Langenbecks Arch Surg (2014) 399:801–810
DOI 10.1007/s00423-014-1242-2
REVIEW ARTICLE
Diagnosis and management of postoperative pancreatic fistula
Giuseppe Malleo & Alessandra Pulvirenti &
Giovanni Marchegiani & Giovanni Butturini &
Roberto Salvia & Claudio Bassi
Received: 4 August 2014 / Accepted: 11 August 2014 / Published online: 31 August 2014
# Springer-Verlag Berlin Heidelberg 2014
Abstract
Background Postoperative pancreatic fistula (POPF) is the
leading complication after partial pancreatic resection and is
associated with increased length of hospital stay and resource
utilization. The introduction of a common definition in 2005
by the International Study Group of Pancreatic Surgery
(ISGPS), which has been since employed in the vast majority
of reports, has allowed a reliable comparison of surgical
results. Despite the systematic investigation of risk factors
and of surgical techniques, the incidence of POPF did not
change in recent years, whereas the associated mortality has
decreased.
Purpose The purposes of this review article were to summa-
rize the current evidence on the diagnosis and management
strategies of POPF and to provide a concise reference for the
practicing surgeons and physicians.
Conclusion The high incidence of POPF was accompanied
by a shift from operative to non-operative management. How-
ever, the current management strategy is driven by the pa-
tient’s condition and local expertise and is generally based on
poor evidence. A randomized trial showed that enteral nutri-
tion is superior to total parenteral nutrition, and pooled data of
randomized trials failed to show any advantage of somatostat-
in analogs for accelerating fistula closure. The choice of
percutaneous versus endoscopic drainage of peripancreatic
collections remains arbitrary, and—when re-operation is need-
ed—there are very few comparative data regarding local
drainage with or without main pancreatic stenting as opposed
to anastomotic revision or salvage re-anastomosis. The con-
tinuous development of specialist, high-volume units with
G. Malleo (*) : A. Pulvirenti : G. Marchegiani : G. Butturini :
R. Salvia : C. Bassi
Unit of Surgery B, The Pancreas Institute, Department of Surgery,
G.B. Rossi Hospital, University of Verona Hospital Trust,
P.Le L.A. Scuro 10, 37134 Verona, Italy
e-mail: giuseppe.malleo@ospedaleuniverona.it
appropriate resources and multidisciplinary experience in
complication management might further improve the evi-
dence and the outcomes.
Keywords Postoperative pancreatic fistula .
Pancreaticoduodenectomy . Distal pancreatectomy .
Image-guided drainage . Postoperative complications
Introduction
Postoperative pancreatic fistula (POPF) remains the most
common major complication after partial pancreatic resection
and is the principal matter of concern among pancreatic sur-
geons. The economic aspect is not secondary, because POPF
leads to prolonged hospital stay and substantial resource uti-
lization [1]. As recently pointed out by Matthews, “in decades
past, a postoperative leak after pancreatic surgery was a
dreaded complication that often spelled disaster” [2]. Fortu-
nately, in current practice, the majority of POPF will resolve
spontaneously with conservative therapy, which includes
drainage of exocrine secretions and/or peripancreatic collec-
tions, with restriction of oral intake and nutritional support [3,
4]. In few cases, re-operation may be required either because
of inaccessibility of infected collections to percutaneous or
endoscopic drainage, or because of clinical instability associ-
ated with uncontrolled sepsis and multi-organ failure. This
latter picture is the risk factor most closely associated with
postoperative mortality, even in highly specialized centers [5].
The consensus definition of POPF, emanated in 2005 by
the International Study Group of Pancreatic Fistula, has been
since employed in most of the studies investigating outcome
measures in pancreatic surgery, thereby allowing a reliable
comparison among different experiences [6]. A recent pooled
analysis of studies employing the International Study Group
of Pancreatic Fistula (ISGPF) definition and including more
802
than 100 patients showed that the incidence of POPF follow-
ing pancreaticoduodenectomy (PD) was between 22 and
26 %, whereas in distal pancreatectomy, it was in excess of
30 % [6]. The highest rate of POPF follows middle-segment
pancreatectomy, which ranges from 20 to 60 %, because of the
creation of two pancreatic remnants and thus of two potential
sites for fistula formation [7].
This review article summarizes the current evidence on the
diagnosis and management of POPF, with the aim to provide a
concise reference for the practicing surgeons and physicians.
Diagnosis of postoperative pancreatic fistula
Definition and grading
According to the type of pancreatic resection, POPF repre-
sents failure of healing/sealing of a pancreatic-enteric anasto-
mosis (in pancreaticoduodenectomy, middle-segment pancre-
atectomy, or—when an anastomosis is fashioned—in left
pancreatectomy), or a parenchymal leak originating from the
raw pancreatic surface after left pancreatectomy, middle-
segment pancreatectomy (at the head stump level), enucle-
ation, or trauma. The primary diagnostic criterion, according
to the International Study Group of Pancreatic Surgery
(ISGPS) definition, is the output via an operatively placed
drain or a subsequently placed percutaneous drain of any
measurable volume of drain fluid on or after postoperative
day 3, with an amylase content greater than three times the
upper normal serum value [8]. The appearance of drain fluid
might be heterogeneous, depending on the presence of a
pancreo-enteric anastomosis and of a biliary anastomosis in
the same enteric loop, as well as on local conditions (e.g.,
presence of a sterile or infected collection). In particular, the
effluent may vary from a dark brown to greenish bilious fluid,
to milky water to clear “spring water” that looks like pancre-
atic juice (Fig. 1). Much effort has been spent to find early
Fig. 1 Different nuances of
pancreatic fistula drain effluent
Langenbecks Arch Surg (2014) 399:801–810
predictors of POPF, and different “fistula scores” based on the
combination of relevant risk factors have been proposed. This
aspect is beyond the scope of this review and has been
addressed in a tandem paper by McMillan and Vollmer [9].
The ISGPF definition was further compounded by the concept
of clinical relevance, to distinguish asymptomatic biochemical
fistulae from those that are associated with clinical illness,
therapeutic intervention, or death. The clinical grading system
(grade A, B, C) is based on nine criteria, including clinical
conditions, specific treatment, ultrasound or computed tomog-
raphy findings, persistent drainage, death related to POPF,
signs of infections, sepsis, and re-admission (Table 1). Indeed,
the data necessary for the grading system cannot be analyzed
before the end of the course of patient care and, therefore, by
definition, cannot be of any predictive value. Criticisms to the
ISGPF grading system rely on its potentially limited practical
relevance. Grade A POPF is of little consequence, and the
distinction between a grade B and C fistula, both of which are
clinically relevant, may be somewhat artificial [2]. Further-
more, POPF suffers from the inability to be compared quan-
titatively. In a North American study, the utilization of a
postoperative morbidity index, based on the Modified Accor-
dion Severity Grading System, showed that each grade of
POPF segregated into discrete Accordion profiles. Clinically
relevant fistulae (grade B–C) usually reflect the patient’s
highest Accordion Score, whereas biochemical POPFs are
often superseded by more severe complications [10].
Clinical picture
Besides the effluent appearance and the measurement of am-
ylase value in drains, which are the mainstays to establish the
diagnosis, the suspicion of POPF begins whenever there is a
deviation in the normal clinical course of a patient who has
just undergone partial pancreatic resection. Non-specific
symptoms include unexpected abdominal discomfort, nausea,
Langenbecks Arch Surg (2014) 399:801–810
Table 1 Parameters for postop-
erative pancreatic fistula grading Parameter
Clinical conditions
Specific treatment*
Reproduced from Bassi et al. [8]
US/CT (if obtained)
POPF postoperative pancreatic
Persistent drainage (after 3 weeks)**
fistula
a Re-operation
Partial (peripheral) or total par-
enteral nutrition, antibiotics, en- Death related to POPF
teral nutrition, somatostatin ana- Signs of infections
logs, and/or minimal invasive Sepsis
drainage
b
Readmission
With or without a drain in situ
abdominal distension with impaired bowel function, or even
overt abdominal pain with tenderness. Localized abdominal or
wound erythema, warmth, or swelling indicates spontaneous
drainage of pancreatic juices through the abdominal wall and
may result in infected wound dehiscence. Delayed gastric
emptying (as defined by the ISGPS paper) is often secondary
to a POPF with a peripancreatic fluid collection [11]. Fever
(>38 °C), increased serum leukocyte count, and increased C-
reactive protein levels may be also present, indicating a sys-
temic inflammatory response and a concomitant infection.
The majority of patients with POPF manifest with immediate-
ly obvious presentations after the index operation. However, a
subgroup of patients presents with latent POPF. As pointed
out by Pratt et al., latent fistulae differ in that they lack an
amylase-rich fluid on initial measurement, yet ultimately ex-
hibit (either acutely or subacutely) the clinical and radiologic
findings indicative of fistula. Subsequent measurement of
amylase activity obtained from drains, wound dehiscence (if
present), or operative samples confirms the diagnosis of
POPF. Advanced age and a small pancreatic duct were the
only predictors of latent fistula. Latent presentations had two
times the infection rate of evident fistulas, required more
aggressive interventions, resulted in longer hospitalizations,
and incurred greater hospital costs [12].
Radiology
Based on the ISGPF definition, radiologic documentation is
neither mandatory nor necessarily recommended for POPF
diagnosis. In the past, different authors did not recommend
routine imaging, especially in patients devoid of clinical con-
cern, because of the high prevalence of transient intra-
abdominal fluid collections in the pancreatic bed or adjacent
to a pancreatic-enteric anastomosis. On the contrary, cross-
sectional imaging is widely used in case of a deviation from
the normal course, especially for interventional purposes (e.g.,
guidance of percutaneous drainage placement). Although the
evidence in this field is limited, the few recent studies avail-
able suggested that fluid collections around the
803
Grade A Grade B Grade C
Well Often well Ill appearing/bad
No Yes/no Yes
Negative Negative/positive Positive
No Usually yes Yes
No No Yes
No No Possibly yes
No Yes Yes
No No Yes
No Yes/no Yes/no
pancreaticojejunostomy site or in the pancreatic bed were
significantly correlated with POPF and that—under postoper-
ative conditions without a POPF—a fluid collection does not
usually accumulate in these areas [13, 14]. In a study on the
accuracy of routine contrast-enhanced CT for the diagnosis of
POPF, the sensitivity was 63 %, and the specificity was 83 %
[15]. Contrast-enhanced CT can differentiate between collec-
tions in contiguity with the pancreatic duct or the anastomotic
suture line and other collections, according to site (superior
recess of lesser sac, subhepatic space, paracolic gutters, root of
mesentery). Furthermore, the radiologist can determine the
number and the imaging features of the collections (shape,
size, attenuation, homogeneity, wall enhancement), and the
possible disruption of pancreaticojejunostomy or
pancreaticogastrostomy (Fig. 2). The presence of air bubbles
in a peripancreatic collection correlates well with a clinical
and biochemical diagnosis of POPF [13]. Finally, contrast-
enhanced CT visualizes acute pancreatitis of the pancreatic
stump and the characteristics of the pancreatic duct. A poten-
tial limitation of CT is represented by false-positive diagnoses
of pancreatic fistula. However, it has been found that in PD,
true-positive and false-positive perianastomotic collections
behaved differently. False-positive collections were smaller
and never contained air bubbles and—most importantly—
disappeared rapidly at follow-up CT in all cases, unlike
perianastomotic collections due to fistula [15].
According to the authors’ experience, conventional
fistulography is a useful examination for the diagnosis of
POPF, especially following PD. Fistulography is a cost-
effective and readily available dynamic study with a low
radiation dose. Twenty-five milliliters of water-soluble con-
trast agent is injected through the surgically placed drain from
which “sinister” effluent was detected. The trajectory of the
contrast agent is dynamically followed during fluoroscopy,
and fistulograms are analyzed to determine the position of the
tip of the drain, the presence or absence of a communication
with a jejunal loop or the stomach, a fistulous tract, fluid
collections, or a communication with the main pancreatic duct
(Fig. 3). When fistulography depicts a primary filling with
804
Fig. 2 a Pancreaticojejunostomy dehiscence. Hypodense gap between
the pancreati c remnant a nd t he jejunal loop ( arrow). b
Pancreaticogastrostomy dehiscence. Hypodense gap between the
contrast agent of the jejunal loop or stomach, amylase-rich
fluid in the drain leaks directly from the gastric or intestinal
lumen, probably due to the erosion and migration of the drain
tube into the anastomosis. The dislocation of the drain into the
anastomosed loop has important consequences for the patient
treatment, because a 3–5-cm retraction of the drainage tube
frequently allows a complete recovery of the clinical compli-
cation within few days [16].
Management of pancreatic fistula
Nutritional support
Most of POPF patients are at increased catabolic processes
and basal energy expenditure. Furthermore, high-output fistu-
lae (those producing >200 mL of exocrine secretion daily) are
associated with fluid, electrolyte imbalances, and nutritional
depletion [3]. Attempts to accelerate POPF closure by
prolonged fasting can be detrimental to the patient clinical
course, because the inadequate nutrient intake increases the
risk of further clinical deterioration. Thus, nutritional support
has been regarded as a key element of conservative therapy.
Fig. 3 a Fistulogram of a pancreaticojejunostomy leak. After injecting
the water-soluble dye from the drain, a small collection and the jejunal
loop are visualized. b Fistulogram of a pancreaticogastrostomy leak.
Langenbecks Arch Surg (2014) 399:801–810
pancreatic stump and the posterior wall of the stomach, fluid collection
near the pancreatic stump extended up to the hepatic hilum (arrow)
Total parenteral nutrition (TPN) blocks the food-induced pan-
creatic secretion, eliminating the release of gastrointestinal
hormones [17]. However, studies suggested that long-term
TPN leads to negative functional and morphological changes
caused by the absence of food in the gastrointestinal tract, lack
of bile salts and proteolytic enzymes, motility dysfunction,
and changes in the serum levels of hormones [18]. Disadvan-
tages of long-term TPN include wound infection and sepsis
[19], metabolic complications (such as hyperglycemia), and
negative functional and morphological change in the gastro-
intestinal mucosa and pancreas (gastrointestinal mucosa atro-
phy, gut barrier dysfunction, pancreas atrophy, decrease en-
zyme secretion or synthesis) [17, 19]. Contrary to TPN, en-
teral nutrition (EN) generally not only avoids pancreatic stim-
ulation but also may stimulate the release of specific gut
peptides forming a negative feedback control system, and thus
inhibiting pancreatic secretion [19]. In patients with clinically
relevant POPF, EN therapy seems to offer important benefits.
Klek et al. demonstrated, by open-label randomized, con-
trolled clinical trial of EN versus TPN, that the EN support
increased by more than twofold the probability of fistula
closure, shortened the time to closure, and was associated with
After injecting the dye from the surgically placed drain, the gastric cavity
is visualized
Langenbecks Arch Surg (2014) 399:801–810
faster recovery, lower rates of nutrition-related complications,
and lower cost than TPN [20]. Accordingly, EN should be
always established in case of clinically relevant pancreatic
fistula, whenever feasible, 2 to 4 h after placement of a
naso-intestinal tube or a feeding jejunostomy into the first
jejunal loop 20 cm below the Treitz ligament or, in patients
after PD, 20 cm below the last jejunal anastomosis. The
European Society for Clinical Nutrition and Metabolism
(ESPEN) guidelines suggest for patients with a clinically
relevant POPF, especially those who are critically ill, an
energy supply not exceeding 20–25 kcal/kg body weight/
day. Patients with a severe malnutrition should receive EN
of up 25–30 total kcal/kg body weight/day. If these target
values are not reached, supplementary parenteral nutrition
may be given [21]. Metoclopramide or erythromycin could
be initially given in patients with intolerance to enteral feeding
(e.g., with high gastric residuals). Regarding the formula, in
elective upper gastrointestinal surgical patients, immune-
modulating formulae (enriched with arginine, nucleotides,
and omega-3 fatty acids) are superior to standard enteral
formulae, although patients who do not tolerate more than
700-mL enteral formulae per day should not receive it [21].
Somatostatin analogs
Over the years, several groups of investigators have proposed
the administration of somatostatin, a naturally occurring
tetradecapeptide with a wide spectrum of actions, to enhance
POPF closure and reduction in time to fistula closure [22].
Somatostatin inhibits pancreatic exocrine, biliary, and small
bowel secretions and increases the net absorption of water
[23]. Thus, the anticipated effects of somatostatin are to re-
duce the output of the digestive fistulae and to potentially
alleviate their natural course. The major disadvantage of so-
matostatin is its very short half-life (1–2 min) that necessitates
continuous intravenous infusion for periods of several weeks,
up to 60 days. Synthetic analogs with longer half-life have
been developed to avoid the need for continuous intravenous
infusion. The octapeptide octreotide has a half-life of 120 min,
which allows intermittent subcutaneous dosing schedules.
Lanreotide, another synthetic octapeptide, is available as an
approved prolonged-release microparticle formulation
(lanreotide 30 mg PR). Administered by the intramuscular
route, this formulation releases the octapeptide over a period
of 10 to 14 days. The pharmacological effect of lanreotide
appears as early as 2 h after the injection and is maintained for
at least 10 days [24]. A recent systematic review and meta-
analysis of somatostatin and somatostatin analogs for the
treatment of fistula of the gastrointestinal tract, including
POPF, analyzed seven randomized clinical trials published
between 1990 and 2009. Of the 297 patients who were pooled,
102 had a POPF. No significant advantage of the use of
somatostatin analogs in terms of the fistula closure rate was
805
shown (OR for fistula closure=1.52, 95 % CI 0.88 to 2.61).
Time to fistula closure was described using different defini-
tions, such as reduction of output expressed as a percentage or
an absolute decrease in fistula output. In addition, studies
failed to report whether median or mean time to closure was
reported, and estimates of precision were not included. Owing
to the inconsistent reporting of data, the authors did not
attempt a pooled analysis on time to closure [22]. Further-
more, because studies were published between 1990 and
2009, substantial differences existed with regard to the defi-
nition of fistula and the criteria applied for a proven POPF.
Definitions varied from no information, fistula based on ra-
diological studies or fistula output criteria. The review was
also limited by the lack of distinction between pancreatic and
enteric fistula. Most studies included both, and POPFs were,
in the majority, in only a single study. Data were not presented
for enterocutaneous and pancreatic fistula separately, making
it impossible to pool data for pancreatic fistula alone. Because
there is no solid evidence that somatostatin analogs result in a
higher closure rate of POPF compared with other treatments,
these drugs should not be considered as a standard treatment.
Interventional techniques
Interventional radiology is increasingly recognized as having
a significant role in the management of complications that
arise after pancreatic surgery [25]. Image-guided percutane-
ous drainage of peripancreatic fluid collections/abscesses sub-
sequent to a POPF was identified as the most common indi-
cation for interventional radiology after pancreatic resections
(Fig. 4) [26]. Peripancreatic collections can be approached
percutaneously as long as the subject is hemodynamically
stable and has an acceptable coagulation panel, and there is
a safe access route for a needle. Although most of the times the
drainage procedure is mandated by serious clinical symptoms,
such as fever, leukocytosis, pain, or sepsis, a number of
clinically silent abscesses have been reported [27]. Normally,
prophylactic broad-spectrum antibiotics are given before all
procedures. Using real-time ultrasound or CT guidance, the
collection is punctured, and fluid is aspirated. A sample is sent
for amylase measurement and microbiology cultures. Using
the Seldinger technique, a drainage catheter is placed into the
fluid collection. The collection is emptied as completely as
possible, and then postdrainage imaging is performed. Cath-
eter exchange or removal is based on clinical improvement as
well as drainage catheter output, catheter malfunction or dis-
lodgement, and evidence of persistent fluid on repeat imaging.
Different studies have shown that more than 85 % of patients
were managed successfully with percutaneous drainages with-
out the need for re-operation [25, 28]. Disadvantages of per-
cutaneous drainages include the need of daily care and regular
flushing of the catheter to maintain patency, as well as fre-
quent monitoring of fluid output to determine appropriate
806
Fig. 4 Computed-tomography-guided percutaneous drainage of a large collection from a pancreaticojejunostomy leak. a The collection is punctured
using a 21-gauge needle. b A pig-tail drain is placed into the collection employing the Seldinger technique
timing for catheter removal. Furthermore, radiologically guid-
ed external drainage may cause localized skin irritation and
infections, prolongs the duration of treatment (especially in
collections associated with distal pancreatectomy and enucle-
ation), delays oral re-feeding, and implies a longer hospital
stay [29]. The use of endoscopic ultrasound (EUS)-guided
transmural drainage, initially applied for management of
pseudocysts secondary to acute pancreatitis, has grown in
frequency to drain collections associated with POPF, after
both PD and distal pancreatectomy [30, 31]. Briefly, an
FNA needle is used to puncture through the gastric wall into
the fluid collection. A guide wire is advanced into the collec-
tion cavity under fluoroscopic guidance and is allowed to
form one coil in the collection. The cyst-gastrostomy tract
can initially be dilated to 4–8 mm, and then up to 10–15 mm,
particularly if endoscopic debridement is considered. One or
multiple pigtail biliary stents, at the discretion of the
endoscopist, are then placed. Varadarajulu et al. recently re-
ported technical and clinical success rate of 100 % with no
associated procedural or delayed complications [32]. Com-
pared with percutaneous drainage, EUS-guided drainage of-
fers several potential advantages. High-resolution, real-time
imaging of the fluid collection, pancreas, and surrounding
vasculature allows for precise and safe cyst-gastrostomy cre-
ation. The drain tract can be safely dilated to a larger diameter
than percutaneous drainage, allowing for rapid evacuation of
the collection and quicker resolution of symptoms. Further-
more, lavage and debridement can be performed as needed. In
theory, reintroducing pancreatic fluid from a leak into the
gastrointestinal tract is more physiologic and may reduce fluid
and enzyme losses [29]. A number of recent retrospective
studies compared the drainage of collections associated with
POPF using endoscopic and percutaneous modalities. The
rates of technical success (90–100 %) and treatment success
(79–100 %) were similar between the two techniques. Recur-
rences were often treated by “salvage” drainage via the other
modality. Hence, endoscopic and percutaneous drainage ap-
peared to be equally effective and complementary interven-
tions for POPF-associated collections [4, 29, 33, 34]. A word
Langenbecks Arch Surg (2014) 399:801–810
of caution: These studies are limited by the retrospective
design and the possible selection bias. In particular, patients
were chosen for EUS-guided drainage based on their imaging
studies and consensus opinion of the surgeon and endoscopist.
Furthermore, the EUS-guided approach seems to be better
deployed for subacute collections with mature wall, although
the timing of elective drainage of POPF-associated collections
has not been well stated in these studies. Most studies exclude
patients with fluid collections <4 weeks old because of the
presumed lack of mature wall [4]. Patients with refractory
POPF after left pancreatectomy (e.g., patients without signs
of improvement after prolonged drainage) may benefit from
sphincterotomy and pancreatic duct stenting, in an attempt to
decompress the pancreatic duct and promote antegrade flow
of pancreatic fluid. The experience with this technique is
limited and, given the potential association of pancreatic stent
placement with acute pancreatitis, it should be employed in
highly selected patients [35].
Surgical treatment
Although most POPF can be managed non-operatively, some
do require re-operative surgical intervention. Indications to
relaparotomy are not uniform across different studies and
depend on institutional expertise and patient’s preoperative
conditions. In general, an operative intervention has to be
considered in case of deteriorating general conditions despite
maximal supporting care, septic intra-abdominal collections
inaccessible to percutaneous or endoscopic drainage,
suspected peritonitis by visceral perforation, and necrosis [5,
36]. Because serious bleeding from pseudoaneurysms can
further complicate POPF, ongoing bleeding after failure or
contraindication of radiologic endovascular procedures man-
dates emergency relaparotomy and operative bleeding control
[37]. In a recent large analysis, failure of non-operative man-
agement requiring surgical re-intervention occurred in 21 % of
patients with POPF and in 36 % of patient with late hemor-
rhage, leading to a considerably increased mortality risk.
Sepsis was identified as the only independent predictor of
Langenbecks Arch Surg (2014) 399:801–810
failure of non-operative management, and the risk of dying
within 5 years after PD was six times greater in patients with
POPF in whom non-operative management failed [38].
Pancreatic stump management at relaparotomy encom-
passes multiple surgical options. Most of them apply to PD:
& Debridment and drainage of the peripancreatic region
(with or without external drainage of the main pancreatic
duct)
& Attempted repair of the site of leakage (e.g., pancreatic-
enteric anastomosis)
& Construction of a new pancreatic-enteric anastomosis
& Resection of the pancreatic-enteric anastomosis with rem-
nant ligation or closure
& Completion pancreatectomy
The choice of the technique is often dictated by intraoper-
ative findings, while the severity of clinical instability will
also set limitations regarding extent and duration of the oper-
ative procedure [36, 39, 40]. Access to the complication site
can be a fortuitous trial, because the operating field is hindered
by significant inflammation from local sepsis and tissue deg-
radation [41]. In PD, the pancreatic anastomosis should be
evaluated for the degree of surrounding inflammation, abscess
or necrosis, the extent of anastomotic dehiscence, the extent of
the gap between the pancreas and the jejunal limb, and the
quality of the pancreatic parenchyma and the jejunal serosa
[42]. A limited disruption of the pancreaticojejunostomy can
be treated by local drainage; oversewing of the anastomosis
seems to have little value, as this method is rarely successful
[41]. When the disruption exceeds half the circumference of
the suture line, or when extended necrosis of the pancreatic
stump is likely to prevent its secondary healing, some authors
consider the pancreatic anastomosis as not preservable and
suggest completion pancreatectomy [43]. Other indications to
completion pancreatectomy after initial PD are the inability to
find the main pancreatic duct despite a careful exploration of a
largely disrupted pancreatic anastomosis, and splenic artery
lesions requiring arterial ligation to ensure hemostasis [44].
Patrons of completion pancreatectomy suggest that this pro-
cedure achieves sterilization of the infection source and has a
decreased need for re-operation. However, it comes at a price
of absolute endocrine and exocrine insufficiency and still has
a significant mortality rate, ranging from 24 to 50 % [5, 36, 45,
46]. Van Berge Henegouwen et al. compared drainage versus
early completion pancreatectomy, suggesting that the latter
approach is the treatment of choice, since they reported no
mortality, in contrast with previous studies [47]. The group
from Heidelberg claimed that completion pancreatectomy
should no longer be considered in patients with POPF, be-
cause almost all the patients can be managed with conserva-
tive treatment [48]. In distal pancreatectomy, an emergency
resection of the pancreatic head to control complicated POPF
807
is a very rare event, as these fistulae can usually be managed
without surgical revision, and the procedure is associated with
a considerable morbidity and mortality that make it an unfa-
vorable approach [49].
Stenting of the pancreatic duct has been proposed as a
salvage technique when a largely dehiscent
pancreaticojejunostomy cannot be properly re-fashioned.
The gap between the pancreatic remnant and the jejunum
can be bridged with a pediatric feeding tube, which can be
externalized several centimeters downstream through the
pancreaticobiliary drainage limb in a Witzel fashion, and then
externalized through the abdominal wall. The ultimate objec-
tive is to enable a channel to develop from reactive repair
around the periphery of the stent, which will shunt secretions
from the pancreatic remnant to the pancreaticobiliary limb
[42]. A similar technique was described by Paye et al., who
place an exteriorized pancreatic stent, and either staple or
exteriorize the jejunal stump. Restorative laparotomy is per-
formed >3 months after the salvage procedure, with the stent
being used as a guide to re-fashion the pancreatic anastomosis.
All these procedures were associated with high success rates
[44]. Bachellier et al. reported the results of salvage
pancreaticogastrostomy for the treatment of POPF following
PD with pancreaticojejunostomy in the rare patients for whom
relaparotomy was necessary, and compared the results with a
historical series of similar patients who underwent completion
pancreatectomy. Mortality was 0 % after salvage
pancreaticogastrostomy and 50 % after completion pancrea-
tectomy. Despite retrospective and affected by selection bias,
this data indicated that salvage pancreaticogastrostomy is
a safe and efficient alternative to completion pancrea-
tectomy for the treatment of grade C POPF after PD with
pancreaticojejunostomy [50].
Whatever the surgical approach is, it has been shown that
subsequent operations were required in nearly 50 % of pa-
tients with redo-operations. The majority of these were nec-
essary in patients with septic complications and the need for
open abdominal lavage with secondary abdominal wall
closure [41].
A surgical option for external refractory POPF is
fistulojejunostomy. The surgical repair consists of precise iden-
tification of the fistula tract around a drainage tube (far from
the pancreatic gland) and its anastomosis with a Roux-en-Y
jejunal loop. This technique has been described by our group
and other authors and has been shown to be associated with a
very high success rate, despite the very limited indications
[51, 52].
Very complex POPF
Occasionally, POPF may present as part of a complex abdom-
inal wound infection, with fluid necessitation through an
underlying fascial defect. Large collections anterior to the
808
pancreatic remnant and rich in pancreatic juice may disrupt
either the small or the large bowel, resulting in an
enterocutaneous fistula (Fig. 5). The clinical course is
protracted and replete with multiple interventional drainage
procedures, and fistula output rich in electrolytes, minerals,
and proteins worsens the ongoing electrolyte imbalance and
malnutrition. Intra-abdominal infection control with antibiotics
(according to microbiologic cultures) is mandatory and allows
the formation of a “controlled fistula,” namely, an
enterocutaneous fistula without evidence of sepsis or localized
infection (e.g., abdominal wall cellulitis). Protection of the skin
against maceration and breakdown caused by the corrosive
action of pancreatic end enteric fistula effluent is important.
Stoma appliances and protective films may also limit the need
for re-iterated wound dressing; negative pressure applied by
vacuum-assisted wound closure devices may be useful [53].
The POPF therapy should be established depending on the
anatomic information obtained once the fistula is well con-
trolled, including the length and the course of the fistula tract,
its relationship with the bowel, the nature of the bowel adjacent
to the fistula (inflammation, stricture, interruption of bowel
continuity), and the presence of an abscess cavity in communi-
cation with the fistula. In the event of complicated and persis-
tent high-output fistulae, surgical approaches involving local
drainage, bowel resection with stoma, dismantling of the dehis-
cent pancreatic anastomosis, and closure of the pancreatic duct,
or completion pancreatectomy must be considered. However,
these procedures are a considerable source of mortality, ranging
from 5 to 35 % [54]. Plastic surgical involvement in fistula
closure is usually necessary when the defect of the abdominal
wall is particularly large, requiring specific expertise.
Fig. 5 Complex pancreatic and enterocutaneous fistula following
pancreaticoduodenectomy. After percutaneous drainage of an infected
collection, two emergency laparotomies for bleeding from the anastomot-
ic site were required. Subsequently, common hepatic artery stenting was
necessary for massive hemorrhage. Then, the patient developed a high-
output pancreatic-enteric fistula and a colonic fistula. a Disrupted
Langenbecks Arch Surg (2014) 399:801–810
Conclusion
POPF is a complex, multivariable phenomenon that con-
tinues to challenge pancreatic surgeons. The introduction in
2005 of a common and worldwide-accepted definition has
allowed standardized report of surgical results. Despite the
technical advancements in perioperative care, the incidence
of POPF has not diminished, presumably because indica-
tions to pancreatic resections have broadened to include
cystic and neuroendocrine neoplasms. These conditions are
in fact associated with soft pancreatic texture, which is the
leading risk factor for POPF formation. Studies have
shown that the high incidence of POPF was accompanied
by a shift from operative to non-operative management,
with a decrease in the length of hospital stay in the non-
operative group [38]. However, the management strategy is
driven by the patient’s condition and local expertise and is
based on poor evidence. A single randomized trial showed
that enteral nutrition is superior to TPN [20], and pooled
data of randomized trials failed to show any advantage of
somatostatin analogs for accelerating POPF closure [22].
The choice of percutaneous versus endoscopic drainage of
peripancreatic collections remains arbitrary, and—when re-
operation is needed—there are very few comparative data
regarding simple local drainage with or without main pan-
creatic stenting as opposed to anastomotic revision or
salvage re-anastomosis. The continuous development of
specialist, high-volume units with appropriate resources
and multidisciplinary experience in complication manage-
ment might further improve the evidence and the
outcomes.
pancreaticojejunostomy (asterisk), associated collection draining through
a large defect of the abdominal wall (arrow). b After persisting drainage,
use of vacuum-assisted wound closure devices, and multiple endoscopic
attempts of fistula closure, the fistula resolved (time to closure 5 months),
and the patient eventually survived
Langenbecks Arch Surg (2014) 399:801–810
Conflicts of interest None.
References
1. Enestvedt CK, Diggs BS, Cassera MA et al (2012) Complications
nearly double the cost of care after pancreaticoduodenectomy. Am J
Surg 204:332–338
2. Matthews JB (2011) Prevention, evaluation, and treatment of leaks
after pancreatic surgery. J Gastrointest Surg 15:1327–1328
3. Bassi C, Malleo G (2011) Pancreas: postoperative pancreatic fistula:
use of enteral nutrition. Nat Rev Gastroenterol Hepatol 8:427–428
4. Tilara A, Gerdes H, Allen P et al (2014) Endoscopic ultrasound-
guided transmural drainage of postoperative pancreatic collections. J
Am Coll Surg 218:33–40
5. Fuks D, Piessen G, Huet E et al (2009) Life-threatening postoperative
pancreatic fistula (grade C) after pancreaticoduodenectomy: inci-
dence, prognosis, and risk factors. Am J Surg 197:702–709
6. Harnoss JC, Ulrich AB, Harnoss JM et al (2014) Use and results of
consensus definitions in pancreatic surgery: a systematic review.
Surgery 155:47–57
7. Goudard Y, Gaujoux S, Dokmak S et al (2014) Reappraisal of central
pancreatectomy a 12-year single-center experience. JAMA Surg 149:
356–363
8. Bassi C, Dervenis C, Butturini G et al (2005) Postoperative pancre-
atic fistula: an international study group (ISGPF) definition. Surgery
138:8–13
9. McMillan MT, Vollmer CM (2014) Predictive factors for pancreatic
fistula following pancreatectomy. Langenbecks Arch Surg Epub
Ahead Print. doi:10.1007/s00423-014-1220-8
10. Miller BC, Christein JD, Behrman SW et al (2013) Assessing the
impact of a fistula after a pancreaticoduodenectomy using the Post-
operative Morbidity Index. HPB 15:781–788
11. Malleo G, Crippa S, Butturini G et al (2010) Delayed gastric empty-
ing after pylorus-preserving pancreaticoduodenectomy: validation of
International Study Group of Pancreatic Surgery classification and
analysis of risk factors. HPB 12:610–618
12. Pratt WB, Callery MP, Vollmer CM Jr (2009) The latent presentation
of pancreatic fistulas. Br J Surg 96:641–649
13. Raman SP, Horton KM, Cameron JL, Fishman EK (2013) CT after
pancreaticoduodenectomy: spectrum of normal findings and compli-
cations. AJR Am J Roentgenol 201:2–13
14. Zink SI, Soloff EV, White RR, et al. (2008) Pancreaticoduodenectomy:
frequency and outcome of post-operative imaging-guided percutane-
ous drainage. Abdom Imaging
15. Bruno O, Brancatelli G, Sauvanet A et al (2009) Utility of CT in the
diagnosis of pancreatic fistula after pancreaticoduodenectomy in
patients with soft pancreas. AJR Am J Roentgenol 193:W175–W180
16. Faccioli N, Foti G, Molinari E et al (2012) Role of fistulography in
evaluating pancreatic fistula after pancreaticoduodenectomy. Br J
Radiol 85:219–224
17. Spiller RC, Trotman IF, Higgins BE et al (1984) The ileal brake–
inhibition of jejunal motility after ileal fat perfusion in man. Gut 25:
365–374
18. Fong YM, Marano MA, Barber A et al (1989) Total parenteral
nutrition and bowel rest modify the metabolic response to endotoxin
in humans. Ann Surg 210:449–56
19. O’Keefe SJD (2006) Physiological response of the human pancreas
to enteral and parenteral feeding. Curr Opin Clin Nutr Metab Care 9:
622–628
20. Klek S, Sierzega M, Turczynowski L et al (2011) Enteral and paren-
teral nutrition in the conservative treatment of pancreatic fistula: a
randomized clinical trial. Gastroenterology 141:157–63
809
21. Meier R, Ockenga J, Pertkiewicz M et al (2006) ESPEN Guidelines
on Enteral Nutrition: pancreas. Clin Nutr 25:275–284
22. Gans SL, van Westreenen HL, Kiewiet JJS et al (2012) Systematic
review and meta-analysis of somatostatin analogues for the treatment
of pancreatic fistula. Br J Surg 99:754–760
23. Dollinger HC, Raptis S, Pfeiffer EF (1976) Effects of somatostatin on
exocrine and endocrine pancreatic function stimulated by intestinal
hormones in man. Horm Metab Res 8:74–78
24. Gayral F, Campion JP, Regimbeau JM et al (2009) Randomized,
placebo-controlled, double-blind study of the efficacy of lanreotide
30 mg PR in the treatment of pancreatic and enterocutaneous fistulae.
Ann Surg 250:872–877
25. Sanjay P, Kellner M, Tait IS (2012) The role of interventional
radiology in the management of surgical complications after
pancreatoduodenectomy. HPB 14:812–817
26. Sohn TA, Yeo CJ, Cameron JL et al (2003) Pancreaticoduodenectomy:
role of interventional radiologists in managing patients and complica-
tions. J Gastrointest Surg 7:209–219
27. Sierzega M, Kulig P, Kolodziejczyk P, Kulig J (2013) Natural history
of intra-abdominal fluid collections following pancreatic surgery. J
Gastrointest Surg 17:1406–1413
28. Munoz-Bongrand N, Sauvanet A, Denys A, Sibert A, Vilgrain V,
Belghiti J (2004) Conservative management of pancreatic fistula after
pancreaticoduodenectomy with pancreaticogastrostomy. J Am Coll
Surg 199:198–203
29. Kwon YM, Gerdes H, Schattner MA et al (2013) Management of
peripancreatic fluid collections following partial pancreatectomy: a
comparison of percutaneous versus EUS-guided drainage. Surg
Endosc 27:2422–2427
30. Reddymasu SC, Pakseresht K, Moloney B et al (2013) Incidence of
pancreatic fistula after distal pancreatectomy and efficacy of endo-
scopic therapy for its management: results from a tertiary care center.
Case Rep Gastroenterol 7:332–339
31. Bartoli E, Rebibo L, Robert B et al (2013) Efficacy of the double-
pigtail stent as a conservative treatment for grade B pancreatic fistula
after pancreatoduodenectomy with pancreatogastric anastomosis.
Surg Endosc 28:1528–1534
32. Varadarajulu S, Wilcox CM, Christein JD (2011) EUS-guided ther-
apy for management of peripancreatic fluid collections after distal
pancreatectomy in 20 consecutive patients. Gastrointest Endosc 74:
418–423
33. Azeem N, Baron TH, Topazian MD et al (2012) Outcomes of
endoscopic and percutaneous drainage of pancreatic fluid collec-
tions arising after pancreatic tail resection. J Am Coll Surg 215:
177–185
34. Onodera M, Kawakami H, Kuwatani M et al (2011) Endoscopic
ultrasound-guided transmural drainage for pancreatic fistula or pan-
creatic duct dilation after pancreatic surgery. Surg Endosc 26:1710–
1717
35. Grobmyer SR, Hunt DL, Forsmark CE et al (2009) Pancreatic stent
placement is associated with resolution of refractory grade C pancre-
atic fistula after left-sided pancreatectomy. Am Surg 75:654–7
36. Balzano G, Pecorelli N, Piemonti L et al (2014) Relaparotomy for a
pancreatic fistula after a pancreaticoduodenectomy: a comparison of
different surgical strategies. HPB 16:40–5
37. Wente MN, Veit JA, Bassi C et al (2007) Postpancreatectomy hem-
orrhage (PPH)–An International Study Group of Pancreatic Surgery
(ISGPS) definition. Surgery 142:20–25
38. Tol JA, Busch OR, van Delden OM, van Lienden KP, van Gulik TM,
Gouma DJ (2014) Shifting role of operative and nonoperative inter-
ventions in managing complications after pancreatoduodenectomy:
what is the preferred intervention? Surg Epub Ahead of Print. doi:10.
1016/j.surg.2014.04.026
39. Gangl O, Fröschl U, Hofer W et al (2011) Unplanned reoperation and
reintervention after pancreatic resections: an analysis of risk factors.
World J Surg 35:2306–2314
810
40. Denbo JW, Orr WS, Zarzaur BL, Behrman SW (2012) Toward defin-
ing grade C pancreatic fistula following pancreaticoduodenectomy:
incidence, risk factors, management and outcome. HPB 14:
589–593
41. Standop J, Glowka T, Schmitz V et al (2009) Operative re-
intervention following pancreatic head resection: indications and
outcome. J Gastrointest Surg 13:1503–1509
42. Kent TS, Callery MP, Vollmer CM Jr (2010) The bridge stent tech-
nique for salvage of pancreaticojejunal anastomotic dehiscence. HPB
12:577–582
43. Dellaportas D, Tympa A, Nastos C et al (2010) An ongoing dispute in
the management of severe pancreatic fistula: pancreatospleenectomy
or not? World J Gastrointest Surg 2:381–384
44. Paye F, Lupinacci RM, Kraemer A et al (2013) Surgical treatment of
severe pancreatic fistula after pancreaticoduodenectomy by
wirsungostomy and repeat pancreatico-jejunal anastomosis. Am J
Surg 206:194–201
45. Cullen JJ, Sarr MG, Ilstrup DM (1994) Pancreatic anastomotic leak
after pancreaticoduodenectomy: incidence, significance, and man-
agement. Am J Surg 168:295–298
46. Farley DR, Schwall G, Trede M (1996) Completion pancreatectomy
for surgical complications after pancreaticoduodenectomy. Br J Surg
83:176–179
Langenbecks Arch Surg (2014) 399:801–810
47. van Berge Henegouwen MI, De Wit LT, Van Gulik TM et al (1997)
Incidence, risk factors, and treatment of pancreatic leakage after
pancreaticoduodenectomy: drainage versus resection of the pancre-
atic remnant. J Am Coll Surg 185:18–24
48. Büchler MW, Wagner M, Schmied BM et al (2003) Changes in
morbidity after pancreatic resection: toward the end of completion
pancreatectomy. Arch Surg 138:1310–4
49. Hackert T, Werner J, Büchler MW (2011) Postoperative pancreatic
fistula. Surgeon 9:211–217
50. Bachellier P, Oussoultzoglou E, Rosso E et al (2008)
Pancreatogastrostomy as a salvage procedure to treat severe postop-
erative pancreatic fistula after pancreatoduodenectomy. Arch Surg
143:966–70
51. Bassi C, Butturini G, Salvia R et al (2000) A single-institution
experience with fistulojejunostomy for external pancreatic fistulas.
Am J Surg 179:203–206
52. Nair RR, Lowy AM, McIntyre B et al (2007) Fistulojejunostomy for
the management of refractory pancreatic fistula. Surgery 142:636–42
53. Piazza RC, Armstrong SD, Vanderkolk W et al (2009) A modified
“fistula-VAC” technique: management of multiple enterocutaneous
fistulas in the open abdomen. Plast Reconstr Surg 124:453e–5e
54. Lee SH (2012) Surgical management of enterocutaneous fistula.
Korean J Radiol 13:S17–S20