Transvaginal vesicovaginal fistula closure: A retrospective study of 28 cases at a single facility in Japan.

Not applicable.

Ippei Kurokawa: Writing – original draft. Yukiko Nomura: Conceptualization; writing – original draft. Chie Nakagawa: Writing – review and editing. Yoshiyuki Okada: Writing – review and editing. Kohzo Fuji: Writing – review and editing. Yasukuni Yoshimura: Conceptualization; supervision; writing – review and editing.

This retrospective study included patients who underwent VVFC at our center from April 2019 to May 2021. The patients were diagnosed using vaginal examination, cystoscopy, cystography, and magnetic resonance imaging at our center or at referral institutions. Age, body mass index (BMI), etiology, site of fistula, and postoperative outcomes were investigated. We chose a transvaginal approach regardless of location, size, or recurrence, as long as the fistula was >5 mm away from each of the bilateral ureteral orifices. This study was conducted as per the principles of the Declaration of Helsinki and approved by the Institutional Review Board (approval number: 21‐034‐B). Informed consent was obtained from all patients. We perform a modified transvaginal multilayer flap repair technique as described below. 3 A good surgical field was obtained by pulling the structure around the fistula to the operative field. A Lone Star ® retractor system (CooperSurgical, Inc.) was used to keep the vaginal cavity open during surgery. A 10‐Fr pediatric urethral catheter was inserted into the bladder through the fistula tract from the vagina. By placing traction on the fistula catheter with the balloon inflated, the vaginal and bladder walls around the fistula tract were moved into an accessible operative field. Saline with epinephrine was injected into the vaginal wall around the fistula. The circumferential incision was placed slightly lateral to the fistula, avoiding the scarred tissue around the fistula. To avoid subsequent dissection difficulties and impaired blood flow due to shallow incisions, the depth of the vaginal wall incision was sought to be full thickness. Bent cardiovascular surgical scissors, as well as regular scissors and small gauze balls, were used for effective and safe dissections between the bladder and vaginal wall. The scar tissue around the fistula tract with poor blood flow was removed as much as possible. However, in cases with large fistulas, the scar tissues were not removed when tension‐free suturing was predicted to be difficult. The bladder wall was sutured in the mucosa and muscular layer. The mucosa was sutured using polyglactic #4/0 rapid‐absorbable sutures to penetrate the mucosal layer. It was necessary to avoid persistent bleeding from the bladder mucosa that causes postoperative bladder tamponade. The muscular layer was sutured using polyglactic #2/0 absorbable sutures. When the Martius flap was not used, the vaginal wall was closed by continuous or interrupted sutures using polyglactic #2/0 absorbable sutures to avoid overlap with the bladder suture. The existence of strong scarring and ischemic condition around the fistula were expected to adversely affect wound healing. In such cases, Martius flap interposition was applied, replacing the well‐vascularized autologous fat pad grafts between the wound of the bladder and vaginal walls. 4 Regarding surgery, an approximately 6‐cm longitudinal incision was made on the labia majora to vertically detach the well‐vascularized subcutaneous adipose tissue. The adipose tissue was ligated and cut at the top (or the bottom). A tunnel from this incision toward the incision of the vaginal wall was created. The pedicled graft was passed through the tunnel to the site of the bladder closure and was fixed to the vaginal wall using polyglactic #3/0 absorbable sutures to cover the bladder wall wound.

This study included 28 patients referred from other institutions for the treatment of VVF. The background of the patients is shown in Table  1 . Patients' background characteristics. Data are expressed as medians (ranges). Data are expressed as n (%) for categorical variables. Twenty‐one (75.0%) fistulas were associated with hysterectomy. Among the 21 hysterectomies, 18 (85.7%) were open surgeries and three (14.3%) were laparoscopic surgeries. Additionally, 16 (76.2%) were simple total hysterectomies and five (23.8%) were radical hysterectomies. Intraoperative bladder injury was noticed and repaired during surgery in eight patients (32.0%). In the remaining 68% of cases, no injury was noticed during surgery. Surgical records and previous intraoperative findings were requested from all referral institutions, and 15 records of hysterectomies were obtained. According to these, in two cases, the dissection between the bladder and uterus was difficult; bladder dissection was successfully performed in most of the cases. Energy devices were used in all 12 patients for whom information on the type of device used for hysterectomy was available. Nine patients (32.1%) had a history of VVF repair at other institutions; of them, seven (77.8%) underwent repair using a transabdominal approach and two (22.2%) using a transvaginal approach. The size of the fistulas was 5 mm (range: 2–20) and the locations in the bladder were the supratrigone in 16 (57.1%), trigone in nine (32.1%), and bladder neck in three (10.7%) cases (Figure  1 ). The locations of fistulas in the bladder. More than half of the fistulas were located in the supratrigone of the bladder. Approximately 30% of the fistulas were located on the trigone and 10% on the bladder neck. Surgery was performed at least 2 months after the diagnosis of VVF to allow for the possibility of spontaneous closure of the fistula and healing of inflammation. Twelve patients (42.9%) underwent Martius flap interposition combined with VVFC. There were no intraoperative or postoperative complications of Clavien–Dindo classification grade 2 or higher. However, postoperatively, infection at the harvest site of the Martius flap was confirmed in two cases and treated conservatively (Table  2 ). We inserted a urethral catheter in all patients for 2 weeks and prescribed a β3 agonist to 21 patients. The median hospital stay was 8.5 days (range: 5–18). Surgical outcomes. Data are expressed as medians (ranges). Data are expressed as n (%) for categorical variables. Twenty patients (71.4%) had a temporary ureteral catheter inserted at the beginning of the surgery to avoid ureteral injury because the fistula was relatively close to the ureteral orifice. In the remaining cases, catheterization was not necessary because of the low likelihood of ureteral injury. The overall success rate of the VVFC was 85.7% (24/28 patients). The success rate was 89.5% (17/19 patients) if the refractory patients who had recurrence after VVFC at another institution were excluded. Recurrence of VVF was diagnosed when urine leakage was observed during postoperative vaginal examination. Recurrence was observed in four of 28 (14.3%) patients; two had a history of previous VVF repair, and one had a history of radiotherapy. The remaining one had obesity (BMI 38.9 kg/m 2 ), although she had no history of previous repair or radiotherapy. In VVFCs of these four patients with postoperative recurrence, Martius flap interposition was determined to be unnecessary due to the achievements of tension‐free closure. No VVF recurrence was observed in the 12 patients who underwent the Martius flap interposition (Figure  2 ). Recurrence free rate after VVFC. The overall success rate of the VVFC was 85.7%. Recurrence of VVF was observed in four patients (14.3%) who did not undergo Martius flap interposition. No recurrence of VVF was observed in all 12 patients who underwent Martius flap interposition. VVFC, vesicovaginal fistula closure; VVF, vesicovaginal fistula.

The protocol for this research project was approved by a suitably constituted Institutional Reviewer Board at Showa University (Approval number, 21‐034‐B).

Consents for participation and publication were acquired from all patients.

Not applicable.

In Japan, as in other developed countries, VVF is mostly iatrogenic and is a very rare disease. This is the first study to describe the experience of a relatively large number of VVFCs in Japan. In developed countries, 83.2% of VVFs occur after surgery, and 95.2% of VVFs in developing countries are childbirth‐related. 5 Additionally, of the 2055 fistulas in well‐resourced countries, hysterectomy by any route was an etiological factor in 62.7% of all fistulas and 75.4% of fistulas resulting from surgery. In reports of urinary tract injuries related to hysterectomy based on Finnish nationwide data, the total incidence of bladder injuries was 1.3/1000 procedures and the incidence of 0.08% for VVFs was reliable: 0.22% in laparoscopic, 0.1% in total abdominal, 0% in supracervical, and 0.02% in vaginal hysterectomies. 6 In a population‐based cohort study that included 296 130 women undergoing hysterectomy for benign indications (open abdominal surgery: 36.9%, laparoscopic: 26.1%, vaginal: 21.2%, and laparoscopic‐assisted vaginal approach: 12.3%), there were 2058 (0.7%) bladder injuries and 834 (0.3%) genitourinary fistulas (80/834 of which developed after an injury repair). The immediate identification and repair of genitourinary injury were associated with a reduced risk of subsequent genitourinary fistula formation. 7 Aoki et al. reported that 92.9% of VVFs occurred during pelvic surgery (mainly hysterectomy) in Japan. 8 In the present study, hysterectomy‐related VVFs accounted for 75.0% of all fistulas. No injury was noticed during previous surgery in 68% of the cases. VVFs might develop due to delayed thermal damage to the vaginal cuff and bladder wall, despite no apparent intraoperative bladder injury. Therefore, we requested all referral institutions to provide the surgical records and intraoperative findings, and 15 institutions responded. According to these 15 records of previous surgery, we found only two cases in which the dissection between the bladder and uterus was difficult; bladder dissection was successfully performed in most cases. Energy devices (including electrocauteries) were used in all 12 patients for whom information on the type of device used for hysterectomy was available. These results might support the hypothesis that not only direct bladder injury, but intraoperative delayed thermal damages to the vaginal cuff and bladder wall were also responsible for the development of VVF. Thermal injury can occur during monopolar or bipolar electrosurgical instrument use during bladder dissection in hysterectomy. Electrosurgical instruments should be used carefully, with low power and voltage settings and for the least amount of time in the vicinity of the bladder. 9 During laparoscopic hysterectomy, electrosurgical instrument use is more frequent than during an abdominal hysterectomy. The higher incidence of VVF in laparoscopic hysterectomy might be due to the thermal injury in the bladder and vagina. As the incidence of VVF might increase as laparoscopic total hysterectomy becomes more common, more careful use of coagulation equipment is desired. The VVF caused by hysterectomy often occurs in the trigone or supratrigone of the bladder. 10 , 11 , 12 Therefore, in most cases, a transvaginal approach is possible if there are sufficient distances between each of the bilateral ureteral orifices and fistula. Transvaginal repair is recommended as the first choice for repairing iatrogenic VVFs because of its simplicity, efficacy, and low mobility. 13 There are no differences in success rates between transabdominal and transvaginal surgeries, and the latter requires a shorter length of stay and has fewer complications. 14 However, the transvaginal approach is limited by the location and size of the fistula. Therefore, transabdominal repair tends to be indicated for cases with VVFs in high position and complex VVFs or a possibility of urinary tract reconstruction. 15 , 16 , 17 In this study, approximately 90% of the fistulas were located on the supratrigone and trigone of the bladder. The transvaginal approach was performed in all cases because the fistula was >5 mm away from the bilateral ureteral orifices. We actively inserted ureteral catheters if the fistula was relatively close to the ureteral orifices to avoid ureteral injury. Securing a good surgical field and paying attention to ureteral orifices could help the success of transvaginal VVFC in most of the iatrogenic VVFs. However, there is no standard surgical treatment algorithm, and the approach of VVFC is chosen according to the surgeon's preference. 5 , 13 In a review of 124 studies, 1379 of 1430 (96.4%) patients were treated surgically. The preferred surgical approach was a transvaginal route, followed by the transabdominal/transvesical approach, laparoscopic/robotic approach, and combined techniques with reported success rates of 93.82%, 97.05%, 98.87%, and 90.70%, respectively. 1 The first attempt at surgical repair is most likely to be successful; therefore, it should be performed by a surgeon with extensive experience in fistula repair. 5 In this study, the overall success rate of VVFC was 85.7%. However, the success rate was 89.5% if the refractory patients who had recurrence after VVFC at another institution were excluded; this is a slightly lower success rate than the findings of other studies. This might be due to the fact that this study included a relatively large proportion of complicated cases with a history of one or more prior failed repair surgeries (32%) and radiotherapy (11%). When limited to patients with a history of prior repair, the success rate decreased to 77.8%, (7/9 patients) indicating the importance of success at the first attempt of repair. Nevertheless, it is difficult to acquire the techniques of VVFC, including Martius flap interposition. A questionnaire study on VVFs in Japan revealed that only 6.5% of surgeons who had ever performed VVF surgery had performed VVF surgery in more than 10 cases. 18 Therefore, it appears desirable to consolidate patients with VVFs to facilities with experts. Ayed et al. evaluated the prognostic factors for recurrence of VVF. 19 The risk of recurrence was significantly higher in the presence of multiple fistulas, larger fistula size (>10 mm), complex VVFs including the bladder neck or urethra, obstetric etiology, and pre‐repair urinary tract infections. Furthermore, the most important factor for successful VVFC was adherence to basic surgical principles, including careful preoperative evaluation, wide exposure of the fistula and surrounding tissues, tension‐free closure, excision of all fibrosed tissue, and initial maintenance of an uninfected and dry suture line. We always perform VVFC per the basic surgical principles described above. However, in this study, recurrence was observed in four patients who had not undergone Martius flap interposition. At the beginning of this study, Martius flap interposition was not applied if tension‐free closure could be achieved with tissues with sufficient blood flow, even with a history of previous repair or radiation therapy or with severe obesity. Currently, due to the experience with these cases of recurrence, the Martius flap is used in all cases with a history of previous repair or radiation therapy, and no recurrence is observed thereafter. There was no recurrence in all the 12 cases in which the Martius flap was used. In one of the recurrent cases, obesity was the only comorbidity. Poor wound healing in overweight cases is common, however, it remains unclear whether overweight is a risk factor for VVFC failure. Conversely, Streit‐Cieckiewcz et al. reported that the success rate of VVFC was higher in patients with high BMI. 20 Therefore, we currently do not consider obesity alone as an indication for the Martius flap. Vascularized tissue flaps are used to reinforce a repair, to fill dead space, and to improve vasculogenesis following a repair, although there is no high level of evidence to confirm the benefit of tissue interposition. 5 The Martius flap is a method of tissue replacement using the subcutaneous tissue of the labia majora as a pedicle flap, which is widely used in transvaginal surgery. It is performed in patients with poor blood flow in the tissue around the fistula due to severe scarring or other intraoperative findings. Ali et al. reported a lower recurrence rate with using a Martius flap in VVFC than without using it (3.33% vs. 13.33%). 21 Conversely, Singh et al. reported that the success rate of VVF repair via the transvaginal route did not differ for use or non‐use of the Martius flap interposition (93.1% vs. 96.43%); the Martius flap might induce numbness and pain in the labia majora. However, complicated VVFs were excluded from the study. 22 In the present study, there were two cases of infection at the labia majora wound. In both cases, the wound was opened and conservatively treated. No patients complained of persistent numbness and pain in the labia. Using a Martius flap might improve the outcomes of VVFC without severe complications, especially in patients with a history of previous repair or radiation therapy. The limitation of this study was that the population was small. In Japan, it has been difficult to report on a large number of VVF cases because the patients tend to be distributed to regional facilities and obstetric VVF is extremely rare. On the other hand, the strength was the use of energy devices in the referral institutions, although not in all cases. Further investigation of more cases in the future is necessary to develop useful methods for prevention of VVFs. In conclusion, we reported on 28 patients with VVFs at our center who underwent VVFC by the transvaginal approach. The success rate of VVFC was 85.7%. Intraoperative delayed thermal damage at the vaginal cuff and bladder wall might be responsible for the development of VVF following hysterectomy. Martius flap interposition might improve the outcome of VVFC, as no recurrence of VVF was observed among patients who underwent this procedure. Further investigation with a larger number of cases is necessary.

Vesicovaginal fistula (VVF) is a relatively rare disease in developed countries, although it is often caused by childbirth in developing countries. Most VVF cases in developed countries are surgically induced; the majority are related to total hysterectomy. 1 However, the incidence of VVF was reported to be only 0.1%–0.2% among total hysterectomies. 2 VVF is often difficult to diagnose and to treat without sufficient experience. We have treated a relatively large number of iatrogenic patients with VVF in Japan. Therefore, this study aimed to report the background of patients with VVFs at our center and our ingenuity and outcomes of vesicovaginal fistula closure (VVFC). Additionally, the causes of VVF are discussed.

The authors declare no conflict of interest. No funding was received for this study.