Covered metallic stents for the palliation of colovesical fistula

Abstract

Colovesical fistula is a distressing condition that is usually managed surgically. For some patients in whom surgery is not feasible, covered colonic stents offer palliation. We present two challenging cases with contrasting outcomes. The first case is a colovesical fistula secondary to malignancy with a successful outcome after stenting and the second a complex diverticular fistula with a poor outcome. From our limited experience, it is a useful technique but careful patient selection is essential to its safe application. There is little published experience of the use of these stents for colovesical fistula.

Colovesical fistula is a distressing condition leading to recurrent urinary tract infections and passage of faeces in the urine. This abnormal communication between the urinary bladder and the bowel is most commonly caused by diverticular disease. Of patients with sigmoid diverticulosis, 6–12% present with symptoms related to fistulation.^1,2 Other causes include malignancy, Crohn’s disease, radiotherapy and iatrogenic injury.³ Colovesical fistulae are more common in males, with a male-to-female ratio of 3:1.

Surgery in the form of resection of the diseased colon or a diversion stoma remains the mainstay of management for colovesical fistula in patients fit for anaesthesia. However, patients with malignant fistulae usually present in the advanced stages of their cancer when radical surgery or even a defunctioning stoma may not be appropriate. Stenting offers an alternative for these patients either as palliation or as a stop-gap measure while optimising the patient for surgery. Covered, self-expanding, metallic stents have been used to treat such patients in only a limited number of cases with varied outcomes.^4–6

The use of self-expanding, metallic colonic stents to relieve acute colonic obstruction has increased in recent times and is more widely accepted as standard practice.^7,8 Current national randomised trials are in progress to establish whether stents in this role reduce morbidity and mortality of emergency colonic surgery.⁹ Though the technique for deploying covered colonic stents is similar to that for their uncovered counterparts, experience of using stents to treat colonic fistula remains limited and published literature regarding their outcomes is rather sparse. We describe two cases with contrasting outcomes following stenting for colovesical fistula.

Case histories

Patient 1 – malignant colovesical fistula

A 62-year-old man with metastatic locally advanced adeno-carcinoma of the sigmoid colon presented with recurrent urinary tract infections and faecaluria. At initial presentation, the patient declined surgery in view of his metastatic disease and his frailty. He had palliative radiotherapy. Computed tomography (CT) scan showed progression of his cancer (Fig. 1) but no definite evidence of fistulation. He was septic with pyrexia, tachycardia and raised inflammatory markers. Intravenous antibiotics were commenced with a good response. A water-soluble contrast enema clearly demonstrated the fistula (Fig. 2).

Figure 1.

CT scan showing locally advanced rectosigmoid tumour invading the bladder.

Figure 2.

Water soluble contrast enema demonstrating colovesical fistula (arrow).

Under image guidance (Fig. 3), a ComVi^® covered colonic stent (10 cm long, 30 mm diameter; Niti S Comvi^®; TaeWong Medical, Korea) was deployed to cover the site of the previously identified fistula (Fig. 4). There were no immediate complications and the gross faecaluria immediately resolved. The patient was discharged home 3 days later. At 4-week follow-up, urine culture was negative for coliforms. Six months after stenting, he continued to enjoy a reasonable quality of life in spite of advanced malignancy.

Figure 3.

Guide-wire passed through tumour prior to stenting. Image guidance confirms the stent will straddle the fistula once deployed (arrow).

Figure 4.

Covered stent deployed successfully.

Patient 2 – benign colovesical fistula

A 74-year-old woman presented with recurrent urinary tract infections, and progressive left groin pain and swelling. She also had pneumaturia for 3 years prior to her presentation. Her medical history was quite complicated including type II diabetes, hypertension, myocardial infarction and atrial fibrillation for which she was on warfarin. Her surgical history included a hysterectomy for endometrial cancer, open cholecystectomy and incisional hernia repair.

Physical findings were a body mass index of 39 kg/m², left lower quadrant abdominal and groin tenderness. During the course of a protracted hospital stay, the groin swelling evolved into a faecal fistula. Urine culture yielded a heavy growth of coliforms. CT scan showed complicated diverticular disease with evidence of colovesical and colocutaneous fistula to the left groin. Flexible sigmoidoscopy showed extensive diverticulosis but no evident fistulous opening.

Surgery was considered initially but was considered too high risk because of her significant co-morbidities. She had a ComVi^® stent (12 cm long, 30 mm diameter Niti-S Comvi^® covered colonic stent; TaeWong Medical, Korea) inserted with reasonable expansion though the procedure was technically very difficult. She had a protracted after stent period in which she had recurrent episodes of diverticulitis. She subsequently perforated her colon 3 months following stent insertion and succumbed to peritonitis. Post-mortem examination confirmed the perforation was stent-related.

Discussion

The Niti-S ComVi^® stent is flexible and retains shape-memory when deployed. It was originally designed for palliation of advanced obstructing colon cancers.¹⁰ It is described as a covered stent because polytetrafluoroethylene is inserted between the stent wires to prevent tumour in-growth. When used in the treatment of colovesical fistula, covered stents are thought to function in a similar manner creating a new barrier between the bowel lumen and the fistula tract there-by alleviating the symptoms associated with fistulation.

Technically, the first case was straightforward due to the presence of a clearly defined stricture and fistula on the procedural imaging, which aided stent positioning. Case two was technically very difficult due to patient immobility and difficulties in demonstrating the exact site of the fistula and correlating procedural appearances with the prior imaging. The contrasting outcomes between these cases could be attributed to the nature of the strictures, the underlying pathology and patient co-morbidities. Diverticular strictures tend to be longer than malignant ones making stenting technically more difficult and decreasing the long-term survival of the stent. Covered stents are thought to help close fistulae by compressing the tissues around them.¹¹ They may, therefore, increase the risk of perforation considering the significant pericolic inflammation that already exists. In contrast, malignant fistulae occur due to tumour invasion and there is not usually festering sepsis around these fistulae.

Conclusions

The use of a covered stent is a valuable option in the palliation of colovesical fistulae secondary to malignancy in cases where surgery is not feasible or appropriate. This includes locally advanced rectal cancer and inoperable metastases. The outcome in patients with inflammatory conditions such as diverticular disease is at present equivocal and complication rates anecdotally appear to be higher than the malignant group. Covered stents could improve quality of life in this select group of patients while avoiding invasive surgical treatment.