Neurourology and Urodynamics 35:90–94 (2016)

Transvaginal Neobladder Vaginal Fistula Repair After Radical Cystectomy With Orthotopic Urinary Diversion in Women Maude E. Carmel,1* Howard B. Goldman,2 Courtenay K. Moore,2 Raymond R. Rackley,2 and Sandip P. Vasavada2 1

UT Southwestern Medical Center, Dallas, Texas 2 Cleveland Clinic, Cleveland, Ohio

Aim: We present the surgical management and outcomes of patients who underwent transvaginal neo-bladder vaginal fistula (NBVF) repair at our institution. Methods: Between 2002 and 2012, eight patients underwent transvaginal NBVF repair. The surgical management entailed placing a Foley catheter into the fistula tract. A circumferential incision was made around the fistula tract after which a plane between the serosa of the neobladder and the vaginal epithelium was created. Interrupted polyglycolic acid sutures were used to close the fistula. An additional layer of vaginal wall, Martius, or omental flap was interposed before vaginal wall closure. A urethral catheter was placed for a minimum of 14 days and removed after a negative cystogram and pelvic exam with retrograde neobladder filling without leakage. Results: All patients presented with a fistula following radical cystectomy with orthotopic ileal neobladder. Two patients had failed two prior transvaginal fistula repairs. A unilateral Martius flap was used in five patients and an omental flap was used in one patient. The surgery was successful in all patients. After a mean follow up of 33 months [4–117], five patients underwent or are waiting to undergo management of stress urinary incontinence with bulking agents. No patient had a recurrent fistula. Conclusions: Management of NBVF is challenging but cure is possible using a transvaginal approach. Most patients will suffer from incontinence after the repair because of a short and incompetent urethra. Patients should be counseled about the high probability of requiring a secondary procedure to achieve continence. Neurourol. Urodynam. 35:90–94, 2016. # 2014 Wiley Periodicals, Inc. Key words: cystectomy; urinary bladder neoplasm; urinary diversion; urinary fistula; vaginal fistula; vesicovaginal fistula INTRODUCTION

Radical cystectomy with orthotopic urinary diversion is a common surgical treatment for women with muscle-invasive bladder cancer and has now become the standard of care. It is estimated that 60–70% of women would be candidates for orthotopic urinary diversion at the time of radical cystectomy.1 Neo-bladder vaginal fistula (NBVF) is an uncommon complication occurring in 0–10% of women undergoing this procedure.2–8 Several risk factors for the development of this complication have been identified such as damage to the anterior vaginal wall during dissection, the presence of poorly vascularized tissue between the neobladder and the anterior vaginal wall, suture line proximity, poor tissue vascularity after radiation therapy,9 and local cancer recurrence.4 Nevertheless, injury to the distal anterior vaginal wall at the level of the urethra appears to be the single most important contributing factor.4 Several technical modifications have been described in the past two decades to decrease the risk of this complication. These include preservation of the anterior vaginal wall,5 double vaginal cuff closure, interposition of a well vascularized flap between the neobladder and the vagina, such as the omentum, and avoidance of overlapping suture lines.2 Unfortunately, not all patients are amenable to a vaginal sparing approach and these strategies cannot prevent fistula development in all women. The surgical management of NBVF is challenging due to vaginal shortening, vaginal atrophy and proximity of the neobladder to the vaginal wall. Treatment options for NBVF are limited and often involve conversion to a continent or #

2014 Wiley Periodicals, Inc.

incontinent urinary diversion with resultant decrease in quality of life. Currently, there are no large series that describe the optimal management of this challenging entity. We present a single institution series of eight patients who underwent transvaginal NBVF repair where we describe the surgical technique and outcomes. MATERIALS AND METHODS

After obtaining institutional review board approval, a retrospective chart review was performed to identify all patients who had undergone transvaginal NBVF repair at our institution. All surgeries were performed by one of four urologists with fellowship training in Female Pelvic Medicine and Reconstructive Surgery at a high volume tertiary care center. Information about patient demographics, comorbidities, bladder cancer history, radical cystectomy and orthotopic Abbreviations: BMI, body mass index; FU, follow-up; mo, months; N, number; NBVF, neobladder vaginal fistula; TMUS, trans-obturator mid-urethral sling; yo, year old. Dirk De Ridder led the peer review process as the Associate Editor responsible for the paper. Potential conflicts of interest: Nothing to disclose. Institution where work was performed: Cleveland Clinic, 9500 Euclid Av, Cleveland, OH, 44195.  Correspondence to: Maude Carmel, M.D., UT Southwestern Medical Center, 5323 Harry Hines Blvd, Mail code 9110, Dallas, TX 75390-9110. E-mail: [email protected] Received 26 June 2014; Accepted 2 September 2014 Published online 18 October 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/nau.22687

Transvaginal Neobladder Vaginal Fistula Repair

91

Fig. 1. Surgical technique of transvaginal NBVF repair. (A) A weighted speculum and a lone-star retractor are used to help with exposure. A large caliber Foley catheter is inserted and left postoperatively. (B) A circumferential incision is made around the fistula tract, thoracic scissors are used to develop a plane between the serosa of the neobladder and the vaginal epithelium. (C) The neobladder is closed transversely using interrupted 2–0 polyglycolic acid sutures.

reconstruction details, procedural details, post-operative management and complications were recorded. The diagnosis of NBVF was made with history, physical examination, cystoscopy, and voiding radiography of the pouch. Outcomes regarding recurrence of fistula, continence status, voiding function, cancer status and other complications at the last clinic visit were recorded. Surgical Technique

The surgical technique of NBVF repair involves a multi-layer transvaginal closure. The patient is put under general anesthesia and placed in the lithotomy position. A weighted speculum and a lone-star retractor with hooks are used to help with exposure (Fig. 1A). A large caliber Foley catheter is placed inside the bladder via the urethra and left postoperatively. If the fistula tract is small, the fistula track is dilated with lachrymal sounds and urethral sounds until a small Foley catheter can be introduced in the fistula tract for traction. The anterior vaginal wall is infiltrated with a solution of lidocaine with epinephrine for hydrodissection. A circumferential incision is made around the fistula tract. Using Metzenbaum scissors and thoracic scissors, a plane is created between the serosa of the neobladder and the vaginal epithelium (Fig. 1B). The 908 angle tip of the

thoracic scissors helps with the dissection of the plane between the vaginal wall and the neobladder especially when a wide mobilization is required or when the access is limited by a tight vaginal introitus. A circumferential vaginal flap is developed until good mobilization of the neobladder and vaginal wall is obtained in order to ensure a tension-free closure. The dissection is limited distally to avoid the urinary sphincter. The fistula tract is not excised. The neobladder is closed transversely using interrupted sutures of 2–0 polyglycolic acid sutures (Fig. 1C). The integrity of the repair is tested by instilling at least 200 ml of sterile water. An additional layer of vaginal wall, Martius flap, or omental flap is interposed between the neobladder and vaginal epithelium (Fig. 2A). We recommend the use of a Martius flap for the second layer of the repair. However, in cases where the omentum was interposed between the vagina and the neobladder during the cystectomy, the omentum can be mobilized and fixed over the repair if it can be reached easily through the vaginal incision. The vaginal incision is closed with a running interlocking suture of 2–0 polyglycolic acid sutures by advancing an inverted-U vaginal flap over the repair to avoid overlapping suture lines (Fig. 2B). A moistened vaginal packing is left in place for 24 hr. The patient is observed overnight and discharged the next day after removal of the vaginal packing. The Foley catheter is

Fig. 2. Surgical technique of transvaginal NBVF repair. (A) An additional layer-Martius or omental flap—is interposed between the neobladder and the vaginal epithelium. (B) The vaginal incision is closed with a running interlocking suture using 2–0 polyglycolic acid sutures by advancing the flap over the repair to avoid overlapping suture lines.

Neurourology and Urodynamics DOI 10.1002/nau

92

Carmel et al.

removed after a minimum of 14 days. The number of days of catheter drainage was left to each surgeon’s discretion. Depending on the size of the fistula and the quality of the vaginal tissue, the surgeon could decide to keep the catheter for more than 14 days. The patient is instructed to perform irrigation of the neobladder to prevent catheter obstruction with mucus. The catheter is removed in the office after a negative cystogram and a pelvic exam with a retrograde bladder filling demonstrating no leakage. We advocate performing the retrograde filling of the neobladder before the catheter removal to have two different tests confirming the absence of leak in order to have the certainty there is no residual fistula before the catheter removal. Contrast can sometimes leak around the catheter during the cystogram, resulting in a false positive. RESULTS

From 2002 to 2012, a total of eight patients were referred to us for management of a NBVF, of which two were referred from outside institutions. Mean age at the time of repair was 57 year old (40–77). Their mean body mass index was 26.3 kg/m2 (19.3– 32.4). Five patients were active smokers and three patients were former smokers. All patients had undergone radical cystectomy and orthotopic ileal bladder (Studer pouch). Six cystectomies had been performed open, one patient had a robotic laparoscopy assisted radical cystectomy and open neobladder reconstruction and one patient had a robotic laparoscopy assisted radical cystectomy with intracorporal neobladder reconstruction. All patients had a concomitant total hysterectomy but vaginal sparing was performed in five patients. Vaginal sparing was recorded when the original surgeon mentioned vaginal sparing in the operative note and described performing an incision around the cervix to preserve the anterior vaginal wall. None of the patients received radiation therapy. One patient had received neo-adjuvant chemotherapy and two patients had received adjuvant chemotherapy. No patient had evidence of local recurrence, however, one patient had evidence of metastatic disease at the time of the fistula repair. Median time from cystectomy to NBVF repair was 310 days (190–1,533). This delay from the time of cystectomy to NBVF repair was due to a delay to diagnosis and to referral to our institution. All patients were referred to our institution more than 6 months after the cystectomy. This surgery was the first

repair for six of these patients. Two patients had already failed two previous transvaginal fistula repairs without tissue interposition at outside institutions. It had been 548 and 1,055 days since their last surgery. The characteristics of the fistulas and details of the NBVF repair surgery are summarized in Table I. A Martius flap was used in five patients, omentum (accessible from the vaginal incision) was used in one patient and a second layer of vaginal wall was used in two patients. Mean operative time was 150 min (97–217 min). Mean estimated blood loss was 74 ml (25–150 ml). All patients were discharged within 23 hr after surgery with a urethral Foley, mean caliber of 20 French (16–30). The catheter was removed following a negative cystogram and negative pelvic examination with retrograde filling of the bladder after a mean duration of 24 days (14–60). Patient #2 had an evidence of a leak at the time of the first cystogram 4 weeks after surgery which subsequently resolved after two additional weeks of catheter drainage. After a mean follow-up of 33 months (4–117), no patient had a recurrent fistula. However, five patients had severe stress incontinence secondary to a short and incompetent urethra. The incontinence was managed with a synthetic transobturator mid-urethral sling in one patient and injection of bulking agents in three patients. One patient was considering injection of urethral bulking agent at the time of her last visit. Patient #2 who underwent a mid-urethral sling was still using less than one pad per day at her last follow-up visit, 8 months after the procedure. Patient #5 required a total of four bulking agent injections in order to use one pad or less per day. Patients #6 and #7 both received one bulking agent injection in order to have 50% improvement. All these patients were last seen 3 months after their last injection. They were also contacted in the following months and all patients described themselves satisfied with their current continence status and did not desire another injection. No patient required intermittent selfcatheterization after the fistula repair or management of stress urinary incontinence. All patients were voiding spontaneously and described themselves as being satisfied with the surgical results. DISCUSSION

NBVF after radical cystectomy with orthotopic urinary diversion is an infrequent but devastating complication. These fistulas rarely heal spontaneously and therefore, a prolonged

TABLE I. Surgical Repair of Neobladder Vaginal Fistula Duration of urethral catheter FU (mo)

Location of fistula

Size of fistula (cm)

Type of flap

1 2

Vaginal apex Distal anterior vaginal wall

0.5 1

2nd layer of vaginal wall 2nd layer of vaginal wall

21 60

117 11

3 4 5

Distal left anterior vaginal wall Distal anterior vaginal wall Distal left anterior vaginal wall

1 3 1

Left Martius flap Omental flap Left Martius flap

13 26 14

97 12 6

6

Distal anterior vaginal wall

2

Left Martius flap

18

4

7

Distal right anterior vaginal wall

3

Right Martius flap

20

5

8

Distal left anterior vaginal wall

1

Left Martius flap

21

7

Patient

FU, follow-up; mo, months; TMUS, trans-obturator mid-urethral sling.

Neurourology and Urodynamics DOI 10.1002/nau

Outcomes No recurrence No recurrence SUI managed with TMUS No recurrence No recurrence No recurrence SUI managed with bulking agents injection No recurrence SUI managed with bulking agents injection No recurrence SUI managed with bulking agents injection No recurrence SUI, awaiting bulking agents injection

Transvaginal Neobladder Vaginal Fistula Repair 10

trial of catheter drainage is not recommended. Treatment of choice for this complication is still controversial and multiple surgical techniques have been described. The repair of this type of fistula repair is very complex because the wall of the neobladder is much thinner than a normal bladder wall, making it more prone to fistula formation. In addition, these patients are often left with a very short and atrophic vagina, making vaginal repair challenging. Despite several small case series reporting the effectiveness of transvaginal repair of NBVF, this is not always possible and some patients require conversion to either an incontinent or a continent catheterizable reservoir diversion.4 This conversion rate has been around 32% of all reported NBVF.4,6,9,11–20 We present the largest series of transvaginal NBVF repairs. This review spans 10 years. Two patients had their fistula repair done in 2002 and 2004. The other six patients were referred to our institution between 2008 and 2012. We cannot explain the difference in patient distribution over time with the possibility of an increasing incidence of this complication because we do not know the total number of cystectomies the referring physicians performed during that period. However, other factors could explain a higher number of fistulas being referred to us since 2008, the main one being the expansion of our department, which increased our referral and surgical volume during this period. In our opinion, transvaginal approach should be favored and this is supported by multiple factors. The location of the majority of these fistulas is on the distal anterior vaginal wall, which is the most frequent site of injury during the urethral dissection. One fistula was located at the vaginal apex. The mechanism of fistula formation at the vaginal apex is probably similar to the usual vesico-vaginal fistulas after hysterectomy. This reinforces the need for adequate closure of the vaginal cuff after the hysterectomy during cystectomy. Ali-El-Dein et al.2 recommended double vaginal cuff closure to reduce the incidence of this fistula location. NBVF whether located at the vaginal apex or the distal vaginal wall are easily accessible via a transvaginal approach. This approach avoids the morbidity of a laparotomy and possible damage to the pouch and bowel. Ali-El-Dein et al.11 reported their experience with NBVF repair and they suggested that an abdominal approach should be considered and used in patients with vaginal atrophy and large fistulas (defined as >2 cm). We disagree that the presence of vaginal atrophy warrants an abdominal approach. Most women undergoing radical cystectomy for bladder cancer have vaginal atrophy. We encourage the use of topical estrogen cream preoperatively, if not contraindicated, to promote tissue healing. However, absorption of the cream might be compromised by the severity of the incontinence but we still instruct patients to use it in order to improve tissue quality as much as possible. We believe that with adequate dissection and mobilization of the vaginal wall, adequate closure can be obtained even in the presence of vaginal atrophy. Tissue interposition during vaginal repair of NBVF is key for success and it has been well established as being effective in treating complex vesico-vaginal fistulas.21 The technique described by Ali-El-Dein11 is different than the one we have described. The author described a two layer repair by closing the excised fistula track and a second layer using fascia and fat surrounding the pouch. This second layer is not as definite and mature as in vesico-vaginal fistula repair, because the wall of the neobladder is much thinner than the native bladder wall and the pubocervical fascia is absent. Instead of creating this second layer, we advocate tissue interposition between the neobladder and the vagina to achieve a successful repair. The Neurourology and Urodynamics DOI 10.1002/nau

93

interposition of a Martius flap allows a better epithelization surface, extra vascularization of a possibly ischemic zone and better lymphatic drainage, as well as avoidance of an overlapping suture line.22 Using this technique, we successfully treated three patients with fistulas measuring more than 2 cm in diameter, two patients with recurrent fistulas after two prior vaginal repairs and six patients with vaginal atrophy. The size of the fistula, previous treatment failure, and the presence of vaginal atrophy should not be a limiting factor for successful vaginal repair. Tissue interposition was not performed in two patients but these patients were done early in our series. These patients had no vaginal atrophy and had good tissue quality, which allowed us to use a second layer of vaginal wall. The technique used was similar to the one described by Ali-El-Dein.11 However, one of these patients (patient #2) had the persistence of a small leak on a cystogram performed 4 weeks after surgery. Fortunately, this leak resolved after an additional 2 weeks of catheter drainage. We believe that this may have been prevented with the interposition of a Martius or omental flap and this is the most optimal technique for NBVF repair. The rate of incontinence after NBVF repair was high in our series. Out of the eight patients presented, five had to undergo additional procedures for the management of severe stress urinary incontinence. Patients were fully counseled prior to the NBVF repair of the high risk of incontinence after surgery and these patients were very motivated to avoid any type of stoma. Ali-El-Dein reported six patients undergoing transvaginal neobladder-vaginal fistula repair and they reported that none of their patients had incontinence after the repair.11 However, they commented that all of these fistulas were small (