- Email: [email protected]
II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter randomized trial
CONN-09089; No of Pages 7 Contraception xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Contraception journal homepage: www.elsevier.com/locate/con
Original research article
Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrelonly contraceptive patches in a multicenter randomized trial☆,☆☆ Carolyn L Westhoff a,⁎, Beatrice A Chen b, Jeffrey T Jensen c, Kurt Barnhart d, Michael Thomas e, Stephanie Teal f, Diana Blithe g a
Columbia University University of Pittsburgh/Magee-Womens Research Institute Oregon Health and Science University d University of Pennsylvania e University of Cincinnati f University of Colorado g National Institute of Child Health and Human Development b c
a r t i c l e
i n f o
Article history: Received 5 January 2018 Received in revised form 6 April 2018 Accepted 8 April 2018 Available online xxxx Keywords: Contraceptive patch Levonorgestrel Cervical mucus Pharmacodynamics Ovulation suppression
a b s t r a c t Objectives: To evaluate the pharmacokinetic and pharmacodynamic effects of two novel levonorgestrel (LNG)only contraceptive patches. Study Design: This multi-center Phase I/II trial randomized participants to use a 7-day transdermal patch designed to deliver a 40 mcg or 75 mcg daily LNG dose for 11 consecutive weeks; randomization was stratified by body mass index (BMI) b32 or 32–40 kg/m2. Evaluations occurred twice weekly for 5 weeks, then once weekly with supplemental visits for participants reporting a urinary LH surge. We evaluated pharmacokinetics via the weekly trough serum LNG concentrations (≥180 pg/mL considered therapeutic). The primary pharmacodynamic outcomes were cervical mucus and ovulation, evaluated using modified Insler cervical mucus scores and serum markers of ovulation. Results: Randomization yielded similar groups. Of 121 enrolled women, 117 contributed analyzable data. Overall, mean LNG concentrations achieved with use of the 75 mcg patch [value] were roughly double those seen with the 40 mcg dose. Concentrations achieved in higher-BMI women (BMI=32–40 kg/m2) were 50% lower compared to lower-BMI women for both patches. During the study, 94/117 (80%) participants had at least one LNG concentration b180 pg/mL. Insler scores were ≥5 at least once in 102/117 (87%) of participants, and ovulation occurred at least once in 81/117 (69%). Sixty participants (51%) had at least one ovulation following an Insler score ≥5; this occurred most frequently among higher-BMI participants using the lower-dose patch (19/25, 76%). Conclusions: Women using two investigational LNG patches did not achieve consistent concentrations of LNG N180 pg/mL, an efficacy benchmark established in other studies with LNG implants, and they experienced high rates of ovulation as well as cervical mucus scores not supportive of high contraceptive efficacy. Implications: The novel LNG-only patches in this study may not adequately decrease pregnancy risk, particularly for higher-BMI women. Future studies of LNG-only contraceptive patches will need to employ higher LNG doses and assess the acceptability and safety of such dosing. © 2018 Elsevier Inc. All rights reserved.
☆ Funding: The National Institute of Child Health and Human Development sponsored this study, and Agile Therapeutics supplied the investigational product. Agile Therapeutics had no involvement in the protocol, study execution, analysis, or preparation of this manuscript. ☆☆ Potential Conflicts of Interest: CW – Consultant to Merck, Bayer, Cooper Surgical, Agile Therapeutics and Allergan. Research support from Estetra SPRL, Leon Farma and Medicines360, all managed through Columbia University.BC – Consultant to Merck. Research support through Agile Therapeutics, Bayer Healthcare, Medicines360, and Merck, all managed through the Magee-Women’s Research Institute.JJ – Consultant and research support from Bayer Healthcare, Merck, Agile Therapeutics, Abbvie, HRA Pharma, Teva, and the Population Council; consulting only from MicroChips and Evofem, and research support only from Estetra SPRL and Medicines360. These companies and organizations may have a commercial or financial interest in the results of this research and technology. These potential conflicts of interest have been reviewed and managed by OHSU.KB – Consultant to Bayer and AbbvieMT – Consultant to Evofem. Research support from Agile, Merck, Evofem, Estetra, Medicines 360ST – Consultant to Merck and Medicines360. University of Colorado Department of OB-GYN receives support for contraceptive research trials from Bayer, ContraMed, Medicines360.DB – None. ⁎ Corresponding author. Tel.: +1 212 305 4805. E-mail address: [email protected] (C.L. Westhoff).
https://doi.org/10.1016/j.contraception.2018.04.007 0010-7824/© 2018 Elsevier Inc. All rights reserved.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
2
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
1. Introduction Progestin (P)-only contraception is a key option for all women, especially those who have medical contraindications to, or who prefer to avoid, estrogen-containing contraceptives. Women with higher body mass index (BMI), a group at increased-risk of venous thromboembolism, may particularly benefit from use of a P-only method [1]. Around the world, available P-only methods include pills (levonorgestrel (LNG), desogestrel, norethindrone), injections (medroxyprogesterone acetate, norethisterone enantate), intrauterine devices (LNG), and implants (LNG and etonogestrel). While LNG pills and implants are FDAapproved, they are not currently available in the United States (U.S.). The World Health Organization (WHO) previously tested an LNG-only vaginal ring, but this product is not marketed in any country [2]. As LNG is a potent progestin, LNG-only products have potential for high efficacy; for example, LNG-containing implants achieve failure rates less than 1% per year for at least 5 years [3]. Sivin et al. evaluated serum LNG concentrations during a 7-year study of the Jadelle® two-rod LNG implant system. While levels declined during the study, pregnancies did not occur in subjects with LNG N180 pg/mL [3]. This level of efficacy is achieved notwithstanding that these implants suppress ovulation only incompletely with as many as 20–40% of users demonstrating possible ovulation based on achieving serum progesterone ≥3 ng/mL [4,5]. This prevalence of ovulation (or possible ovulation) led to the presumption that extra-ovulatory effects are important for contraceptive efficacy, in particular effects on cervical mucus thickening [6]. A commonality between the LNG-containing implant and the LNGonly vaginal ring is a decrease in effectiveness among higher-BMI women; this weight dependence is particularly well documented for the LNG implant and ring [2,3]. LNG emergency contraception (EC) effectiveness is also weight dependent [7,8]. Pharmacokinetic (PK) studies of LNG-containing combined pills and EC, as well as implants, all indicate lower LNG concentrations among higher-BMI women compared to lower-BMI women [9–12]. LNG-containing combined oral contraceptives are among the most widely used contraceptives worldwide, and the effectiveness of LNGonly implants further supports the expanded use of LNG products with non-oral delivery systems; however, the relationship between LNG effectiveness and BMI requires continued evaluation. The present dose-finding study thus evaluated two novel LNG-only contraceptive patches in U.S. women with a BMI up to 40 kg/m2. The benefits of a patch compared to oral P-only contraceptives may include enhanced effectiveness due to more stable serum levels, and greater ease of use due to weekly administration. Patch benefits compared to an implant include greater personal control over initiation and discontinuation. The primary goal of this study was to evaluate cervical mucus scores during patch use. Secondary goals were to assess LNG serum concentrations and ovulation prevalence. 2. Materials and methods This multi-center open-label Phase 1/2, randomized, parallel group study evaluated cervical mucus scores, pharmacokinetic profile and progesterone (P4) levels during use of two P-only patches containing different doses of LNG. The study patches were the AG900–6.5 and AG900–12.5 transdermal contraceptive delivery systems (TCDS) (Agile Therapeutics). Each TCDS is designed for one week of patch wear. AG900–6.5 delivers approximately 40 mcg LNG daily, similar to the daily dose of the LNG-only oral contraceptive, OvretteTM. The AG900–12.5 patch delivers approximately 75 mcg LNG daily. Staff applied each study patch to the participant's abdomen and changed the patch weekly for up to 11 weeks. This clinical trial (NCT01166412) took place at 6 centers, all members of the NICHD Contraceptive Clinical Trials Network (CCTN): these centers were Columbia University, Magee-Womens Research Institute
at the University of Pittsburgh, Oregon Health and Science University (OHSU), University of Cincinnati, University of Colorado, and University of Pennsylvania. Each center obtained local IRB approval, and each participant signed written informed consent prior to any study activities. All recruitment and participant-related study activities took place between December 2010 and July 2012. Individual participants completed study activities in about 4 months. We recruited presumed ovulatory women with a regular menstrual cycle (28 +/− 7 days) in good general health aged 18–44, with a baseline BMI of either b32 kg/m2 or ≥32–40 kg/m2 (deemed here as lowerand higher-BMI) as assessed at screening. We used this BMI cutoff to ensure greater sampling of high BMI women. By design, the study included women with a BMI range from 17.4 to 39.9. Additional eligibility criteria included either heterosexual abstinence, willingness to rely on condoms throughout the study, or sterilization. The main exclusion criteria were undiagnosed bleeding or discharge, an abnormal cervical screening result that would require additional evaluation or treatment during the study, injectable contraception use within 10 months, or other hormonal contraception use within 30 days of enrollment. Current IUD use or medical contraindications to combined hormonal contraception also precluded participation. Following informed consent, participants presented on menstrual cycle day 10–14 for cervical mucus aspiration and transvaginal sonography (TVS) to measure ovarian follicle development. Participants returned on cycle day 20–24, or approximately 7 days after obtaining an eligible cervical mucus score, for a serum P4 level to confirm an ovulatory cycle. Eligibility required a modified Insler score of≥7 and a serum P4 of ≥3 ng/mL. Investigators used the SelectMucus Endocervical Aspirator (Cooper Surgical) to obtain mucus samples, and assessed these according to WHO methodology [13]. The WHO cervical mucus score incorporates volume, consistency (viscosity), ferning, spinnbarkeit and cellularity, scoring each variable on a 4-point scale (0–3, possible total score 0– 15). In the analysis, we rated Insler scores b5 or an inability to aspirate any mucus as “hostile” or unfavorable mucus; conversely, we considered scores ≥5 as being unlikely to offer contraceptive protection (i.e., favorable mucus for conception). After confirming baseline ovulation and eligibility, the study randomized participants to one of the investigational patches. The study coordinating center, Health Decisions (Chapel Hill, NC), prepared a randomization schedule stratified by BMI group and study site, and specified a 1:1 allocation ratio. Sites obtained treatment assignments using an interactive voice response system; the patch for each dose was a visibly different size, precluding blinding. Randomization codes were generated using permuted blocks and assigned by a web-based system. The randomization was stratified by BMI groups (BMI N = 32 kg/m2 but less than 40 kg/m2, BMIb32 kg/m2) and the study was open-label. Participants received a new patch at weekly intervals. Study staff applied the initial patch during the first 5 days of a menstrual cycle and also applied each weekly patch. Subsequent patch applications took place within one hour of the initial patch application time. Patch checks at every study visit were intended to encourage compliance and detect non-compliance. Visits during patch use included blood sampling for P4 and LNG (trough levels at the end of one week, immediately prior to applying the next patch), cervical mucus assessment, and TVS to evaluate ovarian follicular activity. Additional sampling assessed LNG serum levels at 0, 1, 2, 4 and 6 h following application of patches 1 and 5, and following removal of patch 11. Participants made twice weekly visits throughout the first 4 weeks. During weeks 5–11, scheduled visits occurred weekly, but whenever the weekly TVS identified an ovarian follicle N12 mm, the participant would initiate daily urine LH testing using a Clearblue digital ovulation test kit (Proctor and Gamble, Cincinnati, Ohio). Any positive urine LH triggered a return visit for TVS and blood sampling. At each TVS, we identified the largest follicle visible in each ovary and recorded 3 perpendicular dimensions of that follicle.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
Participants completed daily diary cards to report patch-wearing information and use of concomitant medications. Participation ended 1 week after final patch removal. The Endocrine Technologies Support Core (ETSC) at the Oregon National Primate Research Center conducted the hormone analyses. Serum P4 (linear ranges 0.03–60 ng/ml) was measured using a Roche cobas e411 chemiluminescence-based automatic clinical platform (Roche Diagnostics, Indianapolis, IN) with intra- and inter-assay variation consistently less than 7%. Serum levonorgestrel (LNG) was measured by competitive radioimmunoassay (RIA) (Immunometrics, London) (see Appendix A for additional details). We considered an LNG concentration below 180 ng/mL to be inconsistent with reliable contraceptive efficacy [3]. The study planned to enroll 120 women, 60 at each dose level, further stratified to enroll BMI groups equally into each dose level. The sample size is consistent with prior Phase I/II non-confirmatory studies in this therapeutic area and is sufficient to provide relatively precise estimates of the proportion of subjects with ‘hostile’ cervical mucus (95%
3
confidence interval within 8–9% of the point estimate). We provide descriptive results stratified by patch dose (40 mcg vs. 75 mcg LNG daily) and by BMI group (b32 vs. 32–40 kg/m2). In sensitivity analyses we dichotomized BMI at 30; results were similar and thus are not presented further. The calculated LNG population PK used the weekly trough levels during treatment. We summarized LNG concentrations using medians (using geometric means yielded similar results). We describe the proportion of women in each group with Insler scores of 0–4 vs. ≥5, and describe the proportion of participants who were presumed ovulatory during treatment based on measured P4 concentrations ≥3 ng/mL, as well as other thresholds from 2–5 ng/mL. Because ovulation does not present a high pregnancy risk if it occurs in association with hostile cervical mucus, we also quantified ovulations with a mucus score ≥5 during the prior week as being consistent with high pregnancy risk. Additional sensitivity analyses quantified ovulations with mucus scores ≥5 up to 10 days prior to detection of an elevated P4 and a wider range of mucus scores. Sensitivity analyses evaluating different thresholds of mucus scores yielded similar results and are not shown.
Fig. 1. Progress of participants through the trial of two LNG-only contraceptive patches.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
4
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
3. Results One hundred twenty-one women completed screening and underwent randomization (Fig. 1); 104 women completed the study and 17 discontinued early. Of these, 7 withdrew consent, 5 withdrew due to an adverse event, 2 were lost-to-follow-up, 1 used an excluded medication, and the investigator/'sponsor withdrew 2 (elevated cholesterol levels at screening leading to a waiver being denied for continued participation; at risk for incarceration). The 5 adverse events leading to study discontinuation included acne, agitation, hot flashes, mood swings and an incomplete central retinal vein occlusion (which the investigator rated as “possibly related” to patch use). Twelve of the 17 withdrawals occurred among women assigned to the higher-dose patch, and 10 of these withdrawals were among women with BMI b32 kg/m2. The participants who withdrew were more likely to be parous than the women who completed the study (65% vs. 39%), but were otherwise similar (data not shown). Outcome data came from 117 women; the 104 women who completed the study made an average of 17.8 visits during treatment (range 16–18). Seventy-one lower-BMI and 50 higher-BMI women contributed data because enrolling higher-BMI women was more difficult. The treatment groups were similar with respect to participant baseline characteristics, as detailed in Table 1. Participants had a mean age of 31 with 28% greater than 35 years at enrollment; 15% identified as Latina. LNG concentrations approached steady state 2 days after initial patch application, and individual participants maintained consistent levels through the study. We compared the peak LNG concentrations measured after application of patch five to the trough concentrations; the median percent differences were as follows: The low-dose patch Cmax was 19% N trough LNG in the BMI b32 group, and was 18% N trough LNG in the BMI N32 group. The high-dose patch Cmax was 22% N trough LNG in the BMI b32 group, and was 25% N trough LNG in the BMI N32 group. Thus, at steady state LNG concentrations remained quite stable at weekly patch changes. Further, LNG values did not show evidence of accumulation after the application of the second and subsequent patches. Fig. 2 illustrates geometric mean LNG concentrations during patch use stratified by LNG dose and BMI subgroup. At every visit during treatment LNG serum concentrations were lower
among participants with higher BMI. After accounting for BMI, LNG concentrations were unrelated to age or to race/ethnicity. During 11 treatment weeks, trough LNG concentrations below 180 pg/mL occurred at least once among 24/25 (96%) higher-BMI users of the lower-dose patch, and in 14 of these 25 women (56%), all LNG concentrations were below 180 pg/mL (see Table 2). Even among the 33 lower-BMI users of the higher-dose patch, 20 women (61%) had at least one trough LNG value below 180 pg/mL, and one participant in that subgroup (3%) had all trough values below 180 pg/mL. Table 3 presents the frequency of cervical mucus Insler scores ≥5 during treatment in the same dose vs. BMI groups with the median LNG concentrations of each subgroup. Very few participants (15/117, 13%) exhibited consistently unfavorable mucus (b5) throughout the entire study. Sensitivity analyses using alternative thresholds yielded similar results (data not shown). Within each BMI/dose subgroup, the women with mucus scores consistentlyb5 and those with higher mucus scores had similar LNG concentrations. Table 4 presents ovulation frequency, defined as achieving a P4 level ≥3 ng/mL, during treatment stratified by the same dose/BMI subgroups. Most participants (81/117, 69%) ovulated at least once during treatment, particularly those using the lower-dose patch with a higherBMI. Within each group, the women who ovulated and those who did not had similar LNG concentrations. In a sensitivity analysis, we defined ovulation based on P4 thresholds ranging from 2–5 ng/mL; varying the definition of ovulation in this way had little effect on the results. During weeks 5–11, 77/117 (66%) participants started daily LH testing at least once due to detection of a follicle N12 mm. Of these, 19/77 (25%) reported a positive result and presented for an additional examination. Eight of these 19 subjects (42%) achieved a serum P4 ≥3 ng/mL following the LH surge. Ovulation is likely to present a high pregnancy risk only if it occurs in association with or following favorable cervical mucus (favorable meaning here favorable for conception). Table 5 shows the proportion of women who ovulated during the study with an Insler score ≥5 in the week preceding the detection of ovulation. During treatment, 51% of participants met this criterion. As with ovulation or Insler scores assessed alone, these results had no association with LNG concentrations.
Table 1 Baseline characteristics of participants in a Phase I/II trial of two doses of an LNG-only contraceptive patch. Characteristic
Age (years) Mean (SD) Race/Ethnicity, n (%) Asian Black or African American White Other Hispanic or Latina BMI at Screening c Mean (SD) n (%) ≤ 24.9 25.0–29.9 30.0–31.9 ≥ 32.0 Number of Deliveries, n (%) 0 1 2+
Lower-dose patch a BMIb32 (n=36)
Lower-dose patch a BMI 32–40 (n=25)
Higher-dose patch b BMIb32 (n=35)
Higher-dose patch b BMI 32–40 (n=25)
All Enrolled Subjects (N=121)
31.3 (5.9)
30.3 (6.0)
29.3 (7.3)
33.4 (6.5)
30.9 (6.6)
3 (8.3) 11 (30.6)
1 (4.0) 11 (44.0)
1 (2.9) 8 (22.9)
0 8 (32.0)
5 (4.1) 38 (31.4)
17 (47.2) 5 (13.9) 6 (18.2)
10 (40.0) 3 (12.0) 5 (21.7)
24 (68.6) 2 (5.7) 1 (4.0)
15 (60.0) 2 (8.0) 5 (21.7)
66 (54.5) 12 (9.9) 17 (16.3)
25.3 (3.4)
35.1 (2.3)
25.0 (3.9)
36.4 (2.6)
17 (47.2) 16 (44.4) 3 (8.3) 0
0 0 0 25 (100)
18 (51.4) 13 (37.1) 4 (11.4) 0
0 0 0 25 (100)
35 (28.9) 29 (24.0) 7 (5.8) 50 (41.3)
21 (60.0) 3 (8.6) 11 (31.4)
15 (60.0) 5 (20.0) 5 (20.0)
21 (60.0) 3 (8.6) 11 (31.4)
12 (60.0) 4 (16.0) 9 (36.0)
69 (57.5) 15 (12.5) 36 (30.0)
29.6 (6.1)
LNG = levonorgestrel; SD = standard deviation. a LNG Patch AG900–6.5. b LNG Patch AG900–12.5. c BMI = body mass index (kg/m2).
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
5
Fig. 2. LNG concentrations according to BMI and patch-dose subgroups over 11 weeks of active treatment and a final week of washout.
4. Discussion Serum LNG concentrations among women using the progestin-only transdermal contraceptive systems studied here varied by dose and BMI in the expected directions with the lowest serum concentrations seen among higher-BMI women using the lower-dose patch. Overall, the patch users' LNG levels were lower than levels reported among women with LNG implants: LNG levels during contraceptive implant use (both the 6-capsule and the 2-rod systems) between 6 months and 2 years ranged between 300–400 pg/mL [3,14,15]. In contrast, the average levels in the present study ranged from about 90–300 pg/mL, depending on dose and BMI group. Only women in the higher-dose patch/lower-BMI subgroup had LNG levels comparable to those reported in implant users. Women in all four subgroups frequently had LNG concentrations b180 pg/mL, levels associated with a greater risk of pregnancy in implant studies [3]. The finding of lower LNG
concentrations among higher-BMI women seen here is consistent with results of other studies of LNG-containing contraceptives [3,9– 12]. The low observed LNG levels observed here could be partly due to covert patch non-compliance undetected by the study staff. Progestin-only contraceptives act through inhibition of ovulation and where ovulation inhibition is incomplete, through thickening of cervical mucus that may prevent sperm penetration. Previous reports of other progestin-only contraceptives have found a marked effect on cervical mucus indicated by decreasing mucus scores within days or even hours of initiating the method [16–18]. Because of the typical, incomplete ovulation suppression with LNG use alone, Croxatto [6] attributed the contraceptive effectiveness of implants in large part to these cervical mucus changes. Given that expectation, the primary goal of this study was to assess whether the LNG patches were associated with decreases in cervical mucus scores. However, we found that only 10% of participants consistently had cervical mucus Insler scores b5. In
Table 2 Patch users experiencing LNG concentrationsb180 pg/mL during treatment, by dose/BMI subgroup (N=117a). LNG concentrationsb180 pg/mL
Never n (%) Ever n (%) Always⁎
Lower-dose patch b BMIb32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMIb32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
All analyzable subjects N=117
4 (11.4)
1 (4)
13 (39.4)
5 (20.8)
23 (19.7)
31 (88.6)
24 (96)
20 (60.6)
19 (79.2)
94 (80.3)
8 (22.9)
14 (56)
4 (16.7)
27 (23.1)
1 (3)
n (%) LNG concentrations reported as median pg/mL. Insler scores range from 0–15. These cells show number of participant/visits with a score of ≥5. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 enrolled participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5. ⁎ Based on 11 measures per subject, one at end of each patch week.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
6
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
Table 3 Frequency of cervical mucus Insler scores ≥5 during treatment according to treatment dose and BMI, and median LNG concentrations within each subgroup (N=117a). # visits with Insler score ≥5
Parameter Lower-dose patch b BMI b32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMI b32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
None
n (%) LNG n (%) LNG n (%) LNG n (%) LNG
1 (4.0) 47.0 4 (16.0) 147.5 11 (44.0) 84.5 9 (36.0) 101.9
9 (27.3) 310.5 7 (21.2) 377.1 12 (36.4) 368.1 5 (15.1) 290.2
2 (8.4) 49.6 8 (33.3) 194.2 5 (20.8) 147.6 9 (37.5) 206.3
One Two Three
3 (8.6) 202.8 6 (17.1) 168.4 9 (25.7) 169.6 17 (48.6) 184.0
LNG concentrations reported as median pg/mL. Insler scores range from 0–15. These cells show number of participant/visits with a score of ≥5. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 enrolled participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5.
contrast, 2/3 of participants achieved a score N5 at least twice during 11 weeks. Due to infrequent sampling, that 10% may over-estimate the proportion of women with scores b5 throughout the entire 11 weeks. However, we used home LH testing to improve our chance of evaluating peak mucus in all women who ovulated during the once/week sampling interval. An additional limitation is that inter-rater variability may limit the values of Insler score since multiple individuals carried out mucus evaluations in this multi-center study. However, even our imperfect sampling methodology identified a large proportion of the study patch users with mucus that could permit sperm penetration. This finding is discouraging with regard to the contraceptive potential of the study patches. In this study we rated any serum P4 ≥3 ng/mL as a possible ovulation. By this definition, ovulation was common during the study; 85% of women receiving the lower-dose patch had a P4 ≥3 ng/mL at least once during patch use, and 53% of women receiving the higher-dose patch crossed that threshold. As an isolated P4 ≥3 ng/mL may not indicate successful ovulation, we also initiated daily home LH testing whenever a participant had a follicle N12 mm in order to enhance ovulation detection. Even using more stringent criteria (such as a P4 threshold ≥5), ovulation was frequent, and notably more frequent than reported with LNG-implant use. Because contraceptive effects may nonetheless be present in women who continue to ovulate during patch use, we also estimated each participant's possible pregnancy risk based on a P4 increase associated with an Insler score N5. This occurred commonly (as shown in Table 5),
again suggesting that these patches may have limited contraceptive potential. The Insler score was first described as a means to determine preovulatory mucus changes for fertility patients, and this approach may not be a suitable marker to determine contraceptive effectiveness at the level of the cervix; however, at present, we lack other markers. Taken together, the combination of low circulating LNG levels, along with the frequent ovulation and the frequent absence of hostile cervical mucus seen in this study suggests that these LNG patches may have limited potential to achieve contraceptive success. Acknowledgements Thanks to Sarah Godfrey, Clint Dart and Anita Shanker of Health Decisions, the study coordinating center for management of the study and its data. Appendix A. Laboratory methods Serum levonorgestrel (LNG) was measured by competitive radioimmunoassay (RIA) (Immunometrics, Cat # IM-115; London UK). For this assay, samples (200 μl) were extracted in duplicate with 4 ml diethyl ether, dried under forced air and re-dissolved in assay buffer. Hormone values were corrected for extraction losses determined by radioactive trace recovery performed at the same time as sample extraction. The linear range of the LNG assay was 23–1500 fmol/tube, which resulted in a final calculated sensitivity of 47 pg/ml. The overall inter-assay variation was 23% and the intra-assay variations did not exceed 20%. Quality control samples and validations were repeated prior to each assay run. We further validated the low range sensitivity (e.g. b50 pg/mL) of the LNG RIA by using liquid chromatography–tandem mass spectrometry (LC–MS/MS) performed at the Bioanalytical Shared Resource/Pharmacokinetics Core facility at OHSU using protocols adapted from previously published studies. LNG extraction was optimized using diisopropyl ether with a recovery of 90%. The accuracy at 25 pg/ml was 111% with a relative standard deviation (RSD) of 7.4% and precision RSD of 5.4%. A subset of 400 serum samples was analyzed by LC–MS/MS to improve low level sensitivity to 25 pg/mL. The Endocrine Technologies Support Core (ETSC) at the Oregon National Primate Research Center conducted the hormone analyses. Serum estradiol (E2; linear ranges 5–4300 pg/ml), progesterone (P4; 0.03–60 ng/ml), luteinizing hormone (LH; 0.1–200 mIU/ml), follicle-stimulating hormone (FSH; 0.1–200 mIU/ml), and sex hormonebinding globulin (SHBG; 0.033–19 μg/ml) levels were measured using a Roche cobas e411 chemiluminescence-based automatic clinical platform (Roche Diagnostics, Indianapolis, IN) with intra- and inter-assay variation consistently less than 7% for each of the assays. Serum levonorgestrel (LNG) was measured by competitive radioimmunoassay (RIA) (Immunometrics, London).
Table 4 Ovulation frequency during patch use according to dose and BMI, with median LNG concentration by subgroup (N=117a). # Ovulations during patch use
Parameter
Lower-dose patch b BMIb32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMIb32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
None
n (%) LNG n (%) LNG n (%) LNG n (%) LNG
8 (22.9) 182.5 5 (14.2) 129.0 8 (22.9) 199.0 14 (40.0) 158.1
1 (4.0) 70.9 7 (28.0) 119.8 11 (44.0) 72.1 6 (24.0) 78.0
17 (51.5) 319.3 10 (30.3) 325.4 4 (12.1) 221.5 2 (6.1) 176.2
10 (41.7) 197.6 8 (33.3) 168.5 1 (4.2) 70.1 5 (20.8) 152.9
Once Twice Three times
Ovulation inferred by P4 ≥3 ng/mL; LNG concentrations reported as median pg/mL. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
7
Table 5 Ovulation frequency preceded by cervical mucus ≥5 during patch use according to dose and BMI, with median LNG concentration in each subgroup (N=117a). # Ovulations during patch use
Parameter
Lower-dose patch b BMIb32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMIb32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
None
n (%) LNG n (%) LNG n (%) LNG n (%) LNG
10 (28.6) 184.5 9 (25.7) 174.1 10 (28.6) 152.5 6 (17.1) 182.0
6 (24.0) 66.9 11 (44.0) 103.0 7 (28.0) 99.1 1 (4.0) 106.1
26 (78.8) 325.5 4 (12.1) 354.4 1 (3.0) 708.4 2 (6.1) 205.3
15 (62.5) 192.6 6 (25.0) 169.6 1 (4.2) 206.8 2 (8.3) 170.2
Once Twice Three times
Ovulation indicated by P4 ≥3 ng/mL; LNG concentrations reported as median pg/mL. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5.
References [1] Horton LG, Simmons KB, Curtis KM. Combined hormonal contraceptive use among obese women and risk for cardiovascular events: A systematic review. Contraception 2016;94(6):590–604. [2] World Health Organization. Microdose intravaginal levonorgestrel contraception: A multicentre clinical trial: III. The relationship between pregnancy rate and body weight. Contraception 1990;41(2):143–50. [3] Sivin I, Wan L, Ranta S, Alvarez F, Brache V, Mishell Jr DR, et al. Levonorgestrel concentrations during 7 years of continuous use of Jadelle contraceptive implants. Contraception 2001;64(1):43–9. [4] Brache V, Faúndes A, Johansson E, Alvarez F. Anovulation, inadequate luteal phase and poor sperm penetration in cervical mucus during prolonged use of NorplantR implants. Contraception 1985;31(3):261–73. [5] Croxatto HB, Diaz S, Pavez M, Miranda P, Brandeis A. Plasma progesterone levels during long-term treatment with levonogestrel silastic implants. Acta Endocrinol 1982;101:307–8. [6] Croxatto HB, Díaz S, Salvatierra AM, Morales P, Ebensperger C, Brandeis A. Treatment with Norplant subdermal implants inhibits sperm penetration through cervical mucus in vitro. Contraception 1987;36(2):193–201. [7] Glasier A, Cameron ST, Blithe D, Scherrer B, Mathe H, Levy D, et al. Can we identify women at risk of pregnancy despite using emergency contraception? Data from randomized trials of ulipristal acetate and levonorgestrel. Contraception 2011;84(4): 363–7. [8] Kapp N, Abitbol JL, Mathé H, Scherrer B, Guillard H, Gainer E, et al. Effect of body weight and BMI on the efficacy of levonorgestrel emergency contraception. Contraception 2015;91(2):97–104. [9] Edelman AB, Carlson NE, Cherala G, Munar MY, Stouffer RL, Cameron JL, et al. Impact of obesity on oral contraceptive pharmacokinetics and hypothalamic–pituitary– ovarian activity. Contraception 2009;80(2):119–27.
[10] Edelman AB, Cherala G, Blue SW, Erikson DW, Jensen JT. Impact of obesity on the pharmacokinetics of levonorgestrel-based emergency contraception: single and double dosing. Contraception 2016;94(1):52–7. [11] Praditpan P, Hamouie A, Basaraba CN, Nandakumar R, Cremers S, Davis AR, et al. Pharmacokinetics of levonorgestrel and ulipristal acetate emergency contraception in women with normal and obese body mass index. Contraception 2017;95(5): 464–9. [12] Westhoff CL, Torgal AH, Mayeda ER, Pike MC, Stanczyk FZ. Pharmacokinetics of a combined oral contraceptive in obese and normal-weight women. Contraception 2010;81(6):474–80. [13] World Health Organization. WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction. 5th ed. Cambridge University Press; 2010. [14] Olsson SE, Odlind V, Johansson ED, Nordström ML. Plasma levels of levonorgestrel and free levonorgestrel index in women using NORPLANT implants or two covered rods (NORPLANT-2). Contraception 1987;35(3):215–28. [15] Roy S, Mishell Jr DR, Robertson DN, Krauss RM, Lacarra M, Duda MJ. Long-term reversible contraception with levonorgestrel-releasing Silastic rods. Am J Obstet Gynecol 1984;148(7):1006–13. [16] Barbosa IC, Coutinho E, Hirsch C, Ladipo OA, Olssen SE, Ulmsten U. Temporal relationship between Uniplant insertion and changes in cervical mucus. Contraception 1996;54(4):213–7. [17] Dunson TR, Blumenthal PD, Alvarez F, Brache V, Cochon L, Dalberth B, et al. Timing and onset of contraceptive effectiveness in Norplant implant users: Part I. Changes in Cervical Mucus. Fertil Steril 1998;69(2):258–66. [18] Petta CA, Faundes A, Dunson TR, Ramos M, DeLucio M, Faundes D, et al. Timing of onset of contraceptive effectiveness in Depo-Provera users: Part I. changes in cervical mucus. Fertil Steril 1998;69(2):252–7.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
Contents lists available at ScienceDirect
Contraception journal homepage: www.elsevier.com/locate/con
Original research article
Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrelonly contraceptive patches in a multicenter randomized trial☆,☆☆ Carolyn L Westhoff a,⁎, Beatrice A Chen b, Jeffrey T Jensen c, Kurt Barnhart d, Michael Thomas e, Stephanie Teal f, Diana Blithe g a
Columbia University University of Pittsburgh/Magee-Womens Research Institute Oregon Health and Science University d University of Pennsylvania e University of Cincinnati f University of Colorado g National Institute of Child Health and Human Development b c
a r t i c l e
i n f o
Article history: Received 5 January 2018 Received in revised form 6 April 2018 Accepted 8 April 2018 Available online xxxx Keywords: Contraceptive patch Levonorgestrel Cervical mucus Pharmacodynamics Ovulation suppression
a b s t r a c t Objectives: To evaluate the pharmacokinetic and pharmacodynamic effects of two novel levonorgestrel (LNG)only contraceptive patches. Study Design: This multi-center Phase I/II trial randomized participants to use a 7-day transdermal patch designed to deliver a 40 mcg or 75 mcg daily LNG dose for 11 consecutive weeks; randomization was stratified by body mass index (BMI) b32 or 32–40 kg/m2. Evaluations occurred twice weekly for 5 weeks, then once weekly with supplemental visits for participants reporting a urinary LH surge. We evaluated pharmacokinetics via the weekly trough serum LNG concentrations (≥180 pg/mL considered therapeutic). The primary pharmacodynamic outcomes were cervical mucus and ovulation, evaluated using modified Insler cervical mucus scores and serum markers of ovulation. Results: Randomization yielded similar groups. Of 121 enrolled women, 117 contributed analyzable data. Overall, mean LNG concentrations achieved with use of the 75 mcg patch [value] were roughly double those seen with the 40 mcg dose. Concentrations achieved in higher-BMI women (BMI=32–40 kg/m2) were 50% lower compared to lower-BMI women for both patches. During the study, 94/117 (80%) participants had at least one LNG concentration b180 pg/mL. Insler scores were ≥5 at least once in 102/117 (87%) of participants, and ovulation occurred at least once in 81/117 (69%). Sixty participants (51%) had at least one ovulation following an Insler score ≥5; this occurred most frequently among higher-BMI participants using the lower-dose patch (19/25, 76%). Conclusions: Women using two investigational LNG patches did not achieve consistent concentrations of LNG N180 pg/mL, an efficacy benchmark established in other studies with LNG implants, and they experienced high rates of ovulation as well as cervical mucus scores not supportive of high contraceptive efficacy. Implications: The novel LNG-only patches in this study may not adequately decrease pregnancy risk, particularly for higher-BMI women. Future studies of LNG-only contraceptive patches will need to employ higher LNG doses and assess the acceptability and safety of such dosing. © 2018 Elsevier Inc. All rights reserved.
☆ Funding: The National Institute of Child Health and Human Development sponsored this study, and Agile Therapeutics supplied the investigational product. Agile Therapeutics had no involvement in the protocol, study execution, analysis, or preparation of this manuscript. ☆☆ Potential Conflicts of Interest: CW – Consultant to Merck, Bayer, Cooper Surgical, Agile Therapeutics and Allergan. Research support from Estetra SPRL, Leon Farma and Medicines360, all managed through Columbia University.BC – Consultant to Merck. Research support through Agile Therapeutics, Bayer Healthcare, Medicines360, and Merck, all managed through the Magee-Women’s Research Institute.JJ – Consultant and research support from Bayer Healthcare, Merck, Agile Therapeutics, Abbvie, HRA Pharma, Teva, and the Population Council; consulting only from MicroChips and Evofem, and research support only from Estetra SPRL and Medicines360. These companies and organizations may have a commercial or financial interest in the results of this research and technology. These potential conflicts of interest have been reviewed and managed by OHSU.KB – Consultant to Bayer and AbbvieMT – Consultant to Evofem. Research support from Agile, Merck, Evofem, Estetra, Medicines 360ST – Consultant to Merck and Medicines360. University of Colorado Department of OB-GYN receives support for contraceptive research trials from Bayer, ContraMed, Medicines360.DB – None. ⁎ Corresponding author. Tel.: +1 212 305 4805. E-mail address: [email protected] (C.L. Westhoff).
https://doi.org/10.1016/j.contraception.2018.04.007 0010-7824/© 2018 Elsevier Inc. All rights reserved.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
2
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
1. Introduction Progestin (P)-only contraception is a key option for all women, especially those who have medical contraindications to, or who prefer to avoid, estrogen-containing contraceptives. Women with higher body mass index (BMI), a group at increased-risk of venous thromboembolism, may particularly benefit from use of a P-only method [1]. Around the world, available P-only methods include pills (levonorgestrel (LNG), desogestrel, norethindrone), injections (medroxyprogesterone acetate, norethisterone enantate), intrauterine devices (LNG), and implants (LNG and etonogestrel). While LNG pills and implants are FDAapproved, they are not currently available in the United States (U.S.). The World Health Organization (WHO) previously tested an LNG-only vaginal ring, but this product is not marketed in any country [2]. As LNG is a potent progestin, LNG-only products have potential for high efficacy; for example, LNG-containing implants achieve failure rates less than 1% per year for at least 5 years [3]. Sivin et al. evaluated serum LNG concentrations during a 7-year study of the Jadelle® two-rod LNG implant system. While levels declined during the study, pregnancies did not occur in subjects with LNG N180 pg/mL [3]. This level of efficacy is achieved notwithstanding that these implants suppress ovulation only incompletely with as many as 20–40% of users demonstrating possible ovulation based on achieving serum progesterone ≥3 ng/mL [4,5]. This prevalence of ovulation (or possible ovulation) led to the presumption that extra-ovulatory effects are important for contraceptive efficacy, in particular effects on cervical mucus thickening [6]. A commonality between the LNG-containing implant and the LNGonly vaginal ring is a decrease in effectiveness among higher-BMI women; this weight dependence is particularly well documented for the LNG implant and ring [2,3]. LNG emergency contraception (EC) effectiveness is also weight dependent [7,8]. Pharmacokinetic (PK) studies of LNG-containing combined pills and EC, as well as implants, all indicate lower LNG concentrations among higher-BMI women compared to lower-BMI women [9–12]. LNG-containing combined oral contraceptives are among the most widely used contraceptives worldwide, and the effectiveness of LNGonly implants further supports the expanded use of LNG products with non-oral delivery systems; however, the relationship between LNG effectiveness and BMI requires continued evaluation. The present dose-finding study thus evaluated two novel LNG-only contraceptive patches in U.S. women with a BMI up to 40 kg/m2. The benefits of a patch compared to oral P-only contraceptives may include enhanced effectiveness due to more stable serum levels, and greater ease of use due to weekly administration. Patch benefits compared to an implant include greater personal control over initiation and discontinuation. The primary goal of this study was to evaluate cervical mucus scores during patch use. Secondary goals were to assess LNG serum concentrations and ovulation prevalence. 2. Materials and methods This multi-center open-label Phase 1/2, randomized, parallel group study evaluated cervical mucus scores, pharmacokinetic profile and progesterone (P4) levels during use of two P-only patches containing different doses of LNG. The study patches were the AG900–6.5 and AG900–12.5 transdermal contraceptive delivery systems (TCDS) (Agile Therapeutics). Each TCDS is designed for one week of patch wear. AG900–6.5 delivers approximately 40 mcg LNG daily, similar to the daily dose of the LNG-only oral contraceptive, OvretteTM. The AG900–12.5 patch delivers approximately 75 mcg LNG daily. Staff applied each study patch to the participant's abdomen and changed the patch weekly for up to 11 weeks. This clinical trial (NCT01166412) took place at 6 centers, all members of the NICHD Contraceptive Clinical Trials Network (CCTN): these centers were Columbia University, Magee-Womens Research Institute
at the University of Pittsburgh, Oregon Health and Science University (OHSU), University of Cincinnati, University of Colorado, and University of Pennsylvania. Each center obtained local IRB approval, and each participant signed written informed consent prior to any study activities. All recruitment and participant-related study activities took place between December 2010 and July 2012. Individual participants completed study activities in about 4 months. We recruited presumed ovulatory women with a regular menstrual cycle (28 +/− 7 days) in good general health aged 18–44, with a baseline BMI of either b32 kg/m2 or ≥32–40 kg/m2 (deemed here as lowerand higher-BMI) as assessed at screening. We used this BMI cutoff to ensure greater sampling of high BMI women. By design, the study included women with a BMI range from 17.4 to 39.9. Additional eligibility criteria included either heterosexual abstinence, willingness to rely on condoms throughout the study, or sterilization. The main exclusion criteria were undiagnosed bleeding or discharge, an abnormal cervical screening result that would require additional evaluation or treatment during the study, injectable contraception use within 10 months, or other hormonal contraception use within 30 days of enrollment. Current IUD use or medical contraindications to combined hormonal contraception also precluded participation. Following informed consent, participants presented on menstrual cycle day 10–14 for cervical mucus aspiration and transvaginal sonography (TVS) to measure ovarian follicle development. Participants returned on cycle day 20–24, or approximately 7 days after obtaining an eligible cervical mucus score, for a serum P4 level to confirm an ovulatory cycle. Eligibility required a modified Insler score of≥7 and a serum P4 of ≥3 ng/mL. Investigators used the SelectMucus Endocervical Aspirator (Cooper Surgical) to obtain mucus samples, and assessed these according to WHO methodology [13]. The WHO cervical mucus score incorporates volume, consistency (viscosity), ferning, spinnbarkeit and cellularity, scoring each variable on a 4-point scale (0–3, possible total score 0– 15). In the analysis, we rated Insler scores b5 or an inability to aspirate any mucus as “hostile” or unfavorable mucus; conversely, we considered scores ≥5 as being unlikely to offer contraceptive protection (i.e., favorable mucus for conception). After confirming baseline ovulation and eligibility, the study randomized participants to one of the investigational patches. The study coordinating center, Health Decisions (Chapel Hill, NC), prepared a randomization schedule stratified by BMI group and study site, and specified a 1:1 allocation ratio. Sites obtained treatment assignments using an interactive voice response system; the patch for each dose was a visibly different size, precluding blinding. Randomization codes were generated using permuted blocks and assigned by a web-based system. The randomization was stratified by BMI groups (BMI N = 32 kg/m2 but less than 40 kg/m2, BMIb32 kg/m2) and the study was open-label. Participants received a new patch at weekly intervals. Study staff applied the initial patch during the first 5 days of a menstrual cycle and also applied each weekly patch. Subsequent patch applications took place within one hour of the initial patch application time. Patch checks at every study visit were intended to encourage compliance and detect non-compliance. Visits during patch use included blood sampling for P4 and LNG (trough levels at the end of one week, immediately prior to applying the next patch), cervical mucus assessment, and TVS to evaluate ovarian follicular activity. Additional sampling assessed LNG serum levels at 0, 1, 2, 4 and 6 h following application of patches 1 and 5, and following removal of patch 11. Participants made twice weekly visits throughout the first 4 weeks. During weeks 5–11, scheduled visits occurred weekly, but whenever the weekly TVS identified an ovarian follicle N12 mm, the participant would initiate daily urine LH testing using a Clearblue digital ovulation test kit (Proctor and Gamble, Cincinnati, Ohio). Any positive urine LH triggered a return visit for TVS and blood sampling. At each TVS, we identified the largest follicle visible in each ovary and recorded 3 perpendicular dimensions of that follicle.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
Participants completed daily diary cards to report patch-wearing information and use of concomitant medications. Participation ended 1 week after final patch removal. The Endocrine Technologies Support Core (ETSC) at the Oregon National Primate Research Center conducted the hormone analyses. Serum P4 (linear ranges 0.03–60 ng/ml) was measured using a Roche cobas e411 chemiluminescence-based automatic clinical platform (Roche Diagnostics, Indianapolis, IN) with intra- and inter-assay variation consistently less than 7%. Serum levonorgestrel (LNG) was measured by competitive radioimmunoassay (RIA) (Immunometrics, London) (see Appendix A for additional details). We considered an LNG concentration below 180 ng/mL to be inconsistent with reliable contraceptive efficacy [3]. The study planned to enroll 120 women, 60 at each dose level, further stratified to enroll BMI groups equally into each dose level. The sample size is consistent with prior Phase I/II non-confirmatory studies in this therapeutic area and is sufficient to provide relatively precise estimates of the proportion of subjects with ‘hostile’ cervical mucus (95%
3
confidence interval within 8–9% of the point estimate). We provide descriptive results stratified by patch dose (40 mcg vs. 75 mcg LNG daily) and by BMI group (b32 vs. 32–40 kg/m2). In sensitivity analyses we dichotomized BMI at 30; results were similar and thus are not presented further. The calculated LNG population PK used the weekly trough levels during treatment. We summarized LNG concentrations using medians (using geometric means yielded similar results). We describe the proportion of women in each group with Insler scores of 0–4 vs. ≥5, and describe the proportion of participants who were presumed ovulatory during treatment based on measured P4 concentrations ≥3 ng/mL, as well as other thresholds from 2–5 ng/mL. Because ovulation does not present a high pregnancy risk if it occurs in association with hostile cervical mucus, we also quantified ovulations with a mucus score ≥5 during the prior week as being consistent with high pregnancy risk. Additional sensitivity analyses quantified ovulations with mucus scores ≥5 up to 10 days prior to detection of an elevated P4 and a wider range of mucus scores. Sensitivity analyses evaluating different thresholds of mucus scores yielded similar results and are not shown.
Fig. 1. Progress of participants through the trial of two LNG-only contraceptive patches.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
4
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
3. Results One hundred twenty-one women completed screening and underwent randomization (Fig. 1); 104 women completed the study and 17 discontinued early. Of these, 7 withdrew consent, 5 withdrew due to an adverse event, 2 were lost-to-follow-up, 1 used an excluded medication, and the investigator/'sponsor withdrew 2 (elevated cholesterol levels at screening leading to a waiver being denied for continued participation; at risk for incarceration). The 5 adverse events leading to study discontinuation included acne, agitation, hot flashes, mood swings and an incomplete central retinal vein occlusion (which the investigator rated as “possibly related” to patch use). Twelve of the 17 withdrawals occurred among women assigned to the higher-dose patch, and 10 of these withdrawals were among women with BMI b32 kg/m2. The participants who withdrew were more likely to be parous than the women who completed the study (65% vs. 39%), but were otherwise similar (data not shown). Outcome data came from 117 women; the 104 women who completed the study made an average of 17.8 visits during treatment (range 16–18). Seventy-one lower-BMI and 50 higher-BMI women contributed data because enrolling higher-BMI women was more difficult. The treatment groups were similar with respect to participant baseline characteristics, as detailed in Table 1. Participants had a mean age of 31 with 28% greater than 35 years at enrollment; 15% identified as Latina. LNG concentrations approached steady state 2 days after initial patch application, and individual participants maintained consistent levels through the study. We compared the peak LNG concentrations measured after application of patch five to the trough concentrations; the median percent differences were as follows: The low-dose patch Cmax was 19% N trough LNG in the BMI b32 group, and was 18% N trough LNG in the BMI N32 group. The high-dose patch Cmax was 22% N trough LNG in the BMI b32 group, and was 25% N trough LNG in the BMI N32 group. Thus, at steady state LNG concentrations remained quite stable at weekly patch changes. Further, LNG values did not show evidence of accumulation after the application of the second and subsequent patches. Fig. 2 illustrates geometric mean LNG concentrations during patch use stratified by LNG dose and BMI subgroup. At every visit during treatment LNG serum concentrations were lower
among participants with higher BMI. After accounting for BMI, LNG concentrations were unrelated to age or to race/ethnicity. During 11 treatment weeks, trough LNG concentrations below 180 pg/mL occurred at least once among 24/25 (96%) higher-BMI users of the lower-dose patch, and in 14 of these 25 women (56%), all LNG concentrations were below 180 pg/mL (see Table 2). Even among the 33 lower-BMI users of the higher-dose patch, 20 women (61%) had at least one trough LNG value below 180 pg/mL, and one participant in that subgroup (3%) had all trough values below 180 pg/mL. Table 3 presents the frequency of cervical mucus Insler scores ≥5 during treatment in the same dose vs. BMI groups with the median LNG concentrations of each subgroup. Very few participants (15/117, 13%) exhibited consistently unfavorable mucus (b5) throughout the entire study. Sensitivity analyses using alternative thresholds yielded similar results (data not shown). Within each BMI/dose subgroup, the women with mucus scores consistentlyb5 and those with higher mucus scores had similar LNG concentrations. Table 4 presents ovulation frequency, defined as achieving a P4 level ≥3 ng/mL, during treatment stratified by the same dose/BMI subgroups. Most participants (81/117, 69%) ovulated at least once during treatment, particularly those using the lower-dose patch with a higherBMI. Within each group, the women who ovulated and those who did not had similar LNG concentrations. In a sensitivity analysis, we defined ovulation based on P4 thresholds ranging from 2–5 ng/mL; varying the definition of ovulation in this way had little effect on the results. During weeks 5–11, 77/117 (66%) participants started daily LH testing at least once due to detection of a follicle N12 mm. Of these, 19/77 (25%) reported a positive result and presented for an additional examination. Eight of these 19 subjects (42%) achieved a serum P4 ≥3 ng/mL following the LH surge. Ovulation is likely to present a high pregnancy risk only if it occurs in association with or following favorable cervical mucus (favorable meaning here favorable for conception). Table 5 shows the proportion of women who ovulated during the study with an Insler score ≥5 in the week preceding the detection of ovulation. During treatment, 51% of participants met this criterion. As with ovulation or Insler scores assessed alone, these results had no association with LNG concentrations.
Table 1 Baseline characteristics of participants in a Phase I/II trial of two doses of an LNG-only contraceptive patch. Characteristic
Age (years) Mean (SD) Race/Ethnicity, n (%) Asian Black or African American White Other Hispanic or Latina BMI at Screening c Mean (SD) n (%) ≤ 24.9 25.0–29.9 30.0–31.9 ≥ 32.0 Number of Deliveries, n (%) 0 1 2+
Lower-dose patch a BMIb32 (n=36)
Lower-dose patch a BMI 32–40 (n=25)
Higher-dose patch b BMIb32 (n=35)
Higher-dose patch b BMI 32–40 (n=25)
All Enrolled Subjects (N=121)
31.3 (5.9)
30.3 (6.0)
29.3 (7.3)
33.4 (6.5)
30.9 (6.6)
3 (8.3) 11 (30.6)
1 (4.0) 11 (44.0)
1 (2.9) 8 (22.9)
0 8 (32.0)
5 (4.1) 38 (31.4)
17 (47.2) 5 (13.9) 6 (18.2)
10 (40.0) 3 (12.0) 5 (21.7)
24 (68.6) 2 (5.7) 1 (4.0)
15 (60.0) 2 (8.0) 5 (21.7)
66 (54.5) 12 (9.9) 17 (16.3)
25.3 (3.4)
35.1 (2.3)
25.0 (3.9)
36.4 (2.6)
17 (47.2) 16 (44.4) 3 (8.3) 0
0 0 0 25 (100)
18 (51.4) 13 (37.1) 4 (11.4) 0
0 0 0 25 (100)
35 (28.9) 29 (24.0) 7 (5.8) 50 (41.3)
21 (60.0) 3 (8.6) 11 (31.4)
15 (60.0) 5 (20.0) 5 (20.0)
21 (60.0) 3 (8.6) 11 (31.4)
12 (60.0) 4 (16.0) 9 (36.0)
69 (57.5) 15 (12.5) 36 (30.0)
29.6 (6.1)
LNG = levonorgestrel; SD = standard deviation. a LNG Patch AG900–6.5. b LNG Patch AG900–12.5. c BMI = body mass index (kg/m2).
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
5
Fig. 2. LNG concentrations according to BMI and patch-dose subgroups over 11 weeks of active treatment and a final week of washout.
4. Discussion Serum LNG concentrations among women using the progestin-only transdermal contraceptive systems studied here varied by dose and BMI in the expected directions with the lowest serum concentrations seen among higher-BMI women using the lower-dose patch. Overall, the patch users' LNG levels were lower than levels reported among women with LNG implants: LNG levels during contraceptive implant use (both the 6-capsule and the 2-rod systems) between 6 months and 2 years ranged between 300–400 pg/mL [3,14,15]. In contrast, the average levels in the present study ranged from about 90–300 pg/mL, depending on dose and BMI group. Only women in the higher-dose patch/lower-BMI subgroup had LNG levels comparable to those reported in implant users. Women in all four subgroups frequently had LNG concentrations b180 pg/mL, levels associated with a greater risk of pregnancy in implant studies [3]. The finding of lower LNG
concentrations among higher-BMI women seen here is consistent with results of other studies of LNG-containing contraceptives [3,9– 12]. The low observed LNG levels observed here could be partly due to covert patch non-compliance undetected by the study staff. Progestin-only contraceptives act through inhibition of ovulation and where ovulation inhibition is incomplete, through thickening of cervical mucus that may prevent sperm penetration. Previous reports of other progestin-only contraceptives have found a marked effect on cervical mucus indicated by decreasing mucus scores within days or even hours of initiating the method [16–18]. Because of the typical, incomplete ovulation suppression with LNG use alone, Croxatto [6] attributed the contraceptive effectiveness of implants in large part to these cervical mucus changes. Given that expectation, the primary goal of this study was to assess whether the LNG patches were associated with decreases in cervical mucus scores. However, we found that only 10% of participants consistently had cervical mucus Insler scores b5. In
Table 2 Patch users experiencing LNG concentrationsb180 pg/mL during treatment, by dose/BMI subgroup (N=117a). LNG concentrationsb180 pg/mL
Never n (%) Ever n (%) Always⁎
Lower-dose patch b BMIb32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMIb32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
All analyzable subjects N=117
4 (11.4)
1 (4)
13 (39.4)
5 (20.8)
23 (19.7)
31 (88.6)
24 (96)
20 (60.6)
19 (79.2)
94 (80.3)
8 (22.9)
14 (56)
4 (16.7)
27 (23.1)
1 (3)
n (%) LNG concentrations reported as median pg/mL. Insler scores range from 0–15. These cells show number of participant/visits with a score of ≥5. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 enrolled participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5. ⁎ Based on 11 measures per subject, one at end of each patch week.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
6
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
Table 3 Frequency of cervical mucus Insler scores ≥5 during treatment according to treatment dose and BMI, and median LNG concentrations within each subgroup (N=117a). # visits with Insler score ≥5
Parameter Lower-dose patch b BMI b32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMI b32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
None
n (%) LNG n (%) LNG n (%) LNG n (%) LNG
1 (4.0) 47.0 4 (16.0) 147.5 11 (44.0) 84.5 9 (36.0) 101.9
9 (27.3) 310.5 7 (21.2) 377.1 12 (36.4) 368.1 5 (15.1) 290.2
2 (8.4) 49.6 8 (33.3) 194.2 5 (20.8) 147.6 9 (37.5) 206.3
One Two Three
3 (8.6) 202.8 6 (17.1) 168.4 9 (25.7) 169.6 17 (48.6) 184.0
LNG concentrations reported as median pg/mL. Insler scores range from 0–15. These cells show number of participant/visits with a score of ≥5. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 enrolled participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5.
contrast, 2/3 of participants achieved a score N5 at least twice during 11 weeks. Due to infrequent sampling, that 10% may over-estimate the proportion of women with scores b5 throughout the entire 11 weeks. However, we used home LH testing to improve our chance of evaluating peak mucus in all women who ovulated during the once/week sampling interval. An additional limitation is that inter-rater variability may limit the values of Insler score since multiple individuals carried out mucus evaluations in this multi-center study. However, even our imperfect sampling methodology identified a large proportion of the study patch users with mucus that could permit sperm penetration. This finding is discouraging with regard to the contraceptive potential of the study patches. In this study we rated any serum P4 ≥3 ng/mL as a possible ovulation. By this definition, ovulation was common during the study; 85% of women receiving the lower-dose patch had a P4 ≥3 ng/mL at least once during patch use, and 53% of women receiving the higher-dose patch crossed that threshold. As an isolated P4 ≥3 ng/mL may not indicate successful ovulation, we also initiated daily home LH testing whenever a participant had a follicle N12 mm in order to enhance ovulation detection. Even using more stringent criteria (such as a P4 threshold ≥5), ovulation was frequent, and notably more frequent than reported with LNG-implant use. Because contraceptive effects may nonetheless be present in women who continue to ovulate during patch use, we also estimated each participant's possible pregnancy risk based on a P4 increase associated with an Insler score N5. This occurred commonly (as shown in Table 5),
again suggesting that these patches may have limited contraceptive potential. The Insler score was first described as a means to determine preovulatory mucus changes for fertility patients, and this approach may not be a suitable marker to determine contraceptive effectiveness at the level of the cervix; however, at present, we lack other markers. Taken together, the combination of low circulating LNG levels, along with the frequent ovulation and the frequent absence of hostile cervical mucus seen in this study suggests that these LNG patches may have limited potential to achieve contraceptive success. Acknowledgements Thanks to Sarah Godfrey, Clint Dart and Anita Shanker of Health Decisions, the study coordinating center for management of the study and its data. Appendix A. Laboratory methods Serum levonorgestrel (LNG) was measured by competitive radioimmunoassay (RIA) (Immunometrics, Cat # IM-115; London UK). For this assay, samples (200 μl) were extracted in duplicate with 4 ml diethyl ether, dried under forced air and re-dissolved in assay buffer. Hormone values were corrected for extraction losses determined by radioactive trace recovery performed at the same time as sample extraction. The linear range of the LNG assay was 23–1500 fmol/tube, which resulted in a final calculated sensitivity of 47 pg/ml. The overall inter-assay variation was 23% and the intra-assay variations did not exceed 20%. Quality control samples and validations were repeated prior to each assay run. We further validated the low range sensitivity (e.g. b50 pg/mL) of the LNG RIA by using liquid chromatography–tandem mass spectrometry (LC–MS/MS) performed at the Bioanalytical Shared Resource/Pharmacokinetics Core facility at OHSU using protocols adapted from previously published studies. LNG extraction was optimized using diisopropyl ether with a recovery of 90%. The accuracy at 25 pg/ml was 111% with a relative standard deviation (RSD) of 7.4% and precision RSD of 5.4%. A subset of 400 serum samples was analyzed by LC–MS/MS to improve low level sensitivity to 25 pg/mL. The Endocrine Technologies Support Core (ETSC) at the Oregon National Primate Research Center conducted the hormone analyses. Serum estradiol (E2; linear ranges 5–4300 pg/ml), progesterone (P4; 0.03–60 ng/ml), luteinizing hormone (LH; 0.1–200 mIU/ml), follicle-stimulating hormone (FSH; 0.1–200 mIU/ml), and sex hormonebinding globulin (SHBG; 0.033–19 μg/ml) levels were measured using a Roche cobas e411 chemiluminescence-based automatic clinical platform (Roche Diagnostics, Indianapolis, IN) with intra- and inter-assay variation consistently less than 7% for each of the assays. Serum levonorgestrel (LNG) was measured by competitive radioimmunoassay (RIA) (Immunometrics, London).
Table 4 Ovulation frequency during patch use according to dose and BMI, with median LNG concentration by subgroup (N=117a). # Ovulations during patch use
Parameter
Lower-dose patch b BMIb32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMIb32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
None
n (%) LNG n (%) LNG n (%) LNG n (%) LNG
8 (22.9) 182.5 5 (14.2) 129.0 8 (22.9) 199.0 14 (40.0) 158.1
1 (4.0) 70.9 7 (28.0) 119.8 11 (44.0) 72.1 6 (24.0) 78.0
17 (51.5) 319.3 10 (30.3) 325.4 4 (12.1) 221.5 2 (6.1) 176.2
10 (41.7) 197.6 8 (33.3) 168.5 1 (4.2) 70.1 5 (20.8) 152.9
Once Twice Three times
Ovulation inferred by P4 ≥3 ng/mL; LNG concentrations reported as median pg/mL. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007
C.L. Westhoff et al. / Contraception xxx (2018) xxx–xxx
7
Table 5 Ovulation frequency preceded by cervical mucus ≥5 during patch use according to dose and BMI, with median LNG concentration in each subgroup (N=117a). # Ovulations during patch use
Parameter
Lower-dose patch b BMIb32 (n=35)
Lower-dose patch b BMI 32–40 (n=25)
Higher-dose patch c BMIb32 (n=33)
Higher-dose patch c BMI 32–40 (n=24)
None
n (%) LNG n (%) LNG n (%) LNG n (%) LNG
10 (28.6) 184.5 9 (25.7) 174.1 10 (28.6) 152.5 6 (17.1) 182.0
6 (24.0) 66.9 11 (44.0) 103.0 7 (28.0) 99.1 1 (4.0) 106.1
26 (78.8) 325.5 4 (12.1) 354.4 1 (3.0) 708.4 2 (6.1) 205.3
15 (62.5) 192.6 6 (25.0) 169.6 1 (4.2) 206.8 2 (8.3) 170.2
Once Twice Three times
Ovulation indicated by P4 ≥3 ng/mL; LNG concentrations reported as median pg/mL. LNG = levonorgestrel; BMI = body mass index (kg/m2). a N=117, 4/121 participants contributed no outcome data; b LNG Patch AG900–6.5; c LNG Patch AG900–12.5.
References [1] Horton LG, Simmons KB, Curtis KM. Combined hormonal contraceptive use among obese women and risk for cardiovascular events: A systematic review. Contraception 2016;94(6):590–604. [2] World Health Organization. Microdose intravaginal levonorgestrel contraception: A multicentre clinical trial: III. The relationship between pregnancy rate and body weight. Contraception 1990;41(2):143–50. [3] Sivin I, Wan L, Ranta S, Alvarez F, Brache V, Mishell Jr DR, et al. Levonorgestrel concentrations during 7 years of continuous use of Jadelle contraceptive implants. Contraception 2001;64(1):43–9. [4] Brache V, Faúndes A, Johansson E, Alvarez F. Anovulation, inadequate luteal phase and poor sperm penetration in cervical mucus during prolonged use of NorplantR implants. Contraception 1985;31(3):261–73. [5] Croxatto HB, Diaz S, Pavez M, Miranda P, Brandeis A. Plasma progesterone levels during long-term treatment with levonogestrel silastic implants. Acta Endocrinol 1982;101:307–8. [6] Croxatto HB, Díaz S, Salvatierra AM, Morales P, Ebensperger C, Brandeis A. Treatment with Norplant subdermal implants inhibits sperm penetration through cervical mucus in vitro. Contraception 1987;36(2):193–201. [7] Glasier A, Cameron ST, Blithe D, Scherrer B, Mathe H, Levy D, et al. Can we identify women at risk of pregnancy despite using emergency contraception? Data from randomized trials of ulipristal acetate and levonorgestrel. Contraception 2011;84(4): 363–7. [8] Kapp N, Abitbol JL, Mathé H, Scherrer B, Guillard H, Gainer E, et al. Effect of body weight and BMI on the efficacy of levonorgestrel emergency contraception. Contraception 2015;91(2):97–104. [9] Edelman AB, Carlson NE, Cherala G, Munar MY, Stouffer RL, Cameron JL, et al. Impact of obesity on oral contraceptive pharmacokinetics and hypothalamic–pituitary– ovarian activity. Contraception 2009;80(2):119–27.
[10] Edelman AB, Cherala G, Blue SW, Erikson DW, Jensen JT. Impact of obesity on the pharmacokinetics of levonorgestrel-based emergency contraception: single and double dosing. Contraception 2016;94(1):52–7. [11] Praditpan P, Hamouie A, Basaraba CN, Nandakumar R, Cremers S, Davis AR, et al. Pharmacokinetics of levonorgestrel and ulipristal acetate emergency contraception in women with normal and obese body mass index. Contraception 2017;95(5): 464–9. [12] Westhoff CL, Torgal AH, Mayeda ER, Pike MC, Stanczyk FZ. Pharmacokinetics of a combined oral contraceptive in obese and normal-weight women. Contraception 2010;81(6):474–80. [13] World Health Organization. WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction. 5th ed. Cambridge University Press; 2010. [14] Olsson SE, Odlind V, Johansson ED, Nordström ML. Plasma levels of levonorgestrel and free levonorgestrel index in women using NORPLANT implants or two covered rods (NORPLANT-2). Contraception 1987;35(3):215–28. [15] Roy S, Mishell Jr DR, Robertson DN, Krauss RM, Lacarra M, Duda MJ. Long-term reversible contraception with levonorgestrel-releasing Silastic rods. Am J Obstet Gynecol 1984;148(7):1006–13. [16] Barbosa IC, Coutinho E, Hirsch C, Ladipo OA, Olssen SE, Ulmsten U. Temporal relationship between Uniplant insertion and changes in cervical mucus. Contraception 1996;54(4):213–7. [17] Dunson TR, Blumenthal PD, Alvarez F, Brache V, Cochon L, Dalberth B, et al. Timing and onset of contraceptive effectiveness in Norplant implant users: Part I. Changes in Cervical Mucus. Fertil Steril 1998;69(2):258–66. [18] Petta CA, Faundes A, Dunson TR, Ramos M, DeLucio M, Faundes D, et al. Timing of onset of contraceptive effectiveness in Depo-Provera users: Part I. changes in cervical mucus. Fertil Steril 1998;69(2):252–7.
Please cite this article as: Westhoff CL, et al, Phase I/II pharmacokinetic and pharmacodynamic evaluation of two levonorgestrel-only contraceptive patches in a multicenter random..., Contraception (2018), https://doi.org/10.1016/j.contraception.2018.04.007