Are digestive symptoms in women presenting with pelvic endometriosis specific to lesion localizations? A preliminary prospective study

Transcription

1 Human Reproduction, Vol.27, No.12 pp , 2012 Advanced Access publication on September 7, 2012 doi: /humrep/des322 ORIGINAL ARTICLE Gynaecology Are digestive symptoms in women presenting with pelvic endometriosis specific to lesion localizations? A preliminary prospective study Horace Roman 1,2, *, Julie Ness 1, Nicolae Suciu 3, Valérie Bridoux 4, Guillaume Gourcerol 5,6, Anne Marie Leroi 5,6, Jean Jacques Tuech 4, Philippe Ducrotté 7,Céline Savoye-Collet 8, and Guillaume Savoye 6,7 1 Department of Gynecology and Obstetrics, Rouen University Hospital, Rouen, France 2 Research Group EA 4308 Spermatogenesis and Male Gamete Quality, Rouen University Hospital, Rouen, France 3 Department of Gynaecology and Obstetrics, Polizu University Hospital, Bucharest, Romania 4 Department of Digestive Surgery, Rouen University Hospital, Rouen, France 5 Physiology Unit, Rouen University Hospital, Rouen, France 6 Digestive Tract Research Group EA3234/IFRMP23, Rouen University Hospital, Rouen, France 7 Department of Gastroenterology, Rouen University Hospital, Rouen, France 8 Department of Radiology, Rouen University Hospital, Rouen, France *Correspondence address. Clinique Gynécologique et Obstétricale, CHU Charles Nicolle, 1 rue de Germont, Rouen, France. Tel: ; Fax: ; horace.roman@gmail.com Submitted on February 23, 2012; resubmitted on July 29, 2012; accepted on August 6, 2012 study question: What are the types and frequency of digestive symptoms in patients with different localizations of pelvic endometriosis and which specific symptoms are related to rectal stenosis? summary answer: There is a high prevalence of digestive complaints in women presenting with superficial pelvic endometriosis and deep endometriosis sparing the rectum. what is known already: Women presenting with pelvic endometriosis frequently report gastrointestinal complaints of increased intensity during menstruation, which are not necessarily linked to the infiltration of the disease into the rectal wall. Even though intrarectal protrusion of the nodule can have an impact on bowel movement, only a minority of women with rectal nodules seemed to be concerned by significant narrowing of the rectum. study design and size: This three-arm cohort prospective study included 116 women and was carried out over 22 consecutive months. participants, setting and methods: Prospective recording of data was performed for women treated for Stage 1 endometriosis involving the Douglas pouch (n ¼ 21), deep endometriosis without digestive infiltration (n ¼ 42) and deep endometriosis infiltrating the rectum (n ¼ 53). Patient characteristics, pelvic pain and data from preoperative standardized questionnaires The Gastrointestinal Quality of Life Index (GIQLI), the Knowles-Eccersley-Scott-Symptom Questionnaire (KESS) and the MOS 36-Item Short-Form Health Survey (SF-36) were compared according to endometriosis localization. main results: The values of total KESS and total GIQLI score were comparable for the three groups, as were a majority of the digestive complaints. Women presenting with rectal endometriosis were more likely to report an increase in intensity and length of dysmenorrhoea, while deep dyspareunia appeared to be more severe in women with superficial endometriosis. Women presenting with rectal endometriosis were more likely to present cyclic defecation pain (67.9%), cyclic constipation (54.7%) and a significantly longer stool evacuation time, although these complaints were also frequent in the other two groups (38.1 and 33.3% in women with Stage 1 endometriosis and 42.9 and 26.2% in women with deep endometriosis without digestive involvement, respectively). No independent clinical factor was found to be related to infiltration of the rectum by deep endometriosis. Among women with rectal endometriosis, only 26.4% presented with rectal stenosis. These women were significantly more likely to report constipation, defecation pain, appetite disorders, longer evacuation time and increased stool consistency without laxatives. Study presented to the 40th Annual Meeting of the American Association of Gynecological Laparoscopists, Hollywood, FL, USA, & The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please journals.permissions@oup.com

2 Digestive symptoms in endometriosis 3441 limitations: Patients treated for pelvic endometriosis in a tertiary referral centre may not be representative of the general endometriosis population presenting with those lesions. Statistically significant differences were revealed between the three groups; however, the results were based on a small number of subjects, which carries an inherent risk of type II error particularly when comparing variables with closed values. wider implications of the findings: In women presenting with pelvic endometriosis, it seems likely that various digestive symptoms are the consequence of cyclic inflammatory phenomena leading to irritation of the digestive tract, rather than to actual infiltration of the disease itself into the rectum, with the exception of a limited number of cases where the disease leads to rectal stenosis. study funding/competing interest: The North-West Inter Regional Female Cohort for Patients with Endometriosis (CIRENDO) is financed by the G4 Group (The University Hospitals of Rouen, Lille, Amiens and Caen). No financial support was specifically received for this study. The authors declare no conflict of interest. Key words: deep endometriosis / rectal endometriosis / peritoneal endometriosis / digestive symptoms / rectal stenosis Introduction Women presenting with pelvic endometriosis frequently report gastrointestinal complaints of increased intensity during menstruation (Fauconnier et al., 2002). Catamenial changes in bowel movement features are not however necessarily linked to the infiltration of the disease into the rectal wall, as indicated by the strong relationship between vaginal endometriosis and catamenial painful defecation (Fauconnier et al., 2002). Furthermore, the suppression of menses over several months leads to a significant decrease in the intensity of digestive symptoms, such as rectal tenesmus and dyskesia (Fedele et al., 2001), suggesting that digestive complaints are predominantly related to cyclic micro-haemorrhages occurring in endometriosis implants. Although Fauconnier et al. (2002) proposed that the association of painful defecation with specific anatomic localizations may be due to reasons of a mechanical nature, the question remains as to why women with deep endometriosis infiltrating the rectum present with digestive complaints predominantly or exclusively during menses. Catamenial complaints could reasonably be related to cyclic microhaemorrhages and inflammation occurring around or inside the rectum, factors which might be principally responsible for digestive complaints, over and above that of rectal infiltration by the nodule itself (Roman et al., 2011). Endometriosis deposits are generally found on the peritoneum within the posterior pelvic compartment in close proximity to the terminal large bowel, such that local prostaglandin release and inflammation of lesions may explain any altered bowel function (Seaman et al., 2008). To investigate this hypothesis, we carried out a prospective analysis of digestive complaints reported by women referred with three distinct localizations of pelvic endometriosis. This study aimed firstly to assess the relationship between digestive symptoms and endometriosis localizations and secondly to identify those clinical symptoms presented by patients with deep endometriosis infiltrating the rectum that are related to rectal stenosis. Materials and Methods We performed a three arm-comparative study using prospective recording in women undergoing surgical management of pelvic endometriosis from January 2010 to October 2011 at Rouen University Hospital, France. The first arm included women with superficial implants only, intraoperatively identified and histologically confirmed as located on the Douglas pouch peritoneum. The American Fertility Society revised score (AFSr) did not exceed 5, and intra-operative exploration confirmed the absence of deep localizations. The second arm included women presenting with deep endometriotic nodules located on the posterior vagina, rectovaginal space and on uterosacral ligaments, while sparing the digestive tract. The third arm included women with deep endometriosis infiltrating at least the muscular layer of the rectum up to 15 cm from the anus. Deep localizations were confirmed by imaging and intra-operative examinations. Pre-operative assessment of all women referred with endometriosis was based on clinical examination and pelvic magnetic resonance imaging. Women presenting with deep endometriosis also benefited from an endorectal ultrasound to assess the presence and depth of rectal involvement. Those patients with rectal involvement systematically underwent virtual colonoscopy-based computed tomography (CTC) to measure rectal nodule height and any subsequent rectal stenosis (Vassilieff et al., 2011). In order to determine the presence of stenosis, a CTC was performed and measurements were taken of rectal diameter at different heights. The radiologist firstly identified an indentation or protrusion of the nodule through the rectal wall, along with an absence of rectal distension when anal insufflation was performed. This was then correlated with a decrease of antero-posterior rectal diameter measured from a sagittal view at nodule level, compared with the diameter measured immediately above and below the nodule (Figs. 1 3). The goals and strategy of surgical management were established on the basis of imaging data and in conjunction with informed choices of the patient. Prospective recording of data on antecedents, clinical symptoms, findings of clinical and radiological examinations, surgical procedures and postoperative outcomes was performed through the CIRENDO database (the North-West Inter Regional Female Cohort for Patients with Endometriosis ), a prospective cohort financed by the G4 Group (The University Hospitals of Rouen, Lille, Amiens and Caen) and coordinated by the first author of the study (H.R.). Information was obtained from surgical and histological records and self-questionnaires completed before surgery. Patients also completed the MOS 36-Item Short-Form Health Survey, used in the evaluation of quality of life and health status (Ware et al., 1998). Data recording, contact and follow-up were carried out by a clinical research technician. Women were included in the CIRENDO database only when endometriosis was confirmed by both surgical exploration and biopsy performed in a member facility. Post-operative follow-up was based on data from the aforementioned questionnaires completed at 1 and 3 years. Pre- and post-operative assessments of digestive symptoms were carried out using gastrointestinal standardized questionnaires. For the

3 3442 Roman et al. Figure 1 Computed tomography-based virtual colonoscopy (CTC) revealing rectal stenosis due to deep endometriosis infiltrating the anterior rectal wall: 2D sagittal view of the insufflated rectum showing an important reduction in rectal diameter. diagnosis of constipation, the Knowles-Eccersley-Scott-Symptom Questionnaire (KESS) (Knowles et al., 2000) was used; it is compiled of 11 individual items with a maximum possible 39 points. The KESS questionnaire was designed to be sufficiently simple so as to be completed in under 5 min. Each question had four or five possible answers, which were scored on an unweighted linear integer scale to produce a range between zero and three, or between zero and four points. Lower scores represented symptom-free states and higher scores represented increased symptom severity. The KESS total score differentiates patients with constipation, for whom overall values are greater than 10, from healthy controls, for whom the median value averages 2 (range 0 6) (Knowles et al., 2000). The Gastrointestinal Quality of Life Index (GIQLI) (Slim, 1999) is a self-administered questionnaire including 36 questions concerning digestive symptoms, physical status, emotions, social dysfunction and effects of medical treatment. Consequently, it comprises not only questions on gastro-intestinal symptoms but also those on other aspects of quality of life and has been validated for various gastrointestinal diseases. The 36 items of the GIQLI were scored from 0 to 4 with the total score ranging from 0 (worst) to 144 (best quality of life). The GIQLI also measures physical well-being, mental well-being, digestion and defaecation (Nieveen van Dijkum et al., 2000). Total score median values vary around 126 for healthy controls, with scores lower than 100 are recorded for patients suffering from gastro-intestinal diseases. Prospective record and management of data were approved by the French authority of CCTIRS and CNIL (Comité Consultatif pour le Traitement de l Information en matière de Recherche dans le domaine de la Santé, and Conseil National d Informatique et Libertés). Digestive symptoms, pelvic painful symptoms and questionnaires results were compared in relation to the three distinct localizations of Figure 2 CTC revealing rectal stenosis due to deep endometriosis infiltrating the anterior rectal wall: 3D virtual endoluminal view showing narrowing of the lumen corresponding to rectal stenosis. Figure 3 CTC revealing rectal stenosis due to deep endometriosis infiltrating the anterior rectal wall: 3D transparent rendering mode double contrast enema-like view displaying colorectal anatomy and exact location of the stenosis. endometriotic lesions. Pre-operative CTC data then enabled comparison of groups with and without rectal stenosis. On the basis of the results from univariate analysis, those variables shown to have a statistically

4 Digestive symptoms in endometriosis 3443 significant relationship with endometriosis localization were selected for subsequent multivariate analysis. Four and five-class variables were transformed into two-class variables, and a multinomial regression logistic model was used to estimate the adjusted odd ratio (OR), which corresponded to a likely presence of a deep rectal nodule. The multinomial regression logistic model is usually used to model the relationship between an outcome variable of more than two categories (in this case, several distinct localizations of endometriosis) and one of more covariates or risk factors (patient characteristics and questionnaire responses) (Ancel, 1999). Adjusted ORs were then compared using the Wald test. Statistical analysis was performed using the Stata 9.0 software (Stata Corporation, TX, USA). Median values, percentiles, range, mean values and SD were calculated for continuous variables, and percentages were calculated for the qualitative variables. Variable distributions depending on groups were compared by univariate analysis [Fischer s exact test for qualitative parameters and analysis of variance (ANOVA) test, Kruskall Wallis test and Mann and Whitney test for continuous variables]. The post hoc statistical power of univariate tests comparing the values of different items of KESS and GIQLI scores was estimated using the Stata command simpower. P, 0.05 was considered to be statistically significant. Results Population characteristics From January 2010 to October 2011, 212 women were surgically treated for endometriosis in our department and included in the CIRENDO database. Among them, respectively 21, 42 and 53 fulfilled inclusion criteria in the three groups. Table I presents the main patient characteristics. These were comparable with the exception of age and antecedents including miscarriage, ectopic pregnancy and infertility. Table II presents major painful pelvic symptoms related to endometriosis, namely dysmenorrhoea and deep dyspareunia, and their impact on patient social and sexual behaviour and quality of life. Women presenting with deep endometriosis infiltrating the rectum reported a significant increase in intensity and duration of dysmenorrhoea, while deep dyspareunia appeared to be more impaired in women with superficial endometriosis. In all the three groups, there was severe impact of dyspareunia on sexual activity: % of women reported dissatisfaction concerning their sexual activity in the previous month. Table III presents intraoperative findings, disease localizations and AFSr score values. Table I Patient characteristics before surgical management performed by our team. Patient characteristics Stage 1 superficial Deep endometriosis Deep infiltrating endometriosis of P-value endometriosis n 5 21 (%) n 5 42 (%) the rectum n 5 53 (%)... Age (years) BMI (kg/m 2 ) Age of first periods Age of first sexual intercourse Surgical antecedents for pelvic 4 (19.1) 9 (21.4) 21 (39.6) 0.09 endometriosis Obstetrical antecedents Pregnancies 12 (57.1) 20 (47.6) 29 (54.7) 0.71 Deliveries 12 (57.1) 15 (35.7) 20 (37.7) 0.26 Previous use of contraception 16 (76.2) 26 (66.7) 35 (70) 0.62 Contraceptive pills 16 (76.1) 24 (57.1) 30 (56.6) 0.26 Pregnancy intention when 7 (36.6) 18 (52.9) 23 (51.1) 0.54 referred for surgery Infertility 2 (9.5) 18 (43.9) 21 (39.6) 0.01 The MOS 36-item Short-Form Health Survey Physical Functioning (PF) Role Physical (RP) Bodily pain (BP) General Health (GH) Vitality (VT) Social Functioning (SF) Role Emotional (RE) Mental Health (MH) Health Transition (HT) Physical Composite Score (PCS) Mental Composite Score (MCS)

5 3444 Roman et al. Table II Main painful symptoms related to pelvic endometriosis. Symptoms Stage 1 superficial Deep endometriosis Deep infiltrating endometriosis P-value endometriosis n 5 21 (%) n 5 42 (%) of the rectum n 5 53 (%)... Dysmenorrhoea* Dysmenorrhoea 18 (85.7) 40 (95.2) 50 (96.2) 0.39 Primary dysmenorrhoea 11 (61.1) 22 (57.9) 33 (66.0) 0.75 Biberoglou and Behrman score ,0.01 of dysmenorrhoea VAS score of dysmenorrhoea Length of dysmenorrhoea (days) ,0.01 Deep dyspareunia** Having sexual intercourse 19 (90.5) 34 (81.0) 46 (86.8) 0.68 Having deep dyspareunia 19 (100) 27 (79.0) 44 (95.7) 0.04 Biberoglou and Behrman score of deep dyspareunia VAS score of deep dyspareunia ,0.01 Related to position 16 (84.2) 24 (70.6) 37 (80.4) 0.52 Frequency of sexual activity last 0.53 month Five times 3 (15.0) 4 (14.8) 6 (13.3) 3 4 times 2 (10.0) 6 (22.2) 11 (24.4) 1 2 times 4 (20.0) 9 (33.3) 15 (33.3) Not at all 11 (55.0) 8 (29.6) 13 (28.9) Satisfied with sexual activity last 0.89 month Very much 4 (20.0) 8 (30.8) 9 (21.4) Enough 8 (40.0) 6 (23.1) 11 (26.2) Not really 7 (35.0) 10 (38.5) 18 (42.9) Not at all 1 (5.0) 2 (7.7) 4 (9.5) *Percentages and P-values are calculated for 18, 40 and 50 women, respectively, presenting with dysmenorrhoea. **Percentages and P-values are calculated for 19, 34 and 46 women, respectively, reporting sexual activity during the last year. Endometriosis localization and type and frequency of digestive symptoms As summarized in Table IV, women presenting with deep endometriosis infiltrating the rectum were more likely to present cyclic defecation pain (67.9%), cyclic constipation (54.7%) and a significantly longer time to evacuate stools. However, these complaints were also frequent in other groups (38.1 and 33.3% for the superficial endometriosis group and 42.9 and 26.2% for the group with deep endometriosis sparing the rectum, respectively). Women presenting with deep endometriosis infiltrating the rectum were also more likely to present with appetite disorders. Results from the two standardized questionnaires (11 and 36 items) and their comparisons are available upon request. Rectal stenosis and digestive symptoms Preoperative CTC revealed rectal stenosis in 14 out of 53 women presenting with deep endometriosis infiltrating the rectum (26.4%). The group was therefore divided into two separate groups according to the presence of rectal stenosis (39 and 14 women), and the aforementioned comparison was performed between the four groups. When compared with other three groups, women with rectal stenosis were significantly more likely to report constipation (P, 0.01), defecation pain (0.04) and appetite disorders (0.01). They also presented with an increased evacuation time (,0.01) and increased stool consistency without laxatives (0.03), and presented a strong tendency towards increased feelings of incomplete evacuation (P ¼ 0.06). The post hoc statistical power of univariate tests comparing the values of different items of KESS and GIQLI scores varied from 20 to 50%. The multinomial logistic regression model revealed that unsuccessful evacuatory attempts was the only factor independently related to rectal stenosis (P ¼ 0.013). The Wald test allowed comparison of adjusted OR, provided by the multinomial logistic regression model, and showed that patients presenting with rectal stenosis were the most likely to report unsuccessful evacuatory attempts and catamenial appetite disorders, compared with other women affected by deep endometriosis. Discussion The high prevalence of digestive symptoms in all the groups supports the hypothesis that a majority of digestive complaints reported by

6 Digestive symptoms in endometriosis 3445 Table III Intraoperative findings and surgical techniques. Intraoperative findings and Stage 1 superficial Deep endometriosis Deep infiltrating endometriosis of P-value surgical technique endometriosis n5 21 (%) n 5 42 (%) the rectum n 5 53 (%)... Surgical route 0.41 Laparoscopy 21 (100) 41 (97.6) 49 (92.4) Laparotomy 0 1 (2.4) 4 (7.6) AFSr score ,0.01 Deep endometriotic nodule Isolated left uterosacral ligament 10 (23.8) 6 (11.3) 0.16 Isolated right uterosacral 5 (11.9) 2 (3.8) 0.24 ligament Both right and left uterosacral 9 (21.4) 1 (1.9) 0.04 ligaments Rectovaginal space 8 (19.0) 16 (30.2) 0.24 Uterosacral ligaments and 9 (21.4) 26 (49.1) 0.01 rectovaginal space Bladder 1 (2.4) 1 (1.9) 1 Other localizations of the endometriosis Right ovary 0 21 (50.0) 35 (66.0),0.01 Left ovary 0 25 (64.0) 39 (73.6),0.01 Sigmoid colon (37.7),0.01 Diaphragm 0 4 (9.8) 6 (11.3),0.01 Douglas pouch obliteration Partial 0 14 (33.3) 10 (18.9),0.01 Complete 0 11 (26.2) 37 (69.8),0.01 women presenting with deep endometriosis infiltrating the rectum may not be due to the mechanical consequences of the infiltration into the rectal wall, but rather to inflammatory phenomena following cyclic micro-haemorrhages in and around endometriosis foci (Roman et al., 2011). This hypothesis may also explain both an increase in intensity of digestive symptoms during menstruation and their subsequent relief following suppressive hormonal therapy. In patients presenting with deep endometriosis with rectal stenosis, certain digestive complaints are significantly more frequent than for other groups. Significant narrowing of the rectum seems to concern only a minority of women with rectal nodules (Mabrouk et al., 2012); however, there is little consensus on the threshold used to define significant stenosis. In the literature, rectal stenosis often appears overlooked when preoperative assessment is limited to endorectal or vaginal ultrasound, which is used to ascertain the depth of rectal wall infiltration and as a basis for performing colorectal resection. In our department, preoperative assessment of deep endometriosis infiltrating the rectum systematically includes CTC (Vassilieff et al., 2011), and therefore, colorectal resection is generally reserved for women with obvious narrowing of rectal diameter (20 25% in our series). The somewhat high frequency of digestive complaints in women presenting with only superficial implants of the Douglas pouch is most likely due to the enrolment process, which involves mandatory laparoscopic exploration of the pelvis. Women included in Group 1 may therefore not be representative of the general population of women with Stage 1 endometriosis. In our department, women supposed to be affected by superficial endometriosis receive first line-hormonal treatment. Only those who show no complete improvement require laparoscopic exploration and are thus enrolled in the cohort. Furthermore, generally only the most symptomatic women or those with the most severe anatomical conditions are referred to our department, and therefore, it remains a possibility that the patients with deeply infiltrating endometriosis and endometriosis infiltrating the rectal wall are also not representative of the general population. Since the goal of this study was to demonstrate that digestive complaints are not exclusively due to rectal infiltration by endometriosis, we do not believe that this selection bias weakens the study. As numerous differences were found not to be statistically significant, we calculated post hoc statistical power for all the tests comparing the values of different items for KESS and GIQLI scores, using the Stata command simpower. We observed that statistical power varies from 20 to 50%. As the power depends on both subject number in each group (which is constant, respectively, 21, 42 and 53 women) and the difference to be analysed (namely the values of any variable in the three groups), it results in a specific statistical power for each test (lowest when the three values to be compared are close). To provide an overall power calculation, we would need to define one principal outcome and to calculate the number of subjects required by a statistical power of.80%. However, this procedure does not guarantee a statistical power greater than 80% for secondary outcomes which are compared between the three groups. Increasing

7 Table IV Digestive complaints and principal findings using standardized questionnaires. Digestive complaints Stage 1 superficial Deep Deep infiltrating P-value endometriosis endometriosis endometriosis of the n 5 21 (%) n5 42 (%) rectum n 5 53 (%)... CIRENDO self-questionnaire Symptoms associated with catamenial pelvic pain Defecation pain 8 (38.1) 18 (42.9) 36 (67.9) 0.02 Nausea 8 (38.1) 8 (19.1) 19 (35.9) 0.13 Blood in the stools 1 (4.8) 3 (7.1) 8 (15.1) 0.43 Constipation 7 (33.3) 11 (26.2) 29 (54.7) 0.01 Diarrhoea 8 (38.1) 18 (42.9) 25 (47.2) 0.75 Bloating 9 (42.9) 25 (59.5) 27 (50.1) 0.43 Appetite disorders 4 (19.1) 3 (7.1) 18 (34.0),0.01 Tiredness 14 (66.7) 31 (73.8) 41 (77.4) 0.62 Headache 11 (52.4) 16 (38.1) 23 (43.4) 0.55 KESS score Total KESS score value (n ¼ 105 women, ANOVA test) Abnormal KESS score (.6) 13 (65) 28 (73.7) 39 (83) 0.26 Item 1. Duration of constipation (0 18 months ¼ 0; 18 months-5 ys ¼ 1; 5 10 ys ¼ 2; ys ¼ 3;.20 ys ¼ 4) Item 2. Laxative use (none ¼ 0; for short duration ¼ 1; regular, long duration ¼ 2; long duration, ineffective ¼ 3) Item 3. Frequency of bowel movement (1 2 times/1 2 days ¼ 0; 2 or less/week ¼ 1; less than once/week ¼ 2; less than once/2 week ¼ 3) Item 4. Unsuccessful evacuatory attempts (never/rarely ¼ 0; occasionally ¼ 1; usually ¼ 2; always, manual evacuation ¼ 3) Item 5. Feeling incomplete evacuation (never ¼ 0; rarely ¼ 1; occasionally ¼ 2; usually ¼ 3; always ¼ 4) Item 6. Abdominal pain (never ¼ 0; rarely ¼ 1; occasionally ¼ 2; usually ¼ 3; always ¼ 4) Item 7. Bloating (never ¼ 0; perceived by patient only ¼ 1; visible to others ¼ 2; severe causing satiety or nausea ¼ 3; severe with vomiting ¼ 4) Item 8. Enemas/digitation (non ¼ 0; enemata/suppositories occasionally ¼ 1; enematas/ suppositories regular ¼ 2; manual evacuation occasionally ¼ 3; manual evacuation always ¼ 4) Item 9. Time taken to evacuate (,5 min ¼ 0; 5 10 min ¼ 1; min ¼ 2;.30 min ¼ 3) Item 10. Difficulty evacuating causing a painful evacuation effort (never ¼ 0; rarely ¼ 1; occasionally ¼ 2; usually ¼ 3; always ¼ 4) Item 11. Stool consistency without laxatives (soft/loose/normal ¼ 0; occasionally hard ¼ 1; always hard ¼ 2; always hard, usually pellet-like ¼ 3) GIQLI score Total GIQLI score value (n ¼ 115 women, ANOVA test) Abnormal GIQLI score (,125) 19 (90.5) 37 (90.2) 50 (94.3) Roman et al.

8 Digestive symptoms in endometriosis 3447 Item 7. Bowel frequency (always ¼ 0; usually ¼ 1; occasionally ¼ 2; rarely ¼ 3; never ¼ 4) Item 30. Bowel urgency (always ¼ 0; usually ¼ 1; occasionally ¼ 2; rarely ¼ 3; never ¼ 4) Item 31. Diarrhoea (always ¼ 0; usually ¼ 1; occasionally ¼ 2; rarely ¼ 3; never ¼ 4) Item 32. Constipation (always ¼ 0; usually ¼ 1; occasionally ¼ 2; rarely ¼ 3; never ¼ 4) Item 34. Blood in stool (always ¼ 0; usually ¼ 1; occasionally ¼ 2; rarely ¼ 3; never ¼ 4) Item 36. Uncontrolled stools (always ¼ 0; usually ¼ 1; occasionally ¼ 2; rarely ¼ 3; never ¼ 4) Mental well-being subscale (Items 10 14) Gastrointestinal defecation subscale (Items 7, 26, 30, 31, 34, 36) Gastrointestinal digestion subscale (Items 1 6, 27, 28, 32, 35) Physical well-being subscale (Items 8, 15, 16, 18 23, 33) statistical power of the comparison for the overall GIQLI score to 80% would require a 25-fold increase in subject number in each group, while a corresponding increase for the overall value of the KESS score would be obtained by a 3-fold increase in subject number. It has been shown that women with pelvic endometriosis are likely to present with other hidden painful abdominal and pelvic diseases related to a pelvic visceral hypersensitivity (Berkley et al., 2005; Chung et al., 2005). Accurate interpretation of digestive symptoms in women with pelvic endometriosis is challenging, as new data suggest a significant co-morbidity between endometriosis (from minimal to severe) and digestive functional diseases such as irritable bowel syndrome (IBS) (Seaman et al., 2008; Issa et al., 2011). It has been shown that IBS is diagnosed times more frequently in women with pelvic endometriosis than in controls free of endometriosis (Seaman et al., 2008). Similarly, patients with minimal to mild and moderate to severe endometriosis are found to demonstrate a high prevalence of symptoms consistent with IBS (respectively 65 and 50%) and a significantly lower pain threshold related to bowel distension (Issa et al., 2011). This high prevalence of digestive symptoms due to visceral sensitivity is likely to exist in our series too, and to concern women in all groups. As this study focuses on the lack of specificity of digestive complaints for deep endometriosis infiltrating the rectum, we do not believe that this should compromise the results but rather that these results emphasize the care that should be taken before definitively attributing all digestive symptoms to endometriotic nodules arising in the digestive tract, and the importance of taking this into account in decisions involving radical colorectal surgery. We have previously suggested that digestive complaints reported by women presenting with deep endometriosis infiltrating the rectum can be explained by three major consequences of disease development: cyclic micro-hemorrhages and inflammation into the rectal wall, anterior fixation of the rectum to the uterine cervix or vaginal fornix, and rectal stenosis (Roman et al., 2011). Rectal fixation on the uterine cervix or the vaginal fornix is due to the backward development of type II and III deep rectovaginal nodules (Squifflet et al., 2002), and may lead to abnormal angulations of the digestive tract, disturbing stool progression and resulting in defecation pain or constipation. Rectal stenosis is usually due to nodule protrusion into the rectum and may be revealed as an indentation by barium enema or CTC (Fig. 4) (Vassilieff et al., 2011). Revealing intrarectal protrusion, however, does not infer that a nodule has mandatorily developed through the full thickness of the rectal wall, as it can concomitantly infiltrate the rectal musculosa and push the rectal layers inside. Although women presenting with deep nodules responsible for rectal stenosis are significantly more likely to report digestive complaints related to severe constipation, they appear to represent only a minority of patients with deep endometriosis infiltrating the rectum (26% in our series and 12% in that of Mabrouk et al., 2012). In the absence of a unanimous definition of rectal stenosis, we considered stenosis to be present in all situations where the rectal diameter was decreased by intrarectal nodule protrusion. Certain authors, however, reserve the term stenosis uniquely to rectal narrowness exceeding 20 or 30% (Mabrouk et al., 2012). The inflammatory nature of endometriosis deposits located in close proximity to the terminal large bowel and local prostaglandin release may explain altered bowel function (Seaman et al., 2008). Digestive complaints (diarrhoea, constipation, dyschesia, tenesmus, defecation

9 3448 Roman et al. Figure 4 CTC revealing rectal stenosis due to deep endometriosis infiltrating the anterior rectal wall: specimen of colorectal resection confirming infiltration of the rectal wall by the deep nodule shown by CTC. pain) especially or exclusively during menstruation may therefore be explained by cyclic irritation of the rectal wall and inflammation acting as an irritant factor, leading to an increase in the daily number of bowel movements (usually described as a diarrhoea and smooth or liquid stools), in defecation pain and in a feeling of incomplete emptying of the rectum. These symptoms are similar to those encountered in women with inflammatory diseases of the colon and rectum or with Douglas pouch abscesses. However, they are cyclic in nature and significant relief can be obtained with suppressive hormonal therapy. Clearly not all digestive complaints can be related to cyclic phenomena. Defecation pain, though significantly more frequent in women presenting with rectal nodules (68%), is also present in women included in other two groups (respectively 38 and 43%). Fauconnier et al. (2002) drew attention to the lack of high specificity of defecation pain for deep endometriosis infiltrating the rectum and pointed to a particularly complex mechanism for this complaint. In a recent study, Mabrouk et al. (2012) reported on the frequency and the mechanisms of digestive symptoms revealed in women presenting with deep endometriosis infiltrating the rectum and free of hormonal treatment. Although 84% of patients enrolled in the study had deep nodules infiltrating the middle rectum (5 8 cm from the anus), only 12% of the nodules led to rectal narrowness.30%. Questionnaires were completed during the post-menstrual phase and focused on digestive complaints perceived during the previous 24 h. Results revealed constipation (40%), a feeling incomplete evacuation (36%) and stool fragmentation (52%) as the most frequent complaints, while dyschesia was evaluated at using a 10-point visual analogue scale. Anorectal manometry performed in the post-menstrual phase revealed a normal rectoanal inhibitory reflex in all women, suggesting that no patient had abnormal bowel motility, rectal functional disorders or nerve plexus dysfunction at the time of the evaluation. However, the major alteration was an increase in resting pressure for the internal anal sphincter in 80% of women, most likely the consequence of chronic inflammation resulting in pain and muscle spasm, as previously observed in other inflammatory diseases (Andersson et al., 2003). These findings lend support to the hypothesis that cyclic inflammation is a critical factor in digestive complaints and furthermore, it may explain why colorectal resection does not result in the relief of preoperative constipation. Any decrease in rectal capacity is likely to both increase the number of bowel movements and impact on the tonus of the internal anal sphincter. This study highlights several important issues in the management of deep endometriosis infiltrating the rectum. Although our study design cannot clearly demonstrate a causal relationship between cyclic inflammation and digestive symptoms, our findings strongly suggest that various digestive complaints may be unrelated to rectal infiltration by the nodule, as they occur in women free of rectal involvement. Our results add weight to the argument for the use of hormonal suppressive therapy as an alternative to surgery in the management of this complex disease. Key to the decision on goals and type of surgery appears to be the presence of rectal stenosis and of factors that allow long-term suppression of cyclic ovarian function (e.g. age, pregnancy intention, adverse effects of hormonal treatment and the possibility of performing bilateral oophorectomy) and not only the size and depth of rectal infiltration. Colorectal resection may therefore be required only for patients for whom hormonal suppressive therapy has not been found to lead to relief of digestive complaints, i.e. in cases where fibrous endometriotic nodules have led to rectal stenosis. Acknowledgements The authors thank Amélie Bréant, the clinical research technician of the North-West Inter Regional Female Cohort for Patients with Endometriosis, for her management of the database used in our study. Authors roles H.R. wrote the first draft of the report and performed statistical analyses. G.S. revised the manuscript. H.R. and J.J.T. performed the surgical procedures. C.S-C. performed computed tomography based-virtual colonoscopy. H.R. and J.N. checked the data recording. All the authors contributed to the writing of the final manuscript and approved it to be published. Funding No financial support was received for this study. Conflict of interest None declared. References Ancel PY. Value of multinomial model in epidemiology: application to the comparison of risk factors for severely and moderately preterm births. Rev Epidemiol Sante Publique 1999;47: Andersson P, Olaison G, Hallböök O, Boeryd B, Sjödahl R. Increased anal resting pressure and rectal sensitivity in Crohn s disease. Dis Colon Rectum 2003;46:

10 Digestive symptoms in endometriosis 3449 Berkley KJ, Rapkin AJ, Papka RE. The pains of endometriosis. Science 2005; 308: Chung MK, Chung RP, Gordon D. Interstitial cystitis and endometriosis in patients with chronic pelvic pain: The Evil Twins syndrome. JSLS 2005; 9: Fauconnier A, Chapron C, Dubuisson JB, Vieira M, Dousset B, Bréart G. Relation between pain symptoms and the anatomic location of deep infiltrating endometriosis. Fertil Steril 2002;78: Fedele L, Bianchi S, Zanconato G, Portuese A, Raffaelli R. Use of a levonorgestrel-releasing intrauterine device in the treatment of rectovaginal endometriosis. Fertil Steril 2001;75: Issa B, Onon TS, Agrawal A, Shekhar C, Morris J, Hamdy S, Whorwell PJ. Visceral hypersensitivity in endometriosis: a new target for treatment? Gut 2011; doi: /gutjnl Knowles CH, Eccersley AJ, Scott SM, Walker SM, Reeves B, Lunniss PJ. Linear discriminant analysis of symptoms in patients with chronic constipation. Validation of a new scoring system (KESS). Dis Colon Rectum 2000;43: Mabrouk M, Ferrini G, Montanari G, Di Donato N, Raimondo D, Stanghellini V, Corinaldesi R, Seracchioli R. Does colo-rectal endometriosis alter intestinal functions? A prospective manometric and questionnaire-based study. Fertil Steril 2012; doi: / j.fertnstert Nieveen van Dijkum EJM, Terwee CB, Oosterveld P, van der Meulen JHP, Gouma DJ, de Haes JCJM. Validation of the gastrointestinal quality of life index for patients with potentially operable periampullary carcinoma. Br J Surg 2000;87: Roman H, Vassilieff M, Gourcerol G, Savoye G, Leroi AM, Marpeau L, Michot F, Tuech JJ. Surgical management of deep infiltrating endometriosis of the rectum: pleading for a symptom-guided approach. Hum Reprod 2011;26: Seaman HE, Ballard KD, Wright JT, de Vries CS. Endometriosis and its coexistence with irritable bowel syndrome and pelvic inflammatory disease: findings from a national case-controls study Part 2. Br J Obstet Gynaecol 2008;115: Slim K. First validation of the French version of the Gastrointestinal Quality of Life Index (GIQLI). Gastroenterol Biol Clin 1999;23: Squifflet J, Feger C, Donnez J. Diagnosis and imaging of adenomyotic disease of the retroperitoneal space. Gynecol Obstet Invest 2002; 54: Vassilieff M, Suaud O, Savoye-Collet C, Da Costa C, Marouteau-Pasquier N, Belhiba H, Tuech JJ, Marpeau L, Roman H. Computed tomography-based virtual colonoscopy: an examination useful for the choice of the surgical management of colorectal endometriosis. Gynecol Obstet Fertil 2011;39: Ware JE, Gandeck B, Kosinski M, Aaronson NK, Apolone G, Brazier J, Bullinger M, Kaasa S, Leplege A, Prieto L, Sullivan M et al. the equivalence of SF-36 summary health scores estimated using standard and country-specific algorithms in 10 countries: results from the IQOLA Project. J Clin Epidemiol 1998;51: