Analysis of Risk Factors Associated With Persistent Noncyclic Pelvic Pain Despite Hormonal Treatment in Patients With Endometriosis

We used clinical data collected prospectively from 2014 to 2020 at the Clermont‐Ferrand University Hospital Center (CHU) in France. All patients included in the present study provided written consent prior to participating in our previous study, which was approved by the Institutional Review Board (IRB) on April 3, 2013 (IRB approval number AU1026). The patients agreed to allow their samples and clinical data to be used in other research protocols aimed at improving understanding of endometriosis pathophysiology.

These analyses examined the association between several patient characteristics and the presence of persistent noncyclical pain. The data suggested that persistent noncyclical pelvic pain was present in 67 of 164 patients (41%; 95% confidence interval: 33%–49%). The results of the first analysis, which examined the separate association between each factor and pain, are summarized in Table  2 . The univariable analysis results suggested a statistically significant association with pain for rASRM stage [ 8 ], DE lesions, OV lesions, hormonal treatment, noncyclical pelvic pain, dyspareunia, and menarche age (Table  2 ). Women with an rASRM stage [ 8 ] of 3 or 4 had a higher occurrence of pain when compared with patients in stage 1 or 2. Over 40% of patients in stages 3 and 4 had pain, compared to only 24% of those in stage 1 or 2. The odds of pain were 3 times higher for patients in stage 4 compared with those in stage 1 or 2. DE lesions were associated with a much‐reduced occurrence of pain. The odds of pain were approximately 0.2 times as large (or 5 times smaller) in patients with these lesions compared with patients without DE lesions. Conversely, patients with ovarian lesions were at a higher risk of pain. The odds of pain for these patients were over 3 times higher than for patients without ovarian lesions. Pain was least common in patients receiving GnRHa treatment. Only 24% of patients on this treatment had pain, compared with at least 45% of patients on the other two treatments. The odds of pain were less than half for patients receiving GnRHa when compared with the odds for patients receiving progestin therapy. A higher score for both noncyclical pain and dyspareunia pretreatment was associated with an increased risk of pain. Patients with noncyclical pain scores of 31 or higher had odds of pain that were around 7 times higher than for patients with lower scores. Only 24% of patients with a low dyspareunia score (30 or below) had pain, compared with 66% of patients with scores over 60. The results for menarche age suggested a nonlinear relationship with pain; thus, the results are difficult to interpret from the odds ratios and are best viewed graphically. Figure  S1 shows the fitted relationship between menarche age and the probability of pain. The graph suggests that patients with an age at menarche of around 12 had the lowest risk of pain, while patients with a younger and, particularly, older age at menarche had a higher risk of pain. Univariable association with persistent pain. Abbreviations: C/S, C‐section; CO, combined oral contraceptives; DE, deep infiltrating endometriosis; GnRHa, gonadotropin releasing hormone agonists; IVF, in vitro fertilization; OV, ovary; rASRM stage, revised American Society for Reproductive Medicine classification (rASRM) stage. Odds ratios given for a 5‐unit increase in predictor variable. b To limit the number of factors in the multivariable analysis, only those showing an association with the outcome in the univariable analysis ( p < 0.02) were included. Odds ratios given for a 10‐unit increase in predictor variable. Pain score: Visual analog scale (VAS). e Although the results for menarche age did not quite reach statistical significance, menarche age was retained in the multivariable analysis. The second stage of the analysis was to examine the joint association between the factors and pain in a multivariable analysis. The results suggested some association with pain for ovarian lesions, hormonal treatment, noncyclical pelvic pain, dyspareunia, menarche age, and previous cesarean section (Table  3 ). Although the results for menarche age did not quite reach statistical significance, it was retained in the final model. After adjusting for these variables, there was no longer any significant effect of DE lesions, a factor that was significant in the univariable analyses. Ovarian lesions were again associated with an increased risk of pain. The odds of pain for patients with these lesions were 4 times higher than for patients without them. Patients receiving GnRHa hormonal treatment were again at the lowest risk of pain. The odds of pain for this group were only around a third of those for patients on progestin therapy. After adjusting for the other variables, results showed that patients receiving continuous CO were at the highest risk of pain. This group had odds of pain that were around twice those of patients receiving progestin therapy. High values for pre‐treatment noncyclical pelvic pain and dyspareunia were both associated with a higher risk of pain. Women with noncyclical pain scores above 30 had odds of pain that were at least 7 times higher than those with scores under 30. Patients with dyspareunia scores of over 60 had odds of pain that were over 6 times higher than those of women with scores of 30 or under. In the initial analysis, there was a nonlinear relationship between age at menarche and pain. After adjusting for the other factors in the model, the relationship became more linear. Patients with a later onset of menarche had a higher risk of pain. Every year's increase in menarche age was associated with a 36% increase in the odds of pain. Previous cesarean section was not a significant factor in the univariable analysis. However, after adjusting for the other factors, it became significant. Patients with a previous cesarean section had an increased risk of pain, with the odds of pain being over 9 times larger than for patients with no previous cesarean section. Multivariable analysis of factors associated with persistent pain. Abbreviations: C/S, C‐section; CO, combined oral contraceptives; GnRHa, gonadotropin releasing hormone agonists; OV, ovary. Statistically significant. Pain score: Visual analog scale (VAS).

This study showed that persistent pain was the least common among patients receiving GnRHa treatment. These results align with clinical guidelines recommending GnRH agonists for endometriosis patients experiencing persistent symptoms after initial therapy [ 2 ]. However, 24% of patients treated with GnRHa experienced persistent pain, suggesting that other factors are involved. Since our hospital is a tertiary referral hospital, the referring physician decided on the preoperative hormonal treatment. A randomized controlled trial is needed to confirm the effectiveness of hormonal treatment in alleviating persistent pain. The present multivariable analysis revealed that patients with OV lesions were four times more likely to experience pain than those without OV lesions. However, the analysis showed no significant effect of DE lesions, despite them being a significant factor in the univariable analyses. A previous study showed that severe pelvic pain is significantly more prevalent in patients with both DE and OV lesions [ 9 ]. The present results suggest that hormonal treatment may be more effective at relieving pain in patients with DE alone than in those with both DE and OV lesions. Further studies are needed to confirm these findings and to investigate the underlying mechanisms of persistent pain in patients with OV lesions. One potential explanation might be that the hormonal environment in OV differs from that in DE, as the ovarian tissue concentrations of estradiol and progesterone are more than 100‐fold higher than in peripheral serum [ 10 ]. Among the different types of endometrioses, DE, OV, and peritoneal endometriosis, the intratissue estradiol concentration is lowest in DE [ 11 ]. Hormonal treatments may sufficiently block estrogen action in DE to block pain signaling [ 12 , 13 , 14 ] but not in OV. Further studies are needed to validate our speculation and investigate the effects of hormonal treatment on local E2 concentrations in DE and OV. Furthermore, the present study identified another risk factor: severe dyspareunia. Vercellini et al. [ 15 ] showed that surgery and progestin therapy were equally effective in the treatment of deep dyspareunia in women with rectovaginal endometriosis. However, in patients without rectovaginal endometriosis, medical therapy performed significantly better than surgery [ 15 ]. In the present study population, a total of 136 patients (82.9%) had rectovaginal endometriosis. These findings suggested that in the present study population with severe dyspareunia, nociceptive pain arising from rectovaginal nodule could also be involved in the underlying mechanisms of persistent pain despite hormonal treatment. However, it is also necessary to investigate whether the other mechanisms of pain, neuropathic and/or nociplastic pain, may also be involved in the underlying mechanisms to provide personalized pain management [ 16 , 17 ]. The present study identified chronic noncyclic pelvic pain before hormone treatment as a risk factor for persistent noncyclical pelvic pain despite hormone treatment. Hormone therapy, rather than surgery, has been recommended in Cochrane systematic reviews [ 18 , 19 ]. However, patients with persistent noncyclic pelvic pain despite hormonal treatment may have predominant neuropathic, nociplastic, or mixed pain, which may require multidisciplinary or interdisciplinary treatment [ 16 , 17 ]. A recent meta‐analysis showed that multimodal physical therapies may be effective in reducing pain intensity in patients without a clear underlying condition or known disease (cancer, infection, endometriosis, etc.) [ 20 ]. To date, two randomized controlled trials (RCTs) have shown that multimodal physiotherapy may be effective in reducing the intensity of endometriosis‐associated pain [ 21 , 22 ]. Large‐scale RCTs with a well‐described methodology are needed to determine the best methods for improving pain by evaluating the effectiveness of nonpharmacological conservative therapies using different forms of physiotherapy. The present study showed that patients with a previous cesarean section had an increased risk of persistent noncyclical pelvic pain. In the present study, no patients had cesarean‐section scar endometriosis. Since chronic postoperative pain is one of the postoperative complications of cesarean section, it must be interpreted with caution [ 23 , 24 ]. Of the nine patients who had previously undergone a cesarean section, five experienced chronic, noncyclic pelvic pain prior to receiving hormone treatment. This pain could have been related to endometriosis, chronic postoperative pain, or both. Chronic postoperative pain can often have features of neuropathic pain [ 24 ]. The present findings underscore the necessity of more thoroughly evaluating pain in endometriosis patients who have undergone a previous cesarean section and experience chronic, noncyclic pelvic pain. Before initiating hormonal and/or surgical treatment, it is essential to use clinical tools such as Quantitative Sensory Testing to distinguish between nociceptive, neuropathic, and nociplastic pain [ 25 ]. Pain can be classified as nociceptive, neuropathic, or nociplastic [ 26 ]. For patients with endometriosis‐associated pain, the main initial symptom is nociceptive pain [ 27 , 28 ]. Over time, recurrent pain can lead to nociplastic pain and neurogenic inflammation [ 27 , 28 ]. There is considerable overlap in pain mechanisms among patients [ 29 ]. Current treatment options for managing endometriosis‐associated pain include hormone therapy, surgery, and complementary therapies [ 27 , 28 ]. However, nociplastic pain is less responsive to therapies that target the periphery, such as anti‐inflammatory drugs, opioids, and surgery [ 29 ]. Thus, pain phenotyping may help determine the need for hormone therapy, surgery, and/or complementary therapies. Complex cases involving predominantly neuropathic, nociplastic, or mixed pain types may require multidisciplinary or interdisciplinary treatment [ 16 , 17 ]. The present study identified several predictors in patients with histologically proven endometriosis that are useful for better personalized medicine. However, the present study has several important limitations. First, the present study selected only a limited number of risk factors, and therefore we cannot rule out potential confounding by other variables. We did not perform a sample size calculation to minimize type I and/or type II errors, because little is known about predictors of persistent pain despite hormone treatment [ 3 ]. However, the present findings may be useful for planning future studies. Further studies with sample size calculations are needed to investigate whether other factors might be involved in an increased risk of persistent pain. Second, all patients underwent complete laparoscopic removal of endometriotic lesions. However, we did not repeat face‐to‐face interviews to assess pain after surgery, so it is not clear whether surgical treatment could be effective in patients with persistent pain despite hormone treatment. Third, patients with severe dyspareunia and/or noncyclic pelvic pain may have concomitant adenomyosis. However, we did not systematically investigate whether patients had adenomyosis preoperatively. Fourth, there may be a recall bias for pain scores prior to hormone treatment. Fifth, we did not use clinical tools to differentiate between predominant nociceptive, neuropathic, and nociplastic pain. Sixth, as our hospital is a tertiary referral hospital, the present results may not be generalized.

All procedures followed were in accordance with the ethical standards of the committee responsible for human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and its later amendments. Informed consent was obtained from all patients for being included in the study.

The prevalence of pelvic endometriosis is estimated to be 10% in the general female population; in women with pain, infertility, or both, which adversely affect quality of life, the frequency is 35%–50% [ 1 ]. Endometriosis is an estrogen‐dependent disease; thus, hormonal therapies are most commonly used to treat endometriosis‐associated pain [ 2 ]. Progestins and combined oral contraceptives (CO) are considered first‐line therapy [ 2 ]. Gonadotropin‐releasing hormone agonists (GnRHa), because of their higher cost, are considered second‐line treatment [ 2 ]. These treatments are effective at managing pain for most patients. However, some patients still experience persistent pain, which is defined as pain that lasts more than 3 months [ 3 ]. There is insufficient knowledge regarding the clinical characteristics of patients who experience persistent pain during treatment [ 3 ]. However, identifying risk factors that predict inadequate pain management is essential to improving personalized medicine. Elucidating the underlying mechanisms of persistent pain despite hormonal treatment also requires this type of clinical information. In the present study, we attempted to identify risk factors for persistent noncyclic pelvic pain despite hormone therapy.

Sachiko Matsuzaki has nothing to declare. Jean‐Luc Pouly has received consulting fees and honoraria for lectures from Gedeon Richter. He is a stockholder of Ipsen and Sanofi. Michel Canis is the president‐elect of the American Association of Gynecologic Laparoscopists (AAGL). The AAGL covers his travel expenses. He also received travel expense coverage from Gedeon Richter for the 2023 World Congress on Endometriosis.

In this retrospective study, we analyzed clinical data collected prospectively from 2014 to 2020 at Clermont‐Ferrand University Hospital in France. All patients included in the present study provided written consent prior to participating in our previous study, which was approved by the Institutional Review Board (IRB) on April 3, 2013 (IRB approval number AU1026). The patients agreed to allow their samples and clinical data to be used in other research protocols aimed at improving understanding of endometriosis pathophysiology. Data on pain symptoms (before hormonal treatment, during, and on the day of the interview) as well as other clinical parameters were collected via face‐to‐face interviews conducted between hospital admission and surgery. Pain assessment was done using a 100‐point linear visual analog scale (VAS), with 0 representing no pain and 100 representing the worst possible pain [ 4 ]. For patients with endometriosis, a VAS score greater than 30 mm typically indicates significant pain [ 5 , 6 , 7 ]. Some clinical trials also use this threshold as an inclusion criterion to identify endometriosis patients with pain [ 5 , 6 , 7 ]. On the basis of the previous studies [ 5 , 6 , 7 ] and our clinical experience, we defined persistent pain despite receiving hormonal treatment as a VAS score greater than 30. This retrospective analysis included a total of 164 patients with histopathologically confirmed DE and/or OV. Of those patients, 92 had only DE, 28 had only OV, and 44 had both. All of these patients had received at least 3 months of preoperative hormonal treatment with progestins, combined oral contraceptives (CO), or gonadotropin‐releasing hormone agonists (GnRHa), which are the currently available major hormonal treatments. Our hospital is a tertiary referral hospital, and the choice of preoperative hormonal treatment was decided by the referring physician. All patients with endometriosis who underwent preoperative treatment were referred for surgery due to pain, either with or without infertility. We included DE ≥ 1 cm and OV ≥ 3 cm. Patients who had undergone various preoperative hormonal treatments, those who had received other types of hormonal treatment, and those who did not have face‐to‐face interviews or had missing clinical data were excluded from the analysis. Patients with other known gastrointestinal, urological, or chronic pelvic inflammatory diseases that could cause painful pelvic symptoms were also excluded. Patient clinical characteristics are shown in Table  1 . Baseline characteristics: Risk factor analysis. Abbreviations: BMI, body mass index; CO, combined oral contraceptives; DE, deep infiltrating endometriosis, OV, ovary; GnRHa, gonadotropin releasing hormone agonists; IVF, in vitro fertilization; rASRM stage, revised American Society for Reproductive Medicine classification (rASRM) stage; VAS, Visual analog scale. Student's t test. Some patients had both lesions. Fisher's exact test. Pearson's Chi‐square test. Score according to the revised American Fertility Society Classification (rASRM, 1997). Visual analog scale (VAS). Pain scores of pre‐hormonal treatment recalled by patients during face‐to‐face interviews conducted between hospital admission and surgery. The STATA program version 13.1 (StataCorp, College Station, TX, USA) was used for statistical analysis. Groups were compared using Fisher's exact test or Pearson's chi‐square test for categorical variables, and the t test for parametric continuous variables. To identify risk factors associated with persistent noncyclic pain (defined as VAS score > 30) despite hormonal treatments, all analysis was performed using logistic regression, due to the binary nature of the outcome (pain or no pain). The analysis was performed in two stages. First, the separate association between factor and outcome was examined in a series of univariable analyses with a total of 18 variables: age, BMI (body mass index), revised American Society for Reproductive Medicine (rASRM) score (8) (stage, implants score, adhesion score), endometriotic lesions (DE and/or OV), previous surgeries for endometriosis, type of hormonal treatment (continuous hormonal contraceptives, progestin therapy, or GnRHa), pain score by visual analog scale (dysmenorrhea, noncyclic chronic pain, dyspareunia), age at menarche, gravidity, parity, infertility duration, previous in vitro fertilization (IVF), previous cesarean section. For continuous predictor variables, the shape of the association was examined; if required, squared (and potentially) cubic terms were added to best fit the data. Second, the joint association between factors and the presence of pain was examined in a multivariable analysis. To limit the number of factors in this analysis, only factors showing any association with the outcome in the univariable analyses ( p  < 0.2) were included in this stage of the analysis. A backwards selection procedure was performed to choose only the important factors for the final model. This involves the omission of nonsignificant variables, one at a time, until only factors showing an association with the outcome were retained. Statistical significance was accepted at the 0.05 level.

Figure S1: Fitted relationship between menarche age and the probability of persistent pain despite hormonal treatment.