A fistula (a term derived from the Latin word for pipe) is an abnormal connection between 2 epithelialized surfaces that usually involves the gut and another hollow organ, such as the bladder, urethra, vagina, or other regions of the gastrointestinal (GI) tract. Fistulas may also form between the gut and the skin or between the gut and an abscess cavity. Rarely, fistulas arise between a vessel and the gut, resulting in profound GI bleeding, which is a surgical emergency. (See Etiology.)
Most GI fistulas (75-85%) occur as a complication of abdominal surgery. However, 15-25% of fistulas evolve spontaneously and are usually the result of intra-abdominal inflammation or infection. Regardless of their cause, fistulas have a tremendous impact on patients and society. Increased morbidity and mortality rates, greater health care costs for diagnosis and treatment, prolonged hospital stays, and delayed return to work are just a few direct consequences of this condition. (See Etiology and Prognosis.)
Fistulas were formerly associated with considerable mortality rates. In the decades following the 1960s, however, the introduction of intensive care units (ICUs) and parenteral nutrition lowered the mortality rate to approximately 20%; however, prolonged hospital stays and the high cost of medical and surgical care remained unchanged. In addition, the frequency of fistula formation has not decreased, because of advanced and complicated disease, complex surgical techniques, and an aging population.
Several classification systems for fistulas exist, none of which are used exclusively. The 3 most commonly used classification systems are based on anatomic, physiologic (output volume), and etiologic characteristics. Used in combination, these classifications can help to provide an integrated understanding and optimal management scheme for the fistula. (See Etiology, Treatment, and Medication.)
Anatomically, the fistulas are named according to their participating anatomic components, and they can be divided into internal and external fistulas. Internal fistulas connect the GI tract with another internal organ, the peritoneal space, the retroperitoneal space, the thorax, or a blood vessel. External fistulas, which commonly occur postoperatively, are abnormal connections between the GI tract and the skin. (See Presentation.)
Risk factors for intestinal fistulas include the following:
Contrary to common belief, fistulas do not necessarily develop as a consequence of downstream stenosis of the intestine. (See the image below.)
Enterocutaneous fistula after bowel injury from an incisional hernia repair, 6 weeks postinjury.
Gastric fistulas are iatrogenic in most cases (85%). The other cases are usually a consequence of irradiation, malignancy, inflammation, and ischemia. Anastomotic leak after a gastric resection for cancer, peptic ulcer disease, or bariatric surgery can lead to leakage of intestinal or gastric juices, which initiates a cascade of events: localized infection, abscess formation, and, possibly, abscess and fistula formation.
Nearly 80% of small bowel fistulas result from complications of abdominal surgery. These fistulas may occur from disruption of the anastomotic suture line, inadvertent iatrogenic enterotomy, or small bowel injury at the time of closure. Inadequate blood flow from devascularization or tension at the anastomotic suture lines, anastomosis of diseased bowel, or perianastomotic abscess may compromise the integrity of surgical anastomoses.
Crohn disease, malignancy, peptic ulcer disease, and pancreatitis spontaneously cause 10-15% of small bowel fistulas. In patients with Crohn disease, fistulas arise from aphthous ulcers that progress to deep transmural fissures and inflammation, subsequently leading to adherence of the bowel to adjacent structures that eventually penetrate other structures. Microperforation with abscess formation leads to subsequent macroperforation into the adjacent organ or skin, resulting in fistula formation.
Crohn fistulas are more often internal and less commonly external (to the skin). Ileosigmoid fistulas, usually a complication of a diseased terminal ileum that invades the sigmoid colon, are the most common type of fistula between two loops of bowel. Enteroenteric, gastrocolic, duodenocolic, enterovesical, rectovaginal, and perianal fistulas are other potential complications of Crohn disease. Perianal fistulas are the most common external fistulas in patients with Crohn disease. (See the image below.)
Psoas abscess from Crohn disease that later fistulized to the skin.
Colonic fistulas are primarily a consequence of intra-abdominal inflammation but can also occur after surgical intervention for an inflammatory condition. IBD, diverticulitis, malignancy, and appendicitis (especially with the presence of an appendiceal abscess requiring percutaneous drainage) are the most common inflammatory conditions that lead to colonic fistulas.
Aortoenteric fistulas most commonly occur secondarily, usually after the surgical placement of a graft. Aortoenteric fistulas can develop in the following ways:
In developed countries, Crohn disease is the most common cause of spontaneous fistula formation. In their lifetime, as many as 40% of patients with Crohn disease develop a fistula, most often an external or a perianal one.
The incidence of fistula formation in patients with diverticulitis is much lower. Fistula formation complicates diverticulitis in 1-12% of patients. Colovesical fistulas in men and colovaginal fistulas in women are the most common types of fistulas in this population.
Fistulas can complicate radiation therapy weeks to years after treatment. Radiation therapy for malignancy is associated with fistula formation in approximately 5-10% of patients. Notably, surgery and anastomosis in previously irradiated tissue increases the risk of anastomotic leak and, subsequently, fistula formation.
Internationally, the frequency of various types of fistulas may vary in correlation with their prevalence in different populations.
For example, the prevalence of fistulas secondary to Crohn disease may be less prevalent in Africa primarily because the disease is less prevalent in that population. However, the prevalence of obstetric fistulas may be higher in developing countries because of obstructed labor (including malpresentation and cephalopelvic disproportion) and lack of prompt access to emergency obstetric care. Accurate prevalence rates of obstetric fistulas are unavailable, likely because of inaccurate reporting of the medical condition and the stigma of its associated symptoms.
Racial differences in patients with fistulas generally parallel those of the underlying disease or condition that predisposed persons in a specific racial population to developing fistulas. For example, since Crohn disease is more common in whites, patients with Crohn disease who develop fistulas are more likely to be white.
With regard to sex-related prevalences, colovesical fistulas are more common in men and in women who have undergone a hysterectomy. Colovaginal fistulas, of course, occur only in women. Otherwise, fistulas are equally prevalent in males and females.
As with race, age parallels the etiology or underlying condition that predisposes patients to develop fistulas.
The prognosis is based on the etiology of the fistula, as well as the comorbidities of the patient. Pain, wound management, abscess formation, local infection, nutritional deficiencies, and recurrent septic states are just a few of the physical consequences of intestinal fistulas.
Patients with fistulas most likely present with much more than physical discomfort and pain. The stigmas of malodorous fistula drainage, malnutrition, and emotional distress also cause significant psychological consequences. In addition, patients with postoperative fistulas have the added distress of lengthy hospital stays, associated morbidity, a delay in returning to work, and restricted social activities. Considerable mortality is associated with fistulas, primarily from sepsis.
In one study, in patients who developed fistulas after pancreaticoduodenectomy, specific factors were associated with increased mortality. These factors included fistula site, underlying disease, low hospital volume, the surgeon's experience, high intraoperative blood loss, and complications.
Symptoms caused by fistulas that involve 2 segments of the bowel vary depending on the location of the fistula and the amount of bowel bypassed. For this reason, enteroenteric fistulas in which only a short segment of bowel is bypassed may be asymptomatic and diagnosed incidentally based on imaging findings or during surgery. Conversely, ileosigmoid fistula may cause diarrhea, weight loss, or abdominal pain.
Patients with gastrocolic fistulas may present with symptoms of abdominal pain, weight loss, and feculent belching.
Enterovesical and colovesical fistulas are easier to diagnose in patients who present with symptoms of pneumaturia, fecaluria, and recurrent urinary tract infections.
Patients with rectovaginal and anovaginal fistulas may be asymptomatic and present with symptoms only when the bowel movements are more liquid. Possible symptoms include inadvertent passage of stool or gas, dyspareunia, and perineal pain.
Patients with external fistulas generally present with symptoms of drainage through the skin. Patients with aortoenteric fistulas may report rectal bleeding.
Fluid or stool output through the skin, diarrhea, abdominal tenderness, weight loss, signs of malnutrition, and electrolyte imbalances are all possible findings in patients with fistulas.
Rectal bleeding may be a finding in patients with a history of radiation therapy. Hypotension and rectal bleeding may occur in patients with aortoenteric fistulas.
This can be helpful in determining the origin of the bowel disease that caused the fistula, but it is not a particularly helpful or necessary study to reveal a fistula. Biopsy samples may be obtained during the procedure and are useful in diagnosing IBD or Crohn disease and malignancy.
Although it has been reported and described, this procedure is not a widely used modality for diagnostic and therapeutic use with enteric fistulas. A small-caliber endoscope is passed into the lumen of the fistula in an attempt to identify the source of the fistula. Fistuloscopy may identify abscesses and visualize the bowel involved. Therapeutically, a drain can be placed or fibrin glue sealant may be applied to close the fistula.
Useful in the evaluation of suspected enterovesical fistula, cystoscopy may allow visualization of fistulas from within the bladder.
Instilling methylene blue into the rectum and examining a vaginal tampon 15 minutes after placement can often establish the presence of a rectovaginal fistula.
Albumin and prealbumin levels should be obtained, as well as blood urea nitrogen (BUN), creatinine, and electrolyte concentrations. These are used to determine the patient's nutritional status and whether fluid or metabolic disturbances are present (more of a concern for high-output fistulas).
Although complete blood count (CBC) results may be within the reference range, leukocytosis may be present if an undrained abscess or a continued inflammatory process has developed within a segment of the bowel. Anemia may be present with chronic disease or if a malignant process is involved.
Abscess culture findings may be helpful, especially in the presence of sepsis or ongoing infection (the predominant organism involved being Escherichia coli). Cultures of enterocutaneous fistula output may not be of much clinical use, as normal bowel flora often predominates.
For colovesical fistulas, urinalysis usually reveals increased white blood cell (WBC) count and bacteria levels. Urine culture findings may help to direct antibiotic therapy.
Histologic findings of fistula site biopsy are usually consistent with chronic inflammation. In patients with Crohn disease as the causative factor, transmural involvement with noncaseating granulomas and lymphoid aggregates throughout the bowel wall may be observed. In patients with carcinoma, inflammation adjacent to the tumor remains a typical finding. The clinical scenario and test results are usually helpful in determining the diagnosis.
Staging is appropriate when the etiology of the fistula is carcinoma.
Patients can be given charcoal or Congo red dye orally to verify the presence of an enterocutaneous fistula. However, this is not helpful in determining which portion of bowel is involved. This test is often used in postoperative patients with persistent drainage from a wound in whom an enterocutaneous fistula is suspected or in women with persistent vaginal drainage in whom a rectovaginal fistula is suspected.
Abdominal and pelvic computed tomography (CT) scanning is the imaging method of choice to evaluate Crohn disease and possible fistulas. While identification of the fistula is not always possible, CT scanning often reveals perifistular inflammation. This provides additional information regarding the possible etiology of the fistula and the extraluminal involvement of disease.
Revealing abscess cavities or excluding possible sources of sepsis is an important step in the evaluation of patients with suspected fistulas. This information may also prove helpful if surgical intervention is planned. CT angiography may be used in the diagnosis of suspected aortoenteric fistulas if the patient is stable.
Although magnetic resonance imaging (MRI) is reported as an imaging modality that can help identify and characterize enteric fistulas, motion artifact may limit its usefulness, and MRI is not considered a routine adjunctive study in the evaluation of patients with enteric fistulas. T1-weighted images provide information relative to the inflammation in fat planes and possible extension of the fistula relative to the surrounding visceral structures. T2-weighted images can demonstrate fluid collections along the fistula tract and inflammatory changes within the surrounding muscle.
Radiographic study with contrast medium (usually given at the site of fistula output) may be performed to help delineate the extent of the fistula and its communication with the underlying bowel.
Ultrasonography can be used in conjunction with physical examination to identify abscesses and fluid collections along the fistula tract.
Contrast studies to evaluate the stomach, small intestine, and colon may reveal a fistula and may also be helpful in determining the cause of fistula formation by identifying diverticular disease, Crohn disease (characteristic string sign), or evidence of malignancy.
This procedure can help to evaluate for the presence of a possible enterovesical fistula.
Angiography may assist in preoperative planning and evaluation of aortoenteric fistulas in a stable patient or determine the arterial source of bleeding in those with a less common arterioenteric fistula.
Follow-up care is based on disease etiology. In patients with chronic inflammatory conditions, such as Crohn disease, ongoing follow-up care is required.
If patients are simply waiting for definitive surgical therapy and being maintained on TPN, they do not need to be in the hospital. Home infusion or placement in a rehabilitation facility is perfectly acceptable. Close contact is needed, as line sepsis and other infectious complications may occur, and early and aggressive treatment is essential to ensure a good outcome.
In patients who have undergone surgical repair, normal postoperative follow-up care is required. Once the incisions are healed and the drains are removed, the patient may be discharged from care.
Patients with Crohn disease may require ongoing outpatient maintenance therapy with medication.
Several elements are required to successfully treat patients with an intestinal fistula: adequate nutrition, control and maintenance of the fistula drainage site, appropriate treatment of infection, and avoidance of sepsis.[11, 12]
Spontaneous closure of a proportion of GI fistulas with nonoperative management is well documented. Although dependent on the etiology of the fistula, 60% or more close if they are iatrogenic, if no distal obstruction is present, if no foreign body is involved, if the tract is long, if there is a low output, and if there is no active infection. Numerous studies have delineated the important determinants associated with decreasing the time to closure of a fistula and decreasing a patient's overall associated morbidity and mortality.
Conservative management of enteric fistulas has been described for periods of up to 3 months. One study demonstrated that 90% of the fistulas that spontaneously closed did so within the first month, once management of sepsis had been established. Of note, none of the fistulas spontaneously closed after 3 months.
Factors to consider for fistulas that do not spontaneously close include the following:
In these cases, surgical repair may be the definitive treatment. (See the image below.)
Status post-pancreatic debridement for necrotizing pancreatitis. The patient had a colonic injury with attempted closure using a skin graft. The patie....
The following consultations can aid in treatment:
In the initial period, patients are maintained on total parenteral nutrition (TPN) and are given nothing by mouth (NPO).
In patients with low-output, distal fistulas, elemental diets may be initiated as long as they do not profoundly increase the fistula output.
Aggressive physical therapy provides long-term benefits to patients. Typically, patients do not require prolonged bedrest (which only adds to comorbidities) unless this is necessary for some other reason.
Initial fistula management should address each of the following resuscitation and stabilization issues in patients with a GI fistula.
TPN has long been regarded as an essential therapy (especially in high-output fistulas) to decrease output and to maintain good nutritional status.
Malnutrition is a significant cause of morbidity and mortality, especially with enterocutaneous fistulas. Typically, patients with low-output (< 200 mL/24 h) fistulas should receive their full resting expenditure, 1-1.5 g of protein/kg daily and a lipid intake that accounts for approximately 30% of daily caloric intake. Patients with high-output (>500 mL/24 h) fistulas should receive 1.5-2 times their resting energy expenditure, 1.5-2.5 g of protein/kg daily, and twice the recommended daily allowance of lipids.
Control of enteric contents draining from the fistula continues to be a topic of ongoing research and development. Standard ostomy supplies and other methods of skin care and drainage control can be used in an attempt to reduce or eliminate the persistent tissue inflammation and infection surrounding the fistula, which can lead to sepsis. Use of the vacuum-assisted closure (VAC) device to better manage output has been reported to help improve the perifistular environment. All of these techniques and devices are used not to close the fistula, but rather to help keep the surrounding tissues healthy and to allow the fistula to heal on its own.
The use of CT scanning and ultrasonography can help to determine if fluid collections or abscesses are present along the abscess tract. Identification of these fluid collections often allows for CT-guided drainage of these loci to prevent infection. Along with better drainage control and appropriate antibiotic treatment, this helps to decrease the morbidity and mortality associated with enteric fistulas and allows for a safer period of conservative management. Radiologically placed catheters have been demonstrated to safely and successfully drain most abscesses.
In patients with proximal, high-output fistulas, the volume depletion associated with the drainage can be a significant problem. Whether medications such as octreotide (a synthetic substitute of somatostatin that suppresses the release of many GI hormones) help to close fistulas remains unclear. Research has demonstrated significant and nonsignificant effects of medication on the closure of fistulas, yet these studies agree that octreotide does decrease overall fistula output.
The period of nonoperative management of an enteric fistula, while allowing for spontaneous closure of the fistula, also provides time to optimize nutritional status and to heal the wound site from the patient's initial surgery (if the enteric fistula occurred postoperatively). Thus, definitive surgery for fistula repair is generally delayed for several months until physiologic deficits have been restored and intra-abdominal conditions are less hostile. However, if diffuse peritonitis with ongoing sepsis is observed, immediate operative exploration may be necessary to stabilize the patient.
The preferred procedure involves excision of the fistula tract, with segmental resection of involved bowel and anastomosis of the remaining bowel. If an unexpected abscess is encountered or the quality of the bowel wall is suboptimal, some surgeons may consider a primary anastomosis unsafe, instead choosing to perform a staged procedure, with exteriorization of the ends of the bowel during the first procedure. A staged repair may also be more appropriate in cases in which advanced malignancy or severe radiation changes are expected. If the procedure is performed for a malignancy, preferably, the involved segment of bowel is removed to negative margins.
The abdominal wall may not be able to be closed due to a lack of fascia. In these cases, biologic meshes, which are manufactured from porcine or bovine dermis, human cadaveric dermis, or porcine small intestinal submucosa, may be used to help close the abdominal cavity. These meshes are not at risk for infection.
Perianal abscesses should be drained and anal strictures dilated. Patients with low anal fistulas can be treated with fistulotomy. Some surgeons are in favor of a noncutting seton, especially in the presence of an active inflammation of the rectosigmoid colon. Noncutting setons may be placed in fistula tracts in patients with rectal inflammation, and endorectal advancement flap procedures for high perianal fistulas and rectovaginal fistulas may be performed in patients without rectal inflammation. Another option that preserves sphincteric function is using a fistula plug, which provides a matrix for soft-tissue healing.[19, 20]
If a rectovaginal fistula persists after the patient has received medical therapy and anorectal stricture or active rectal disease is not evident, then surgical repair may be performed with either (1) transanal or transvaginal advancement flaps or (2) laparotomy with primary closure or sleeve advancement flap.
Patients with colovesical fistulas can almost always be treated with resection of the involved segment of colon and primary reanastomosis, with or without closure of the bladder defect. Healing of the bladder is usually managed easily with temporary urethral catheter drainage.
The overall incidence of aortoenteric fistulas has changed with the advent of endovascular repairs of abdominal aortic aneurysms, but the criterion standard remains open excisional repair and extra-anatomic bypass for revascularization in the case of secondary aortoenteric fistulas (which occur after open repair of abdominal aortic aneurysms). Endovascular repair of primary aortoenteric fistulas in high-risk patients (ie, those at risk for chronic infection of the endograft) has been reported.
Intestinal fistulas carry high morbidity and mortality rates. If medical and nonoperative treatments are not effective, the risks of surgery need to be discussed with patients and their families.
Complications are routine, as dense, fibrotic adhesions are likely to be encountered during surgery. Infection, bleeding, and injury to adjacent organs, as well as recurrence of the fistula, are all possibilities. To reduce these complications, patients must receive optimal nutrition and must be treated by an experienced surgeon.
As previously stated, it is not yet clear whether medications such as octreotide (a synthetic substitute of somatostatin that suppresses the release of many GI hormones) help to close fistulas. Although contradictory findings on the role of medication in fistula closure have been found, research has shown that octreotide decreases overall fistula output. Immunosuppressive agents used in fistula therapy include azathioprine and infliximab.
Clinical Context: Octreotide is an inhibitory hormone consisting of 2 peptides (14 amino acids and 28 amino acids in length) secreted by hypothalamus and delta cells of the stomach, intestines, and pancreas. It inhibits the release of growth hormone (GH) and thyroid-stimulating hormone (TSH) and suppresses the release of many GI hormones (gastrin, cholecystokinin, secretin, motilin, vasoactive intestinal polypeptide [VIP], glucose-dependent insulinotropic polypeptide [GIP]).
Octreotide results in decreased gastric emptying and reduces smooth muscle contractions and blood flow in the intestines. It is approved for use in treating acromegaly (since it blocks GH release) and symptoms related to carcinoid syndrome and VIPomas.
These agents inhibit the release of serotonin and the secretion of many hormones involved in GI function.
Clinical Context: Azathioprine is a prodrug that is converted in the body to 6-mercaptopurine (6-MP). A member of a group of medicines called antimetabolites, azathioprine is a chemotherapy medication that inhibits the activity of the immune system, consequently reducing inflammation. Originally developed to treat certain forms of leukemia, the drug has been used to treat some conditions in which the immune system is overly active, such as Crohn disease. Clinical studies of azathioprine therapy have demonstrated a significant improvement in closure rates or improvement of the fistula site compared with placebo.
Clinical Context: Infliximab is a monoclonal antibody with murine variable regions that specifically bind human tumor necrosis factor alpha (TNF-alpha), which has important role in promoting inflammation. By blocking the action of TNF-alpha, infliximab reduces signs and symptoms of inflammation. Clinical studies have demonstrated that infliximab significantly improved closure rates of fistulas in patients with Crohn disease and was as effective as maintenance therapy in reducing the number of relapses compared with previous medical treatment therapies.
These agents inhibit the activity of key factors in the immune system.