Effect of pain management, supervised exercise, and pelvic floor muscle training on sexual function in women with endometriosis: A randomized controlled trial

All authors were responsible for the conception and design of this study. Rakel Gabrielsen conducted data collection and extraction and performed the analyses. Rakel Gabrielsen, Merete Kolberg Tennfjord, and Tina Tellum drafted the manuscript. Rakel Gabrielsen, Merete Kolberg Tennfjord, Tina Tellum, Marie Ellstrøm Eng, and Kari Bø reviewed and revised it for important intellectual content. All authors approved the final draft.

The study was approved by the Regional Medical Ethics Committee (REK South East 2019/1236, date of approval: October 22, 2019), Data Protect Officer (2019_135), and registered at ClinicalTrials.gov ( NCT05091268 ). All participants gave written informed consent before enrolling the study.

Norwegian Fund for Post‐Graduate Training in Physiotherapy under Grant number 139512.

Participants were recruited and randomized as previously described. 11 In brief, 81 women met the inclusion criteria and were randomized to either the exercise group ( n  = 41) or the control group ( n  = 40). At the end of the 4‐month intervention, four participants (10%) in each group were lost to follow‐up. At 12‐month follow‐up, questionnaire data were available for 26 participants in the exercise group (63%) and 29 in the control group (73%). A CONSORT flow diagram and detailed information on adherence to the intervention protocol have been published previously. 11 Background characteristics of the patient population are shown in Table  1 . No important differences between the groups at baseline were found. Background characteristics at baseline. Abbreviations: N , number; SD, standard deviations. Employment categories were not mutually exclusive, and participants could report more than one employment status. Other includes homemakers, job seekers, and other employment categories. At baseline, there were no significant differences between the exercise and control groups in any of the FSFI domain scores or in the total FSFI score (all p  > 0.05). All confidence intervals included zero, indicating that the groups were well balanced at the start of the study. No statistically significant between‐group difference was observed in disease extent distribution (Fisher–Freeman–Halton exact test, p  = 0.147). At the end of the 4‐month intervention, the exercise group demonstrated a significantly higher total FSFI score compared with the control group (MD = 3.52, 95% CI 0.49–6.54, p  = 0.023) (Table  2 ), indicating a short‐term improvement in overall sexual function. Significant between‐group differences were also observed for the subdomains lubrication (MD = 0.90, 95% CI 0.17–1.63, p  = 0.016) and orgasm (MD = 0.81, 95% CI 0.07–1.55, p  = 0.033). No significant between‐group differences were found for desire, arousal, or pain at the end of the 4‐month intervention (Table  S1 ). At 12 months follow‐up, all between‐group differences had dissipated, with no significant effects on total FSFI or any subdomains. Baseline values and adjusted between‐group differences in Female Sexual Function Index (FSFI) total score. Note : Mean differences (MD) were estimated using ANCOVA adjusted for baseline values. Positive value indicates higher scores in the exercise group. Abbreviations: CI, 95% confidence interval; MD, mean difference. At the end of the 4‐month intervention, the mean FSFI total score increased from 20.4 to 21.9 in the exercise group, whereas it decreased from 18.2 to 16.2 in the control group. Similarly, lubrication scores increased from 3.72 to 3.99 and orgasm scores from 3.53 to 3.74 in the exercise group, while corresponding scores decreased from 3.09 to 2.61 and from 3.19 to 2.60, respectively, in the control group. At baseline, 83% (34/41) of participants in the exercise group and 65% (26/40) in the control group reported being sexually active. At the end of the 4‐month intervention, the corresponding proportions were 92% (34/37) and 61% (22/36), respectively. At the end of the 4‐month intervention, a higher proportion of women in the exercise group were sexually active compared with the control group (34/37, 92% vs. 22/36, 61%). Transition analyses showed that three women in the exercise group and four women in the control group became sexually active during follow‐up. No women in the exercise group became sexually inactive, whereas four women in the control group changed from sexually active to sexually inactive (Table  3 ). Changes in sexual activity status from baseline to 4 months in the exercise and control groups. A total of 56 participants were sexually active at the end of the 4‐month intervention and were included in a subgroup analysis. In this subgroup, no significant between‐group difference was observed in total FSFI score (mean difference 0.05, 95% CI −2.21 to 2.31, p  = 0.965), and exploratory analyses of FSFI domains did not demonstrate consistent between‐group differences (Table  4 ). Subgroup analysis of participants sexually active at 4 months ( n  = 56): Adjusted between‐group mean differences (95% confidence intervals) and p ‐values for the Female Sexual Function Index (FSFI) total score and subdomains. Note : Values are adjusted mean differences (ANCOVA) with 95% confidence intervals, adjusted for baseline values. Subgroup analyses of FSFI domains were exploratory.

The present study examined sexual function in women with endometriosis‐related pelvic and genital pain participating in a multimodal intervention consisting of supervised exercise, PFMT, and pain management. At the end of the 4‐month intervention, women allocated to the intervention group reported higher FSFI scores than that receiving pain management alone. Differences were also observed in FSFI subdomains lubrication and orgasm, whereas no between‐group differences were found for the rest of the subdomains. However, a subgroup analysis among sexually active women found no between‐group differences in FSFI scores. Furthermore, a transition analysis showed that no women in the exercise group became sexually inactive during follow‐up, whereas four women in the control group changed from sexually active to sexually inactive. Together, these findings suggest that the observed between‐group differences in FSFI scores may have been influenced by differences in sexual activity status rather than by differences in sexual function among sexually active women. 16 This finding represents an important challenge to the interpretation of sexual function outcomes measured by the FSFI. 16 Because the FSFI is partly dependent on recent sexual activity, differences in the proportion of sexually active participants between groups may influence total scores and complicate comparisons of sexual function across treatment groups. 16 Previous research has shown that physical activity may improve sexual function by enhancing body image and psychological well‐being. 8 In addition, previous research suggests that PFMT can improve sexual function. 9 However, earlier work has not focused specifically on endometriosis‐associated pelvic and genital pain. PFMT is thought to act through several mechanisms, including PFM strength and neuromuscular control, particularly improved ability to fully relax the muscles after contraction. 17 , 18 Supervised PFMT may also increase PFM volume and functional capacity, 19 which could in turn facilitate orgasmic response by strengthening muscles attached to the corpus cavernosum of the clitoris, thereby enhancing arousal and orgasmic PFM contraction. 20 Additionally, greater pelvic blood flow and improved clitoral sensitivity, as well as heightened body awareness and muscle control, may contribute to enhanced lubrication and sexual satisfaction. 21 Although these mechanisms are plausible, empirical evidence remains limited, and the extent to which improved PFM function directly translates into better sexual function still remains unclear. A statistically significant difference in total FSFI score was observed at the end of the 4‐month intervention; however, the magnitude of change was modest. The mean FSFI score in the exercise group increased from 20.4 at baseline to 21.9 post‐intervention. Nevertheless, mean FSFI scores in both groups remained below the established cut‐off for normal sexual function at all‐time points, indicating that many participants continued to experience sexual dysfunction despite the observed improvement. The effect from general exercise and PFMT on sexual function was not maintained at 12 months, indicating that the initial benefits of the exercise intervention diminished over time. This loss of effect may reflect insufficient PFMT following the initial 4 months of supervised exercise, shifting motivation, or the naturally fluctuating course of endometriosis symptoms. The attenuation of effects underscores the importance of long‐term follow‐up and strategies that promote continued behavioral engagement, adherence, and self‐management to sustain exercise and PFMT routines to preserve the short‐term improvements. Overall, the findings support an integrated rehabilitation approach combining pain education, tailored exercise, and structured long‐term follow‐up to achieve durable improvements in sexual function and quality of life. A limitation of this study was the relatively low adherence to the PFMT protocol. The reported mean of 1.8 sessions per week refers to supervised sessions and home‐based exercise sessions, as shown in the primary data analysis of the RCT. 11 A previous systematic review reported limited knowledge about the effect of PFMT in women with endometriosis 17 ; therefore, the optimal dosage, frequency, and intensity for PFMT in this population remain uncertain. In the present study, the findings suggest that the combined intervention of supervised exercise and PFMT, when added to pain management, may lead to short‐term improvements in sexual function. Whether this effect is attributable to PFMT, general exercise, or their combination cannot be determined from the present study. Another limitation is that sexual function was assessed as a secondary outcome, and no sample size calculation was performed specifically for the FSFI. The trial was powered to detect differences in pain outcomes rather than sexual function. 11 Consequently, the study may have been underpowered to detect clinically relevant differences in sexual function, and the results should be considered exploratory. Furthermore, multiple comparisons were performed across the FSFI domains, increasing the risk of findings by chance. Therefore, the domain‐specific results should be interpreted as exploratory. The FSFI, although not specifically validated for endometriosis, is a widely used and psychometrically robust instrument for assessing female sexual function across diverse populations. 22 Its established domains (desire, arousal, lubrication, orgasm, satisfaction, and pain) capture the key aspects of sexual function relevant to women with endometriosis‐related pain. Sexual pain was assessed using the FSFI pain domain, which refers to pain related to vaginal penetration in general and does not differentiate between superficial and deep dyspareunia, which could reflect different pain mechanisms. Also, as FSFI scoring is contingent on recent sexual activity involving vaginal penetration, participants who were not sexually active received a score of zero, which may affect interpretation of total scores rather than the validity of the instrument itself. 22 Also, in this study, sexual activity was not explicitly defined as vaginal penetration, which may have introduced some variability in participant interpretation. The interpretation of the FSFI findings is complicated by the substantially higher proportion of sexually active women in the exercise group at the end of the 4‐month intervention compared with the control group (92% vs. 61%). Transition analyses showed that a similar number of women in each group became sexually active during follow‐up (3 vs. 4 women), whereas no women in the exercise group became sexually inactive compared with four women in the control group. Furthermore, no between‐group differences were observed when analyses were restricted to sexually active participants. Together, these findings suggest that differences in sexual activity status between the groups may have contributed to the higher FSFI scores observed in the exercise group, rather than reflecting differences in sexual function among sexually active women. Therefore, the FSFI findings should be interpreted with caution. Almost 80% of participants had minimal or mild disease, and pain in this population may have reflected mechanisms beyond nociceptive input from visible lesions alone. This is increasingly recognized in endometriosis and was part of the rationale for this exercise‐based RCT. We did not assess central sensitization or other nociplastic pain mechanisms directly, which should be acknowledged as a limitation. It is also uncertain whether the findings apply to more severe disease. A detailed anatomical classification such as #Enzian was not available, limiting exploration of differences across phenotypes. Although none had deep vaginal endometriosis, variation in other anatomical subtypes cannot be excluded, which may affect symptom presentation, treatment response, and generalizability to specific deep endometriosis phenotypes. The observed effects were modest and were not maintained at long‐term follow‐up, which might be expected given the supervised nature of the intervention. The program was also relatively resource‐intensive, involving multiple contacts with health care providers, which. raises questions about cost‐effectiveness and opportunity costs in clinical practice. A formal economic evaluation will be reported separately. The interpretation of the 12‐month follow‐up data are limited by the lack of ongoing monitoring of PFMT and high loss to follow‐up. During this period, uncontrolled factors such as dosage, frequency, and intensity of PFMT, life changes, or natural symptom fluctuation may have influenced outcomes, potentially confounding the long‐term effects of the intervention. 23 Lastly, the intervention did not include specific strategies targeting sexual function, such as the use of vibrators. Emerging evidence suggests that such approaches may positively influence sexual function, 24 and this may represent a relevant avenue for future research. Strengths of this RCT include the novel aspects that have not been investigated before, the randomized design with concealed allocation, low loss to follow‐up at 4 months post‐intervention, and the inclusion of supervised group training with experienced women's health physiotherapists. Furthermore, the exercise program was customized to each participant's exercise level with individually tailored intensity progression and a detailed exercise program to ensure reproducibility. 11 The clinical significance of this study lies in the limited availability of established treatment options for women with endometriosis who experience persistent pelvic or genital pain, dyspareunia, or sexual dysfunction. As research on conservative interventions, including exercise and pain management strategies, is recommended, 25 this study may help determine if supervised exercise and PFMT could be considered a potentially beneficial option for alleviating symptoms in women with endometriosis.

The findings suggest that the combined intervention of supervised exercise and PFMT, when added to pain management, may lead to short‐term improvements in sexual function at the end of the 4‐month intervention. Whether this effect is attributable to general exercise, PFMT, or their combination cannot be determined from the present study. However, the higher proportion of sexually active women in the exercise group and the absence of between‐group differences among sexually active participants suggest that these findings may partly reflect differences in sexual activity status rather than improved sexual function. Further studies with sexual function as a prespecified primary outcome are needed to clarify the effects of exercise‐based interventions on sexual function in women with endometriosis.

Pain during sexual intercourse is one of the most distressing and least openly discussed consequences of endometriosis. 1 , 2 This chronic, estrogen‐dependent inflammatory condition affects millions of women worldwide and extends far beyond the pelvic lesions visible on imaging or during surgery. 3 , 4 Endometriosis can profoundly alter sexual functioning, intimacy, and quality of life, often long before diagnosis and despite treatment. 1 , 4 In line with these clinical consequences, prior research has identified a significant association between endometriosis and sexual dysfunction. 5 , 6 A sexual dysfunction has been defined as «a departure from normal sensation and/or function experienced by a woman during sexual activity». 7 Whereas general exercise has been shown to enhance sexual desire and sexual function among women with antidepressant‐induced sexual dysfunction and women who have undergone hysterectomy, 8 limited knowledge exists about women with endometriosis. A systematic review and meta‐analysis concluded that pelvic floor muscle training (PFMT) improved overall sexual functioning using the Female Sexual Function Index (FSFI) total score and several subdomains in women with sexual dysfunctions, 9 but to our knowledge, this knowledge remains unexplored among women with endometriosis. Currently, there are no established treatment protocols for women with endometriosis who experience persistent pelvic and genital pain, dyspareunia, or sexual dysfunction. The National Institute for Health and Care Excellence, therefore, recommends further research into conservative interventions, including exercise and pain‐management strategies, to reduce pain and limit negative effects on relationships, sexual functioning, and overall well‐being. 10 A recent paper from our research group showed how pain management combined with supervised exercise and PFMT compared to pain management alone in women with endometriosis significantly reduced current pelvic and genital pain. 11 Given the close relationship between pelvic and genital pain and sexual function in this population, the aim of this study was to explore whether this intervention improves sexual function among these women.

TT has received speaker fees from GE Healthcare and Gedeon Richter. The other authors declare no conflicts of interest.

This is a planned secondary analysis of a two‐arm, assessor‐blinded parallel‐group randomized controlled trial (RCT) that examined the effect of adding supervised group and individual exercise training, including PFMT, to pain management compared with pain management alone, in women with endometriosis‐associated pelvic and genital pain. 11 The details of the participants, study design, recruitment, randomization, blinding, data collection, and intervention procedures are outlined in the previously published study. 11 Briefly, 81 women with histologically confirmed endometriosis presenting with moderate or severe pelvic and genital pain corresponding to ≥4/10 on a numeric rating scale (NRS), 12 and able to participate in an exercise group were eligible for inclusion. Disease extent was classified by two experienced gynecologists blinded to baseline characteristics and group allocation, reducing the risk of classification bias. Disease extent was based on available surgical information, as standardized classification systems were not consistently reported. Four categories were defined based on established clinical descriptions: minimal disease (few isolated endometriotic lesions); mild disease (superficial but more widespread pelvic endometriosis involving multiple pelvic sites); moderate disease (deep infiltrating endometriosis without major distortion of pelvic anatomy and with only mild adhesions); and severe disease (deep infiltrating endometriosis with bowel involvement and/or marked anatomical distortion, including frozen pelvis, kissing ovaries, or extensive adhesions). Differences in the distribution of disease extent between groups were assessed using Fisher–Freeman–Halton exact test. The RCT was conducted between January 2022 and January 2024 at Akershus University Hospital, Norway, and included an electronic questionnaire at baseline, at the end of the 4‐month intervention, and 12 months follow‐up. The intervention was planned and reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) statement. From March 2022 to January 2023, all participants attended a 4‐h multidisciplinary pain management course. The course was conducted by an interdisciplinary team including a physiotherapist, a gynecologist, and a psychiatrist. In addition to information about possible pharmacological and surgical interventions for disease management, the course emphasized information on general exercise, PFMT, and relaxation techniques, as well as deep abdominal breathing exercises. The participants in the control group received no further follow‐up after completing the pain management course. The intervention consisted of a weekly 60‐min group training session over 4 months. Three physiotherapists were trained in the intervention protocol; each of them was responsible for an exercise group. 11 A planned secondary analysis of sexual function, measured at baseline, at the end of the 4‐month intervention, and 12 months follow‐up using a Norwegian translated version of the Female Sexual Function Index (FSFI). 13 A Norwegian translation of the FSFI has been used in research settings and is recognized in Norwegian clinical instrument compilations; formal psychometric validation for Norwegian populations is currently underway. The FSFI is a brief, multidimensional questionnaire designed to assess sexual function in women and demonstrates strong psychometric properties. 14 The FSFI is easy to administer and has been shown to distinguish between clinical and non‐clinical populations, with a total score of 26.55 proposed as the cut‐off for identifying sexual dysfunction. 14 , 15 It consists of 19 items rated on a five‐point Likert scale, with higher scores indicating greater sexual functioning in each domain. The FSFI assesses sexual function over the past 4 weeks. In this study, both the FSFI total score and the six subdomains (desire, arousal, lubrication, orgasm, satisfaction, and pain) were evaluated at the end of the 4‐month intervention and 12 months follow‐up. To calculate the total score, the sum of the scores for each domain is multiplied by a domain‐specific factor (0.6 for desire, 0.3 for arousal, 0.3 for lubrication, 0.4 for orgasm, 0.4 for satisfaction, and 0.4 for pain) to standardize domain scores, and these values are then summed to produce the total FSFI score. Fifteen items include a zero option to indicate either “no sexual activity” (12 items) or “did not attempt intercourse” (3 items) within the past 4 weeks. A prerequisite for calculating a total sexual functioning score using the FSFI is that participants have engaged in sexual activity and attempted vaginal penetration within the past 4 weeks. 15 Data were analyzed using IBM SPSS software (SPSS Statistics for Windows, V29.0, IBM Corp, Armonk, NY, USA). The distribution of data was assessed for normality using histograms and plots. Background variables were reported as means with standard deviations (SD) or as numbers and percentages. For the statistical analysis, the FSFI was scored according to standard procedures as described above. 15 Scoring was conducted in SPSS by creating new variables for each subdomain and for the total score. Within‐group differences were calculated using paired t ‐tests and reported as the change from baseline to the end of the 4‐month intervention, and from baseline to 12 months follow‐up, with results reported as mean differences (MD) with 95% confidence intervals (CI) and p ‐values. Between‐group differences at 4 and 12 months were analyzed using analysis of covariance (ANCOVA), adjusting for baseline FSFI total score. The assumption of homogeneity of regression slopes was assessed and met. Results are reported as MD with 95% CI and p . All tests were two‐sided with α  = 0.05. Given the exploratory nature of this planned secondary analysis, no formal adjustment for multiple testing was applied. Missing data were not imputed; analyses were performed on available cases. No sample size calculation was performed for this outcome, as it was a secondary endpoint; the sample size was determined based on the primary outcome measure as reported previously. 11 A subgroup analysis was conducted including participants who reported being sexually active at the end of the 4‐month intervention, corresponding to the 4‐week recall period of the FSFI. Between‐group differences in FSFI total and domain scores were analyzed using ANCOVA, adjusting for baseline FSFI values. Analyses of FSFI domains were considered exploratory. A post hoc transition analysis was performed to examine changes in sexual activity status from baseline to 4 months. Participants were classified as sexually active or sexually inactive at both time points, and transitions between categories (remaining active, remaining inactive, becoming active, or becoming inactive) were summarized descriptively for both the exercise group and the control group.

Supplementary Table 1. Baseline values and adjusted between‐group differences in Female Sexual Function Index (FSFI) total and subdomain scores.