COLLECTIVE REVIEW

Diagnosis and Management of PancreaticP viJer , FA

Paencreorcuplianthomespetheprodevieuamaserclusigsup

DoproFircysmaoncohclothadebaforlia

WhindthaPadusectissscrofIn1).stuliteorianInevapashoterthelocpooginto

foracuresAPpeproofThthiforthapsethe original Atlanta Classification, because it describes aunique clinical entity and has been used liberally in the

Dis

Received January 18, 2009; Revised February 17, 2009; Accepted April 13,2009.FronioicalCoSto

2Pubrecent literature,10,11 we include this term in this review.In the absence of glandular necrosis, these terms readily

apply. But the Atlanta Classification unfortunately doesnot address fluid collections that develop in the setting of

m the Departments of Surgery, Wilford Hall Medical Center, San Anto-, TX (Cannon) and Beth Israel Deaconess Medical Center, Harvard Med-School, Boston, MA (Callery, Vollmer).rrespondence address: Charles Vollmer Jr, MD, Department of Surgery,neman 9th Floor, 330 Brookline Ave, Boston, MA 02215.

385009 by the American College of Surgeons ISSN 1072-7515/09/$36.00lished by Elsevier Inc. doi:10.1016/j.jamcollsurg.2009.04.017seudocysts: What is the EemyW Cannon, MD, SM, FACS, Mark P Callery, MD

ncreatic pseudocysts represent organized collections ofzyme-rich fluid that persist after an episode of acute pan-atitis (AP), an exacerbation of chronic pancreatitis (CP),pancreatic trauma. These mature collections require ac-rate diagnosis and expert management by a multidisci-nary team of dedicated surgeons, gastroenterologists,d radiologists to minimize morbidity and mortality. Al-ugh most data on the topics of diagnosis and manage-nt of pseudocysts are classified as level IV evidence, pro-ctive studies and cohort data have recently appeared inliterature, calling our historic understanding of thisblem into question. Using the Oxford Levels of Evi-nce and Grades of Recommendation as recently re-wed by Ridgway and Guller,1 this review critically eval-tes the current surgical literature on the diagnosis andnagement of pancreatic pseudocysts in the context of aies of clinically oriented questions.2 Each question con-des with the authors recommendation and a grade as-ned to that recommendation based on the quality of theporting literature.

es the cause of pancreatitis influence thebability of pseudocyst formation?st described in 1761 by Morgagni, pancreatic pseudo-ts represent a widely recognized result of both inflam-tory and traumatic pancreatic ductal disruption. Basedexisting case series, most pseudocysts develop after al-olic pancreatitis, with gallstone pancreatitis ranking ase second. But numerous case series and reports indicatet any cause of pancreatic injury can lead to pseudocystvelopment. Patients with CP who develop acute exacer-tions appear to have a higher incidence of pseudocystmation than patients with AP3-5 while patients with bil-ry AP seem to have the lowest incidence6 (Grade: C).

closure Information: Nothing to disclose.dence?CS, Charles M Vollmer Jr, MD, FACS

at features of an acute fluid collectionicate it will progress to a pseudocyst rathern resolve?ncreatic pseudocysts develop when the main pancreaticct or one of its radicals is disrupted, excreting pancreaticretions into the retroperitoneum or the peripancreaticue planes. A number of different terms are used to de-ibe this accumulated fluid depending on the chronicitythe collection and the underlying pancreatic pathology.1992, the Atlanta Classification was proposed (Table7 Although this terminology is well known, a recentdy showed it has not been universally applied in therature.8 These investigators called for refinement of theginal system to reflect the many variations on imagingd clinical features that exist in patients with pancreatitis.addition, an interobserver agreement study designed toluate a series of nine morphologic descriptors of acutencreatitis as seen on CT has been reported.9 This studywed a high degree of interobserver agreement on sevenms evaluated including presence of a collection, relation ofcollection with the pancreas, content, shape, mass effect,ulated gas bubbles, and air-fluid levels. It has been pro-sed that such terms should supplant the clinical terminol-y presently in use. But because this updated scheme remainsthe developmental phases, the following review will adherethe original Atlanta Classification where possible.According to this system, within the first 4 weeks ofmation, accumulated peripancreatic fluid is labeled ante fluid collection. The majority of these collectionsolve spontaneously, but in 5% to 15% of patients withand in as many as 40% of patients with CP, the fluid

rsists. In these patients, the acute collection produces afound inflammatory response along the serosal surfacesthe adjacent organs, resulting in a fibrous pseudocapsule.is process takes between 4 and 8 weeks, at which points collection becomes a pseudocyst. A pseudocyst thatms after an episode of AP is an acute pseudocyst; onet develops in the setting of CP is labeled a chronicudocyst. Although this latter term was not included in

paqucrohapsewi

salqucrobuevacresho

WhindOnpapesymharemThcatorsivincqurelpse

WhpapaBecupsenesivtheanan

tiothelinatisetresmipaiallesvergroles

tiocatdiamepirOnvidmanoan

Table 1. Summary of 1992 Atlanta Classification Terminology7

Pathology Characteristics

Acuc

Pann

Acup

Pana

AP,

386 Cannon et al Diagnosis and Management of Pancreatic Pseudocysts J Am Coll Surgncreatic necrosiseither sterile or infected. Conse-ently, numerous terms such as walled off pancreatic ne-sis, collection in evolution, organized necrosis and necromave spawned to fill this void (Fig. 1). In this review, alludocysts are considered to be associated with an other-se viable gland.There are no case-control or cohort studies that defineient risk factors for pseudocyst development. One fre-ently referenced study suggests significant pancreatic ne-sis (25%) as a risk factor for pseudocyst development,t this study was a retrospective case series designed toluate the utility of endoscopic retrograde cholangiopan-atography (ERCP) in AP.12 Extrapolations of these datauld be made with caution (Grade: D).

at features of an established pseudocysticate it will persist or become symptomatic?e early observational report found that the majority ofncreatic pseudocysts larger than 6 cm in diameter, whichrsist longer than 6 weeks, result in significant clinicalptoms and complications.13 But subsequent case series

ve found that approximately half of acute pseudocystsain asymptomatic regardless of size or duration.6,14-16

e other half either manifest symptoms or become compli-ed by infection, rupture, hemorrhage, vascular thrombosis,obstruction of adjacent structures.To date, no comprehen-e cohort study has been conducted to evaluate the trueidence of pseudocysts or their natural history. Conse-ently, no prospective indicators have been identified thatiably predict the natural history of an already establishedudocyst (Grade: D).

at preinterventional studies reliably differentiatencreatic pseudocysts from cysticncreatic neoplasms?fore treating any peripancreatic fluid collection, an ac-rate diagnosis must be established. Most importantly, audocyst must be distinguished from a cystic pancreaticoplasm.17Making this distinction requires a comprehen-e assessment of the patient: understanding the history ofdisease process, reviewing available imaging studies,

d in some cases, performing biochemical and cytologicalysis of the peripancreatic fluid.

Abbreviations and Acronyms

AP acute pancreatitisCP chronic pancreatitisERCP endoscopic retrograde cholangiopancreatographyEUS endoscopic ultrasoundMRCP magnetic resonance cholangiopancreatographyWhen an episode of AP results in an acute fluid collec-n that persists on serial imaging over a period of weeks,diagnosis of an acute pseudocyst is assured. This directk between pancreatitis and development of a peripancre-c fluid collectionmay bemore difficult to establish in theting of CP. In addition, cystic pancreatic neoplasms mayult in a low-grade chronic inflammatory process thatmics CP. It is important for the clinician to review thetients complete radiographic history because earlier ax-imaging may define the presence or absence of the cysticion over time. If a clear-cut diagnosis of an inflammatorysus a neoplastic process cannot be made on clinicalunds alone and there is no evidence for a preexistingion, further imaging is indicated.MRI or endoscopic ultrasound (EUS) may reveal septa-ns, solid components within the cyst(s), or a communi-ion between the cyst and themain pancreatic duct. If thegnosis still remains uncertain, more invasive diagnosticasures should be undertaken. Typically this involves as-ation of the cyst for cytology and biochemical testing.e multicenter, retrospective case-controlled study di-ed 112 cases of cystic pancreatic lesions that were ulti-tely surgically resected into mucinous and nonmuci-us groups and compared pre-resection EUS, cytology,d fluid tumor marker levels between the groups.18 A

te fluidollections

Occur early in the course of AP, are located inor near the pancreas, and always lack a wallof granulation or fibrous tissue.

creaticecrosis

Diffuse or focal area(s) of nonviablepancreatic parenchyma, which is typicallyassociated with peripancreatic fat necrosis;nonenhanced pancreatic parenchyma3cm or involving more than 30% of thearea of the pancreas.

teseudocysts

Collection of pancreatic juice enclosed by awall of fibrous or granulation tissue, whicharises as a consequence of AP, pancreatictrauma, or CP; usually round or ovoid andhave a well-defined wall; require 4 or moreweeks from the onset of AP.

creaticbscess

Circumscribed intraabdominal collection ofpus, usually in proximity to the pancreas,containing little or no pancreatic necrosis,which arises as a consequence of AP orpancreatic trauma; occurs later in thecourse of severe AP, often 4 weeks or moreafter onset; the presence of pus and apositive culture for bacteria or fungi, butlittle or no pancreatic necrosis, differentiatea pancreatic or peripancreatic abscess frominfected necrosis.

acute pancreatitis; CP, chronic pancreatitis.

CEnoEUfor

DochchplasyOnevameandeimprenaapageberosrittrefinagatioimnoinjarc

To date, however, there are no studies directly comparingthever(G

WhofpaInsizthaifeplisolofonpainvpasar

bepethatomSypseincsistherupvasPsedointtioriepreplien

IsanTrapseRodraoficaledmeize

Fig(lumneccredeficul

387Vol. 209, No. 3, September 2009 Cannon et al Diagnosis and Management of Pancreatic PseudocystsA level of 192 ng/mL was found to distinguish muci-us from nonmucinous pathologies more accurately thanS morphology and cytology (79% accuracy versus 51%EUS morphology and 59% for cytology) (Grade: C).

endoscopic retrogradeolangiopancreatography or magnetic resonanceolangiopancreatography have any role innning the management of patients withmptomatic pseudocysts?ce the diagnosis of a pancreatic pseudocyst is made,luation of the ductal architecture may affect manage-nt. Several schemes for classifying pancreatic ductalatomy in the setting of a veritable pseudocyst have beenveloped, but no consensus exists on patient selection foraging, the optimal timing of imaging studies, or theferred imaging modality. ERCP and magnetic reso-nce cholangiopancreatography (MRCP) have both beenplied. If considering surgical versus percutaneous drain-, one case series suggests preintervention ERCP shouldperformed to guide clinical management.19 Another ret-pective assessment of an ERCP-based treatment algo-hm showed fewer adverse events in patients in whom theatment algorithm was applied.20 But the benefit of de-ing the ductal anatomy with ERCP must be weighedinst the risk of potentially infecting a sterile fluid collec-n. Because of this concern, ERCP is often performedmediately before a planned intervention.MRCP offers aninvasive alternative and can now be paired with secretinection to provide a functional assessment of the ductalhitecture and physiologic capacity of the parenchyma.

ure 1. Complex lesser-sac fluid collection abutting the stomachen demonstrated with superior arrow) in the setting of extensiverosis of the distal pancreas after an episode of gallstone pan-atitis 2 weeks earlier. The nomenclature for this entity is poorlyned and contributes to confusion among practitioners and diffi-ty in developing comparative studies to address this condition.quality of diagnostic information obtained by MRCPsus ERCP in the setting of pancreatic pseudocystsrade: C).

at is the risk of expectant managementan established, asymptomaticncreatic pseudocyst?contrast to traditional management guided by arbitrarye and duration parameters, current evidence indicatest intervention should be reserved for patients who man-st symptoms or who develop a pseudocyst-related com-cation. But existing management guidelines are basedely on a few level III and IV studies, making the strengththese recommendations limited at best.11 Currently, onlye registered trial on pseudocyst management is enrollingtients,21 which suggests this field remains ripe for clinicalestigations. Because of the limited numbers of eligibletients, though, multicenter collaboration will be neces-y to accrue sufficient power to answer this question.For acute fluid collections, no intervention is requiredcause the majority of these resolve. If the acute collectionrsists to form a pseudocyst, current knowledge suggestst these can still be managed expectantly unless symp-s manifest or complications develop (Grade: D).

mptoms generally stem from the local mass effect of theudocyst or the associated inflammatory response. Theselude abdominal pain, early satiety, weight loss, and per-tent fevers. Potential complications include infection ofpseudocyst, biliary or gastric outflow obstruction, freeture of the pseudocyst into the peritoneal cavity, orcular thrombosis leading to sinistral hypertension.udocyst erosion into adjacent vessels may result in pseu-aneurysm formation or even catastrophic hemorrhageo the gastrointestinal tract or peritoneal cavity. In addi-n to intervening once a complication develops, an expe-nced pancreatic surgeon should not hesitate to interveneemptively if imaging features suggest an imminent com-cation such as erosion into the splenic hilum, a threat-ing pseudoaneurysm, or evolving sinistral hypertension.

endoscopic drainage of pseudocysts as safed effective as surgical drainage?ditional open surgical approaches to acute, symptomaticudocysts include cyst-gastrostomy, cyst-duodenostomy,ux-en-Y cyst-jejunostomy, and, in rare cases, externalinage. No studies exist to guide patient selection or typeoperation, although anatomic cyst topography is a crit-l factor. Recent advances in endoscopic capabilities haveto the development of an array of nonsurgical drainagethods that bear consideration as well. To date, random-d comparison of endoscopic management of pseudo-

cysts versus surgical management has not been performed,sostapatanpathoingropauntiewhstrplidethesho(2.of

chrosachtowe92740.0crowiNSofpachcyspsewarecteclednedawi

ingsurproIndores

20 procedures, respectively) and fewer days to resolution(33cyspseanageloowi

entheneEUadinjtifycysthepselatplorecturrensteaftcumopesmthemotheclosurenpsecysgutat(G

IspaPerecsocgreanneprapa

388 Cannon et al Diagnosis and Management of Pancreatic Pseudocysts J Am Coll Surgthe advantages of one approach over the other cannot beted with certainty. A recent retrospective study com-red 79 patients who suffered complications from percu-eous drainage, endoscopic drainage, or both, with 100tients who underwent surgical intervention alone.22 Al-ugh this study suggested that fewer complications occurpatients undergoing primary surgical intervention, theseups cannot be directly compared because the number oftients undergoing successful nonsurgical intervention isknown. Another retrospective cohort study of 10 pa-nts who underwent surgical cyst-gastrostomy versus 20o underwent endoscopic transgastric drainage demon-ated no difference in treatment success, procedural com-cations, or reintervention, although the study was un-rpowered to detect clinically important differences inse measures.23 But it did demonstrate a significantlyrter hospital length of stay in the endoscopic group65 versus 6.5 days, p 0.008) and a mean cost savings$5,738 per patient for endoscopic drainage.Endoscopic drainage has been applied to both acute andronic pseudocysts and pancreatic necrosis.24 In this ret-pective review, resolution of the fluid collection wasieved in 113 of 138 patients (82%), with a median timeresolution of 40 days. Patients with chronic pseudocystsre more likely to have resolution (59 of 64 patients,%) than those with acute pseudocysts (23 of 31 patients,%, p 0.02) or necrosis (31 of 43 patients, 72%, p 06). Complications were greatest in patients with ne-sis (37%) versus chronic pseudocysts (17%, p 0.02),th a similar trend versus acute pseudocysts (19%, p ). After a median of 2.1 years, recurrence occurred in 18113 patients (16%), with the greatest recurrence seen intients with necrosis (29%), which was higher than withronic pseudocysts (12%, p 0.047) and acute pseudo-ts (9%, p NS). On multivariate analysis, chronicudocysts were a marker for successful drainage; necrosiss a marker for unsuccessful drainage, complications, andurrence. These authors demonstrated viability of thishnique for treating symptomatic pseudocysts, which hasto further development of this approach. For pancreaticcrosis, however, surgical debridement remains the stan-rd, although efforts to refine this technique for patientsth necrosis continue.25

Endoscopic pseudocyst drainage has been described us-both a transpapillary and a transenteric approach. Notprisingly, increased endoscopic experience with thesecedures correlates with improved patient outcomes.26

this retrospective study, performance of 20 or more en-scopic drainage procedures afforded improved rates ofolution (93% versus 45% for more than versus less than.5 days versus 50 days) in patients with chronic pseudo-ts. The transpapillary approach requires that theudocyst communicate with the main pancreatic ductd that it have few septations to permit complete drain-. Pancreatic ductal strictures, if identified, may be bal-n dilated, after which a single 57 F stent is placedthin the pancreatic duct.The transenteric endoscopic approach requires either andolumenal bulge or EUS evidence of adherence betweengastric or duodenal wall and the cyst without associatedcrosis, but such simplicity rarely exists. Navigation withS theoretically permits localization and avoidance ofjacent vessels that could lead to significant hemorrhage ifured during attempted endoscopic drainage. After iden-ing the pseudocyst, aspiration confirms access to thet cavity. A contrast injection can be performed for fur-r confirmation if required. Once access is established, audocystotomy is performed and the tract balloon di-ed. One or more double pigtail stents can then be de-yed to maintain patentcy of the cyst-enterostomy. Oneent randomized, prospective study suggests that prema-e removal of these stents leads to pseudocyst recur-ce.27 Fifteen patients were randomized to have theirnts left in place indefinitely; 13 underwent stent retrievaler a median of 2 months. The primary pseudocyst re-rred in five patients after stent removal at a median of 6nths after initial drainage versus none in the group withrsistent stent-facilitated drainage (p 0.013). Althoughall and potentially underpowered to detect recurrence instent maintenance group, this study suggested that re-val of endoscopically placed transenteric stents increasesrate of recurrence, likely from obstruction or prematuresure of the cyst-enterostomy. An ongoing trial using agical stapling device to create the endoscopic cyst-terostomy may address this limitation of endoscopicudocyst management by creating a larger opening fort decompression.21 At present, without clear data toide treatment selection, the approach used is often dic-ed by the skill set of the physician caring for the patientrade: C).

there any role for percutaneous drainage ofncreatic pseudocysts?rcutaneous pseudocyst drainage has also been used. But aent cohort study showed percutaneous drainage was as-iated with higher mortality, longer hospital stay, andater complications than surgical drainage.28 This resultd numerous observational studies indicate that percuta-ous drainage, although seemingly convenient for thectitioner and the patient, should be performed only intients with an acute pseudocyst and radiographically

proven normal ductal anatomy or in cases in which comor-bidterdranicunroscystiethedra(7%weage

paproiednetiowhearmaquconcysstascodra

IsovLawhmiA nap(G

WhpreMwebaenabearBupecysrimges

thenediatolIntivtercasthe(Fimeindbeint

canassentenpseincristicofquanofbe

thapatma

Figlesshoan

389Vol. 209, No. 3, September 2009 Cannon et al Diagnosis and Management of Pancreatic Pseudocystsconditions or physiologic exhaustion make surgical in-vention ill advised. Predictors of failure of percutaneousinage include a pancreatic duct cut-off, direct commu-ation of the pseudocyst with the pancreatic duct, andderlying CP, according to one case series.19 Another ret-pective case series compared 66 patients with pseudo-ts who underwent percutaneous drainage with 41 pa-nts who were observed and 66 who underwent surgicalrapy.29 Of those who underwent initial percutaneousinage, 38 of 66 (58%) failed as compared with 3 of 41) who were observed and 8 of 66 (12%) who under-

nt operations. Of the 38 who failed percutaneous drain-, 33 (87%) required operations and 4 died.The practice of using percutaneous drainage to treat ancreatic abscess in order to avoid the theoretic risk ofgression to retroperitoneal sepsis has not been well stud-. Percutaneous drainage has been used in early infectedcrosis as a guide for minimally invasive surgical interven-n either immediately30 or in a delayed fashion in patientso fail to resolve.31 Such an approach has merit in cases ofly necrosis with uncontrolled sepsis because the patienty resolve with percutaneous drainage or with subse-ent minimally invasive removal of the necrotic debris. Intrast, patients with a symptomatic or infected pseudo-t (ie, a pancreatic abscess) are typically physiologicallyble and are best managed with a single surgical or endo-pic procedure without the preamble of percutaneousinage (Grade: B).

there any advantage to laparoscopic drainageer open surgical drainage?paroscopic drainage of pseudocysts has been performedere the minimally invasive surgical approach essentiallymics traditional open cyst-enteric drainage techniques.umber of case series have been reported that suggest thisproach is safe, although definitive evidence is lackingrade: D).

at features of pseudocysts indicate adilection for developing complications?anagement of pseudocyst complications has not beenll studied to date. Consequently, current approaches aresed primarily on observational data and surgeon experi-ce. Infection of the pseudocyst resulting in a pancreaticscess can rapidly progress to retroperitoneal sepsis, soly recognition and prompt intervention are required.t distinguishing those clinical features associated withripancreatic inflammation from infection of the pseudo-t poses significant challenges. Imaging findings such asenhancement or gas within the pseudocyst clearly sug-t an infection (Fig. 2). In the absence of such findings,treating physician may consider percutaneous fine-edle aspiration to evaluate for infection. But this invasivegnostic test should be pursued only if the patient canerate a major surgical intervention based on the results.such cases, infection of the pseudocyst can be defini-ely managed with direct surgical drainage into the en-ic system. This approach is especially important in thee of the disconnected pancreatic segment, in whichdrainage of the distal gland is completely disrupted

g. 3). Although percutaneous drainage has been recom-nded by some for treatment of pancreatic abscess, asicated above, this strategy should be used with cautioncause it typically only forestalls the inevitable surgicalervention.Other complications, including hemorrhage and rupture,present as life-threatening emergencies. Pseudocyst-

ociated hemorrhage generally should be controlled withdovascular embolization of the affected vesselmost of-a branch of the splenic or gastroduodenal artery. Someudocysts, including those in the pancreatic tail, whichorporate into the splenic hilum, carry an especially highk of acute hemorrhage and should be treated prophylac-ally with surgical resection. Although rare, free rupturethe pseudocyst typically results in peritonitis and re-ires emergency open surgery with abdominal washoutd external drainage of the pseudocyst cavity. Predictorsthese events have not been evaluated in the literaturecause of their relative infrequency.Giant pseudocysts have been variably defined as greatern 15 cm32 or greater than 10 cm.33 Initial observations thatients with pseudocysts larger than15cm require a uniquenagement approach have not been corroborated with case-

ure 2. Infected pancreatic pseudocyst. This well-circumscribedion, in direct connection with a viable pancreatic head (notwn), illustrates evidence of a gas-forming infection 5 weeks afterattack of alcohol-induced pancreatitis.

conandcmpatinmetecstuorinc

WhchChchtheobpadofrosicstaonan

maaspsetretiechwiwistu

procedure to the longitudinal pancreaticojejunostomy in-cretrawotiewicrerespsetheheinefftessplpaanbelor

WhtreAfveistadascrthepafolummecliasy

DIExrelinagedraforthehintheinsageseradfut

top

Figpseetychytinuenl

390 Cannon et al Diagnosis and Management of Pancreatic Pseudocysts J Am Coll Surgtrolled or cohort data. In the case-control study by Solianicolleagues,33 41 patients with pseudocysts greater than 10(including 19 greater than 14 cm)were comparedwith 30ients with pseudocysts smaller than 10 cm. No differencesmorbidity,mortality, or recurrenceweredetected after treat-nt, although this study was arguably underpowered to de-t clinically significant differences. Based on these limiteddies, it appears that there are no reliable clinical indicatorsanatomic features that predict pseudocyst complications,luding pseudocyst size (Grade: C).

at treatment approach should be used for aronic pancreatic pseudocyst?ronic pseudocyststhose arising in the setting ofronic pancreatitisbear special consideration becausey are often associated with an underlying stricture orstruction in the main pancreatic duct. In these cases, thencreatic ductal pathology must be addressed either en-scopically or surgically to avoid pseudocyst recurrencem a persistent distal obstruction. Because of this intrin-difference between acute and chronic pseudocysts, thetus of the underlying gland should be firmly establishedclinical grounds or with confirmatory imaging beforey planned intervention.Once the diagnosis of a chronic pseudocyst has beende, treatment options include transampullary stentingdescribed earlier, longitudinal pancreaticojejunostomy,udocyst-enteric drainage, or resection. Endoscopicatment is generally used as a first line treatment in pa-nts without chronic pain, biliary obstruction, or largeronic pseudocysts involving the pancreatic tail. Patientsth a large main pancreatic duct (7mm) can be treatedth a longitudinal pancreaticojejunostomy. One recentdy indicates that adding a separate pseudocyst drainage

ure 3. Disconnected pancreatic segment. This well-developedudocyst progressed in size over 6 weeks, leading to early sati-. Notice the viable (enhancing) remnant of pancreatic paren-ma in the tail (circle). The drainage of this segment is in discon-ity with the pancreatic head and is the source of persistentargement of the cyst over time.ases operative time, hospital days, complications, andnsfusion requirements with no benefit added.34 In otherrds, drainage of the ductal system alone suffices. In pa-nts with a chronic pseudocyst in the pancreatic headth or without biliary or gastric outlet obstruction, pan-aticoduodenectomy or duodenal-preserving pancreaticection should be considered. Finally, as described above,udocysts in the pancreatic tail carry a risk of rupture intospleen or splenic vasculature, resulting in catastrophic

morrhage, so operative removal is advised. In addition,patients with associated splenic vein thrombosis, theects of sinistral hypertension can lead to upper gastroin-tinal bleeding. Consequently, distal pancreatectomy andenectomy should be strongly considered for these selecttients. Because of the variability in pseudocyst locationd ductal anatomy, chronic pseudocyst treatment cannotnarrowly proscribed because the treatment must be tai-ed to the individual clinical situation (Grade: C).

at followup imaging should be obtained afteratment of a pancreatic pseudocyst?ter definitive treatment of pancreatic pseudocysts, sur-llance imaging is often used but has not been studied orndardized with regard to either timing or imaging mo-lity. Ultrasonography, CT, and MRI have all been de-ibed for this purpose. Although no consensus exists onneed for or timing of post-treatment studies, if the

tients symptoms do not improve shortly after treatment,lowup imaging is obviously warranted. In addition, doc-entation of cyst resolution at some interval after treat-nt should also be considered. There is no clearly definednical or economic value for longitudinal followup in themptomatic, postintervention patient (Grade: D).

SCUSSIONisting studies on pancreatic pseudocysts are hampered byative infrequency, inconsistent terminology, differencesmethodology, selection bias, and lack of uniform man-ment principles that limit the conclusions that can bewn from them and preclude combining study groupsfurther analysis. With a paucity of randomized studies,state of the art of pancreatic pseudocyst managementges largely on observational data from case series. None-less, a few contemporary studies afford some reliableights into the correct approach for diagnosis and man-ment of this challenging clinical problem and shouldve as the foundation for future work in this field. Indition, this review has identified a number of areas forure study that bear summarizing.Results of the limited case-control and cohort studies onics related to pancreatic pseudocysts are presented in

Table 2. Summary of Pancreatic Pseudocyst Studies*

First author YearLevel ofevidence

Randomizedgroups, n Intervention/design

Medianfollowup Major endpoint Minor endpoint Interpretations/comments

Bru

Ahe

Mo

Var

Bar

391Vol. 209, No. 3, September 2009 Cannon et al Diagnosis and Management of Pancreatic Pseudocystsgge18 2004 IIIb Retrospective analysisof mucinous (n 68) versusnonmucinous (n 44) pancreatic cystsincluding 27inflammatory cysts

Unknown Receiver-operatorcharacteristiccurves of tumormarkers

Comparison ofsensitivity,specificity,andaccuracy oftumormarkerswithmorphologyandcytology

This retrospective case-control studyprovides the current bestevidence for distinguishingneoplastic pancreatic cysts frombenign cysts and suggests thatCEA was more accurate thanEUS morphology and cytology atmaking this distinction.

arne20 1992 IIIb Retrospectiveapplication of atreatment algorithm

Unknown Adverse outcomes(persistent orrecurrentpseudocyst orcomplicationrequiringadditionaltreatment orhospitalization

In this retrospective algorithmassessment, the authors providethe first evidence that pancreaticductal anatomy should guidetherapy. Adverse events occurredin 6 of 14 patients who did notfollow the algorithm (43%)versus 3 of 26 who did (12%,p 0.04). This algorithmapplied percutaneous drainagemore liberally than those ofsubsequent investigators.

rton28 2005 IIb Cohort analysis of TheNational InpatientSample

Unknown Complication rates,length of stay,disposition,inpatientmortality

This study includes 14,530 patientswith pancreatic pseudocyststreated with either surgical orpercutaneous drainage. Aftercontrolling for confoundingvariables, surgical versuspercutaneous drainage had alower mortality (2.8% versus5.9%), shorter length of stay (15versus 21 days), and many fewercomplications. These data alsoindicated a protective benefit ofERCP (odds ratio 0.68, 95% CI0.510.9).

adarajulu23 2008 IIIb Nonrandomizedretrospective reviewof patientsundergoing surgical(n 10) versusendoscopic (n 20) cyst-gastrostomy

Treatment success,proceduralcomplications,reinterventions,postprocedurelength of stay,mean direct cost

This small retrospective, matched,case-controlled study comparedthe periprocedure outcomes,length of stay, and cost ofendoscopic cyst-gastrostomy in20 patients with surgical cyst-gastrostomy in 10 patients. Therewas no detected difference inperiprocedural outcomes; lengthof stay (2.65 versus 6.5 d) andcosts ($9,077 versus $14,815)were lower with endoscopicversus surgical intervention.

on24 2002 IIIb Nonrandomizedretrospective reviewof endoscopicdrainage of acutepseudocysts (n 31), chronicpseudocysts (n 64), and pancreaticnecrosis (n 43).

2.1 y Acute resolution,complications,and recurrence

This is a relatively large series of 138patients with pancreaticpseudocysts and pancreaticnecrosis who underwentattempted endoscopic drainage.Acute resolution was achieved in113 of 138 patients (82%).Comparisons between groupsshowed that complications andinterval recurrence were greaterin patients with necrosis versuschronic pseudocysts, with similartrends for both when comparingnecrosis with acute pseudocysts.Endoscopic management ofpseudocysts warrants furtherstudy but such techniques shouldnot be applied to patients withnecrotic debris within theperipancreatic fluid.

(continued)

Taguanacca timchateIf aanhatheimonpsesul

toanreqwihotifyusiacupro

whife

Table 2. Continued

FirsLevel of Randomized Median

wup

Har own

Ary o

Soli own

*FirER sound

392 Cannon et al Diagnosis and Management of Pancreatic Pseudocysts J Am Coll Surgble 2. In summary, these studies show that in distin-ishing pseudocysts from cystic pancreatic neoplasms,elevated CEA level within the cyst fluid is the mosturate predictor of a neoplastic process.18 In designingreatment approach, ERCP-based management mayprove outcomes by defining the pancreatic ductal ar-itecture.20 Percutaneous drainage is generally associ-d with worse outcomes than surgical management.28

n endoscopic management approach is attempted foracute or chronic pseudocyst,24 the endoscopist shouldve significant experience with the procedure,26 andendoscopically placed stent should not be removedmediately on pseudocyst resolution.27 Finally, baseda small case-control series, management of giantudocysts with surgical cyst-gastrostomy did not re-t in worse clinical outcomes.33

Although these limited studies do provide some insight,truly understand the natural history and multiple nu-ces of this disease process, a well-designed cohort study isuired. A large dataset encompassing a sizable populationth patients at risk (primarily patients who consume alco-l or who have cholelithiasis) would be required to iden-sufficient numbers of affected patients. Such a study

ng standard terminology could define the incidence ofte fluid collections, the frequency with which thesegress to form a pseudocyst, and the frequency with

t author Year evidence groups, n Intervention/design follo

ewood26 2003 IIIb Nonrandomizedretrospective reviewof results beforeand after 20endoscopicdrainage procedures

Unkn

anitakis27 2007 IIIb 2 Prospective case-controlled series

14 m

ani33 2004 IIIb Nonrandomizedretrospective reviewof necroticpseudocysts froman attack of acutepancreatitis

Unkn

st author, Level IV studies excluded.CP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultraich established pseudocysts become complicated, man-st symptoms, or even regress.Then, as noted earlier, a multicenter group of collabo-ing investigators should design and conduct studies todress existing controversies and questions related to theluation and management of established pseudocysts.me of these unanswered questions include optimal man-ment of pancreatic abscess, the accuracy of pancreaticctal anatomy documented by MRCP, the merits of en-scopic versus surgical therapy for acute and chronicudocysts, defining clinical and imaging features thatticipate future complications in an asymptomaticudocyst, and the need for, and timing of, postinterven-n imaging of pseudocysts, just to name a few. As withny other clinical entities for which there are numerousatment options, a clinical pathway for pseudocyst man-ment should also be considered because this may allowstreamlined first-line therapy, accelerated recognition ofatment failures, and improved patient outcomes. Asth many surgical problems, existing studies on the man-ment of pancreatic pseudocyst have laid the ground-rk for our current clinical thinking but in no way repre-t a scientifically robust definitive answer on how toproach this condition and all of its nuances. Future stud-must seek to better define this clinical entity and itstimal management.

Major endpoint Minor endpoint Interpretations/comments

Pseudocystresolution, timeto resolution,procedurefailure,complications,length of stay,recurrence

This study showed a significantimprovement in frequency ofresolution and decreased lengthof stay for endoscopic drainage ofchronic pseudocysts when theendoscopist had performed over20 procedures.

Recurrence of theindex pseudocyst

This is a well-designed study of 15patients who had percutaneousendolumenal stents maintainedversus 13 patients where thestents were removed onresolution of the cyst. Kaplan-Meier curves show a recurrencerate of 40% at 10 months in thelatter group. Because of the smallsize of the endoscopic cyst-enterostomy, this result shouldcome as no surprise.

Morbidity,mortality, andrecurrence

This study does not clearly excludepatients with associatedpancreatic necrosis. It doessuggest, however, that patientswith giant pseudocysts can safelyundergo cyst-gastrostomy,although it is significantlyunderpowered to detectimportant differences in any ofthe endpoints.

.ratadevaSoagedudopseanpsetiomatreagefortrewiagewosenapiesop

REFERENCES

1. Ridgway PF, Guller U. Interpreting study designs in surgicalresearch: a practical guide for surgeons and surgical residents.J Am Coll Surg 2009;208:635645.

2. Oxford Centre for Evidence-Based Medicine Levels of Evidence

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19. Nealon WH, Walser E. Main pancreatic ductal anatomy candirect choice of modality for treating pancreatic pseudocysts(surgery versus percutaneous drainage). Ann Surg 2002;235:751758.

20. Ahearne PM, Baillie JM, Cotton PB, et al. An endoscopic ret-

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

393Vol. 209, No. 3, September 2009 Cannon et al Diagnosis and Management of Pancreatic Pseudocysts(May 2001). Available at: http://www.cebm.net/index.aspx?o1025. Accessed January 9, 2009.Imrie CW, Buist LJ, Shearer MG. Importance of cause in theoutcome of pancreatic pseudocysts. Am J Surg 1988;156:159162.Ammann RW, Akovbiantz A, Largiader F, Schueler G. Courseand outcome of chronic pancreatitis. Longitudinal study of amixed medical-surgical series of 245 patients. Gastroenterology1984;86:820828.Kourtesis G, Wilson SE, Williams RA. The clinical significanceof fluid collections in acute pancreatitis. Am Surg 1990;56:796799.Maringhini A, Uomo G, Patti R, et al. Pseudocysts in acutenonalcoholic pancreatitis: incidence and natural history. DigDis Sci 1999;44:16691673.Bradley EL, 3rd. A clinically based classification system for acutepancreatitis. Summary of the International Symposium onAcute Pancreatitis, Atlanta, GA, September 1113, 1992. ArchSurg 1993;128:586590.BollenTL, van Santvoort HC, Besselink MG, et al. The AtlantaClassification of acute pancreatitis revisited. Br J Surg 2008;95:621.van Santvoort HC, Bollen TL, Besselink MG, et al. Describingperipancreatic collections in severe acute pancreatitis using mor-phologic terms: an international interobserver agreement study.Pancreatology 2008;8:593599.Baillie J. Pancreatic pseudocysts (part I). Gastrointest Endosc2004;59:873879.Jacobson BC, BaronTH, Adler DG, et al. ASGE guideline: Therole of endoscopy in the diagnosis and the management of cysticlesions and inflammatory fluid collections of the pancreas. Gas-trointest Endosc 2005;61:363370.Neoptolemos JP, London NJ, Carr-Locke DL. Assessment ofmain pancreatic duct integrity by endoscopic retrograde pancre-atography in patients with acute pancreatitis. Br J Surg 1993;80:9499.Bradley EL, Clements JL Jr, Gonzalez AC.The natural history ofpancreatic pseudocysts: a unified concept of management. Am JSurg 1979;137:135141.Vitas GJ, Sarr MG. Selected management of pancreatic pseudo-cysts: operative versus expectant management. Surgery 1992;111:123130.Warshaw AL, Rattner DW.Timing of surgical drainage for pan-creatic pseudocyst. Clinical and chemical criteria. Ann Surg1985;202:720724.Yeo CJ, Bastidas JA, Lynch-Nyhan A, et al. The natural historyof pancreatic pseudocysts documented by computed tomogra-phy. Surg Gynecol Obstet 1990;170:411417.Warshaw AL, Rutledge PL. Cystic tumors mistaken for pancre-atic pseudocysts. Ann Surg 1987;205:393398.Brugge WR, Lewandrowski K, Lee-Lewandrowski E, et al. Diag-nosis of pancreatic cystic neoplasms: a report of the cooperativepancreatic cyst study. Gastroenterology 2004;126:13301336.rograde cholangiopancreatography (ERCP)-based algorithm forthe management of pancreatic pseudocysts. Am J Surg 1992;163:111115; discussion 115116.Natural orifice translumenal endoscopic surgery (NOTES) cys-tgastrostomy for the treatment of pancreatic pseudocysts(NCT00541593). Available at: http://www.clinicaltrials.gov.Accessed June 29, 2008.Nealon WH, Walser E. Surgical management of complicationsassociated with percutaneous and/or endoscopic managementof pseudocyst of the pancreas. Ann Surg 2005;241:948957;discussion 957960.Varadarajulu S, Lopes TL, Wilcox CM, et al. EUS versus surgi-cal cyst-gastrostomy for management of pancreatic pseudocysts.Gastrointest Endosc 2008;68:649655.Baron TH, Harewood GC, Morgan DE, Yates MR. Outcomedifferences after endoscopic drainage of pancreatic necrosis,acute pancreatic pseudocysts, and chronic pancreatic pseudo-cysts. Gastrointest Endosc 2002;56:717.Papachristou GI, Takahashi N, Chahal P, et al. Peroral endo-scopic drainage/debridement of walled-off pancreatic necrosis.Ann Surg 2007;245:943951.Harewood GC, Wright CA, Baron TH. Impact on patient out-comes of experience in the performance of endoscopic pancre-atic fluid collection drainage. Gastrointest Endosc 2003;58:230235.Arvanitakis M, Delhaye M, Bali MA, et al. Pancreatic-fluid col-lections: a randomized controlled trial regarding stent removalafter endoscopic transmural drainage. Gastrointest Endosc2007;65:609619.Morton JM, Brown A, Galanko JA, et al. A national comparisonof surgical versus percutaneous drainage of pancreatic pseudo-cysts: 19972001. J Gastrointest Surg 2005;9:1520; discus-sion 2021.Heider R, Meyer AA, Galanko JA, Behrns KE. Percutaneousdrainage of pancreatic pseudocysts is associated with a higherfailure rate than surgical treatment in unselected patients. AnnSurg 1999;229:781787; discussion 787789.CarterCR,McKayCJ, ImrieCW.Percutaneous necrosectomy andsinus tract endoscopy in the management of infected pancreaticnecrosis: an initial experience. Ann Surg 2000;232:175180.Horvath KD, Kao LS, Wherry KL, et al. A technique forlaparoscopic-assisted percutaneous drainage of infected pancre-atic necrosis and pancreatic abscess. Surg Endosc 2001;15:12211225.Johnson LB, Rattner DW,Warshaw AL. The effect of size of giantpancreatic pseudocysts on the outcome of internal drainage proce-dures. Surg Gynecol Obstet 1991;173:171174.Soliani P, Ziegler S, Franzini C, et al. The size of pancreaticpseudocyst does not influence the outcome of invasive treat-ments. Dig Liver Dis 2004;36:135140.Nealon WH, Walser E. Duct drainage alone is sufficient in theoperative management of pancreatic pseudocyst in patients withchronic pancreatitis. Ann Surg 2003;237:614620; discussion620621.

Diagnosis and Management of Pancreatic Pseudocysts: What is the Evidence?Does the cause of pancreatitis influence the probability of pseudocyst formation?What features of an acute fluid collection indicate it will progress to a pseudocyst rather than resolve?What features of an established pseudocyst indicate it will persist or become symptomatic?What preinterventional studies reliably differentiate pancreatic pseudocysts from cystic pancreatic neoplasms?Do endoscopic retrograde cholangiopancreatography or magnetic resonance cholangiopancreatography have any role in planning the management of patients with symptomatic pseudocysts?What is the risk of expectant management of an established, asymptomatic pancreatic pseudocyst?Is endoscopic drainage of pseudocysts as safe and effective as surgical drainage?Is there any role for percutaneous drainage of pancreatic pseudocysts?Is there any advantage to laparoscopic drainage over open surgical drainage?What features of pseudocysts indicate a predilection for developing complications?What treatment approach should be used for a chronic pancreatic pseudocyst?What followup imaging should be obtained after treatment of a pancreatic pseudocyst?DISCUSSIONREFERENCES