The etiology of bleeding secondary to progestin-only contraception such as the etonogestrel implant is poorly understood. One leading theory is that sustained exposure to a progestin can lead to endometrial angiogenesis disruption, resulting in the development of a dense venous network that is fragile and prone to bleeding.1 In a recent study evaluating reasons for discontinuation of a contraceptive method within 6 months of initiation, unscheduled bleeding was reported in 53% of women using the etonogestrel implant.2 Treatment of unscheduled bleeding may increase implant acceptability and improve continuation rates. Pharmacologic therapies such as nonsteroidal antiinflammatories and estrogen have shown mixed results and may be contraindicated as a result of medical comorbidities.3,4
Ulipristal acetate, a selective progesterone receptor modulator, is marketed for emergency contraception in the United States. Outside the United States, it is available for treatment of abnormal uterine bleeding associated with uterine leiomyoma. Uterine bleeding is reduced or eliminated by antiproliferative effects resulting from mixed agonist and antagonist effects of ulipristal acetate within the endometrial and myometrial tissue.5–9 Ulipristal acetate may also displace local progestin within the uterus to counteract bleeding secondary to the mentioned creation of a dense, fragile venous network.
The primary objective of this study was to evaluate the effectiveness of ulipristal acetate in decreasing bleeding days resulting from the etonogestrel implant over a 30-day period as compared with placebo. We hypothesized that ulipristal acetate would reduce total days of unscheduled bleeding. For our secondary outcomes, we also hypothesized that ulipristal acetate would increase satisfaction with bleeding and the implant and not affect ovulatory status.
MATERIALS AND METHODS
We performed a randomized, double-blind, placebo-controlled trial at Washington University School of Medicine in St. Louis, Missouri, from May 2017 to January 2018. This study was approved by the Washington University School of Medicine Human Research Protection Office (February 8, 2017). Before initiation of the trial, a U.S. Food and Drug Administration (FDA) Investigational New Drug application (FDA IND #134150) for ulipristal acetate was approved and the trial was registered at ClinicalTrials.gov (NCT03118297). We used Consolidated Standards of Reporting Trials guidelines in the planning of this study.10
We used poster advertisements and health care provider referrals from the St. Louis area to recruit English-speaking women aged 18–45 years using the etonogestrel 68-mg subdermal implant. Women must have had their implant placed greater than 90 days and less than 3 years before enrollment and report dissatisfaction with their bleeding pattern of greater than one bleeding episode in the previous 24-day period. Previous studies investigating therapies for unscheduled bleeding have used this definition.11 Women must have been willing to complete a 30-day bleeding diary and have the ability to do so by text message. Finally, because ovulation suppression could theoretically be affected by addition of ulipristal acetate, participants were asked to either abstain from intercourse or use condoms during the study period, which were provided at no cost. Participants were excluded if they reported a contraindication to ulipristal acetate (ie, breastfeeding, use of a CYP3A4 inducer such as carbamazepine and topiramate, use of oral glucocorticoids for uncontrolled severe asthma).
After informed consent, eligible participants underwent a staff-administered baseline survey in which demographic information, medical history, length of etonogestrel implant use, bleeding history in the 30 days before enrollment, and current level of satisfaction with bleeding pattern were collected. Current satisfaction with and intent to discontinue the etonogestrel implant as a result of unscheduled bleeding was assessed.
Participants received a pill bottle containing the study drug that was created and dispensed by the Investigational Drug Pharmacy at Barnes-Jewish Hospital Washington University in St. Louis containing either seven 15-mg ulipristal acetate capsules or seven identical-appearing placebo capsules according to the randomization scheme. Ulipristal acetate tablets in a 30-mg dose were purchased from a manufacturer and delivered directly to the pharmacy. Ulipristal acetate tablets were cut in half, weighed, and overencapsulated with gelatin capsules. Placebo capsules were made to look identical to ulipristal acetate capsules. Pill bottles were labeled by participant identification number and participant name.
Creation of the randomization scheme was performed by the study statistician in a one-to-one fashion with block sizes of eight using a computer-generated random allocation rule. Once consent was obtained, a copy of the consent as well as a prescription for study medication was sent to the research pharmacy. The study medication was delivered by the pharmacist to the site of study enrollment. The pharmacy and study statistician were the only individuals with access to the randomization scheme. Unblinding occurred only after completion of all data collection.
Participants took their first dose of medication before leaving their enrollment visit and then daily thereafter for a total of 7 days. Before medication administration, a verification text message (through a third-party service, Twilio, through Research Electronic Data Capture12) was sent to the participant to verify that she could receive text messages. Condoms and the remainder of the study medication were given to the participant at the end of her enrollment visit.
Participants completed daily bleeding diaries through automated text message to evaluate the effectiveness of ulipristal acetate in reducing bleeding days resulting from the etonogestrel implant over a 30-day period as compared with placebo. For 30 days, participants used single digits to code if they had taken their medication (for first week of study period only), if they had bleeding, and, if so, how much based on the Pictorial Blood Assessment Chart. If the initial text message was not answered, it was automatically sent up to two subsequent times that day. Diaries were monitored every 72 hours to assess for completeness by the study team. Participants with missing data were contacted by phone to complete the missing logs and assess desire to continue the study. A participant was considered lost to follow-up after 2 weeks of no contact.
Participants were called weekly for 4 weeks to complete a follow-up survey. This survey assessed ease of medication use (first week only), medication satisfaction, any side effects experienced, and whether participants used condoms if sexually active. Twenty participants from both groups volunteered to take part in a study subset assessing ovulation status. These participants had weekly in-person follow-up visits where the same survey was administered in addition to a blood draw testing serum progesterone levels. Blood was drawn by a medical assistant or nurse and immediately sent to the Core Laboratory at Washington University in St. Louis for analysis. Potential for ovulation was defined as a progesterone level above 5 ng/mL.13 All participants were called at the end of the study to complete an exit survey assessing bleeding pattern satisfaction and whether their bleeding was bothersome enough to have the etonogestrel implant removed. Participants were compensated with gift cards for their time and effort.
Participant characteristics are presented as means, SDs, frequencies, and percentages. Length of implant use and number of bleeding days were evaluated using medians and interquartile ranges. We used a χ2 test for categorical data and Student t test or Mann-Whitney U test for continuous variables, where appropriate (Table 1). Number of bleeding days during the study calculated from daily bleeding diaries was compared using a Mann-Whitney U test. Participant satisfaction with bleeding pattern, side effects experienced, medication satisfaction, and desire to continue using the contraceptive implant are presented with frequencies and percentages. χ2 test and Fisher exact test were used, where appropriate. All analyses were performed using SAS 9.4. A P value <.05 was considered statistically significant.
We calculated that 104 women (52 in each group) would be required to detect a 30% reduction in number of bleeding days with an α of 0.05 and 80% power. Previous literature demonstrates that women with the contraceptive implant have an average of 20 bleeding days per month14; therefore, our effect size was a decrease by 6 days, which we felt to be a clinically relevant estimate. Because there was no literature regarding the use of ulipristal acetate for bleeding secondary to the etonogestrel implant, we extrapolated from literature regarding other therapies used for unscheduled bleeding secondary to the etonogestrel implant to define our standard deviation of 10. We assumed a 15% loss to follow-up rate. In December 2017, we were notified by the FDA that a hold was being placed on all research evaluating the use of ulipristal acetate. We had previously performed an interim analysis halfway through enrollment that showed a pooled SD of 7, which was lower than previously accounted for. Communication with the FDA indicated a likely lengthy delay. The data safety and monitoring board was then consulted and given the findings of the interim analysis, the decision was made to perform a second interim analysis when the last currently enrolled participant completed their 30-day follow-up window. The pooled SD was again found to be 7. Additionally, we also noted a much smaller loss to follow-up rate (3%) than we previously planned for. Using the observed SD of 7, loss to follow-up rate of 3%, and the completed enrolled sample size of 65, we determined the study had achieved 90% power to assess for a 30% reduction in bleeding days and, thus, enrollment was stopped.
We assessed a total of 336 women for eligibility. Sixty-five women were ultimately randomized with 33 allocated to the placebo group and 32 to the ulipristal acetate group. Two participants in the placebo group were lost to follow-up (no longer answered text messages or telephone calls), leaving 63 participants for analysis, 31 in the placebo group and 32 in the ulipristal acetate group (Fig. 1). All participants reported taking all seven pills as dispensed. There were three participants missing a total of nine bleeding diaries from the 30-day period for 0.48% missing data. As a result of the extremely low rate of loss to follow-up, imputation of missing values was not performed. Table 1 compares baseline demographic characteristics of the analysis sample. Groups were similar in all aspects, including length of implant use (median use in months, placebo 9.2, ulipristal acetate 10.9, P=.84) and reported number of bleeding days the month before enrollment (median days, placebo 20.0, ulipristal acetate 20.5, P=.79).
At the end of the 30-day follow-up period, the median number of bleeding days was 12.0 (interquartile range 6–21) in the placebo group and 7.0 (interquartile range 4.5–11) in the ulipristal acetate group (P=.002). Three women in the placebo group (9.7%) and 11 women in the ulipristal acetate group (34.4%) had bleeding cessation by day 10 (P=.03). Figure 2 presents participant satisfaction with bleeding pattern at baseline and at the end of the study. At baseline, both groups had similar satisfaction levels. Women in the ulipristal acetate group were more likely to be satisfied with their bleeding pattern at the end of the study; 71.9% of women in the ulipristal acetate group were “very happy” with their bleeding as opposed to 26.7% in the placebo group (P<.001). Whereas 27 of 32 women in the ulipristal acetate group at study initiation reported they were “not at all happy” or “somewhat unhappy” with their bleeding at baseline, there were no participants reporting these levels of dissatisfaction at the conclusion of the study. At baseline, groups were similar in their desire to have their implant removed as a result of bothersome bleeding with the majority stating they would desire removal (placebo group 63.3%, ulipristal acetate group 67.7%, P=.72). At the end of the study, more women in the ulipristal acetate group desired to keep their implant than in the placebo group (89.7% vs 63.3%, P=.03).
There were few side effects reported during the treatment period, including headache (19.4% in the placebo group, 9.4% in the ulipristal acetate group) and nausea and vomiting (9.7% in the placebo group, 3.1% in the ulipristal acetate group). Groups were similar in side effects experienced with the majority experiencing no side effects (placebo group 74.2%, ulipristal acetate group 81.3%, P=.5). We did not encounter any serious adverse events. All participants from both groups felt their medication was easy to use. One hundred percent of ulipristal acetate users stated they would use the medication again for bleeding, which was significantly different than placebo users (66.7%; P=.001).
Although not powered to detect a difference for ovulatory status, no participants in either group (subset of 20 per group) were noted to have ovulatory progesterone levels (placebo group range: less than 0.2–1.3 ng/mL; ulipristal acetate group range: less than 0.2–4.4 ng/mL).
We found that administration of ulipristal acetate was associated with a reduction in the number of bleeding days resulting from the etonogestrel implant over a 30-day period as compared with placebo. Women who used ulipristal acetate had higher levels of satisfaction with their bleeding pattern at the end of the study period compared with placebo and would be more likely to continue the implant as their contraceptive method. Placebo users did become more satisfied with their bleeding and implant, which may be the result of objective measuring of bleeding days as well as a placebo effect. Ulipristal acetate users identified minimal side effects, found the medication easy to use, and would use the medication again, if necessary. Ovulatory progesterone levels were not noted; therefore, etonogestrel implant efficacy was not felt to be altered, but more research is needed to confirm because we were not powered for this outcome.
Numerous studies have investigated therapies for the treatment of unscheduled bleeding with progestin-only contraceptive methods; however, none included the use of ulipristal acetate with the etonogestrel implant.3,4 The mechanism by which ulipristal acetate may reduce unscheduled bleeding associated with the etonogestrel implant is unclear, although it may be the result of antiproliferative effects.5
Our study has limitations. First, no pilot data were available to determine the ideal dosing and duration for ulipristal acetate for bothersome bleeding in implant users. Our regimen was chosen based on feasibility; the currently available dose of ulipristal acetate in the United States is 30 mg, allowing the pharmacy to easily cut the pill in half for a 15-mg dose. We felt a 7-day regimen would be acceptable to women. We did not design this study with a run-in period; therefore, the number of bleeding days in the 30 days before enrollment was self-reported, which may be less accurate as a result of recall bias. We felt that a run-in period would have led to difficulties with recruitment and higher rates of loss to follow-up because many women who present with unscheduled bleeding want immediate treatment or implant removal. Although we ultimately found a lower number of days in our placebo group than reported, it is possible that women unhappy with their bleeding may overestimate their number of bleeding days. It is important to note, however, that both groups reported the same amount of bleeding days initially. We did need to end our study early as a result of a FDA hold regarding liver toxicity in some women using Esmya (5 mg ulipristal acetate daily for 3 months for uterine leiomyoma) overseas. Since our study completion, the European Medicines Agency has determined that Esmya should not be used in women with known liver problems and that liver function tests should be followed before, during, and after use. Esmya is a much longer course of ulipristal acetate than we have proposed, but further research should likely adhere to these recommendations.15 Finally, we were not powered to detect an effect on ovulatory status, which was used as a proxy for etonogestrel implant efficacy. To detect such a small difference would likely require a sample size larger than 500. We are reassured, however, that no ovulatory progesterone levels were found in the subset of either treatment group.
Our study also has many strengths, which include being a randomized, double-blind, placebo-controlled trial. Use of automated text messaging for daily bleeding diaries led to low dropout and missing data rates, leading to greater internal validity. Additionally, ulipristal acetate is a novel therapy for unscheduled bleeding associated with the etonogestrel implant. Unlike many other therapies used for unscheduled bleeding resulting from progestin-only contraception, ulipristal acetate has few contraindications and, thus, could be used more widely. Finally, because we had very few exclusion criteria and participants were racially and economically diverse, we feel the results are generalizable to reproductive-aged women.
Our study demonstrates that in women with unscheduled bleeding resulting from the etonogestrel implant, ulipristal acetate use led to a greater decrease in bleeding days compared with placebo as well as increased satisfaction with the bleeding pattern and the etonogestrel implant. Increased satisfaction with the etonogestrel implant may lead to a decrease in discontinuation rates and potentially a decrease in unintended pregnancy rates in this population. The only other study evaluating ulipristal acetate for progestin-mediated bleeding (50 mg ulipristal acetate for 3 days beginning on days 21, 49, and 77 after levonorgestrel intrauterine device insertion) showed benefit in the short term but not long term.16 However, a different progestin was used in this study and therapy was given to prevent bleeding, not treat it. Further investigation of ulipristal acetate in different dosages and regimens as well as for longer follow-up periods is warranted.
1. ESHRE Capri Workshop Group. Ovarian and endometrial function during hormonal contraception. Hum Reprod 2011;16:1527–35.
2. Grunloh DS, Casner T, Secura GM, Peipert JF, Madden T. Characteristics associated with discontinuation of long-acting reversible contraception within the first 6 months of use. Obstet Gynecol 2013;122:1214–21.
3. Abdel-Aleem H, d’Arcangues C, Vogelsong KM, Gaffield ML, Gulmezoglu AM. Treatment of vaginal bleeding irregularities induced by progestin only contraceptives. The Cochrane Database of Systematic Reviews 2013, Issue 10. Art. No.: CD003449. DOI: 10.1002/14651858.CD003449.pub5.
4. Zigler RE, McNicholas C. Unscheduled vaginal bleeding with progestin-only contraceptive use. Am J Obstet Gynecol 2017;216:443–50.
5. Donnez J, Tatarchuk TT, Bouchard P, Puscasiu L, Nataliya F, Zakharenko T, et al. PEARL I Study Group. Ulipristal acetate versus placebo for fibroid treatment before surgery. N Engl J Med 2012;366:409–20.
6. Donnez J, Tomaszewski J, Vázquez F, Bouchard P, Lemieszczuk B, Baró F, et al. PEARL II Study Group. Ulipristal acetate versus leuprolide acetate for uterine fibroids. N Engl J Med 2012;366:421–32.
7. Kalampokas T, Kamath M, Boutas I, Kalampokas E. Ulipristal acetate for uterine fibroids: a systematic review and meta-analysis. Gynecol Endocrinol 2016;32:91–6.
8. Barlow DH, Lumsden MA, Fauser BC, Terrill P, Bestel E. Individualized vaginal bleeding experience of women with uterine fibroids in the PEARL I randomized controlled trial comparing the effects of ulipristal acetate or placebo. Hum Reprod 2014;29:480–9.
9. Nieman LK, Blocker W, Nansel T, Mahoney S, Reynolds J, Blithes D, et al. Efficacy and tolerability of CDB-2914 treatment for symptomatic uterine fibroids: a randomized, double-blind, placebo-controlled, phase IIb study. Fertil Steril 2011;95:767–72.e1–2.
10. Moher D, Hopewell S, Schulz KF, Montori V, Gotzsche PC, Devereaux PJ, et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. J Clin Epidemiol 2010;63:e1–37.
11. Mansour D, Korver T, Marintcheva-Petrova M, Faser IS. The effects of Implanon on menstrual bleeding patterns. Eur J Contracept Reprod Health Care 2008;13(suppl 1):13–28.
12. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377–81.
13. Leiva R, Bouchard T, Boehringer H, Abulla S, Ecochard R. Random serum progesterone threshold to confirm ovulation. Steroids 2015;101:125–9.
14. Simmons KB, Edelman AB, Fu R, Jensen JT. Tamoxifen for the treatment of breakthrough bleeding with the etonogestrel implant: a randomized controlled trial. Contraception 2017;95:198–204.
15. European Medicines Agency. PRAC recommends new measures to minimise risk of rare but serious liver injury with Esmya for fibroids. London (UK): European Medicines Agency.
16. Warner P, Guttinger A, Glasier AF, Lee RJ, Nicerkson S, Brenner RM, et al. Randomized placebo-controlled trial of CDB-2914 in new users of a levonorgestrel-releasing intrauterine system shows only short-lived amelioration of unscheduled bleeding. Hum Reprod 2010;25:345–53.