Endoscopic Therapy for the Treatment of Inflammatory Bowel Disease-Related Fistula, Sinus, and Abscess

List of Abbreviations


Crohn’s disease


The Crohn’s disease activity index


Computed tomography


Endoscopic balloon dilation


Enterocutaneous fistula


Enteroenteric fistula


Examination under anesthesia


Endoscopic ultrasound




Inflammatory bowel disease




Ileal pouch-anal anastomosis


Interquartile range


Magnetic resonance tomography


Needle knife


Needle-knife fistulotomy


Needle-knife sinusotomy


Operating room


Over-the-scope clip


Procedure-associated complications


Pouch-vaginal fistula


Quality of life


Rectal-vaginal fistula


Self-expandable metal stent


Tumor necrosis factor


Through-the-scope clip


Ulcerative colitis


The author is supported by the Ed and Joey Story Endowed Chair.


Fistula and abscess can result from underlying Crohn’s disease (CD) or ulcerative colitis (UC) or from CD- or UC-related surgery. Fistula and abscess represent the most advanced phenotypes of CD, as compared with inflammatory or fibrostenotic counterparts of the disease, which has a great impact on the disease outcome and patient’s general wellbeing and quality of life. In a majority of patients with CD, fistula results from persistent mucosal and transmural inflammation and stricture, which is evidenced by longitudinal studies and clinical coexistence of inflammation, stricture, and fistula in a given patient. Medical therapy, including the use of antibiotics, immunomodulators, antitumor necrosis factor (TNF), anti-integrins, or anti-interleukin (IL)12/IL23 agents, has been showed to be effective in reducing fistula drainage, particularly in the presence of concurrent bowel inflammation. However, it appears that the efficacy of medical therapy is limited to perianal fistulas and to a much less extent, enterocutaneous fistula (ECF). It is generally believed that the medical therapy has a limited role in the treatment of hollow organ-to-hollow organ fistulae, such as rectal bladder fistula, rectal-vaginal fistula (RVF), pouch-vaginal fistula (PVF), enteroenteric fistula (EEF), and gastrocolonic fistula from CD. Currently, the standard treatment of those hollow organ-to-hollow organ fistulae is surgical intervention. The main disadvantages of surgical treatment are its invasiveness and complexity, postoperative infectious and technique-associated complications, limited efficacy, and frequent postoperative disease recurrence. Surgery in patients with inflammatory bowel disease (IBD) carries a higher risk for the development of stricture, anastomotic leak, sinus, fistula, and abscess, than that for non-IBD colorectal diseases. In addition, the distinction between disease-associated complications and surgery-related complications can be difficulty, as the two situations can present with similar clinical, endoscopic, and radiographic features.

For the last decade, endoscopic therapy has evolved to a valid option for the treatment of IBD. Endoscopic balloon dilation (EBD) has become a part of routine clinical practice in the treatment of IBD-associated or IBD surgery–associated strictures. The role of endoscopic therapy for CD fistula and IBD surgery-related fistula has also been explored, despite the fact that the application of endoscopic therapy to IBD lags behind that for diseases of the upper gastrointestinal (GI)_track and pancreaticobiliary system. The gap in endoscopic therapy in IBD may attribute to the lack of deep knowledge and interest in disease process and management of IBD from therapeutic endoscopists and/or the lack of technical skills in advanced endoscopy from the IBD specialists. There is a growing need for endoscopic management of IBD, as the latter has become an integral part of multidisciplinary approach to this complex disorder in major tertiary-care centers ( Table 16.1 ).

Table 16.1

Indications and Precautions of Endoscopic Treatment of IBD-Associated Fistula, Sinus, and Abscess

Modality Indications Precautions
Endoscopic fistulotomy Short, shallow, single track fistula Avoid : internal and external anal sphincter area; anterior wall of distal bowel; long, deep fistula
Fibrin glue and plug Feasible Efficacy and safety remain to be proven
Topical therapy with stem cell or stromal cells Theoretically feasible Efficacy and safety remain to be proven
Seton placement Feasible and effective in simple, perianal fistula Not feasible: complex or branched fistula
Pigtail stent placement Feasible in intraabdominal or pelvic abscess with an orifice in the bowel side. Particularly useful in treating abscess which is not feasibly treated by drainage via interventional radiology due to overlaying bowel Guidance with endosonography or other imaging modalities preferred
Through-the-scope clip Acute bowel perforation; small, fresh fistula Not effective: long and/or fibrotic fistula
Over-the-scope clip Acute or chronic bowel perforation/leak; enterocutaneous fistula, esp. related to surgical anastomotic leak Not effective: large, chronic fistula, especially related to Crohn’s disease
Endoscopic suturing Acute or chronic bowel perforation or defect Questionable efficacy: CD-related fistula
Endoscopic self-expandable metal stent Acute anastomotic leak Questionable efficacy: CD-related fistula, vaginal fistula

Principles of Endosocpic Treatment of Fistula, Sinus, and Abscess

General principles of therapeutic endoscopy in IBD are discussed in a separate chapter ( Chapter 10 ). Like any management plan for any disease, the successful treatment relies on the knowledge of principles, the development of techniques by the implementation of the principles and the proper use of tools, devices, and equipment. Therefore, the doctrine should be “principle first, technique second, and tools, device, and equipment last.”

We should have a solid knowledge in pathogenesis, disease process, and natural history of IBD. This author has summarized the natural history of CD as follows: “no inflammation, no stricture; no stricture, no fistula; and no fistula, no abscess.” This leads to proper approaches to various conditions in IBD. We should understand the advantages and disadvantages of medical versus endoscopic versus surgical therapies. The goals of the current medical treatment for CD have been the control inflammation and prevention of development of stricture and subsequent fistula and abscess and postoperative disease recurrence. The main goals of surgical treatment in CD are the relief of mechanical obstruction, repair or removal of fistula, and drainage of abscess. Generally speaking, medical therapy for IBD is less effective and also less invasive than surgical therapy, and vice versa. The “happy median” is the endoscopic therapy.

Endoscopic therapy is less invasive than surgery and may be more effective than medical therapy, especially for IBD-associated stricture. The endoscopic therapy plays a role in bridging and facilitating the efficacy medical and surgical therapy. As endoscopists, we should get the first-hand knowledge of mechanical complication of IBD, by observing surgical procedure in operating room (OR) and participation in the examination of histopathology of bowel resection specimens with our pathologists and surgeons. We should understand indications, contraindications, and limitations of endoscopic therapy for IBD. When we deliver endoscopic therapy for IBD-associated fistula and abscess, we should also address concurrent bowel inflammation with proper medical therapy and concurrent strictures with EBD or endoscopic stricturotomy. Concurrent intraabdominal or perianal abscess may be treated with radiographic, endoscopic, or surgical drainage.

Fistula and abscess often result from underlying CD or UC or from IBD-associated surgery. Before the initiation of medical, endoscopic, or surgical therapy, it is important to delineate the anatomy and nature of fistula and abscess. A combined assessment of clinical, endoscopic, radiographic, histopathological features is required. It is important to evaluate the location, number, length, complexity of fistula; the number, size, and associated abscess; and concurrent bowel inflammation and stricture. We should carefully review previous operative report(s). The interaction and collaboration with GI pathologist, GI radiologists, and surgeons is critical, especially for complex cases.

The main principles for the endoscopic management for fistula are to (1) open up the fistula track, that is, fistulotomy, whenever possible; (2) close the primary opening of the fistula and keep the secondary opening; (3) fill up or inject the fistula track with various agents, such as stem cells, fibrin glue, or plug, if possible. We need to make sure that the closure of the primary opening of a fistula should precede the closure of the secondary opening or at least the primary and secondary openings are closed the same time. If the sequence is reversed, that is, the closure of the secondary opening prior to the closure of the primary opening, the patient may develop abscess or branching of a simple fistula track into complex tracks. The outcomes can be measured by a complete or partial closure of the primary and/or secondary openings, absence of fistula drainage, complete scarring off of fistula track, and prevention from the formation of abscess or branched fistula.

Those principles have helped to develop and implement proper techniques of endoscopic therapy for IBD-associated fistula, sinus, and abscess, and for IBD-surgery–associated fistula, suture line, or anastomotic leaks or abscess.

Endoscopic Fistulotomy

Endoscopic fistulotomy may be defined as the one performed with the use of an endoscope or performed in the endoscopy unit with an endoscopic device. For any fistula, complete fistulotomy provides the best outcome as compared with other means of fistula treatment. However, not all the fistulae are amenable to endoscopic fistulotomy. Endoscopic fistulotomy may be attempted in following conditions: (1) suture line or anastomotic leak in the distal bowel, leading to bowel-bowel fistula; and (2) the track should be superficial and short. The short fistula in the distal bowel is often associated with prior surgery in the area, such as ileal pouch-anal anastomosis (IPAA), coloanal anastomosis, and colorectal anastomosis. It is important to measure the length of fistula track and thickness of overlaying bowel wall, by procedural abdominal imaging as well as a soft-tip guidewire via endoscopy. This author has gained some experience in endoscopic fistulotomy in patients with distal bowel-to-distal bowel fistula, including ileal pouch-to-pouch fistula. Deeper fistula track which undergoes endoscopic fistulotomy may need the deployment of endoclips along the electroincised edges of the fistula track to prevent the reformation of the track ( Fig. 16.1 ).

Figure 16.1

Endoscopic fistulotomy at ileocolonic anastomosis in Crohn’s disease. (A) Fistula opening at the distal neoterminal ileum; (B) fistula track ( green dotted line ) from the neoterminal ileum to proximal colon detected by a guidewire; (C) needle-knife fistulotomy; (D) clipping of edges of the incised bowel wall to prevent reformation of the fistula.

Shen B. Gastrointest Endosc 2017; 85 :1133–43.

Endoscopic treatment of perianal fistula may be more complicated. Complete fistulotomy can be safely performed in patients with superficial perianal fistula outside the external sphincter, that is, extrasphincteric fistula ( Fig. 16.2 ). Partial fistulotomy may be performed in patients with a long perianal fistula with the secondary orifice located way beyond the edge of external anal sphincter ( Fig. 16.3 ). Fistulotomy with endoscopic needle knife or isolated-tip knife with proper topical anesthesia has the obvious advantage over surgical scalpel knife with minimum bleeding of the former approach.

Figure 16.2

Endoscopic fistulotomy of an extrasphincteric perianal fistula in Crohn’s disease. (A) A newly developed fistula track ( green dotted line ) despite seton treatment; (B) postprocedure view of needle-knife fistulotomy.

Shen B. Gastrointest Endosc . 2017; 85 :1133–43.

Figure 16.3

Treatment of perianal abscess/fistula in endoscopy suite: (A) Endoscopic needle knife ( red arrow ) was used to drain the abscess ( green arrow ) and deliver fistulotomy; (B) postprocedure view.

Shen B. Gastrointest Endosc . 2017; 85 :1133–43.

Endoscopic fistulotomy can be used in combination with other endoscopic treatment modalities, such as closure of suture line leak ( Fig. 16.4 ), stricture dilation, and needle-knife stricturotomy.

Figure 16.4

Endoscopic fistulotomy and closure of the tip of the J pouch leak. Seton placement in perianal fistula. (A) A fistula track ( green dotted line ) near the tip of the “J” leak; (B) complete endoscopic needle-knife fistulotomy with placements of endoclips; (C) the tip of the “J” leak detected with a guidewire; (D) deployment of an over-the-scope clip to close the leak.

Endoscopic Sinusotomy

Anastomotic leak is common in surgery for CD and UC. The leak can be acute and chronic. Acute anastomotic leak can cause abdominal and pelvic sepsis. Acute leak can be managed medically, endoscopically, and surgically (see below). Chronic anastomotic leak can form a sinus cavity. The most common location of sinus is the presacral area, which has been reported often in IPAA surgery. Pouch sinus is defined as a blind tract resulting from chronic pouch-anal anastomotic leak or suture/staple line leak. It occurs in 2.8%–8% of IPAA patients. Most patients required surgical intervention with incision and drainage, debridement, unroofing, occlusive treatment with fibrin glue, pouch diversion, or surgically redo pouch.

We developed novel endoscopic needle-knife sinusotomy (NKSi) to treat the pouch sinus with remarkable outcome. We reported of the first cohort of 65 patients, with the mean depth of the pouch sinus was 4.4 ± 1.8 cm. Twenty patients (30.8%) had complex pouch sinuses. After a median of 2.0 needle-knife therapies (interquartile range [IQR]: 1.5–3.5) during a follow-up period of 1.1 years (IQR: 0.4–2.8), 28 patients (43.1%) with pouch sinus had complete healing, 27 (41.5%) had a partial healing. The risk of complication was low. The principle of endoscopic NKSi is the incision of the wall of posterior pouch body next to the sinus, making the latter become a part of pouch and eventual epithelialization of sinus cavity. The key of success of the procedure is the deployment of endoclips along the incised edges of the sinus ( Fig. 16.5 ).

Figure 16.5

Endoscopic needle-knife sinusotomy. (A) A presacral sinus with pus ( green arrow ); (B) complete sinusotomy with needle knife; (C and D) placement of endoclips along the incised edges of sinus.

Endoscopic Injection of Filling Agents

Various agents have been used to fill the track of to inject to the wall of perianal fistula in IBD or non-IBD patients. The procedure is normally performed in a combination with examination under anesthesia (EUA) in OR by colorectal surgeons. Theoretically, the administration of these agents into the fistula track can be achieved via endoscopy.

Fibrin Glue

Fibrin glue injection, with its simple and repeatable advantage, has been extensively used to treat perianal fistula in OR. In a study of 30 consecutive patients with a complex anal fistula, fibrin glue was injected directly into the fistula track following seton placement for 8 weeks. A complete closure of fistulas was achieved observed 17 patients (57%) at 1 month. After a mean follow-up of 12 months, 15 (50%) patients reported the cure for their fistulas. A randomized controlled trial of fibrin glue versus fistulotomy, seton placement, and advancement flap demonstrated that the former approach was less effective in healing simple fistula but more effective in healing complex fistula, than the latter approach.

It appears that the success rate of the treatment was no different in those with or without CD. Fibrin glue treatment has been studied in patients with CD-related fistulae only in a smaller scale. In a case series of 14 patients with CD-related anal fistula, fibrin glue was injected into the fistula tract under endoscopic ultrasound (EUS) during EUA, fistulas completely stopped drainage in 10 patients (71%), the leakage decreased in one (7%) at a 3-month follow-up. At a mean of 23-month follow-up, fistula drainage resolved in eight patients (57%). In a retrospective study of fibrin glue injection for complex anal fistulas in 39 patients, including cryptoglandular, CD-related, and coloanal and ileal pouch-anal anastomotic, and deep transsphincteric, superficial transsphincteric, supralevator, or rectovaginal fistulas. Most patients had an early response with “closure” of fistulae. Recurrence was common and a durable healing was only observed in 12 patients (31%). A multicenter, open-label, randomized controlled trial was conducted in CD patients with fistulas between the anus or low rectum and perineum, vulva, or vagina. Patients were stratified into the simple or complex fistula groups and then randomly assigned to receive fibrin glue injection (n = 36) versus observation only (n = 41), after setons were pulled out. Clinical remission at week 8 was achieved in 13/34 patients (38%) in the fibrin glue group and 6/37 (16%) in the control group.

The long-term outcome of anorectal or perianal fistulas of fibrin glue administration has been disappointing. Multiple factors may contribute to the poor long-term outcomes, including the location, length, and complexity of the fistula track underlying diseases, and concurrent medical and endoscopic (such as endoscopic balloon dilation of concurrent strictures) therapies. Potential foreign body reaction to CD may be a concern. For CD-related fistula, fibrin glue treatment alone may not be adequate, and it may be used in a combination with medical therapy of underlying disease, endoscopic therapy, and surgical flap repair. In a randomized controlled trial of repair alone versus repair and fibrin glue in 58 patients with transsphinteric fistula from CD or non-CD conditions, the recurrence rate was found to be 20% and 46%, respectively.

Endoscopic administration of fibrin glue is technically feasible. Fibrin can be injected through the operating channel of an endoscope. The efficacy of fibrin glue in the treatment of perianal fistula provides the rationale for the endoscopic approach in selected patients. In addition, EUS may be used for the guidance of the therapeutic endoscopic procedure as well as for the monitoring of the outcome. This author has been using hyperosmolar glucose (50%) and doxycycline, as fibrosing agents to treat long-track superficial fistulae and chronic anastomotic sinus. This author has attempted to use n-butyl-2-cyanoacrylate (Histoacryl) via endoscopy embolization to treat non-CD, refractory colorectal surgery–associated anastomotic leak, and fistula. My success rate in the closure of the leak or fistula appears to be low. The results were different from what reported in the treatment of enteric or biliary fistulae.


A variety of forms of plugs, for example, suturable bioprosthetic plug (Surgisis, Cook Surgical, Inc.), have been developed. The plug is typically introduced through the primary opening of a fistula tract. A prospective study of comparison between fibrin glue versus the plug was conducted to treat high transsphincteric fistulae in 25 non-CD patients. The suturable plug was found to have a better outcome for the closure of fistula. A prospective, multicenter study of 93 patients with non-CD complex cryptoglandular transsphincteric anal fistulas was performed. After exclusion of 13 patients who lost to follow up and 21 who withdrew primarily undergoing alternative treatments, the fistula healing rates at 6 and 12 months were 41% and 49%, respectively. The efficacy and safety of the fistula plug have also been studied for the closure of CD-related anorectal fistula. In a prospective study of 20 consecutive patients with 36 fistulae, fistula tracts were successfully closed in 16 of 20 (80%) patients at a final follow-up. The successful closure was found to be correlated with the presence of a seton or the use of anti-TNF agents. The deployment of the plug has traditionally been performed by surgeons in OR. In fact, endoscopic deployment may also be feasible.

Topical Therapy of Stromal Cells or Stem Cells

There has been a growing interest in transplantation of bone marrow–derived stromal cells or adipose tissue–derived mesenchymal stem cells for the treatment of perianal fistulae in patients with CD. A small randomized placebo controlled study of 21 patients with allogeneic bone marrow–derived mesenchymal stromal cells showed a greater rate of healing in the study group than controls. A recent multicenter, randomized placebo control study of 212 patients with allogeneic adipose-derived mesenchymal stem cells showed a significantly greater number of patients in the study group (53/107 [50%]) achieved the combined remission than that by the placebo group (36/105 [34%]) in the intent-to-treat analysis. The most common side-effect is perianal abscess. Those treatments were normally delivered through topical injection of fistula wall and opening in the OR setting. Technically, the therapy may also be administered endoscopically in selected patients in experienced hands. The stem-cell therapy for CD fistula is discussed in Chapter 18 .

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Dec 30, 2019 | Posted by in GASTROENTEROLOGY | Comments Off on Endoscopic Therapy for the Treatment of Inflammatory Bowel Disease-Related Fistula, Sinus, and Abscess

Full access? Get Clinical Tree

Get Clinical Tree app for offline access