You are on page 1of 47

Management of Complex

Gastrointestinal Fistula
Despite all the medical advances over the last 2 decades the management
of gastrointestinal fistula still remains a significant challenge and carries
a mortality rate of up to 10%.1 This mortality rate is even higher when the
fistula is associated with an open abdominal wound.2-4 These fistulas are
most often seen in the postoperative period due to anastomotic leakage.
They typically occur following surgery for intestinal obstruction, cancer,
or inflammatory bowel disease.5,6 Thus, 75% to 85% are iatrogenic in
origin.7 Their management requires input from a wide range of personnel,
with attention to the control of sepsis, fluid and electrolyte balance,
maintenance of nutrition, and attention to wound/stoma care. The prin-
cipal participants in patient management include nutritionists, enteros-
tomal therapists (ET), radiologists, psychiatrists/psychotherapists, nurses,
internists, surgeons, and other personnel (Table 1). It is a condition that
places a considerable economic burden on the healthcare provider.
Regardless of the pathogenesis, the management most often requires
considerable lengths of hospital stay and extensive multidisciplinary
input. The development of an intestinal fistula following surgery is a
devastating complication for the patient and their family. It may lead to
significant anxiety, loss of self-esteem, depression, and considerable loss
of earnings and financial hardship.
Fistulas are defined as an abnormal communication between 2 epithe-
lized surfaces. This article focuses on acquired as opposed to congenital
fistulas. Several classification systems have been used in their descrip-
tion.8-10 The anatomic classification names the fistula according to the
organs involved. The high pressure organ from which the fistula arises is
named first (eg, colovesical, aortoenteric, or gastrocutaneous). The
physiological classification is based on their output over 24 hours.
High-output fistulas produce more than 500 mL/24 h and lead to
considerable difficulties with fluid management and skin care. These
generally originate from the small bowel and the patient may require total
parenteral nutrition (TPN). Moderate-output fistulas produce 200 to 500

Curr Probl Surg 2009;46:384-430.


0011-3840/2009/$36.00 0
doi:10.1067/j.cpsurg.2008.12.006

384 Curr Probl Surg, May 2009


TABLE 1. Members of the multidisciplinary team
Enterostomal therapists
Surgeons/medical personnel
Nurses
Radiologists
Nutritionists
Infectious disease team
Psychiatrists/psychologists

mL/24 h. In contrast, low-output fistulas produce less than 200 mL/24 h,


are generally of colonic origin, and the patient may be able to tolerate
normal oral intake. There is still considerable debate and differing
opinions about whether there is a correlation between the volume of
fistula output and the potential for closure.9-11 Several publications have
shown an association between the extent of fistula output and mortality
rate, with high-output fistulas having a greater potential for mortality than
low-output fistulas.12,13 Another classification system is based on fistula
etiology (eg, diverticular fistula, malignant fistula, etc.).8 In addition,
fistulas may be classified as internal or external, simple, or complex.
Internal fistulas arise when there is communication between adjacent
bowel loops or organs. If the intestinal fistula only bypasses a short
segment of bowel then the patient may be relatively asymptomatic
without any electrolyte disturbance. However, if the intestinal segment
bypassed by the fistula is significant (eg, gastrocolic fistula) then the
patient will present with major electrolyte abnormalities and nutritional
deficiencies.
Potentially any intra-abdominal organ may be affected (Fig 1). In the
case of spontaneously occurring fistulas one must have a high index of
suspicion for underlying Crohns disease or malignancy.5,14-16 These, in
addition to fistulas that develop after radiotherapy, tend not to close
spontaneously and generally require surgical intervention. Other coexist-
ing factors that make spontaneous fistula closure unlikely include distal
obstruction, complete anastomotic dehiscence, ongoing intra-abdominal
sepsis, malnourishment, diseased bowel, and the formation of an epithe-
lized tract.7-8 In addition, when an enterocutaneous fistula (ECF) occurs
in a patient with an intra-abdominal mesh, it is unlikely to close until the
offending foreign material is removed.17
A very useful acronym in the management of patients with intestinal
fistula, which has evolved from the experience gleamed from several
centers down the years, is SNAP; S is for control of sepsis and
appropriate skin care; N is for nutrition, ideally via the enteral route (if
Curr Probl Surg, May 2009 385
FIG 1. A left ureterocolic fistula that occurred following surgery for retroperitoneal fibrosis. (Courtesy
of the Department of Colorectal Surgery & Radiology, Cleveland Clinic, Cleveland, OH.)

not, then parenterally or a combination of both); A is to define the


underlying anatomy; P is for a definite plan to deal with the fistula18
(Table 2).
The majority of fistulas that respond to conservative management will
close within 6 weeks. If the fistula does not close spontaneously, then the
time to surgical intervention is critical to the outcome. It is our experience
386 Curr Probl Surg, May 2009
TABLE 2. SNAP: A useful acronym in management of patients with complex intestinal fistula
S: Sepsis/skin care
N: Nutrition (enteral/parenteral)
A: Define underlying anatomy with appropriate imaging
P: Definitive plan to deal with the fistula

that intra-abdominal adhesions achieve a maximum density from the tenth


postoperative day to the sixth week. Surgery within this time frame is ill
advised because of the difficult dissection with potential for multiple
enterotomies with ensuing fistulas. In addition, there is the risk for
mesenteric vascular injury, necessitating extensive small bowel resection
and the potential for short gut syndrome. We would delay surgery until all
the physiologic and metabolic parameters have stabilized and generally
like to wait 6 months from the time of previous surgery. Occasionally one
may have to intervene within this critical time period because of ongoing
sepsis and failure to stabilize or progress. One may also be forced to
operate if wound problems lead to an intolerable situation for the patient
and family. On occasions one may encounter a very hostile abdomen, in
which case the creation of a proximal stoma with minimal distal
dissection may be the safest solution until one returns for definitive
surgical intervention at a later date. If one is unable to mobilize the bowel
and exteriorize it without tension at the marked stoma site then it may
have to be brought through the midline wound (Fig 2). When operating on
patients with intestinal fistulas surgeons must be aware of the predictors
of further recurrence. The Cleveland Clinic experience supports resection
and primary anastomosis over wedge repair or oversewing.19
Some fistulas may produce quite dramatic symptoms, such as the
discharge of feces per urethra or vagina or onto the skin in the case of a
colocutaneous fistula. This will lead to a high level of anxiety for the
patient and anyone unfamiliar with such pathology. However, in the
absence of underlying malignancy and sepsis these fistulas may be
managed conservatively in a subset of patients with extensive comorbidi-
ties. One must balance the impingement on patients quality of life with
the risk of surgical intervention. A fistula associated with an ostomy site
may cause considerable wound problems and difficulty with pouching
(Fig 3). In patients with internal Crohns fistulas that are asymptomatic
one should consider the common wisdom that it is impossible to make an
asymptomatic patient feel better. A percentage of these patients will
become symptomatic and ultimately need surgical intervention, but a
proportion may be spared an unnecessary operation with its associated
risks.20
Curr Probl Surg, May 2009 387
FIG 2. On occasions one may have to be prepared to bring a stoma through the midline laparotomy
wound if tension is a problem. (Courtesy of the Enterostomal Nursing Department, Cleveland Clinic,
Cleveland, OH.)

Given that the majority of intestinal fistulas arise following an abdom-


inal procedure, careful attention to intraoperative detail and meticulous
hemostasis are important if one is to reduce the incidence. We would
liken the developments in the care of intestinal fistula to that of pancreatic
necrosis, where the trend is to drain any pancreatic sepsis, maintain
adequate nutrition, and delay surgical intervention until a reasonable
amount of time has elapsed from the primary event.21-23 If the fistula does
not close spontaneously then we would recommend that such patients
should be managed in a specialized unit with appropriate expertise, given
the difficulties with care and multidisciplinary input required.
Etiology and Prevention
The majority of fistulas are due to preceding surgery or trauma.
Typically the patient has undergone a bowel anastomosis that has
subsequently leaked. Thus when performing any anastomosis one must
pay careful attention to the basic principles of ensuring a tension-free
repair and good blood supply to the bowel segments (Table 3). In
addition, intraoperative and patient circumstances must be favorable for
an anastomosis (ie, a well-nourished patient in the absence of extensive
388 Curr Probl Surg, May 2009
FIG 3. An ileostomy with associated fistula may give rise to a wound that is difficult to manage.
(Courtesy of the Enterostomal Nursing Department, Cleveland Clinic, Cleveland, OH.)

TABLE 3. Conditions required when creating an anastomosis


Adequate exposure
Tension-free anastomosis
Good blood supply
Well-nourished patient
Patient must not be immunocompromised
Minimal contamination of operative field
Hemodynamically stable patient

comorbidities or immunosuppression). If the patient has risk factors for


leakage then a diverting stoma is more prudent.24 The authors favor
performing a temporary diverting loop stoma if the patient has had
preceding chemoradiotherapy for rectal cancer, if one is performing a
coloanal anastomosis, or in high-risk patients who would not tolerate the
clinical consequences of a leak. We would also generally favor diverting
an ileal pouch-anal anastomosis and would only consider omission of the
ileostomy under optimal conditions.25 It is important to close any
mesenteric defects to prevent postoperative internal hernia that may lead
Curr Probl Surg, May 2009 389
to bowel obstruction and anastomotic dehiscence. Before closure one
should examine the entire small bowel several times to identify serosal
tears or inadvertent enterotomies. In addition, particularly after a long
case, one must pay utmost attention to the fascial closure. One must avoid
catching small bowel or the stoma in the closure suture. Operative times
extending beyond 2 hours and contamination of the field from spillage of
intestinal contents are risk factors for anastomotic leakage. Many sur-
geons favor wrapping the omentum around the anastomosis, but there is
no randomized trial to prove its validity. Instinct would suggest that it
may help to control a small leak. The anastomosis should be placed well
removed from the abdominal wound. The immediate postoperative period
is critical. All organs must be well perfused. Any episode of hypotension
may reduce blood supply to the anastomosed bowel segments with
subsequent ischemia, perforation, and fistulization.
Patients who require several abdominal procedures for the control of
sepsis (eg, pancreatic necrosis) or surgery for extensive trauma may
develop considerable edema of all tissues and organs. In this scenario it
may not be possible to close the abdomen and therefore one may have to
perform a laparostomy because of the risk of abdominal compartment
syndrome. However, the open abdomen and exposed intestinal contents
carry a risk for fistulization.26 In addition, retraction of the abdominal
wall tissues means that the patient may ultimately require abdominal wall
reconstruction, which may necessitate the insertion of a biological mesh,
which, despite all the advances, is still susceptible to infection.27
All abdominal procedures carry the inherent possibility of the develop-
ment of a fistula and this should be explained to the patient as part of the
informed consent. Although the incidence is low, the devastating effects
of an intestinal fistula mean that the patient should have some knowledge
of this potential complication. Iatrogenic causes due to instrumentation/
interventional procedures are another relatively common cause of intes-
tinal fistulas. One of the most common is a gastrocolic or colocutaneous
fistula arising from percutaneous endoscopic gastrostomy (PEG) tube
insertion.28,29 With advances in interventional radiology, patients with
significant gastrointestinal bleeding may be managed by mesenteric
angiography with embolization. Although successful in the control of
bleeding it may render the treated segment of bowel ischemic with
resulting perforation and fistulization. In patients with extensive comor-
bidities or metastatic disease, palliative stents are often used to deal with
the impending obstruction from a primary malignancy. However, one of
the risk factors with these indwelling stents, whether esophageal, colonic,
or biliary, is the potential for erosion and fistula formation.30-32 Even
390 Curr Probl Surg, May 2009
FIG 4. Computed tomographic enterography identifying an ileosigmoid fistula with arrow pointing to
site of communication. (Courtesy of the Department of Colorectal Surgery & Radiology, Cleveland
Clinic, Cleveland, OH.)

ventricular-peritoneal shunts have been reported to migrate into the


intestines.33 Iatrogenic fistulas tend to present quickly, in contrast to
spontaneously occurring fistulas, which have a more insidious onset.
Fistulas may also arise following blunt or penetrating abdominal
trauma. In particular, injuries to the upper abdomen may result in a
duodenal or pancreatic fistula. A fistula may be the primary presentation
of almost any intra-abdominal condition. Spontaneous fistulas may occur
due to inflammatory bowel disease, in particular Crohns disease, which
causes a transmural inflammation.5,8 The typical scenario is the presence
of terminal ileal disease, which becomes adherent and then fistulizes to
the sigmoid colon (Fig 4). Often the point of communication with the
sigmoid colon is subtle and requires careful examination. In many cases
it may fistulize to the sigmoid at its mesenteric border. At laparotomy, if
the site of fistulization is not obvious then submerging the sigmoid colon
in water and performing a flexible sigmoidoscopy may help. In addition
to identifying the point of communication this will help to confirm that
the fistula is due to terminal ileal disease and that the sigmoid colon is free
Curr Probl Surg, May 2009 391
of Crohns disease. This is very important when deciding the subsequent
surgical management. Crohns disease may also give rise to an ECF or
extend retroperitoneally to involve the psoas muscle, causing an abscess.
Crohns fistulas tend not to close spontaneously and, if symptomatic, may
require surgical intervention. Other causes for spontaneous fistulas
include diverticular disease, bowel ischemia, as sequelae to radiation
enteritis, pancreatitis, perforated duodenal ulcers, and erosion of indwell-
ing catheters.30 The management of a fistula following radiotherapy may
be particularly difficult. A perforated tumor giving rise to an ECF is
associated with a poor prognosis. In rare cases a fistula may arise from
appendicitis or following an appendectomy, which is quite traumatic for
the patient and their family because their initial perception was that they
were undergoing a simple, nearly risk-free procedure.34,35
Special Fistulas
The development of a fistula following the creation of a pelvic pouch is
disappointing and may lead to pouch loss (Figs 5 and 6). The incidence
varies between 4.5% and 10%.36,37 Risk factors for leakage include
prolonged steroid use, hypoalbuminemia, anemia, and tension at the
anastomotic site. The existence of these factors or any intraoperative
technical difficulties again mandates the need for a diverting ileostomy.38
Although a diverting ileostomy will not prevent anastomotic leakage it
will reduce the clinical consequences of the ensuing sepsis, reducing the
re-laparotomy rate and potential for pouch loss. In the review of our
institutes experience we have found that the use of a double-stapled
technique over a handsewn anastomosis is associated with fewer septic-
related problems and better functional results.39 In addition, any surgeon
contemplating the formation of an ileal pouch anal anastomosis (IPAA)
must be aware of all the clinical steps that may be required to ensure the
pouch reaches the pelvis without any undue tension on the mesentery.
Technical tips include dividing the ileocolic artery just distal to the
superior mesenteric artery, mobilizing to the level of the duodenum,
scoring the mesentery, and ensuring that the pouch reaches before
dividing the distal rectum (Fig 7). If reach is a significant problem (that
may particularly occur in tall men) then one may use an S pouch since
this will afford additional length (Fig 8). However, the efferent limb
should be only 2 to 3 cm since excessive length may give rise to
obstructed defecation.40,41 Undue tension on the anastomosis and accom-
panying pouch mesentery may result in pouch ischemia and the devel-
opment of a pouch fistula. At the end of the procedure we recommend
inserting a drain into the pelvis. This will drain any collection and reduce
392 Curr Probl Surg, May 2009
FIG 5. Pouchography demonstrating a pouch vaginal fistula with an obvious fistulous tract (arrow).
(Courtesy of the Department of Colorectal Surgery & Radiology, Cleveland Clinic, Cleveland, OH.)

Curr Probl Surg, May 2009 393


FIG 6. A pouch-cutaneous fistula. (Courtesy of the Department of Colorectal Surgery & Radiology,
Cleveland Clinic, Cleveland, OH.)

the potential for the development of a presacral hematoma. Any pelvic


hematoma may become infected and discharge through a pouch suture
line, giving rise to a fistula.
The clinical manifestations of pouch leakage may be mild including
tachycardia, leukocytosis, back pain, or difficulty with urination. The
presence of persistent ileus is often a telltale sign. On other occasions the
presentation may be more dramatic, including peritonitis or a pouch
cutaneous fistula. The fistula may also occur between the bladder, vagina,
or perineal skin. Occasionally it may exit at a previous drain site.
Presentation early in the postoperative course suggests a technical factor
such as when the posterior wall of the vagina is inadvertently caught in
the anterior portion of the IPAA staple line. We recommend that when
performing the anastomosis using a circular stapler, the trocar should
emerge posterior to the distal staple line to reduce the potential for vaginal
entrapment. If one has an index of suspicion for a fistula, then perform a
computed tomographic (CT) scan of the abdomen and pelvis using
intravenous, oral, and pouch contrast, providing renal function is satis-
factory. This may confirm the presence of a leak, but more importantly
will help identify if there is any undrained pelvic sepsis. The occurrence
394 Curr Probl Surg, May 2009
FIG 7. Techniques for ensuring adequate pouch reach. (Courtesy of the Art Department, Cleveland
Clinic, Cleveland, OH.)

of a pouch vaginal fistula carries a significant risk of pouch loss.42 Initial


management again consists of eradication of sepsis and may require the
insertion of draining setons for a considerable period of time. One should
only contemplate surgery once the inflammatory effects have attenuated.
On occasion a temporary ileostomy may be required to control the sepsis
and to restore the patient to a satisfactory quality of life before
considering intervention. If the internal opening of the fistula is accessible
from the perineum then a local repair in the form of a vaginal or pouch
advancement flap may be considered.43 A delayed presentation, typically
12 months after the reversal of the ileostomy or the indexed pouch
surgery, in the absence of complications should raise an index of
suspicion for underlying Crohns disease. One should carefully examine
for other manifestations of perianal Crohns disease including large,
edematous skin tags, painless fissures or those in atypical places, and anal
canal strictures (Fig 9). The presence of ulcers or inflammation within the
pouch may be due to pouchitis or the use of nonsteroidal anti-inflamma-
tory drugs (NSAIDs). However, the presence of ulcers within the
proximal small bowel or efferent limb of the pouch would be more
suggestive of Crohns disease.44 This is critically important since a
Curr Probl Surg, May 2009 395
FIG 8. Creation of an S pouch reservoir. (Courtesy of the Art Department, Cleveland Clinic,
Cleveland, OH.)

diagnosis of Crohns disease of the pouch will alter the management


options. Rather than offering an advancement flap or redo pouch for the
fistula one may consider the insertion of draining setons as a means of
controlling anorectal sepsis (Fig 10). This may then be combined with
396 Curr Probl Surg, May 2009
FIG 9. Elephant skin tags and external os of a fistula tract in a female with perianal Crohns disease.
(Courtesy of the Department of Colorectal Surgery, Cleveland Clinic, Cleveland, OH.)

Curr Probl Surg, May 2009 397


FIG 10. Pouch-vaginal fistula with a loose draining seton to control sepsis. (Courtesy of the Department
of Colorectal Surgery, Cleveland Clinic, Cleveland, OH.)

infliximab therapy.45,46 The patient should be counseled on the potential


need for a diverting ileostomy and the risk of pouch loss. If conservative
measures fail or the patients quality of life is significantly impaired then
one may consider surgical intervention. In patients with a pouch-vesical
fistula the communication is usually between the pouch and bladder
dome. Generally, disconnection of the fistulous communication between
the pouch and bladder, followed by bladder and pouch repair, omental
interposition, and a diverting ileostomy is associated with a successful
398 Curr Probl Surg, May 2009
outcome. However, if the origin of the fistula is from the anastomosis
between the pouch and anal transition zone, this usually requires pouch
disconnection with resection of the diseased segment and a new pouch-
anal anastomosis. The presentation of a pouch fistula beyond 12 months
is associated with a poor prognosis. Again, this may be attributed to the
potential underlying diagnosis of Crohns disease.47
Intestinal fistulas may occur due to complicated diverticular disease.
These may be internal or external. The patient typically develops a local
diverticular perforation with involvement of an adjacent organ.48 This
condition is seen in up to 20% of patients hospitalized with acute
diverticulitis. Most patients present in the elective setting with symptoms
attributed to the involved appendage. A small percentage may present in
a critical condition with malnutrition and sepsis due to smoldering
intra-abdominal infection. A colocutaneous fistula may also arise follow-
ing percutaneous drainage of a diverticular abscess. The presence of feces
in the percutaneous drain may create a sense of panic, with the patient and
family pressuring the surgeon to intervene. However, in the absence of
uncontrollable sepsis, surgical intervention may be planned electively
when the patients nutritional status and medical condition is opti-
mized.49,50 The fistula may also arise following an elective procedure for
diverticular disease. In either scenario the longer one leaves before the
ensuing intervention the easier the subsequent surgery should be. In some
instances the fistula may close with conservative management, especially
if there is no distal blockage, no intra-abdominal sepsis, and the patient
has had a true colorectal anastomosis. Our usual practice is to perform a
colonoscopy 6 weeks after resolution of diverticular symptoms to exclude
malignancy, which will determine whether one performs a standard
sigmoid colectomy or oncological resection.
Patients with colovesical fistulas may present with frank feces per
urethra, pneumaturia, or recurrent urinary tract infections. There is a
higher incidence in male patients attributed to the protective effect of the
uterus and broad ligament in women. In many cases, unless there is a very
large communication, it may be difficult to identify the fistulous tract
despite radiological and endoscopic investigations. In these cases one
may have to rely on clinical symptoms. The most sensitive diagnostic test
is the CT scan, which may report air in the bladder that should not be
present unless the patient has undergone recent instrumentation. It may
also identify the adjacent implicated bowel loop and determine whether
there is any intra-abdominal abscess. Barium enema is rarely diagnostic
since one cannot generate enough pressure to fill the fistulous tract.
Patients with a colovesical fistula should undergo a diagnostic colonos-
Curr Probl Surg, May 2009 399
copy and cystoscopy to ensure there is no associated colorectal malig-
nancy, bladder cancer, or undiagnosed Crohns disease, which would
influence the ensuing management.51 On cystoscopy the findings may be
subtle, including edema, erythema, and a bullous lesion in the bladder
mucosa. It is important to determine if the trigone of the bladder is
involved. In elderly patients with comorbidities one may adopt a
conservative approach, particularly where the operative risks outweigh
the extent of symptoms. They can be maintained on long-term antibiotics
and, contrary to perception, the potential for bacteremia and sepsis is
minimal.
The majority of patients diagnosed with a colovaginal fistula have
undergone a previous hysterectomy. Again the perforated diverticulum
may give rise to an inflammatory phlegmon and abscess, which becomes
adherent to the vault of the vagina. The patient typically complains of
feces per vagina. In some instances they may present with incontinence.
On examination the sphincters are intact and functional. It is the leakage
of stool from the fistulous tract that causes the sensation of incontinence.
Examination of the vagina is again imperative. It may identify an opening
in the vault, but more importantly it helps to rule out an associated
malignancy. Vaginoscopy allows one to assess the tissues and take
biopsies if required. Vaginography, in which a Foley catheter is inserted
into the vagina and filled with contrast, has a high sensitivity and will help
to determine if there are multiple tracts involved.
In the management of a fistula due to complicated diverticular
disease an abdominal approach is required. Regardless of whether this
is done using an open or laparoscopic technique, it is imperative that the
entire sigmoid colon is resected.48 This generally requires splenic flexure
mobilization to allow a tension-free anastomosis and resection to the
upper rectum ensures that no high pressure zone of distal sigmoid colon
is left in situ with the potential for recurrence. Anatomically the
confluence of the tenaie coli will help to identify the upper rectum. The
colon used for the colorectal anastomosis must be soft and without muscle
hypertrophy. For colovesical fistulas, generally no treatment of the
bladder is required once the adherent sigmoid loop is resected. If there is
a large fistula to the bladder then the edges may be trimmed and the
bladder defect closed in 2 layers. If one suspects that the area of
communication is in the region of the bladder trigone then involvement of
an urologist is strongly suggested. In complex diverticular communica-
tions one may also consider the insertion of ureteric stents, which will aid
in identification of the ureters and reduce the incidence of unrecognized
injury. Postoperatively, we will leave the Foley catheter in place for 5
400 Curr Probl Surg, May 2009
days with a cystogram before removal to ensure there is no urinary
leakage. We would also tend to leave a pelvic drain in situ until the Foley
catheter is removed.
An aortoenteric fistula (AEF) occurs when a communication arises
between the aorta and gastrointestinal tract. It may be a primary or
secondary event, is extremely rare, and usually fatal if not recognized
early.52 A primary AEF occurs when an abdominal aortic aneurysm
(AAA), either atherosclerotic or inflammatory, erodes in to the intestines.
This typically occurs between the abdominal aorta and the third or fourth
parts of the duodenum. Less common causes include tuberculosis,
radiotherapy, malignancy or foreign bodies.53,54 Secondary AEF gener-
ally occur following aortic reconstructive procedures.55 The presence of
an upper gastrointestinal bleed in a patient with a previous aortic
aneurysm repair, whether open or endovascular, should raise an index of
suspicion for an aortoenteric fistula. On upper endoscopy one may see the
aortic graft fistulizing through the abdominal wall. The patient will
typically present with massive hematemesis and hemorrhagic shock.
However, a percentage of patients may have a herald bleed, which is a
rapid large volume hemorrhage followed by a period of stabilization.
When diagnosed, this condition requires urgent definitive intervention.
The traditional repair consists of laparotomy, excision of the native graft,
ligation of the aorta, and creation of an extra-anatomic bypass. Many
centers are now using endovascular stents to cover the fistula. The risk of
graft infection is balanced against operating in a hostile abdomen with
dense adhesions. This technique allows stabilization of the patient with
definitive repair performed in a more elective, hemodynamically stable
setting.56
Gastrointestinal fistulas may arise following radiotherapy. Radiotherapy
as a primary or adjuvant therapy plays an important role in the
management of pelvic malignancies. Its tumorcidal ability in most cases
is dose-dependent. It is often associated with collateral damage to
adjacent organs. The small bowel, with its rapid cell turnover, is
particularly sensitive to its deleterious effects arising from DNA damage
and the generation of oxygen-free radicals. The risk of small intestine-
related complications is significant from 50 Gy and above and also
correlates with the volume of organ exposed. Early manifestations of
injury include diarrhea, cramping, and tenesmus. These symptoms are
most often self-limiting and not predictive of long-term complications.
Long-term radiotherapy-induced sequelae in relation to small bowel
injuries include malabsorption syndromes, fistulas, strictures, and perfo-
ration.57,58 These complications arise 6 to 12 months following radio-
Curr Probl Surg, May 2009 401
therapy and the underlying pathology is attributed to changes in the
microvasculature and the development of an obliterative vasculitis.
When operating on such patients one will often encounter dense adhe-
sions and a hostile abdomen. One must proceed with caution, particularly
if there is small bowel grossly adherent within the pelvis, in which case
it may be safer to bypass the affected segment rather than resect. If an
anastomosis is required it is best to avoid using the cecum and terminal
ileum, which most often suffers the most severe effects of pelvic
radiotherapy. Affected small bowel has a gray appearance; its mesentery
is thickened and it loses its mobility (Fig 11). Today we are more
conscious of excluding small bowel from the pelvic radiotherapy field. If
noninvasive methods are insufficient then one may consider the laparo-
scopic placement of a pelvic mesh, which will elevate the small bowel out
of the radiation field for the duration of therapy.59
Bevacizumab is a humanized monoclonal antibody that inhibits vascu-
lar endothelial growth factor (VEGF). It binds to the 2 VEGF receptors
(VEGF receptor-1 and VEGF receptor-2) found on the surface of vascular
endothelial cells. Vascular endothelial growth factor plays a central role
in angiogenesis, which drives tumor growth and the development of
metastasis. Clinical trials have demonstrated that bevacizumab will
augment the effects of chemotherapy in the treatment of metastatic
colorectal cancer. However, its effectiveness in inhibiting blood supply is
not limited to the tumor site and there have been recent reports of it
causing spontaneous gastrointestinal perforations and anastomotic leak-
ages in the postoperative period, leading to ECFs.60 In several cases the
fistulas developed several months to years after a previously unproblem-
atic anastomosis.61 These serious gastrointestinal complications have led
to a warning label from both the manufacturer and the FDA, identifying
these risks. The presence of diverticulitis, bowel obstruction, a history of
abdominal radiotherapy, and existing tumor on the bowel are additional
risk factors for perforation in patients receiving bevacizumab therapy.
Microthrombosis has been postulated as a reason for these effects. August
and colleagues postulated that in many cases patients may have an occult
leak, despite the absence of clinical signs and that the addition of
bevacizumab may lead to complete anastomotic dehiscence.61 Thus one
should delay elective gastrointestinal surgery until at least 4 weeks have
elapsed from the cessation of bevacizumab therapy.
The management of pancreatic, esophageal, or gastric fistula generally
comes under the care of surgeons specializing in upper gastrointestinal
pathology. A pancreatic fistula most commonly occurs following surgery
for pancreatic neoplasms, pancreatitis, and neuroendocrine tumors.62 It
402 Curr Probl Surg, May 2009
FIG 11. Small bowel affected by radiation enteritis with associated stricturing disease. (Courtesy of
the Department of Colorectal Surgery, Cleveland Clinic, Cleveland, OH.)

Curr Probl Surg, May 2009 403


may also occur due to blunt or penetrating abdominal trauma.63 It is a
difficult condition to manage and is associated with a prolonged hospital
stay and the potential for mortality. Although the true definition of a
fistula is a communication between 2 epithelized surfaces, the persistent
escape of pancreatic enzymes from a transected pancreatic surface or a
pancreatic-enteric anastomosis is referred to as a pancreatic fistula. In
most patients the diagnosis is established when the amylase content of a
high-output drain effluent is measured. The amylase content of the drain
is only considered significant if it is 3 times the upper normal serum
amylase value. The fistula is only considered clinically relevant if it is
associated with fever, leukocytosis, organ dysfunction, and the need for
drainage of an associated collection. Many authors refer to a biochemical
pancreatic fistula as one in which there is elevated amylase content in
the drain at the third postoperative day in an asymptomatic patient.64
These tend to resolve spontaneously. A grading system that differen-
tiates the severity has been proposed based on 9 clinical criteria. Grade A
fistulae are transient and characterized by an elevated drain amylase level
in an asymptomatic patient. Grade B fistulae are symptomatic and require
active clinical management, most often in the form of antibiotics,
nutrition, percutaneous drainage, and so forth. Grade C fistulae are those
associated with significant postoperative problems including organ dys-
function that requires major deviation from the planned postoperative
course, may necessitate reintervention, and are associated with significant
risk of morbidity and mortality.65
Pancreatic fistulas have been reported to communicate with the pulmo-
nary system.66,67 The underlying pathology is thought to be a pancreatic
pseudocyst arising following pancreatitis, which may decompress into the
pleura, giving rise to a pancreatico-pleural fistula. Alternatively, it may
decompress into the bronchial tree, giving rise to a pancreatico-bronchial
fistula. If one has an index of suspicion for this type of fistula then one
should analyze the amylase content of broncho-alveolar lavage. In most
patients, one would begin with conservative management with tube
thoracostomy insertion.68 Some authors advocate stenting of the pancre-
atic duct.
In patients with an advanced esophageal malignancy, a fistula may arise
between the mid esophagus and left mainstem bronchus. Less commonly
it may involve the trachea and right mainstem bronchus. The patient may
present with recurrent pneumonitis or a lung abscess. It is a very sinister
development and may arise de novo or following radiotherapy. The
mainstay of care is palliative to avoid ongoing pulmonary soiling, which
is extremely distressing for the patient and family. An esophageal stent
404 Curr Probl Surg, May 2009
would appear to offer the best palliation using the least invasive means.
By sealing the fistula it prevents respiratory contamination and restores
the ability to swallow, thus improving the patients quality of life and
survival period.69,70
Management Principles in Patients
with Intestinal Fistula
The initial management of affected patients is to ensure the eradication
and control of sepsis. In several studies this has been found to be the most
important determinant of outcome.2,8,11 The majority of deaths from
intestinal fistula are related to uncontrolled intra-abdominal sepsis, which
leads to increased catabolism, ongoing nutritional losses, and impaired
immune function. In this scenario the fistula will not close spontaneously
and the patient is ill equipped for the subsequent surgery. Any patient
with an intestinal fistula and evidence of organ dysfunction such as
cardiac, respiratory, or renal failure will most likely have an undrained
focus of sepsis. Very often these patients will not present with the typical
inflammatory response. Rather than the classical signs of infection they
present with significant weight loss, jaundice, and hypoalbuminemia.18,71
Operating on these patients in such a catabolic state may very quickly
lead to multi-organ failure and death.72 Thus one requires input from an
experienced radiologist, with the resources for frequent imaging and
percutaneous drainage of any collections, as clinically indicated. We
would have a low tolerance to repeat CT scans at appropriate intervals.
Percutaneous drainage using either CT scan or ultrasound guidance is the
least invasive means of draining any collection and avoids the consider-
able inflammatory second-hit response that would be generated by
surgical intervention.73 The initial catheter can be upsized if required. In
addition, contrast studies may be performed through the catheter (tubo-
gram) to confirm that the abscess cavity is decreasing in size. Although it
will not allow fistula closure, it can downsize a complex fistula to a
simpler one that makes elective resection more feasible. On occasion,
percutaneous drainage may not be feasible or may be inadequate. This can
occur when the abscess is surrounded by critical structures or is a
multi-loculated collection. In these cases open surgical drainage may be
required and can be guided by the CT findings.74
We would generally not give antibiotics for an abscess that is draining
adequately unless there is accompanying cellulitis. However, by the time
of referral to a tertiary unit, many patients with intestinal fistulas are
immunocompromised and have received several different courses of
antibiotics, all of which provide a fertile environment for opportunistic
Curr Probl Surg, May 2009 405
FIG 12. On occasion patients may come with horrendous abdominal wounds that are extremely
difficult to manage. (Courtesy of the Enterostomal Nursing Department, Cleveland Clinic, Cleveland,
OH.)

infections. In this patient group the potential for superimposed infection


with methicillin-resistant Staphylococcus aureus (MRSA) or fungal
infection is considerably higher. Most units will have an infectious
disease team that will aid one in making appropriate decisions on the need
for and type of antimicrobial therapy.
Skin protection is a critical part of the care pathway. The fistula effluent
may be acidic or alkaline and very quickly lead to skin excoriation.75 The
enzymes within the enteral succus may digest the abdominal wall leading
to an almost unmanageable wound (Fig 12), with the fistula exiting within
the center. Indeed, in a spontaneously occurring fistula, spreading
cellulitis may be the first manifestation of its presentation. The enteros-
tomal therapy team must be consulted immediately since appropriate
appliances and skin protection may prevent this downward spiral of skin
excoriation and resulting difficulties in getting containment of the fistula
effluent. One of the long-term complications of patients with a continent
ileostomy (K-pouch) is the development of a fistula between the pouch
and nipple valve (Fig 13). This leads to pouch incontinence and
excoriation of the surrounding skin. Insertion of a Waters tube with the
application of a face-plate is a means of controlling the effluent until
406 Curr Probl Surg, May 2009
FIG 13. Origin of a fistula tract (arrow) at the base of the nipple valve in a patient with a continent
ileostomy. (Courtesy of the Department of Colorectal Surgery, Cleveland Clinic, Cleveland, OH.)

pouch revision can be performed. A well-managed fistula ensures that the


patient has a good quality of life and may even be able to return to work
during the recovery period. In this regard patient education is important.
Patients, their partners, and caregivers should be taught how to change the
stoma and associated appliances. When the patient becomes comfortable with
changing the stoma appliance, this provides them with independence and
reduces anxiety levels when the appliance leaks and there is no one available
to help.76 It is important to be aware that the presence of a fistula discharging
malodorous body contents onto the skin leads to loss of self-esteem, poor
body image, anxiety, and depression. Prolonged hospital admission, partic-
Curr Probl Surg, May 2009 407
ularly for young patients, isolates them from their friends and removes them
from the home environment at a time that may be critical to their develop-
ment. Thus they may especially require psychological support.
Eradication of sepsis is performed in tandem with meeting the nutri-
tional needs of the patient. Ideally the patient would be fed using the oral
route, but if this is not possible then parenteral nutrition may be required.
Hypoalbuminemia is a significant risk factor for mortality.14 Fazio and
colleagues found that for ECF patients with an albumin less than 2.5 g/dL
the mortality rate was 42% in contrast to 0% mortality for patients with
albumin greater than 3.5 g/dL.77 Serum transferrin is also a strong
predictor of fistula-associated mortality.78 It is important that one assesses
early in the course of the illness whether the patients nutritional needs are
being met. Generally these patients are quite debilitated and catabolic at
the time of presentation. This malnourishment leads to an impaired
immune system with reduced ability to fight underlying and opportunistic
infections. These patients very quickly lose more than 10% of their body
weight and many at presentation look cachectic with the appearance of a
patient with metastatic cancer. As a rule, nutritional support is required
when the duration of illness is anticipated to be longer than 10 days. In
this patient subset the time to recovery may be weeks or months.
Malnutrition can led to depression, lethargy, and a poorly motivated
patient with little enthusiasm for recovery.
Whether enteral nutrition is appropriate depends on the nature of the
gastrointestinal fistula. In some patients without underlying sepsis and a
low-output fistula there may be no change in their basal metabolic rate.
Generally these patients are managed on the ward unless there is
associated multiorgan failure requiring supportive therapy. Markers of the
patients nutritional status include proteins with a short half-life such as
albumin, prealbumin, transferrin, and retinol binding protein. The majority of
hospitals managing such complex patients will have a nutritional team in
place that will be able to guide the patients caloric requirements. There must
be very close interaction between all teams to ensure that the appropriate
calories and nutrients are being administered to the patients since excess,
inappropriate nutrients may have damaging effects. The basal metabolic rate
accounts for the greatest energy expenditure and in normal individuals is
estimated at 25 kcal/kg per day. There are several formulas such as the
Harris-Benedict formula that allow precise calculations. Underlying sepsis is
a major cause of increased metabolism and these patients will require greater
nutritional support until the infection is eradicated.
As a rule, if the gastrointestinal tract functions, then nutrition should be
administered orally.79,80 With low-output fistula (200 mL/day), enteral
408 Curr Probl Surg, May 2009
nutrition is generally well tolerated. Very often, supplementation of the
diet with high-calorie supplements is sufficient to meet the calorie
requirements. In patients with dysphagia, whether due to neurological or
iatrogenic causes, insertion of a percutaneous endoscopic gastrostomy
tube (PEG)/jejunostomy tube with enteral nutrition is preferable to TPN.
In some cases one may encounter patients who require a small bowel
resection with a proximal jejunostomy and the orifice of the distal small
bowel brought to the surface as a mucus fistula. These patients may still
have a considerable amount of distal bowel in circuit. In this scenario one
may consider the administration of enteral nutrition via the distal limb of
the efferent bowel (fistuloclysis).81,82 More than 75 cm of distal small
bowel is required to facilitate absorption. The feeds are generally
administered at night, which allows the patient to pursue normal activities
throughout the day. The patient is also allowed oral intake, which will
help psychologically but will be of no nutritional value.
Understanding the underlying pathology and associated anatomy is the
next part of this management algorithm. The radiologic modality chosen
depends on the fistula type. Although a CT scan performed in the early
stages gives information on the nature of the fistula, its principal function
is to identify any intra-abdominal sepsis that may require drainage. With
its ability to delineate mucosal detail, barium is considered the ideal
medium to use when investigating the lumen of the gastrointestinal tract.
However, extravasation of barium within the thoracic or peritoneal cavity
can induce an intense inflammatory response. Thus, water-soluble con-
trast medium is considered safer, but its lower radiographic density means
it has less reliability in identifying small leaks or fistulous tracts. We find the
images provided by CT enterography to be of very high quality (Fig 14).
Fistulograms provide invaluable information in patients with a fistula
communicating with the skin. The study should only be performed once the
tract has matured. In addition to identifying the origin of the fistula it will help
determine if the gastrointestinal tract is in continuity. In some cases it may be
very useful for the surgeon to be present while the procedure is being
performed since one can interact with the radiologist to ensure the pertinent
information is obtained. Imaging (whether radiologic, endoscopic, or both) is
important so that one can rule out coexisting pathology before considering
definitive intervention. When operating on a patient with a fistula one should
always anticipate the possibility of intraoperative findings that do not
correlate with the preoperative radiographic findings.
Naturally the patient will be anxious to know the subsequent course of
action and management plan. Knowledge of the cause of the fistula,
associated anatomy, and presence or absence of sepsis will allow one to
Curr Probl Surg, May 2009 409
FIG 14. Computed tomographic enterography identifying an ileovesicular fistula with arrow pointing
to site of communication. (Courtesy of the Department of Colorectal Surgery & Radiology, Cleveland
Clinic, Cleveland, OH.)

be able to counsel the patient about the potential outcomes. Some fistulas
are more likely to close in response to conservative management than
others. This includes those arising from the esophagus, pancreas, jeju-
num, and duodenal stump, and fistulas arising from the lateral aspect of
the gastrointestinal tract that is in continuity.8,11,83 In favorable circum-
stances the general consensus is that 80% to 90% of fistulas will close
within 6 weeks. Less likely to close are gastric, ileal, and lateral duodenal
fistulas. Fistulas associated with inflammatory bowel disease, radiother-
apy, malignancy, the presence of a foreign body such as mesh, bowel
discontinuity, or a large adjacent abscess are unlikely to close of their own
accord.1,84 Also, fistulas involving multiple sites or organs are unlikely to
close without surgery. In this scenario one must plan an appropriate
management course with the patient, giving a potential time frame of
410 Curr Probl Surg, May 2009
greater than 6 months to subsequent reintervention. The patient and his
family must understand the difficult situation and have realistic goals. In
some cases one may be forced to intervene earlier than anticipated at
which time the surgeon is likely to encounter a hostile abdomen with
dense adhesions. This carries the risk of further enterotomies and fistulas. We
advise entering the abdomen away from the initial incision. This may be
above or below the initial laparotomy site or alternatively via a totally
separate incision. Rather than attacking the fistula head-on one should
identify and free up the afferent and efferent limbs. In optimal conditions the
fistula may be resected and a primary anastomosis performed, but in
suboptimal conditions or when the surrounding bowel is diseased then an end
stoma may be more prudent. In some patients the adhesions are so dense that
the only safe course of action is the creation of a very proximal stoma.
A subset of patients in addition to the fistula may have a small bowel
obstruction. If symptomatic this may require the insertion of a PEG tube to
decompress the small bowel until resolution of the obstruction or definitive
surgery.85 In many cases the small bowel obstruction will resolve once the
inflammatory process has settled. On occasions we have encountered
significant gastrointestinal bleeding in association with the fistula. If it is an
ECF, then the wound must be carefully examined. This is done with optimal
lighting. One may identify a bleeding point arising from the wound edges or
the abdomen as it undergoes granulation. One should consider the possibility
of a proximal gastrointestinal source and perform an upper endoscopy, if
indicated. If the bleeding is persistent, significant, and the source cannot be
identified, then mesenteric angiography may be both diagnostic and thera-
peutic. A laparotomy should only be considered as a last resort.
Everybody involved in patient care should understand the complex nature
of these cases. A large percentage of patients will require a prolonged
hospital course with numerous interventions. The triad of sepsis, malnutri-
tion, and electrolyte disturbance as reported in the 1960s are still the 3 major
determinants of mortality in patients who develop intestinal fistula today.12
Stoma Considerations in Patients
with Intestinal Fistula
When considering surgical intervention in a patient with intestinal
fistula whether in the elective or emergency setting, close coordination
with the ET team is important in achieving a successful outcome.
Although one may not be planning on diversion it is always prudent to
mark the patient for a stoma before any intervention.86 This will take into
account the intended incision site and location of the fistula and will
reduce the potential for a poorly positioned stoma that gives difficulty
Curr Probl Surg, May 2009 411
FIG 15. A recessed stoma may be impossible to pouch and require revisional surgery. (Courtesy of
the Enterostomal Nursing Department, Cleveland Clinic, Cleveland, OH.)

with pouching and may ultimately require revision (Fig 15). If the patient
has a fistula involving the mid-small bowel then it is likely that any
ensuing stoma will be placed within the upper quadrants. This is in
contrast to surgery for a colonic fistula in which case one could
reasonably assume that the diverting stoma will be formed in the lower
quadrants. However, one should anticipate all possibilities. Preoperative
marking allows one to take into account body habitus, where the patient
wears his belt, and the ability of the patient to see the stoma. This is
associated with improved patient satisfaction postoperatively. It is impor-
tant that one avoids placing the stoma within deep crevices, scars, or near
bony prominences or folds as this will create difficulties with pouching.
One should always take into account that the patient may require the
stoma for a considerable amount of time before reversal and that in a
percentage of cases it may be permanent. If the patient is confined to a
wheelchair then the optimal stoma site should be marked with them
sitting in the chair. On some occasions patients may have had so many
operations that it is difficult to identify a suitable site. In this scenario the
enterostomal therapist may mark nontraditional sites. We still follow the
mantra of placing the stoma through the rectus abdominis muscle
whenever possible since this is associated with a lower incidence of
412 Curr Probl Surg, May 2009
FIG 16. We ideally like to mark our stoma site with India ink to leave a permanent tattoo. (Courtesy
of the Enterostomal Nursing Department, Cleveland Clinic, Cleveland, OH.)

parastomal herniation.87 It is important that the abdominal wall opening


is of sufficient caliber to allow the stoma to be brought to the surface
without tension. Often in these cases the mesentery of the bowel is
extremely thickened and requires a considerable opening in the rectus
sheath to allow the identified bowel loop to be brought to the surface
without tension. As a rule we would prefer to manage a parastomal
hernia, in contrast to the ill effect of an ischemic stoma, in patients who
are already critically ill. In patients with increased abdominal girth a loop
end ileostomy may be required. In some situations it may be necessary to
bring the stoma through the laparotomy wound. We prefer to mark our
selected sites with India ink (Fig 16). This provides a permanent tattoo
and avoids the intended site being lost.
Octreotide and Gastrointestinal Fistula
In 1979, Klempa and colleagues reported on the beneficial effects of
somatostatin in reducing the complication rate following Whipple resec-
tion.88 The limitation of its short half-life (2 to 3 minutes) was overcome
with the development of a synthetic analog, octreotide, which had similar
pharmacological properties and a longer half-life (2 hours), allowing it to
be administered subcutaneously 3 times a day.89 They work by inhibiting
Curr Probl Surg, May 2009 413
both endocrine and exocrine pancreatic secretion and by decreasing
splanchnic blood flow. Additional effects include the inhibition of other
gastrointestinal hormones, gastrointestinal secretions, gallbladder empty-
ing, and gut motility. They have been typically used in conjunction with
TPN. A multicenter trial by Torres and colleagues reported that the
continuous intravenous infusion of somatostatin in combination with TPN
reduced the time to healing and associated morbidity in comparison to
TPN alone in patients with postoperative gastrointestinal fistulas.90
Subsequent prospective randomized, double-blinded, placebo-controlled
trials investigated the merits of octreotide in patients with postoperative
ECF. Only 1 of the studies demonstrated a beneficial effect reducing the
fistula output after 24 hours of treatment (53% reduction with octreotide
TPN vs 9% TPN placebo). However, the consensus of other trials by
Sancho and colleagues, and Scott and colleagues, was that octreotide,
when combined with TPN, did not aid fistula closure or reduce the closure
time.91,92
Overall, the general consensus is that they provide a modest benefit
only. In a subset of patients they may reduce the fistulous output and limit
the electrolyte disturbance, helping in wound management. If a beneficial
effect is seen it typically occurs within the first 48 hours. In our own
practice we use octreotide only to control fistula output when other
antimotility agents have failed.
Operative Intervention
Once the patient has been stabilized and a trial period of conservative
management has failed then one may have to plan a surgical course.
Generally, if the patient is free of sepsis, well nourished, and the fistula
has not closed spontaneously within 6 weeks then it is unlikely to do so
without operative intervention. Operative treatment is successful in more
than 80% of patients allowing restoration of gastrointestinal continu-
ity.1,19 Of critical importance when considering intervention is that one
delays the ensuing surgery until a minimum of 6 months has passed from the
prior surgery. This reduces the potential for encountering a frozen abdomen
due to dense adhesions. Fazio and colleagues have shown that operative
intervention between 10 days and 6 weeks carries a significantly higher
mortality rate in comparison to intervention outside this critical time period.77
Intervention within this period may result in multiple enterotomies and
damage to the mesentery of the bowel with resulting loss of significant bowel
segments. This may leave the patient with a short gut syndrome.
Exposure is critical. The majority of cases require a major laparotomy
via an incision extending from the xiphisternum to the symphysis pubis,
414 Curr Probl Surg, May 2009
FIG 17. Typical wound following the repair of multiple enterocutaneous fistulas with a diverting
ileostomy and Penrose drains (arrows) to aid drainage at the site of communication between fistulas
and skin. (Courtesy of the Department of Colorectal Surgery, Cleveland Clinic, Cleveland, OH.)

leaving the patient with a major wound and often several drains (Fig 17).
Care must be taken to avoid an inadvertent enterotomy at the time of
peritoneal entry.93 Thus one should enter the abdomen well removed from
existing fistula sites with particular care at the point of peritoneal entry as
the underlying bowel may be fused to the peritoneum and posterior rectus
Curr Probl Surg, May 2009 415
fascia. We recommend using a scalpel at this point. The majority of
enterotomies at re-laparotomy are made on reentering the abdomen. The
intra-abdominal anatomy must be clearly identified. Ureteric stents should
be considered if there is the possibility of retroperitoneal or pelvic
dissection. Although many surgeons lament the wasted time associated
with their insertion, in complex cases they greatly aid the identification of
the ureters. Many centers are now inserting ureteric stents selectively if at
the time of the laparotomy one encounters significant intra-abdominal
pathology that increases the risk of ureteric damage.94 The addition of a
ureteric injury to the already complex situation may be fatal. In these
cases one generally must mobilize the entire small bowel from the
ligament of Trietz to the ileocecal junction. This may require a consid-
erable amount of time. We occasionally use hydrodissection for densely
adherent bowel loops. If there is a segment of bowel that is particularly
diseased then it may have to be resected. If resection carries considerable
risks then the diseased segment may have to be bypassed. Some authors
suggest serosal patching and Roux-en-Y drainage of fistula defects, but
these are techniques we would not generally use. Primary closure of the
abdomen should be attempted. This may require extensive lateral dissec-
tion or fascia splitting incisions to allow midline approximation. If there
is a considerable abdominal defect then we would involve the plastic
surgery team to aid the abdominal wall reconstruction. They may use a
myocutaneous flap or other autologous tissue repair.95 If the tension is too
great and carries the risk of abdominal compartment syndrome then one
may have to consider mesh insertion. However, several studies have
reported significantly higher complication rates, including fistulization
and incisional hernia formation, when one uses prosthetic mesh in
comparison to native tissues. Connolly and colleagues reported on the
outcomes for 61 patients undergoing 63 operations to close ECF associ-
ated with open abdominal wounds. These were patients who had several
preceding laparotomies and ultimately required a laparostomy at the
initial presentation since abdominal closure was not feasible. Their
postoperative mortality rate was 4.8%, again highlighting that surgery in
this group is high risk. Their use of a biological mesh was strongly
associated with further fistulization and incisional hernia formation. The
refistulization rate was 11%. None of the 34 patients who had primary
closure of the abdomen developed fistula. This was in contrast to 7 of 29
(24.1%) reconstructed with a prosthetic mesh, with the highest rate of
refistulization in patients in whom a porcine collagen mesh was used.
They concluded that simultaneous reconstruction of the intestinal tract
and abdominal wall was associated with a high complication rate and that
416 Curr Probl Surg, May 2009
such patients should be managed in a specialized unit.96 However, in the
absence of a more favorable alternative most surgeons resort to using inert
materials such as Permacol or Alloderm if abdominal wall reconstruction is
required. Previous reports indicated that these biological materials appear to
offer greater resistance to infection despite the presence of a contaminated
field.97-99 The use of composite meshes is also quite popular. If mesh
insertion is required then we would advise mobilizing the omentum off the
transverse colon to a sufficient degree so that it may be interposed between
the abdominal wall mesh and underlying abdominal contents. When operat-
ing on such patients one must always be cognizant that the small bowel is a
limited resource and extensive resection may commit the patient to lifelong
TPN and its associated complications.
Management of Intraoperative Complications
It is imperative when undertaking operative intervention in this patient
group that one has an exit strategy in case of intraoperative difficulties.
This must be a key part of the preoperative discussion with the patient and
their family. Despite waiting a reasonable period of time following their
previous intervention, in a percentage of patients, tissues may be densely
adherent and result in multiple enterotomies.93 Loss of defined planes
may also result in damage to the mesentery with excessive bleeding. The
resulting injury may necessitate removal of an excessive amount of small
bowel and the potential for short bowel syndrome, condemning the patient to
lifelong TPN. These patients may also very quickly develop disseminated
intravascular coagulation leading to significant blood loss and multiorgan
failure. If this situation is encountered or anticipated it may be best to
abandon the laparotomy. If possible one may try and bring a more proximal
loop of bowel to the surface to make the downstream fistula more manage-
able in terms of wound care. This may be the first loop of jejunum distal to
the ligament of Trietz, which ensures that it is a proximal loop and not distal
to the existing fistula sites. Preoperatively, patients should be marked for an
ostomy site in all abdominal quadrants. This is done in conjunction with the
ET team and will improve patient satisfaction with the resulting stoma if
required. One can then make a plan for further intervention in 6 to 12 months
when the adhesions may be more manageable.
Laparoscopic Surgery for ECF
At one time the presence of a fistula was considered a contraindication
to pursuing a laparoscopic approach.48 However, there are now several
studies reporting the successful use of minimally invasive surgery in the
management of a subset of patients with colovaginal, colovesical, and
Curr Probl Surg, May 2009 417
Crohns-related fistula.100-103 However, before undertaking these com-
plex cases the surgeon must have a considerable amount of experience
with laparoscopy and, in particular, laparoscopic colorectal surgery.
Preoperatively patients should undergo endoscopy and radiological im-
aging as appropriate to rule out cancer. The first port is inserted under
direct vision at a site removed from previous scars and underlying
pathology. Great care should be taken since missed enterotomies are a
significant cause of morbidity in laparoscopic colorectal surgery.104 One
should have a low index to convert to an open approach in the presence
of dense adhesions, an inflammatory phlegmon, associated abscesses, or
failure to make progress in a timely fashion.105 When patients have a
diverticular fistula the affected segment of sigmoid colon is resected. This
generally requires splenic flexure mobilization and subsequent anastomo-
sis to the upper rectum. The site of bladder or vaginal fistulization is
typically small and does not require repair. However, the site of
fistulization and extent of the defect can be assessed by distending the
bladder with methylene blue diluted in saline. If a large defect is identified
it can be closed using intracorporeal suturing. Laparoscopic fistula
takedown can be assisted using an endoscopic stapling device.106 At the
end of the case the greater omentum is placed between the small or large
bowel anastomosis and bladder or vagina. The operative principles in the
laparoscopic approach are in keeping with open surgery. We would
particularly advise caution in patients with their fistula arising from small
bowel Crohns disease since the mesentery is often quite thick with a
significant risk for bleeding. Thus, once the bowel is sufficiently mobi-
lized it may be safer for the resection and ligation of mesenteric vessels
to be performed extracorporeally. In many cases the laparoscopic ap-
proach allows one to identify the area of pathology and then make a small
incision over this critical region, avoiding the need for a major laparot-
omy.
Novel Techniques in the Management
of Gastrointestinal Fistula
Proponents of novel methods are of the opinion that the techniques are
minimally invasive and associated with low morbidly and that, although
outcomes are variable and still under investigation, they provide an
attractive alternative option to a major laparotomy. Khairy and colleagues
reported on the use of percutaneous gelfoam embolization in the man-
agement of duodenal fistula.107 Lisle and colleagues subsequently re-
ported its use in the management of intestinal and colonic fistulas using
a percutaneous radiologically-guided method. An initial CT scan is
418 Curr Probl Surg, May 2009
required to ensure that there is no undrained sepsis. In addition, there
must be no distal obstruction or intra-abdominal foreign body such as
mesh. A fistulogram is then performed to outline the anatomy. This will
provide important information on the length, tortuosity, and the origin of
the tract. After the fistulogram the embolization procedure is performed
using fluoroscopic guidance. A guidewire is passed through the fistula
into the bowel lumen. An introducer sheath is then passed over the
guidewire until it lies at the point that the fistula tract exits the bowel
lumen. The guidewire and central introducer are then removed, leaving
the sheath. Gelfoam pledgets are then pushed down the sheath with the
introducer until they occlude the fistula. The sheath is then removed. They
reported its successful use in 3 patients with ECF.108
Kumar and colleagues reported on the endoscopic closure of fecal
colocutaneous fistulas using metal clips in 2 patients. One patient had a
fistula arising from a cecal perforation due to a stab injury and the second
patient developed a colocutaneous fistula following a left hemicolectomy
for an adenocarcinoma of the descending colon. The origin of this concept
developed with the successful closure of colonic perforations occurring
following colonoscopy using metal clips.109
There have also been some reports of ECF closure using fibrin glue. In
1996 Hwang and Chen reported on the successful use of fibrin glue in 6
patients with ECF.110 Gage and colleagues reported on the treatment of 3
ECFs arising in pancreas transplant recipients using percutaneous drain-
age and fibrin sealant. In all patients the anastomotic leak presented as an
intra-abdominal fluid collection that was converted to a controlled fistula
by percutaneous drain insertion. The technique itself consisted of insert-
ing a guidewire before removing the percutaneous drain. A 5-Fr Beren-
stein catheter was then advanced over the guidewire to the origin of the
fistula tract from the bowel and its position was confirmed by injection of
contrast. The tissue seal was then injected and the catheter slowly
removed until the entire tract was obliterated to the level of the skin.111
These were all low-output ECFs.
Aside from the use of fibrin glue we have no personal experience in the
use of these novel techniques, but the low associated morbidly and the
fact that one may avoid an operation does create some interest. However,
in the vast majority of ECF cases, a major laparotomy is required to deal
with the problem definitively.
Skin Grafting and VAC Dressing
An ECF that drains through a large open abdominal wound can lead to
major problems with fluid loss, soiling, and skin excoriation.7,83 Tradi-
Curr Probl Surg, May 2009 419
tionally these wounds were managed using large ostomy appliances often
connected to a sump system to draw off effluent in an attempt to reduce
skin maceration. In more recent years we are beginning to explore other
alternatives for wound management. Skin grafting is a useful adjunct in
these patients.112 If the graft is successful then it reduces fluid loss and
bacterial colonization of the wound. Before grafting the granulated
abdominal wound is debrided of any necrotic or purulent material, taking
care not to cause an inadvertent enterotomy. If successful it makes the
wound more manageable until the time of definitive fistula repair. One of
the dangers is that the fistula effluent may get under the skin graft and lift
it from the bed. Temporary intubation of the fistula lumen while the graft
is healing may reduce the incidence of this problem. Once the skin graft
takes, an ostomy bag can be applied to collect the fistulous output.113 In
our experience if one has a very large open abdominal wound and a skin
graft can be applied to 70% or 80% of the area then this will greatly aid
the patient and wound management. Contrary to ones instinct and the
presence of bacterial contamination a large percentage of these grafts will
take with little morbidity to the donor site. If a patient is a suitable
candidate for a skin graft then the involvement of a plastic surgery
colleague will aid in the decision making. The graft may or may not be
done in conjunction with the vacuum-assisted closure (VAC) system.114
The VAC system has revolutionized the management of open wounds
that previously may have taken months to heal. The device works by
applying a negative subatmospheric pressure.115 Since its introduction it
has gained widespread popularity and has been applied to the manage-
ment of a wide range of wounds in different anatomical locations.116-118
Reported success has been achieved for pilonidal, sacral, facial wounds,
and so forth. It consists of a polyurethrane foam dressing that is cut to the
required size and inserted onto the exposed wound. This is covered by a
clear dressing. Typically a negative pressure is applied at 125 mm Hg, but
this can be varied. It aids the removal of purulent material, reduces tissue
edema, and encourages angiogenesis with ensuing granulation. The
removal of purulent material reduces the bacterial count within the
wound. It has equal efficacy in adults and children. It is changed every 2
to 3 days. It has been found to be superior to the traditional management
of open abdominal wounds and reduces the nursing care involved. Costs
involved have been raised as an issue, but it significantly promotes time
to closure and thus in the long-term is more cost-effective than the use of
traditional saline-soaked gauze dressings. Also it allows one to discharge
a patient to community heath care, whereas traditionally the patient was
kept in the hospital until significant wound healing was achieved.
420 Curr Probl Surg, May 2009
One of the concerns with the use of the VAC dressing is that contact
with unprotected bowel may give rise to additional ECFs.119 We believe
that it has a role in the management of patients with large open wounds
and associated fistulas. However, the underlying fistula must be carefully
isolated from the VAC device and this requires a lot of care and expertise
from the wound care team (Fig 18). Some authors advocate if the patient
has an open wound plus a fistula then the VAC may be applied directly
to the fistula. Gunn and colleagues reported on its use in 15 patients with
ECFs. In 11 patients with no visible intestinal mucosa they achieved
closure with a mean time of 14 days. It was also successful in 1 patient
who had preceding neoadjuvant therapy.120 Although the series was small
they found no correlation between the volume of fistula output and the
potential for closure. In the 4 patients with exposed intestinal mucosa no
closure was achieved, which is not surprising since these fistulas are
generally well epithelialized and not amenable to conservative measures.
However, we would advise caution in this scenario and agree with the
sentiments of Fischer that the application of a VAC device directly onto
exposed bowel has the potential to lead to the development of further
fistula and a more difficult surgical problem to manage.121
Summary
Large series reporting outcomes in the management of complex
gastrointestinal fistulas have greatly improved our knowledge of this
subject. Patients with intestinal fistulas may present with a myriad of
symptoms and diagnostic difficulties. The initial principle of care is to
control and eradicate underlying sepsis. With radiologic advances this is
typically achieved with CT- or ultrasound-guided percutaneous drainage.
Most specialized centers will have interventional radiologists experienced
in the management of these complex cases. We would advocate, if the
patient has an open wound plus a fistula, an aggressive approach with
repeat scanning and insertion or upsizing of drains as clinically indicated.
On occasion open surgical drainage may be indicated due to overwhelm-
ing sepsis, lack of appropriate radiological resources, or where the septic
focus is in a location that is inaccessible or surrounded by structures that
may be damaged with a radiological approach. If the patient has an
external fistula then early involvement of the enterostomal therapist is
required to gain control of the wound, prevent skin excoriation, and to
provide a means of measuring the fistulous output. Often at this initial
stage the surgeon involved in the patients care comes under considerable
pressure from the family and patient to provide a quick resolution. When
the fistula is iatrogenic the feelings of guilt often lead surgeons to
Curr Probl Surg, May 2009 421
FIG 18. (A) Open abdominal wound with
enterocutaneous fistulae and a left sided
ileostomy (arrow). (B) The same patient with
a vacuum-assisted closure system applied
and the enterocutaneous fistula opening
carefully isolated. This helped to control the
abdominal wound until time for definitive
surgery. (Courtesy of the Enterostomal Nurs-
ing Department & Wound Care Team,
Cleveland Clinic, Cleveland, OH.)

422 Curr Probl Surg, May 2009


consider early operative intervention. In most cases this is inappropriate
and may lead to further fistulas when masterly inactivity with appro-
priate supportive care would have been the correct course of action. In
this situation a second opinion from a supportive colleague may be
valuable.
In most cases one will be able to decide relatively quickly whether
nutritional support is required. Knowledge of the anatomical origin of the
fistula, its etiology, and physiological output will help in this decision. If
the patient has sufficient gastrointestinal tract, then use the enteral route.
In the initial catabolic stage additional calories may be required above
the patients basic metabolic needs. Again most centers have experienced
nutritional personnel who will help assess the patients status and
determine if TPN is indicated. Total parenteral nutrition may be required
because of associated bowel obstruction, ongoing weight loss, and
high-output fistulas giving rise to electrolyte disturbances, nutritional
deficiencies, and difficulties with wound care. At the time of definitive
operative intervention a sufficient amount of time must be allocated to the
intended procedure. Division of dense adhesions may take a consid-
erable amount of time until one reaches the area of relevant pathology. If
the fistula arises from the small or large bowel then the segment of origin
will generally have to be excised and a primary anstomosis performed as
appropriate. Strong consideration should be given to creating a diverting
stoma. These patients most often have been very sick for a considerable
period of time and are ill equipped medically and psychologically to deal
with a further anastomotic problem. However, the decision to form a
stoma will be guided by the patients clinical status and operative
findings. In addition, if the fistula arises from the proximal bowel a stoma
to protect this will invariably commit the patient to a further 3 months of
TPN. Stoma reversal also carries the potential for complications.
The preceding monograph provides many guidelines and principles to
follow when dealing with intestinal fistula. However, the care of each
patient must be individualized according to their underlying pathology,
age, comorbidities, and response to treatment. These patients require a
significant amount of time and commitment from personnel involved in
their care. Advances in health care mean that people are living longer and
we are pushing the boundaries of surgical intervention. Therefore, there is
going to be an inevitable percentage of patients who will develop
intestinal fistula. The necessary experience required to deal with many of
these complex cases can only be acquired with exposure and time. A
multidisciplinary approach with input from radiology, psychiatry, enter-
Curr Probl Surg, May 2009 423
ostomal therapy, and nutritional therapy is critical to achieving a
successful outcome.

REFERENCES
1. Draus JM Jr, Huss SA, Harty NJ, Cheadle WG, Larson GM. Enterocutaneous
fistula: are treatments improving? Surgery 2006;140:570-6; discussion 576-8.
2. Bosscha K, Hulstaert PF, Visser MR, van Vroonhoven TJ, van der Werken C. Open
management of the abdomen and planned reoperations in severe bacterial perito-
nitis. Eur J Surg 2000;166:44-9.
3. Schein M. Intestinal fistulas and the open management of the septic abdomen. Arch
Surg 1990;125:1516-7.
4. Fischer JE. The management of high-output intestinal fistulas. Adv Surg 1975;9:
139-76.
5. Fischer JE. The pathophysiology of enterocutaneous fistulas. World J Surg 1983;
7:446-50.
6. Chintapatla S, Scott NA. Intestinal failure in complex gastrointestinal fistulae.
Nutrition 2002;18:991-6.
7. Berry SM, Fischer JE. Enterocutaneous fistulas. Curr Prob Surg 1994;31:469-566.
8. Berry SM, Fischer JE. Classification and pathophysiology of enterocutaneous
fistulas. Surg Clin North Am 1996;76:1009-18.
9. Evenson AR, Fischer JE. Current management of enterocutaneous fistula. J
Gastrointest Surg 2006;10:455-64.
10. Soeters PB, Ebeid AM, Fischer JE. Review of 404 patients with gastrointestinal
fistulas. Impact of parenteral nutrition. Ann Surg 1979;190:189-202.
11. Datta V, Windsor AC. Surgical management of enterocutaneous fistula. Br J Hosp
Med (Lond) 2007;68:28-31.
12. Edmunds LH Jr, Williams GM, Welch CE. External fistulas arising from the
gastro-intestinal tract. Ann Surg 1960;152:445-71.
13. Levy E, Frileux P, Cugnenc PH, Honiger J, Ollivier JM, Parc R. High-output
external fistulae of the small bowel: management with continuous enteral nutrition.
Br J Surg 1989;76:676-9.
14. Soeters PB, Fischer JE, Franklin C. One hundred and nineteen patients with
gastrointestinal fistulas. Arch Chir Neerl 1977;29:19-31.
15. Segar RJ, Bacon HE, Gennaro AR. Surgical management of enterocutaneous
fistulas of the small intestine and colon. Dis Colon Rectum 1968;11:69-73.
16. Felley C, Mottet C, Juillerat P, Pittet V, Froehlich F, Vader JP, et al. Fistulizing
Crohns disease. Digestion 2007;76:109-12.
17. Disa JJ, Goldberg NH, Carlton JM, Robertson BC, Slezak S. Restoring abdominal
wall integrity in contaminated tissue-deficient wounds using autologous fascia
grafts. Plast Reconstr Surg 1998;101:979-86.
18. Kaushal M, Carlson GL. Management of enterocutaneous fistulas. Clin Colon
Rectal Surg 2004;17:79-88.
19. Lynch AC, Delaney CP, Senagore AJ, Connor JT, Remzi FH, Fazio VW. Clinical
outcome and factors predictive of recurrence after enterocutaneous fistula surgery.
Ann Surg 2004;240:825-31.
20. Broe PJ, Bayless TM, Cameron JL. Crohns disease: are enteroenteral fistulas an
indication for surgery? Surgery 1982;91:249-53.
424 Curr Probl Surg, May 2009
21. Carter R. Percutaneous management of necrotizing pancreatitis. HPB (Oxford)
2007;9:235-9.
22. Frossard JL, Steer ML, Pastor CM. Acute pancreatitis. Lancet 2008;71:143-52.
23. Larvin M. Management of infected pancreatic necrosis. Curr Gastroenterol Rep
2008;10:107-14.
24. Bax TW, McNevin MS. The value of diverting loop ileostomy on the high-risk
colon and rectal anastomosis. Am J Surg 2007;193:585-7; discussion 587-8.
25. Remzi FH, Fazio VW, Gorgun E, Ooi BS, Hammel J, Preen M, et al. The outcome
after restorative proctocolectomy with or without defunctioning ileostomy. Dis
Colon Rectum 2006;49:470-7.
26. Tsuei BJ, Skinner JC, Bernard AC, Kearney PA, Boulanger BR. The open
peritoneal cavity: etiology correlates with the likelihood of fascial closure. Am Surg
2004;70:652-6.
27. Carbonell AM, Matthews BD, Dreau D, Foster M, Austin CE, Kercher KW, et al.
The susceptibility of prosthetic biomaterials to infection. Surg Endosc 2004;19:
430-5.
28. Friedmann R, Feldman H, Sonnenblick M. Misplacement of percutaneously
inserted gastrostomy tube into the colon: report of 6 cases and review of the
literature. J Parenter Enteral Nutr 2007;31:469-76.
29. Roozrokh HC, Ripepi A, Stahlfeld K. Gastrocolocutaneous fistula as a complication
of peg tube placement. Surg Endosc 2002;16:538-9.
30. Baraza W, Lee F, Brown S, Hurlstone DP. Combination endo-radiological
colorectal stenting: a prospective 5-year clinical evaluation. Colorectal Dis 2008
Apr 9 (Epub ahead of print).
31. Cerisoli C, Diez J, Gimenez M, Oria M, Pardo R, Pujato M. Implantation of
migrated biliary stents in the digestive tract. HPB 2003;5:180-2.
32. Rogart J, Greenwald A, Rossi F, Barrett P, Aslanian H. Aortoesophageal fistula
following polyflex stent placement for refractory benign esophageal stricture.
Endoscopy 2007;39:E321-2 (suppl 1).
33. Verbo A, Manno A, Mattana C, Pedretti G, Rizzo G, Coco C. Colonic perforation
due to ventricular-peritoneal shunt migration: a laparoscopic approach. Tech
Coloproctol 2006;10:65-6.
34. Killelea BK, Arkovitz MS. Perforated appendicitis presenting as appendicoumbili-
cal fistula. Pediatr Surg Int 2006;22:286-8.
35. Agrawal V, Prasad S. Appendico-cutaneous fistula: a diagnostic dilemma. Trop
Gastroenterol 2003;24:87-9.
36. Tekkis PP, Fazio VW, Remzi F, Heriot AG, Manilich E, Strong SA. Risk factors
associated with ileal pouch-related fistula following restorative proctocolectomy.
Br J Surg 2005;92:1270-6.
37. Tsujinaka S, Ruiz D, Wexner SD, Baig MK, Sands DR, Weiss EG, et al. Surgical
management of pouch-vaginal fistula after restorative proctocolectomy. J Am Coll
Surg 2006;202:912-8.
38. Remzi FH, Fazio VW, Gorgun E, Ooi BS, Hammel J, Preen M, et al. The outcome
after restorative proctocolectomy with or without defunctioning ileostomy. Dis
Colon Rectum 2006;49:470-7.
39. Ziv Y, Fazio VW, Church JM, Lavery IC, King TM, Ambrosetti P. Stapled ileal
pouch anal anastomoses are safer than handsewn anastomoses in patients with
ulcerative colitis. Am J Surg 1996;171:320-3.
Curr Probl Surg, May 2009 425
40. Sandborn WJ. Pouchitis and functional complications of the pelvic pouch. In: Fazio
VW, Church JM, Delaney CP, eds. Current Therapy in Colon and Rectal Surgery.
2nd ed. Philadelphia: Elsevier, Mosby, 2005:229-33.
41. Sagar PM, Dozois RR, Wolff BG, Kelly KA. Disconnection, pouch revision and
reconnection of the ileal pouch-anal anastomosis. Br J Surg 1996;83:1401-5.
42. Shah NS, Remzi F, Massmann A, Baixauli J, Fazio VW. Management and
treatment outcome of pouch-vaginal fistulas following restorative proctocolectomy.
Dis Colon Rectum 2003;46:911-7.
43. Fazio VW, Tjandra JJ. Pouch advancement and neoileoanal anastomosis for
anastomotic stricture and anovaginal fistula complicating restorative proctocolec-
tomy. Br J Surg 1992;79:694-6.
44. Shen B, Remzi FH, Lavery IC, Lashner BA, Fazio VW. A proposed classification
of ileal pouch disorders and associated complications after restorative proctocolec-
tomy. Clin Gastroenterol Hepatol 2008;6:145-58; quiz 124.
45. Colombel JF, Ricart E, Loftus EV Jr, Tremaine WJ, Young-Fadok T, Dozois EJ,
et al. Management of Crohns disease of the ileoanal pouch with infliximab. Am J
Gastroenterol 2003;98:2239-44.
46. Ricart E, Panaccione R, Loftus EV, Tremaine WJ, Sandborn WJ. Successful
management of Crohns disease of the ileoanal pouch with infliximab. Gastroen-
terology 1999;117:429-32.
47. Raval MJ, Schnitzler M, OConnor BI, Cohen Z, McLeod RS. Improved outcome
due to increased experience and individualized management of leaks after ileal
pouch-anal anastomosis. Ann Surg 2007;246:763-70.
48. Kockerling F, Schneider C, Reymond MA, Scheidbach H, Scheuerlein H, Konradt
J, et al. Laparoscopic resection of sigmoid diverticulitis. Results of a multicenter
study. Laparoscopic colorectal surgery study group. Surg Endosc 1999;13:567-71.
49. Szojda MM, Cuesta MA, Mulder CM, Felt-Bersma RJ. Management of divertic-
ulitis [review]. Aliment Pharmacol Ther 2007;26:67-76 (suppl 2).
50. Singh B, May K, Coltart I, Moore NR, Cunningham C. The long-term results of
percutaneous drainage of diverticular abscess. Ann R Coll Surg Engl 2008;90:
297-301.
51. Garcea G, Majid I, Sutton CD, Pattenden CJ, Thomas WM. Diagnosis and
management of colovesical fistulae; six-year experience of 90 consecutive cases.
Colorectal Dis 2006;8:347-52.
52. Roche-Nagle G, ODonnell DH, Brophy DP, Barry MC. Primary aortoenteric
fistula. Am J Surg 2008;195:506-7.
53. Tsai TJ, Yu HC, Lai KH, Lo GH, Hsu PI, Fu TY. Primary aortoduodenal fistula
caused by tuberculous aortitis presenting as recurrent massive gastrointestinal
bleeding. J Formos Med Assoc 2008;107:77-83.
54. Caes F, Vierendeels T, Welch W, Willems G. Aortocolic fistula caused by an
ingested chicken bone. Surgery 1988;103:481-3.
55. Bergqvist D, Bjorck M, Nyman R. Secondary aortoenteric fistula after endovascular
aortic interventions: a systematic literature review. J Vasc Interv Radiol 2008;19(2
Pt 1):163-5.
56. Danneels MI, Verhagen HJ, Teijink JA, Cuypers P, Nevelsteen A, Vermassen FE.
Endovascular repair for aorto-enteric fistula: a bridge too far or a bridge to surgery?
Eur J Vasc Endovasc Surg 2006;32:27-33.
57. Onodera H, Nagayama S, Mori A, Fujimoto A, Tachibana T, Yonenaga Y.
426 Curr Probl Surg, May 2009
Reappraisal of surgical treatment for radiation enteritis. World J Surg 2005;29:
459-63.
58. Sher ME, Bauer J. Radiation-induced enteropathy. Am J Gastroenterol 1990;85:
121-8.
59. Joyce M, Thirion P, Kiernan F, Byrnes C, Kelly P, Keane F, et al. Laparoscopic
pelvic sling placement facilitates optimum therapeutic radiotherapy delivery in the
management of pelvic malignancy. Eur J Surg Oncol 2008 Mar 19 (Epub ahead of
print).
60. Collins D, Ridgway PF, Winter DC, Fennelly D, Evoy D. Gastrointestinal
perforation in metastatic carcinoma: a complication of bevacizumab therapy. Eur
J Surg Oncol 2008 Apr 15 (Epub ahead of print).
61. August DA, Serrano D, Poplin E. Spontaneous, delayed colon and rectal
anastomotic complications associated with bevacizumab therapy. J Surg Oncol
2008;97:180-5.
62. Veillette G, Dominguez I, Ferrone C, Thayer SP, McGrath D, Warshaw AL, et al.
Implications and management of pancreatic fistulas following pancreatico-
duodenectomy: the Massachusetts General Hospital experience. Arch Surg 2008;
143:476-81.
63. Subramanian A, Dente CJ, Feliciano DV. The management of pancreatic trauma in
the modern era. Surg Clin North Am 2007;87:1515-32.
64. Kong J, Gananadha S, Hugh TJ, Samra JS. Pancreatoduodenectomy: role of drain
fluid analysis in the management of pancreatic fistula. ANZ J Surg 2008;78:240-4.
65. Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative
pancreatic fistula: an international study group (ISGPF) definition. Surgery 2005;
138:8-13.
66. Fujiwara T, Kamisawa T, Fujiwara J, Tu Y, Nakajima H, Egawa N. Pancreatico-
pleural fistula visualized by computed tomography scan combined with pancre-
atography. JOP 2006;7:230-3.
67. Oh YS, Edmundowicz SA, Jonnalagadda SS, Azar RR. Pancreaticopleural fistula:
report of two cases and review of the literature. Dig Dis Sci 2006;51:1-6.
68. Dhebri AR, Ferran N. Nonsurgical management of pancreaticopleural fistula. JOP
2005;6:152-61.
69. Balazs A, Kupcsulik PK, Galambos Z. Esophagorespiratory fistulas of tumorous
origin: non-operative management of 264 cases in a 20-year period. Eur J Cardio-
thorac Surg 2008 Aug 2 (Epub ahead of print).
70. Shin JH, Song HY, Ko GY, Lim JO, Yoon HK, Sung KB. Esophagorespiratory
fistula: long-term results of palliative treatment with covered expandable metallic
stents in 61 patients. Radiology 2004;232:252-9.
71. Carlson GL. Surgical management of intestinal failure. Proc Nutr Soc 2003;62:
711-8.
72. Fazio VW, Coutsoftides T, Steiger E. Factors influencing the outcome of treatment
of small bowel cutaneous fistula. World J Surg 1983;7:481-8.
73. Thomas HA. Radiologic investigation and treatment of gastrointestinal fistulas.
Surg Clin North Am 1996;76:1081-94.
74. Brolin RE, Nosher JL, Leiman S, Lee WS, Greco RS. Percutaneous catheter versus
open surgical drainage in the treatment of abdominal abscesses. Am Surg 1984;
50:102-8.
75. Metcalf C. Enterocutaneous fistulae. J Wound Care 1999;8:141-2.
Curr Probl Surg, May 2009 427
76. Metcalf C. Stoma care: empowering patients through teaching practical skills. Br J
Nurs 1999;8:593-600.
77. Fazio VW, Coutsoftides T, Steiger E. Factors influencing the outcome of treatment
of small bowel cutaneous fistula. World J Surg 1983;7:481-8.
78. Kuvshinoff BW, Brodish RJ, McFadden DW, Fischer JE. Serum transferrin as a
prognostic indicator of spontaneous closure and mortality in gastrointestinal
cutaneous fistulas. Ann Surg 1993;217:615-22; discussion 622-3.
79. Rombeau JL, Rolandelli RH. Enteral and parenteral nutrition in patients with
enteric fistulas and short bowel syndrome. Surg Clin North Am 1987;67:551-71.
80. Lloyd DA, Gabe SM, Windsor AC. Nutrition and management of enterocutaneous
fistula. Br J Surg 2006;93:1045-55.
81. Ham M, Horton K, Kaunitz J. Fistuloclysis: case report and literature review. Nutr
Clin Pract 2007;22:553-7.
82. Mettu SR. Fistuloclysis can replace parenteral feeding in the nutritional support of
patients with enterocutaneous fistula [correspondence]. Br J Surg 2004;91:1203.
83. Dionigi G, Dionigi R, Rovera F, Boni L, Padalino P, Minoja G, et al. Treatment of
high output entero-cutaneous fistulae associated with large abdominal wall defects:
single center experience. Int J Surg 2008;6:51-6.
84. Gonzalez-Pinto I, Gonzalez EM. Optimising the treatment of upper gastrointestinal
fistulae. Gut 2001;49:iv22-31 (suppl 4).
85. Hoffer EK, Cosgrove JM, Levin DQ, Herskowitz MM, Sclafani SJ. Radiologic
gastrojejunostomy and percutaneous endoscopic gastrostomy: a prospective, ran-
domized comparison. J Vasc Interv Radiol 1999;10:413-20.
86. Colwell JC, Gray M. Does preoperative teaching and stoma site marking affect
surgical outcomes in patients undergoing ostomy surgery? J Wound Ostomy
Continence Nurs 2007;34:492-6.
87. Israelsson LA. Parastomal hernias. Surg Clin North Am 2008;88:113-25.
88. Klempa I, Schwedes U, Usadel KH. Prevention of postoperative pancreatic
complications following duodenopancreatectomy using somatostatin. Chirurg
1979;50:427-31.
89. Beglinger C, Drewe J. Somatostatin and octreotide: physiological background and
pharmacological application. Digestion 1999;60:2-8 (suppl 2).
90. Torres AJ, Landa JI, Moreno-Azcoita M, Arguello JM, Silecchia G, Castro J, et al.
Somatostatin in the management of gastrointestinal fistulas. A multicenter trial.
Arch Surg 1992;127:97-9; discussion 100.
91. Scott NA, Finnegan S, Irving MH. Octreotide and postoperative enterocutaneous
fistulae: a controlled prospective study. Acta Gastroenterol Belg 1993;56:266-70.
92. Sancho JJ, di Costanzo J, Nubiola P, Larrad A, Beguiristain A, Roqueta F, et al.
Randomized double-blind placebo-controlled trial of early octreotide in patients
with postoperative enterocutaneous fistula. Br J Surg 1995;82:638-41.
93. van Goor H. Consequences and complications of peritoneal adhesions. Colorectal
Dis 2007;9:25-34 (suppl 2).
94. Pokala N, Delaney CP, Kiran RP, Bast J, Angermeier K, Fazio VW. A randomized
controlled trial comparing simultaneous intra-operative vs sequential prophylactic
ureteric catheter insertion in re-operative and complicated colorectal surgery. Int J
Colorectal Dis 2007;22:683-7.
95. de Vries Reilingh TS, Bodegom ME, van Goor H, Hartman EH, van der Wilt GJ,
428 Curr Probl Surg, May 2009
Bleichrodt RP. Autologous tissue repair of large abdominal wall defects. Br J Surg
2007;94:791-803.
96. Connolly PT, Teubner A, Lees NP, Anderson ID, Scott NA, Carlson GL. Outcome
of reconstructive surgery for intestinal fistula in the open abdomen. Ann Surg
2008;247:440-4.
97. Buinewicz B, Rosen B. Acellular cadaveric dermis (AlloDerm): a new alternative
for abdominal hernia repair. Ann Plast Surg 2004;52:188-94.
98. Kolker AR, Brown DJ, Redstone JS, Scarpinato VM, Wallack MK. Multilayer
reconstruction of abdominal wall defects with acellular dermal allograft (AlloDerm)
and component separation. Ann Plast Surg 2005;55:36-41; discussion 41-2.
99. Parker DM, Armstrong PJ, Frizzi JD, North JH Jr. Porcine dermal collagen
(Permacol) for abdominal wall reconstruction. Curr Surg 2006;63:255-8.
100. Pokala N, Delaney CP, Brady KM, Senagore AJ. Elective laparoscopic surgery for
benign internal enteric fistulas: a review of 43 cases. Surg Endosc 2005;19:222-5.
101. Young-Fadok TM, HallLong K, McConnell EJ, Gomez Rey G, Cabanela RL.
Advantages of laparoscopic resection for ileocolic Crohns disease. improved
outcomes and reduced costs. Surg Endosc 2001;15:450-4.
102. Zapletal C, Woeste G, Bechstein WO, Wullstein C. Laparoscopic sigmoid
resections for diverticulitis complicated by abscesses or fistulas. Int J Colorectal Dis
2007;22:1515-21.
103. Engledow AH, Pakzad F, Ward NJ, Arulampalam T, Motson RW. Laparoscopic
resection of diverticular fistulae: a 10-year experience. Colorectal Dis 2007;9:632-4.
104. Binenbaum SJ, Goldfarb MA. Inadvertent enterotomy in minimally invasive
abdominal surgery. JSLS 2006;10:336-40.
105. Rotholtz NA, Laporte M, Zanoni G, Bun ME, Aued L, Lencinas S, et al. Predictive
factors for conversion in laparoscopic colorectal surgery. Tech Coloproctol 2008;
12:27-31.
106. Watanabe M, Hasegawa H, Yamamoto S, Hibi T, Kitajima M. Successful
application of laparoscopic surgery to the treatment of Crohns disease with fistulas.
Dis Colon Rectum 2002;45:1057-61.
107. Khairy GE, al-Saigh A, Trincano NS, al-Smayer S, al-Damegh S. Percutaneous
obliteration of duodenal fistula. J R Coll Surg Edinb 2000;45:342-4.
108. Lisle DA, Hunter JC, Pollard CW, Borrowdale RC. Percutaneous gelfoam
embolization of chronic enterocutaneous fistulas: report of three cases. Dis Colon
Rectum 2007;50:251-6.
109. Kumar R, Naik S, Tiwari N, Sharma S, Varsheney S, Pruthi HS. Endoscopic
closure of fecal colo-cutaneous fistula by using metal clips. Surg Laparosc Endosc
Percutan Tech 2007;17:447-51.
110. Hwang TL, Chen MF. Randomized trial of fibrin tissue glue for low output
enterocutaneous fistula. Br J Surg 1996;83:112.
111. Gage EA, Jones GE, Powelson JA, Johnson MS, Goggins WC, Fridell JA. Treatment
of enterocutaneous fistula in pancreas transplant recipients using percutaneous drainage
and fibrin sealant: three case reports. Transplantation 2006;15:82:1238-40.
112. Dumanian GA, Llull R, Ramasastry SS, Greco RJ, Lotze MT, Edington H.
Postoperative abdominal wall defects with enterocutaneous fistulae. Am J Surg
1996;172:332-4.
113. Cheong EC, Ong WC, Lim TC, Lim J. Successful split-thickness skin grafting in
Curr Probl Surg, May 2009 429
a contaminated wound with an enterocutaneous fistula. Plast Reconstr Surg 2005;
115:1221-2.
114. Goverman J, Yelon JA, Platz JJ, Singson RC, Turcinovic M. The Fistula VAC,
a technique for management of enterocutaneous fistulae arising within the open
abdomen: report of 5 cases. J Trauma 2006;60:428-31; discussion 431.
115. Thompson JT, Marks MW. Negative pressure wound therapy. Clin Plast Surg
2007;34:673-84.
116. Delman KA, Johnstone PA. Vacuum-assisted closure for surgical wounds in
sarcoma. J Surg Oncol 2007;96:545-6.
117. Siegel HJ, Long JL, Watson KM, Fiveash JB. Vacuum-assisted closure for
radiation-associated wound complications. J Surg Oncol 2007;96:575-82.
118. Chen Y, Almeida AA, Mitnovetski S, Goldstein J, Lowe C, Smith JA. Managing
deep sternal wound infections with vacuum assisted closure device. ANZ J Surg
2008;78:333-6.
119. Rao M, Burke D, Finan PJ, Sagar PM. The use of vacuum-assisted closure of
abdominal wounds: a word of caution. Colorectal Dis 2007;9:266-8.
120. Gunn LA, Follmar KE, Wong MS, Lettieri SC, Levin LS, Erdmann D. Manage-
ment of enterocutaneous fistulas using negative-pressure dressings. Ann Plast Surg
2006;57:621-5.
121. Fischer JE. A cautionary note: the use of vacuum-assisted closure systems in the
treatment of gastrointestinal cutaneous fistula may be associated with higher
mortality from subsequent fistula development. Am J Surg 2008;196:1-2.

430 Curr Probl Surg, May 2009

You might also like