Abuse history as a modifier of symptom phenotype in endometriosis: Insights from a multidimensional retrospective cohort

SMF contributed to conceptualization, data curation, formal analysis, and drafted the manuscript. SC, MC, RS, and SA assisted with data curation, analysis, and figures and critically revised the manuscript. ML provided supervision, methodological input, and ultrasound imaging expertise, and contributed to manuscript revision and final approval.

This study was reviewed and approved by the Hamilton Integrated Research Ethics Board (HiREB) on March 10, 2025 under project number 18671.

This research did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors.

A total of 247 participants were evaluated. Among these, 235 participants underwent TVUS within the study period, completed the intake form alongside questionnaires, and were included in the analysis (Figure  1 ). Participant stratification by four groups included: Endo+/Abuse+ ( n  = 42), Endo+/Abuse− ( n  = 105), Endo−/Abuse+ ( n  = 48), and Endo−/Abuse− ( n  = 40). Among the participants, the prevalence of the subtype of disease on TVUS was DE (+/− SE) 24.2% (57/235), only SE 19.1% (45/235), OE (+/− SE) 2.6% (6/235), DE and OE 14.0% (33/235), whereas 40.0% (94/235) did not have any visible diagnosis of endometriosis. Among those with endometriosis, 62.4% (88/141) had a previous or subsequent surgery where endometriosis was surgically and pathologically confirmed. Rates of current HT use differed significantly across groups ( χ 2  = 23.7, p  < 0.001), with the highest prevalence in the Endo+/Abuse+ group (81.0%) and the lowest in the Endo−/Abuse+ group (33.3%). Group demographics, symptoms, and questionnaire scores are summarized in Table  1 . Flow chart of included participants. Endo, endometriosis; n , number; TVS, transvaginal ultrasound. Demographic characteristics and symptom frequencies in patients stratified by endometriosis and abuse history. Between‐group comparisons of symptomatology are summarized in Table  2 . CPP scores differed significantly across the four groups ( H  = 12.498, p  = 0.006, η 2  = 0.07, 95% CI: 0.02–0.12). Comparisons using the Mann–Whitney U ‐test revealed that patients in the Endo+/Abuse+ group reported significantly higher CPP scores compared to both the Endo−/Abuse+ ( p  = 0.001) and Endo−/Abuse− ( p  = 0.004) groups. No significant differences were found between the Endo+/Abuse+ and Endo+/Abuse− groups ( p  = 0.019; not significant after Bonferroni correction). All other pairwise comparisons were non‐significant ( p  > 0.05). No significant differences were found across groups for dysmenorrhea ( p  = 0.193), pain with intercourse ( p  = 0.113), or pain with bowel movements ( p  = 0.067) using Kruskal–Wallis tests. Symptom severity and questionnaire scores across study groups. Group differences were assessed for multiple binary GI symptoms using Chi‐squared or Fisher's Exact tests, as appropriate. Nausea was significantly associated with group assignment ( χ 2  = 15.438, p  < 0.001). Adjusted standardized residuals revealed that the Endo+/Abuse+ group reported significantly more nausea than expected, while other groups did not differ significantly. Bloating was also significantly associated with group assignment ( χ 2  = 11.913, p  = 0.008). The Endo+/Abuse+ group was more likely to report bloating, while the Endo+/Abuse− group was significantly less likely to report bloating than expected. Constipation ( p  = 0.442), diarrhea ( p  = 0.646), and bowel regularity issues ( p  = 0.067) were not significantly associated with the group. CSI scores differed significantly across the four groups ( H  = 9.18, p  = 0.027, η 2  = 0.04, 95% CI: 0.01–0.09). Mann–Whitney U ‐tests indicated that the Endo+/Abuse+ group had significantly higher CSI scores than the Endo+/Abuse− group ( p  = 0.005, Z  = −2.778). This difference remained significant after Bonferroni correction. No other pairwise comparisons were significant, including comparisons between Endo−/Abuse+ and any other group. GIQLI differed significantly across the four study groups ( H  = 11.3, p  = 0.010, η 2  = 0.05, 95% CI: 0.01–0.10). Comparisons using Mann–Whitney U ‐tests with Bonferroni correction (adjusted α  = 0.0083) revealed that participants in the Endo+/Abuse+ group reported significantly lower GIQLI scores compared to those in the Endo−/Abuse− group ( p  < 0.0083). Total FSFI scores did not differ significantly across the four study groups (H = 6.46, p  = 0.091, η 2  = 0.02, 95% CI: 0.00–0.03). Comparisons using the Mann–Whitney U ‐test revealed no statistically significant differences between any group pairs after Bonferroni correction (adjusted α  = 0.0083).

In this retrospective cohort study, we found that individuals with both endometriosis and a documented history of abuse (Endo+/Abuse+) reported a distinct symptom profile characterized by higher chronic pelvic pain scores compared to participants without sonographic evidence of endometriosis, alongside significantly elevated central sensitization and gastrointestinal symptom burden. Notably, chronic pelvic pain intensity did not differ significantly between individuals with endometriosis with versus without a history of abuse, whereas central sensitization scores were significantly higher in the Endo+/Abuse+ group. Specifically, this group demonstrated higher CPP scores, a greater prevalence of GI symptoms (notably nausea and bloating), elevated CSI and lowered GIQLI scores. These findings support the hypothesis that the co‐occurrence of endometriosis and a history of abuse may be associated with a distinct, high symptom burden phenotype. Our findings build upon a growing body of literature suggesting that chronic pain conditions, including endometriosis, are influenced not only by lesion burden but also by psychosocial factors such as a history of abuse. Figure  1 illustrates the proposed biopsychosocial pathways by which a history of abuse may interact with endometriosis‐related inflammation to perpetuate central and peripheral sensitization, thereby amplifying pain and GI symptoms. Prior research has linked abuse to increased pain sensitivity, dysregulation of the HPA axis, and features of central sensitization, all mechanisms that may amplify pain perception independently of anatomical disease. 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 18 , 19 , 20 , 21 , 27 While CPP scores were elevated across all groups in our tertiary cohort, endometriosis status alone did not differentiate symptom severity. However, the co‐existence of abuse history was associated with compounded increases in CPP and CSI scores. Our findings are consistent with Bourdon et al., who demonstrated that childhood and adolescent sexual abuse was associated with severe pelvic pain irrespective of endometriosis status. This aligns with our observation that abuse history appears to compound symptom severity. However, we cannot disentangle whether this effect is directly attributable to endometriosis, to pain pathways more broadly, or to their intersection. While prior studies have demonstrated associations between childhood sexual abuse and severe pelvic pain irrespective of endometriosis, our study extends this literature by using advanced TVUS to phenotype visible disease and by incorporating multidimensional outcomes beyond pain, including GI symptoms, central sensitization, and validated quality‐of‐life indices. This imaging‐based stratification enables us to distinguish between anatomical and psychosocial contributions to symptom profiles in a real‐world, front‐line cohort. Importantly, our findings suggest that the compounded symptom burden observed in the Endo+/Abuse+ group is not limited to pain intensity but extends to GI morbidity and central sensitization. Given the binary classification of abuse exposure, these findings should not be interpreted as reflecting trauma severity or duration, but rather as indicating that a history of reported abuse, regardless of subtype, is associated with altered symptom profiles in this cohort. The dissociation observed between CPP intensity and central sensitization scores among individuals with endometriosis warrants particular consideration. While CPP severity was comparable between the Endo+/Abuse+ and Endo+/Abuse− groups, significantly higher CSI scores in the Endo+/Abuse+ group suggest differences in pain processing rather than pain magnitude. This pattern may reflect greater nociplastic pain mechanisms, altered central pain modulation, or heightened somatic symptom amplification in individuals with a trauma history, even when self‐reported pain intensity is similar. In this context, central sensitization may represent a more sensitive marker of trauma‐associated pain vulnerability than pain severity alone, helping to explain why patients with comparable CPP scores may differ substantially in symptom complexity, treatment response, and functional impact. The association between abuse and GI symptoms, including nausea and bloating, also aligns with prior work 28 , 29 , 30 on stress‐mediated alterations in autonomic nervous system function and brain‐gut axis dysregulation. However, we did not observe significant group differences for dysmenorrhea, dyspareunia, bowel pain, or other bowel symptoms, suggesting that abuse history may contribute more prominently to centrally mediated symptoms than to those typically attributed to local inflammatory pathology. 31 , 32 However, psychiatric and GI comorbidities were not adjusted for, which may confound associations between abuse history and symptom outcomes. These findings underscore the specific importance of abuse‐informed care in the context of endometriosis, a condition already marked by delayed diagnosis, fragmented care, and high symptom burden. While trauma‐informed practices are increasingly integrated into general gynecology, our data suggest that individuals with both endometriosis and a history of abuse report a compounded symptom phenotype, including greater pain intensity, GI symptoms, and central sensitization that may not be fully addressed by lesion‐focused management alone. Routine screening for abuse history in patients with endometriosis can facilitate earlier identification of those at risk for complex pain presentations and guide referrals for interdisciplinary support. For patients, this may translate to improved symptom management, more targeted therapies, and reduced frustration in navigating their care. For healthcare providers, it offers a framework to better stratify treatment pathways and reduce unnecessary interventions. At the systems level, addressing the multidimensional needs of this high‐symptom‐burden subgroup may help reduce repeat consultations, procedural overuse, and healthcare costs associated with unmanaged chronic pain. This study supports the integration of psychosocial variables into endometriosis phenotyping frameworks, particularly in recognizing abuse‐associated symptom clusters. While clinical awareness of trauma is growing, prospective research is needed to define the temporal and mechanistic pathways linking abuse exposure to symptom development, including pain chronification, GI dysregulation, and central sensitization. Longitudinal cohort studies could clarify whether the presence of early abuse exposure precedes or exacerbates symptom onset, and whether biological intermediaries such as autonomic dysfunction, inflammatory signaling, or altered nociceptive processing are involved. Mechanistic investigations may also help determine why conventional lesion‐targeted treatments are less effective in this subgroup. Additionally, future interventional studies should explore whether multimodal approaches, including cognitive behavioral therapy, 33 pelvic floor physical therapy, and integrative pain programs, could attenuate symptoms in patients with trauma‐associated endometriosis phenotypes. While these interventions are not trauma‐specific protocols, they target pain, central sensitization, and functional impairment, which may be particularly relevant in this subgroup. In parallel, trauma‐focused therapies (e.g., trauma‐informed CBT or eye movement desensitization and reprocessing [EMDR]) also warrant investigation to address psychological sequelae more directly. Identifying responders to these interventions could inform more personalized, stratified care models in endometriosis research and practice. Ultimately, the routine use of validated tools, such as the CSI, in both clinical trials and cohort studies may facilitate earlier identification and subgroup stratification, thereby enabling a better alignment between symptom drivers and treatment approaches. 34 , 35 In addition to the routine use of the CSI, the integration of validated trauma‐screening tools (e.g., Childhood Trauma Questionnaire, 36 Adverse Childhood Experiences survey 37 ) should be considered in both research and clinical settings. These measures would enable more nuanced stratification of patients by type, severity, and timing of abuse, facilitating clearer mechanistic insights and tailored care pathways. We also recognize the ethical importance of framing. Trauma‐informed care must complement, rather than replace, the biomedical investigation and treatment of endometriosis. Misinterpretation of trauma–symptom associations risks reinforcing psychosomatic dismissal, which has historically harmed patients with endometriosis. Our intention is to emphasize that both biomedical pathology and psychosocial context require integrated attention in care models. The strengths of this study include the use of a well‐characterized cohort, which utilized standardized intake forms and validated questionnaires, as well as advanced ultrasound diagnosis of endometriosis by an expert sonologist. Although prior studies have examined trauma and pelvic pain, our study is among the first to use imaging‐based phenotyping and multidimensional symptom measures to evaluate this intersection. However, several limitations must be acknowledged. Abuse history was analyzed as a binary construct without accounting for type, severity, duration, or timing of exposure, representing an important limitation of this study. This approach reflects the constraints of retrospective clinical documentation and ethical considerations surrounding trauma disclosure in routine care; however, it assumes equivalence across heterogeneous abuse experiences. Prior literature suggests that specific trauma characteristics, particularly childhood, sexual, or repeated abuse, may be more strongly associated with chronic pelvic pain and central sensitization than others. 9 , 10 , 11 As such, the inability to stratify abuse exposure may have obscured differential associations between trauma subtypes and symptom outcomes or contributed to exposure misclassification. Importantly, this limitation constrains mechanistic inference. Our findings should therefore be interpreted as associations with reported trauma exposure broadly, rather than with specific trauma characteristics or severity. Future prospective studies incorporating validated trauma instruments (e.g., Childhood Trauma Questionnaire, Adverse Childhood Experiences) are necessary to enable more granular stratification and to clarify dose–response relationships between trauma exposure and symptom phenotypes. Second, we were unable to control for key potential confounders, including timing of trauma relative to symptom onset, ethnicity, and surgical confirmation for all cases. This limitation increases the risk of inflated associations, particularly as trauma prevalence in the general population is high, with lifetime exposure estimated at ~25%–50% among women. 15 , 16 , 17 , 18 Third, trauma was captured as a broad construct without differentiation between transient versus enduring consequences. For instance, PTSD arises in only ~5.6% of trauma‐exposed individuals and affects ~3.9% of the global population, 38 highlighting the importance of distinguishing trauma subtypes and sequelae. Fourth, the use of current HT differed significantly between groups, raising the possibility of confounding. While we documented this imbalance, we did not adjust for HT in regression models, as intake forms only captured current use and lacked detail on timing, duration, or type of therapy relative to trauma exposure or symptom onset. As such, statistical adjustment risked introducing bias without improving interpretability. Nonetheless, given evidence that HT may modulate PTSD symptoms via fear extinction pathways, 39 the uneven distribution of HT across groups should be recognized as a potential confounder. Additionally, the CSI is a self‐report screening tool with overlap with psychological distress and does not provide objective confirmation of altered nociceptive processing; results involving CSI should be interpreted accordingly. Future prospective studies should incorporate detailed HT data into analytic models to better isolate its effects. We were not sufficiently powered to perform multivariable analyses adjusting for disease extent, lesion location, or psychiatric comorbidities. Although these factors may influence symptom presentation, particularly for centrally mediated outcomes, stratified or multivariable modeling would have risked overfitting given sample size constraints and variable distributions. Accordingly, analyses were limited to bivariate comparisons and interpreted as exploratory and hypothesis‐generating. Future prospective studies with larger, well‐characterized cohorts will be needed to enable robust multivariable adjustment. Lastly, although SE was evaluated using emerging ultrasound techniques, its limited diagnostic sensitivity may have led to misclassification, particularly among participants without sonographic evidence of disease who may still have had superficial lesions.

Individuals with both endometriosis and a history of abuse exhibited a distinct symptom profile characterized by greater GI symptom burden and higher central sensitization compared to other groups. While CPP severity was elevated among individuals with endometriosis overall, pain intensity did not differ significantly by abuse history within the endometriosis subgroup. Endometriosis alone did not significantly elevate symptom burden. These findings highlight the importance of integrating psychosocial assessment into endometriosis care and support an abuse‐informed, multidisciplinary approach for patients with complex symptom profiles.

Endometriosis is a chronic, inflammatory disease characterized by the presence of endometrial‐like tissue outside the uterus, affecting 1 in 10 people presumed female sex at birth and gender diverse individuals. 1 Clinically, it presents with a wide range of symptoms, including dysmenorrhea, dyspareunia, chronic pelvic pain (CPP), and gastrointestinal (GI) disturbances such as bloating and nausea. 2 Despite its substantive prevalence, the degree of symptom burden among people with endometriosis often varies widely and does not consistently correlate with disease stage or lesion location. 3 This discrepancy suggests that symptom severity in endometriosis is influenced not only by anatomical findings but also by additional physiological and psychosocial mechanisms. CPP is increasingly understood as a complex pain condition that may persist in the absence of active pelvic pathology, including endometriosis. 4 The pathophysiology of CPP includes not only nociceptive pain associated with lesion activity but also neuropathic and nociplastic components, particularly in individuals with persistent, widespread pain. 5 , 6 Central sensitization, characterized by an amplification of neural signaling within the central nervous system, has emerged as a critical mechanism underlying chronic pain states, including CPP. 5 , 6 , 7 , 8 Among the emerging contributors to symptom variability among individuals with CPP is the role of psychological trauma, including a history of abuse. 9 , 10 , 11 Prior literature has identified associations between abuse history and increased pain sensitivity, altered stress reactivity, and higher rates of central sensitization in chronic pain syndromes. 9 , 10 , 11 Notably, Bourdon et al. found that childhood and adolescent sexual abuse was associated with severe pelvic pain irrespective of endometriosis diagnosis, underscoring that trauma exposure may exacerbate pain symptomatology independent of lesion presence. Abuse‐related trauma has been linked to persistent activation of the hypothalamic–pituitary–adrenal (HPA) axis and dysregulation of the autonomic nervous system, which may amplify pain perception and exacerbate GI symptoms. 12 The relationship between abuse history and chronic pain is well‐documented, with early adverse experiences linked to increased vulnerability to central sensitivity syndromes such as fibromyalgia, irritable bowel syndrome (IBS), and CPP. 9 , 10 , 11 , 13 , 14 Approximately 1 in 4 to nearly half of people presumed female sex at birth have experienced some form of abuse, including physical, psychological, sexual, emotional, financial, verbal, and/or neglect at some point in their lives. 15 , 16 , 17 , 18 Despite this, relatively few studies have systematically examined the intersection between endometriosis and abuse history, 9 , 19 , 20 , 21 particularly overlapping non‐pain related symptoms, central sensitization, and quality of life. This study seeks to provide hypothesis‐generating evidence by evaluating whether a history of self‐reported abuse is associated with symptoms beyond pain and patient phenotypes among individuals with and without endometriosis seen on advanced transvaginal ultrasound (TVUS). Specifically, we aim to compare pain intensity, GI symptoms, and validated questionnaire scores across four groups: individuals with both endometriosis and a self‐reported history of abuse (Endo+/Abuse+), individuals with endometriosis but no abuse history (Endo+/Abuse−), individuals with a history of abuse but no endometriosis seen on ultrasound (Endo−/Abuse+), and those with neither condition (Endo−/Abuse−). By identifying an abuse‐associated symptom phenotype, this study aims to inform future research on symptom stratification and the underlying biopsychosocial mechanisms linking trauma history to endometriosis‐related morbidity.

The authors have no relevant financial or non‐financial conflicts of interest to disclose.

This retrospective observational cohort study was conducted at the McMaster University Medical Center Gynecology Health Clinic, a tertiary care center specializing in advanced gynecologic assessment and minimally invasive gynecological surgery. Data were extracted from standardized patient intake forms at the time of referral for gynecologic consultation and “one‐stop shop” advanced TVUS between January 2023 and June 2024. Standardized ultrasound reporting forms and validated questionnaires were utilized as part of routine care, thereby reducing missing data and ensuring the accurate interpretation of medical records. Participant inclusion criteria were: (1) age ≥18 years; (2) referral for any gynecological referral; (3) completion of baseline intake forms, including abuse history and validated symptom questionnaires; and (4) underwent a TVUS performed by a single expert minimally invasive gynecologic surgeon and sonologist (ML). Exclusion criteria were: (1) incomplete intake forms or missing abuse history; (2) pregnancy at the time of evaluation; and (3) inability to complete questionnaires in English. Endometriosis was diagnosed and characterized using TVUS protocols aligned with the International Deep Endometriosis Analysis (IDEA) consensus, enabling standardized reporting and identification of ovarian (OE) and deep endometriosis (DE). Superficial endometriosis (SE) was also evaluated using targeted scanning techniques described in recent advancements in ultrasound‐based detection of SE. 22 , 23 Advanced TVUS, when performed by experienced providers, offers high diagnostic accuracy for OE and DE, 23 and is increasingly recognized as a reliable first‐line assessment tool. 24 , 25 , 26 At our center, recent validation has demonstrated the diagnostic accuracy of TVUS for DE and OE, in line with international consensus standards, 26 with sensitivities ranging from approximately 85%–100% and specificities exceeding 97%–100%. 23 The use of TVUS in this study reflects its role as a first‐line diagnostic for OE and DE, enabling standardized phenotyping of patients based on visible disease. The inclusion of a ‘no endometriosis on ultrasound’ group (Endo−) does not imply the complete absence of pathology, as SE may remain undetected. Instead, this comparator group represents a clinically meaningful subset of patients who present with endometriosis‐like symptoms but lack sonographic evidence of OE or DE. This stratification allowed us to disentangle the contributions of psychosocial factors such as abuse history from the anatomical burden of endometriosis, thereby examining whether trauma exposure is associated with distinct symptom phenotype independent of visible disease. Although disease subtype (e.g., DE, OE, SE) was documented, the study was not designed or powered to examine symptom differences stratified by disease extent or location, nor to assess balance of extent across abuse strata. History of abuse was classified as present or absent, based on self‐reported documentation of physical, emotional, or sexual abuse in standardized intake forms or clinician notes. More detailed information regarding the type, frequency, timing, or severity of abuse was not systematically collected, both due to privacy considerations in routine clinical intake and the retrospective design of this study. While some patients provided informal details, these were inconsistently captured and therefore deemed unreliable for research purposes. For this reason, abuse was analyzed as a binary variable rather than stratified by subtype, frequency, or severity. Participants were stratified into four groups based on the presence (confirmed via TVUS) or absence of endometriosis (no visible endometriosis seen on TVUS) and the presence or absence of documented abuse history: Endometriosis with abuse history (Endo+/Abuse+), Endometriosis without abuse history (Endo+/Abuse−), No endometriosis seen on TVUS with abuse history (Endo−/Abuse+), or No endometriosis seen on TVUS and no abuse history (Endo−/Abuse−). Data were collected from a standardized clinical intake form developed in our center, which patients complete upon referral. The form includes detailed assessments of pelvic pain symptoms, including dysmenorrhea, dyspareunia, CPP, and related factors such as GI disturbances (nausea, bloating, diarrhea, constipation, and pain during defecation). CPP was assessed as non‐cyclic pelvic pain and characterized using a VAS scale 1–10. Patients completed three validated questionnaires, including the Central Sensitization Inventory (CSI), the GI Quality of Life Index (GIQLI), and the Female Sexual Function Index (FSFI), which reflected the impacts on quality of life and pathophysiological effects among participants. Descriptive statistics were used to summarize demographic and clinical characteristics across the four study groups (Endo+/Abuse+, Endo+/Abuse−, Endo−/Abuse+, Endo−/Abuse−). Continuous variables were assessed for normality using the Shapiro–Wilk test and reported as medians with interquartile ranges, as they exhibited non‐normal distributions. Categorical variables were summarized using frequencies and proportions. Between‐group differences in symptom severity, including CPP, dysmenorrhea, dyspareunia, and pain with bowel movements, were evaluated using the Kruskal–Wallis test. Where overall significance was identified, pairwise Mann–Whitney U ‐tests were performed, with Bonferroni correction applied to account for multiple comparisons (adjusted α  = 0.0083). Binary outcomes such as nausea, bloating, constipation, diarrhea, dyschezia, and dyspareunia were analyzed using the chi‐squared test of independence. When expected cell counts were less than five, Fisher's Exact test was used. For significant Chi‐Square results, adjusted standardized residuals were reviewed to identify which group(s) contributed to the observed associations, using a threshold of ±1.96 for statistical significance. Cumulative questionnaire scores (CSI, QIQLI, and FSFI) were analyzed individually, assessed using the Kruskal–Wallis test, with Mann–Whitney U ‐tests and Bonferroni correction applied as appropriate. For nonparametric tests, effect sizes were calculated ( η 2 for Kruskal–Wallis, r for Mann–Whitney U ) and 95% confidence intervals were provided where possible. Group differences in hormonal therapy (HT) use were evaluated using Chi‐squared testing, and proportions were reported, given the potential relevance of HT as a confounder. No formal correction was applied across families of outcomes; findings are interpreted as exploratory and hypothesis‐generating. All statistical analyses were performed using SPSS V29 software (SPSS Inc., Chicago, IL, USA). A total of 235 participants were included in the analysis and stratified into four study groups. Given the retrospective design, no a priori sample size calculation was conducted. Instead of post‐hoc power analyses, effect sizes with confidence intervals were reported to contextualize the magnitude of observed between‐group differences. Given the exploratory design and sample size constraints, analyses were limited to bivariate comparisons rather than multivariable modeling.