pancreatic and intestinal fistulas

7/24/2019 Pancreatic and Intestinal Fistulas

1/8

281Chapter 25

Pancreatic fistulas (PFs) and intestinal fistulas (IFs)are troublesome, occasionally significant, and not un-common sequelae of necrotizing pancreatitis (NP).hey account for increased morbidity and sometimesmortality, and prolonged hospital stay, and they arecostly both f inancially and with regard to resources.

Incidence varies between 5% and over 50% amongpublished series, but there is a definite decreasingtrend recently across the literature. he wide varia-tion in incidence reflects not only different levels ofexpertise among the authors, but also the strikinglack of a universally accepted and applied definitionof PF specifically in the context of NP. Although re-cently a consensus definition and staging of postop-erativePF was published [1], there has been no similarunifying attempt in the setting of NP, as is our topic.he decreased incidence of PF/IF in the recent litera-ture, in addition to improved surgical expertise, cer-

tainly ref lects in part the recent change in the overallmanagement strategy of NP, as will be discussed be-low.

Higher imaging precision has led to more accuratediagnoses by the delineation of fine, but crucial ana-tomic details of both PF and IF. In addition, advancedtechnology and refined operative and interventionalor minimally invasive techniques have contributed toan improved outcome in these patients.

In this chapter we will discuss the pathogenesisand management of PF and IF separately, but prior to

this, it is essential to briefly outline a very significantchange in the management scheme of NP that hastaken place during the last decade, which has crucialimplications in both the incidence and the treatmentof PF and IF.

Modern Management of NPand its Implications

Since it has been recognized that the early peak ofmortality in the biphasic mortality pattern of NP isdue to the overwhelming systemic inflammatory re-sponse syndrome (SIRS; not sepsis), whereas the latersecond peak is due to sepsis, two major componentsof modern management have emerged: (1) very ag-gressive hemodynamic, ventilatory, metabolic, andnutritional support and avoiding operative treatmentin the early phase, and (2) delayed operative treatment(where necessary) for as long as possible. his ap-proach, which has been substantiated by cornerstoneclinical studies [2,3] and is now the preferred man-agement strategy in patients with NP [4], has led tooptimized hemodynamics early after NP, much fewerreoperations for debridement (usually just one), es-

sentially no gauze packing, and placement of fewerdrains. As will be discussed in detail below, these fac-tors have substantially decreased the incidence of NP/IF.

Pathogenesis

Although the pathogenesis of PF/IF is multifactorial,the most common factor in their development is thepresence of pancreatic parenchymal necrosis, as this

results in the disruption of small or large pancreaticducts with subsequent extravasation of exocrine se-cretions into the retroperitoneum [5]. Operative ne-crosectomy and local drainage allow for externalegress of these extravasated secretions and the poten-tial for a pancreaticocutaneous fistula. he concur-rent pancreatic and peripancreatic inflammatoryprocess may also lead to stenosis of the pancreaticducts, which represents a substantial element for thechronicity of fistulas. he importance of pancreaticparenchymal necrosis as the main risk factor in thepathogenesis of PF is stressed by the finding in onestudy that all patients who developed pancreaticocu-

Pancreaticand Intestinal Fistulas

G. Tsiotos J. Tsiaoussis


2/8

Pancreatic and Intestinal Fistulas G. Tsiotos J. Tsiaoussis282

taneous fistulas had proven pancreatic parenchymalnecrosis, while none of the patients with peripancre-atic retroperitoneal fat necrosis with an inflamed but

viable pancreas developed a PF [6].

Compromised Blood Supply

Compromised blood supply (in the form of vascularthrombosis) to the colon and the duodenum has beenpostulated as a pathogenetic mechanism for the for-mation of gastrointestinal fistulas in particular. En-zyme-rich fluids and inflammatory products releasedor produced early in the course of the necrotizingprocess can dissect throughout the retroperitonealtissues and into the transverse mesocolon to involvethe vascular supply to the colon or duodenum, withconsequent vascular thrombosis. he subsequentischemia (if extensive enough) may lead to segmentalcolonic or duodenal necrosis and eventually the for-mation of a gastrointestinal fistula. Colonic ischemiamay occur as a result of a low-flow state [7] caused byinadequate initial resuscitation or as a consequence ofthe hemodynamic response to the sepsis syndrome.his speculation would explain the tendency for fis-tulas to arise from the left transverse colon and thesplenic flexure, where collateral f low is more compro-mised in low-flow states. However, the modern ag-gressive hemodynamic resuscitation in the early

phase of NP (see above) has minimized a low-flowstate as a cause of a fistula.

Autodigestion

Autodigestion of adjacent organs as a cause of IF inthe course of NP has more of a theoretical background.According to this hypothesis, the extravasated exo-crine secretions may result in transmural necrosis ofthe stomach and small intestine in a way similar to

peripancreatic fat necrosis. his concept for the devel-opment of gastrointestinal tract fistulas seems muchless likely because, unlike fat, the stomach and smallintestine have a much better vascular supply and,thus, associated protective mechanisms [6].

Choice of Operative Technique

he development of fistulas may be related to thechoice of operative technique, since repeated localtrauma to the surface of an organ, as might occurwith repeated open packing of the lesser sack or dur-

ing planned relaparotomies, may lead to intestinalwall erosions. o reduce this possibility, covering ofthe exposed viscera and major blood vessels with aform of nonadherent interface before applying the in-tra-abdominal gauze packing has been proposed [5].In addition, recent studies indicate that necrosectomy

followed by closed packing or by closed continuouslavage may lead to a lower incidence of PF/IF forma-tion [8,9], as this approach requires fewer intra-ab-dominal interventions for the repeated removal ofpancreatic necrotic material.

It is interesting, however, to note that the discus-sions about individual techniques and the compari-sons among them tend to become obsolete, since to-days management strategy of patients with NPconsisting of aggressive nonoperative initial manage-ment followed by necrosectomy as late as possible (re-moval of well-demarcated necrotic tissue withoutcompromising viable viscera usually 1 month afterthe onset of NP) leads to a more accurate distinctionbetween viable and necrotic tissue and a more com-plete necrosectomy, with a lower risk of leaving non-

viable infected debris behind. hus, a much lowernumber of relaparotomies is required, with a lowerincidence of adjacent organ injury and, as a result, alower frequency of gastrointestinal fistulas.

Choice of Operative Approach

to the Retroperitoneal Space

he choice of operative approach to the retroperito-neal space during necrosectomy could predispose tofistula development as a result of adjacent organ in-

jury. he lesser sac can be approached through thetransverse mesocolon, the gastrocolic omentum, orthe gastrohepatic omentum. Because of the inflam-matory process, the stomach and transverse colonmay have been densely adherent to the inflammatorymass. Consequently, accessing the lesser sac through

an avascular area of the mesocolon to the left of theligament of reitz seems quicker and safer, avoidingany inadvertent injury to the adjacent organs. Fer-nandez-del Castillo et al. recognized the fact that nocolonic fistula developed in their series and attributedthis to their preferred access via the mesocolon [3].

Pressure Necrosis

Another iatrogenic mechanism of fistula formationmay be from pressure necrosis of a segment of bowelfrom an adjacent drain and such a mechanism could


3/8

Chapter 25 Pancreatic and Intestinal Fistulas 283

reflect the late development of certain colonic, small-bowel, or gastric fistulas. o avoid this, the peripan-creatic drains should not be positioned directly on theduodenum or ascending colon when placed from thepatients right side, or on the descending colon whenplaced from the left side. he ideal placement of the

drains is behind the splenic flexure of the colon andbelow the lower pole of the spleen, thereby exiting theabdominal wall in the left anterior axillary line. Also,large, hard, stiff sump drains should be avoided as anadditional means of prevention of gastrointestinal fis-tula.

Minimally invasive techniques (percutaneous orendoscopic drainage) and minimally invasive surgery(retroperitoneoscopic debridement) are alternativesto open surgery in select cases, promising lower mor-bidity including lower incidence of PF [10]. hese in-novations have not been popularized yet and only asmall number of series (with highly selected patients)have been published [11]. With the currently availableexperience, laparoscopy-assisted necrosectomy mightbe followed by a higher trend for significant injuriesto intra-abdominal viscera including a higher inci-dence of PF/IF, as has been reported (2060%) [9].

Pancreatic Fistulas

Definition and Incidence

he lack of a widely accepted definition of this com-plication in the setting of NP has contributed in partto the major discrepancy in its reported incidenceamong published series. Criteria such as amylase levelin the excreted fluid, daily output of the fistula, andits duration vary among studies. Despite the varyingdefinitions, a PF can be conceptualized anatomicallyas an abnormal communication between a pancreaticduct (major or minor) and the skin (pancreaticocuta-neous fistula or external fistula), or between the pan-

creatic duct and peritoneal or pleural cavity, or an-other hollow viscus (internal fistula). Because of itsclinical significance we will deal with external fistu-las in this chapter. he presence of infected NP (ver-sus sterile NP) seems to favor the development of PF,as PF is far more common (up to 76%) in the formergroup [6,12].

Diagnosis and Imaging

he diagnosis of the presence of a PF can be easilymade by measuring the amylase activity of the fluid

excreted through a drain tube; this is higher than1000 IU/dl and usually up to a few thousands of IU/dl.Amylase activity levels of a few hundreds of IU/dlgenerally do not reflect a PF. he diagnosis of a PFshould be followed by the precise delineation of itsanatomic details. Complete mapping of the PF and its

relationship to the pancreatic ductal system deter-mines its prognosis and dictates the management op-tions.

he two issues of paramount importance to bestudied and looked for by imaging are:

1. Communication of the PF with the main pancreaticduct or one of its minor branches.

2. Integrity of the pancreatic duct downstream (i.e.,between the ductal disruption and the sphincter ofOddi). Te possibilities are the following:a. proximal stenosis (i.e., stenosis of the main

pancreatic duct between its disruption and thesphincter of Oddi, in which case the PF is fedprimarily, but not exclusively, by the portion ofthe pancreatic duct of the distal pancreas)

b. disconnected pancreatic duct (i.e., no communi-cation between the PF and the proximal pancre-atic duct, in which case the PF is exclusively fedby the portion of the pancreatic duct of the dis-tal pancreas. Tis distal portion of the pancreasis then an isolated pancreatic segment drainingsolely through the fistula)

c. normal pancreatic duct (i.e., the PF is fed bya rather small disruption of the pancreatic duct,but its proximal and distal portions are in conti-nuity and there is no stenosis along the length ofthe duct).

Imaging techniques that can be employed in order toextract this fine information include: (1) contrast si-nogram, (2) contrast-enhanced computed tomogra-phy (CEC), (3) magnetic resonance cholangiopan-creatography (MRCP), and (4) endoscopic retrograde

cholangiopancreatography (ERCP).he water-soluble contrast sinogram is the first

imaging study that should be performed and it maywell be the only one required. It is very easy to per-form, dynamic, low-cost, and noninvasive. It may re-

veal a communication of the PF with the main or aside pancreatic duct. In cases of right or mid-bodypancreatic necrosis, the sinogram may show filling ofthe distal pancreatic duct without opacification of theproximal pancreatic duct or the duodenum. A spiralCEC scan with thin cuts may also demonstrate thepresence and anatomy of a PF, but there is no com-parative study that favors CEC scan over sinogram.


4/8


MRCP has been utilized increasingly to demonstratepancreatic ductal anatomy and certainly does have arole in the imaging of PF as they relate to the ductalsystem. It should be kept in mind, however, that al-though CEC and MRCP represent today the mostmodern modalities of cross-sectional anatomy and

are readily available in most institutions, they maynot necessarily provide more information pertinentto the precise anatomy of a PF compared to an ex-pertly performed sinogram. Interestingly, it has beenshown, for example, that although MRCP is capableof identifying major pancreatic ductal injuries, its re-liability to discriminate subtle anomalies in anatomyor to demonstrate a communication between the pan-creatic duct and a pseudocyst is not high [13]. ERCP isgenerally reserved when sinogram and cross-section-al modalities have not provided all of the necessaryanatomic information, or when an endoscopic thera-peutic procedure is contemplated after sinogram hasprecisely demonstrated the anatomy.

Management

PFs are quite often complicated by fluid and electro-lyte abnormalities, malnutrition, skin erosion, andless often hemorrhage and sepsis. he initial manage-ment of PF is conservative and its fundamental prin-ciples are the following:

1. Provision of optimal drainage to avoid intra-ab-dominal fluid collections.

2. Maintenance of fluid and electrolyte balance.3. reatment of local infection.4. Optimization of nutritional status by parenteral or

preferably enteral feeding.5. Skin care.

An additional appealing line of PF management is thereduction of pancreatic secretion (and thus PF out-

put) by possible administration of octreotide. herole of octreotide in the prevention and the treatmentof PF has been studied in the postpancreatectomy set-ting (for tumor or chronic pancreatitis), but not in theNP setting. It is fair to note that five of nine prospec-tive randomized studies demonstrate a favorable ef-fect of octreotide over placebo, whereas the remain-ing four did not [14]. Again, all nine studies includedpatients who had undergone an elective pancreaticoperation and not patients with PF in the context ofNP. Our experience with the use of octreotide in aneffort to accelerate the closure of inflammatory PFhas not been encouraging [6]. It seems unlikely that

this issue will be definitively resolved in the near fu-ture; today, the use of octreotide in this setting is notevidence-based and should be reserved for use onlywithin a clinical protocol.

Spontaneous closure of a PF should be the primarytherapeutic goal. Compared to postoperative PF, post-

NP PF tends to have less chances of spontaneous clo-sure (53% vs 86%) and longer duration for those thatdo selfresolve (22 weeks vs 11 weeks) [15]. Althoughgood nutritional status, optimal drainage, and ab-sence of local and systemic infection certainly providea favorable background for spontaneous PF closure,the single most important factor determining prog-nosis and dictating definitive management is the pat-tern of ductal disruption that has given rise to thedevelopment of the PF. Failure of spontaneous closureis generally due to anatomic factors such as down-stream ductal obstruction and disconnected ductsyndrome (isolated pancreatic tail). For example, rec-ognition of an intact pancreatic duct without a down-stream ductal obstruction indicates a high possibilityof spontaneous closure. On the contrary, surgery isnecessary when the PF is associated with a leak fromthe pancreatic duct that is not joined with the gastro-intestinal tract. Indeed, no such PF closed after amean of 26 weeks of aggressive medical therapy, andall of these patients required surgical intervention[15].

In general, when a PF persists for more than

2 weeks and its daily output remains essentially un-changed over this period (i.e., without significant,meaningful decrease), a sinogram should be per-formed, followed possibly by MRCP. he choiceamong further treatment options depends on the spe-cific findings:

Communication of the PF with a Minor (Side)

Pancreatic Duct

Nonoperative management is justif ied and the chanc-

es are that such a PF will eventually close, even after afew months, provided that the duct drains with nostenosis (stricture) toward the main pancreatic ductand to the duodenum.

Communication of the PF with the Main

Pancreatic Duct that has a Proximal Stenosis

Spontaneous closure is unlikely because the pancre-atic duct distal to the stricture is preferentially drain-ing to the PF. hese patients are ideal for endoscopicbridging of the proximal and distal parts of the pan-creatic duct (traversing the stricture, as well as the


5/8


ductal disruption feeding the PF) and stent place-ment. Small-diameter (57 mm) stents with variablelength across the site of the ductal stricture and dis-ruption may be used. Success rates are high (75100%)[16,17]. Suboptimal stent placement, stent migration,or stent occlusion may require repetition of the tech-

nique to achieve complete closure of the PF (up to fivestent placements in one patient have been reported forPF resolution). Interestingly, even when the stent onlytraverses the stricture, but does not bridge the site ofductal leakage, the outcome may be successful [18].he duration of endoscopic therapy to close the PF

varies from a few days to several weeks, and once it hasbeen rendered, stents are retrieved after 1014 days.

Communication of a PF with a Disconnected

Pancreatic Duct

Nonoperative therapy is doomed to failure and endo-scopic stent placement is highly unlikely to be suc-cessful since there is no communication between theproximal and the distal portions of the pancreaticduct; the distal isolated pancreatic segment is drain-ing exclusively via the PF. Operative treatment shouldbe planned after the surrounding inflammation hasceased and the general condition of the patient is im-proved (usually it is after several weeks or even a fewmonths before operative treatment takes place). Sur-gical options include distal pancreatectomy with or

without splenectomy (realistically, the latter is techni-cally hard given the recent extensive retropancreaticinflammation and the fibrosis resulting from NP)and distal pancreaticojejunostomy using a defunc-tionalized Roux-en-Y loop. A more rarely utilizedthird option is a fistulojejunostomy, provided that thefistulous tract has well matured. Operative manage-ment, when indicated, is generally successful (>90%),but is associated with a not insignificant mortality(6%) [19]. Failures are due to inadequate resolution ofthe inflammatory process of the NP.

Communication of the PF with a Normal Main

Pancreatic Duct

Although the main pancreatic duct has no strictures,it does have a disruption, but its portions proximaland distal to this disruption are in continuity. Enoughtime should be provided for spontaneous closure. Ifthis does not take place, endoscopic stent placementshould be performed and this is expected to be thedefinitive therapy.

Minimally invasive endoscopic techniques forstent placement today represent the definitive therapy

of PF in the vast majority of patients with appropri-ately defined fistulas and pancreatic ductal anatomy.However, in the minority in whom this approach fails,an operation is mandatory. For patients with no dis-connected pancreatic segment, a side-to-side pancre-aticojejunostomy at the area of the origin of the fistu-

lous tract is usually the optimal operative option.

Intestinal Fistulas

Definition and Incidence

As with PFs, the incidence of IFs varies widely amongstudies (143%) [10]. However, in the case of Ifs, thiswide variation is not primarily due to variability ofdefinition, but rather reflects differences in the tech-nique of the initial necrosectomy and further opera-tive debridements. It is especially with IFs where thechanged management scheme of NP (i.e., delayed firstnecrosectomy with much lower number of subsequentdebridements required) has resulted in a recent sig-nificant reduction in their incidence.

IFs should be conceptualized anatomically in up-per-gut fistulas and colonic fistulas. his distinctionis clinically relevant because the former have a gener-ally milder course and tend to close nonoperatively, asopposed to the latter, which can be associated withsignificant morbidity and often require operative

management. he pathogenesis of IF has been alreadydiscussed in detail earlier in this chapter.

he diagnosis of the presence of IF is easily sus-pected, solely by the nature of the fluid coming out

via the drain tube (or the incision). Low-viscosity, bil-ious fluid, higher-viscosity, green-brownish fluid,and obvious fecal material obviously reflect duode-nal/proximal jejunal, small intestinal, and colonic fis-tulas, respectively. Amylase activity level of the IFfluid is high, but certainly far lower than the levelsassociated with PFs; it is usually a few hundreds of

IU/dl. When amylase activity level from an IF is in therange of thousands of IU/dl, one should suspect a co-existing PF draining through the same fistuloustract.

Imaging

Water-soluble contrast infusion through the fistuloustract under fluoroscopy (sinogram) is the first imag-ing modality to perform. his will show the lengthand width of the fistulous tract, it will light up thehollow viscous where the tract originates from (and


6/8


thus define whether it is duodenum, proximal or dis-tal small bowel, or colon), and will demonstrate thepresence or absence of intestinal stenosis or obstruc-tion distal to the intestinal disruption and the sourceof the fistula. A CEC scan should always be per-formed primarily to rule out an undrained fluid col-

lection that often coexists at the time that the IF isfirst diagnosed.

Management

As with PF, the principles of IF management are: pro-vision of excellent drainage, optimization of fluid,electrolyte, and nutritional status, treatment of localinfection, and skin care. In patients with upper-gutfistulas, nutrition should be provided either parenter-ally or enterally distal to the site of the IF using a na-sojejunal tube or a tube placed operatively.

A lot of the decision-making regarding furthermanagement depends upon the source of the fistula(upper gut versus colonic) and on the specific ana-tomic details of the IF:

1. Every undrained fluid collection seen in CECshould be well drained by percutaneously placedtubes under radiologic guidance. Tis will alleviatesystemic infection and also simplify the fistuloustract by providing a direct communication between

the intestine and the skin without pooling of intes-tinal content into the surrounding tissues.

2. Te presence of a luminal obstruction or stenosisdistal to the luminal disruption feeding the IF es-sentially guarantees failure of nonoperative manage-ment. On the contrary, absence of a distal stenosis

justifies nonoperative management for long time.3. Te pattern of the fistulous tract is of paramount

importance. Narrow and long fistulas are far morelikely to close spontaneously compared to the wideand short ones. Prolonged nonoperative manage-

ment is the preferred option for the former, whereasit is not justified for the latter.

4. Te role of IF output is important and twofold. Ahigh output (i.e., >200 ml/day persistently) may befirst associated with hard-to-maintain electrolytebalance, and second with distal luminal stenosisand/or a wider and shorter fistulous tract. It is forthese reasons that a persistently high-output IF isless likely to close spontaneously.

Provided that no absolute contraindication (i.e., distalluminal obstruction) is present for nonoperative man-agement, the prognosis of IF is closely related to the

part of the gastrointestinal tract that the IF originatesfrom.

Duodenal Fistulas

Duodenal fistulas are very rare [20]. When present,

they originate from the medial aspect of the duode-num as a result of extension of the pancreatic inflam-matory process and duodenal wall necrosis. If wellcontrolled, despite their potential for a high initialoutput, most close spontaneously. In patients wherethe duodenal fistula is diagnosed during the first or asubsequent necrosectomy, it is reasonable to proceedat that setting with pyloric exclusion (pyloric staplingwith a noncutting device and gastrojejunostomy), orsimple tube duodenostomy and excellent periduode-nal drainage. Duodenal fistulas manifested and diag-nosed prior to, or following necrosectomy should becontrolled by interventional radiology means (percu-taneous drainage). In patients where a duodenal fis-tula persists (longer than 23 months with essentiallyunchanged daily output) despite optimal nonopera-tive management, a Roux-en-Y duodenojejunostomyat the site of the duodenal wall defect is usually thera-peutic. Needless to say, such a procedure should bedelayed enough in relation to the time of onset of NPso that the inflammatory process has completelyceased and a safe operation can be performed.

Fistulas of the Small Intestine

In general, enteric fistulas tend to close spontaneously(more often than their duodenal counterparts). Plentyof time should be allowed (months) before the sur-geon decides to proceed with an operation, provided(as previously emphasized) that the fistula and the in-testine are well studied and there is no distal stenosis.In patients with a persisting high-output enteric fis-tula, operative management consists of resection ofthe fistulous tract, usually with the corresponding

bowel segment and an anastomosis between the prox-imal and the distal bowel. Resection of the fistuloustract and oversewing of the bowel wall defect wherethe fistula originated from, although appealing, maybe realistically more technically challenging and mayin fact compromise the diameter of the bowel at thatpoint.

Colonic Fistulas

Colonic fistulas are associated with more severe formsof NP, and the mortality of NP among patients with aconcomitant colonic fistula is higher, highlighting


7/8


the severity of this condition [21]. he portions of thecolon more frequently involved in colonic fistulas arethe transverse because of its close proximity to thelesser sac and the splenic flexure because of its vascu-lar pattern and the subsequent suboptimal blood sup-ply in low-flow states. Colonic fistulas not only reflect

a more severe episode of NP, but also predispose pa-tients to further comorbidity in and of themselves.his is why it has been highly recommended to pro-ceed with immediate fecal diversion [12] as soon as acolonic fistula is diagnosed.

he increased morbidity due to a colonic fistula isusually secondary to undrained intra-abdominal fe-cal collections around the colonic wall defect; thismay give rise to systemic sepsis. Sufficient and opti-mal percutaneous drainage cannot be always achieveddue to the high viscosity and particulate nature of thefecal material. he combination of these factors usu-ally dictate an urgent operation, during which proxi-mal fecal diversion (loop ileostomy preferably, or loopcolostomy) should be the first priority, followed by ex-cellent debridement and some type of colectomy (de-pending on the site of the colonic defect and the ex-tent of colonic wall necrosis).

It is important to realize, however, that it is not themere presence of a colonic fistula that translates tooperative treatment, but rather the concomitant pres-ence of undrained fecal material pooled around thecolonic necrosis and causing sepsis. If such conditions

are proved by CEC scan not to be present and thepatient is not septic, a colonic fistula may in fact beconceptualized as a colostomy and no immediate ac-tion needs to be taken. In such a setting, where co-lonic fistulas are clinically asymptomatic, well con-trolled, have a low output, and there is no distalcolonic obstruction, they can be treated conservative-ly, and spontaneous closure may ensue.

Summary

PFs and IFs are notorious complications of NP, buttheir incidence has decreased due to the recent changeof management scheme for NP consisting of aggres-sive initial hemodynamic support and delayed necro-sectomy, which has led to a reduction in the numberof reoperations required. Precise delineations of thefistulous tract in relation to the pancreatic ductal sys-tem and the pattern of pancreatic duct disruption dic-tate the prognosis and the appropriate managementoption for a PF. Many PFs close spontaneously, somerequire advanced endoscopic intrapancreatic proce-dures (ductal stent placement), and few require opera-

tive intervention. Most upper-gut fistulas close spon-taneously provided that there is no distal intestinalstenosis, whereas colonic fistulas are generally trou-blesome, may lead to sepsis, and require urgent opera-tive treatment.

References

1. Bassi C, Dervenis C, Butturini G, Fingerhut A, et al (2005)Postoperative pancreatic fistula: an international study group(ISGPF) definition. Surgery 138:813

2. Mier J, Luque-de Leon E, Castillo A, et al (1997) Early versuslate necrosectomy in severe necrotizing pancreatitis. Am JSurg 173:7175

3. Fernandez-Del Castillo C, Rattner DW, Makary MA, Mosta-favi A, McGrath D, Warshaw AL (1998) Debridement andclosed packing for the treatment of necrotizing pancreatitis.Ann Surg 228:676684

4. Uhl W, Warshaw A, Imrie C, Bassi C, et al (2002) IAP Guide-lines for the surgical management of acute pancreatitis. Pan-creatology 2:565573

5. Bradley EL III (1993) A clinically based classification systemfor acute pancreatitis. Arch Surg 128:586590

6. siotos GG, Smith CD, Sarr MG (1995) Incidence and man-agement of pancreatic and enteric fistulas after surgicalmanagement of severe necrotizing pancreatitis. Arch Surg130:4852

7. Albridge MC, Francis ND, Glazer G, Dudley HAF (1989)Colonic complications of severe acute pancreatitis. Br J Surg76:362367

8. Buchler MW, Gloor B, Muller CA, et al (2000) Acute necro-tizing pancreatitis: treatment strategy according to the status

of infection. Ann Surg 232:619626 9. Werner J, Feuerbach S, Uhl W, Buchler MW (2005) Manage-

ment of acute pancreatitis: from surgery to interventionalintensive care. Gut 54:426436

10. Connor S, Alexakis N, Raraty MG, et al (2005) Early andlate complications after pancreatic necrosectomy. Surgery137:499505

11. Connor S, Ghaneh P, Raraty MG, et al (2003) Minimallyinvasive retroperitoneal pancreatic necrosectomy. Dig Surg20:270277

12. Ho HS, Frey CF (1995) Gastrointestinal and pancreaticcomplications associated with severe pancreatitis. Arch Surg130:817823

13. Nealon WH, Walser E (2005) Surgical management of

complications associated with percutaneous and/or endo-scopic management of pseudocyst of the pancreas. Ann Surg241:948960

14. Alexakis N, Sutton R, Neoptolemos JP (2004) Surgical treat-ment of pancreatic fistula. Dig Surg 21:262274

15. Howard J, Stonerock CE, Sarkar J, et al (1998) Contem-porary treatment strategies for external pancreatic fistulas.Surgery 124:62733

16. Boerma D, Rauws E, van Gulik , Huibregtse K, Obertop H,Gouma D (2000) Endoscopic stent placement for pancreati-cocutaneous fistula after surgical drainage of the pancreas.Br J Surg 87:15061509

17. Costamagna G, Mutignani M, Ingrosso M, et al (2001) Endo-

scopic treatment of postsurgical external pancreatic fistulas.Endoscopy 33:317322


8/8


18. Kozarek R, Ball , Patterson D, et al (1997) ranspapillarystenting for pancreaticocutaneous fistulas. J GastrointestSurg 1:357361

19. Voss M, Ali A, Eubanks E, Pappas (2003) Surgical man-agement of pancreaticocutaneous fistula. J Gastrointest Surg7:542546

20. Sakorafas GH, siotos GG, Sarr MG (1999) Experience withduodenal necrosis: a rare complication of acute necrotizingpancreatitis. Int J Pancreatol 25:139142

21. Kriwanek S, Gschwantler M, Beckerhinn P, et al (1999) Com-plications after surgery for necrotizing pancreatitis: risk fac-tors and prognosis. Eur J Surg 165:952957

pancreatic and intestinal fistulas

Documents