Urinary Fistulae: Etiology, Diagnosis and Management

Abstract

This chapter covers the most common and clinically relevant types of urinary fistulae. Fistulae are classified by the involved organs and discussed separately. For each type of fistula, its pathogenesis, clinical presentation, recommended workup and management options are reviewed. Advantages and disadvantages of different surgical approaches are explored, however detailed instruction on surgical techniques is not provided. This chapter should prepare the clinician to recognize risk factors and symptoms of urinary fistula, initiate the appropriate diagnostic steps, and guide patients toward informed decisions between the available therapeutic options.

1. Introduction

A fistula is an abnormal passage between a hollow organ and the body surface, or between two hollow organs. Urinary fistulae are those that involve any part of the urinary tract—kidneys, ureters, bladder or urethra. Most urinary fistulae in the industrialized world are iatrogenic and are caused by surgery, but they may also be caused by malignancy, radiation, inflammation, prolonged childbirth and a variety of other pathologies. The resultant leakage of urine often leads to great psychological distress and social isolation in afflicted patients. Although some fistulae heal with only conservative management, the vast majority require definitive surgical repair in the operating room. Surgical techniques for the repair of urinary fistula have evolved over time, their merits intensely debated in the medical literature such that the optimal approach for a given clinical scenario is often difficult to ascertain. Therefore, patients should be evaluated on a case-by-case basis and selection of treatment should take into account location and etiology of the fistula, patient goals of care and functional status, and surgeon experience and preference.

2. Vesicovaginal fistula

2.1 Etiology

Vesicovaginal fistula (VVF) is an abnormal connection between the bladder and vagina that allows continuous, involuntary leakage of urine into and out of the vagina. The incidence of VVF has been estimated at 0.3 to 2% [1]. In developing countries, the predominant cause of VVF (97%) is prolonged obstructed labor [2]. It has been estimated there are between two and three million women with untreated fistula worldwide, with >95% of obstetric etiology and in developing nations [2]. The presenting fetal part presses against the pelvic floor causing extensive ischemic injury, subsequent soft tissue necrosis and fistula formation. In such cases, fistula is often only one component of devastating multifaceted pathology referred to as obstructed labor injury complex, which includes rectal atresia, incompetence of the anal sphincter, destruction of the urethra or cervix, infertility, renal failure, and nerve injury [3]. Treatment of obstetric fistulae are hindered by poor access to care in endemic regions.

In the developed world, advances in obstetric care have reduced the incidence of obstetric fistulae dramatically, and the vast majority of VVF are iatrogenic. The most common cause is gynecological surgery, principally hysterectomy [2], though pelvic malignancy and radiation are also significant causes. In a series of 348 cases of urinary fistula in the UK from 2011, over half of the cases resulted from hysterectomy [4]. Risk of post-hysterectomy fistula formation appears to be influenced considerably by surgical approach and indication. In a report on 343,771 women undergoing hysterectomy in the English National Health Service between 2000 and 2008, the lowest rate of fistula formation was observed after vaginal hysterectomy for prolapse (1 in 3861), and the highest rate was observed after radical hysterectomy for cervical cancer (1 in 87) [5]. VVF involve the posterior surface of the bladder abutting the reproductive tract. A systematic review of 15 studies found that 55.7% of VVF involved the supra-trigonal posterior bladder wall, 27.54% involved the trigone, and the remainder had varied descriptions including “juxta-cervical” [6].

2.2 Presentation and workup

Post-hysterectomy patients with VVF typically present with continuous urinary incontinence, and in the majority of cases onset is delayed between 3 days and 8 weeks postoperatively, suggesting that fistula formation is less often the result of direct operative injury and more often the result of compromised blood supply from thermal cautery or dissection and delayed tissue necrosis [4]. VVF resulting from pelvic radiation may not present for years following completion of therapy due to the slow progression of the underlying obliterative endarteritis. Careful patient history helps to differentiate VVF from other causes of incontinence such as stress, urgency and overflow incontinence. A pelvic exam should always be performed to assess for masses or other concurrent pelvic pathology. At this time a simple dye test may be performed wherein a gauze roll is inserted into the vagina and methylene blue or other dye instilled into the bladder per urethra. After several minutes the gauze is removed and examined for staining. Alternatively, cystography or CT cystography can be performed, however in both methods a small fistula may evade detection if the bladder is not distended adequately to force the dye or contrast agent through the fistula. Cystoscopy should be performed to delineate any concurrent bladder pathology and to attempt to localize the fistula, especially in relation to the ureteral orifices as ureteral reimplantation may be required if the ureter(s) are involved. The most common location of the injury is at the vaginal cuff in the supratrigonal posterior bladder wall [1]. Retrograde pyelogram performed at the time of cystoscopy or CT urogram can be used to evaluate for ureteral involvement, which is present in up to 12% of cases. Cross-sectional imaging of the pelvis is recommended if malignancy is suspected.

2.3 Management

2.3.1 Conservative management

An initial trial of catheter drainage with anticholinergic or beta-3 agonist medications is warranted in select patients with small, uncomplicated fistulae (<1 cm) and no history of pelvic radiation. Conservative management is most likely to be successful when the patient presents early on before the fistula has fully epithelialized, after which it is unlikely to heal spontaneously. However, patients should be forewarned that surgery is likely to become necessary. A meta-analysis of 124 studies reported that of 239 women managed with catheter drainage for 2–12 weeks, only 8% had spontaneous resolution of the fistula and the rest ultimately underwent surgical repair [2].

2.3.2 Early vs. late repair

Perhaps the most controversial aspect of VVF repair is the ideal timing of surgery. There is no good consensus on the definitions of early and late repair, with “early” ranging broadly from 1 to 3 months in the literature and “late” ranging from 2 to 4 months, and these intervals overlapping across studies [7]. One faction of experts advocates for early repair, drawing attention to the diminished quality of life of women awaiting surgery, while other groups advocate delaying repair for at least 6–8 weeks to allow for resolution of inflammation and edema as well as demarcation and sloughing of necrotic tissue [1, 7]. Yet other studies report no difference in outcomes. In a meta-analysis by Bodner et al., no statistically significant difference could be detected in the success rate between 241 patients undergoing repair <12 weeks from fistula occurrence to repair, and 223 patients undergoing repair >12 weeks [2]. In general, it is encouraged to repair uncomplicated postsurgical fistula as soon as possible in order to minimize patient suffering. Notable exceptions are post-obstetric and radiation fistulae, in which cases a waiting period of 3 to 6 months or longer may be appropriate.

2.3.3 Vaginal vs. abdominal approach

Another contentious aspect of VVF repair is the choice between vaginal and abdominal (open, laparoscopic or robotic) approaches. In reality, this decision is largely dictated by the surgeon’s experience, and the best approach is likely the one that he or she is most comfortable with. There are no randomized trials comparing the two approaches, but there is substantial data from multiple retrospective studies indicating that both approaches offer excellent outcomes with success rates typically exceeding 90% and approaching 100% in some series. The transvaginal approach is associated with less morbidity and shorter hospital stay and is therefore preferred in most cases, with the abdominal approach reserved for cases where there is concurrent pelvic pathology requiring additional procedures, such as ureteric involvement requiring reimplantation or diminished bladder capacity requiring augmentation cystoplasty [1, 2, 4, 7]. Traditionally the fistulous tract is excised during the repair, but some have criticized this step for causing increasing bleeding and creating an unnecessarily large defect [1, 7].

2.3.4 Interposition grafts

In most cases, an uncomplicated VVF requires only a multilayer, tension-free repair. In a series of 207 patients that underwent VVF repair, the success rate for fistulas <2 cm without tissue interposition was 99% after the first attempt [1]. When tissue quality is poor due to radiation or previous surgery, interposition grafts are used to provide a well-vascularized barrier between bladder and vaginal suture lines and reduce the risk of recurrence. Commonly omentum is used in the abdominal approach, and a Martius flap of fibrofatty tissue from the labia is used for vaginal approach. In cases of a high (proximal) VVF it may be difficult to mobilize a Martius flap adequately, and in these cases gracilis muscle or peritoneal flaps may be used instead [7]. Use of a peritoneal flap, which is readily accessible from the proximal vaginal vault, has been shown to result in comparable outcomes to use of the Martius flap without the need for a second skin incision [1]. Some studies detected no statistically significant difference in success rates based on use of interposition grafts; nonetheless, some surgeons prefer to incorporate these grafts routinely.

2.3.5 Post-obstetric fistula repair

It is worth emphasizing the difference in pathophysiology between post-obstetric fistulae, which are surrounded by large areas of de-vascularized tissue, and postsurgical fistulae which are typically surrounded by healthier tissue. The management of post-obstetric fistulae thus presents a more complex surgical challenge. First, repair of post-obstetric VVF necessitates the use of tissue flaps with adequate vascular pedicle to bring new blood supply to the surgical site in virtually all cases [3]. Second, total loss of the urethra occurs in about 5% of cases [3] and is complicated by a lack of surrounding viable tissue with which to reconstruct a neourethra. Third, urinary incontinence persists even after successful fistula closure in a significant portion of post-obstetric fistula patients [3] and may be due to damage to the bladder neck, urinary sphincter or the pudendal branches that provide its innervation. Fourth, rectovaginal fistula may occur concurrently with VVF, and the rate of successful combined repair has been reported to be equivalent to that of isolated rectovaginal fistula (78%) [3], which is comparatively lower than that of VVF.

2.3.6 Postoperative considerations

Postoperatively, continuous urinary drainage with suprapubic and/or urethral catheterization is paramount to prevent distention of the bladder and tension and breakdown of suture lines. A voiding cystogram is recommended at 2–3 weeks to rule out extravasation or leak, at this time contrast and/or methylene blue is instilled into the catheter and a tampon or vaginal packing is examined for staining to confirm closure of the fistula.

3. Ureterovaginal fistulae

3.1 Etiology

Ureterovaginal fistula (UVF), like VVF, most often occurs as a result of injury during gynecologic surgery, though it may also arise after vaginal delivery, infertility treatment involving oocyte retrieval, pelvic radiation and pessary use [8]. Hysterectomy for a benign indication is most often the offending procedure, with the incidence estimated at 1 UVF per 250 hysterectomies [8]. Furthermore, total abdominal, total laparoscopic and robotic hysterectomy are more likely to result in UVF than supracervical hysterectomy [8].

3.2 Presentation and workup

Patients present with a chief complaint of continuous urinary leakage per vagina. The normal urge to void may differentiate UVF from VVF, as in cases of UVF the uninjured ureter continues to fill the bladder normally, whereas in VVF the bladder is continuously drained via the fistula; however, it should be remembered that VVF and UVF may occur concurrently. The recommended initial workup is a dual-dye tampon test, which is inexpensive and may be performed in the outpatient setting. Patients are prescribed oral phenazopyridine to take before the visit. In the office, a gauze roll is placed in the vagina and methylene blue dye is instilled into the bladder via a Foley catheter. After approximately 30 minutes the gauze roll is removed and examined for presence of orange staining (indicates UVF) and/or blue staining (indicates VVF).

3.3 Management

3.3.1 Primary management

Historically, UVF were managed with ureteroneocystostomy (ureteral reimplantation), urinary diversion with percutaneous nephrostomy, or even nephrectomy [8, 9, 10]. However, more recent studies have reported that endoscopic ureteral stenting results in high cure rates for post-hysterectomy UVF and should be considered first-line therapy [8, 9]. Cystoscopy with retrograde pyelography is first used to confirm the diagnosis and localize the fistula, as well as to rule out concurrent VVF. Retrograde stent placement may then be attempted at the same time as retrograde pyelogram. If retrograde stent placement is unsuccessful, antegrade stent placement may be attempted by an interventional radiologist. Rates of successful stent placement vary widely in the literature from 5–92% in different series [8]. The optimal length of ureteral stenting is unclear, with studies reporting protocols from 4 to 8 weeks [9, 10] to 3 months [8]. Rates of resolution of UVF after successful stent placement have been reported from 32% [11] to 100% [8, 9]. A meta-analysis of 17 retrospective studies found that for 285 patients undergoing endoscopic stent placement for UVF, the overall pooled success rate was 32%. However, when stratified by time from fistula diagnosis to intervention, it was shown that success rates were 95% if stenting was performed within 2 weeks, 63% within 6 weeks and 20% after 6 weeks [11]. This suggests that the success of endoscopic stent placement as the sole treatment modality for UVF decreases with time from diagnosis, and that stent placement could potentially obviate the need for more invasive surgery if performed early enough. In contrast, success rate of surgical intervention approached 100% regardless of whether or not it was performed within 6 weeks from diagnosis [11]. Management with urinary diversion with percutaneous nephrostomy is not effective [10]. Several criteria for primary surgical intervention with ureteral reimplantation have been suggested including concurrent VVF, history of ureteral injury with repair at the time of index surgery, and extremely long delay from index surgery to diagnosis [8].

3.3.2 Post-operative considerations

Ureteral strictures have been reported in 6–38% of UVF managed conservatively and may have a delayed presentation. For this reason, annual follow-up to rule out stricture formation and urinary obstruction is recommended. Renal ultrasound, CT urogram, Mag-3 Lasix renogram and retrograde pyelogram may be used as surveillance studies.

4. Urethrovaginal fistula

4.1 Etiology

Urethrovaginal fistula is most often caused by operations of the vagina and urethra including urethral diverticulectomy, urethral sling procedures and anterior colporrhaphy. Excision of urethral diverticulum is an uncommon cause of urinary fistula overall, but the most common cause of urethrovaginal fistulae, accounting for 39.5% of all urethrovaginal fistulae [4]. Other causes include traumatic pelvic fractures, obstetric trauma, vaginal and urethral malignancies, pelvic radiation, and urethral erosion from chronic catheterization or pessary use [12].

4.2 Presentation and workup

Presentation depends on the size and location of the fistula. Large or proximal urethrovaginal fistulae may cause stress incontinence or continuous incontinence akin to VVF. If the fistula is distal to the external urinary sphincter, the patient may be asymptomatic, or may complain of “vaginal voiding” wherein urine which leaks into the vagina is emptied as soon as the patient stands from a seated position. Dyspareunia and urinary tract infection (UTI) may also occur. Physical exam and cystoscopy should be performed but it may be difficult to demonstrate smaller fistulas. Voiding cystourethrogram (VCUG) is the most useful radiographic study; importantly, voiding-phase imagery must be obtained in order to diagnose urethrovaginal fistula in a patient with a competent bladder neck.

4.3 Management

Simple excision of the fistula with layered closure has a high recurrence rate, and for this reason interposition grafts are often used to reinforce the repair. A Martius labial fat pad may be tunneled beneath the labia minora and into the vaginal lumen. Use of other grafts including rectus muscle flaps, vaginal flaps, autologous fascial and synthetic grafts have all been described. Postoperatively, urinary diversion with suprapubic catheterization is more appropriate than urethral catheterization and is recommended along with a period of abstinence from sexual intercourse for 6 weeks [12]. If the initial insult causes significant damage to the bladder neck or external sphincter, fistula repair alone may fail to resolve incontinence, and in these cases a midurethral or bladder neck sling may be appropriate [12].

5. Vesicouterine fistula

5.1 Etiology

Vesicouterine fistula (VUF) is the least common form of urinary fistula, accounting for 1–4% [13, 14]. Over 75% occur after cesarean delivery [14], and there is an increase in both the absolute number of these fistulas as well as their relative proportion among urinary fistulae due to the increasing number of cesarean deliveries [15]. Other causes include prolonged labor, vaginal delivery after previous cesarean section, placenta percreta, anterior colporrhaphy, endometrial ablation, and excision or embolization of uterine fibroma [15]. Many VUF patients experience onset of symptoms with resumption of menses after childbirth, which suggests the attendant increase in circulating estrogen and resultant changes in endometrium may play a role in the pathophysiology of VUF formation [16]. In a review of 795 cases of VUF, a subset of women treated with hormonal suppression with oral contraceptives or LHRH analogs experienced resolution of their fistulae [16]. The authors speculated that hormonal suppression may be a possible treatment, but their sample size was extremely small.

5.2 Presentation and workup

Menouria is the hallmark of VUF, differentiated from hematuria as the cyclic expulsion of menstrual blood, endometrial tissue and/or blood clots with the urine. Patients may present with any combination of menouria, amenorrhea, and urinary incontinence, though not all of these features may be present simultaneously [15]. Notably, menouria and amenorrhea without urinary incontinence is referred to as Youssef Syndrome. In contrast to other urinary fistula, the absence of urinary incontinence as a reliable symptom makes the diagnosis more difficult. One proposed continence mechanism in these patients is that the uterine isthmus acts as a sphincter that restricts the passage of urine; additionally, studies have shown that the intrauterine pressure is lower than intravesical pressure for the majority of the menstrual cycle [15]. In most cases the diagnosis can be made swiftly on cystoscopy, while cystography, hysterography, and ultrasound are also often obtained.

5.3 Management

The rate of VUF resolution with bladder drainage has been estimated at 5% or less [14, 16]. Thus, the standard treatment is surgical closure using an interposition graft (omentum or peritoneum), with or without hysterectomy. While laparoscopic, transvaginal and robotic approaches have been described, the best outcomes have been reported with an open abdominal approach [15]. Given its relatively low prevalence there is a lack of studies on VUF and its impact on future fertility. Among the studies that have been published, many did not report the proportion of affected women who attempted another pregnancy. However, multiple studies have reported successful full-term pregnancies after vesicouterine fistula repair, suggesting that fertility can be recovered after treatment [15].

6. Colovesical fistula

6.1 Etiology

Colovesical fistula is an abnormal communication between the bladder and large intestine, usually sigmoid colon [17], that occurs as a sequela of inflammatory or neoplastic disease. The most common cause is diverticulitis, which accounts for 65–79% of cases [17, 18]. The relative risk of developing a colovesical fistula in the presence of diverticular disease is estimated at 2–4% [17, 18]. The next most common cause is malignancy (both colon and bladder), which accounts for 10–20% [17, 18]. Other causes include inflammatory bowel disease, radiation, appendicitis and trauma. While inflammatory bowel disease is often cited as a cause of CVF, it should be specified that it is Crohn’s disease rather than ulcerative colitis that has a greater tendency to fistulize, as the entire thickness of the bowel wall is affected rather than just the mucosa. Moreover, Crohn’s more often results in ileovesical fistula than CVF.

6.2 Presentation

The presenting symptoms are typically urinary, with the findings of pneumaturia (air bubbles in the urine in the absence of recent transurethral instrumentation or catheterization) and fecaluria being pathognomonic. Pneumaturia is present in 50–70% of cases [13, 17] and should be considered indicative of CVF until proven otherwise. While pneumaturia is the most common symptom, it may not be distressing to the patient and is infrequently the chief complaint. Rather, it is typically fecaluria or less-specific symptoms such as cystitis, dysuria, urinary frequency, or hematuria which drive a patient to seek medical attention. GI symptoms and leakage of urine per rectum are rare. Patients with persistent UTIs, especially those with inflammatory bowel disease, should be suspected of having a CVF and worked up appropriately [13].

6.3 Workup

6.3.1 Oral tests

There is considerable debate in the literature regarding the ideal initial workup for CVF and it should be emphasized that there is no consensus on a ‘gold standard.’ The classic “poppy seed test” involves oral intake of poppy seeds, which remain intact through the GI tract and may appear in the urine within 48 hours. Chen et al. [18] reported that the detection rate of the poppy seed test was 100% for surgically-confirmed fistulae, superior to CT scan. Variations of this test use activated charcoal or a variety of orally-ingested dyes. While inexpensive, these oral tests do not provide any insight into the location of the fistula or surrounding structures.

6.3.2 Contrast studies

Cystography and barium enema are likely to be abnormal in the presence of a CVF, however, the reported rates of unequivocal fistula demonstration are low at 20–34% for cystography and 30–35% for barium enema [13, 18]. Intravenous pyelogram is not as reliable for definitive diagnosis likely due to the higher intracavitary pressures generated during cystography and barium enema [13], and this test has largely fallen out of favor with the advent of newer cross-sectional imaging modalities. The Bourne Test refers to radiography of a centrifuged urine sample immediately after non-diagnostic barium enema and has been shown to increase the diagnostic yield of barium enema [17].

6.3.3 Cross-sectional imaging

Since the proliferation of accessible CT imaging, its diagnostic accuracy for detecting colovesical fistulae on CT has been reported at over 90%, higher than cystography or barium enema, and many have suggested this should be the initial diagnostic tool in cases of suspected colovesical fistula [17, 18]. It is recommended to perform CT after administration of oral or rectal contrast but without administration of IV contrast, as identification of enteric contrast media in the bladder provides a conclusive diagnosis. Not only is CT highly sensitive for fistula, but also it is able to rule out adjacent pathology including colonic strictures, masses, abscesses, and to localize the lesion for surgical planning. Findings on CT scan suggestive of fistula include air within the bladder lumen, as well as a thickened segment of colon in intimate contact with the bladder wall at the location of the fistula (see Figure 1). False positives may result after cystoscopy and in the presence of a gas-forming UTI or indwelling catheter; for this reason cystoscopy should be postponed until after CT. MRI has excellent sensitivity and specificity approaching 100% and is a suitable alternative to CT scan as the choice initial study [18].

Figure 1.

CT scan showing a colovesical fistula from diverticulitis.

6.3.4 Endoscopy

Cystoscopy is likely to be abnormal in the presence of a CVF, though findings may be nonspecific. It may reveal pieces of undigested food or feces in the bladder, which is diagnostic of CVF. Focal erythema and edema termed a herald patch may also indicate colovesical fistula, especially in the context of suspicious patient history and location on the bladder dome or left side nearest the sigmoid colon; however, similar lesions can be seen with intrinsic bladder pathology such as urothelial malignancy. Importantly, cystoscopy may fail to demonstrate fistula in 54–65% of cases [18] and therefore when the clinician’s index of suspicion is high, the diagnosis of fistula should not be excluded on the basis of a negative cystoscopy alone. Colonoscopy has a reported detection rate of 8.5–55% for CVF [18] and is not particularly helpful in making a diagnosis. However, as 10–20% of CVF are due to bladder and bowel malignancy [17, 18], both cystoscopy and colonoscopy are considered mandatory parts of the workup regardless.

6.4 Management

Nonoperative management is an option in minimally symptomatic patients, particularly those with Crohn’s. In these cases, a trial of bowel rest with parenteral nutrition, steroids or other immunomodulators along with catheter drainage is sometimes sufficient to allow for resolution of the fistula. However, due to reports of deconditioning, sepsis and increased mortality among patients with conservatively managed fistulae, this option is usually reserved for those unfit for surgery.

The recommended management in most cases involves resection of the fistula-containing segment of bowel and reconstruction of contiguous bowel anatomy. Many surgical approaches have been described including single-stage resection with primary anastomosis, as well as procedures in two stages (proximal bowel diversion and resection, followed by colostomy closure) and three stages (diversion, resection, and later closure). Advocates of the staged approach have argued that it is safest and that single-stage procedures should be reserved for patients with small fistulae of inflammatory etiology and good preoperative functional status [13]. Others have reported data suggesting that staging repairs does not reduce the risk of complications and therefore most CVF should be treated in one stage [17, 18]. An exception to either approach involves fistulae that occur after pelvic radiation, which are both more difficult to remove and more likely to recur. In these cases, a palliative bowel diversion (without resection of the fistulized bowel) may be preferable to attempting definitive repair and risking fistula recurrence and the need for repeated surgeries.

7. Ureteroarterial fistula

7.1 Etiology

Ureteroarterial fistula (UAF) is a rare condition with only a few hundred cases reported in the literature, however it is becoming more common with the increasing life expectancy of patients with pelvic malignancy, many of whom are managed with chronic ureteral stents [19, 20]. Incidence of UAF has been estimated at 2–5 patients per year [20]. Mortality has been reported to range from 7 to 23% [19]. The most common predisposing factors are pelvic surgery, radiation and ureteral stent placement which have been reported in 89–100%, 43–74% and 67–84% of UAF patients, respectively [19, 20], and all play a role in the proposed pathogenesis of UAF. Surgery and radiation causes fibrosis and fixation of the ureter to the iliac artery or aorta. Radiation causes ischemic injury to the arterial wall by damaging the vasa vasorum. Lastly, ureteral pressure necrosis results from arterial pulsation against an indwelling ureteral stent.

7.2 Presentation and workup

The main presenting symptom is hematuria, which may range from mild and intermittent to life-threatening hemorrhagic shock. Hematuria often occurs at the time of ureteral stent exchange. Flank pain has been reported in over 50% of cases [19] and may be due to obstructing ureteral blood clots. Cystoscopy may localize bleeding to one of the ureteral orifices; pulsatile bleeding on cystoscopy is indicative of UAF. Pelvic angiography is the best modality for the diagnosis and localization of UAF. Therefore, if hematuria occurs in the presence of the triad of radiation, pelvic surgery and ureteral instrumentation, the physician should consider pelvic angiography to rule out UAF [19, 20]. In a native ureter, the most common site of UAF is where the ureter crosses the bifurcation of the common iliac artery or just distal to this at proximal external iliac artery [19]. In patients with urinary diversion e.g., ileal conduit, the site of fistula may be the proximal common iliac artery or the distal abdominal aorta [19].

7.3 Management

Open surgical repair of UAF is difficult because patients have a history of pelvic surgery, radiation or both; moreover, they may be hemodynamically unstable at presentation. For these reasons, the endovascular approach has become the treatment of choice for patients with UAF regardless of age or comorbidities [19, 20]. The most common treatment is stent-graft exclusion of the fistula, with or without embolization of the affected artery and/or placement of a bypass graft. Ureteral management includes nephrostomy tube drainage, maintenance of an indwelling ureteral stent, or nephroureterectomy in the case of an atrophic or nonfunctioning ipsilateral kidney.

Lower extremity ischemia is common post-procedurally and occurs in over 50% of patients, but is somewhat less common following endovascular repair (50%) than open repair (67%) [19, 20]. Sequalae of lower extremity ischemia include skin ulcers, DVT, chronic leg pain, femoral head osteonecrosis and limb amputation. Anticoagulation has not been shown to prevent lower extremity morbidity in the setting of acute treatment of UAF. Stent-graft infection is another rare but potentially serious complication, which may be caused by either commensal skin organisms such as Staph. Aureus or by urinary pathogens such as E. Coli in the case of a UAF.

8. Conclusions

Urinary fistulae result from iatrogenic and pathologic processes leading to tissue inflammation or ischemia in or near the urinary tract. The physical and psychosocial impacts of various fistulae are often devastating for patients. Thorough history, physical exam and diagnostic testing will lead to the correct diagnosis for the patient. Conservative management may be attempted, but most patients end up undergoing definitive surgical repair. The surgical approach is tailored to each patient’s anatomy and goals of care, as well as surgeon experience.

Conflicts of interest

The authors declare no conflicts of interest.