Vascular‐ and nerve‐sparing bowel resection for deep endometriosis: A retrospective single‐center study

Surgical management of bowel endometriosis is still controversial. Recently, many authors have pointed out the potential benefits of preserving the superior rectal artery, thus ensuring better perfusion of the anastomosis. The aim of this study was to evaluate the complication rate and functional outcomes of a bowel resection technique for deep endometriosis (DE) involving a nerve‐ and vascular‐sparing approach.


| Ethics statement
This study received approval from the institutional review board of the "Dipartimento Universitario Scienze della Vita e di Sanità Publica" (protocol number DIPUSVSP- 15-11-2238) and was performed in accordance with the Declaration of Helsinki.During preoperative evaluation, patients were asked in advance to sign a consent form regarding the subsequent use of their anonymized data.

| Variables and procedures
Segmental bowel resection was performed in patients for whom medical therapy had failed to control symptoms, in patients with simultaneous bowel obstruction or nodule residue >3 cm after a shaving technique, or in patients with multiple bowel nodules.Usually, before surgery women try at least one progestin such as dienogest norethisterone acetate, or desogestrel.All of the women had a histologically confirmed diagnosis of endometriosis.We excluded all women aged <18 years who had previous discoid or segmental bowel resection for any benign or malignant diseases, pelvic external beam radiotherapy/brachytherapy, or concomitant diagnosis of diabetic microangiopathy/vasculopathies.Retrieved data included medical and surgical history at preoperative evaluation.Moreover, all of the women were subjected to rectovaginal examination, advanced transvaginal ultrasonography, and/or pelvic magnetic resonance imaging.In cases of subocclusive symptoms, either a colonoscopy or double-contrast barium enema was also required to evaluate stenosis.Along with the above, interviews on pain symptoms and questionnaires on pelvic organ function were also conducted.Specifically, we focused on the main demographic, anthropometric, and clinical data (i.e.age, body mass index, and previous surgery); clinical variables (pain symptoms, urinary, and gastrointestinal and sexual function both before and after intervention); surgical findings (operating time, estimated blood loss, any intraoperative complications, length of resection, distance of the nodule from anal verge, and the need for ileostomy); and perioperative data (days of hospitalization, need for self-catheterization, and postoperative complications).
Postoperative complications, occurring within 30 days after surgery, were described using the Clavien-Dindo classification. 11Six months after surgery, patients underwent rectovaginal evaluation, transvaginal, and transabdominal ultrasonography.Interviews regarding pain symptoms and questionnaires were also reassessed.
The severity of pain symptoms (dysmenorrhea, dysuria, dyschezia, and dyspareunia) were assessed using visual analog scale (VAS) scores (ranging from 0 to 10, from absence of pain to most severe).

K E Y W O R D S
bowel endometriosis, bowel resection, complications, deep endometriosis, functional outcomes, vascular-sparing Information regarding functional outcomes was assessed using validated questionnaires: the Knowles-Eccersley-Scott Symptom (KESS) questionnaire, 12 the Gastro-Intestinal Quality of Life Index (GIQLI), 13 the Bristol Female Lower Urinary Tract Symptoms (BFLUTS) questionnaire, 14 and the Female Sexual Function Index (FSFI). 15The KESS questionnaire was used to assess bowel function and specifically determine whether the patient experienced constipation (0 to 39 points).We used a cutoff criterion of ≥10 points in the total KESS score to define constipation. 12The GIQLI was used to describe the health-related quality of life (QoL) of patients with gastrointestinal disease (0 to 144 points).The questionnaire consists of 36 items and a higher score indicates a better QoL. 13 The BFLUTS questionnaire was instead used to assess urinary function, 14 with total scores ranging from 0 to 45, where higher scores imply decreased bladder function.Finally, for the assessment of sexual function, we applied the validated Italian translation of the FSFI, i.e. 19 questions exploring all domains of sexual function.A total FSFI score <26.5 was considered to be female sexual dysfunction. 15Urinary retention was defined by a postvoiding residual volume of 100 mL.In these cases, self-catheterization was recommended until the post-urinary residual volume was <100 mL at three consecutive measurements.

| End points and outcome assessment
The primary end point of the study was to evaluate surgical outcomes, such as intraoperative, perioperative and postoperative complications in women who underwent vascular-and nerve-sparing bowel resection.As secondary end points, we looked at functional outcomes related to pelvic organs (bowel, urinary, and sexual function), assessed using validated questionnaires, and pain symptoms both at baseline and at 6-month follow-up, in order to highlight potential improvement after the intervention.Furthermore, we evaluated the correlation between functional outcomes at follow-up, by means of KESS, GIQLI, FSFI, and BFLUTS questionnaires, and surgical, anthropometric, and intraoperative findings.

| Surgical technique and postoperative care
When preparing for surgery, all patients followed a 5-day residuefree diet and received mechanical bowel preparation in the form of a 4-L split dose of Macrogol 2 L 2 days before surgery and 2 L the day before surgery.Intravenous cefuroxime and metronidazole were administered intraoperatively as antibiotic prophylaxis.
All patients were operated on by a multidisciplinary surgical team highly experienced in the laparoscopic surgical excision of bowel endometriosis, including a gynecologist and a colorectal surgeon.The severity of the disease was intraoperatively classified using the revised American Society for Reproductive Medicine (r-ASRM) score 16 and the Enzian classification. 179][20] When DE involved the lateral and/or posterior parametrium, a nerve-sparing parametrectomy was performed, as previously described. 2,21In the case of further ureteral involvement due to the disease, ureterolysis was performed first and, if this failed to solve ureteral infiltration, ureteroneocistostomy was performed. 22gmental bowel resection was performed following the same steps, i.e. a 5-mm trocar was added in the right hypochondrium.
Then, the peritoneum of the mesosigma was opened above the root of the IMA, as close to the bowel wall as possible.Sigmoid vessels, which supply the bowel segment to resect, were progressively identified and selectively coagulated.The dissection was performed until the rectal wall below the endometriotic nodule was reached, and then the rectum was transected with a linear stapler (Echelon Flex Endopath Stapler 60 mm; Ethicon).Colorectal anastomosis was performed by extracting the segment of bowel to be resected through a suprapubic mini-Pfannenstiel incision (4-5 cm) and following the classic steps, 18,23 or with a totally intracorporeal anastomosis (TICA). 24See Video S1 in the supplementary material.The choice of a totally intracorporeal anastomosis (introduced to our institution in 2021) or of a mini-Pfannenstiel incision was made at the discretion of the gynecologist and colorectal surgeon.
Following the TICA technique, before anastomosis, the anvil of the circular stapling device (EEA circular stapler with Tri-Staple technology, 28 or 31 mm Medium/Thick; Covidien) was prepared with a 0 vicryl suture, bound at the hole of the tip (Figure 1).The anvil was brought into the abdominal cavity through the opening for the 12-mm port in the right abdominal flank.A colotomy was performed at the colonic wall just proximal to the endometriotic nodule (Figure 2), and then the anvil was introduced into the colon through the colotomy (Figure 3).The linear stapler was arranged to include the whole colostomy.The suture attached to the rod of the anvil needed to be taken from the superior edge of the colotomy, keeping the vicryl suture out of the linear stapler.The colon was then transected with a linear stapler (Figure 4) and the anvil extracted through the colon next to the suture line, pulling on the thread tied to it (Figure 5).Then, the circular stapler was introduced in the rectum and end-to-end anastomosis was performed.The specimen was extracted through the 12-mm port on the right flank F I G U R E 1 Anvil, prepared with a 0 vicryl suture, bound at the hole of the tip.
or through the vagina in cases of hysterectomy.At the end of the procedure, an air leak test was performed to evaluate anastomosis integrity.One drainage was left in place.In the postoperative period, at 3 and 5 postoperative days, white cell count and C-reactive protein measurements were performed to look at potential early postoperative septic complications.Fast-track diet resumption was followed for nutrition.

| Statistical analysis
Only a few studies have investigated vascular-and nerve-sparing bowel resection for DE.Given the exploratory nature of the end points, there was no need for a formal sample size calculation.
Consequently, we enrolled a convenience sample of 61 women in the study, following the time-window chosen for enrollment.
The sample was described regarding its clinical and demographic characteristics using descriptive statistical techniques.Specifically, qualitative data sets were expressed as absolute and relative percentage frequencies, whereas quantitative variables were expressed as either mean and standard deviations or median and interquartile ranges (IQRs), as appropriate.To verify the Gaussian distribution of quantitative variables, a Shapiro-Wilk test was applied.Missing values in quantitative data (all <5%) were treated using the imputeR package (R Project for Statistical Computing), with multiple imputations using Lasso regression methods centered on the mean. 25e-versus post-differences in the overall VAS scales and in the four questionnaires on sexual function and quality of life were analyzed by Student t test or Wilcoxon rank sum test for paired data, as appropriate.[28] Statistical significance was set at a P value <0.05.Suggestive P values were further reported (0.05 ≤ P < 0.10).All analyses were conducted using R software version 4.2.0 (CRAN, R Core 2022). 29

| RE SULTS
In the end, we included 61 women in the study, with a mean age of 38.4 ± 5.9 years and a median body mass index of 22.9 kg/m 2 (IQR, 20.4-24.6).A total of 39 (67.2%) women had not had any previous  1.
Endometriosis was further mapped by the Enzian classification system, as reported in Table 2.

| Intervention and complications rate
All of the women had surgery performed in laparoscopy.More specifically, intestinal nodules were mainly 43 single (73.8%) versus 16 multiple (26.2%), with a mean size of 3.6 ± 0.8 cm.The median resected intestinal tract was 8 cm (IQR, 6-9 cm) and was mainly referred to the rectum (n = 56; 91.8%).

| Complications rate
The rate of intraoperative complications was extremely low (n = 2; 3.3%).Fever was observed in 11 cases (18%).Notably, none of the patients experienced rectovaginal, vesicovaginal, ureteral or vesical fistula, or ureteral stenosis and both uroperitoneum and hemoperitoneum.No cases of reintervention were recorded, while bladder voiding deficit was observed in five patients (8.3%) and urinary tract infections were seen in six (9.8%).Data sets are shown in Table 4.

| Postoperative improvement of symptoms and quality of life
Finally, concering pain symptoms at 6-month follow-up evaluations, all VAS scales related to dysuria, dysmenorrhea, dyspareunia, and dyschezia were significantly reduced, with an overall disappearance of symptoms perception.Moreover, a significant improvement was observed on the constipation index, i.e. the KESS score (mean 18.7 ± 7.6 vs. 13.5 ± 7.5 at 6 months; P < 0.001), and for overall QoL (mean GIQLI 57.6 ± 22.1 vs. 82.8± 20.5; P < 0.001), but urinary function (i.e. the BFLUTS score) did not significantly improve.Furthermore, sexual function worsened in general, although without reaching statistical significance.All data sets are reported in Table 5.

| DISCUSS ION
To our knowledge, our study is one of the few in the literature that aims to investigate complication rates and functional outcomes in segmental bowel resection for DE, focusing on the simultaneous use of Hypothetically, the preservation of the SRA makes it possible to reduce the risk that potentially arises from hypoperfusion of the proximal stump of the rectum, due to a marginal artery of Drummond (anastomosis between the sigmoid arteries and the SRA) that is, for example, insufficient for the anastomosis requirement. 9 However, there is no adequate evidence to establish its superiority compared with the classic technique. 8 our series, there were no major Clavien-Dindo complications 3 or 4.More specifically, no rectovaginal, ureteral, or vesicovaginal fistulas were reported, nor was there ureteral or colorectal anastomosis stenosis.These data findings are consistent with what was reported in a recent review by Darici et al., 8 which showed a rate between 0% and 3.6% for rectovaginal fistulas and between 0% and 6.1% for ureteral fistulas in the group of patients undergoing segmental resection using nerve-and vascular-sparing techniques.
Anastomotic strictures, instead, ranged from 0% to 1.2%.Similarly, rates of postoperative pelvic abscesses (4.9%, three patients) and postoperative hemorrhages requiring transfusions (4.9%, three patients) were similar to those reported by Darici et al., i.e. 0.6% and 4.6%, respectively. 8 did not observe any cases of reintervention, as the only anastomotic leak occurred in a patient for whom a temporary ileostomy had previously been performed due to an anastomosis <5 cm from the anal margin.As such, the patient was treated conservatively, maintaining the stoma for 70 days, and it then closed without complications after performing a barium enema, which confirmed the healing of the millimetric colorectal dehiscence.
Although the most frequent postoperative complication was bladder voiding deficit (8.6%), which was resolved through the use of intermittent self-catheterization within 45 days in each of the five patients, its rate was lower than in other studies on nerve-sparing techniques (0%-22%). 8However, this comparison is not reliable because only a few studies specified parametrectomy, which is in itself considered a risk factor for postoperative urinary retention. 2,18The number of preventive ileostomies, although seemingly high (31.9%), is comparable to the previous literature 18,30 and attributable to our center's choice to perform them in all cases of ultra-low resections, as well as in simultaneous ureteral resection/reimplantation.
Bowel function, on the other hand, improved significantly in our casuistic, as confirmed by the considerable enhancement of both KESS (P < 0.001) and GIQLI (P < 0.001) after surgery for bowel endometriosis, as also previously reported in a recent multicenter study by our group. 2Conversely, other studies have not shown any relief from digestive complaints after segmental bowel resection for DE. 31,32In addition, in a recent study on discoid resection for bowel DE, Roman et al. reported a significant improvement in gastrointestinal function only regarding the GIQLI but not with all validated questionnaires used before and after surgery (KESS, Wexner score, and Bristol scale), though their findings are not reliably comparable due to the absence of segmental resections in his series. 33though surgery was generally associated with an improvement in dyspareunia, as already reported by several authors, [34][35][36] we ob-  median distance of resection from the anal margin (6 cm).Both of these factors can lead to the lesion of a number of autonomic nerves responsible for decreasing blood flow to the vagina and aiding lubrication, as previously reported. 2These data sets, in fact, are consistent with other findings from authors who reported that segmental resection can be linked to either lower sexual quality of life 37 or reduction of sexual pleasure. 18e main strength of our study is that, despite several other articles having investigated the role of surgery for bowel endometriosis, to our knowledge this is the first scientific report focused on complications and functional outcomes, evaluated using well-recognized questionnaires, resulting from a surgical technique based on the preservation of the vascular branch of the IMA, as well as the pelvic orthosympathetic and parasympathetic innervation.
Obviously, our study is not without limitations.First, its retrospective nature and, second, its relatively small sample size did not allow us to fully assess complication rates.Furthermore, the short follow-up period may also constitute a limitation, even though we designed the study to focus on complications and functional outcomes directly connected to the surgical technique of vascular-and nerve-sparing bowel resection, without dwelling on the risk of endometriosis relapse.

| CON CLUS IONS
This study should be considered an initial observation that needs further confirmation through larger-scale and prospective longitudi-

F
I G U R E 2 Colotomy, performed cranially and proximally to the endometriotic.F I G U R E 3 Anvil, introduced into the colon through the colotomy, leaving the thread tied to the anvil out of the colon.F I G U R E 4The colon is transected with a linear stapler including the colotomy in the suture.F I G U R E 5 The anvil is extracted through the colon next to the suture line, pulling on the thread tied to it.surgery for endometriosis.Furthermore, 44 patients (72.1%) had undergone previous hormone therapy at least 3 months before surgery.All patients were classified as having stage III or IV disease according to American Society for Reproductive Medicine (ASRM) (27 [44.3%] and 34 [55.7%]patients, respectively).There were no conversions to laparotomy.At baseline, VAS pain scales disclosed a remarkably high perception of dysmenorrhea (median, 8 [IQR, 5-9]) and dyschezia (median, 8 [IQR 5-9]), whereas pain due to dysuria was low overall.Moreover, there was a mean KESS score of 18.7 ± 7.6, thus highlighting an overall index of constipation.QoL was scarce, expressed by a mean GIQLI of 57.6 ± 22.1.Urinary function was instead good overall, with a median BFLUTS of 9 (IQR, 4-13), but there was an overall indication of sexual dysfunction, as shown by a mean FSFI of 17.5 ± 9.3.All data sets are reported in Table

TA B L E 1
Abbreviations: ASRM, American Society for Reproductive Medicine; BFLUTS, Bristol Female Lower Urinary Tract Symptoms; BMI, body mass index; FSFI, Female Sexual Function Index; GIQLI, Gastro-Intestinal Quality of Life Index; KESS, Knowles-Eccersley-Scott Symptom; VAS, visual analog scale.a Descriptive statistics are expressed as mean (standard deviation) or median (interquartile range) for quantitative variables and as absolute and relative percentage frequencies for qualitative variables.

a
served an overall worsening of sexual function, even though it was not statistically significant.This finding might depend on the high number of posterolateral parametrectomies performed in association with segmental bowel resection (78.7% of patients) and on the TA B L E 4 Intraoperative and postoperative complication rates (N = 61).Descriptive statistics are expressed as median (interquartile range) for quantitative variables and as absolute and relative percentage frequencies for qualitative variables.TA B L E 5 Pain VAS scale and questionnaires evaluation before intervention and at 6-month follow-up (N = 61).
nal studies.Nonetheless, the low rate of complications and the good gastrointestinal outcomes we observed support a more widespread use of surgical techniques that enhance the sparing of the SRA, as well as pelvic orthosympathetic and parasympathetic innervation, in rectus-sigmoid DE surgery.AUTH O R CO NTR I B UTI O N S Manuel Maria Ianieri and Francesco Santullo were responsible for study conception and design and manuscript drafting.Pia Clara Pafundi was responsible for statistical analysis and manuscript drafting.Alessandra De Cicco Nardone contributed to the study design.Greta Benvenga, Pierfrancesco Greco, Maria Vittoria Alesi, and Federica Campolo contributed to data acquisition.Claudio Lodoli and Carlo Abatini contributed to data analysis and interpretation.Giovanni Scambia and Fabio Pacelli supervised the project.All authors have read and approved the final manuscript.