Post Hoc Analysis of a Previous Study of Estetrol 15 mg/Drospirenone 3 mg Combination for Dysmenorrhea and Uterine Bleeding in Adenomyosis Patients

The previous study protocol (FSN‐013P‐03) was approved by the institutional review board at each study site. Throughout the course of the study, monitors made regular visits to each study site to ensure adherence to the protocol.

This work was supported by the Fuji Pharma Co., Ltd.

Of the 162 participants included in the full analysis set, 36 women diagnosed with adenomyosis were included in each of the E4/DRSP and placebo groups. One participant in the placebo group was withdrawn from the study before initiation of the open label phase due to poor compliance with the study protocol. Demographic characteristics were comparable between treatment groups, with a mean age of 35.8 years and a mean body mass index of 21.84 kg/m 2 . Neither gravidity nor parity differed significantly between groups, and approximately 50% of participants graded bleeding episodes as “heavy” (Table  S1 ). Of note, approximately 71% of participants with adenomyosis had neither endometriosis nor fibroids as a comorbidity, which means that participants with adenomyosis would make up the major population, representing the ad‐hoc analysis results. Adenomyosis was diagnosed using TVUS or MRI. All of the participants with endometriosis had endometrioma and cul‐de‐sac endometriosis (deep endometriosis), with the exception of one participant. On average, E4/DRSP reduced the total dysmenorrhea score at Week 16 by 2.2 from baseline (4 treatment cycles). The point estimate of the difference (E4/DRSP—placebo) was inferred to be −1.8 with a two‐sided 95% CI of −2.5 to −1.1 ( p  < 0.001; Table  1 ). Responder rates were significantly higher in the E4/DRSP group (66.7%, two‐sided 95% CI 47.2%–82.7%) than in the placebo group (20.0%, two‐sided 95% CI 8.4%–36.9%; p  < 0.001) at Week 16 (Table  1 ). Changes in dysmenorrhea scores from baseline and responder rate following 16‐week double‐blinded treatment in participants with adenomyosis. Abbreviations: E4/DRSP, estetrol 15 mg/drospirenone 3 mg; LSM, least squares mean. Values represent LSM from ANCOVA model inferring missing data by last observation carried forward method. Values represent the proportion of subjects who achieved a reduction of ≥ 2 in dysmenorrhea score from baseline. Values represent mean ± standard deviation. Two‐sided 95% confidence interval. Two‐sided 95% confidence interval with the Clopper–Pearson method. Following the four cycles of double‐blinded treatment, the VAS score for adenomyosis‐associated pelvic pain (lower abdominal and back pain) during menstrual bleeding periods decreased by 47.3 mm from baseline, and the difference from placebo was −32.3 mm (two‐sided 95% CI −46.3 to −18.3 mm; p  < 0.001). Similar results were obtained from the VAS for dysmenorrhea symptoms (Table  2 ). Effect of E4/DRSP on VAS scores (symptoms, pelvic pain) in participants with adenomyosis following 16‐week double‐blinded treatment. Abbreviation: E4/DRSP, estetrol 15 mg/drospirenone 3 mg. Values represent mean ± standard deviation, after excluding participants with no bleeding events. Values represent mean from two‐sample t ‐test based on data obtained at last visit. Two‐sided 95% confidence interval. Pelvic pain VAS during the menstrual bleeding period improved steadily during the additional 36‐week open‐label extension for both continued E4/DRSP treatment and placebo treatment switched to E4/DRSP in Week 16 (Figure  1 ). In addition, a slightly larger but non‐significant alleviation appeared in the VAS score for non‐menstrual pelvic pain following 16‐week E4/DRSP treatment compared to placebo treatment (Table  2 ). Alleviation of pelvic pain over the course of E4/DRSP treatment. Placebo was switched to E4/DRSP following placebo treatment for 16 weeks. Missing data due to immature termination was not imputed. E4/DRSP, estetrol 15 mg and drospirenone 3 mg. At Week 16, use of rescue medication tended to be slightly decreased (2.15 ± 3.611 days) with E4/DRSP treatment compared to placebo treatment (3.51 ± 2.934 days). However, no significant difference was observed. Both daily activity and sleep quality were significantly improved after 16 weeks of E4/DRSP treatment. The proportion of participants reporting the top two categories of “not disturbed at all” and “not much disturbed” increased by 61.1% for daily activity and 75.0% for sleep quality, significantly better than with placebo treatment ( p  = 0.0329 for daily activity, p  = 0.0287 for sleep quality) (Figure  2 ). Similarly, an advantage was seen for E4/DRSP treatment in terms of PGI‐I, translating to 55.6% of participants reporting the top two categories (“very much improved” and “much improved”), significantly better than with placebo (8.3%, p  < 0.001) (Figure  2 ). Further, 16 weeks of E4/DRSP treatment tended to improve the 11‐point scale for working efficiency by 54% compared to baseline, slightly larger than that with placebo (31%). However, this result was not significant. Proportion of participants rated as improved. Left, middle and right sets of two bars indicate interference with daily activity and sleeping (“not disturbed at all” or “not much disturbed”), and PGI‐I (“markedly improved” or “much improved”), respectively. E4/DRSP, estetrol 15 mg and drospirenone 3 mg; PGI‐I, patient global impression of improvement. TEAEs were reported in 34 of the 36 participants taking E4/DRSP (94.4%) and 29 of the 36 participants with placebo (80.6%) over the four treatment cycles (Table  S2 ). Intermenstrual bleeding was the most frequently reported TEAE with E4/DRSP, and the incidence decreased with increasing treatment cycles. Drug‐related TEAEs were characterized as those commonly reported for COC products, including nausea (11.1%) and headache (5.6%). At Week 52, overall profiles for TEAEs were almost the same as those at Week 16 (Table  S2 ). Throughout the 16‐week placebo treatment period, 58% of participants graded menstrual bleeding as “more than normal”, but this proportion was markedly diminished to 6% after the switch to E4/DRSP for 36 weeks (Figure  3 ), similar to that following 52 weeks of E4/DRSP. No deaths or VTE occurred in any participants, as reported previously [ 24 ]. At Week 16, the proportion of participants with D‐dimer level ≥ 1 μg/mL (the upper limit of the reference range) was comparable between treatments, with one participant in each (2.8%). Proportion of participants who graded bleeding events as “more than normal” with the originally allocated placebo treatment. Placebo was switched to E4/DRSP at the initiation of Cycle 5. E4/DRSP, estetrol 15 mg and drospirenone 3 mg.

This post hoc analysis of the previous study [ 24 ] demonstrated that E4/DRSP significantly decreased total dysmenorrhea scores in participants with adenomyosis compared to placebo. The between‐treatment difference in change in total dysmenorrhea score from baseline was estimated to be −1.8 after four cycles of E4/DRSP. The responder rate, defined as the proportion of participants with total dysmenorrhea score reduced by ≥ 2 points from baseline, was significantly larger in the E4/DRSP group (66.7%) than in the placebo group (20.0%). Further, the VAS score for menstrual pelvic pain after 16 weeks of treatment was 32.3 mm lower with E4/DRSP than with placebo ( p  < 0.001), appearing comparable to results for progestin as reported in a previous clinical study (−37.8 mm) [ 29 ]. Also of note was the finding that the between‐treatment difference in non‐menstrual pelvic pain VAS score was approximately −11 mm, close to the minimal clinically important difference [ 30 ]. These desirable effects would likely underpin the improvements in daily activity, sleep quality, and PGI‐I following 16 weeks of E4/DRSP treatment, supporting the importance of managing pelvic pain in maintaining overall QoL. Chronic pain conditions also increase the stress associated with health conditions such as sleeping disturbance, which would translate into increased lesional and systemic levels of pain‐producing substances such as prolactin acting on nociceptors [ 31 , 32 , 33 ]. No formal guidelines have been established to prioritize treatment modalities from one to the other; instead, long‐term therapy is essentially required. Adenomyosis is hormone‐dependent, so endocrine therapies, estrogen/progestin combinations, GnRHa and progestin alone are used to treat adenomyosis. The peak age for susceptibility to adenomyosis is considered to be around the late 30s, but advances in imaging techniques have revealed that the disease can be diagnosed even in approximately 30% of adolescents and younger women with dysmenorrhea and/or HMB [ 8 ]. Other researchers have referred to the same considerations for the onset of adenomyosis, so medical treatments should also consider bone health, for which a certain level of estrogen needs to be maintained. GnRHa negatively affect BMD, as estradiol levels decline due to the mode of action, raising concerns regarding the use of progestin alone. E4/DRSP also diminishes estradiol levels, but E4 exerts weak estrogenic activity and, more importantly, a previous clinical study indicated that E4/DRSP had no clear adverse impact on insulin‐like growth factor I, one of the key factors relevant to bone metabolism [ 20 , 34 ]. No particular TEAEs were reported following E4/DRSP treatment among participants with adenomyosis in this study. One typical symptom appears to be increased menstrual bleeding, as AUB. Placebo treatment was continued over the course of the double‐blinded phase in this study, then was switched to E4/DRSP. This extension part clearly indicated that E4/DRSP decreased the proportion of participants experiencing AUB, from approximately 60% in the double‐blinded phase with placebo to 6% at the end of the open‐label phase with E4/DRSP. This proportion was quite similar to that for the E4/DRSP group over the entire 52‐week course of treatment, at approximately 7%. Caution is required regarding the use of DNG because frequent AUB can induce severe anemia, so the present findings suggest that E4/DRSP may be favorable as an adenomyosis treatment without specific caution regarding AUB leading to anemia. The blood coagulation/fibrinolysis system is reported to be enhanced in patients with adenomyosis, accompanied by an increase in the potential risk of VTE. The present study showed a reduced impact on D‐dimer levels, consistent with results from early clinical studies and suggesting that E4/DRSP may be favorable even with the disturbed hemostatic condition of patients with this pathology. This post hoc analysis supports E4/DRSP as a suitable treatment for adenomyosis due to alleviation of pelvic pain as robustly determined from several measurement indices, including dysmenorrhea score, VAS, QoL‐related questionnaires, menstrual bleeding and impact on hemostasis. Even so, several limitations should be carefully considered for this post hoc analysis. First, no randomization was performed, meaning that potential biases were not necessarily excluded. Second, the sample size was not estimated from an assumed statistical hypothesis, so the statistical power may have been insufficient. Finally, no morphological evaluations (such as asymmetrical thickening of the myometrium) were performed. Nevertheless, the results suggest that further clinical studies are warranted to confirm whether E4/DRSP is effective for the amelioration of various adenomyosis‐associated symptoms.

All patients provided written informed consent.

Adenomyosis is a common benign condition that often manifests as dysmenorrhea, chronic pelvic pain (lower abdominal and back pain), abnormal uterine bleeding (AUB), heavy menstrual bleeding (HMB) and infertility [ 1 , 2 , 3 , 4 ]. Despite tremendous efforts, the pathogenesis of adenomyosis remains poorly understood and classification of the disease has thus remained uncertain. Chapron et al. proposed the “outside to inside invasion” theory [ 5 ]. Patients with deep infiltrating endometriosis show a high prevalence of posterior focal adenomyosis of the outer myometrium, which could mean that ectopic endometrial debris after retrograde menstruation might infiltrate the pelvic organs, which would support the “outside to inside invasion” hypothesis. Nonetheless, the variability of clinical phenotypes cannot be comprehensively explained by this theory alone. As such, established guidelines remain inadequate not only for diagnosing and classifying the disease, but also for tailoring adequate therapeutic modalities. As a result, individualization of therapeutic management is currently only available based on age, symptoms and desire for fertility [ 6 ]. Similar to endometriosis, an estrogen‐dependent, progesterone‐resistant condition, endocrine therapies are first used to alleviate the symptoms caused by adenomyosis [ 7 ]. Advanced imaging techniques such as transvaginal ultrasonography (TVUS) and magnetic resonance imaging (MRI) have allowed modification of the etiological backgrounds to adenomyosis. Although age over 40 years is included as a common risk factor for adenomyosis, the pathology is increasingly being diagnosed among young women with pelvic pain and AUB. A recent observational cohort study of symptomatic women between 13 and 25 years old found that concomitant dysmenorrhea and HMB were significantly associated with adenomyosis [ 8 ]. From the perspective of the balance between risk and benefit, these etiological modifications essentially necessitate medical treatment not only to preserve fertility, but also for life‐long management of the disease. Gonadotropin‐releasing hormone analogs (GnRHa) appear likely to ease the symptoms of adenomyosis, but their clinical efficacy remains contentious due to the lack of data on potential disease progression and the scarcity of relevant literature [ 9 ]. Further, GnRHa treatments can lead to menopause‐like symptoms such as hot flushes and impaired bone mineral density (BMD), requiring consideration of the trade‐off between maintaining estrogen levels and clinical efficacy. This has led to GnRHa treatment being limited to 6‐month use for adenomyosis. While estrogen and progestin combinations containing GnRHa have been developed, their use remains limited to 24 months. For younger women with adenomyosis, in particular, GnRHa use requires attention to effects on bone health due to the impairment of estrogen levels. Similar concerns are seen with medical treatments using progestin alone, such as dienogest (DNG) [ 10 , 11 ]. Further, as the primary phenotype of adenomyosis symptoms, HMB can lead to the development of severe anemia, which would then pose questions as to the adequacy of bleeding control. Patients on DNG are known to frequently experience AUB, reported as immature medical control in patients with pre‐existing anemia over the course of long‐term use [ 9 ]. In addition, DNG likely has the potential to negatively affect bone health in association with declining systemic levels of estrogen [ 10 , 11 ], and a next‐generation sequencing investigation revealed certain genetic mutations related to non‐response to DNG therapy [ 12 ]. Combined oral contraceptives (COCs) are also known to be effective for ameliorating adenomyosis‐associated symptoms, but the literature on their use for this purpose remains limited [ 9 ]. Aside from clinical efficacy, safety concerns such as the risk of thromboembolic events need to be carefully considered over the course of disease treatment, as blood coagulation and fibrinolysis are likely to be enhanced in women with adenomyosis [ 13 ]. Likewise, caution is required regarding the safety profile of GnRHa preparations containing estrogen and progestin combinations when used as add‐back therapy. Estetrol (E4) is a native estrogen exclusively produced in the fetal liver, showing greater affinity to estrogen receptor (ER)α than to ERꞵ. E4 also shows a unique mode of action, acting as an agonist to nuclear ERα but an antagonist to membrane ERα, manifesting as various estrogenic activities [ 14 , 15 , 16 ]. Early clinical studies have revealed that the E4/Drospirenone (DRSP) combination shows reduced effects on the blood coagulation and fibrinolysis system, disturbance of which is likely to increase the risk of thromboembolic events such as venous thromboembolism (VTE) [ 17 ]. Kobayashi et al. recently reported that E4/DRSP had less impact on acquired activated protein C resistance, a potential index for the risk of VTE, compared to the ethinyl estradiol and DRSP combination in both patients with endometriosis and healthy young women [ 18 , 19 ]. Further, E4/DRSP was found to have minimal impact on lipid and glucose metabolism [ 20 , 21 ]. The combination of E4 15 mg and DRSP 3 mg has demonstrated excellent contraceptive efficacy, good bleeding control, and no safety concerns in two different Phase III studies [ 22 , 23 ]. Regulatory approval for E4/DRSP as a COC was granted by both the European Union and the United States in 2021. In Japan, E4/DRSP was launched in 2024 as a medical treatment for primary and secondary dysmenorrhea. This study aimed to clarify whether E4/DRSP ameliorated adenomyosis‐associated pain and AUB and eventually improved patient quality of life (QoL), by conducting post hoc analyses of data obtained from a randomized clinical trial of Japanese patients with dysmenorrhea [ 24 ].

Tasuku Harada, M.D., Ph.D., Takao Kobayashi, M.D., Ph.D., and Yutaka Osuga, M.D., Ph.D. received consulting fees from Fuji Pharma Co., Ltd. The other authors declare no conflicts of interest.

The study design has been provided in detail elsewhere [ 24 , 25 ]. In brief, the study comprised two parts: a 16‐week, double‐blinded phase to confirm the superior clinical efficacy of E4/DRSP over placebo; and an additional 36‐week, open‐label phase to clarify long‐term safety and efficacy, switching all patients previously receiving placebo to E4/DRSP at Week 16. The original study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice protocols and met all local legal and regulatory requirements. The study protocol was reviewed and approved by the institutional review boards of each study site. Written, informed consent was obtained from all participants prior to enrolment. Women ≥ 20 years old who had been diagnosed with primary or secondary dysmenorrhea were enrolled. Another critical eligibility criterion was total dysmenorrhea score ≥ 3 during the baseline observation period over two menstrual cycles. The primary inclusion criterion of participants with secondary dysmenorrhea was diagnosis results obtained from ultrasonography (TVUS), magnetic resonance imaging (MRI), laparotomy, or laparoscopy, indicating adenomyosis, endometriosis with endometrioma, or uterine fibroids. Adenomyosis was diagnosed using either TVUS or MRI [ 25 ]. Other eligibility criteria were as described in a previous report [ 24 , 25 ]. In the double‐blinded phase, eligible participants were randomly allocated to either the E4/DRSP group or the placebo group at an equivalent ratio balanced by dysmenorrhea classification (primary or secondary). Participants took one tablet daily from the first day of menstruation after randomization in a double‐blinded manner. Participants in the E4/DRSP group were treated with E4/DRSP in a cyclic regimen (1 cycle = E4/DRSP for 24 days, followed by a 4‐day hormone‐free interval receiving placebo) for four cycles over 16 weeks. The placebo group received oral placebo tablets daily for 28 days per cycle over four cycles. After the 16‐week, double‐blinded phase, all participants received an additional 36 weeks of E4/DRSP treatment in the same cyclic regimen as the open‐label extension phase. In both groups, participants were allowed to take non‐steroidal anti‐inflammatory drugs (loxoprofen or ibuprofen) as rescue medications as appropriate throughout the study. Randomization codes were developed and managed by an office independent of the clinical investigators and other study stakeholders to maintain blindness. Post hoc analyses were conducted for the following measures: (1) change in total dysmenorrhea score, as the sum of “interference with working/daily activities” and “use of analgesics” scores [ 26 ], from baseline to Week 16 of treatment; (2) responder rates of patients achieving a ≥ 2.0‐point reduction in total dysmenorrhea score from baseline; (3) visual analog scale (VAS) scores for adenomyosis‐associated pelvic pain (lower abdominal and back pain) during menstrual and non‐menstrual bleeding periods; (4) daily use of rescue medication; (5) 5‐point scales for daily activity and sleep quality; (6) a 7‐point patient global impression scale (PGI‐I); and (7) an 11‐point scale for work efficiency impairment due to health problems over the past 7 days. All variables were collected from an e‐diary device used by participants, and from examinations by investigators at clinical sites. For safety evaluations, all treatment‐emergent adverse events (TEAEs) were reported throughout the study, and investigators determined their severity and causality. Bleeding events were also recorded using the e‐diary device accompanying severity. Clinical laboratory tests (including hemostasis parameters) were also conducted. As stated in a previous report [ 24 ], the sample size estimation considered a between‐group difference in changes to total dysmenorrhea score of −1.0 from baseline with 90% power at an α level of 0.05 (two‐sided) for confirmatory purposes [ 24 ]. This post hoc analysis was conducted to clarify whether E4/DRSP provides favorable clinical efficacy against adenomyosis‐associated symptoms. For dysmenorrhea score, between‐treatment differences (E4/DRSP—placebo) were estimated using analysis of covariance (ANCOVA) with treatment type as a fixed effect, adjusting for the corresponding baseline score. Within the framework of ANCOVA, two‐sided 95% confidence intervals (CIs) were inferred. Two samples t ‐tests were applied to VAS scores for both dysmenorrhea symptoms and pelvic pain between‐treatment differences (E4/DRSP—placebo) with two‐sided 95% CIs. The responder rate was defined as the proportion of participants for whom total dysmenorrhea score was reduced ≥ 2 points from baseline. Likert‐type scales (i.e., daily activity, sleep quality and PGI‐I) were collapsed by the two highest scores, using the remaining scores to categorize the case as “improved” or “unchanged/worsened”. Fisher's exact test was applied for dichotomous variables, inferring two‐sided 95% CIs. Working efficiency was characterized by changes in an 11‐point scale (from 0, “working not affected at all”, to 10, “working completely prevented”) from initiation of treatment [ 27 , 28 ], and the two‐sample t ‐test was applied including days of rescue use at Week 16. All efficacy analyses were conducted on the analysis set of participants who had received one or more study drugs and for whom all evaluation measures were obtained. Safety analyses included the frequencies of any TEAEs, drug‐related TEAEs, and severity and causality using the analysis set of participants who consumed at least one study tablet. Descriptive statistics were determined for continuous variables and expressed as mean ± standard deviation. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Table S1: Demographic and baseline characteristics (FAS). Table S2: Any TEAEs.