <0.001 0.39 (0.31, 0.49) <0.001 NVP 6 (36) 460.01 330.23 0.72 0.84 (0.67, 1.04) 0.17 0.93 (0.74. 1,18) 0.64 Etonogestrel ART-naive 22 (140) 255.74 204.79 0.80 Ref – Ref – EFV 3 (16) 218.38 132.91 0.61 0.57 (0.38, 0.84) 0.02 0.51 (0.34, 0.76) 0.006 NVP 7 (36) 305.05 218.82 0.72 0.88 (0.67, 1.15) 0.42 1.07 (0.77, 1.50) 0.73

ART, antiretroviral therapy; CI, confidence interval; GM, geometric mean; GMR, geometric mean ratio; EFV, efavirenz; NVP, nevirapine. aAdjusted only for days from ART initiation. bAdjusted for days from ART initiation, age, BMI, nationality, and closest viral load. cThe median time for each ART-naive woman was used to split the observations into ‘pre-ART’ and ‘post-ART’ so as to facilitate intrasubject comparisons in this group. dThe observed GMR is intrasubject, calculated as the ratio between geometric means of observed pre and post-ART initiation samples. eThe no ART (reference) category GMRs from models include all samples from ART-naive women as well as pre-ART samples from those who initiated ART during study, to better estimate changes due to initiating ART while accounting for background change in concentrations over time. Thus, these models incorporate information from within women and well as between women. initiation sample available for analysis. All women in the CYP3A53 6986A>G, and CYP3A57 27131insT. No NVP and the ART-naive groups had at least two pre and allele variants were significantly associated with differ- post-ART initiation samples available for analysis. The ences in the progestin GMRs in the NVP group relative median age of all included women was 33 (IQR 30– to the no ART group. 36.5). Parity, marital status, years of formal schooling, nationality, BMI, and follow-up time in this analysis were comparable between the two ART groups and varied Endogenous progesterone evaluation somewhat from the ART-naive group. Among women using levonorgestrel implants, using either cutoff of 3 or 5 ng/ml, high endogenous Pharmacokinetic evaluation progesterone concentrations were detected in two The geometric means for levonorgestrel and etonogestrel (14%), one (5%), and one (1%) samples in the EFV, concentrations in the pre and post-ART samples per NVP, and no ART groups, respectively (Fig. 1). Among woman (i.e., intrasubject comparisons) are described in women using etonogestrel implants, using a lower cutoff Table 2. The adjusted progestin GMRs comparing EFV of 3 ng/ml, high endogenous progesterone concentra- and the NVP groups to the no ART group (i.e., tions were detected in three (38%), zero, and one (1%) intersubject comparisons) were 0.39 [(0.31, 0.49); samples in the EFV, NVP, and no ART groups, P < 0.001] and 0.93 [(0.74, 1.18); P ¼ 0.64], respectively. respectively; and using a higher cutoff of 5 ng/ml, high The adjusted progestin GMRs of etonogestrel comparing endogenous progesterone concentrations were detected the EFV and NVP groups to the no ART group (i.e. in two (25%) samples in the EFV group only. intersubject comparisons) were 0.51 [(0.34, 0.76); P ¼ 0.006] and 1.07 [(0.77, 1.50); P ¼ 0.73], respectively. Pregnancy and concentrations Pharmacogenetic evaluation The one woman who became pregnant while concomi- Of the allele variants we evaluated for changes in tantly using an implant and EFV-containing ART had the levonorgestrel or etonogestrel concentrations in the main following timeline. She first reported initiation of EFV- effects models, no alleles met our significance threshold containing ART on 2 March 2011, then first reported but the number of variant copies for one allele, CYP2B6 implant use on 30 June 2011. One sample was available 15582C>T, was associated with a 48% increase in for analysis from 22 September 2011, approximately 3 progestin concentrations after adjusting for ART use months after her first report of implant use, and showed a (P ¼ 0.03; Table 3). In our interaction models, only one levonorgestrel concentration of 140 pg/ml and an allele, ABCB1 3435C>T(P < 0.001), met our signifi- endogenous progesterone concentration of more than cance threshold and was associated with 88% greater 10.0 ng/ml; her second sample available from 14 March reduction in progestin GMRs in the EFV group relative 2012, showed a levonorgestrel concentration of 180 pg/ to the no ART group. Nonetheless, the following variants ml and an endogenous progesterone concentration of were associated with 64–85% greater reductions in 3.3 ng/ml. On 16 April 2012, her urine test was positive progestin GMRs in the EFV group relative to the no for a pregnancy and on 24 April 2012, she experienced a ART group at P less than 0.05: CYP2B6 983T>C, pregnancy loss.

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. 2000 AIDS 2019, Vol 33 No 13 d value P Interaction copy (90% CI) Progestin GMR additional variant fold difference per 12 women ¼ Nevirapine, n 12 ¼ MAF, n e, all women from the sample are ith each additional variant allele copy. with 0 (0) 0.00 – – 2 (4) 0.08 1.12 (0.50, 2.52) 0.81 4 (9)0 (0) 0.17 1.70 (0.91, 3.20) 0.00 0.16 – – variant (samples) No. of women b erapy; CI, confidence interval; MAF, minor allele change in progestin GMR of efavirenz/nevirapine d 0.001 value 0.02 0.004 0.007 < P Interaction Interaction PG models difference per copy (90% CI) 11 women additional variant Progestin GMR fold Efavirenz, ¼ n 11 ¼ MAF, n No. of women (samples) with variant values less than 0.05. P 59 ¼ 59 MAF, n ¼ n women 9 (42) 0.08 2 (6) 0.14 1.47 (0.85, 2.52) 0.24 1 (5) 0.04 0.68 (0.23, 2.01) 0.56 0.0001). No. of No ART, women (samples) ¼ with variant P d value 0.03 P 0.003, but italicize here all a ¼ P less than 0.001 ( P Progestin GMR (per copy) (90% CI) additional variant 59 women ¼ All women, n 59 ¼ MAF, n Main effect PG models No of women 39 (226) 0.33 0.80 (0.61, 1.06) 0.06 39 (185) 0.33 8 (17) 0.36 0.65 (0.30, 1.42) 0.36 8 (24) 0.33 0.88 (0.45, 1.71) 0.75 (samples) with variant e G (rs3842) 24 (131) 0.25 1.19 (0.96, 1.47) 0.19 24 (112) 0.25 4 (6) 0.18 0.78 (0.37, 1.61) 0.57 6 (13) 0.29 0.86 (0.53, 1.39) 0.60 T (rs4803419) 9 (53) 0.08 1.48 (1.11, 1.99) T (rs2472677) 44 (245) 0.42 1.08 (0.86, 1.36) 0.59 44 (199) 0.42 9 (18) 0.45 0.61 (0.36, 1.03) 0.12 10 (28) 0.46 0.62 (0.17, 2.20) 0.53 G (rs776746) 16 (88)G (rs1523130) 0.15 14 (78) 1.26 (0.97, 1.63) 0.12 0.14 0.83 (0.61, 1.12) 16 (73) 0.30 0.15 14 (63) 3 (6) 0.12 0.14 5 (8) 0.36 (0.17, 0.76) 0.23 1.79 (0.88, 3.62) 0.18 2 (7) 0.08 0.94 (0.42, 2.14) 0.91 T (rs8192726) 8 (47) 0.08 0.98 (0.71, 1.35) 0.90 8 (40) 0.08 2 (6) 0.14 0.61 (0.36, 1.03) 0.12 1 (1) 0.04 0.62 (0.17, 2.20) 0.53 T (rs1045642) 7 (39) 0.06 1.12 (0.74, 1.69) 0.65 7 (31) 0.06 2 (4) 0.09 0.12 (0.05, 0.29) A (rs10264272) 26 (154) 0.28 1.00 (0.83, 1.20) 0.99 26 (119) 0.28 5 (11) 0.32 1.63 (1.05, 2.53) 0.07 7 (24) 0.33 0.63 (0.39, 1.02) 0.11 G (rs2740574) 56 (323) 0.77 1.11 (0.89, 1.38) 0.45 56 (268) 0.77 10 (20) 0.77 0.59 (0.33, 1.04) 0.12 12 (35) 0.71 0.68 (0.36, 1.27) 0.30 > T (rs2032582t) 6 (33) 0.05 1.03 (0.66, 1.59) 0.92 6 (25) 0.05 1 (4) 0.05 2.72 (0.85, 8.75) 0.16 2 (4) 0.08 1.34 (0.58, 3.10) 0.57 C (rs3003596) 46 (267) 0.52 0.93 (0.77, 1.13) 0.55 46 (221) 0.52 11 (16) 0.55 0.76 (0.49, 1.18) 0.30 12 (30) 0.50 0.97 (0.62, 1.51) 0.90 A (rs2032582a) 0 (0) 0.00 – – 0 (0) 0.00 0 (0) 0.00 – – 0 (0) 0.00 – – T (rs2307424) 13 (76) 0.11 1.18 (0.86, 1.62) 0.39 13 (56) 0.11 2 (5) 0.09 2.49 (1.06, 5.86) 0.08 4 (15) 0.17 1.15 (0.59, 2.24) 0.73 Variant T (rs3745274) > > > > C (rs28399499) 7 (43) 0.07 1.14 (0.81, 1.60) 0.52 7 (40) 0.07 1 (3) 0.05 0.15 (0.05, 0.43) > > > > > > > > SNP (rsID) > > 4046A 392A 540C 6986A 516G 3435C 983T 1089T 1836G 2677G 15582C 63396C 2677G 44477A 27131insT (rs41303343) 16 (100) 0.14 0.82 (0.61, 1.10) 0.26 16 (95) 0.14 2 (5) 0.09 0.26 (0.12, 0.59) 14690G Assesses effect modification of the association between ART category and progestin concentrations by variant allele; model estimates indicate fold We used a Bonferroni adjustment of significance threshold of Assesses the association of the variant allele with progestin concentrations independentIncludes of all ART samples category; from model women estimates who indicate did progestin not GMR initiate w ARTAllele and exceeds pre-ART Hardy–Weinberg samples from equilibrium women threshold who of initiated regimens with efavirenz or nevirapine; therefor Gene ABCB1 represented. Table 3. Effect modification of the association between antiretroviral therapy use and progestin concentrations by possession of variant alleles. CYP2B6 Models adjusted for ART category,frequency; days from GMR, ART geometrica initiation, mean implant type, ratio; BMI, SNP, logb of single closest nucleotide viral polymorphism; load measure, rsID, and reference country. SNP ART, antiretroviral ID;c th PG, pharmacogentics. d e CYP2A6 CYP3A4 CYP3A5 NR1I2 (PXR) vs. no ART groups with each additional variant copy. NR1I3

Fig. 1. Endogenous progesterone concentrations by antiretroviral therapy category and implant type.

Discussion implants faced nearly three times higher rates of pregnancies than women using NVP with implants [8]. Utilizing pre and post-ART plasma samples among However, given the overall effectiveness of the implants women on contraceptive implants, we demonstrate that (<1% failure rate) [23], the absolute rates of pregnancy concomitant use of EFV significantly reduces levonor- with EFV and implants were still lower than those faced gestrel and etonogestrel concentrations by 61 and 49%, by women using EFV and most other forms of respectively, as compared with women not yet on ART. contraception (e.g. 3.3 pregnancies/100 women-years Wealso report on a new finding of a greater array of allelic of implant use vs. 5.4 of depot medroxyprogesterone variants in hepatic enzymes associated with these acetate use, both reported with concomitant EFVuse) [8]. reduced concentrations. Current HIV and contraceptive guidelines recommend caution regarding implant use with EFV. However, with Our pharmacokinetic findings demonstrating a signifi- multiple pharmacokinetic and clinical studies consistently cant reduction of hormone concentrations with EFV demonstrating reduced concentrations or effectiveness of initiation are similar to prospective, parallel group implant hormones when used with EFV, national and comparisons. Scarsi et al. [10] reported a 57% reduction international authorities should consider recommending in levonorgestrel concentrations among women on EFV offering an alternative ARTregimen – where available – vs. no ART at 48 weeks after levonorgestrel implant to these women. insertion. Three of twenty (15%) women became pregnant between 36 and 48 weeks after implant We report on novel interactions showing greater insertion. Among women already on EFV-containing reduction in progestin concentrations with EFV use ART and initiating etonogestrel implants, Vieira and associated with allelic variations in ABCB1, CYP2B6, colleagues noted a 2.9-fold reduction in the etonogestrel CYP3A53,andCYP3A57. Although prior studies area under the concentration curve over the first 24 weeks have identified genetic polymorphisms in CYP450 of implant use compared with the no ART group [9]. enzymes that may contribute to differences in con- Collectively, these studies and our pharmacokinetic data traceptive pharmacokinetics [19,20], ours is the first support the findings of the largest cohort study to date study to identify an association between single nucleotide examining the clinical outcome of pregnancies, which polymorphism (SNPs) in ABCB1-encoding genes and demonstrated that women concomitantly using EFV and larger reductions in progestin concentrations among

EFV compared with ART-naive implant users. This study of implants and EFV [9]. Our finding of finding is surprising given that ABCB1 encodes the P- endogenous progesterone concentrations consistent with glycoprotein drug efflux pump, an ATP-binding cassette luteal activity further supports the hypothesis that implant (ABC) membrane protein that plays a role in pumping efficacy is reduced with concomitant EFV use due to drugs out of the cell, of which EFV is not a known lower circulating progestin concentrations. The one case substrate [24]. A previous pharmacogenetic study we analyzed where a pregnancy occurred with concomi- evaluated the effects of the ABCB1 3435C>Tvariant tant implant and EFV use supports this hypothesis; the on ART drug concentrations and clinical outcomes woman had high endogenous progesterone concentra- among white individuals living with HIVand found that tions consistent with luteal activity, coupled with patients who possessed the TT genotype had lower EFV relatively low levonorgestrel concentrations, 1–6 months plasma concentrations relative to those with TC and CC before she became pregnant. genotypes, which does not align with our findings given established pharmacokinetic interactions between EFV HIV treatment guidelines in resource-rich settings now and implant progestins [25]. Another study showed an include integrase inhibitor-containing ART, particularly association between possession of the ABCB1 3435C>T with dolutegravir, as first-line ART regimens [27,28]. variant and slower recoveryof CD4þ cell count after EFV Similarly, dolutegravir-containing ART is now recom- initiation among 50 individuals living with HIV in mended as first-line globally, but with the significant Belgium, but no association was observed between the exception for women of reproductive age due to the SNP and EFV pharmacokinetic parameters. This study, potential risk of neural tube defects [6]. Limited data however, did observe an association between another suggest that dolutegravir does not significantly influence ABCB1 polymorphism (3842T>C) and a higher EFV contraceptive progestin concentrations [29], and dolute- accumulation ratio [26]. Thus, the implications of gravir’s concomitant use with hormonal contraceptives is ABCB1 polymorphism, antiretroviral metabolism, and supported by relevant guidelines [6,30]. Although the any drug–drug interactions requires further exploration. rollout for dolutegravir has been paused for women of As for CYP enzymes, Neary et al. [19] have previously reproductive age, offering women who are already on reported that Ugandan women who possessed the implants and EFV the choice to switch to dolutegravir- variants for CYP2B6 516G>T, resulting in slow EFV containing ART is a potential opportunity to avoid metabolism, had lower levonorgestrel concentrations unintended pregnancies with EFV. Similarly, women on when combined with EFV than those without these hormonal contraception should not be denied access to polymorphisms. We did not find a significant interaction dolutegravir, as these women are highly unlikely to between CYP2B6 516G>Tand ARTuse in our analysis, become pregnant due to the implants’ high effectiveness. but did see reduced plasma progestin concentrations with EFV use among women possessing another CYP2B6 Although our pharmacokinetic and pharmacogenetic variant (983T>C). Ultimately, given our small sample evaluations made opportune use of existing samples and is size, caution is advised in interpreting these pharmaco- strengthened by longitudinal intrasubject comparisons genetic findings as definitive. Of note, we find varying pre and post-ART for each woman, several limitations results in the main effects and the interaction models, exist to our work. First, implant insertion dates are not which is anticipated given that the main effects model known for the women limiting our intersubject examines the relationship between allele variations and comparisons, although the comparisons pre and post hormone concentrations after adjustment for ART use, ART initiation are valid within each woman. Second, the while the effect modifier model examines whether a number of samples and the total duration of time each variant allele modifies the association between ARTuse woman contributes varies both within and across ART and hormone concentrations. Further studies with a groups; our adjusting for days from ART initiation and more diverse and larger population could allow us to use of a correlation structure that accounts for spatial determine whether women with these specific poly- autocorrelation should help mitigate any marked morphisms are indeed at higher risk of drug–drug differences between the groups. Intersubject and intra- interactions between implants and ART. subject variability can be better characterized by future studies with consistent sampling. Third, sporadic plasma Implants are believed to prevent pregnancies by gradually collections, irrespective of menstrual cycles, pose marked releasing progestins into the serum, thereby suppressing challenges in interpretation of endogenous progesterone ovulation, increasing cervical mucus viscosity, and concentrations; however, more resource-intensive possibly altering the endometrium. Although our approaches to detecting ovulation, such as frequent numbers are small, that higher proportions of women ultrasounds, are less feasible, particularly in sub-Saharan in the EFV group had high endogenous progesterone Africa. Finally, the multiple testing in our pharmacoge- concentrations, associated with luteal activity, suggests netic analyses may increase the likelihood of Type I error, that the efficacy of the implants via ovulation suppression and while we made a Bonferroni adjustment, a may be compromised with concomitant EFV use. A significance threshold of P value of 0.003 may be similar finding was reported in another pharmacokinetic too conservative.

Conclusion Contributions by each author: R.C.P. conceived and Our study demonstrates that among women already on designed the study with overall guidance from K.K.S. contraceptive implants, concomitant use of EFV signifi- R.M.S. conducted the analyses with guidance from cantly reduces levonorgestrel or etonogestrel concentra- R.C.P., K.K.T., and B.T. S.W.B. and D.W.E. quantified tions and could increase the risk of ovulation and the progestin concentrations. C.J.K. and E.J.K. geno- subsequent pregnancy. We also discovered a greater array typed the alleles. J.R.L. and J.M.B. helped obtain the of allelic variants in hepatic enzymes and membrane funding to conduct the analyses. The remainder of the transporters that modify the association between ARTuse authors provided guidance on the analyses as needed. All and serum progestin concentrations. With multiple authors reviewed the article and approve of pharmacokinetic and clinical studies consistently dem- its publication. onstrating reduced hormone concentrations or effectiveness of implants when used concomitantly with EFV, US Centers for Disease Control and Prevention (CDC, national and international authorities could recommend U48 DP 005013 SIP 14-023), the US National Institutes offering alternative but superior ART regimens, such as of Health (NIH, NICHD R21HD074439 and NIAID dolutegravir-containing ART, to women living with HIV P30-AI-027757 and UM1-AI-106701, OD011092), and already using or desiring contraceptive implants. the Bill & Melinda Gates Foundation (OPP47674) supported the study. R.C.P. was supported by the NIH National Institute of Allergy and Infectious Diseases (K23AI120855). K.K.S. was supported by the Eunice Acknowledgements Kennedy Shriver National Institute of Child Health and Human Development (R01HD085887). The findings We thank all the women and the clinical trial site teams and conclusions in this report are those of the authors and who participated in the Partners PrEP Study. do not necessarily represent the official position of the funding agencies, the Centers for Disease Control and Partners PrEP Study Team Prevention, or the United States Government.

University of Washington Coordinating Center and Central Laboratories, Seattle, USA: Connie Celum Conﬂicts of interest (principal investigator, protocol cochair), Jared M. Baeten There are no conﬂicts of interest. (medical director, protocol cochair), Deborah Donnell (protocol statistician), Robert W. Coombs, Jairam R. Lingappa, M. Juliana McElrath. References

1. UNAIDS. Global report: UNAIDS report on the global AIDS Study sites and site principal investigators: Eldoret, Kenya epidemic 2013. Geneva, Switzerland: UNAIDS; 2013. (Moi University, Indiana University): Kenneth H. Fife, 2. Desgrees-du-Lou A, Msellati P, Viho I, Yao A, Yapi D, Kassi P, Edwin Were; Kabwohe, Uganda (Kabwohe Clinical et al. Contraceptive use, protected sexual intercourse and incidence of pregnancies among African HIV-infected women. Research Center): Elioda Tumwesigye; Jinja, Uganda DITRAME ANRS 049 Project, Abidjan 1995–2000. Int J STD (Makerere University, University of Washington): Patrick AIDS 2002; 13:462–468. Ndase, Elly Katabira; Kampala, Uganda (Makerere 3. Halperin DT, Stover J, Reynolds HW. Benefits and costs of expanding access to family planning programs to women living University): Elly Katabira, Allan Ronald; Kisumu, Kenya with HIV. AIDS 2009; 23 (Supp 1):S123–S130. (Kenya Medical Research Institute, University of 4. UNAIDS. Global plan towards the elimination of new HIV California San Francisco): Elizabeth Bukusi, Craig R. infections. Geneva, Switzerland: UNAIDS; 2011. 5. Khu NH, Vwalika B, Karita E, Kilembe W, Bayingana RA, Sitrin Cohen; Mbale, Uganda (The AIDS Support Organiza- D, et al. Fertility goal-based counseling increases contraceptive tion, CDC-Uganda): Jonathan Wangisi, James D. Camp- implant and IUD use in HIV-discordant couples in Rwanda and bell, Jordan W. Tappero; Nairobi, Kenya (University of Zambia. Contraception 2013; 88:74–82. 6. WHO. Updated recommendations on first-line and second-line Nairobi, University of Washington): James Kiarie, Carey antiretroviral regimens and postexposure prophylaxis and re- Farquhar, Grace John-Stewart; Thika, Kenya (University commendations on early infant diagnosis of HIV: interim gui- of Nairobi, Kenya Medical Research Institute, University dance. Geneva: World Health Organization (WHO); 2018. 7. WHO. Statement on dolutegravir: potential safety issue affecting of Washington, Jomo Kenyatta University): Nelly R. women living with HIV using dolutegravir at the time of con- Mugo, Kenneth Ngure; Tororo, Uganda (CDC-Uganda, ception. Geneva, Switzerland: World Health Organization (WHO); 2018. The AIDS Support Organization): James D. Campbell, 8. Patel RC, Onono M, Gandhi M, Blat C, Hagey J, Shade SB, et al. Jordan W. Tappero, Jonathan Wangisi. Pregnancy rates in HIV-positive women using contraceptives and efavirenz-based or nevirapine-based antiretroviral therapy in Kenya: a retrospective cohort study. Lancet HIV 2015; Data management was provided by DF/Net Research, 2:e474–e482. Inc. (Seattle, Washington, USA) and site laboratory 9. Vieira CS, de Souza BM, Brito RM, Rocha Prandini MB, Amaral oversight was provided by Contract Lab Services TR, Bahmondes E, et al. Effect of antiretroviral therapy including lopinavir/ritonavir or efavirenz on etonogestrel-releasing (University of the Witwatersrand, Johannesburg, South implant pharmacokinetics in HIV-positive women. J Acquir Africa). Immune Defic Syndr 2014; 66:378–385.

10. Scarsi KK, Darin KM, Nakalema S, Back DJ, Byakika-Kibwika P, 20. Moreno I, Quinones L, Catalan J, Miranda C, Roco A, Sasso J, Else LJ, et al. Unintended pregnancies observed with combined et al. Influence of CYP3A4/5 polymorphisms in the pharma- use of the levonorgestrel contraceptive implant and efavirenz- cokinetics of levonorgestrel: a pilot study. Biomedica 2012; based antiretroviral therapy: a three-arm pharmacokinetic 32:570–577. evaluation over 48 weeks. Clin Infect Dis 2016; 62:675–682. 21. Blue SW, Winchell AJ, Kaucher AV, Lieberman RA, Gilles CT, 11. Chappell CA, Lamorde M, Nakalema S, Chen B, Mackline H, Pyra MN, et al. Simultaneous quantitation of multiple contra- Riddler S, et al. Efavirenz decreases etonogestrel exposure: a ceptive hormones in human serum by LC–MS/MS. Contra- pharmacokinetic evaluation of implantable contraception with ception 2018; 97:363–369. antiretroviral therapy. AIDS 2017; 31:1965–1972. 22. Practice Committee of the American Society for Reproductive 12. Kreitchmann R, Stek A, Best B, Capparelli E, Wang J, Shapiro Medicine. Diagnostic evaluation of the infertile female: a DE, et al. Interaction between etonogestrel-releaseing implant committee opinion. Fertil Steril 2015; 103:e44–e50. and three antiretroviral regimens. In: Conference on Retro- 23. Trussell J. Contraceptive failure in the United States. Contra- viruses and Opportunistic Infections (CROI). Seatte, WA; 13– ception 2011; 83:397–404. 17 February 2017. 24. Stormer E, von Moltke LL, Perloff MD, Greenblatt DJ. Differ- 13. Coukell AJ, Balfour JA. Levonorgestrel subdermal implants. A ential modulation of P-glycoprotein expression and activity by review of contraceptive efficacy and acceptability. Drugs nonnucleoside HIV-1 reverse transcriptase inhibitors in cell 1998; 55:861–887. culture. Pharm Res 2002; 19:1038–1045. 14. Sivin I, Wan L, Ranta S, Alvarez F, Brache V, Mishell DR Jr, et al. 25. Fellay J, Marzolini C, Meaden ER, Back DJ, Buclin T, Chave JP, Levonorgestrel concentrations during 7 years of continuous et al. Response to antiretroviral treatment in HIV-1-infected use of Jadelle contraceptive implants. Contraception 2001; individuals with allelic variants of the multidrug resistance 64:43–49. transporter 1: a pharmacogenetics study. Lancet 2002; 15. Baeten JM, Donnell D, Ndase P, Mugo NR, Campbell JD, 359:30–36. Wangisi J, et al. Antiretroviral prophylaxis for HIV prevention 26. Elens L, Vandercam B, Yombi JC, Lison D, Wallemacq P, in heterosexual men and women. NEnglJMed2012; 367:399–410. Haufroid V. Influence of host genetic factors on efavirenz 16. Mujugira A, Baeten JM, Donnell D, Ndase P, Mugo NR, Barnes plasma and intracellular pharmacokinetics in HIV-1-infected L, et al. Characteristics of HIV-1 serodiscordant couples en- patients. Pharmacogenomics 2010; 11:1223–1234. rolled in a clinical trial of antiretroviral preexposure prophy- 27. Panel on Antiretroviral Guidelines for Adults and Adolescents. laxis for HIV-1 prevention. PLoS One 2011; 6:e25828. Guidelines for the use of antiretroviral agents in HIV-infected 17. Pyra M, Heffron R, Mugo NR, Nanda K, Thomas KK, Celum C, adults and adolescents. Editor: Department of Health and et al. Effectiveness of hormonal contraception in HIV-infected Human Services. May 2018. women using antiretroviral therapy. AIDS 2015; 29:2353–2359. 28. European AIDS Clinical Society. Guidelines for treatment of 18. Wenzl R, van Beek A, Schnabel P, Huber J. Pharmacokinetics of HIV-positive adults. October 2017. etonogestrel released from the contraceptive implant Impla- 29. Song IH, Borland J, Chen S, Wajima T, Peppercorn AF, Piscitelli non. Contraception 1998; 58:283–288. SC. Dolutegravir has no effect on the pharmacokinetics of oral 19. Neary M, Lamorde M, Olagunju A, Darin KM, Merry C, By- contraceptives with norgestimate and ethinyl estradiol. Ann akika-Kibwika P, et al. The effect of gene variants on levonor- Pharmacother 2015; 49:784–789. gestrel pharmacokinetics when combined with antiretroviral 30. Curtis KM, Tepper NK, Jatlaoui TC, Berry-Bibee E, Horton LG, therapy containing efavirenz or nevirapine. Clin Pharmacol Zapata LB, et al. U.S. medical eligibility criteria for contra- Ther 2017; 102:529–536. ceptive use. MMWR Recomm Rep 2016; 65:1–103.