EUS versus surgical cyst-gastrostomy for management of pancreatic pseudocysts

ORIGINAL ARTICLE: Clinical Endoscopy

EUS versus surgical cyst-gastrostomy for management of pancreatic pseudocysts Shyam Varadarajulu, MD, Tercio L. Lopes, MD, MSPH, C. Mel Wilcox, MD, Ernesto R. Drelichman, MD, Meredith L. Kilgore, PhD, John D. Christein, MD Birmingham, Alabama, USA

Background: Although EUS-guided cyst-gastrostomy is increasingly being performed, there are no studies that compare the clinical outcomes and cost-effectiveness with surgical cyst-gastrostomy. Objectives: To compare the clinical outcomes of EUS-guided cyst-gastrostomy with surgical cyst-gastrostomy for the management of patients with uncomplicated pancreatic pseudocysts and to perform a cost analysis of each treatment modality. Design: A retrospective case-controlled study. Setting: A tertiary-referral center. Patients: Consecutive patients with uncomplicated pancreatic pseudocysts managed by surgical and EUSguided cyst-gastrostomy. Methods: An independent observer blinded to all clinic outcomes matched each patient who underwent a surgical cyst-gastrostomy with 2 patients who underwent an EUS-guided cyst-gastrostomy for age, etiology of pancreatitis, and the size of the pseudocyst. Main Outcome Measurements: Rates of treatment success, complications, and reinterventions; length of postprocedure hospital stay; and cost associated with each treatment modality. Results: Ten patients (6 men; mean age 42.3 years, range 22-65 years) who underwent surgical cyst-gastrostomy were matched with 20 patients who underwent an EUS-guided cyst-gastrostomy. There were no significant differences in demographics, major comorbidities, and clinical characteristics between both cohorts. Although there were no significant differences in rates of treatment success (100% vs 95%, P Z .36), procedural complications (none in either cohort), or reinterventions (10% vs 0%, P Z.13) between surgery versus an EUS-guided cyst-gastrostomy, the mean length of a postprocedure hospital stay for an EUS-guided cyst-gastrostomy was significantly shorter than for surgical cyst-gastrostomy (2.65 vs 6.5 days, P Z.008). The average direct cost per case for EUS-guided cyst-gastrostomy was significantly less when compared with surgical cyst-gastrostomy ($9077 vs $14,815, P Z .01), which corresponded to a cost savings of $5738 per patient. Limitations: Retrospective, nonrandomized design; patients with pancreatic abscess or necrosis were not evaluated; a limited sample size and a short duration of follow-up. Conclusions: EUS-guided cyst-gastrostomy should be considered as a first-line treatment approach for patients with uncomplicated pancreatic pseudocysts, because the procedure is cost saving and is associated with a shorter length of a postprocedure hospital stay when compared with surgical cyst-gastrostomy. There was no significant difference in clinical outcomes between both treatment modalities. (Gastrointest Endosc 2008;68:649-55.)

See CME section; p. 731. Copyright ª 2008 by the American Society for Gastrointestinal Endoscopy 0016-5107/$34.00 doi:10.1016/j.gie.2008.02.057

Management of pancreatic pseudocysts has traditionally been surgical.1-3 More recently, transmural endoscopic drainage of pancreatic pseudocysts has been increasingly accepted as a minimally invasive alternative to surgical drainage, with good clinical outcomes in patients in whom it is technically successful.4-6 The procedure entails the creation of a fistula between the pseudocyst and the

www.giejournal.org

Volume 68, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 649

Abbreviations: CBD, common bile duct; DPEJ, direct percutaneous jejunostomy; IV, intravenous; PFC, peripancreatic fluid collections; UAB, University of Alabama at Birmingham.

Management of pancreatic pseudocysts

Varadarajulu et al

gastric lumen (cyst-gastrostomy) or duodenal lumen (cystduodenostomy). A transmural stent or a nasocystic catheter is then deployed within the pseudocyst to facilitate drainage. The 2 major drawbacks of this relatively ‘‘blind’’ approach are perforation and hemorrhage.7,8 This limitation is currently overcome with the development of EUS that permits proper selection of an optimal site for the needle puncture, thus avoiding adjacent vasculature and organs.9-11 Also, unlike conventional transmural drainage, EUS permits drainage of pseudocysts that do not cause luminal compression.12 The major advantage of EUS-guided cyst-gastrostomy over surgery is that, besides being minimally invasive, patients can be discharged home earlier, thus potentially resulting in cost savings. Debilitated patients and those with prohibitive comorbid medical conditions can also be treated by this approach without the need for general anesthesia and the stress of an open abdominal operation. However, to our knowledge, there are no studies that analyzed cost or compared the clinical outcomes between surgical and EUS-guided cyst-gastrostomy. The aim of this retrospective study was to compare the clinical outcomes between surgical and EUS-guided cystgastrostomy and to perform a cost analysis of both treatment modalities.

by a radiologist. None of the patients who underwent a surgical cyst-gastrostomy had a preoperative ERCP or EUS for assessment of ductal anatomy or pseudocyst morphology.

PATIENTS AND METHODS

Protocol for the surgical cyst-gastrostomy approach

A retrospective case-controlled study was conducted that included consecutive patients (O18 years of age) who underwent surgical cyst-gastrostomy and EUS-guided cyst-gastrostomy at a tertiary-referral center between July 2005 and June 2007. Patients underwent an EUS or a surgical cyst-gastrostomy based on the clinical service (medical vs surgical gastroenterology) to which they were admitted. Patients were identified from the surgical and endoscopy databases, and patient medical records and CTs were individually reviewed to verify procedural indication, patient demographics, pseudocyst characteristics, technical and treatment outcomes, and complications. Financial records were accessed from the University of Alabama (UAB) Hospital Accounting Department. For every patient who underwent surgical cyst-gastrostomy, 2 patients who underwent EUS-guided cyst-gastrostomy were matched by an independent observer for all of the following variables: patient age (range to not exceed 10 years, patients O70 years were excluded), etiology of pancreatitis, and pseudocyst size (range not to exceed O2 cm). The observer was blinded to all clinical outcomes. All procedures were performed by one pancreaticobiliary surgeon and one endosonographer. All patients underwent a contrastenhanced CT of the abdomen at our institution before undergoing a cyst-gastrostomy. All patients with a pancreatic abscess or necrosis as suspected by CTwere excluded. The pancreatic-fluid collection was categorized according to the Atlanta classification,13 based on CT imaging reviewed

Patients were placed in the supine position, and intravenous (IV) cefazolin (1 g) was administered before incision. A limited upper midline incision was made, approximately 10 cm in length at the middle third of the distance from the umbilicus to the xiphoid process, to allow access to the abdomen. The pseudocyst was palpable posterior to the stomach. Cautery was used to create an approximate 5-cm longitudinal gastrotomy near the greater curvature of the fundus. Palpation and pseudocyst localization with 18-gauge needle aspiration guided the posterior gastric incision. Cautery was used to incise an approximate 2 cm opening in the posterior gastric wall. The pseudocysts were aspirated and irrigated. No attempt was made to dissect the posterior gastric wall from the inflammatory reaction in the lesser sac. An incisional biopsy specimen of the pseudocyst wall was routinely evaluated with frozen section to rule out neoplasm. An endovascular stapler with a 2.5-mm staple load was used to create an approximate 6-cm cyst-gastrostomy. A nasogastric tube was left in the stomach. The anterior gastrotomy was then closed in one layer with a running 3-0 suture. Based on patient nutritional status, a feeding jejunostomy was placed. Patients were then transferred to the surgical floor after routine postoperative monitoring. The nasogastric tube was removed on postoperative day 1, and clear liquids were started on postoperative day 2. Patients were discharged from the hospital when a soft diet was tolerated and pain control was adequate. Outpatient follow-up occurred 3 to 4 weeks after discharge.

650 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 4 : 2008

www.giejournal.org

Capsule Summary What is already known on this topic d

Transmural endoscopic drainage of pancreatic pseudocysts is accepted as a minimally invasive alternative to surgical drainage, and the addition of EUS guidance lessens the risk of perforation and hemorrhage.

What this study adds to our knowledge d

d

In a retrospective study that compared patients with uncomplicated pancreatic pseudocysts managed by surgical or EUS-guided cystogastrostomy, no significant differences were seen in treatment success rates, procedural complications, or reinterventions. Cost was lower and postprocedure length of hospital stay was shorter for an EUS-guided cystogastrostomy.

Varadarajulu et al

Protocol for EUS-guided cyst-gastrostomy approach

Management of pancreatic pseudocysts

they were managed similarly to patients in whom treatment was successful. Those with treatment failure underwent surgery.

After administration of one dose of IV ciprofloxacin (400 mg), an EUS-guided cyst-gastrostomy was performed at the endoscopy suite, with the patient under conscious sedation with a combination of midazolam, meperidine, and ketamine administered by the endoscopist. All procedures were performed by using a therapeutic linear array echoendoscope (GF-UC 140T; Olympus America Inc, Center Valley, Pa). At EUS, the pseudocyst was accessed by using a 19-gauge needle (EUSN-19-T; Cook Endoscopy, WinstonSalem, NC), and a 0.035-inch guidewire (X-wire; CONMED Industries, Billerica, Mass) was coiled within the pseudocyst under fluoroscopic guidance. The tract was then sequentially dilated by first passing a 4.5F ERCP cannula (Proforma Cannula/Apollo 3AC, CONMED Industries) and then a 10F ERCP inner guiding catheter (OASIS system; Cook) over the guidewire. Further dilation of the tract was performed by using an 8-mm biliary balloon dilator (Eliminator; CONMED), and two 7F or 10F, 4-cm double-pigtail plastic stents (Cook) were deployed. A sample of the cyst aspirate was sent for assessment of carcinoembryonic antigen, amylase, and lipase levels in all patients. An ERCP was routinely attempted before EUS-guided cyst-gastrostomy in all patients, unless the extrinsic compression caused by the pseudocyst precluded duodenoscope passage to the second portion of the duodenum. A pancreatogram was attempted to define communication between the duct and the pseudocyst. In cases when the pancreatic duct was completely disrupted and the proximal duct was accessible with a guidewire, or in patients with a ductal stricture, a transpapillary bridging stent was placed.14 All patients with pancreatic pseudocyst in the setting of smoldering pancreatitis15 underwent placement of direct percutaneous endoscopic jejunostomy (DPEJ) feeding tubes by using a previously described technique16 or underwent placement of a percutaneous gastrojejunostomy feeding tube by interventional radiologists. The rationale was to provide symptomatic relief via strict pancreatic rest in these patients. Patients were discharged home when they were afebrile, had a normal white cell count, did not require IV narcotics for pain control, were able to tolerate a low-fat diet or enteral nutrition via a feeding tube, and were symptomatically better. All patients were evaluated with a contrast-enhanced CTof the abdomen and outpatient clinic visit at 6 weeks after EUSguided drainage. In patients with treatment success, stents placed at EUS-guided drainage were retrieved by an endoscopy, and the enteral feeding tubes were discontinued. Also, the transpapillary pancreatic stents were retrieved if the pancreatogram revealed no leak. Stent removal was undertaken in all patients the same day after an outpatient clinic visit. Those patients with a partial decrease in size of the pseudocyst underwent replacement of transmural stents and were reevaluated after 1 month with a repeated contrast-enhanced CT. If the fluid collection had then resolved,

Treatment success for a surgical cyst-gastrostomy was defined as clinical resolution of symptoms and radiologic resolution of the pseudocyst by US or CT at a 4-week outpatient follow-up. Treatment success for an EUS-guided cyst-gastrostomy was defined as complete resolution or a decrease in the size of the fluid collection to %2 cm on CT, in association with clinical resolution of symptoms at a 6-week follow-up. Because this was a retrospective study, there was no uniformity on the duration of follow-up for patients who underwent a cyst-gastrostomy by either technique. Reinterventions were defined as need for repeated surgery or EUS-guided cyst-gastrostomy because of persistent symptoms in association with a residual pseudocyst (O4 cm) on follow-up imaging within a 3-month period. The length of the postprocedure hospital stay was defined as the time to hospital discharge from the day of surgery or an EUS-guided cyst-gastrostomy. EUS-related bleeding was defined as any hemorrhagic event that required endotherapy, blood-product transfusion, or inpatient observation. Perforation after an EUS was diagnosed when a pneumoperitoneum was evident on imaging studies in association with peritoneal signs. Infection after EUS or surgery was defined as any septic event after a cyst-gastrostomy proven by new-onset fever or positive blood cultures.

www.giejournal.org

Volume 68, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 651

Cost analysis We performed a cost analysis from the institution and/ or health system perspective. Hospital-associated costs were obtained from the UAB Hospital Accounting Department. Professional fees and outpatient visit costs were calculated based on the Medicare reimbursement fee structure. All relevant costs for both surgical and EUSguided cyst-gastrostomy were taken into consideration. They included the cost of inpatient hospital stays; medications, procedural fees, materials, professional fees, procedural facility fees, anesthesia, imaging studies (eg, CT, US), other relevant procedures (eg, ERCP), feedingjejunostomy-tube placement, follow-up endoscopy for stent retrieval, and outpatient clinic visits. To confirm that all costs associated with both treatment modalities were captured in their entirety, the UAB Accounting Department audited individual costs for each patient enrolled in this study. All costs are expressed in 2007 U.S. dollars.

Consent Informed consent was obtained after the risks and benefits of the procedures and the alternative treatment options were explained to the patients. Approval for retrospective chart review was obtained from the UAB Institutional Review Board.

Definitions

Management of pancreatic pseudocysts

Varadarajulu et al

Outcome measures This study compared the rates of treatment success, complications, reinterventions, length of postprocedure hospital stay, and cost associated with each treatment modality.

TABLE 1. Patient characteristics of surgery versus EUS groups

Statistical analysis

Mean age (y)

Statistical analysis was performed by using Stata/MP 10.0 for Windows (StataCorp LP, College Station, Tex). An analysis of variance was performed to compare patients within each matched triad: one undergoing surgery, with two others undergoing EUS-guided drainage. Heteroscedasticity between surgical and EUS-guided drainage procedures was corrected by using Huber-White robust standard errors.

No. men (%)

RESULTS

Surgery EUS (n Z 10) (n Z 20)

Characteristics

42.3

No. white (%) 2

Mean pseudocyst size (mm ) Mean diameter of the pseudocyst (mm)*

43.1

P .884

6 (60)

15 (75)

.403

7 (70)

15 (75)

.772

6179

7588

.238

89

98

.381

Location (%)

.766

Head

20

10

Body

10

15

Tail

40

50

Multiple

30

25

Twenty-eight patients underwent surgical cyst-gastrostomy during the 2-year study period. Sixteen patients were excluded because of the presence of a pancreatic abscess or necrosis, and 2 patients were excluded because of Roux-en-Y cyst-gastrostomy. The remaining 10 patients with uncomplicated pancreatic pseudocysts were matched with 20 patients who underwent an EUS-guided cyst-gastrostomy. Three patients who underwent a surgical cystgastrostomy and 8 patients who underwent an EUS-guided cyst-gastrostomy were outpatients and were admitted after their procedures, whereas the remainder were inpatients transferred from outside hospitals. Patient demographics, the number of major comorbidities, and the characteristics of the pancreatic pseudocyst were not significantly different between both cohorts (Table 1). The clinical outcomes of patients who underwent surgical and EUS-guided cyst-gastrostomy are shown in Table 2. The rates of treatment success were not significantly different between surgery and EUS-guided cyst-gastrostomy (100% vs 95%, P Z .36). One patient who underwent EUS-guided cyst-gastrostomy continued to experience persistent fever, with an elevated white cell count, after the procedure. Because CT imaging revealed no change in the size of the pseudocyst, despite placement of two 10F transmural stents, the patient underwent a surgical cyst-gastrostomy during the same hospital admission, with complete relief of symptoms. Also, there was no significant difference in the rates of reintervention between the surgical and the EUS groups (10% vs 0%, P Z.13). One patient who underwent a surgical cyst-gastrostomy was seen with abdominal pain after 70 days. Because a CT revealed a persistent pseudocyst (O10 cm) at the same location, the patient underwent another operation in which a Roux-en-Y cystjejunostomy was performed that resulted in complete resolution of the pseudocyst and the symptoms. No complications related to surgery or an EUS were encountered in either cohort.

Only 2 of 10 patients in this study had gallstone pancreatitis and required a cholecystectomy at the time of surgery. We excluded the costs associated with a cholecystectomy from our analysis. Although 3 patients underwent placement of jejunostomy feeding tubes at the time of surgery, 10 of 20 patients who underwent an EUS-guided cyst-gastrostomy had a temporary percutaneous feeding tube placed for enteral nutrition. Also, 16 of 20 patients who underwent an EUS had an ERCP for evaluation of the pancreatic-ductal system. Transpapillary pancreatic stenting was successful in 12 of 16 patients; common bile duct (CBD) stones were not found at an ERCP in any patient. When compared with a surgical cyst-gastrostomy, the mean length of postprocedure hospital stays was significantly shorter for EUS-guided cyst-gastrostomy (6.5 vs 2.65 days; P Z .008; range 4 to 20 vs 1 to 11 days). The median length of hospital stay for an EUS-guided cystgastrostomy was 1 day versus 5 days for a surgical cystgastrostomy. One patient in whom EUS-guided drainage was a treatment failure and who required surgical drainage had a postprocedure hospital stay of 11 days; all others were discharged within 1 to 6 days. In the surgical cohort,

652 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 4 : 2008

www.giejournal.org

Etiology of pancreatitis (%)

1.000

Idiopathic

60

60

Gallstones

20

20

Alcohol

20

20

No. major comorbidities (median)y

1.1 (1)

1.47 (1)

Median serum albumin (g/dL) 3.05 (3.3) 2.79 (2.8)

.511 .447

*Based on largest dimension of the pseudocyst axis. yVascular diseases (coronary, cerebral, or peripheral), diabetes mellitus, chronic obstructive pulmonary disease, end-stage kidney disease, systemic infections, cancer, and AIDS.

Varadarajulu et al

Management of pancreatic pseudocysts

TABLE 2. Outcomes of surgical versus EUS-guided cyst-gastrostomy

Clinical outcomes

Surgery (n Z 10)

TABLE 3. Mean itemized costs for surgery and EUS groups

EUS (n Z 20)

P value

Costs

Surgery (US$) (n Z 10)

EUS (US$) (n Z 20)

Hospital stay

5224

2193

Technical success (%)

100

100

1.000

Treatment success (%)

100

95

.364

Procedure*

2522

1538

Reinterventions (%)

10

0

.132

Professional feesy

3143

1824

Complications (%)

0

0

.000

Pharmacy

1826

930

Radiologyz

238

1445

Anesthesiaz

928

41

Laboratory

372

168

Consults

418

609

Other

134

326

Length of stay (d) Mean (range) Median Mean cost (US$)

6.5 (range 4-20)

2.6 (range 1-11)

5

1

14,815

9077

.008

.016

1 patient required a postprocedure hospital stay of 20 days, because he developed respiratory insufficiency. The mean cost for an EUS-guided cyst-gastrostomy was significantly less when compared with surgical cyst-gastrostomy ($9077 vs $14,815, P Z.016), with the average cost savings per patient being $5738. Despite inclusion of the patient in the EUS cohort who failed EUS-guided drainage and who subsequently underwent a surgical cyst-gastrostomy, the EUS-guided approach was less costly than the surgical approach (Table 2). This difference in cost between the 2 modalities was mainly because of higher professional and facility fees, the anesthesia fee, medications (including general anesthesia drugs), and a longer inpatient hospital stay for the surgical cohort (Table 3).

*For facility and supplies. yIncludes procedure and follow-up. zIncludes professional fees.

In this study, the EUS-guided cyst-gastrostomy approach was associated with a shorter length of postprocedure hospital stay and was less costly when compared with surgical cyst-gastrostomy. Also, the clinical outcomes were not significantly different between the 2 cohorts of patients. Increasing experience with endoscopic and EUS-guided pseudocyst drainage has led to its use even in complex clinical settings such as pancreatic necrosis and multiple pseudocysts.17,18 In addition, these procedures are being performed even earlier than the traditional 6-week waiting period required for the pseudocyst wall to mature. However, the cost-effectiveness and the clinical outcomes of this approach have, to our knowledge, never been compared with a surgical cyst-gastrostomy. This study demonstrated that the clinical outcomes for management of uncomplicated pancreatic pseudocysts are not significantly different between the surgical and EUS approaches. Because this was a retrospective study, it was not possible

to compare the rates of technical success between surgery and EUS cohorts. Although surgery is likely to be technically successful in almost all patients who meet the criteria for a cyst-gastrostomy, not all pseudocysts may be amenable for EUS-guided drainage. The presence of intervening vasculature and the distance of the pseudocyst from the GI lumen are potential limitations.19 In our experience of more than 100 cases, in only 5 cases was drainage of peripancreatic-fluid collections not possible under EUS guidance: 3 patients had necrotic debris that was not amenable to drainage, and, in 2 patients, the pseudocysts tracked deep into the pelvic cavity (unpublished observation). The short-term treatment success rate in this series was 100% for surgery versus 95% for an EUS, which is consistent with prior series.9-12 However, this degree of success for the EUS-guided approach is unlikely to be encountered in patients with a pancreatic abscess or necrosis.18 In such patients, surgery not only facilitates drainage of infective material but also permits pancreatic debridement and thereby provides more definitive therapy. Sixty percent of patients in this study had idiopathic pancreatitis, which made this an unusual study cohort. However, we believe that this is unlikely to impact the study findings, because the clinical outcomes of these patients were not different from others who undergo cyst-gastrostomy, as demonstrated in a recent study.12 We did not encounter any procedural complications in either cohort. This may be because the study cohorts were composed of relatively healthy patients without infective pseudocysts or necrosis. An endoscopic or surgical intervention in patients with a pancreatic abscess or necrosis is likely to be associated with increased morbidity and complications.18,20 Also, we did not encounter any pseudocyst recurrence after drainage in the EUS cohort compared with 1 of 10 patients

www.giejournal.org

Volume 68, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 653

DISCUSSION

Management of pancreatic pseudocysts

Varadarajulu et al

who had a recurrent pseudocyst after surgery. One advantage of performing an ERCP before an EUS is that it enables transpapillary pancreatic stenting of any ductal stricture or disruption, thereby minimizing chances of a pseudocyst recurrence.14 Also, in patients with gallstone pancreatitis, an ERCP facilitates CBD stone clearance and thereby enables a laparoscopic cholecystectomy to be performed on an elective basis. Although 80% of patients in the EUS cohort underwent an ERCP and 60% had a transpapillary pancreatic stent placement, none of the patients in the surgical cohort underwent an ERCP. The decision to perform pancreatic stenting was based on our practice pattern and prior favorable experience with patient management when using this strategy.14 Although we did not encounter stent-induced ductal changes in our patient cohort, this was reported as a significant problem in prior reports.21 Only a long-term clinical follow-up will provide meaningful data on the implications of pancreatic stenting in these patients. An EUS, being a minimally invasive procedure, was associated with a shorter length of a postprocedure hospital stay than surgery. The median and mean length of a postprocedure hospital stay was 1 day versus 5 days and 2.65 days versus 6.5 days for an EUS-guided cyst-gastrostomy and surgery, respectively. In patients who are undergoing EUS-guided drainage, in almost all cases, we attempt an ERCP before an EUS and place a DPEJ feeding tube in those patients with smoldering pancreatitis. In those patients in whom a DPEJ was technically not feasible, a gastrojejunostomy feeding tube was placed by interventional radiologists the same day in most patients. Patients were discharged the following day if they could tolerate a low-fat diet or after initiation of enteral nutrition. On the contrary, postsurgical patients required a longer recovery time before they could ambulate, discontinue IV narcotics, and resume oral intake. One added advantage of EUS is that an alternate diagnosis is established in approximately 5% of patients in whom drainage of pseudocyst is attempted.12 Mucinous cyst neoplasm has been reported to mimic a pseudocyst by other imaging studies, eg, CT, and can be readily diagnosed by an EUS or EUS-guided FNA. In our study, an EUS-guided cyst-gastrostomy was found to be less costly than the surgical approach, with an average savings of $5738 per patient. These savings were mostly because of the early discharge of patients from the hospital, a lesser cost for endoscopic procedures than for surgery, a lower cost for medications used for conscious sedation versus anesthesia, and a lower professional fee for gastroenterologists versus surgeons. The EUS approach cost less, despite the fact that all patients required a repeated CT for assessment of response to endotherapy and another endoscopy session for stent retrieval. Because pancreatic-fluid collections are being increasingly managed by means of an endoscopy, a close collaboration between surgeons, endoscopists, and radiologists is required to identify the best strategy for management

of these patients. The radiologists at our institution routinely comment on their CT report if the pseudocysts were amenable for drainage by endoscopic means. Based on clinical experience, we currently perform an endoscopic cyst-gastrostomy in all patients with pancreatic pseudocysts and reserve surgery only for patients with a pancreatic abscess or necrosis unless they are poorrisk surgical candidates. There are several limitations to this study. First, the retrospective design of this study had inherent limitations. It is possible that patients managed surgically were sicker and required more definitive therapy. We believe that this was unlikely to be a major factor, because the mean size of the pseudocyst was comparable between both cohorts, as were the number of major comorbidities (Table 1). Serum albumin levels, known to adversely impact prognosis when low, were not significantly different between the surgical versus the EUS group (mean 3.05 vs 2.79 g/dL). Second, not all patients were admitted as outpatients to make a definitive comparison between both cohorts. A majority of patients in both groups were transferred from outside facilities. However, the postprocedure hospital stay was significantly shorter for the EUS cohort. Third, the number of patients enrolled in this study was small. Hence, it would not be prudent to make inferences on cost-effectiveness based on our data alone. Fourth, the costs used in this study were specific to our institution and may not be applicable to all other centers. Nevertheless, an analysis of direct costs is a commonly used method and is regarded as being more reliable and generalizable. It certainly is superior to analyzing charges or collections, which vary greatly among institutions, different geographic areas, patient demographics, and insurance coverage. Fifth, the study was undertaken at a busy tertiary-referral center with expertise in therapeutic endoscopy. Hence, the experience reported herein may not be applicable to all centers that practice EUS. Sixth, we do not have long-term followup on all patients to assess durable response to therapy. A minimum of 18-month to 24-month follow-up will be required to evaluate pseudocyst recurrence rates in these patients. Seventh, the study cohort is composed only of patients with uncomplicated pancreatic pseudocysts and not those with an abscess or necrosis, which could explain the superior clinical outcomes reported in this study. In conclusion, when available, EUS-guided cyst-gastrostomy should be the first-line treatment approach for patients with uncomplicated pancreatic pseudocysts, because the procedure is cost saving and is associated with a shorter length of postprocedure hospital stay when compared with surgical cyst-gastrostomy. There was no significant difference in clinical outcomes pertaining to treatment success, complications, and reinterventions between the treatment modalities. A close interaction among the endoscopist, surgeon, and radiologist is necessary to maintain a high level of treatment success and safety.

654 GASTROINTESTINAL ENDOSCOPY Volume 68, No. 4 : 2008

www.giejournal.org

Varadarajulu et al

DISCLOSURE The authors report that there are no disclosures relevant to this publication.

REFERENCES 1. Apostolou C, Krige JE, Bornman PC. Pancreatic pseudocysts. S Afr J Surg 2006;44:148-5. 2. Yin WY. The role of surgery in pancreatic pseudocyst. Hepatogastroenterology 2005;52:1266-73. 3. Teh SH, Pham TH, Lee A, et al. Pancreatic pseudocyst in children: the impact of management strategies on outcome. J Pediatr Surg 2006;41: 1889-93. 4. Yusuf TE, Baron TH. Endoscopic transmural drainage of pancreatic pseudocysts: results of a national and an international survey of ASGE members. Gastrointest Endosc 2006;63:223-7. 5. Monkemuller KE, Baron TH, Morgan DE. Transmural drainage of pancreatic fluid collections without electrocautery using the Seldinger technique. Gastrointest Endosc 1998;48:195-200. 6. Howell DA, Elton E, Parsons WG. Endoscopic management of pseudocysts of the pancreas. Gastrointest Endosc Clin N Am 1998;8:143-62. 7. Smits ME, Rauws EA, Tytgat GN, et al. The efficacy of endoscopic treatment of pancreatic pseudocysts. Gastrointest Endosc 1995;42:202-7. 8. Sharma SS, Bhargawa N, Govil A. Endoscopic management of pancreatic pseudocyst: a long-term follow-up. Endoscopy 2002;34:203-7. 9. Giovannini M, Pesenti CH, Rolland AL, et al. Endoscopic ultrasound guided drainage of pancreatic pseudo-cyst and pancreatic abscess using a therapeutic echoendoscope. Endoscopy 2001;33:473-7. 10. Seifert H, Faust D, Schmitt TM, et al. Transmural drainage of cystic peripancreatic lesions with a new large–channel echoendoscope. Endoscopy 2001;33:1022-6. 11. Antillon MR, Shah RJ, Stiegmann G, et al. Single-step EUS-guided transmural drainage of simple and complicated pancreatic pseudocysts. Gastrointest Endosc 2006;63:797-803. 12. Varadarajulu S, Wilcox CM, Tamhane A, et al. Role of EUS in drainage of peripancreatic fluid collections not amenable for endoscopic transmural drainage. Gastrointest Endosc 2007;66:1107-19.

www.giejournal.org

Management of pancreatic pseudocysts 13. Bradley EL 3rd. A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis, Atlanta, GA, September 11 through 13. Arch Surg 1993;128: 586-90. 14. Varadarajulu S, Noone TC, Tutuian R, et al. Predictors of outcome in pancreatic duct disruption managed by endoscopic transpapillary stent placement. Gastrointest Endosc 2005;61:568-75. 15. Varadarajulu S, Noone TC, Hawes RH, et al. Pancreatic duct stent insertion for functional smoldering pancreatitis. Gastrointest Endosc 2003;58:438-41. 16. Varadarajulu S, Delegge MH. Use of a 19-gauge injection needle as a guide for direct percutaneous endoscopic jejunostomy tube placement. Gastrointest Endosc 2003;57:942-5. 17. Bhasin DK, Rana SS, Udawat HP, et al. Management of multiple and large pancreatic pseudocysts by endoscopic transpapillary nasopancreatic drainage alone. Am J Gastroenterol 2006;101:1780-6. 18. Baron TH. Treatment of pancreatic pseudocysts, pancreatic necrosis, and pancreatic duct leaks. Gastrointest Endosc Clin N Am 2007;17: 559-79. 19. Fockens P, Johnson TG, van Dullemen HM, et al. Endosonographic imaging of pancreatic pseudocysts before endoscopic transmural drainage. Gastrointest Endosc 1997;46:412-6. 20. Mofidi R, Lee AC, Madhavan KK, et al. Prognostic factors in patients undergoing surgery for severe necrotizing pancreatitis. World J Surg 2007;31:2002-7. 21. Rashdan A, Fogel EL, McHenry L Jr, et al. Improved stent characteristics for prophylaxis of post-ERCP pancreatitis. Clin Gastroenterol Hepatol 2004;2:322-9.

Received November 12, 2007. Accepted February 11, 2008. Current affiliations: Division of Gastroenterology-Hepatology (S.V., T.L.L., C.M.W.), Department of Surgery (E.R.D., J.D.C.), University of Alabama at Birmingham School of Medicine, Department of Health Care Organization and Policy, School of Public Health (M.L.K.), University of Alabama at Birmingham School of Public Health, Birmingham, Alabama, USA. Reprint requests: Shyam Varadarajulu, MD, Division of GastroenterologyHepatology, University of Alabama at Birmingham, 410 Lyons Harrison Research Bldg, 1530 3rd Ave South, Birmingham, AL 35294-0007.

Volume 68, No. 4 : 2008 GASTROINTESTINAL ENDOSCOPY 655