Introduction
Deep rectovaginal endometriosis nodules in young women may involve pelvic nerves, particularly when there is infiltration of the parametrium, with or without adjacent vagina, rectum or ureters. Parametrial nodules may compress, surround or infiltrate nerves of large diameter, such as sacral roots, sciatic or pudendal nerves, or thin fibres such as splanchnic or hypogastric nerves and the inferior or superior hypogastric plexus. Depending on the role of nerve fibres, somatic and vegetative symptoms may occur, such as pain in buttocks, legs or perineum, skin hypo or hyperesthesia, bladder and rectal dysfunction or vaginal dryness. These relatively rare symptoms may be overlooked by gynaecologists or considered unrelated to nerve involvement in a severe endometriosis1.
Over the past two decades, several authors have reported series of patients surgically managed for endometriosis with various degrees of involvement of sacral roots and the sciatic nerve1-3. Large diameter nerve involvement generally occurs in two main ways, which in some patients may be associated. i) The first and most frequent situation concerns large rectovaginal or uterosacral nodules, which develop laterally through the parametrium to the lateral pelvic wall, pyriformis and levator ani muscles, where they come into contact with the sacral plexus (Fig 1a). These nodules involve not only the sacral roots, but also the inferior hypogastric plexus, splanchnic and hypogastric nerves, as well as low rectum, vagina or homolateral ureter. Patients presenting with this type of nodule usually present bladder, rectum and left colon dysfunction in addition to vaginal dryness, more likely due to the involvement of the inferior hypogastric plexus and thin nerves than to that of sacral roots. They may also report somatic pain, particularly in the buttock, perineum and leg, due to involvement of sacral root S2, S3 and S4, which send fibres into the sciatic and pudendal nerves; ii) the second less frequent situation involves deep endometriosis nodules which occur more laterally and cranially, in contact with the pelvic wall and sciatic nerve, before exit through the greater sciatic foramen (Fig 1b). Patients present somatic, sensory (buttock and leg pain) and motor complaints (alteration of the Achillean reflex and foot drop, due to involvement of S1). Conversely, they usually have no vegetative symptoms, due to no involvement of sympathetic hypogastric nerves and parasympathetic fibres originating from S2-S4, and no associated infiltration of rectum, ureters and vagina4.
Outcomes of surgical management of deep endometriosis infiltrating the large nerves have been reported by several authors in the literature, and considered satisfactory in the majority of cases1,2,4. To further knowledge in this field, our study aimed to assess preoperative complaints, intraoperative findings, surgical procedures and postoperative outcomes in a series of consecutive patients managed for deep endometriosis involving sacral roots and the sciatic nerve.

Patients and methods

We enrolled in this series all patients with deep endometriosis involving sacral roots and the sciatic nerve, who had benefited from surgical management carried out over 40 consecutive months from October 2016 to April 2019 at Rouen University Hospital, the Endometriosis Centre, Clinique Tivoli-Ducos, Bordeaux, France and Aarhus University Hospital, Denmark. Inclusion criteria were: i) deep endometriosis nodule with limits that come into contact with, entrap or infiltrate sacral roots or the sciatic nerve shown on pelvic MRI, in axial, coronal and sagittal views; ii) presence of vegetative (bladder, rectal dysfunction) or somatic complaints (pain in the leg, buttock or perineum; motor weakness) attributable to pelvic nerve involvement; iii) surgical procedure requiring complete dissection, shaving or intraneural excision of endometriosis of sacral roots or sciatic nerve; iv) patient inclusion and follow up in CIRENDO database. Exclusion criteria were: i) non-French nationality and non-English speaking patients, ii) patients not fulfilling at least one inclusion criteria.
All patients were preoperatively examined by a gynaecologist surgeon whose practice exclusively focuses on deep endometriosis (H.R). Preoperative assessment was performed by radiologists with considerable experience in deep endometriosis and included pelvic MRI, along with endorectal/transvaginal ultrasound, computed tomography based virtual colonoscopy, uro-CT or uro-MRI when required. Urodynamic assessment was performed in patients with baseline symptoms suggesting significant bladder dysfunction. Surgery was proposed in patients for whom symptoms were not relieved by medical therapy. Principles of surgical approach were preoperatively discussed, and patients were fully informed about the aims, risks and expected benefits of the surgery.
All patients were managed laparoscopically. The approach for deep endometriosis infiltrating the parametrium down to the sacral plexus is standardised (Supplemental Video 1) and follows successive steps: 1) ureterolysis; 2) opening of the pararectal space on contact with the rectum in a sagittal plane; 3) dissection prolonged toward the rectovaginal space, including the section of rectovaginal nodules in one fragment infiltrating the rectum, and another infiltrating the vagina and the parametrium, down to levator ani muscles; 4) dissection of presacral space, identification of the superior hypogastric plexus and hypogastric nerve; 5) incision of the peritoneum at promotorium level, prolonged laterally above the origin of hypogastric vessels; 6) anterograde dissection of hypogastric artery and identification of hypogastric vein; 7) anterograde dissection of hypogastric vein and opening of Okabayashi space, successive identification and when required, ligation of pelvic veins reaching the hypogastric vein. This step is of major importance because the surgeon should be able, at any moment, to ligate the hypogastric vein in case of dissection-related injury; 8) dissection is prolonged behind venous network, with identification of pyriform muscles and sacral root S2, S3 and S4, as well as that of splanchnic nerves originating from sacral roots and reaching the inferior hypogastric plexus; 9) anterograde dissection of the nerve network and inferior hypogastric plexus, up to posterior limits of the deep endometriosis nodule; 10) excision of deep endometriosis nodule, from the posterior limit, to the inferior limit on contact with sacral roots, which should be released or shaved, then to the lateral limit on contact with pyriform muscle and lateral pelvic wall. During this step, dissection is less standardised, and follows nodule limits step by step, with sacrifice of involved fibres depending on inferior hypogastric plexus and ligation of several parametrial veins; 11) excision of adjacent vaginal infiltration; 12) removal of rectal infiltration using shaving, disc excision or segmental resection.
Management of deep endometriosis nodules infiltrating the sciatic nerve employs a different approach, laterally from external iliac vessels (5). The procedure includes the following steps (Supplemental Video 2): 1) longitudinal incision of the peritoneum covering external iliac artery; 2) opening and dissection of the space identified laterally by the psoas muscle and medially by the external iliac artery and vein; 3) the dissection progresses more deeply, with identification of the obturatory nerve which is pushed medially, on contact with the psoas muscle; 4) opening of the lumbosacral space, below the level of obturatory nerve, and identification of the sciatic nerve, resulting from the confluence of L5, S1 and S2 roots. During this step the deep endometriosis nodule is identified; 5) dissection of the nodule, in lateral direction on contact with the muscle, up to the greater sciatic notch; 6) removal of the nodule by progressive dissection on contact with sciatic nerve; haemostasis of surrounding gluteal veins should employ clips as coagulation may be imperfect and veins may retract after section, resulting in hidden bleeding which may be challenging to control; 7) complete removal of nodule allows exposure of pudendal nerve, located medially and caudally, just before its exit through the lesser sciatic foramen.
Associated localisations of the disease were managed during the same procedure. We performed shaving, disc excision or segmental resection to remove colorectal nodules, depending on nodule features6. Deep endometriosis responsible for stenosis of ureters was managed by ureterolysis or ureteral resection, followed by either end-to-end ureteral anastomosis or reimplantation of the ureter into the bladder, depending on the extrinsic or intrinsic infiltration of the ureter. Ovarian endometriomas were treated by vaporisation using plasma energy, cystectomy, sclerotherapy or drainage, depending on ovarian reserve and further intention to conceive. Other endometriosis localisations were routinely removed by excision, until all macroscopic lesions were treated. Following the procedure, the surgeon completed a surgical questionnaire with responses recorded in the CIRENDO database. An item “sacral plexus involvement” allowing recording of patients managed with this localisation was added in January 2016.
Postoperative hospitalisation varied from 4 to 6 days. Clinical symptoms and body temperature were recorded 3 times/day, and assessment of blood values of C-reactive protein (CRP) and white blood cells (WBC) was routinely performed at day 4, 5 and 6. Administration of pregabalin by progressive doses began day 1 after surgery and stopped after two weeks, only if neuropathic pain, paraesthesia or hypo/hyperaesthesia were completely absent. The urinary catheter was removed on postoperative days 1 or 2 and post-voiding bladder residue (PVR) was recorded using an ultrasonic bladder scanner (BladderScan BVI 6100, Verathon Inc, Bothell, WA, USA) after each spontaneous void or every 4 hours. Any PVR greater than 100 mL was considered abnormal and further bladder scanning was performed. If PVR was greater than 400mL, the bladder was emptied using an ‘in out’ intermittent catheter. If the patient was completely unable to urinate, routine intermittent catheterisation was carried out 5 times per day (at 7am, 11am, 2:30 pm, 6pm and 10pm). When PVR remained high (over 100-150 ml) 3 to 4 days after surgery, patients were asked to perform intermittent self-catheterisation 5 times per day following discharge. Alfuzosine 10mg/day was prescribed for 3 months to reduce urethral opening pressures. One month postoperatively, patients were seen with a 48-hour record of urinary volumes voided spontaneously or drained via self-catheterisation. If PVR was less than 100ml, self-catheterisation was ceased; if PVR was greater than 100ml, self-catheterisation was continued for a month or more.
PVRs during hospitalisation and during the period of self-catheterisation were carefully recorded in patient medical charts. Postoperative complications and the duration of self-catheterisation required were recorded in the CIRENDO database, as well as follow up data (pain, desire to conceive, pregnancies, standardised questionnaires on gastrointestinal and urinary outcomes and quality of life).
Prospective recording of data concerning antecedents, clinical symptoms, findings of clinical examinations and diagnostic imaging, surgical procedures and postoperative outcomes was performed through the CIRENDO (North-West Inter Regional Female Cohort for Patients with Endometriosis) database (NCT02294825). This prospective cohort is financed by the G4 Group (The University Hospitals of Rouen, Lille, Amiens and Caen) and coordinated by one of the authors (H.R.). Information was obtained using self-questionnaires and surgical and histological records, while data recording, contact and follow-up were carried out by 2 clinical research technicians. To ensure follow up, women enrolled in the database were asked to answer questionnaires 1, 3, 5, 7 and 10 years after surgery. Standardised gastrointestinal questionnaires were routinely used to assess pre- and post-operative digestive and urinary function: the Gastrointestinal Quality of Life Index (GIQLI)7, the Knowles-Eccersley-Scott-Symptom Questionnaire (KESS)8, the Wexner anal continence score9 and the Urinary Symptom Profile scale (questions 8th, 9th and 10th exploring bladder dysfunction)10. Quality of life was assessed using the SF-36 score11. Prospective recording of data was approved by the French authority CCTIRS (Advisory Committee on information processing in healthcare research).
Statistical analysis was performed using Stata 11.0 software (StatCorp). Patient characteristics, surgical procedures, postoperative outcomes and score values were presented as numbers and percentages (qualitative variables) or mean and SD (continuous variables). Comparison between baseline and 1-year follow up data was performed using paired Student t test (continuous variables) and McNemar test with correcting for continuity (qualitative variables). The study was approved by the Rouen University Hospital Institutional Ethics Committee for Non-Interventional Research (E2020-21, April 13, 2020).