Original Paper Received: March 27, 2006 Accepted: July 6, 2006 Published online: August 28, 2006
Dig Surg 2006;23:250–254 DOI: 10.1159/000095398
Endoscopic Transpapillary Stenting of Pancreatic Duct Disruption P.M. Brennan a T. Stefaniak b K.R. Palmer a R.W. Parks a a Department of Surgical Sciences, Royal Infirmary, Edinburgh, UK, and b Department of Gastroenterological Surgery, City Hospital, Gdynia, Poland
Key Words Pancreas Stenting, pancreatic Fistula, pancreatic Pseudocyst, pancreatic Duct stenosis, pancreatic Endoscopic retrograde cholangiopancreatography
Abstract Background/Aims: Endoscopic transpapillary stenting of the pancreatic duct is increasingly being used in the management of pancreatic duct disruption. In contrast to its more established role in pancreatic duct obstruction, little is reported on the spectrum of indications and outcome in management of pancreatic duct disruption. Methods: The indication for and outcome of transpapillary pancreatic duct stenting was analysed retrospectively in a UK supra-regional specialist pancreatobiliary centre, between January 1998 and August 2004. Results: Data were obtained on 30 patients (19 male, 11 female, median age 53 years). The main indications for pancreatic duct stenting were: pancreatic pseudocyst, pancreatic ascites, pancreatic duct leak following necrosectomy, and pancreaticopleural fistula. The median duration of stenting was 6 weeks for fistulae and 10 weeks for pseudocysts. Twenty-one patients (70%) had complete resolution. After a median follow-up of 45 months, no recurrence was noted in successfully treated patents. Conclusion: Endoscopic transpapillary pancreatic duct stenting is an increasingly valuable treatment option in the management of pancreatic fistulae and pseudocysts. Copyright © 2006 S. Karger AG, Basel
© 2006 S. Karger AG, Basel 0253–4886/06/0234–0250$23.50/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/dsu
Introduction
With the increasing availability and application of non-invasive imaging, such as computed tomography and magnetic resonance cholangiopancreatography, endoscopic retrograde cholangiopancreatography (ERCP) is used less commonly for the diagnosis of pancreaticobiliary conditions. However, transpapillary endoscopic intervention has an increasingly important role, involving the use of sphincterotomy, lithotripsy and stenting [1]. The indications for transpapillary pancreatic duct stenting are varied [2], but include the management of pancreatic duct disruption and pancreatic duct stenosis. The rationale for stenting patients with pancreatic duct stenoses is to provide pain relief, when the cause of pain is thought to be due to increased pancreatic intraductal pressure [3]. Pancreatic duct stenting has also been used as a prophylactic technique to protect against post-ERCP pancreatitis in patients undergoing endoscopic sphincterotomy for pancreatic sphincter hypertension [4]. Pancreatic duct disruption, requiring transpapillary stenting, most commonly is associated with acute or chronic pancreatitis [4–8]. Indications for pancreatic duct stenting include treatment of pancreatic pseudocysts, pancreatic ascites, pancreaticobiliary fistulae, pancreaticopleural fistulae, pancreaticocutaneous fistulae, pancreaticoenteric fistulae, as well as leaks after pancreatic necrosectomy [5, 7, 9–11]. Pancreatic duct disruption
Mr. R.W. Parks Surgical Unit, Royal Infirmary of Edinburgh 51 Little France Crescent Edinburgh, EH16 4SA (UK) Tel. +44 131 242 3616, Fax +44 131 242 3617, E-Mail
[email protected]
Fig. 1. Resolution of pancreatic duct leak via transpapillary pancreatic duct stenting. a ERCP demonstrating leak from pancreatic duct. b ERCP following placement of pancreatic stent.
following major blunt pancreatic trauma may also require endoscopic transpapillary treatment [12]. Conventional treatment of pancreatic duct disruption includes establishing adequate drainage and fistula control, and where appropriate, aggressively treating sepsis, providing nutritional support with total parenteral nutrition, and reducing pancreatic exocrine secretions by administration of octreotide [2]. This strategy results in fistula closure in approximately 75% of patients [2, 5]. Pancreatic duct stenting may complement these existing therapies. The aim of this study was to retrospectively analyse the experience of ERCP and transpapillary pancreatic duct stenting in a UK supra-regional specialist centre treating patients with pancreatic duct disruption.
Patients and Methods Patients who had undergone transpapillary pancreatic duct stenting for pancreatic duct disruption at the Royal Infirmary of Edinburgh, UK, between January 1998 and August 2004 were identified from endoscopy records. Stents placed via other approaches, such as transmurally, were not included in the study. Case notes were obtained for all patients identified. These were examined to determine the clinical course before and after endoscopy, including the indication for stent insertion, pancreatogram appearance before and after treatment, length and calibre of stent used, and duration of stent insertion. Stenting was regarded as successful when there was clinical resolution of the fistula or pseudocyst.
Endoscopic Transpapillary Stenting of Pancreatic Duct Disruption
Pancreatic duct stenting was performed during ERCP as described by Tarnasky [4]. Pancreatic sphincterotomy was undertaken with an Erlangen sphincterotome, or using a needle knife over a stent, and a catheter was introduced into the pancreatic duct. After injection of contrast, the nature of the abnormality was observed by fluoroscopy and a therapeutic strategy was established. The endoprosthesis was then inserted into the duct over a 0.24-mm hydrophilic guide wire (fig. 1).
Results
Thirty-three patients underwent 36 transpapillary pancreatic duct stent insertions. Notes were missing for 2 patients. One further patient was not considered because the stent was inserted prophylactically rather than as a therapeutic procedure, and removed only 1 day after insertion. Data were therefore available for 30 patients with pancreatic duct disruption. There was no procedure-related morbidity or mortality. Three patients underwent a repeat procedure. In 1 patient this was undertaken because of stent occlusion. A second patient had no initial resolution of pancreatic ascites, where a proximal stricture had limited the initial stent placement to the pancreatic duct within the head of the gland. A repeat procedure advanced the stent into the more distal pancreatic duct. Data were missing in the 3rd patient. There were 19 men and 11 women, with a mean age of 53 (range 31–82) years. The median follow-up period was Dig Surg 2006;23:250–254
251
Table 1. Indications for pancreatic stenting
Indication Pancreatic duct leak (n = 30) Related to acute or chronic pancreatitis (n = 28) Not related to pancreatitis (n = 2)
Underlying aetiology
Number of cases
Pseudocyst Ascites post-pancreatic necrosectomy Pancreatic ascites Pancreaticopleural fistula Pancreaticojejunostomy for pseudocyst Hepaticojejunostomy for cholangiocarcinoma
11 7 6 4 1 1
45 (range 5–78) months. The median time to stent insertion from the date of admission was 3 weeks. Prior to stent placement, patients received standard non-interventional treatment for pancreatic duct fistulae. The aetiology of pancreatic disease requiring pancreatic duct stenting is shown in table 1. Twenty-eight patients developed pancreatic duct disruption following an episode of acute pancreatitis, or in the setting of chronic pancreatitis. This manifested as either pancreatic pseudocyst (n = 11), persistent amylase-rich drainage following pancreatic necrosectomy (n = 7), pancreatic ascites (n = 6), or pancreaticopleural fistula (n = 4). The aetiology of the pancreatitis was alcohol excess (n = 14), gallstones (n = 7), pancreas divisum (n = 1), azothioprine treatment (n = 1), or idiopathic (n = 5). Two additional patients developed a pancreatic duct leak following surgical intervention. Of the 11 patients stented for pseudocyst drainage, 8 remained symptom-free during follow-up, of whom 4 patients had documented resolution of the pancreatic duct communication at repeat pancreatography. One of these patients, with a pancreatic pseudocyst on the background of acute on chronic alcoholic pancreatitis, first underwent distal pancreatectomy, but required ERCP and pancreatic duct stenting post-operatively when the pseudocyst recurred, which was successful. A further patient had documented resolution of their original pseudocyst after pancreatic duct stenting, but had persistent epigastric pain. Following diagnosis of a recurrent pseudocyst he proceeded to cyst-jejunostomy, but without resolution of pain. One patient, with acute pancreatitis secondary to azothioprine therapy, developed a pseudocyst after pancreatic necrosectomy that did not resolve after 6 months of endoscopic transpapillary stenting. A pig-tail drain was then inserted transmurally from the stomach with no immediate benefit, however, further follow-up was 252
Dig Surg 2006;23:250–254
unavailable. One patient had missing data and it was not possible to determine the outcome following treatment. Seven patients had persistent amylase-rich drainage following pancreatic necrosectomy for acute pancreatitis. One of these patients had a documented pancreatic duct leak prior to surgery. The fistulae persisted after pancreatic duct stent removal in 2 of the 7 patients. Six patients, including one with a persistent leak, had resolution of their symptoms. The remaining patient had stenting of a pancreatic duct leak after failure of resolution with conservative therapy, 21 months after necrosectomy, following an episode of alcohol-related acute pancreatitis. A persistent pancreatic duct leak was demonstrated after 12 weeks, but the stent was removed. A second stent was inserted 7 months later, however the leak was still demonstrated after a further 2 months when the second stent was removed. The patient succumbed to end-stage liver disease 20 months after the last intervention. Six further patients had a diagnosis of pancreatic ascites, 5 on the background of alcohol excess, and 1 related to gallstone pancreatitis. In 4 patients pancreatic duct stent insertion facilitated resolution of the duct leak. At repeat pancreatography following transpapillary pancreatic duct stent insertion, 1 patient was found to have a persistent leak. A transmural stent was inserted into the residual retroperitoneal collection, resulting in resolution of the leak. A final patient died from end-stage alcoholic liver disease 5 days after insertion of a second transpancreatic duct stent. In 4 patients, pancreatic duct stents were inserted for pancreaticopleural fistulae. In 2 patients the leak resolved. In a 3rd patient the pleural collection persisted, although to a lesser extent, and this was thought likely to be an effusion due to an underlying malignancy. The 4th patient proceeded to distal pancreatectomy, splenectomy Brennan /Stefaniak /Palmer /Parks
Table 2. The duration of stenting,
period of follow-up, and success rate, after pancreatic duct stenting
Indication
n
Median duration of stenting, weeks (95% CI)
Median period of follow-up, months
Success rate, %
Pancreatic duct rupture Pseudocyst
19 11
6 (4.8–7.2) 10 (8.6–11.4)
51 43
88 63
and successful fistula management 7 days after stent insertion, because of a deteriorating clinical condition. Two patients had successful pancreatic duct stenting for leakage after surgical intervention, in 1 case following cyst-jejunostomy and in another following hepaticojejunostomy for cholangiocarcinoma. In the latter, the pancreatic parenchyma was traumatised during tumour resection and the patient subsequently developed an amylase-rich discharge from the abdominal drain. A transpapillary pancreatic duct stent was placed after only partial success of non-operative management of the leak. The median period of stenting for duct disruption was 6 weeks, and for pseudocysts was 10 weeks (table 2). Data on stent length and calibre were available for 16 of the pancreatic stent insertion episodes. Thirteen of these (81%) were in the range of 5–7 cm in length and 7–9 Fr in calibre. The success, as measured by clinical improvement with or without radiological resolution of the leak, was 87.5 % for patients with pancreatic duct rupture and 62.5% for patients with pseudocysts, followed up over an average of 51 and 43 months, respectively.
Discussion
With the advent of improved non-invasive diagnostic imaging modalities for pancreatic disease, such as magnetic resonance pancreatography, the role of ERCP is increasingly directed to the therapeutic management of identified lesions. In this analysis of pancreatic stenting in a specialist centre, the different indications for transpapillary pancreatic duct stenting have been identified and the outcome analysed. Pancreatic duct stenting was mainly used in patients where duct disruption, most commonly following alcohol-related pancreatitis, had led to the development of a pancreatic pseudocyst, pancreatic ascites, or a pancreaticopleural fistula. Stenting was particularly useful where duct disruption was identified following pancreatic necrosectomy, or as a consequence of other pancreatic surgery. Postoperative pancreatic fistulae resistant to conEndoscopic Transpapillary Stenting of Pancreatic Duct Disruption
servative treatment remains a particularly challenging problem, and is reported to occur in 23–29% of patients after pancreatic necrosectomy [2]. Boerma et al. [13] studied 15 patients who had developed external pancreatic fistulae following pancreatic necrosectomy. They noted that drainage stopped in all patients after transpapillary pancreatic duct stent insertion, although 4 patients subsequently required further definitive surgery (3 for pseudocyst formation). Pancreatic stenting for the treatment of duct disruption may be employed in one of two ways. Firstly, the endoprosthesis can be introduced in a transpapillary fashion and placed across the pancreatic duct disruption to occlude the defect, which may lead to closure of the fistula. Alternatively, a short stent may be placed across the sphincter of Oddi, reducing the pressure gradient between the pancreatic duct and duodenum, facilitating healing of the fistula [5]. From this study, no relationship between the position of the pancreatic stent and the eventual clinical success of the procedure was identified. More often, the stent was placed as distally or as close to any leak as possible. However, where a proximal stricture restricted passage of the stent, even after dilatation, the stent was placed more proximally. Rosch et al. [14] found no correlation between initial technical success of transpapillary pancreatic duct stenting and outcome, measured as pain, in a retrospective comparative analysis of patients with pancreatic duct stenosis in chronic pancreatitis. There were insufficient data in this analysis to establish any definitive relationship between the size of the pancreatic stents used and the outcome of the procedure. The timing of surgical or endoscopic intervention for pancreatic duct disruption remains controversial [13]. Published series report initial conservative treatment lasting between 2 and 4 weeks [8, 9, 15], compared to an average of 3 weeks in this analysis. Similarly, the duration of pancreatic duct stenting is variable. The limited number of large studies in this area makes it difficult to determine the optimal period of stenting. Boerma et al. [13] removed the stents after a median of 10 (range 2–64) days, obtaining fistula healing in 13 of 15 patients. KozaDig Surg 2006;23:250–254
253
rek et al. [16] reported 9 patients with pancreaticocutaneous fistulas, 8 of which closed successfully after pancreatic duct stenting, including 5 within 48 h. Stents were removed 10–14 days after fistula closure and no patient had recurrence at a median follow-up of 3 years [16]. Saeed et al. [9] removed stents after 6 weeks, while still others have suggested 4–12 weeks before removal [17]. The median duration of stent placement for duct disruption in this study was 6 weeks. The mainstay of therapy for symptomatic pancreatic pseudocysts has been surgical drainage. However, this is associated with 5% mortality, 25% morbidity, and 10% recurrence rates. Endoscopic techniques include wide transmural incision, transmural puncture and stenting, and transpapillary stenting. Overall, the endoscopic experience in expert hands is associated with 94% initial technical success, 90% cyst resolution, and 16% recurrence rates. Additional non-endoscopic interventions, mostly surgical, are necessary in 17% of patients [18]. The results for management of pancreatic pseudocyst reported in this study are similar to the report of Catalano et al. [19] who treated 21 patients using 33 transpapillary en-
doprostheses. Thirteen patients had the stent placed directly into the pseudocyst, and in 8 patients the stent was placed beyond the stricture but not into the pseudocyst. Initial resolution of the pseudocyst was seen in 17 patients (80%), with 16 patients (76%) free of pseudocyst recurrence at a mean follow-up of 37 months. Factors predictive of success included the absence of duct strictures, size of pseudocyst of 66 cm, location in the body of the pancreas, and duration of pseudocyst of !6 months [19]. Establishing evidence-based criteria for the use of transpapillary pancreatic duct stenting is limited by the lack of randomised controlled trials in this area. As a result, for the small number of patients who undergo pancreatic stenting, practice until now has been guided by local experience and the limited number of case studies in the literature. Endoscopic pancreatic duct stenting is a safe and effective procedure for treatment of various pancreatic duct lesions. Transpapillary pancreatic duct stenting should be considered as first-line treatment in patients with pancreatic duct disruption [6, 20].
References 1 Dobosz M, Dymecki D, Babicki A, Hac S, Gruca Z, Wajda Z: Management of postoperative pancreatic fistula resistant to octreotide therapy. Wiad Lek 2001;54:380–385. 2 Ridgeway MG, Stabile BE: Surgical management and treatment of pancreatic fistulas. Surg Clin North Am 1996;76:1159–1173. 3 Bornman PC, Marks IN, Girdwood AW, Berberat PO, Gulbinas A, Buchler MW: Pathogenesis of pain in chronic pancreatitis: ongoing enigma. World J Surg 2003;27:1175– 1182. 4 Tarnasky PR: Mechanical prevention of post-ERCP pancreatitis by pancreatic stents: results, techniques, and indications. JOP 2003;4:58–67. 5 Kozarek RA, Traverso LW: Pancreatic fistulas: etiology, consequences, and treatment. Gastroenterologist 1996;4:238–244. 6 Martin FM, Rossi RL, Munson JL, ReMine SG, Braasch JW: Management of pancreatic fistulas. Arch Surg 1989;124:571–573. 7 Knecht GL, Kozarek RA: Double-channel fistulotome for endoscopic drainage of pancreatic pseudocyst. Gastrointest Endosc 1991;37:356–357. 8 Howard TJ, Stonerock CE, Sarkar J, Lehman GA, Sherman S, Wiebke EA, Madura JA, Broadie TA: Contemporary treatment strategies for external pancreatic fistulas. Surgery 1998;124:627–632.
254
9 Saeed Z, Ramirez F, Hepps K: Endoscopic stent placement for internal and external pancreatic fistulas. Gastroenterology 1993; 105:1213–1217. 10 Wolfsen HC, Kozarek RA, Ball TJ, et al: Pancreaticoenteric fistula: etiology and role of pancreatic stents. J Clin Gastroenterol 1992; 14:117–121. 11 Kozarek RA, Jiranek G, Traverse LW: Endoscopic treatment of pancreatic ascites. Am J Surg 1994;168:223–228. 12 Lin BC, Chen RJ, Fang JF, Hsu YP, Kao YC, Kao JL: Management of blunt major pancreatic injury. J Trauma 2004;56:774–778. 13 Boerma D, Rauws EA, van Gulik TM, Huibregtse K, Obertop H, Gouma DJ: Endoscopic stent placement for pancreaticocutaneous fistula after surgical drainage of the pancreas. Br J Surg 2000;87:1506–1509. 14 Rosch T, Daniel S, Scholz M, Huibregste K, Smits M, Schneider T, Ell C, Haber G, Riemann JF, Jakobs R, Hintze R, Adler A, Neuhaus H, Zavoral M, Zavada F, Schusdziarra V, Soehendra N; European Society of Gastrointestinal Endoscopy Research Group: Endoscopic treatment of chronic pancreatitis: a multicenter study of 1000 patients with long-term follow-up. Endoscopy 2002; 34: 765–771.
Dig Surg 2006;23:250–254
15 Rosso E, Alexakis N, Ghaneh P, Lombard M, Smart HL, Evans J, Neoptolemos JP: Pancreatic pseudocyst in chronic pancreatitis: endoscopic and surgical treatment. Dig Surg 2003;20:397–406. 16 Kozarek RA, Ball TJ, Patterson DJ, Raltz SL, Traverso LW, Ryan JA, Thirlby RC: Transpapillary stenting for pancreaticocutaneous fistulas. J Gastrointest Surg 1997;1:357–361. 17 Kochhar R, Goenka MK, Nagi B, Singh K: Pancreatic ascites and pleural effusion treated by endoscopic pancreatic stent placement. Indian J Gastroenterol 1995; 14:106–107. 18 Lo SK, Rowe A: Endoscopic management of pancreatic pseudocysts. Gastroenterologist 1997;5:10–25. 19 Catalano MF, Geenen JE, Schmalz MJ, Johnson GK, Dean RS, Hogan WJ: Treatment of pancreatic pseudocysts with ductal communication by transpapillary pancreatic duct endoprosthesis. Gastrointest Endosc 1995; 42:214–218. 20 Lipsett PA, Cameron JL: Internal pancreatic fistula. Am J Surg 1992;163:216–220.
Brennan /Stefaniak /Palmer /Parks
