Enterocutaneous fistulas, defined as an abnormal communication between the small bowel and skin, are among the most daunting problems for an intestinal surgeon. The impact of an enterocutaneous fistula on a patient varies from a minor inconvenience to fatal malnutrition and dehydration. Depending on the cause and output of the fistula and the comorbidity of the patient, enterocutaneous fistulas can be very challenging to manage. They must be handled correctly. Colocutaneous fistulas are a different proposition with different, usually less major challenges. They are sometimes included if the term “enterocutaneous” is used more generically. However, this chapter deals primarily with fistulas relating to the small bowel.
Most enterocutaneous fistulas are iatrogenic as a result of events such as failed anastomoses, leaks, and unrecognized inadvertent enterostomies. In the 15% to 20% of fistulas that are not iatrogenic, disease results from a perforation of the bowel with surrounding inflammation that quarantines the leak and prevents peritoneal contamination with fecal peritonitis. Instead, an abscess is usually present that, when drained, results in a fistula or erodes through tissues as it works its way to the surface and drains spontaneously. Deep sepsis will always seek drainage via the path of least resistance. This mechanism of development is typical of fistulas due to Crohn disease or perforating cancers. Fistulas that arise from diseased bowel will not heal until the diseased bowel is treated or resected, and ischemic or malignant fistulas are unlikely to heal spontaneously. Iatrogenic fistulas that occur where the bowel is healthy may well heal with time as long as no distal obstruction and no associated abscess cavity or foreign body are present and the bowel has not matured itself to the skin, as in a stoma. When fistula output is low (<200 mL effluent per day), fistulas often heal spontaneously as long as other factors are favorable.
General Aspects of Care
The care and repair of enterocutaneous fistulas require meticulous attention to detail. Control of the associated sepsis, protecting the skin against the corrosive effects of the effluent, and optimization of nutritional and metabolic status are all important aspects of patient care. Treatment also requires patience. It is tempting to perform a second operation immediately to fix a postoperative fistula and make the patient normal, but this temptation should be resisted. Because of the nature of enterocutaneous fistulas (especially those occurring after surgery), it is proper to allow 4 to 5 months (if possible) between surgeries that are designed to fix the fistula. By this time, intra-abdominal adhesions soften, and it is much easier and less dangerous to operate again.
Enterocutaneous fistulas always result in loss of fluid, protein, trace minerals, and electrolytes. Their effects and complications are related primarily to their output:
High-output fistulas: An output of more than 500 mL per 24 hours normally indicates a fistula in the proximal small bowel.
Moderate-output fistulas: An output of 200 to 500 mL per 24 hours indicates that the fistula is likely to be more distal in the small bowel.
Low-output fistulas: An output of less than 200 mL per 24 hours suggests that most of the stool is passing through the small bowel normally and the fistula is diverting a small fraction of it.
The three major complications seen in these patients are sepsis, fluid and electrolyte imbalance, and malnutrition. Their occurrence is directly related to fistula output; higher output corresponds to a higher morbidity and mortality. Mortality rates have historically been in the range of 20% to 40%, with higher rates if the patient has associated cancer. With good treatment, mortality should now be much lower.
Most enterocutaneous fistulas appear postoperatively, often at about the time that bowel function resumes after the usual postoperative ileus. Five to six days after the procedure the patient has pain and a fever, with leukocytosis. The patient may have abdominal tenderness that is increasingly localized to the wound, a drain hole, or an old incision. Drainage may be required, or drainage may occur spontaneously. The effluent often begins as pus that changes to bile or stool. Although the patient is in no acute danger, a lengthy and possibly complicated course of treatment begins that is summarized in the Plan of Care (see the following sections) and in Table 78-1 .
|Phase||Goal||Time Prior to Disease|
Volume adequate—preferably albumin
Nutrition normal—transferrin, albumin
Pulmonary status—chest physical therapy
Potential status re: sepsis—treated
Cardiac status normal or treated
Hematocrit and hemoglobin—normal
Physical therapy—stamina adequate
Perioperative antibiotics (30-60 min before incision)
Bowel preparations: cathartics, nonabsorbed antibiotics
Hibiclens, chlorhexidine washes (72 hr)
Prevent pulmonary emboli
|30-60 min before incision |
Prevent pulmonary emboli
Prophylactic antibiotics—30-60 min before incision
|30-60 min before incision|
|Presentations||Recognition and stabilization |
Volume resuscitation—colloid, crystalloid
Correct anemia: factors, red blood cells
Drain obvious sepsis
Correct electrolyte abnormalities
Initial nutritional support
Total parenteral nutrition
Begin enteral nutrition
Control fistula drainage
Institute local skin care
Engage stoma nurses
Protect gastric, esophageal, duodenal mucosa with a proton pump inhibitor or H + inhibitors
Use nasogastric tubes only if necessary
Estimate or measure nutritional needs
|Up to 72 hr|
|Investigation/elucidation||Delay 10 days or more |
Drainage and radiology: pointing abscess
|After 7-10 days|
|Therapeutic decisions||Will it close? |
Site of fistula
Trend of drainage
Decision to operate
Optimum time 5-6 mo
Other considerations—malignant fistula
|4-6 wk |
|Definitive therapy||Time of surgery |
State of adhesions (estimated)
Presence or absence of sepsis
State of abdominal wall
Septic challenges—emergency intervention
Plastic surgery help
Approach to adhesions
Freeing up bowel
Approach to fistula
Sacrifice 8-18 inches
Anastomosis or anastomoses
Type of suture
2-layer interrupted sutures
Gastrostomy and feeding jejunostomy
Tapering enteral and parenteral nutrition
Preferences—resection and 2-layer interrupted nonabsorbable anastomosis
|Healing||Mortality, prognosis, and complications |
The central nervous system
|Up to 18 mo|
Plan of Care
Six phases of care are required for the patient with gastrointestinal (GI) cutaneous fistulas:
Because most enterocutaneous fistulas are iatrogenic, the best treatment is prevention. Technical issues may exist that should be correctable, such as bowel damage during lysis of adhesions or wound closure, incarceration of bowel in a Richter hernia, inadequate anastomotic technique, or unrecognized enterotomies. Performing a repeat operation in patients with severe adhesions increases the risk of bowel damage, and sometimes a repeat laparotomy must be deferred until postoperative adhesions soften. Operating on diseased bowel can increase the risk of leaks and fistulas, especially if an anastomosis is unwisely attempted. Diverting an anastomosis does not necessarily prevent leaks and fistulas, but not making an anastomosis at all can be wise and preventative. The same comments apply to patients in poor condition—that is, those with an obstruction, sepsis, or anemia, who are malnourished, or who are taking steroids. The choice of strategy to minimize risk is a matter of recognizing increased risk and taking appropriate steps. Sometimes the best strategy is to delay any surgery and operate when conditions are more favorable. In some cases, risk factors can be improved. For example, abscesses can be drained, anemia can be corrected, and the biochemical effects of malnutrition can be reversed. Patients who have inadvertently lost 10% to 15% of their well body weight over a 3- to 4-month period are at risk of poor healing and other complications of malnutrition. Preoperative nutritional support for 5 to 10 days will not restore nutrition but will likely decrease the risk of a poor outcome.
Initial management of an enterocutaneous fistula is geared toward resuscitation. Fluid and electrolyte imbalances should be identified and corrected. In persons with a chronic fistula, trace metals and vitamins should be administered as well.
Concurrent with the resuscitation, control of fistula drainage and skin care should begin. The fistula should be managed with a pouch as if it were a stoma so the effluent can be controlled and its volume recorded. Sometimes the fistula is easy to manage with a pouching system, but often the effluent exits at the base of a complex wound, thus creating challenges in pouching. In such cases, the services of specialized enterostomal and wound nurses can be very helpful. Closed suction dressings also can be helpful in managing deep complex wounds by controlling the effluent and allowing it to be measured.
Rarely, a fistula may reach the skin in the midst of an abscess. When this situation occurs, the abscess should be drained and the pus cultured. A urologic latex catheter in which an extra hole is cut works well, and a No. 14 intracatheter is placed near the end to be a suction catheter that will not erode. This approach will help protect the skin. Drainage of abscesses should proceed, and 24 hours should elapse before a central line is placed for the purpose of nutrition. If the line is placed before the abscess is drained, bacteremia may infect the catheter. Administration of antibiotics is not necessary unless the patient has a clinically significant infection.
Septic patients who display evidence of mental status change, hemodynamic instability, high fever, or signs of impaired organ function should undergo a computed tomography (CT) scan and may need intensive care. In most instances, systemic organ dysfunction is due to an undrained septic focus that must be effectively drained.
Once the fistula is established, the next priority is to measure the output. Enterocutaneous fistulas sometimes gush when they first appear. Patients are placed on nothing by mouth status to minimize stimulation of the GI tract and define baseline output. The fistula output can then be characterized as high or low, and decisions can be made about nutrition: parenteral for patients with high-output fistulas and perhaps elemental oral diets for patients with low-output fistulas. Monitoring and recording fistula output is important to show the likelihood of spontaneous closure and to judge the adequacy of supportive treatment.
Breakdown of the skin around the fistula may make control of the effluent and ultimate repair much more difficult. Wound and enterostomal therapy nurses are very helpful in implementing techniques to provide effective drainage while protecting the skin.
Nutritional support can start after the sepsis is controlled and the fistula is stabilized and effectively pouched. If the gut cannot be used at all because of the effect of oral intake on fistula output, total parenteral nutrition is indicated. Low-output fistulas may be treated with enteral nutritional support. Approximately 4 feet of relatively normal bowel is necessary to sustain nutrient absorption, and a tube gastrostomy or a feeding jejunostomy sometimes facilitates enteral nutrition.
The goal of nutritional support of a patient with an active enterocutaneous fistula should be to provide 30% to 40% more protein and calories than the requirements calculated by normal weight and gender. Nutritional support may not be wholly parenteral but rather a mixture of enteral and parenteral nutrition. It is important for the patient to get the most appropriate doses and mixtures of protein and calories. Because new components and products are now available for enteral nutrition, 60% to 70% of nutritional support should be given enterally if possible.
Monitoring of enteral and parenteral nutrition is essential. The principle is to begin enteral feeding with a dilute solution, with an osmolality no greater than 150 mOsm. If the stomach is used, the concentration of the enteral feeding solution can be gradually increased until the required calories are delivered in a reasonable volume. If nutritional support is given directly into the small bowel, osmolality should be 150 mOsm, and initially the volume rather than the osmolality should be increased. Once the volume is at an appropriate level and can be tolerated, the osmolality can be increased. Giving a maximum osmolality of 250 mOsm helps prevent diarrhea.
Glucose is the primary caloric source for parenteral nutrition. Starting slowly, a dose of 1.8 to 2.5 g of protein per kilogram is provided. This amount will be adequate for protein replacement, but higher amounts may be needed if there is protein loss. At least 10% to 20% of intravenous calories should be given as lipid, provided it is tolerated (see Chapter 83 ).
Nasogastric Tubes and Other Drainage Tubes
Little evidence exists to show that use of nasogastric tubes or suctioning of the GI tract promotes spontaneous closure of enterocutaneous fistulas. If some element of obstruction is present or the nasogastric tube delivers 500 to 1500 mL of gastric or upper GI secretions, this valuable material includes the protein that is synthesized in the stomach and upper gut. It is usually discarded but has the potential to be placed distally into a feeding jejunostomy. However, because the material being suctioned is often contaminated, it is best to avoid this practice.
Protection of the Gastric, Duodenal, and Upper Gastrointestinal Tract Mucosa from Ulceration
Patients with a fistula should be protected from high acid reflux under stress, which may result in Barrett esophagus or an esophageal stricture. A therapeutic dose of an H2 antagonist or an H + K + adenosine triphosphatase inhibitor is administered provided there are no contraindications. Stress and prolonged periods of taking very little by mouth predispose patients to ulceration, and treatment with liquid antacids such as Gelusil or Maalox may cause diarrhea. In addition, a decrease in gastric acid secretion may inadvertently result in an indirect decline in pancreatic biliary secretion.
The use of somatostatin analogues in patients with enterocutaneous fistulas has received a great deal of attention. Long-acting preparations are currently available in a dosing schedule of 10 to 30 mg given intramuscularly every month. The main area where somatostatin may be of help is in the case of pancreatic fistulas. Some studies suggest that treatment with parenteral nutrition alone leads to a pancreatic fistula closure rate between 60% and 75%, whereas adding somatostatin leads to closure rates of 60% to 92%. Although somatostatin may be helpful in treating persons with pancreatic fistulas, it does not help close fistulas that are unlikely to undergo spontaneous closure and those caused by radiation or neoplasia. However, somatostatin is usually effective in decreasing the volume of effluent and making the fistula more manageable. In addition, it may decrease the time to closure of fistulas that are likely to close. The average time of closure seen in patients treated with parenteral nutrition alone is 50 days, but this may be decreased to 5 to 10 days in selected patients when somatostatin is added to the regimen.
There is no rush to investigate an enterocutaneous fistula and determine whether it is likely to close. Closure is certainly not going to happen within 7 to 10 days except perhaps in pancreatic fistulas alone. The situation can be investigated after the acute complications of fistula development have settled, after supportive care is established, and when the patient’s general state is stable. In patients with a postoperative fistula, investigation should wait for all the sequelae of the surgery to resolve, unless the fistula demands urgent attention.
A fistulogram performed using a No. 5 or No. 8 French pediatric feeding tube and water-soluble contrast material is helpful, with the early films yielding accurate and detailed information regarding bowel continuity, location of the fistula, presence of an abscess, presence of intestinal obstruction, the quality of the bowel, the length of the fistula tract, the size of the bowel wall defect, and perhaps the cause of the fistula.
Other studies such as an upper GI series, small bowel follow-through, and a barium enema are often redundant, although if an accurate picture of the anatomy of the fistula in relation to the bowel is not achieved with a fistulogram alone, other studies are indicated. The precise series of studies depends on the context of the patient, his or her underlying disease, and the likely causes of the fistula. The use of CT scans or magnetic resonance imaging (MRI) is usually limited in the evaluation of the patient fistula without sepsis; however, a CT or MRI scan may be a valuable tool in the search for abdominal abscesses in a patient with a fistula who has sepsis and sometimes can be used to place catheters, perhaps to facilitate drainage.
Will It Close?
Management of the patient with a GI cutaneous fistula will ultimately lead to a decision about whether the fistula will close. Esophageal and lateral duodenal fistulas typically close in 15 to 25 days, and colonic fistulas typically close in 30 to 40 days; small bowel fistulas (especially ileum) may take 40 to 60 days to close, if they close at all. Only a third of the fistulas in complicated cases close spontaneously, and a mere 10% to 20% will close without surgery if they are still open after 4 to 5 sepsis-free weeks with adequate nutritional support. At this point the fistula is likely to become lined with epithelium growing toward the skin, in which case it is unlikely to close. Thus only a third of complicated fistulas will resolve spontaneously. Predicting which fistulas will close spontaneously is difficult. Table 78-2 lists the factors associated with fistula closure—favorable and unfavorable—to assist with this prediction.