SURGICAL MANAGEMENT OF

GASTROINTESTINAL FISTULAS 0039-6109/96 $0.00 + .20

SURGICAL MANAGEMENT AND

TREATMENT OF PANCREATIC
FISTULAS
Michael G. Ridgeway, MD, and Bruce E. Stabile, MD

Pancreatic fistulas occur when the pancreatic duct or one of its branches is
disrupted either by direct trauma or as a result of inflammatory disease. They
may communicate externally with the skin or, less frequently, internally with a
variety of hollow organs or a body cavity. Complications associated with pancre-
atic fistulas are many and include sepsis, fluid and electrolyte losses, bleeding,
pulmonary problems, malabsorption, skin breakdown, and autodigestion or
erosion of adjacent viscera. There is a substantial mortality risk of 8% to 10%
associated with the development of a pancreatic f i ~ t u l a .Most
~ , ~ deaths are due
to intra-abdominal sepsis or hemorrhage.
The therapy for pancreatic fistulas has largely been conservative, with
operation being reserved for those patients with prolonged outputs or life-
threatening complications. Despite pharmacologic suppression of pancreatic exo-
crine secretion and advances in endoscopic and percutaneous therapeutic tech-
niques, pancreatic fistula continues to be a source of morbidity and mortality
following pancreatic surgery, splenectomy, pancreatic trauma, and pancreatitis.

CLASSIFICATION

Pancreatic fistulas have traditionally been classified as either internal or

external (Table 1).The vast majority of external fistulas occur as complications
of elective upper abdominal surgery. Postoperative fistulas are externalized
because of either (1) the placement of drains intraoperatively, (2) spontaneous
drainage via the wound, or (3) percutaneous drainage of postoperative fluid
collections. External pancreatic fistulas are further classified as either high output

From the Department of Surgery, Harbor-UCLA Medical Center (MGR, BES), and UCLA
School of Medicine (BES), Torrance, California

SURGICAL CLINICS OF NORTH AMERICA

-
VOLUME 76 * NUMBER 5 OCTOBER 1996 1159
1160 RIDGEWAY & STABILE

Table I.CLASSIFICATION OF PANCREATIC FISTULAS

Internal External
Pseudocyst Pancreaticocutaneous
Pancreaticoenteric
Pancreatic ascites
Pancreaticopleural
Pancreaticobronchial
Pancreaticovascular

or low output based on the total daily drainage volume. High-output fistulas
(>ZOO mL/day) are more problematic because of their association with pancre-
atic ductal abnormalities that may preclude closure with conservative therapy.
Low-output fistulas (<ZOO mL/day) are more amenable to nonoperative man-
agement.
Internal pancreatic fistulas are uncommon clinical entities and involve com-
munication of the pancreatic duct with an internal organ or space. The two most
frequently encountered variants are pancreatic ascites and pancreatic pleural
effusion. Pancreaticoenteric fistulas are unusual, and pancreaticobiliary and pan-
creaticovascular fistulas are exceedingly rare. The majority of internal pancreatic
fistulas occur as a result of pancreatitis or pancreatic trauma. Lipsett and Cam-
eronZ1recently described the Johns Hopkins Medical Center experience with
internal pancreatic fistulas, of which there were only 50 over a 27-year period.
The majority (82%)of the patients presented with pancreatic ascites. Of note is
that 42% of the patients had no antecedent history of pancreatitis.
Pancreatic pseudocyst constitutes a special type of internal pancreatic fistula
that is most often associated with pancreatitis. Peripancreatic inflammation is
thought to account for the confined nature of these fluid leaks. In cases in which
the wall of the pseudocyst ruptures, a classic internal fistula results. If the
pseudocyst is located anteriorly, rupture typically leads to pancreatic ascites,
whereas a posteriorly situated pseudocyst may rupture into the mediastinum
and/or pleural cavity, causing pancreatic pleural effusion. Therapeutic percuta-
neous drainage of a pancreatic pseudocyst creates an iatrogenic external pancre-
atic fistula if the pseudocyst is in continuity with the pancreatic ductal system
(Fig. 1).

ETIOLOGY AND PRESENTATION

Most pancreatic fistulas occur as complications of surgery (Table 2). They

are most often initially noted between postoperative days 2 and 7 and are
heralded by an increased surgical drain output of serous to cloudy fluid with a
high amylase content. Any upper abdominal invasive procedure involving or
near the pancreas places the patient at risk for fistula development. Operations
on the pancreas itself (including percutaneous pancreatic biopsy) carry signifi-
cant potential for development of pancreatic fistula. Pancreatic fistula compli-
cates 6% to 25% of elective pancreaticoduodenectomies and is the second most
common complication of the procedure (after delayed gastric em~tying).~, 7, 9, 11, 41

Distal pancreatectomy is associated with an even higher incidence of fistula

formation in one large series.” The reasons for this may be the lack of a
pancreaticoenteric anastomosis or imprecise closure of the pancreatic duct fol-
lowing distal pancreatectomy. Fistula formation has also been described follow-
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1161

Figure 1. Abdominal CT scans performed before (A) and after (B) percutaneous catheter
drainage of a large pancreatic pseudocyst. An external pancreatic fistula resulted.

ing gastric and biliary tract surgery, as well as after splenectomy, particularly
when the latter is performed for trauma.
Both blunt and penetrating pancreatic trauma have been associated with
fistula formation. The predisposing injury is usually detected at laparotomy,
with fistulas occurring either as a complication of pancreatic resection or as a
result of an undetected pancreatic ductal disruption, with pancreatic secretions
noted from the drains postoperatively. Blunt trauma to the pancreas usually
involves the body of the gland where it overlies the body of the second lumbar
vertebra (Fig. 2).
Pancreatitis is the other major cause of the development of pancreatic
fistulas. Pancreatic pseudocyst is the most common manifestation of this disor-
der and is an entity that has its own treatment algorithms. External fistulas
are frequent accompaniments to pancreatic dbbridement for severe necrotizing
pancreatitis. In two recent series, external fistulas developed in 23% and 29% of
patients undergoing pancreatic necro~ectomy.'~, 39 In one of these series, pancre-
1162 RIDGEWAY & STABILE

Table 2. ETIOLOGY OF PANCREATIC FISTULAS

Trauma
Iatrogenic
Pancreatic resection
Pancreaticoenteric anastomosis
Internal drainage of pseudocyst (cystoenteric anastomosis)
Pancreatic debridement and drainage
Pancreatic biopsy or aspiration
Inadvertent operative injury
Percutaneous catheter drainage of pseudocyst or fluid collection
Noniatrogenic
Penetrating injury
Blunt injury
Pancreatitis
Acute
Chronic
Pseudocyst
Neoplasm

atic ductal disruptions and external fistulas were more common when pancreatic
necrosis was associated with a peripancreatic fluid collection on CT scan.I4
Internal fistulas are frequently associated with chronic pancreatitis in which
there is disruption of the pancreatic duct, often in association with a ruptured
pseudocyst.26Pancreaticoenteric fistulas result when pseudocysts erode into
adjacent segments of the alimentary tract and most frequently involve the
stomach, duodenum, or transverse colon. Various uncommon to rare internal
pancreatic fistulas have been reported and include connections to the jejunum,
biliary tree, bronchial tree, splenic artery, portal vein, and other vessels (Fig. 3).

Figure 2. Abdominal CT scan showing a complete transection of the mid-body of the

pancreas with an associated retrogastric pseudocyst caused by blunt abdominal trauma.
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1163

Figure 3. Operative pancreatogram demonstrating a rare fistula to the portal vein.

DIAGNOSIS

Because most pancreatic fistulas are external and occur in the setting of
recent upper abdominal surgery, the diagnosis is usually straightforward. In-
creased peripancreatic drain output is the most common initial manifestation. If
the drainage is nonbilious and has a very high amylase content, the finding is
pathognomonic for a pancreatic fistula. However, the drainage may be bilious
if the fistula derives from a pancreaticoenteric anastomotic dehiscence through
which bile can also drain.
When postoperative pancreatic leaks are not adequately drained, problems
in diagnosis can arise. Such leaks are most often due to disruption of the
pancreaticojejunostomy following Whipple resection or inadequate closure of
the pancreatic duct following distal pancreatectomy, and the signs and symp-
toms may be subtle. This is particularly true when an iatrogenic pancreatic
injury occurs during an elective nonpancreatic operation, as the subsequent leak
is usually inadequately drained and rarely suspected by the surgeon. Increasing
abdominal pain, fever, tachycardia, leukocytosis, and delayed gastric emptying
are all nonspecific signs that should alert the surgeon to the possibility of a
pancreatic leak. CT scanning of the abdomen is the most accurate means of
detecting postoperative pancreatic fluid collections. Collections that are symp-
tomatic, enlarging, or suspected to be infected should undergo percutaneous
drainage for both diagnostic and therapeutic reasons. The presence of amylase-
rich fluid confirms the diagnosis of a pancreatic leak.
The diagnosis of pancreatic ascites should be suspected in any patient with
a history of pancreatitis and an acute onset of ascites. The diagnosis is confirmed
by paracentesis yielding fluid of high amylase content. Pancreatic pleural effu-
sion is a rare manifestation of internal pancreatic fistula that is sometimes
associated with concomitant pancreatic ascites and is confirmed by thoracentesis.
Pancreaticoenteric fistula should be suspected in the patient with active pancre-
1164 RIDGEWAY & STABILE

atitis or a pseudocyst who develops signs and symptoms of intra-abdominal

infection. Erosion of a peripancreatic or pseudocyst drain into the bowel can
also produce such a fistula if a pancreatic ductal disruption is already present.
If the colon is eroded, precipitous development of clinical sepsis may occur.
Pancreaticocolonic fistulas can also be associated with pancreatic exocrine insuf-
ficiency and significant fluid and electrolyte losses. However, pancreaticoenteric
fistulas are often asymptomatic, particularly when they involve the proximal
small bowel or stomach. The diagnosis of pancreaticoenteric fistula is made by
radiologic confirmation of pancreatic ductal communication with the bowel,
either by gastrointestinal (GI) contrast studies or by some form of pancreatic
ductography or fistulography.

IMAGING STUDIES

A number of imaging techniques have emerged as important adjuncts

in the detection and management of pancreatic fistulas. Ultrasonography and
especially CT scanning are very sensitive in detecting undrained postoperative
fluid collections and pancreatic pseudocysts. Fistulography (usually achieved
via contrast studies through drains) can demonstrate pancreatic ductal commu-
nications as well as the proximity of drains to the duct and to adjacent viscera
(Fig. 4). Upper GI tract contrast studies are sometimes helpful in demonstrating
the integrity of the anastomoses used in reconstruction after pancreaticoduode-
nectomy. Although the vast majority of postoperative pancreatic fistulas follow-
ing pancreaticoduodenectomy are related to breakdown of the pancreaticojejunal
anastomosis, drainage of amylase-rich fluid can also occur as a result of disrup-
tion of the gastrojejunostomy or hepaticojejunostomy.
The use of endoscopic retrograde cholangiopancreatography (ERCP) has
emerged as the most valuable means of assessing the patient with an ongoing
pancreatic leak, particularly when an operative intervention is being considered.

Figure 4. Pancreatic fistulogram performed by way of a percutaneous catheter drain

showing communication of a residual pseudocyst cavity with the pancreatic duct.
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1165

Figure 5. ERCP demonstrating pancreatic duct and a partially filled large fistula that
extended into the mediastinum, causing bilateral pleural effusions.

Pancreatic ductal anatomy is clearly defined by this procedure, and in cases of

persistent pancreatic fistula, it is usually the ductal anatomy that determines
further treatment. Often the exact anatomy of the fistulous tract is also defined
by ERCP (Fig. 5). Even more important is the identification of ductal obstruction
that defines those patients at high risk for continued fistula output (Fig. 6). This
information aids in planning resection when operative therapy is required.

TREATMENT

Nonoperative Management

Regardless of cause, several important principles are applicable to the treat-

ment of pancreatic fistula. Unimpeded free drainage of the fistula is essential.
In cases of internal pancreatic fistula, this may require repeated paracenteses,
thoracenteses, or tube thoracostomy. Any sepsis associated with the fistula must
be controlled for spontaneous closure to occur. Bowel rest is critical to achieving
the reduction of pancreatic exocrine secretion to basal levels that promote fistula
closure. Adequate nutritional support with total parenteral nutrition is crucial,
as many pancreatic fistulas require prolonged bowel rest.
Fluid and electrolyte losses are often substantial in cases of high-output
pancreatic fistulas. Excessive losses of bicarbonate lead to a metabolic acidosis
that must be corrected. Maintenance of adequate volume status is facilitated by
measuring fistula outputs carefully over a several-day period and replacing
fluid losses with isotonic crystalloid solution. Daily measurement of patient
weight is also helpful in determining total body fluid status, as insensible losses
may be high.
Skin care should be meticulous in patients with external pancreatic fistulas.
1166 RIDGEWAY & STABILE

Figure 6. ERCP showing complete obstruction of the pancreatic duct in a patient with a
catheter-drained pancreaticocutaneous fistula resistant to medical management.

Skin breakdown from exposure to pancreatic digestive enzymes can be avoided

if there is effective control of the fistula by means of closed suction drains. When
this is not possible, fluid collection appliances can be used to minimize skin
exposure to the fistula drainage. Consultation with an enterostomal therapist is
often helpful in this regard.
Approximately 80% of external pancreatic fistulas close spontaneously when
the principles of conservative management are meticulously The
success with internal fistulas has been somewhat less, with spontaneous closure
rates in the 40% to 60% 26 although anecdotal reports have occasionally
noted much higher closure rates.36Neither prophylactic nasogastric tube decom-
pression nor H, receptor antagonist therapy has been shown to have a significant
impact on results.
Problems associated with the conservative management of pancreatic fistu-
las include prolonged hospitalization and the high cost of care. Bacterial translo-
cation and sepsis associated with intestinal mucosal atrophy from prolonged
fasting contribute substantially to overall morbidity and cost. Since the introduc-
tion of clinically efficacious somatostatin analogues, it has become possible to
achieve spontaneous closure in a more expedient and cost-effective manner.

Somatostatin and Its Analogues

Somatostatin is a tetradecapeptide produced by a number of neuroendo-

crine tissues that has been shown to have a wide variety of inhibitory gastroin-
testinal effects.'y,2y
Its administration decreases gut motility and reduces splanch-
nic blood flow.2,2y Of particular interest to surgeons is its dramatic effect on
pancreatic secretion. Boden et a14 found that somatostatin had an inhibitory
effect on both basal and stimulated pancreatic exocrine secretion. It was this
property of somatostatin that led many investigators to the opinion that it
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1167

would have great efficacy in the treatment of established pancreatic fistulas.

Unfortunately, the very short half-life of somatostatin (<3 minutes) meant that
it needed to be administered as a continuous intravenous infusion in order to
be clinically effective. An early trial of intravenous somatostatin in the treatment
of external pancreatic fistulas conducted by Pederzoli et alZ7showed promising
results. In that trial the rate of fistula closure in the treatment group was no
higher, but there was a significant reduction in the time to closure, from a mean
of 32 days in the untreated group to 7 days in the group receiving somatostatin.
Octreotide acetate is a long-acting synthetic octapeptide analogue of somato-
statin with a half-life of approximately 2 hours.20It possesses properties similar
to somatostatin but has the advantage of a much longer serum half-life and thus
the ability to be administered subcutaneously. Its toxicity profile is essentially
the same as that of native somatostatin, and it has relatively few side effects.2,20,
32, 37 Pharmacologic studies confirm that octreotide dramatically reduces basal
and stimulated pancreatic secretion when given in doses as small as 50 k g twice
daily.2Since its introduction a number of case reports of its effectiveness in the
treatment of pancreatic fistulas have been published.', 28, 35
In 1988 Prinz et alZ8published a report of five cases of pancreatic fistula
treated with the new somatostatin analogue. The patients had a mixture of high-
and low-output chronic external pancreatic fistulas. Octreotide significantly de-
creased fistula outputs in all five patients and led to closure in four cases. One
patient's course was complicated by sepsis, nausea, and vomiting, and another
ultimately required operation to close the fistula. Other investigators have subse-
quently reported experiences with octreotide that have confirmed these early
results. Segal et aP6 reported decreased fistula outputs and closure in seven of
eight patients with high-output chronic external pancreatic fistulas after a mean
of 23 days. Importantly, an additional nine of ten patients with pancreatic ascites
from internal pancreatic fistulas experienced resolution with octreotide treatment
as well.
Randomized studies of the use of somatostatin or its analogues in pure
pancreatic fistulas are lacking because of the generally small numbers of patients
with the disorder. Several trials have investigated their use in GI fistulas, but
the results are difficult to interpret because data for enterocutaneous and pancre-
atic fistulas are often not separated and the numbers of patients with pure
pancreatic fistulas are small.31,38, 39 In the study of Tsiotos et al,39 9 of 14
pancreatic fistulas closed spontaneously, with octreotide being used only sporad-
ically in 7 patients. The authors noted that none of the 7 demonstrated any
response with regard to fistula output or time to closure. Torres et a138found a
significant reduction in the time to closure in a somatostatin-treated group, but
only seven patients with pancreatic fistulas were included in the study and no
separate data were provided for those patients. The recently reported study by
Sancho et aP1 was a randomized placebo-controlled trial of octreotide in the
treatment of GI fistulas that found no significant increase in the rate of fistula
closure or decrease in fistula outputs in the treated group. However, the number
of patients was small and there was a trend toward an improved closure rate in
the octreotide-treated group. There was a total of five pure pancreatic fistula
patients in the study, and all healed with or without octreotide therapy.
Despite the limited number of prospective studies reported, sufficient exper-
imental and case report data support the use of octreotide in patients with
established pancreatic fistulas. An initial dose of 50 kg administered subcutane-
ously three times per day is used, and the dose is titrated based upon fistula
output. The maximal dose employed is 200 kg three times daily. Careful mea-
surement of fistula output prior to the initiation of octreotide therapy is im-
1168 RIDGEWAY & STABILE

portant so that an accurate assessment of treatment response can be made.

Because of the high rate of closure of pancreatic fistulas with conventional
therapy alone, it is difficult to prove any benefit of octreotide therapy with
respect to closure rate. However, it seems clear that treatment with octreotide
significantly reduces fistula output and decreases the time to fistula closure.

Endoscopic Stenting

In cases in which fistula closure is not achieved with octreotide-based

conservative therapy, there has been an emerging role for ERCP to assess
pancreatic ductal anatomy. Strictures of the pancreatic duct are often found in
such cases and their definition serves to guide subsequent therapy.
Recently favorable results have been reported with the use of endoscopic
stenting of the pancreatic duct in cases of persistent pancreatic fistula resistant
to octreotide therapy. Transpapillary stents can bypass the high resistance of the
sphincter of Oddi, ductal strictures, and calculi, thereby reducing intraductal
pressure and the driving force behind the fistula. In theory, stents may also
block the ductal openings of some pancreatic fistulas (Fig. 7). Saeed et a130
reported the successful use of stenting in five patients with persistent fistulas
and ductal strictures in whom conservative therapy had failed. The stents were
routinely removed 6 weeks after placement, and no complications were reported.
Others have documented similarly encouraging results.'O, 17, 40 However, long-
term follow-up of these short-term successes is required before the appropriate
role of endoscopic pancreatic ductal stenting can be more accurately defined.
Pancreatic stents clearly have no efficacy in instances of complete ductal disconti-
nuity.

Figure 7. Transpapillaty stent in the pancreatic duct of a patient with chronic pancreatitis,
ductal stricture, and a pancreatic fistula. The fistula healed and did not recur following
stent removal.
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1169

Operative Management

Surgery for persistent pancreatic fistulas is indicated when nonoperative

management fails. In general, internal pancreatic fistulas require operative inter-
vention more often than external fistulas. The spontaneous closure rate for
internal pancreatic fistulas has been reported to be approximately 50% in the
larger series, but recent reports have shown a higher rate when octreotide
therapy was used.z1,36 Parekh and SegalZ6have suggested that patients with
internal pancreatic fistulas secondary to advanced pancreatic disease be selected
for early operative intervention, whereas those associated with mild to moderate
pancreatitis be managed medically.
The definition of failure of medical management can be difficult, however,
and the surgeon must consider a number of questions prior to embarking on a
course of operative intervention. Has the medical management been optimal? Is
there a clearly defined anatomic reason for the failure of the fistula to heal? Has
the patient's nutritional status been adequately addressed? Has sepsis been
successfully controlled? Is the patient fit for surgery? If these questions can be
answered in the affirmative and the fistula has not responded to a prolonged
trial of maximal medical therapy, operation should be performed.
The type of operation is dictated by the anatomy. Preoperative evaluation
with CT scanning and ERCP or fistulography to define ductal anatomy is crucial.
These studies identify the level of resection when this procedure is indicated.
Undrained pseudocysts can be demonstrated as well as the location and severity
of pancreatic ductal strictures. Intraoperative assessment of the pancreatic duct
to ensure patency of the retained segment is also prudent. This is particularly
true if preoperative studies have failed to clearly define the ductal anatomy. We
frequently employ intraoperative pancreatography using a small feeding tube
and water-soluble contrast to visualize the remaining pancreatic duct after
resecting the distal part of the gland.
The choice of operation for pancreatic fistula depends upon the site of
origin of the fistula and the presence and location of any strictures of the
pancreatic duct. Fistulas arising from the body and tail of the pancreas not
associated with ductal strictures in the head of the gland are managed by distal
pancreatectomy. If a large pseudocyst not amenable to resection is present, or if
a ductal stricture cannot be encompassed by the resection, then internal drainage
of the pseudocyst or the actual fistula must be performed (Fig. 8). Fistulas
arising from the head of the gland are, in general, also treated by an internal
drainage procedure, whether or not an associated pseudocyst or ductal stricture
is present. Pancreaticoduodenectomy or distal subtotal pancreatectomy opera-
tions are rarely justified for fistulas because of their higher morbidity and
mortality. Internal drainage is usually accomplished by use of a Roux-en-Y
pancreaticojejunostomy or cystojejunostomy.Closed suction drains are routinely
placed adjacent to a pancreaticojejunostomy or distal pancreatic resection margin
to capture any postoperative anastomotic or pancreatic stump leakage.
Operative management of persistent pancreatic fistula is generally safe and
effective, as has been confirmed by numerous investigator^.'^, 26 Ihse et all5
recently reported their experience in the surgical treatment of 13 patients with
persistent pancreatic fistulas resistant to medical management. Six patients had
external fistulas and the remaining 7 had internal fistulas. Eleven were treated
by pancreaticojejunostomy and 2 by distal pancreatectomy. There was no mortal-
ity, and no recurrent fistulas were noted postoperatively after a follow-up of 6
to 96 months.
1170 RIDGEWAY & STABILE

Figure 8. Chronic nonhealing pancreatic fistula, with plastic catheter inserted into fistula
lumen following excision of fibrous tract (A), and after completion of a Roux-en-Y
pancreaticojejunostomy (B).

PREVENTION

Technical Considerations

Because the majority of pancreatic fistulas are iatrogenic, occurring most

often after elective pancreatic resections, emphasis has been placed on prophy-
lactic measures that might decrease the incidence of this morbid and costly
complication. An accumulating data base from high-volume centers has sug-
gested that the experience of the operating surgeon may be of primary
importance.", 41 The results from these centers have documented progressively
declining rates of complications and mortality over time.
With varying success, a number of surgeons have attempted intraoperative
occlusion of the pancreatic duct with several materials in order to reduce the
rate of postoperative fistula formation.8,l8 Konishi et all8 used prolamine gel to
occlude the pancreatic duct in 51 consecutive cases of distal pancreatectomy
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1171

performed in concert with gastrectomy for gastric cancer, with no reported

postoperative fistulas. Of note, however, is that none of the patients had primary
pancreatic or periampullary lesions that are known to predispose to postopera-
tive fistula.42In a randomized prospective trial, D’Andrea et alMhad less success
with the use of a human fibrin sealant injected into the pancreatic ducts of
patients undergoing pancreatic resections for primary pancreatic disorders. No
significant difference in fistula occurrence was noted between the treated and
untreated groups.
A particular emphasis has been placed on the type and technique of pancre-
aticoenteric anastomosis performed following pancreaticoduodenectomy with
regard to subsequent fistula formation. Experimental studies in dogs have sug-
gested that a duct-to-mucosa technique of pancreaticoenterostomy is superior to
invagination of the transected end of the gland into the bowel and that stenting
is superior to no stenting.13 However, no prospective randomized trials in
humans have yet been performed. A large meta-analysis of 2684 cases suggested
that pancreaticogastrostomy was associated with lower rates of morbidity and
fistula formation than was pancreaticojejunostomy.3 However, in the only pro-
spective randomized trial comparing pancreaticogastrostomy to pancre-
aticojejunostomy, Yeo et a14’ showed no significant differences between the two
techniques when used to reconstruct patients undergoing pancreaticoduodenec-
tomy. Notably, this study demonstrated that the strongest predictors of postop-
erative fistula following pancreaticoduodenectomy were concomitant ampullary
or duodenal disease and inexperience of the operating surgeon.

Octreotide Prophylaxis

In addition to technical considerations, a number of investigators have

studied the use of prophylactic octreotide therapy as a means of decreasing the
rate of postoperative complications attending elective pancreatic operations.
There have been two prospective, placebo-controlled, randomized trials of pro-
phylactic octreotide that have shown significant reductions in complications
and, specifically, pancreatic fistulas in the treated groups.5.23 In both studies,
octreotide or placebo was administered in a double-blinded fashion beginning
either 1 or 1 day5 prior to surgery and continued for a period of 7 days
postoperatively. Both studies included all types of pancreatic resections such
that separate data regarding specific procedures are not available. Furthermore,
the rates of pancreatic fistula development among the treated patients in these
studies were 17.6% and 9%, respectively, which are higher than the rates re-
ported from several high-volume centers where prophylactic octreotide is not
routinely used.” This observation lends further support to the contention that
the best prevention for the development of the complication is meticulous
technique by an experienced pancreatic surgeon.

SUMMARY

Pancreatic fistulas most commonly derive as complications of elective surgi-

cal procedures on the pancreas and as sequelae of pancreatitis or pancreatic
trauma. The majority of external pancreatic fistulas can be managed nonopera-
tively, with an expected rate of closure exceeding 8OYO. Internal fistulas are
somewhat less likely to close with conservative measures alone. Octreotide has
been shown to significantly reduce fistula output and to hasten the closure of
1172 RIDGEWAY & STABILE

both internal and external pancreatic fistulas without affecting the overall rates
of closure. Operative therapy is reserved for the treatment of fistulas that do not
respond to conservative medical management. In randomized prospective trials,
prophylactic octreotide has been shown to reduce the morbidity of elective
pancreatic resections with respect to overall complication and fistula formation
rates. Surgical experience and technique appear to be the most important factors
in determining the overall complication rates following elective pancreatic sur-
gery.

References

1. Ahren 8, Tranberg KG, Bengmark S Treatment of pancreatic fistula with somatostatin

analogue SMS 201-995. Br J Surg 75:718, 1988
2. Anad BS, Goodgame R, Grahm DY: Pancreatic secretion in man: Effect of fasting,
drugs, pancreatic enzymes, and somatostatin. Am J Gastroenterol 89:267-270, 1994
3. Bartoli FG, Arnone GB, Ravera G, et al: Pancreatic fistula and relative mortality in
malignant disease after pancreaticoduodenectomy. Review and statistical meta-analy-
sis regarding 15 years of literature. Anticancer Res 11:1831-1848, 1991
4. Boden G, Sivitz MC, Owen OE: Somatostatin suppresses secretion and pancreatic
exocrine secretion. Science 190:163-165, 1975
5. Buchler M, Friess H, Klempa I, et al: Role of octreotide in the prevention of postopera-
tive complications following pancreatic resection. Am J Surg 163:125-131, 1992
6. Chan KL, Lau WY, Sung JY Octreotide therapy for pancreaticopleural fistula. J Gas-
troenterol Hepatol 9:530-532, 1994
7. Cullen JJ, Sarr MG, Ilstrup DM: Pancreatic anastomotic leak after pancreaticoduodenec-
tomy: Incidence, significance, and management. Am J Surg 168:295-298, 1994
8. D’Andrea AA, Costantino V, Speri C, et al: Human fibrin sealant in pancreatic surgery:
Is it useful in preventing fistulas? A prospective randomized study. Ital J Gastroenterol
26283-286, 1994
9. Edis AJ, Kiernan PD, Taylor WF: Attempted curative resection of ductal carcinoma of
the pancreas. Review of the Mayo Clinic experience, 1951-1975. Mayo Clin Proc
55:531-536, 1980
10. Fabre JM, Bauret P, Prudhomme M, et al: Posttraumatic pancreatic fistula cured by
endoprosthesis in the pancreatic duct. Am J Gastroenterol 90:804-806, 1995
11. Fernandez-del Castillo C, Rather DW, Warshaw AL Standards for pancreatic resection
in the 1990s. Arch Surg 130:295-299, 1995
12. Grauer L, Barkin JS: Role of somatostatin and octreotide in the treatment of pancreatic
pseudocyst, fistula and ascites. Digestion 55(Suppl 1):24-28, 1994
13. Green BS, Loubeau JM, Peoples JB, et al: Are pancreaticoenteric anastomoses improved
by duct-to-mucosa sutures? Am J Surg 161:45-50, 1991
14. Ho HS, Frey CF: Gastrointestinal and pancreatic complications associated with severe
pancreatitis. Arch Surg 1302317-822, 1995
15. Ihse I, Larsson J, Lindstrom E: Surgical management of pure pancreatic fistulas.
Hepatogastroenterology 41:271-275, 1994
16. Jenkins SA, Nott DM, Baxter J N Fluctuations in the secretion of pancreatic enzymes
between consecutive doses of octreotide: Implications for the management of fistulae.
Eur J Gastroenterol Hepatol 7:255-258, 1995
17. Kill J, Ranning H: Pancreatic fistula cured by an endoprosthesis in the pancreatic duct.
Br J Surg 80:1316-1317, 1993
18. Konishi T, Hiraishi M, Kubota K, et al: Segmental occlusion of the pancreatic duct
with prolamine to prevent fistula formation after distal pancreatectomy. Ann Surg
221:165-170, 1995
19. Krulich L, Dhariwal APS, McCann SM: Stimulatory and inhibitory effects of purified
hypothalamic extracts on growth hormone release from rat pituitary in vivo. Endocri-
nology 83:783-790, 1968
 SURGICAL MANAGEMENT AND TREATMENT OF PANCREATIC FISTULAS 1173

20. Kuta K, Nuesch E, Rosenthaler J: Pharmacokinetics of SMS 201-995 in healthy subjects.

Scand J Gastroenterol (Supp1)119:65-72, 1986
21. Lipsett PA, Cameron JL: Internal pancreatic fistula. Am J Surg 163:216-220, 1992
22. Maki HS, Kolts RL, Kuehner ME: Prevention of pancreatic fistula by modified
pancreaticojejunal anastomosis. Am J Surg 160:533-534, 1990
23. Montorsi M, Zago M, Mosca F, et a1 Efficacy of octreotide in the prevention of
pancreatic fistula after elective pancreatic resections: A prospective, controlled, ran-
domized clinical trial. Surgery 11726-31, 1995
24. Oktedalen 0, Nygaard K, Osnes M: Somatostatin in the treatment of pancreatic ascites.
Gastroenterology 99:1520-1521, 1990
25. Paran H, Neufeld D, Mayo A, et al: Preliminary report of a prospective randomized
study of octreotide in the treatment of severe acute pancreatitis. J Am Coll Surg
181:121-124, 1995
26. Parekh D, Segal I: Pancreatic ascites and effusion: Risk factors for failure of conserva-
tive therapy and the role of octreotide. Arch Surg 127707-712, 1992
27. Pederzoli P, Bassi C, Falconi M, et al: Conservative treatment of external pancreatic
fistulas with parenteral nutrition alone or in combination with continuous intravenous
infusion of somatostatin, glucagon or calcitonin. Surg Gynecol Obstet 163:428432,1986
28. Prinz RA, Pickleman J, Hoffman JP: Treatment of pancreatic cutaneous fistulas with a
somatostatin analog. Am J Surg 155:36-41, 1988
29. Reichlin S:Somatostatin, N Engl J Med 309:1556-1563, 1983
30. Saeed ZA, Ramirez FC, Hepps KS: Endoscopic stent placement for internal and
external pancreatic fistulas. Gastroenterology 105:1213-1217, 1993
31. Sancho JJ, DiCostanzo J, Nubiola P, et al: Randomized double-blind placebo-controlled
trial of early octreotide in patients with postoperative enterocutaneous fistula. Br J
Surg 82:638-641, 1995
32. Sargent AI, Overton CC, Kuwik RJ, et al: Octreotide-induced hyperkalemia. Pharmaco-
therapy 14:497-501, 1994
33. Satoh H, Hirohashi Y, Katano M, et al: Percutaneous pneumatic balloon dilation of the
obstructed pancreaticojejunal anastomosis in the management of a case of intractable
pancreatic cutaneous fistula. Gastroenterol Jpn 28:317-321, 1993
34. Saubanet A, Belghiti J, Panis Y, et al: Pancreaticogastrostomy after pancreatoduodenec-
tomy. Hepato-Pancreato-Biliary Surgery 6:91-95, 1992
35. Secchi A, KiCarlo V, Martinenghi S, et al: Octreotide administration in the treatment
of pancreatic fistulae after pancreas transplantation. Transplant Int 5201-204, 1992
36. Segal I, Parekh D, Lipschitz J, et al: Treatment of pancreatic ascites and external
pancreatic fistulas with a long-acting somatostatin analogue (Sandostatin). Digestion
54(S~ppl):53-58, 1993
37. Tauber MT, Harris AG, Rochiccioli P: Clinical use of the long-acting somatostatin
analogue octreotide in pediatrics. Eur J Pediatr 153:304-310, 1994
38. Torres AJ, Landa JI, Moreno-Azcoita M, et al: Somatostatin in the management of
gastrointestinal fistulas: A multicenter trial. Arch Surg 127:97-99, 1992
39. Tsiotos GG, Smith CDE, Sarr MG: Incidence and management of pancreatic and
enteric fistulas after surgical management of severe necrotizing pancreatitis. Arch Surg
130:48-52, 1995
40. Wolfsen HC, Kozarek RA, Ball TJ, et al: Pancreaticoenteric fistula: No longer a surgical
disease? J Clin Gastroenterol 14:117-121, 1992
41. Yeo CJ: Management of complications following pancreaticoduodenectomy. Surg Clin
North Am 75:913-924, 1995
42. Yeo CJ, Cameron JL, Maher MM, et al: A prospective randomized trial of pancreatico-
gastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy. Ann Surg
222:580-592, 1995
Address reprint requests to
Bruce E. Stabile, MD
Department of Surgery, Box 25
Harbor-UCLA Medical Center
1000 West Carson Street
Torrance, CA 90509