Evaluation of the association between self-reported pre-operative symptoms with surgically diagnosed endometriosis using the #ENZIAN classification in a multi-centre cohort

Endometriosis is a chronic gynaecological condition characterized by the presence of endometrial glands and stroma outside the uterus, with a prevalence of up to 10% in women of reproductive age ( Bulun, 2009 ). Typical endometriosis symptoms include dysmenorrhea, dyspareunia, dysuria, and dyschezia, as well as the development of chronic pelvic pain and infertility. Over the past two decades, studies have attempted to correlate the common symptoms of endometriosis to the size and localization of lesions, with only moderate success ( Gruppo Italiano per lo Studio dell’Endometriosi, 2001 ; Chapron et al. , 2003 ; Vercellini et al. , 2007 ). Notably, the severity of symptoms does not directly correlate with the extent or number of lesions found within the pelvic or extrapelvic regions, highlighting the heterogeneous nature of the disease. The coexistence of lesions in multiple locations and of various depths of invasion have complicated the process. In addition, the use of the revised ASRM classification system for endometriosis ( ASRM, 1997 ) has been discussed as being suboptimal for the operative description of lesion localization, especially for assessing the extent and depth of deep infiltrating endometriosis (DE) lesions ( Haas et al. , 2013 ; Andres et al. , 2018 ; Montanari et al. , 2022 ). For this reason, this point-based four-group staging system used both clinically and in most published studies may not suffice for studying the correlation between severity of endometriosis and symptoms. Additional frequent limitations in the literature include retrospective study designs without the use of a standardized questionnaire for assessing the different types of symptoms and the lack of a standardized pain scale. Chapron et al. (2003) reported that the only factors related to the severity of dysmenorrhea, after multivariable analysis, were the presence of rectal or vaginal infiltration by DE in the posterior cul-de-sac and extensiveness of adnexal adhesions. They concluded that the concept of ‘very deep infiltrating endometriosis’, defined as implants invading the wall of the pelvic organs, should be tested in future classification systems specifically addressing the prediction of endometriosis-related pain. Similarly, dysuria and dyspareunia have been associated, respectively, with lesions on the bladder peritoneum and uterosacral ligaments: localizations that are not specifically depicted with the ASRM classification ( Fauconnier et al. , 2002 ; Montanari et al. , 2019 ). Another important factor is the frequent co-existence of adenomyosis with DE, a condition that was previously often overlooked and frequently under-diagnosed ( Di Donato et al. , 2014 ). Adenomyosis has been associated with pain, in some studies more than DE, but is not commonly included in the operative assessment and classification of the disease ( Perello et al. , 2017 ). Montanari et al. (2022) used the operative ENZIAN score ( Keckstein et al. , 2021 ), as shown in Supplementary Fig. S1 , to demonstrate that specific compartments affected by DE (uterosacral ligaments and parametrium), rectum and sigmoid, and bladder were significantly associated with dyspareunia, dyschezia and dysuria, respectively. However, as this was a retrospective study, the compartments were only assessed individually, and not in combination, as is frequently seen clinically. In a recent prospective study, Pashkunova et al. (2024) focused on DE of the bowel, in an attempt to correlate lesion size with pre-operative symptoms. Somewhat unexpectedly, rectal lesion size or distance from the anal ridge did not correlate with dyschezia or gastrointestinal symptoms, but the concomitant involvement of other pelvic compartments such as the rectovaginal space and/or uterosacral ligaments was associated with more symptoms, especially dyspareunia. In a retrospective study focusing on fertility in women with endometriosis, the authors found that the more pelvic compartments involved with DE, the higher the pain reported by patients ( Rocha et al. , 2023 ). On the basis of these findings, we devised a prospective multi-centre study with the aim of assessing whether self-reported pre-operative symptoms correlated to the size, localization, and depth of infiltration of surgically visualized endometriosis lesions. For the standardization of operative assessment, we used the #ENZIAN classification system that classifies peritoneal, ovarian, and deep infiltrating endometriosis according to compartments and allows for documentation of extra-pelvic endometriosis and adenomyosis ( Hudelist et al. , 2021b ; Keckstein et al. , 2021 ).

In total, 838 patients agreed to participate in the study. We excluded patients in whom no endometriosis was seen intraoperatively (n = 172) as well as those who did not fully complete the preoperative questionnaires (n = 145), leaving 521 patients eligible for final analysis, all of whom were surgically confirmed to have endometriosis. Median age was 31.0 (27.0–36.0) years. Nearly all patients (n = 513) (98.5%) suffered from dysmenorrhea whereas 294 (56.4%), 208 (39.9%), and 102 (19.6%) patients reported dyspareunia, dyschezia, and dysuria, respectively. Median duration from onset of symptoms to the day of surgery was 12 (6.0–17.0) years. Analgesics were used in 54.3% of patients with a mean intake duration of 5.3 (SD 4.9) days per month at the time of surgery. Of the patients, 21.9% were using hormone therapy at the time of surgery. Patients were scheduled for surgery when endometriosis was suspected either on the basis of symptoms (70.5%) or due to already visualized endometriosis clinically, on ultrasound or MRI (35.4%). Another reason was endometriosis-associated infertility (26.9%). Table 2 shows distribution of the number of #ENZIAN compartments assessed with surgery according to location and lesion size. The most frequent form of disease was peritoneal endometriosis (#ENZIAN compartment P) (86.4%), followed by adenomyosis (compartment FA) (43.0%) and DE in compartment B (left or right) (34.9%). Distribution of lesions according to #ENZIAN compartments during surgery in absolute numbers (percent). P, peritoneum; O, ovary; T, tube; A, rectovaginal space, vagina, rectocervical area; B, sacrouterine ligaments, cardinal ligaments, pelvic sidewall; C, rectum; F, other localization. We correlated the compartment affected with the prevalence of the specific symptoms reported by affected subjects, summarizing these analyses in Table 3 . We used coloured shading to depict lower prevalence of symptoms (green) to highest prevalence (red), using yellow and orange for intermediate prevalences. Distribution of symptoms according to #ENZIAN compartment, reported as percentage of subjects reporting each. Colour corresponds to prevalence, from red (high prevalence) through orange and yellow to dark green (low prevalence). DE, deep infiltrating endometriosis; P, peritoneum; O, ovary; T, tube; A, rectovaginal space, vagina, recto-cervical area; B, uterosacral ligaments, cardinal ligaments, pelvic sidewall; C, rectum; F, other localization. Compartment P lesions were associated with a high prevalence of all symptoms. However, only 31–42% of those with solely P lesions reported these symptoms, indicating that it is the other compartments concomitantly involved that exacerbate the symptoms. Peritoneal lesions, DE in compartment B as well as adenomyosis were most frequently associated with symptoms, while ovarian (O) lesions and tubal (T) adhesions and DE in compartments A and C were less so. Adenomyosis was likewise associated with a very high prevalence of all four symptoms, between 86% and 90%, especially when in combination with DE and other localizations, as shown in Table 3 . Similarly, patients with DE lesions in any of the compartments A, B and/or C reported all four types of symptoms with a high prevalence of 45–55%. We observed that the larger the endometriosis lesion within a compartment, the higher the prevalence of patients reporting symptoms. Subjects with compartment P lesions had a high prevalence of all four symptoms, further increasing in prevalence according to lesion size. Specifically, patients with P1 lesions suffered from dysmenorrhea, dysuria, dyschezia, and dyspareunia in 33.2%, 30.1%, 29.4%, and 32.2% of cases compared to 53.6%, 59.3%, 55.9%, and 52.2% of patients with P2 and P3 lesions. Similarly, larger compartment B lesions were associated with a higher prevalence of symptoms, especially when located on the left ( Table 3 ). In a next step, we evaluated each symptom individually followed by combinations of symptoms to assess the association with each #ENZIAN compartment (singly and in the case of multiple localizations). Subsequently, we took into account only severe symptoms, performing sub-analyses with VAS symptoms reported as ≥5 and ≥8, as well as when multiple symptoms were reported. We summarize these analyses as odds ratios in Supplementary Table S1 . Due to the fact that endometriosis in one compartment does not preclude its localization in another compartment, the interpretation of these results is challenging. Most notably, severe dysmenorrhea defined as VAS ≥ 8 and severe dyspareunia ≥8 was found more often in patients with ovarian endometriosis and adenomyosis only, respectively. To corroborate the correlation between symptoms and endometriosis localizations, we created three mutually exclusive groups, namely POT, FA, and any DE A/B/C to reflect peritoneal, ovarian and tubal lesions, adenomyosis and DE, respectively. We show these results in Table 4 . Dyspareunia ≥8 was 3.5-fold more often reported in patients with adenomyosis only (OR 3.56 [1.38–9.17]) versus the other groups, while dyschezia was almost twice as likely in those with DE (OR 1.86 [1.3–2.65]). Univariate logistic regression analysis of specific symptoms and #ENZIAN compartment groups. Dyschezia, dysuria, and dyspareunia were first considered as present, if VAS ≥1 was stated. This also applies for the combination with other symptoms. Data are shown as Odds Ratio and 95% interval of confidence (95% CI OR). P  < 0.05. DE, deep infiltrating endometriosis; P, peritoneum; O, ovary; T, tube; FA, adenomyosis; VAS, visual analogue scale. We designed radar charts to help visualize the association of symptoms with #ENZIAN compartments as shown in Figs 1 and 2 . All localizations of endometriosis were associated with dysmenorrhea (prevalence of over 98%), making this symptom universal and not location-specific, as seen in Fig. 1 . There was a high prevalence of dyspareunia among patients with peritoneal endometriosis (P only) and adenomyosis (FA) (61–67%), even in the absence of deep pelvic lesions. In women with DE, the prevalence of dyschezia approached 50%, nearly as high as their reporting of dyspareunia. Interestingly, patients with ovarian endometriosis in addition to peritoneal endometriosis but without DE had a high prevalence of dyspareunia (47%), which we attribute to the concomitant P endometriosis lesions, as reported above. Radar charts depicting the association between single and combined #ENZIAN compartments and individual symptoms. DE, deep infiltrating endometriosis; P, peritoneum; O, ovary; T, tube; A, rectovaginal space, vagina, recto-cervical area; B, uterosacral ligaments, cardinal ligaments, pelvic sidewall; C, rectum; F, other localization. Radar charts depicting the association between individual and combinations of symptoms and single #ENZIAN compartments. DE, deep infiltrating endometriosis; P, peritoneum; O, ovary; T, tube; A, rectovaginal space, vagina, recto-cervical area; B, uterosacral ligaments, cardinal ligaments, pelvic sidewall; C, rectum; F, other localization. Nearly all subjects were affected by peritoneal endometriosis (n = 450, 86.4%) (#ENZIAN P) in addition to lesions in other compartments. Thus, for the symptom-based radar charts ( Fig. 2 ), we excluded compartment P, with the goal of identifying the locations of lesions inducing symptoms ‘on top’ of those generated by compartment P. We found that DE in compartments A and C were less often associated with the four symptoms, with an intraoperative diagnosis in 22–29% and 13–23% patients, respectively. Conversely, all four symptoms and combinations of symptoms were associated with a diagnosis of adenomyosis in a majority of patients (50–67%). Similarly, all four symptoms were associated with DE lesions in compartment B which was diagnosed in 50–59% of subjects. Dysuria affected only ∼20% of subjects, but when present was most associated with adenomyosis.

The institutional ethics committee of the Medical University of Innsbruck approved the study (1131/2022) and the participating centres received approval from their local ethics committees. This prospective, multicentre study was conducted between September 2022 and January 2024 at 17 endometriosis centres in Austria, Germany, and Switzerland endometriosis certified by the Scientific Endometriosis Foundation (Stiftung Endometriose Forschung—SEF; https://endometriose-sef.de/aktivitaeten/klassifikationen/enzian/ ; 12 March 2024, date last accessed) and the European Endometriosis League (EEL). One additional centre is undergoing the certification process. Consecutive patients with suspected endometriosis, and scheduled for surgery within a pre-defined recruitment period of up to 6 months per centre, were invited to participate. This was a non-interventional observational study as the indication for surgery was made independent of, and prior to, study participation. Additional inclusion criteria were age >18 years and adequate knowledge of German to give consent and fill out questionnaires. The exclusion criterion was surgery due to recurrent endometriosis. After providing informed consent, subjects completed a comprehensive pre-operative questionnaire to specify the type, severity, and duration of symptoms, as detailed below. Subjects received a published questionnaire, the Modified AGEM Questionnaire, (AGEM; https://www.ag-endometriose./downloads ; 12 March 2024, date last accessed) to specify age at menarche, duration and frequency of menstruation, as well as the onset and type of symptoms (dysmenorrhoea, dyspareunia, dyschezia, dysuria) including intensity on the visual analogue scale (VAS), and details regarding the frequency and duration of specific symptoms. Surgeons performed a comprehensive intraoperative assessment and documentation of all endometriosis lesions using the #ENZIAN classification, according to standard procedure and as required for endometriosis centre certification by the SEF/EEL. This classification system is based on compartments (peritoneum, ovary, tube, rectovaginal space, uterosacral and cardinal ligaments, rectum, bladder, and other sites), as published and shown in Supplementary Fig. S1 ( Keckstein et al. , 2021 ). Adenomyosis was diagnosed visually during surgery and/or preoperatively via ultrasound according to the MUSA criteria or MRI ( Van den Bosch et al. , 2015 ). Any combination of compartments is possible, resulting in a descriptive disease phenotype but no staging system. Participating study physicians underwent a refresher video training in the #ENZIAN classification prior to the study begin (SEF). Quantitative variables were presented as mean ± SD for normally distributed raw data and as median (interquartile range) for non-normal data distribution. Qualitative or categorical variables were expressed as number of cases (n) and relative rate (%). To compare means, the paired t -test was used. For each specific compartment and combinations of compartments, the prevalence of patient-reported symptom types (dysmenorrhea, dyspareunia, dysuria, and dyschezia) was calculated. Combinations of compartments were constructed as summarized in Table 1 . Study groups based on the localization of endometriosis according to #ENZIAN classification. DE, deep infiltrating endometriosis; P, peritoneum; O, ovary; T, tube; A, rectovaginal space, vagina, rectocervical area; B, sacrouterine ligaments, cardinal ligaments, pelvic sidewall; C, rectum. For the sample size calculation, we evaluated the association of each of three #ENZIAN compartments with the most prevalent symptom, as measured by the VAS: compartment B (DE in uterosacral ligaments, cardinal ligaments) and dyspareunia; compartment C (DE in rectum) and dyschezia, and compartment FA (other location: adenomyosis) and dysmenorrhea. The expected correlations and their coefficients were based on the results of Montanari et al. (2019) . A 0.05 one-sided Fisher’s z -test of the null hypothesis that the Pearson correlation coefficient ρ = 0.1 will have 80% power to detect a ρ of 0.3 when the sample size is 145 per ENZIAN compartment. Taking potential F1-inflation and a 15% drop-out rate into account, we would need a minimum of 145×3 = 435 participants in the study. Chi-squared test was used to calculate the prevalence of specific symptoms, singly or in combination, according to the localization of lesions within compartments. Subsequently, univariate logistic regression analysis was used to examine the association between localization of endometriosis lesions and type of pain. Data were expressed as odds ratios and 95% interval of confidence (95% CI). Radar charts were designed to illustrate these associations. Conversely, we constructed radar charts to depict which compartments were affected according to type of symptom. Since nearly all subjects were affected by peritoneal endometriosis (n = 450, 86.4%) (#ENZIAN P compartment) in addition to lesions in other compartments, we excluded P-compartment only patients for the latter analyses in the hope of distinguishing the peritoneal lesion-related pain from that due to other localizations. A significance level of α = 0.05 was set as significant for all statistical evaluations. The statistical analyses were performed using IBM SPSS Statistics for Windows, Version 29.0 (IBM Corp., released in 2019; Armonk, NY: IBM Corp).

To our knowledge, this is the first prospective study to evaluate the association between self-reported pre-operative symptoms with surgically evaluated endometriosis using the #ENZIAN classification in a large multi-centre cohort. We chose to use the #ENZIAN classification instead of the rASRM staging system to allow for more detailed and specific localization and size of lesions, especially those that are deeply infiltrating. A patient’s specific symptoms should raise suspicion for where lesions are located, helping to hone examinations and more detailed sonographic evaluations to those suspected regions. This, in turn, could help plan the surgery. We found that dysmenorrhea is a universal symptom and, even when severe, does not help to distinguish between localizations of lesions. While we expected dyschezia to be associated with a high prevalence of DE, especially compartment C (rectum) lesions, as reported by Montanari et al. (2019) , we in fact found that it was more commonly associated with adenomyosis. In fact, adenomyosis was highly prevalent in our cohort, documented in 43.0% of women, and appeared to exacerbate the symptoms reported with endometriosis at other localizations. This is an important finding since only the #ENZIAN classification and not the ASRM staging system accounts for the presence of adenomyosis during surgery. Interestingly and somewhat surprisingly, dyspareunia was not only seen in women with DE but also in those with solely peritoneal disease as well as those with solely adenomyosis. Dysuria was the least prevalent symptom overall but, when present, was most often in association with adenomyosis and DE B lesions, yet also seen in patients with peritoneal endometriosis. In contrast to the rASRM classification, ENZIAN and #ENZIAN allow precise description of lesions’ localization, especially of DE. High accordance between preoperative transvaginal ultrasound and surgical findings has been reported ( Hudelist et al. , 2021a ). When comparing preoperative symptoms with endometriosis localization, Montanari et al. found an association between rectal involvement (ENZIAN compartment C) and the presence of dyschezia. In addition, patients with adenomyosis (ENZIAN compartment FA) showed greater severity of dysmenorrhoea than those without adenomyosis ( Montanari et al. , 2019 ). We likewise found that adenomyosis was associated with all symptom types, and that adenomyosis was prevalent in women with dyschezia, only preceded in prevalence by DE B lesions, with DE C lesions being less prevalent. We found larger-size lesions to be somewhat associated with more symptoms, but not in all compartments. This was especially true for lesions in peritoneal, ovarian, and DE B compartments but not tubal, DE A, and DE C lesions as shown in our colour-coded table. We know of no previous studies that have tried to assess this, likely due to the limitations of the ASRM staging system in making this assessment. Only the study of Pashkunova et al. (2024) , which likewise used the #ENZIAN classification, evaluated whether the size of rectal lesions correlates with the severity of symptoms. They found no significant correlation between increasing lesion size in compartment DE C lesions and the presence of dyschezia. One of the biggest challenges in evaluating the association of symptoms with endometriosis localization is the concomitant implantation of lesions in several or even multiple anatomical sites. This results in a myriad of possible combinations of lesions, both superficial and deep infiltrating, and it is difficult to differentiate whether lesions in one compartment cause more or different symptoms than those in another. In the present study, we tried to form clinically relevant groups (POT only, any DE, FA only) in order to better understand the symptomatology of these patient groups. We found that women with dyschezia had a higher likelihood of having DE and those with severe dyspareunia had a higher chance of adenomyosis. Pain seems to operate additively: adenomyosis only or peritoneal endometriosis only caused symptoms in 11–43%, but in combination with other localizations the prevalence of the four symptoms increased to nearly 90%. Thus pain is not a simple linear summation of individual pain localizations, but appears rather to summate exponentially. Especially the presence of peritoneal endometriosis adds an additional level of pain. Nevertheless, we would like to highlight that, based on our results, the pattern of symptoms does not allow for the prediction of precise lesion sites, as only certain symptoms could be associated with a specific localization. The strengths of this study were its prospective and consecutive recruitment and multi-centre participation, strengthening its generalizability. Nonetheless, a larger study population would be necessary to achieve larger patient numbers in clinically relevant sub-groups. Another possible limitation lies in the #ENZIAN classification itself with its non-exclusive allocation of endometriosis sites. In contrast to the ASRM staging system, patients cannot be allocated to exclusive groups based on disease severity or stage. Our study may suffer from selection bias based on severity of symptoms, as women with fewer symptoms might be less likely to undergo surgery. The potential development of central sensitization (CS) due to longstanding symptoms was not assessed in the present study, a further limitation. Furthermore, the operative diagnosis of adenomyosis is challenging for those lacking expertise, a fact that we hope to have overcome by including data from certified endometriosis experts. Pain from endometriosis is a complex entity. Recent reviews, including studies of the group of Mechsner ( Gruber & Mechsner, 2021 ; Velho et al. , 2023 ) have drawn attention to the neo-angiogenesis, neo-innervation, and localized inflammation that play a role in the signalling and processing of endometriosis-induced pain and its perception ( Yan et al. , 2017 ). Thus, the pain and symptoms of endometriosis are multi-factorial, especially when they become chronic, with the resulting peripheral sensitization complicating the specific associations between lesions localization and specific symptom type. Another theory explaining the aetiology of endometriosis-related pain points to possible modification of central nervous system functioning due to the chronic exposure to pain symptoms, known as CS ( As-Sanie et al. , 2013 ; Mechsner, 2022 ). Using objective tools to assess for CS, several studies have shown altered pain thresholds not only in sites of endometriosis, but also in other locations not related to the disease ( As-Sanie et al. , 2013 ). This is likely the reason why we failed to find stronger or more clear associations between lesions localization and symptoms. In summary, dysmenorrhea was found in nearly all patients and therefore does not help to distinguish between sites of endometriosis lesions found surgically. Women with dyschezia have a higher chance of having DE and those with severe dyspareunia have a high chance of adenomyosis. In fact, the findings of the present study identify adenomyosis as a strong driver of pain, especially dyspareunia. Therefore, awareness of the diagnosis of adenomyosis prior to surgery is of utmost importance, although, but especially because, it can rarely be treated by surgery. In follow-up studies, we intend to evaluate the improvement in symptoms after surgery, as it relates to the type, localization, and size of lesions.