Pancreatitis is the most common and potentially serious ENDOSCOPY CORNER

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2009;7: ENDOSCOPY CORNER Short 5Fr vs Long 3Fr Pancreatic Stents in Patients at Risk for Post-Endoscopic Retrograde Cholangiopancreatography Pancreatitis PRABHLEEN CHAHAL,* PAUL R. TARNASKY,* BRET T. PETERSEN, MARK D. TOPAZIAN, MICHAEL J. LEVY, CHRISTOPHER J. GOSTOUT, and TODD H. BARON *Digestive Health Associates of Texas, PA, Dallas, Texas; and Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, Minnesota BACKGROUND & AIMS: Prophylactic placement of pancreatic duct (PD) stents reduces the risk of post-endoscopic retrograde cholangiopancreatography pancreatitis (PEP) in high-risk patients. Some endoscopists prefer longer length, unflanged 3Fr PD stents because they are supposedly more effective and have a higher rate of spontaneous dislodgement; we compared outcomes of patients with these 2 types of stents. METHODS: Patients at high risk for PEP were randomly assigned to groups given either a straight, 5Fr, 3 cm long, unflanged PD stent (n 116) or a 3Fr, 8 cm or longer, unflanged PD stent (n 133). Abdominal radiographs were obtained at 24 hours, 7 days, and 14 days following stent placement to assess spontaneous stent dislodgement. PEP was defined according to consensus criteria. RESULTS: After 14 days, the spontaneous stent dislodgement rates were 98% for 5Fr stents and 88% for 3Fr stents (P.0001). PEP occurred in 12% of patients. The incidence of PEP was higher in the 3Fr group (14%) than the 5Fr group (9%), although this difference was not statistically significant (P.3). Placement failure did not occur in any patients in the 5Fr stent group, but did occur in 11 of the 133 patients in the 3Fr stent group (P.0003). CONCLU- SIONS: Among patients at high-risk for PEP, the spontaneous dislodgement rate of unflanged, short-length, 5Fr PD stents is significantly higher than for unflanged, longlength, 3Fr stents. This decreases the need for endoscopic removal. A higher rate of PD stent placement failure and PEP was observed in patients with 3Fr stents. View this article s video abstract at Pancreatitis is the most common and potentially serious complication of endoscopic retrograde cholangiopancreatography (ERCP) occurring after 4% to 20% of procedures, depending on patient and procedure-related factors. 1 The underlying pathogenesis of post-ercp pancreatitis (PEP) is thought to be multifactorial, 2 with frequent implication of pancreatic outflow obstruction. 3 6 Multiple clinical trials and a meta-analysis have demonstrated that placement of pancreatic stents in high risk patients effectively reduces the incidence of PEP. 3,7 14 This has led to widespread adoption of prophylactic pancreatic stent placement in high risk patients. Moreover, a recent cost effectiveness analysis supported prophylactic insertion of pancreatic stents in high risk patients to help prevent PEP. 15 However, their use is not risk free. Although designed to pass spontaneously, they may remain in place and require endoscopic removal. Pancreatic ductal and parenchymal changes have been reported to occur with prolonged stenting, and stent placement may be complicated by proximal (inward) migration and difficulty with removal. Furthermore, controversy exists as to best type of pancreatic stent. The authors of a recent large, retrospective study suggested that small diameter stents (3 Fr) are superior to 5 Fr stents as they observed a significantly higher rate of spontaneous dislodgement of 3Fr stents compared with 5Fr pancreatic stents. This study also suggested the superiority of longer length 3Fr stents compared with shorter length 5Fr stents in preventing PEP although the difference was not statistically significant. 16 There are no randomized prospective trials to ascertain the superiority of small diameter pancreatic stents. This prospective randomized multicenter trial evaluated the spontaneous distal migration rates and efficacy for unflanged long length 3Fr versus short length 5Fr pancreatic stents in patients at high risk for PEP. Methods Adult ( 18 years) patients undergoing ERCP at Mayo Clinic, Rochester, Minnesota, and Methodist Hospital, Dallas, Texas, were recruited. Pancreatic duct (PD) stents were placed in patients deemed to be at high risk for developing PEP prior to or during ERCP. Following informed consent, enrolled subjects were randomized to receive either a straight, 5Fr polyethylene stent (2 cm or 3 cm length with a single external flap) or an unflanged 3Fr polyethylene stent (8 cm or 10 cm length with single external pigtail) using consecutively numbered sealed opaque envelopes prepared by a study coordinator. Subjects were stratified based on indications for prophylactic pancreatic stent placement to avoid chance imbalance in the randomized groups. The 5 stratification variables were: (1) known or suspected sphincter of Oddi dysfunction (SOD); (2) endoscopic ampullectomy; (3) precut sphincterotomy; (4) pancreatic (major papilla) sphincterotomy; and (5) other (including difficult cannulation and Abbreviations used in this paper: ERCP, endoscopic retrograde cholangiopancreatography; PD, pancreatic duct; PEP, postendoscopic retrograde cholangiopancreatography pancreatitis; SOD, sphincter of Oddi dysfunction by the AGA Institute /09/$36.00 doi: /j.cgh

2 August Fr vs 3Fr PANCREATIC STENTS 835 prior history of PEP). When a subject fit more than 1 stratification group, he or she was assigned to the first appropriate group. Random allocations for each stratification group were placed in 26 consecutively numbered opaque sealed envelopes. The endoscopists were blinded to the randomization allocation, which was only revealed during ERCP, after the endoscopist had decided that PD stenting was indicated to prevent PEP, and before attempts at stent placement. The endoscopist determined the indication for prophylactic PD stent placement, and the sealed envelope was drawn from the consecutively numbered appropriate stratified block. Patients with known main pancreatic duct stricture toward the head of pancreas (n 18), patients with features of severe chronic pancreatitis as defined by Cambridge Classification 17 (n 38), and persons unwilling or unable to give informed consent for this study (n 57) were excluded. Urine pregnancy test was performed on all consented female patients of childbearing age prior to inclusion. A successful PD stent placement was established when the stent was appropriately positioned within the PD with its distal end in the duodenal lumen. A successful placement of 3Fr stent involved positioning of the stent beyond the pancreatic genu. Follow-up plain abdominal radiographs were obtained to assess for spontaneous stent dislodgement the day following ERCP and at 7 days (if not passed within 24 hours), and 14 days (if not passed by 7 days). If the stent had not passed spontaneously by 2 weeks, endoscopic removal was performed. Analysis involving spontaneous dislodgement was repeated for all time points (1 day, 7 days, and 14 days) and both study populations (intentto-treat and treatment-received). The rates of post-ercp pancreatitis (per consensus criteria) 1 ascertained by intention-totreat and treatment-received analysis were compared between the 5Fr and 3Fr PD stent groups. Univariate analyses were performed to determine the effect of procedure-related and patient-related variables on the incidence of PEP. Data Collection Patient demographics (age, gender, indication for ERCP, stent placement indication) were collected at the time of the procedure. Biliary cannulation was graded as difficult if there were 5 or more inadvertent pancreatic duct injections. If a pancreatogram was obtained, it was graded for evidence of chronic pancreatitis by the Cambridge Classification. 17 Specific Measured Outcomes The primary outcome was rate of spontaneous stent dislodgement. The secondary outcome was the incidence of PEP. The diagnosis of PEP was established if a patient had all of the following persisting 24 hours after ERCP: (1) typical pancreatitis pain (epigastric radiating to the back) associated with tenderness to palpation not present before ERCP; and (2) elevation of serum amylase and/or lipase levels 3 times upper limit of normal. Severity of PEP was graded as mild, moderate, or severe according to consensus criteria. 1 Other data collected were: number and type of accessories used for PD stent placement, abdominal computerized tomography findings (as clinically indicated), length of hospital stay including readmissions (if applicable), need for subsequent procedure to manage post-ercp complications (endoscopic therapy, surgical, or percutaneous therapy), success rate of stent placement, reason for stent placement failure, and need for crossover. Statistical Analysis Standard univariate statistical methods, such as 2 or the Fisher exact test for categorical data, and Wilcoxon rank sum test or the Student t test (JMP statistical software; SAS Institute, Cary, NC) for continuous variables, were used. These methods were applied to determine the effect of early stent migration ( 24 hours), retained PD stent at the end of 2 weeks, number of guidewires utilized for PD stent placement, and subject stratification group on the incidence of PEP. A prior study (Rashdan et al, 2004) 16 provided the only available data comparing 3Fr and 5Fr stents with respect to spontaneous dislodgement. Using these results as alternatives to the null hypothesis of no difference, we were able to calculate the sample size (N 103 patients per group) for our primary outcome of spontaneous dislodgement of pancreatic stent with 90% power. Throughout the analysis, we used a 2-sided type I error rate of.05, which gave us 90% power to calculate the primary aim of stent dislodgement. A P value of.05 was considered statistically significant, and supporting risk ratios were calculated. Results Two-hundred forty-nine patients (182 women, 67 men; mean age 50 years) were randomized to 5 Fr (116) and 3Fr (133) PD stent groups. Comparisons for baseline characteristics for both study groups are illustrated in Table 1. Except for age, the 2 groups were similar in baseline characteristics. Table 1. Patient Demographics and Procedure-Related Data 3Fr 5Fr P value Gender Female Male Age (y), mean SD Indication(s) for stent placement a PD (major) sphincterotomy PD (minor) sphincterotomy Ampullectomy SOD Bile duct precut Balloon sphincteroplasty Difficult cannulation Prior PEP Indication(s) for ERCP a Unexplained pancreatitis Biliary pancreatitis Abnormal liver test Unexplained pain Specific planned endotherapy Other b a Some patients had more than one indication. b Dilated bile duct or pancreatic duct on radiological imaging, ampullary polyp, post operative bile leak, cholangiocarcinoma, cholangitis, query chronic pancreatitis.

3 836 CHAHAL ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 8 Stent Placement The stent placement failure rate in patients randomized to 5Fr stents was 0% compared to patients randomized to 3Fr stents, which was 9% (P.0003). Among the 133 patients randomized to receive 3Fr stents, 11 required crossover to 5Fr stents during the procedure due to an angulated, tortuous pancreatic duct preventing deep passage of a inch guidewire into the pancreatic duct beyond the genu and thus preventing successful 3Fr PD stent placement (Figure 1). One patient randomized to receive a 5Fr stent was crossed over to receive a 3Fr stent after intraprocedural iatrogenic guidewire-induced perforation of main pancreatic duct occurred in the region of tail, and it was elected to place a longer 3Fr (8 cm) stent to promote leak closure. Spontaneous Dislodgement The spontaneous dislodgement rate for the 5Fr group at 2 weeks was 98% compared with 88% for the 3Fr stent group (P.0007) with treatment-received analysis, and was 97% and 89% (P.001) with intention-to-treat analysis. The relative risk for retained PD stent at 2 weeks was 4.07 for 3Fr stents compared with 5Fr stents. Among the groups, 5Fr stents migrated Figure 2. Histogram displaying spontaneous dislodgement rates of 3Fr and 5Fr stents at 3 time points (1 day, 7 days, and 14 days). Spontaneous stent dislodgement rate of 3Fr and 5Fr stents, P.0001 for all time points among 3Fr vs 5Fr stent groups. within the first 24 hours at a significantly higher rate than 3Fr stents (50% vs 13%; P.0001). At 1 week, 91% of 5Fr stents were dislodged, compared with only 69% of 3Fr stents (P.0001) (Figure 2). Figure 1. Pancreatogram demonstrating an acutely angulated main pancreatic duct prohibiting deep guidewire access. Post-ERCP Pancreatitis (PEP) The overall incidence of PEP was 11.6% (29 cases). Of these, 25 were mild and 4 were moderate in severity; there were no severe PEP cases. The median duration of hospital stay was 2 days (interquartile range 1 4 days). One case of moderate pancreatitis was due to guidewire-induced perforation resulting in peripancreatic fluid collection. The patient was randomized to receive a 5Fr stent but was crossed over to a 3Fr stent following the guidewire-induced perforation. The stent failed to dislodge and required endoscopic removal at 2 weeks. The peripancreatic fluid collection resolved spontaneously after 8 weeks. The incidence of PEP with intention-to-treat analysis (Figure 3) in the patients randomized to receive a 5Fr stent was 9% (11 patients) and in those randomized to receive a 3Fr stent was 14% (18 patients) with a nonsignificant P value of.3. The incidence of PEP with treatment-received analysis was 10% (12 patients) and 16% (19 patients) in the 5Fr group and the 3Fr group, respectively (P.06). Among the 11 patients randomized to the 3Fr stent group who required crossover to a 5Fr stent, only 1 (9%) developed mild PEP. However, among the 5 high risk patients where pancreatic duct could not be accessed for stent insertion, 2 (40%) developed mild PEP. Univariate analyses were applied to determine the effect of early stent migration (within 24 hours), retained PD stent at end of 2 weeks, number of guidewires utilized for PD stent placement, stratification group, and need for crossover from 3Fr stent to 5Fr stent on the incidence of PEP (Tables 2 and 3). On univariate analysis, a significant association was found between higher incidence of PEP and a higher number ( 1) of guidewires utilized during PD stent placement, retained PD stent at 2 weeks, and minor papilla intervention. However, early (within 24 hours) stent migration, SOD, precut sphincterotomy, major

4 August Fr vs 3Fr PANCREATIC STENTS 837 Table 3. Univariate Analysis to Determine Effect of Listed Variables on Incidence of PEP Variable N PEP, n P value Relative risk 1 guidewire exchange Retained stent at 2 weeks Minor papilla intervention (sphincterotomy) a 3-day hospitalization, n 1). Four patients with mild postprocedure bleeding required a single day hospitalization. Figure 3. Histogram displaying post-ercp pancreatitis incidence in 3Fr and 5Fr stent groups (intention-to-treat analysis). Incidence of PEP in 5Fr and 3Fr groups; P.3. PD sphincterotomy, and ampullectomy as stent placement indication(s), as well as failure to place a 3Fr stent requiring crossover to 5Fr stent were not associated with higher incidence of PEP (Tables 2 and 3). Among the 79 patients in whom the pancreatic stent dislodged within 24 hours, 7 (9%) developed PEP. These 7 patients had received a 5Fr stent. Other Complications The overall complication rate (including PEP) was 15%. There were 3 cases of retroperitoneal perforation, 1 case of iatrogenic guidewire-induced pancreatic duct perforation, and 5 cases of procedural bleeding. One patient with a retroperitoneal perforation also had mild PEP. This patient was randomized to receive a 3Fr stent, which failed to migrate by 2 weeks and required endoscopic removal. All cases were managed conservatively without the need for operative intervention. There were 5 cases of intraprocedural bleeding (4 mild, 1 moderate) (post ampullectomy treated endoscopically, n 2; bleeding at sphincterotomy site managed endoscopically, n 1; Mallory-Weiss tear with spontaneous resolution of bleeding, n 1, and a case of moderate bleeding at precut sphincterotomy site requiring 3 upper endoscopies for control of bleeding and Table 2. Univariate Analysis to Determine Effect of Listed Variables on Incidence of PEP Variable n PEP, n P value Early stent ( 24 hour) migration Failure to place intended stent requiring crossover SOD as stent insertion indication PD (major) sphincterotomy Precut sphincterotomy Ampullectomy Prior PEP Discussion Evidence for the efficacy of temporary pancreatic duct stenting for prophylaxis of PEP in high risk patients has been demonstrated in multiple trials 3,7 13 and in 1 meta-analysis. These studies led to widespread adoption of temporary PD stent placement for prevention of PEP. However, controversy exists regarding which type of pancreatic stent is best. An ideal PD stent should dislodge spontaneously, completely prevent the occurrence of PEP, deploy easily, and not cause stentinduced ductal or parenchymal pancreatic changes. A large, retrospective study suggested that unflanged, longer length (8 cm to 10 cm), 3Fr polyethylene stents with a single duodenal pigtail were associated with significantly higher spontaneous dislodgement rates when compared with larger caliber, shorter length unflanged 4Fr and 5Fr stents. The authors of this retrospective study also reported a lower incidence of PEP in patients who received a 3Fr stent compared with patients who received a 5Fr stent, although the difference was not statistically significant. 16 In this randomized prospective trial, we identified a significantly higher rate of spontaneous dislodgement of 5Fr, short length stents (98%). The rate of spontaneous migration of 3Fr stents in the current study is comparable to the reported rates of migration of smaller caliber but longer length 3Fr stents (88% vs 86%). Our results also demonstrate that approximately 90% of straight, unflanged short length 5Fr stents and 70% of longer length 3Fr stents dislodge by 1 week. There is a clinically important further increment in spontaneous dislodgement of these stents at 2 weeks (98% and 88%). Thus, an abdominal radiograph obtained at 2 weeks after stent placement will ascertain spontaneous dislodgement in the majority of patients. Although stent-induced pancreatic duct changes remain a concern, it is probably reasonable to assume relative safety of short duration (because the majority dislodge within 1 week) of short length 5Fr stent insertion in the head region of pancreas downstream to the pancreatic genu. An important technical factor, which may influence the type of temporary prophylactic PD stent chosen is the technical ease of placement, as prolonged and unsuccessful attempts at placement may increase the risk for development of PEP. 18,19 Deployment of long length 3Fr pancreatic stents may be technically more difficult because of the need for a smaller caliber inch wire which can be difficult to maneuver around tortuous pancreatic ducts when compared with a hydrophilic tip inch wire. A long stent placement also requires a deeper guidewire access in the main pancreatic duct, which may not be possible in patients with

5 838 CHAHAL ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 8 highly angulated or tortuous pancreatic ducts (Figure 1). Thus, we found a significantly higher stent placement failure rate (9% vs 0%) in patients when randomized to receive 3Fr PD stents compared with those randomized to receive 5Fr PD stents. Indeed, in 1 case series, attempted but failed PD stent placement markedly increased the pancreatitis risk in an existing high risk patient population. 19 In our series, in all cases where a longer-length 3Fr stent placement attempt failed, a 5Fr PD stent was successfully placed for prophylaxis of PEP. Patients who crossed over to 5Fr stents did not have an increase in incidence of PEP by univariate analysis presumably because these patients received a 5Fr stent for prophylaxis of PEP (rather than no stent). Failure to place a pancreatic stent in enrolled patients occurred in 5 cases: inability to deeply access the pancreatic duct occurred in 4, and in 1 case a standard biliary sphincterotome tip used to push a 3Fr stent over a guidewire became embedded in the tip of the stent, requiring removal of the stent and guidewire with the sphincterotome. The endoscopist elected not to replace the stent. Of these 5 patients, 2 patients in whom pancreatic duct access was not achieved developed mild PEP. This highlights the probability of markedly increased risk of PEP associated with failed stent placement. 19 Univariate analysis demonstrated a significant association between the use of more than one guidewire and higher risk of PEP. This is in part because of the fact that several patients had a deep pancreatic duct cannulation with a inch guidewire prior to PD stent randomization, and required exchange to a inch guidewire. In addition, in some subjects who were randomized to 3Fr stents but failed pancreatic duct guidewire cannulation with a inch wire, a hydrophilic inch guide wire was advanced to the pancreatic tail and then exchanged for a inch wire over a standard catheter. These manipulations could have contributed to a higher incidence of PEP in the 3Fr group. Moreover, there was a case of guidewireinduced ductal perforation illustrating the fact that pancreatic stent placement is not risk-free. Caution should be exercised while advancing hydrophilic guidewires in the pancreatic duct, and the duct contour and course should be defined by pancreatography. As evident from the indications for stent placement, our patients were considered to be at high risk for the development of PEP. There is a wide variation among endoscopists with regard to the patient selection for stent placement. 20 However, the majority of experts recommend, and evidence from the literature supports, that the following patients are at high risk for PEP and should undergo temporary prophylactic PD stent placement: suspected or known SOD 21 undergoing either biliary or pancreatic sphincterotomy with or without sphincter of Oddi manometry, pancreatic sphincterotomy (major or minor papilla), 22 ampullectomy, 23 precut sphincterotomy, difficult cannulation, and balloon dilation of an intact papilla. 24,25 In our heterogeneous, high risk patient population, the rates of PEP were comparable to the PEP rates reported in the literature in high risk patient populations in whom PD stents were placed. 3,9,16 The majority of cases were classified as mild. Among the various high risk stent placement indications for prophylaxis of PEP, only minor papilla intervention (sphincterotomy) was associated was significant risk ratio (4.05) of increased incidence of PEP despite prophylactic stent insertion. This might be due to the synergistic effect of multiple risk factors in this subgroup of patients, as all of them had prior history of recurrent pancreatitis, 26 which in itself was not associated with higher incidence of PEP after PD stent insertion in our study. There are few data to indicate the optimal duration of PD stent retention for a stent to be effective in reducing the risk of PEP. There is also a concern that early dislodgement of shorter length stents in high risk patients may result in a higher incidence of delayed onset PEP. In our study, 32% (79 out of 249) of stents dislodged spontaneously at 24 hours in both groups. The incidence of PEP was similar among the patients in whom the stent dislodged at 1 day versus 1 week and 2 week dislodgement times. This also corroborates our retrospective data, which also failed to reveal any association between early stent migration and incidence of PEP. 27 However, all the 7 cases of mild PEP in early ( 24 hours) stent migration patients occurred in the 5Fr group. Thus, due to preponderance for early dislodgement in short, unflanged 5Fr stents, there might be a role for short, small caliber flanged stents in certain high risk patients. On the contrary, retained pancreatic stents at 2 weeks postinsertion, the majority of which were 3Fr, were associated with a higher incidence of PEP. The exact cause of this is unclear and could be a chance association. Caution should be exercised at the time of stent removal, and it should be performed as atraumatically as possible to prevent immediate occurrence of pancreatitis after stent extraction. 28 This occurred in 1 of our patients. Even though patients with known or suspected SOD are considered to have the highest incidence of PEP when compared with other indications, we found the incidence of PEP was unchanged in SOD vs non-sod patients, probably because PD stents equalize the PEP risk. In summary, we found a significantly higher spontaneous dislodgement rate of unflanged, straight, small 5Fr stents compared with longer 3Fr stents at 2 weeks, thus obviating the need for endoscopic removal. Obtaining an abdominal radiograph at 2 weeks can demonstrate the stent dislodgement in the majority of patients who receive either 3Fr or short 5Fr stents. Early migration of 5Fr stents did not appear to be associated with higher incidence of PEP. Even though both stents appear to be equally effective in preventing PEP, there is a significantly higher stent placement failure rate with 3Fr stents, particularly in patients with ansa pancreaticus or extremely angulated pancreatic ducts. Thus, an endoscopist should consider placing a shorter PD stent in such cases once the above-mentioned ductal variants become evident on a pancreatogram to minimize pancreatic duct manipulation. Supplementary data To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi: / j.cgh References 1. Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37: Freeman ML. Adverse outcomes of ERCP. Gastrointest Endosc 2002;56:S273 S Tarnasky P, Palech Y, Cunningham J, et al. Pancreatic stenting prevents pancreatitis after biliary sphincterotomy in patients with

6 August Fr vs 3Fr PANCREATIC STENTS 839 sphincter of Oddi dysfunction. Gastroenterology 1998;115: Jacob L, Geenen JE, Catalano MF, et al. Prevention of pancreatitis in patients with idiopathic recurrent pancreatitis: a prospective non-blinded randomized study using endoscopic stents. Endoscopy 2001;33: Ohshio G, Saluja A, Steer ML. Effects of short-term pancreatic duct obstruction in rats. Gastroenterology 1991;100: Lerch MM, Saluja AK, Runzi M, et al. Pancreatic duct obstruction triggers acute necrotizing pancreatitis in the opossum. Gastroenterology 1993;104: Tarnasky PR, Cunningham JT, Cotton PB, et al. Pancreatic sphincter hypertension increases the risk of post-ercp pancreatitis. Endoscopy 1997;29: Smithline A, Silverman W, Rogers D, et al. Effects of prophylactic main pancreatic duct stenting on the incidence of biliary endoscopic sphincterotomy-induced pancreatitis in high-risk patients. Gastrointest Endosc 1993;39: Fogel EL, Eversman D, Jamidar P, et al. Sphincter of Oddi dysfunction: pancreaticobiliary sphincterotomy with pancreatic stent placement has a lower rate of pancreatitis than biliary sphincterotomy alone. Endoscopy 2002;34: Elton E, Howell DA, Parsons WG, et al. Endoscopic pancreatic sphincterotomy: indications, outcome, and a safe stentless technique. Gastrointest Endosc 1998;47: Fazel A, Quadri A, Catalano MF, et al. Does a pancreatic duct stent prevent post-ercp pancreatitis? A prospective randomized study. Gastrointest Endosc 2003;57: Tsuchiya T, Itoi T, Sofuni A, et al. Temporary pancreatic stent to prevent post endoscopic retrograde cholangiopancreatography pancreatitis: a preliminary, single-center, randomized controlled trial. J Hepatobiliary Pancreat Surg 2007;14: Aizawa T, Ueno N. Stent placement in the pancreatic duct prevents pancreatitis after endoscopic sphincter dilation for removal of bile duct stones. Gastrointest Endosc 2001;54: Singh P, Das A, Isenberg G, et al. Does prophylactic pancreatic stent placement reduce the risk of post-ercp acute pancreatitis? A meta-analysis of controlled trials. Gastrointest Endosc 2004; 60: Das A, Singh P, Sivak MV, et al. Pancreatic stent placement for prevention of post-ercp pancreatitis: A cost-effectiveness analysis. Gastrointest Endosc 2007;65: Rashdan A, Fogel EL, McHenry L, et al. Improved stent characteristics for prophylaxis of post-ercp pancreatitis. Clin Gastroenterol Hepatol 2004;2: Axon ATR, Classen M, Cotton PB, et al. Pancreatography in chronic pancreatitis: International definitions. Gut 1984;25: Freeman ML. Post-ERCP pancreatitis: Patient and techniquerelated risk factors. JOP 2002;3: Freeman ML, Overby C, Qi D. Pancreatic stent insertion: consequences of failure and results of a modified technique to maximize success. Gastrointest Endosc 2004;59: Brackbill S, Young S, Schoenfeld P, et al. A survey of physician practices on prophylactic pancreatic stents. Gastrointest Endosc 2006;64: Masci E, Mariani A, Curioni S, et al. Risk factors for pancreatitis following endoscopic retrograde cholangiopancreatography: a meta-analysis. Endoscopy 2003;35: Aronson N, Flamm CR, Bohn RL, et al. Evidence-based assessment: patient, procedure, or operator factors associated with ERCP complications. Gastrointest Endosc 2002;56(Suppl 6): S294 S Harewood GC, Pochron NL, Gostout CJ. Prospective, randomized, controlled trial of prophylactic pancreatic stent placement for endoscopic snare excision of the duodenal ampulla. Gastrointest Endosc 2005;62: DiSario JA, Freeman M, Bjorkman D, et al. Endoscopic balloon dilation compared with sphincterotomy for extraction of bile duct stones. Gastroenterology 2004;127: Baron TH, Harewood GC. Endoscopic balloon dilation of the biliary sphincter compared to endoscopic biliary sphincterotomy for removal of common bile duct stones during ERCP: A metaanalysis of randomized controlled trials. Am J Gastroenterol 2004; 99: Freeman ML. Pancreatic stents for prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis. Clin Gastroenterol Hepatol 2007;5: Chahal P, Baron TH, Petersen BT, et al. Pancreatic stent prophylaxis of post-ercp pancreatitis: Spontaneous migration rates and clinical outcomes. Minerva Gastroenterol Dietol 2007;53: Tarnasky PR. Mechanical prevention of post-ercp pancreatitis by pancreatic stents: results, techniques, and indications. JOP 2003; 4: Reprint requests Address requests for reprints to: Todd H. Baron, MD, Mayo Clinic College of Medicine, Gastroenterology, Charlton 8, 200 First Street Southwest, Rochester, Minnesota baron.todd@ mayo.edu; fax: (507) Conflicts of interest The authors disclose no conflicts.