Prognostic factors of progesterone resistance in symptomatic adenomyosis: impact of lesion localization on treatment outcome of levonorgestrel intrauterine system

Abstract

Adenomyosis often causes dysmenorrhea in women of reproductive age. Progestins such as levonorgestrel intrauterine system (LNG-IUS) are often used for treatment, but some patients experience progesterone resistance, showing poor treatment response. However, the clinical characteristics of progesterone-resistant and progesterone-sensitive patients with symptomatic adenomyosis remain unclear. We analyzed data of 69 patients with adenomyosis treated with LNG-IUS. Dysmenorrhea was quantified using linear visual analog scale (VAS) scoring, and progesterone resistance was interpreted as continued dysmenorrhea during LNG-IUS treatment. The rate of change in VAS scores of dysmenorrhea was calculated: patients with the bottom 25% improvement were defined as progesterone-resistant group, and those with the top 25% improvement as progesterone-sensitive group. The localization of adenomyosis lesions was evaluated by magnetic resonance imaging (MRI) and classified as advanced (localized in all layers of the myometrium), extrinsic (localized on the uterine serosa side), and intrinsic (localized on the endometrial side) subtypes. Progesterone-resistant group had a significantly lower incidence of intrinsic adenomyosis (7.7% vs. 69.2%, p = 0.004) and a tendency toward a higher incidence of advanced adenomyosis (61.5% vs. 23.1%, p = 0.111) compared with progesterone-sensitive group. Progesterone-sensitive group showed significant improvement of dysmenorrhea 1 month after starting LNG-IUS treatment (p < 0.001). These findings indicate that the responsiveness to LNG-IUS treatment can be determined 1 month after starting the treatment and that intrinsic adenomyosis is a favorable prognostic factor for progestin treatment with LNG-IUS, while advanced and extrinsic adenomyosis are predictors for progesterone resistance.

Introduction

Adenomyosis is a benign gynecological disease characterized by pathological displacement of the endometrium in the myometrium, thereby triggering hypertrophy of the surrounding myometrium and causing uterine enlargement and various symptoms, including dysmenorrhea and menorrhagia [1]. Adenomyosis is found in approximately 20% of reproductive-age women, and uterine-sparing treatment is often desired [2]. For patients who wish to preserve their uterus, hormonal treatments such as low-dose estrogen-progestin, progestins, and gonadotropin-releasing hormone analogues (GnRH-a) are administered, and progestins such as Levonorgestrel Intrauterine System (LNG-IUS) and dienogest (DNG) are considered good options in terms of long-term treatment with less adverse events [3,4,5]. However, some patients respond poorly to progestins, resulting in severe dysmenorrhea [4, 6]. This poor response is known as progesterone resistance, a condition of the endometrium that involves a disruption of the progesterone signaling pathway, resulting in unresponsiveness to progesterone [7,8,9]. Progesterone resistance is a significant clinical challenge in the management of adenomyosis; however, the clinical characteristics of patients with adenomyosis that result in progesterone resistance remain unclear. Therefore, in this study, we retrospectively assessed predictive factors associated with progesterone resistance in patients with adenomyosis who were treated with LNG-IUS.

Materials and methods

Data collection

This study was performed in accordance with the Declaration of Helsinki and the Ethical Guidelines for Medical and Biological Research Involving Human Subjects formulated by the Japanese government. This study was reviewed and approved by the Research Ethics Committee of the Faculty of Medicine of the University of Tokyo (IRB number: 3128). The collection of informed written consent was substituted with the informed opt-out procedure because of the retrospective nature of the study. Information about this study was posted on the website of the University of Tokyo Hospital to give participants the opportunity to opt out; those who did not opt out were considered to have provided tacit consent for study participation. Anonymous clinical data were used for the analysis, and individuals cannot be identified based on the data presented. Waiving of written consent and the use of the informed opt-out procedure were approved by the Research Ethics Committee of the Faculty of Medicine of the University of Tokyo.

This study retrospectively analyzed deidentified medical records of 69 patients diagnosed with symptomatic uterine adenomyosis detected with magnetic resonance imaging (MRI) and treated with LNG-IUS (Mirena [52 mg], Bayer Yakuhin) between 2015 and 2024 at the University of Tokyo Hospital. Patients who were lost to follow-up, those with submucosal fibroids, endometrial polyps, or endometrial cancer, those without dysmenorrhea, and those who experienced expulsion of LNG-IUS within 6 months were excluded. Patients who were followed for longer than 6 months without expulsion of LNG-IUS after the insertion were included in the analyses. The degree of dysmenorrhea was evaluated using linear visual analog scale (VAS) scoring, which ranged from 0 cm (no pain) to 10 cm (worst pain possible) [10]. Clinical data used included age, nulliparity, symptoms, types and localization of adenomyosis, uterine size, the diameter of adenomyotic lesion, complications of leiomyoma or endometriosis, treatment history, hemoglobin level before LNG-IUS treatment and at 6 months after starting LNG-IUS treatment, and the VAS scores of menstrual pain and chronic pelvic pain before LNG-IUS treatment, and at 1 month, 3 months, 6 months, and 1 year after starting LNG-IUS treatment. The subtypes of adenomyosis were classified based on lesion localization on MRI according to the method of Kishi et al. [11]. Since no patients who were enrolled in this study were found to have an intramural adenomyosis (namely subtype 3 of the Kishi’s classification), the patients were classified into one of the following three subtypes: advanced (localized in all layers of the myometrium), extrinsic (localized on the uterine serosa side), and intrinsic (localized on the endometrial side) subtypes [11]. The uterine size (cm³) was estimated with the ellipsoid formula:

$$\begin{array}{c}= \,anteroposterior\,diameter\, \times \,longitudinal\,diameter\,\\\times \,transverse\,diameter\, \times \,0.5236\end{array}$$

(1)

[12].

Definition of progesterone resistance

Progesterone resistance is considered when patients with adenomyosis respond poorly to progestins, resulting in severe gynecological symptoms such as dysmenorrhea. However, there are no standardized clinical criteria for progesterone resistance in adenomyosis, as is also the case in endometriosis [9]. In this study, progesterone resistance was interpreted as the state in which dysmenorrhea continued after starting LNG-IUS treatment. Based on previous reports assessing changes in pain [13,14,15], the rate of change in VAS scores for menstrual pain was calculated as follows and used as an indicator of progesterone resistance:

$$= \,\frac{{(VAS\,at\,6\:months\, after\, LNG-IUS\,treatment)\, - \,(VAS\, before\, LNG-IUS\,treatment)}}{{\left( {VAS\, before\, LNG-IUS\,treatment} \right)}}$$

(2)

Based on previous reports on the classification of scoring systems [16, 17], patients with the bottom 25% improvement were defined as progesterone-resistant group, while those with the top 25% improvement were defined as progesterone-sensitive group. Thirteen patients were classified in each group and the clinical characteristics obtained from the medical records were compared.

Statistical analysis

All statistical data were analyzed using R version 4.3.2. Two groups were compared using Fisher’s exact test and Mann–Whitney U test. Pre- and post-treatment comparisons were performed using Wilcoxon signed-rank sum test. Furthermore, a p-value < 0.05 was considered statistically significant.

Results

Patient characteristics

Data of 69 patients with symptomatic adenomyosis detected with MRI and treated with LNG-IUS were reviewed. As the flowchart of this study (Fig. 1) shows, of the 69 patients, four patients who were lost to follow-up, four without dysmenorrhea, and eight with expulsion of LNG-IUS within 6 months were excluded. The characteristics of the remaining 53 patients are shown in Table 1. The average VAS scores before LNG-IUS treatment of 53 patients were significantly reduced 6 months after starting LNG-IUS treatment, showing the general effectiveness of LNG-IUS for the treatment of symptomatic adenomyosis (7.9 [6.4–8.9] vs. 2.3 [0.4–5.8], p < 0.001). As Fig. 2 shows, progesterone-resistant group and progesterone-sensitive group were classified based on the rate of change in VAS scores.

Fig. 1

The flowchart of this study

Table 1 Characteristics of 53 patients analyzed in this study. SD, standard deviation

Fig. 2

Classification of progesterone resistance

The 53 dots signify the rates of the change in VAS scores of menstrual pain of the 53 patients. Patients with the bottom 25% improvement were defined as progesterone-resistant group, while those with the top 25% improvement were defined as progesterone-sensitive group. Thirteen patients were classified in each group.

Intrinsic adenomyosis is a good prognostic factor, while extrinsic and advanced adenomyosis are predictors for progesterone resistance.

As Table 2 shows, the comparison between progesterone-resistant group and progesterone-sensitive group showed no differences in age, parity, hemoglobin level, complications with leiomyoma and endometriosis, localization of adenomyosis, the diameter of adenomyotic lesion, and uterine size. On the other hand, the rate of intrinsic adenomyosis was significantly lower in progesterone-resistant group (7.7% vs. 69.2%, p = 0.004), while the rate of advanced adenomyosis tended higher in progesterone-resistant group (23.1% vs. 61.5%, p = 0.111), indicating that intrinsic adenomyosis is a good prognostic factor for progesterone response and advanced and extrinsic adenomyosis are predictive factors for progesterone resistance.

Table 2 Clinical characteristics of the progesterone resistance group. SD, standard deviation

Progesterone resistance can be determined 1 month after starting LNG-IUS treatment

Figure 3a shows the change over time in the VAS scores of menstrual pain in the two groups. The VAS scores before LNG-IUS treatment did not differ between progesterone-resistant group and progesterone-sensitive group (6.4 [2.8-8.0] vs. 8.2 [6.4–8.9], p = 0.158), but progesterone-sensitive group showed significantly higher improvement rates compared to progesterone-resistant group 1 month after starting LNG-IUS treatment (6.2 [5.4-7.0] vs. 1.3 [0.0-1.7], p = 0.003), and this trend continued at 3 months, 6 months, and 1 year after starting LNG-IUS treatment. The same trend was also observed in terms of chronic pelvic pain: the VAS scores 1 month after starting LNG-IUS treatment were significantly different between the two groups, and the difference continued until 1 year after starting LNG-IUS treatment (Fig. 3b).

Fig. 3

(a) The comparison of the time-dependent change in the VAS scores of menstrual pain, (b)The comparison of the time-dependent change in the VAS scores of chronic pelvic pain

Discussion

In this study, we analyzed patients treated with LNG-IUS, a progestin medication for dysmenorrhea due to adenomyosis, to investigate the factors contributing to progesterone resistance. We found that intrinsic adenomyosis is a favorable prognostic factor, while advanced and extrinsic adenomyosis are predictors for progesterone resistance. Analysis of the VAS scores also indicated that progesterone resistance can be estimated at about 1 month after starting LNG-IUS treatment.

This study has revealed that intrinsic adenomyosis is a good prognostic factor, and subtypes other than intrinsic adenomyosis including advanced adenomyosis are predictors for progesterone resistance in terms of pain. However, the size of the uterus or adenomyosis and the presence of uterine fibroids or ovarian endometriomas are not predictive factors. Intrinsic adenomyosis is characterized by the development of adenomyotic lesions from the endometrial side of the uterus, with lesions confined to the junctional zone and the inner myometrium from the endometrium [11]. This means that chronic adhesion of the lesions to the pelvic peritoneum is less likely to occur than in extrinsic and advanced subtypes, where lesions are found on the serosal side of the uterus and are often complicated with deep infiltrating endometriosis [18, 19]. Even though we have to take into account that the pain of patients with extrinsic and advanced adenomyosis may be influenced by endometriosis with central sensitization that has progressed gradually since early reproductive years [20, 21], these findings suggest that intrinsic adenomyosis may be more responsive to progestin by LNG-IUS because there are fewer areas where progestin-mediated inhibition via the progesterone pathway is not active. Additionally, since the patients were treated with LNG-IUS placed in the uterine lumen, it is possible that intrinsic adenomyosis is easier for the progestin to physically reach because the lesions are localized closer to the endometrium. It is important for gynecologists to recognize the possibility of inadequate response to LNG-IUS treatment in advanced adenomyosis and better response in intrinsic adenomyosis. These are important findings guiding the selection of LNG-IUS for patients whose main symptoms are dysmenorrhea and/or chronic pelvic pain. Moreover, since the risk of bleeding with DNG is considered high in intrinsic adenomyosis, LNG-IUS is the optimal progestin therapy for intrinsic adenomyosis [6].

Furthermore, the progesterone resistance or the good response to progestin can be determined 1 month after starting LNG-IUS treatment, and in the absence of progesterone resistance, the effect of LNG-IUS on adenomyosis can be sustained for at least 1 year. Although there have been reports showing the efficacy of LNG-IUS for uterine adenomyosis [4], this is the first report showing the efficacy of LNG-IUS at 1 month, making it possible to determine the efficacy of treatment at an earlier point after starting LNG-IUS treatment. Furthermore, since the rate of patients who eventually underwent hysterectomy was significantly higher in progesterone-resistant group in this study, it may be necessary for patients with progesterone-resistant adenomyosis who wish to preserve their uterus to select a different treatment option, such as the use of GnRH-a or adenomyomectomy, at an early stage [22,23,24,25].

However, even though gynecological surgeries aimed at improving women’s quality of life have been developed for a variety of conditions [26,27,28,29], there are cases in which performing adenomyomectomy presents significant challenges. Since progestin is the first option to treat adenomyosis, progesterone resistance is clinically important. Progesterone has been reported to suppress the progression of adenomyosis. There have been reports on the efficacy of progestin therapy for adenomyosis and a report on the shrinkage of adenomyosis during pregnancy, during which progesterone is abundant [3,4,5,6, 30]. There are also some reports about the molecular mechanisms of progesterone resistance in adenomyosis, such as the decreased expression of progesterone receptor induced by KRAS mutations in the adenomyotic lesions and the activation of IL-6/STAT-3 signaling pathway, but further research is necessary for the elucidation and the development of the better treatment for the progesterone-resistant adenomyosis [7, 8, 31, 32]. The strength of this study is the identification of clinical factors of progesterone resistance: intrinsic adenomyosis is a favorable prognostic factor, while subtypes other than intrinsic adenomyosis are predictive factors for progesterone resistance. To find effective treatment for progesterone-resistant adenomyosis, it will also be necessary to elucidate the pathophysiology of progesterone resistance, which is still not fully understood.

There are also some limitations in this study. First, this is a retrospective study with a small number of patients. A certain number of patients with adenomyosis suffer from poor responsiveness to treatment with progestins in clinical practice. However, to analyze this poor responsiveness objectively and quantitatively, it is necessary to evaluate scores such as VAS continuously and regularly, and thus, the number of analyzed cases was limited. Regular, continuous, and quantitative observation of pain was used to improve the quality of this study, and MRI, rather than ultrasonography alone, was used to more accurately assess the type of adenomyosis, uterine size, and location of lesions [33]. The results of this study must be interpreted with caution due to limitations such as the small number of patients and the retrospective nature of the study, including the lack of washout periods for prior hormonal treatments. Further accumulation of the cases and consideration of other treatment strategies for progesterone resistant-adenomyosis, such as earlier administration of GnRH-a, are desirable for the establishment of better evidence and treatment strategies. Second, the observation period was limited to 12 months. Since our hospital refers patients to a nearby clinic when their symptoms are well-controlled with treatment, there were not enough patients treated with progestins for longer than 1 year. It is also important to accumulate cases with long-term follow-up to examine the possibility of progesterone resistance emerging during long term treatments. Third, other quantitative scores, such as the menorrhagia multi-attribute scale (MMAS) as a quality-of-life indicator during menstruation and the SF-36 as a quality-of-life indicator in general, were lacking [34, 35]. A subset of patients in this study population was assessed using the MMAS, and a difference in the rate of change in MMAS scores was observed between the progesterone-resistant and progesterone-sensitive groups (before LNG-IUS treatment: 26.1 [8.3–52.0] vs. 28.2 [26.1–28.9]; after LNG-IUS treatment: 35.4 [21.7–72.5] vs. 92.6 [62.0–100.0]). However, data on MMAS were missing for half of the patients, making it difficult to statistically assess these findings due to the limited sample size. To objectively evaluate patients’ symptoms, this study assessed not only the VAS score for dysmenorrhea but also the VAS score for chronic pelvic pain, aiming to capture a wide range of symptoms associated with uterine adenomyosis. The VAS score has been reported to correlate with the SF-36 score in studies on endometriosis, suggesting that it can serve as a proxy for evaluating quality of life in patients with adenomyosis [36]. In the future, it may be possible to define a more effective measure of progesterone resistance by collecting and evaluating quality-of-life indices together with the VAS scores.

In conclusion, the subtype of adenomyosis is the only key prognostic factor of progesterone resistance: intrinsic adenomyosis is a favorable prognostic factor for progestin treatment with LNG-IUS, while advanced and extrinsic adenomyosis are predictive factors for progesterone resistance in terms of dysmenorrhea. The responsiveness to progestin can be determined 1 month after starting LNG-IUS treatment.