Early surgical outcomes of 550 consecutive patients treated for benign gynecological conditions by transvaginal natural orifice transluminal endoscopic surgery.

Yannick Hurni: conceptualization, data curation, formal analysis, methodology, writing—original draft, visualization. Colin Simonson: data curation, supervision, writing—review & editing. Marcello Di Serio, Régine Lachat, Pauline Bodenmann, Stéphanie Seidler: data curation, writing—review & editing. Daniela Huber: project administration, supervision, conceptualization, data curation, writing—original draft, writing—review & editing.

All patients gave written informed consent. The project received approval from the local ethical committee (Commission cantonale d'éthique de la recherche sur l'être humain CER‐VD, Lausanne, Switzerland), with registration number 2021‐02346 on January 21, 2022.

Between May 2020 and March 2024, 600 patients underwent vNOTES gynecological procedures, 550 for benign disorders and 50 for malignant or premalignant conditions (surgical stagings for endometrial and tubo‐ovarian diseases). Of the 550 included patients, 365 (66.4%) underwent a vNOTES hysterectomy, 167 (30.4%) a procedure limited to the adnexa, and 18 (3.3%) other interventions, including myomectomy, pelvic adhesiolysis, post‐hysterectomy pelvic hematoma drainage, pelvic organ prolapse repair, and appendectomy (Figure  1 ). The mean age was 49.4 ± 12.2 years, and the mean BMI was 26.2 ± 5.8 kg/m 2 . The clinical and demographic characteristics of the patients are summarized in Table  1 . Distribution of different surgical interventions. Clinical and demographic characteristics. Note : Data are presented as mean ± standard deviation or number (%). Abbreviations: BMI, body mass index; vNOTES, transvaginal natural orifice transluminal endoscopic surgery. The total complication rate was 6.5% (36 cases), of which 2.7% (15 cases) were intraoperative complications and 4.0% (22 cases) were postoperative complications. Patients presented postoperative complications classified as CD grade I in 4 cases (0.7%), grade II in 10 cases (1.8%), and grade III in 8 cases (1.5%). We observed no CD grade IV and V complications. Multiple complications were observed in only one patient (0.2%), who presented a pelvic hematoma (CD grade II) and a vaginal cuff infection (CD grade II) following a hysterectomy. Three patients (0.5%) were rehospitalized for postoperative complications management. The other complications were managed during the first hospitalization or on an outpatient basis. The conversion rate was 1.6%, with nine cases of conversion to conventional laparoscopy and none to laparotomy. All conversions were due to technical difficulties entering the abdomen vaginally or performing the procedure by vNOTES, and no conversions were due to manage complications. A hybrid approach combining vNOTES with a transabdominal trocar was used in four cases (0.7%). In all of them, transabdominal access was necessary to improve the exposition in cases of large pelvic masses or occult bleeding. The type, rate, and characterization of complications and conversions are reported in Table  2 and Table  S1 . Surgical procedure, operative characteristics, and perioperative outcomes. Note : Data are presented as mean ± standard deviation or number (%). Abbreviations: CD, Clavien–Dindo classification; vNOTES, transvaginal natural orifice transluminal endoscopic surgery.. The mean age of the 365 patients who underwent a hysterectomy was 50.2 ± 11.3 years, and the mean BMI was 26.6 ± 5.6 kg/m 2 . The mean uterine weight was 223.7 ± 214.9 g (range 35–1520 g), and the mean operative time was 70.1 ± 37.0 min, with a mean estimated blood loss of 90.2 ± 125.8 mL. Intraoperative complications were observed in 12 cases (3.3%). They included a bladder injury in six cases (1.6%), a rectal serosal tear in two cases (0.5%), a bowel injury in two cases (0.5%), a small bowel mesenteric tear in one case (0.3%), and a bleeding needing a transfusion in one case (0.3%). Conversion to conventional laparoscopy was performed in six cases (1.6%). Postoperative complications CD grade I was observed in two cases (0.5%), CD grade II in eight cases (2.3%), and CD grade III in eight cases (2.3%). The most frequent postoperative complication was pelvic hematomas, observed in eight cases (2.2%) and needing surgical drainage in five cases (1.4%). All surgical drainages were successfully performed by vNOTES. The mean age of the 167 patients who underwent an intervention limited to the adnexa was 48.4 ± 14.0 years, and the mean BMI was 25.7 ± 6.3 kg/m 2 . The mean operative time was 43.8 ± 29.0 min, and the mean estimated blood loss was 28.3 ± 50.1 mL. Intraoperative complications were observed in three cases (1.8%) and included a rectal serosal tear in two cases (1.2%) and a bowel injury in one case (0.6%). Conversion to conventional laparoscopy was performed in three cases (1.8%). Postoperative complications were observed in four cases (2.4%) and were classified as CD grade I in two cases (1.2%) and CD grade II in two cases (1.2%).

We reported the perioperative outcomes of 550 consecutive patients operated by vNOTES for benign gynecological conditions. Overall, we observed an intraoperative complication rate of 2.7%, a postoperative complication rate of 4.0%, and a conversion rate of 1.6%. These results align with previous findings from three other large observational studies focused on intra‐ and postoperative complications associated with vNOTES gynecological procedures. 8 , 9 , 10 In 2021, Baekelandt and Kapurubandara reported the outcomes of 1000 vNOTES interventions operated by a single surgeon, including 73% of hysterectomy and 25% of adnexal procedures. 8 They reported an intraoperative complication rate of 1.0%, a postoperative complication rate of 2.9%, and a conversion rate of 0.4%. In 2022, Huang et al. reported the results of 1147 vNOTES procedures performed by a single institution surgical team, with a vast majority of adnexal surgeries (79%), 9% of myomectomies, and only 7% of hysterectomies. 9 Similarly, in 2023, Hou et al. reported data from a cohort of 2000 patients operated by vNOTES by a single institution surgical team, with a 79% rate of adnexal surgeries, 8% of myomectomies, and 7% of hysterectomies. 10 In these two studies, the authors reported an intraoperative complication rate of 0.4%–0.8%, a postoperative complication rate of 2.5%–4.1%, and a conversion rate of 0.4%–1.6%. In our study, iatrogenic bladder injury was the most frequently encountered intraoperative complication and was exclusively associated with vNOTES hysterectomies with a rate of 1.6% (6 cases). Similar results were also reported by Baekelandt and Kapurubandara, who observed a bladder injury rate of 1.2% over 730 vNOTES hysterectomies for benign indications. 8 Huang et al. and Hou et al. reported a limited number of hysterectomy cases, and their rates are, therefore, not representative of the risk of bladder injury in vNOTES. Whereas similar rates of bladder injuries have been reported for vaginal hysterectomies (1.1%), 12 this type of complication seems to be less frequently encountered for laparoscopic hysterectomies, with a rate ranging from 0.05% to 0.9%. 13 , 14 As for a conventional vaginal approach, in vNOTES hysterectomies, bladder injuries frequently occur during vesicovaginal space opening. These injuries are often immediately recognized, easily repaired transvaginally, and do not prevent the main intervention from being completed. In our experience, we always performed an intraoperative cystoscopy to evaluate the bladder's inside wall and the ureteral patency. Bladder injuries at the time of abdominal access through the anterior colpotomy could be associated with factors of surgical difficulty such as limited uterus descent, enlarged uteri, especially in the case of anterior isthmic myomas, deep and narrow vaginas, and in the case of previous cesarean sections with vesicovaginal septum adherences. A possible solution for complex anterior entries is accessing the peritoneal cavity only posteriorly through the pouch of Douglas, limiting the anterior dissection to the most distal part of the vesicovaginal space. The remaining part can then be dissected endoscopically once the vNOTES port has been inserted. To allow a safe vesicovaginal space dissection, surgeons should begin by creating bilateral windows into the paravesical spaces, later moving more medially to free the bladder completely. 15 Ureteral injuries were never reported in our and the three other major vNOTES studies 8 , 9 , 10 while being observed in up to 0.8% of cases in laparoscopic hysterectomies. 13 , 14 , 16 During vaginal hysterectomies, ureteral injuries are observed in around 0.02%–0.04% of cases, and such a low rate might also reflect the presumable low risk of ureteral injuries for vNOTES hysterectomies. 16 , 17 As for the conventional vaginal route, the vNOTES approach seems to reduce the risk of ureteral injuries compared to conventional laparoscopy. This could be related to some specific vNOTES features that allow the protection of the ureter in its most at‐risk portions. On the one hand, the vesicovaginal space opening, one of the first surgical steps in vNOTES hysterectomies, allows the ureters to move laterally and ventrally from the uterine artery pedicles. The Alexis retractor helps maintain this distance constantly during the coagulation and transection of the uterine arteries, possibly reducing the risk of ureteral injury. In addition, the caudal and dorsal approach of vNOTES allows an easier, closer, and constant view of the ureters along their entire course in the pelvis up to the prevesical portion, potentially reducing the risk of injuries during both hysterectomies and adnexal surgeries. However, ureteral injury risk could rise considerably in cases of anatomical distortion, and surgeons should pay particular attention to patients presenting conditions such as severe pelvic endometriosis. We observed four cases (0.7%) of superficial rectal serosal tears associated with both vNOTES hysterectomies and adnexal procedures. Rectal serosal tears were observed after vNOTES port insertion. They were probably related to divergent traction forces caused by the Alexis retractor and the cephalic mobilization of the bowels when freeing the pelvis to improve surgical field exposure. All cases were repaired by vNOTES with absorbable sutures, and the patients presented no subsequent complications. Huang et al. and Hou et al. reported rectal wall injuries in 0.3%–0.4% of cases, mostly occurring in patients with pelvic adhesions during posterior access through the pouch of Douglas and the installation of the vNOTES port, but also at the time of ovarian cystectomy in case of severe pelvic endometriosis. 9 , 10 To reduce these complications, a careful preoperative assessment with history collection, gynecological evaluation, and pelvic ultrasound exam should be performed to detect patients contraindicated for the vNOTES approach. 10 , 18 Characteristics associated with an increased risk of adhesions in the pouch of Douglas or adherences into the rectovaginal space should retain the surgeons to perform interventions by vNOTES. These include a history of pelvic radiations, suspected rectovaginal or retrocervical endometriosis, severe pelvic inflammatory disease, a negative sonographic uterine posterior “sliding sign,” or a firm posterior vaginal fornix which appears impossible to pull once grasped immediately before performing the posterior colpotomy. 10 , 18 , 19 We observed 3 cases (0.5%) of bowel injury associated with both vNOTES hysterectomies and adnexal procedures. A superficial serosal thermal lesion was observed in one case due to inadequate exposure. Immediate vNOTES suturing was performed to reinforce the intestinal wall, and the patient presented no consequent complications. In the second case, the patient presented at postoperative day 1 with peritonitis due to intestinal contents leakage from a 7–8 mm small bowel perforation unrecognized at the time of the vNOTES intervention. Laparoscopic washing was performed, and the intestinal defect was treated with direct wall suturing with no further complications. In the third case, a hysterectomy with bilateral salpingectomy was performed on a patient presenting severe adhesions due to pelvic endometriosis with an unrecognized bowel injury. Due to the impossibility of performing the intervention by vNOTES, we converted early to conventional laparoscopy just after performing anterior and posterior colpotomies with bilateral uterosacral ligaments section. Therefore, it is unclear whether the bowel injury occurred during the transvaginal or the conventional laparoscopic surgical stage. The patient presented at postoperative day 12 with a pelvic abscess associated with a bowel lesion requiring laparoscopic partial small bowel resection with a stoma and a postponed re‐anastomosis 48 days later. Bowel injuries not related to abdominal cavity entry through the pouch of Douglas seem rare in vNOTES. Baekelandt and Kapurubandara and Huang et al. reported no cases over 1000 and 1147 interventions, respectively, while Hou et al. reported only two cases (0.1%) of rectal injury associated with pelvic dissection in patients with severe endometriosis over 2000 interventions. 8 , 9 , 10 During vNOTES interventions, good exposure is essential to reduce the risk of such lesions. On some occasions, the typical vNOTES caudal view does not allow to exclude the presence of bowels adhering to the uterus or adnexa, in contrast to the conventional laparoscopic view from above. In our experience, in addition to appropriate pneumoperitoneum pressures and Trendelenburg positions, we found that using intrabdominal gauze to move and protect sensible neighboring structures such as bowels allowed us to keep them at a safe distance from the sectioned or coagulated tissues, which appeared especially important in cases with difficult exposure (eg in the case of obese patients, enlarged uteri, or large myomas and annexal masses). Whereas bowel injuries are rarely reported during vaginal hysterectomies (0.01% 20 ), these are observed in 0.1%–1.0% of cases of laparoscopic gynecological surgeries. 14 , 20 , 21 Bowel injuries during laparoscopy, as vascular ones, are frequently associated with abdominal cavity entry, especially in patients presenting abdominal adhesions. 22 The vNOTES approach could then be a valid alternative for avoiding transabdominal access in these high‐risk patients once the pouch of Douglas has proven to be free of lesions. Hemorrhage‐associated complications were observed in nine cases (1.8%), with one case (0.2%) of intraoperative bleeding needing a transfusion and eight cases (1.6%) of postoperative pelvic hematoma, of which 5 (1.0%) required surgical drainage, and 3 (0.6%) were treated conservatively with surveillance only. All cases were observed in patients operated for a hysterectomy and were associated with bleeding at the site of the vaginal cuff. Similarly, Baekelandt and Kapurubandara and Hou et al. observed hemorrhage‐associated complications in 1.2% and 1.7% of cases, respectively. 8 , 10 These ratios appear similar to those reported for vaginal and laparoscopic hysterectomies. 14 , 23 , 24 In our case series, three patients (0.5%) presented a vaginal wound infection requiring systemic antibiotic treatment in two cases and surgical management in one case. Similarly, Baekelandt and Kapurubandara observed a postoperative surgical site infection in only two cases (0.2%). 8 Conversely, Hou et al. reported a 0.6% ratio of postoperative infections requiring antibiotics or surgical drainage and a supplementary 0.45% of postoperative fever requiring no specific treatment. 10 Huang et al. reported a 1.0% rate of patients needing postoperative antipyretic treatments, even though they did not explain the source of their fever. 9 Potential differences in postoperative infection rates could be associated with different prophylactic antibiotic therapies, which included a single dose of intravenous cephalosporin in the studies reported by Huang et al. and Hou et al., while all cases reported by Baekelandt and Kapurubandara and in our study received a combination of Clindamycin vaginal cream, intravenous Cefazolin, and intravenous metronidazole. 8 , 10 Similar surgical site infection rates have been reported for both conventional vaginal and laparoscopic gynecological surgeries. 14 , 25 Like Baekelandt and Kapurubandara, we observed a vaginal wound dehiscence ratio of 0.4%, with all cases following a hysterectomy. 8 Similar results have been associated with poor vaginal wound healing in cases operated by vaginal surgery (0.2%) and conventional laparoscopy (0.6%). 14 , 26 In our experience, all patients presented healing defects of less than 1 cm without visualization of the intraabdominal organs and were always treated on an outpatient basis. No specific risk factors associated with poor healing capacity were encountered (eg smoking, diabetes, etc.). The impact of vNOTES on sexuality and the fear of developing postoperative dyspareunia are some frequently encountered concerns of patients preparing to undergo vNOTES adnexal surgeries. 27 , 28 Unlike laparoscopic techniques, vNOTES adnexal procedures imply a posterior colpotomy, intraoperative vaginal distension, and the opening of the pouch of Douglas, which could be associated with the patients' concerns mentioned above. Although only a few studies reported sexuality assessments following vNOTES adnexal procedures, these suggest a very limited risk of developing postoperative sexual dysfunctions. 29 , 30 Only four randomized control trials exist comparing laparoscopic vs vNOTES hysterectomy 31 , 32 and laparoscopic vs vNOTES adnexal surgeries. 33 , 34 These have proved the non‐inferiority of the vNOTES approach compared to laparoscopy for successfully performing hysterectomies and adnexal surgeries without conversion, with similar complication rates, shorter operating times, a higher ratio of hysterectomies performed on 1‐day surgery, and lesser postoperative pain. 31 , 32 , 33 , 34 , 35 Similar results have also been reported by other nonrandomized studies comparing vNOTES and laparoscopy. 36 , 37 , 38 In addition, other advantages associated with vNOTES and the absence of transabdominal accesses include better cosmetics results and the avoidance of risk associated with abdominal access and postoperative incisional hernias. At the moment, no randomized control trials exist comparing vNOTES and vaginal surgery, but a randomized controlled trial comparing the vaginal, vNOTES, and laparoscopic approach for hysterectomy in women with benign conditions has recently been registered ( NCT05971875 ). We acknowledge some limitations of this study, mainly resulting from its single‐institution character and a potentially limited number of patients included, which should be larger to analyze rates and causes of rare complications potentially associated with the vNOTES approach. However, our study represents some important strengths, being one of the most significant studies focused on the analyses of intra‐ and postoperative complications associated with vNOTES gynecological surgeries. Despite coming from a single center, our results appear relatively generalizable as they represent the varied surgical activity of a 3.5‐year experience of a surgical team made up of six gynecological surgeons with different levels of experience with minimally invasive and vaginal surgery, including the outcomes of the training phase for each of them. This could allow other surgical teams to estimate their potential surgical outcomes when introducing vNOTES in their institution or helping them in patient counseling.

Our observational study focused on the surgical complications of vNOTES gynecological benign procedures and demonstrated a profile similar to previous studies. Our results add important information for both research and clinical practices.

In recent years, transvaginal natural orifice transluminal endoscopic surgery (vNOTES) has increasingly been used to treat multiple gynecological conditions, including emergencies and malignancies. 1 , 2 , 3 , 4 , 5 , 6 , 7 However, this surgical approach remains a relatively new technique for many surgeons, and the associated scientific evidence is still being established. Evidence about intra‐ and postoperative complications related to vNOTES gynecological interventions is based on data often reported in relatively small case series. Only three large studies focusing on perioperative complications have been published, reporting the outcomes of cohorts including between 1000 and 2000 patients. 8 , 9 , 10 To improve knowledge in this field, we report and analyze data concerning intra‐ and postoperative outcomes of 550 consecutive patients operated by vNOTES for benign gynecological conditions.

The authors have no relevant financial or non‐financial interests to disclose.

Since May 2020, vNOTES has been introduced in our institution. From January 2022, we started collecting, both retrospectively and prospectively, data concerning patients who underwent vNOTES procedures to create an institutional database. This database was developed using Research Electronic Data Capture (REDCap) software. From this database, we identified the first 550 consecutive patients who underwent vNOTES interventions for benign gynecological conditions. Patients from the first case operated by each surgeon were included without omitting surgeries performed during their training phase. We excluded patients who underwent vNOTES surgical stagings for malignant or premalignant conditions. Patients were operated by one of the six surgeons trained to perform vNOTES interventions at our institution. Surgeons were all senior, fully trained gynecologists specialized in gynecological surgery and minimally invasive techniques but with different levels of experience (from 5 to more than 20 years of experience). They were gradually introduced to vNOTES, from the more experienced to the less experienced, with the latter being fully accompanied during their training phase. All surgeons attended a theoretical and practical training program, including simulations and observing live vNOTES surgeries (Applied Medical vNOTES training program). Approximately 75% of the interventions were realized by a single surgeon (DH), while the others performed around 3%–10% of the surgeries. Interventions were performed with patients in a dorsal lithotomy position under general anesthesia. Patients received prophylactic intravenous antibiotics with cefuroxime 1.5 g and metronidazole 500 mg, and a Foley catheter was placed to keep the bladder empty. To perform hysterectomies, access to the abdominal cavity was achieved through anterior and posterior colpotomies, with transvaginal transection of the uterosacral ligaments in order to achieve posterior access. To perform interventions limited to the adnexa or posterolateral myomectomies, access was gained with a posterior colpotomy through the pouch of Douglas. Anterior colpotomy alone was rarely used to perform anterior uterine wall myomectomies. A vNOTES port (GelPoint vPath, Applied Medical, Rancho Santa Margarita, CA, USA) with an adapted diameter (7, 9.5, or 11 cm) was placed in the abdominal cavity through the anterior and/or posterior colpotomies. Carbon dioxide was insufflated to create a pneumoperitoneum with an intraperitoneal pressure of 10–15 mmHg. Three trocars were used to insert a 10‐mm rigid 30° scope, 5‐mm instruments such as atraumatic and bipolar graspers, and sealing devices. To perform hysterectomies, uterine vessels and broad and round ligaments were sealed and cut caudally to cranially. To perform other procedures, we followed the same technique used by conventional laparoscopy. Specimens were extracted through the vagina. Large uteri were morcellated with a cold knife or cold scissors into an Alexis Contained Extraction System (Applied Medical, Rancho Santa Margarita, CA, USA). Large adnexal lesions were retrieved into an Alexis Contained Extraction System to avoid intraabdominal spilling. At the end of the procedure, the colpotomy was closed under direct vision with a running suture using Vicryl 0. Clindamycin vaginal cream was administered once a day on the evening before the surgery, the day of the surgery, and for the first seven postoperative days to reduce the risk of postoperative vaginal suture infections associated with eventually unrecognized bacterial vaginosis. Demographic features, perioperative information, and data concerning short‐term outcomes were analyzed from the database. Intraoperative data included total operating time (from the initial incision to vaginal closure), time to insert the vNOTES port (from the initial incision to the intrabdominal insufflation of CO 2 ), estimated blood loss, intraoperative complications including bleeding requiring transfusion or procedural iatrogenic organ damage, and the need to convert to conventional laparoscopy or laparotomy. Postoperative data included pain evaluation with the visual analog scale graded from 0 to 10 at 12, 24, and 48 h after surgery, the use of opioid analgesics, hospital stay, and postoperative complications that occurred up to 8 weeks. Postoperative complications were graded using the Clavien–Dindo classification (CD). 11 Continuous variables were presented as mean and standard deviation. Dichotomous variables were presented as absolute numbers and percentages (%). Results are reported for the entire group, for the subgroup of patients who underwent a hysterectomy, and for the subgroup of patients who underwent an intervention limited to the adnexa. No intergroup statistical comparisons were made.

Table S1.