ORIGINAL ARTICLE: Clinical Endoscopy

Post-ERCP pancreatitis rates do not differ between needle-knifeand pull-type pancreatic sphincterotomy techniques:a multiendoscopist 13-year experience

Christopher Lawrence, MD, Joseph Romagnuolo, MD, FRCPC, MSc(Epid), Peter B. Cotton, MD, FRCP, FRCS,K. Mark Payne, MD, Robert H. Hawes, MD

Charleston, South Carolina, USA

Background: Pancreatic sphincterotomy is one of several factors associated with an increased risk of post-ERCPpancreatitis (PEP). The needle-knife pancreatic sphincterotomy technique (NKS) is purported to result in less-frequent post-ERCP pancreatitis compared with a standard pull-type sphincterotomy (PTS).

Objective: Our purpose was to analyze the experience with both endoscopic pancreatic sphincterotomy (EPS)techniques with respect to post-ERCP pancreatitis at a single tertiary-level referral center.

Design: Retrospective analysis.

Setting: Tertiary-care medical center (Charleston, South Carolina).

Patients: Patients without chronic pancreatitis and with normal retrograde pancreatogram who underwent EPSbetween 1994 and 2007 were identified. Patients were excluded for the following reasons: pancreatic stent notplaced, both sphincterotomy techniques used, any balloon dilation of the ampullary orifice, precut or accesspapillotomy, pancreas divisum.

Results: A total of 481 patients were identified and underwent 510 ERCPs. Indications for ERCP were recurrentpancreatic-type pain (n Z 353) or pancreatitis (n Z 157). NKS was used for 395 of 510 (77.5%) cases versus 115of 510 (22.5%) in which PTS was used. The incidence of post-ERCP pancreatitis was no different between NKS(25/395, 6.4%) and PTS (9/115, 7.8%). Most cases were mild pancreatitis; a single episode of severe PEP occurredin each group.

Conclusions: The risk of post-ERCP pancreatitis does not differ between EPS techniques when performed ata high-volume pancreaticobiliary referral center when using routine prophylactic pancreatic duct stent place-ment. (Gastrointest Endosc 2009;69:1271-5.)

Pancreatitis remains the most frequent complicationafter ERCP, occurring in approximately 5% of patients.1 Bothpatient and endoscopist factors play a role in the occur-rence of post-ERCP pancreatitis (PEP). The former includeage, sex, suspected sphincter of Oddi dysfunction (SOD),and history of pancreatitis.1-3 Factors dependent on theendoscopist include difficult biliary cannulation, recurrent

Abbreviations: EPS, endoscopic pancreatic sphincterotomy; NKS,

needle-knife sphincterotomy; PEP, post-ERCP pancreatitis; PTS, pull-

type sphincterotomy; SOD, sphincter of Oddi dysfunction.

DISCLOSURE: All authors disclosed no financial relationships relevant

to this publication.

Copyright ª 2009 by the American Society for Gastrointestinal Endoscopy

0016-5107/$36.00

doi:10.1016/j.gie.2008.10.015

www.giejournal.org

pancreatic duct injections, failed pancreatic stent insertion,and pancreatic sphincterotomy.1-5

Endoscopic pancreatic sphincterotomy (EPS) reducesthe frequency of recurrent pancreatitis attacks in manypatients with an obstructive etiology (eg, pancreas divisum,SOD) and facilitates interventions such as pancreatoscopyand removal of pancreatic ductal lithiasis.6-10 However,pancreatic sphincterotomy in itself is a risk factor for PEP,and the risk becomes additive as additional risk factorsaccumulate. The incidence of PEP can be ameliorated tosome extent with prophylactic pancreatic duct stenting.11-13

The technique used to effect the pancreatic sphincter-otomy has been purported to be associated with therisk of PEP,12,14 although data to the contrary have beenpublished.9 The use of any lower-risk technique wouldobviously provide the endoscopist an opportunity to

Volume 69, No. 7 : 2009 GASTROINTESTINAL ENDOSCOPY 1271

Post-ERCP pancreatitis rates in sphincterotomy Lawrence et al

mitigate the possibility of PEP. EPS is performed by use of1 of 2 techniques. The first method uses a standard (oftensmaller caliber) wire-guided sphincterotome oriented inthe typical 1 o’clock position at the major papilla. Thesecond method entails placing a small-caliber pancreaticduct stent followed by needle-knife sphincterotomy usingthe stent as a guide. Both methods may be applied at themajor as well as the minor papilla.

A retrospective study of patients at high risk of PEPshowed a reduced incidence of PEP when pancreaticsphincterotomy was done with the needle-knife papillo-tome as opposed to a pull-type device.12 A more recentprospective study likewise showed needle-knife sphincter-otomy to be superior.14 The current study reviews the13-year experience comparing these 2 techniques in a high-risk homogeneous population at a tertiary-care universitymedical center specializing in complex pancreaticobiliarydisease.

PATIENTS AND METHODS

PatientsAll patients with a normal ERCP undergoing EPS at the

major papilla between January 1994 and January 2007were identified with use of the Medical University of SouthCarolina endoscopy database (GI-Trac, Akron SystemsDevelopment, Charleston, SC). Patients were excludedfrom analysis for any of the following reasons: evidenceof chronic pancreatitis (cross-sectional imaging, EUS R5criteria, pancreatography), complete or partial pancreasdivisum, no prophylactic pancreatic duct stent placement,use of both sphincterotomy techniques to complete thepancreatic sphincterotomy, concurrent balloon dilation ofthe ampullary orifice, or incomplete ‘‘access’’ papillotomy.

MethodsPancreatic sphincterotomy was performed with a nee-

dle-knife sphincterotome (NKS) or a pull-type sphinctero-tome (PTS). The sphincterotomy technique was at thediscretion of the endoscopist. All patients had transpapil-lary pancreatic duct stent placement (before sphincterot-omy in the NKS group, after sphincterotomy in the PTSgroup) placed for PEP prophylaxis. Since 2000, 3F flange-less pig-tail stents have been the preferred stent at our in-stitution; before 2000, 5F stents were generally preferred.For NKS, the cut was performed with the pancreatic ductstent as a guide. PTS was made over a guidewire in theupstream pancreatic duct with the cutting wire orientedtoward the 1 o’clock position in the usual fashion. Thebiliary sphincterotomy was generally completed duringthe same procedure unless a prior biliary sphincterotomyhad been performed. The ERBE electrosurgical generator(ERBE USA, Marietta, Ga) set at 150W/ENDOCUT/effect3 was used for all sphincterotomies November 1997to present; a conventional electrosurgical generator

1272 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 7 : 2009

Capsule Summary

What is already known on this topic

d Endoscopic pancreatic sphincterotomy reduces thefrequency of recurrent pancreatitis attacks in manypatients with an obstructive etiology but is also a riskfactor for post-ERCP pancreatitis.

What this study adds to our knowledge

d In a retrospective review of 510 procedures at a singlehigh-volume tertiary referral center, the risk of post-ERCPpancreatitis did not differ between needle-knifepancreatic sphincterotomy and standard pull-typesphincterotomy (6.4% and 7.8%, respectively).

(SSE2L, ValleyLab, Boulder, Colo) was used before thatdate.

Sphincter of Oddi manometry was performed with anaspirating triple-lumen catheter (Cook, Winston-Salem,NC). Continuous aspiration with simultaneous water per-fusion was standard for pancreatic manometry. An abnor-mal manometry run was defined as baseline basalsphincter pressure elevation more than 40 mm Hg, pres-ent in sequential leads, and sustained for a minimum of30 seconds.

Patients undergoing EPS were routinely admitted toa short-stay 23-hour observation unit under the care ofthe pancreaticobiliary attending physician. PEP was deter-mined by the presence of leukocytosis and typical pancre-atic-type pain the morning after the procedure. Pancreaticenzymes were checked if PEP was clinically suspected. Allcomplications were defined and graded according toestablished consensus criteria.15 Complications are pro-spectively entered into the GI-Trac database.

Statistical analysisStatistical analysis was performing with the c2 test for

categorical variables and the Student t test for continuousdata, with 95% binomial CIs calculated (Stat a, version 7.0,College Station, Tex). The study was approved by the Med-ical University of South Carolina Institutional ReviewBoard.

RESULTS

A total of 481 patients meeting the study criterion un-derwent EPS on 510 occasions. NKS was used for 395 of510 (77.5%) pancreatic sphincterotomies; PTS was usedfor 115 of 510 (22.5%) sphincterotomies. Pancreaticmanometry was performed in 437 of 510 (85.7%) cases.A total of 336 of 437 (76.9%) were treated with NKS,and 101 of 437 (24.4%) were treated with PTS.

The indication for ERCP was pancreatitis (idiopathicrecurrent, severe, or smoldering) in 157 of 510 (30.8%)

www.giejournal.org

Lawrence et al Post-ERCP pancreatitis rates in sphincterotomy

cases. ERCP was performed because of recurrent pancre-atic-type pain (ie, pancreatic SOD) in an additional 353(69.2%) instances. There were no statistically significantdifferences between the 2 groups (Table 1).

The incidence of post-ERCP pancreatitis was not signif-icantly different regardless of pancreatic sphincterotomytechnique (Table 2). Post-ERCP pancreatitis occurred in25 of 395 (6.4%) cases where the NKS technique wasused. The majority of PEP was mild (19/25, 76%), but therewere 5 episodes of moderate PEP and a single episode ofsevere (fatal) PEP in the NKS group. In comparison, withthe PTS technique PEP developed in 9 of 115 (7.8%) cases.Again, the majority of cases were mild (7/9, 77.8%); a singlecase of moderate and severe PEP was observed in thisgroup. The frequency of post-ERCP pancreatitis was notsignificantly different between the 5F and 3F stent periods(9.7% before 2000 vs 5.8% 2000-present, P Z .20).

Post-ERCP complications other than PEP were alsocomparable between the 2 sphincterotomy techniques.In the NKS group, there were 6 cases of postsphincterot-omy bleeding (severe 1, moderate 2, mild 3) and 1 sphinc-terotomy-related retroperitoneal perforation managednonoperatively, resulting in a non-PEP complication rateof 1.8%. In the PTS group, immediate bleeding in a singlepatient was the only additional complication (non-PEPcomplication rate 0.9%). The overall complication ratesfor NKS and PTS were 8.4% and 8.7%, respectively.

DISCUSSION

Pancreatitis is the most common adverse event afterERCP, occurring in approximately 5% of patients.1 Certainhigh-risk patients have been identified, and ERCP is best

TABLE 1. Comparison of NKS and PTS groups

NKS

(n Z 395)

PTS

(n Z 115)

P

value

Age (y)

(median, range)

50, 13-88 49, 18-84 .19

Sex (female/male) 326:69 94:21 .84

Indication .92

Pancreatitis 122 (31%) 35 (30%)

SOD 273 (69%) 80 (70%)

Manometry 336 (85%) 101 (88%) .46

Prophylactic stent 395 (100%) 115 (100%)

Cholecystectomy 315 (80%) 95 (83%) .50

Prior PEP 19 (4.8%) 4 (3.5%) .54

NKS, Needle-knife sphincterotomy; PTS, pull-type sphincterotomy;

SOD, sphincter of Oddi dysfunction; PEP, post-ERCP pancreatitis.

www.giejournal.org

avoided in this group without a strong indication. In addi-tion to appropriate patient selection, the endoscopist cancontrol other factors associated with post-ERCP pancreatitis.Prophylactic pancreatic duct stenting, especially in the high-risk patient, is one such factor.11-13

EPS is used to treat recurrent pancreatitis resultingfrom an obstructive etiology (eg, pancreas divisum,SOD) and to facilitate additional endoscopic pancreaticinterventions. However, pancreatic sphincterotomy isassociated with a greater likelihood of PEP.1,2,4 The tech-nique of pancreatic sphincterotomy and its relationshipto PEP is addressed in the current study.

Our data show that the incidence of PEP in a high-riskpatient cohort without evidence of chronic pancreatitis isno different whether the needle-knife or pull technique isused to perform pancreatic sphincterotomy. The rate ofPEP in our study was 6.4% and 7.8% for these 2 groups,respectively. The overall complication rate was 8.4%in the NKS group and 8.7% in the PTS group. There wasa single instance of severe PEP in each group.

Limited published data exist for rates of PEP after nee-dle-knife or pull-type sphincterotomy. Some have theo-rized that early pancreatic duct stent placement (ie,before pancreatic sphincterotomy) may result in moreprompt pancreatic duct drainage and perhaps a lowerrisk for pancreatitis. The most recent study addressingthis issue was a well-designed prospective randomizedevaluation of pancreatic sphincterotomy in a high-riskcohort.14 The study was prematurely halted when theinterim analysis strongly favored the NKS group, althoughthe PEP rates were unusually low with the NKS and rather

TABLE 2. Complications between NKS and PTS groups

NKS PTS

P

value

PEP 25 (6.4%) 9 (7.8%) NS

Mild 19 7

Moderate 5 1

Severe 1 1

PEP

Pre-2000 (5F era) 10/95 (10.5%) 1/18 (5.6%) NS

2000 (3F era) 15/300 (5.0%) 8/97 (8.2%) NS

Non-PEP

complications

7 (1.8%) 1 (0.9%) –*

Bleeding 6 1

Perforation 1 0

Total complications 32 (8.4%) 10 (8.7%) NS

NKS, Needle-knife sphincterotomy; PTS, pull-type sphincterotomy;

PEP, post-ERCP pancreatitis; NS, not significant.

*No statistical analysis performed given sample size.

Volume 69, No. 7 : 2009 GASTROINTESTINAL ENDOSCOPY 1273

Post-ERCP pancreatitis rates in sphincterotomy Lawrence et al

high in the comparison group (PEP 0% NKS vs 29% PTS).A retrospective study from the Indiana University groupfound a PEP rate of approximately 11% with the NKS tech-nique and a rate of 19% with the PTS technique; the differ-ence was not statistically significant.12

A distinctive feature of the current study is the inclusionof ERCPs performed by a number of endoscopists. Over thecourse of our study, 8 attending physicians performedERCPs included in our analysis. All ERCPs were performedby 2 endoscopists in the study by Varadarajulu et al14; simi-larly, 98.7% were done by 2 endoscopists in the otherstudy.12 We consider the breadth of participating endoscop-ists a strength because it likely removes the preference foror proficiency in a single technique as a potential con-founder in analyses of complications and clinical responseoutcomes. Although randomization balances confoundersthat may threaten the internal validity in a study, forcingan endoscopist to perform a technique with which he orshe may be less comfortable, especially in the context ofan individual patient/anatomy, can be artificial.

The sample size of the current study is by far the largestto date addressing EPS in a high-risk cohort. We included481 patients (undergoing 510 pancreatic sphincteroto-mies) with at least 1 PEP risk factor (suspected SOD, his-tory of pancreatitis, etc). This is approximately twice thesample size of the retrospective analysis and, not surpris-ingly, considerably larger than the single-center random-ized study mentioned earlier.

Across more than a decade of ERCP at our institution,electrocautery equipment and prophylactic pancreaticstent caliber have shifted. Despite an early report demon-strating a lower risk of PEP with use of pure cut electro-cautery current,16 the majority of subsequent studies,including several meta-analyses, did not demonstratea benefit with use of pure cut electrocautery current.4,17-19

Our shift in electrosurgical generators, therefore, likelyhad no effect on PEP incidence. The preferred prophylac-tic pancreatic stent evolved from 5F to 3F over the courseof the study. A recent large retrospective study of morethan 1800 patients could not show a significant differencein PEP (9.8% with 5F stents, 7.5% with 3F stents).20 Like-wise, PEP rates at our institution, irrespective of the pan-creatic sphincterotomy technique, were not significantlydifferent in the 5F stent period compared with the 3Fstent era (9.7% and 5.8%, respectively). Additionally, PEPoccurred at a similar rate between the respective tech-niques when each era was analyzed separately, as demon-strated in Table 2.

The results are limited by their retrospective nature.PEP is likely to be identified more frequently in prospec-tive studies. However, all patients were observed in hospi-tal for 24 hours; by definition, all PEP should be capturedwith our approach. Non-PEP complications may not havebeen captured so reliably. Still, the relative incidence be-tween groups of the non-PEP events is likely unaffected,with missed cases equally distributed between the 2

1274 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 7 : 2009

groups. It should also be highlighted that these resultsare from a tertiary regional pancreaticobiliary referral cen-ter and, as such, may not be generalizable to communityERCP practice. The other two studies12,14 suffer fromthis same issue.

The results of this analysis suggest that in routine prac-tice, with the choice of pancreatic sphincterotomy tech-nique left to the discretion of the endoscopist, bothtechniques of EPS are safe and apparently equivalentwith regard to the incidence of post-ERCP pancreatitis ina high-risk group of patients. A multicenter study shouldbe pursued to address this issue given the limited andconflicting data. In the absence of such data, individualswho use both EPS techniques might wish to explore theirown respective PEP pancreatitis rates. It appears that theendoscopist comfortable using a pull-type sphincterotomytechnique need not feel obligated to use the needle-knifesphincterotome.

REFERENCES

1. Freeman ML, Nelson DB, Sherman S, et al. Complications of endo-

scopic biliary sphincterotomy. N Engl J Med 1996;335:909-18.

2. Freeman ML, DiSario JA, Nelson DB, et al. Risk factors for post-ERCP

pancreatitis: a prospective, multicenter study. Gastrointest Endosc

2001;54:425-34.

3. Vandervoort J, Soetikno RM, Tham TC, et al. Risk factors for compli-

cations after performance of ERCP. Gastrointest Endosc 2002;56:

652-6.

4. Masci E, Mariani A, Curioni S, et al. Risk factors for pancreatitis following

endoscopic retrograde cholangiopancreatography: a meta-analysis.

Endoscopy 2003;35:830-4.

5. Freeman ML, Overby C, Qi D. Pancreatic stent insertion: consequences

of failure and results of a modified technique to maximize success.

Gastrointest Endosc 2004;59:8-14.

6. Costamagna G, Gabbrielli A, Mutignani M, et al. Extracorporeal

shock wave lithotripsy of pancreatic stones in chronic pancreatitis:

immediate and medium-term results. Gastrointest Endosc 1997;46:

231-6.

7. Heyries L, Barthet M, Delvasto C, et al. Long-term results of endoscopic

management of pancreas divisum with recurrent acute pancreatitis.

Gastrointest Endosc 2002;55:376-81.

8. Lehman GA, Sherman S, Nisi R, et al. Pancreas divisum: results of minor

papilla sphincterotomy. Gastrointest Endosc 1993;39:1-8.

9. Attwell A, Borak G, Hawes R, et al. Endoscopic pancreatic sphincterot-

omy for pancreas divisum by using a needle-knife or standard pull-

type technique: safety and reintervention rates. Gastrointest Endosc

2006;64:705-11.

10. Hara T, Yamaguchi T, Ishihara T, et al. Diagnosis and patient manage-

ment of intraductal papillary-mucinous tumor of the pancreas by

using peroral pancreatoscopy and intraductal ultrasonography. Gas-

troenterology 2002;122:34-43.

11. Fazel A, Quadri A, Catalano MF, et al. Does a pancreatic duct stent pre-

vent post-ERCP pancreatitis? A prospective randomized study. Gastro-

intest Endosc 2003;57:291-4.

12. 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;34:280-5.

13. Tarnasky PR, Palesch YY, Cunningham JT, et al. Pancreatic stenting

prevents pancreatitis after biliary sphincterotomy in patients with

sphincter of Oddi dysfunction. Gastroenterology 1998;115:1518-24.

www.giejournal.org

Lawrence et al Post-ERCP pancreatitis rates in sphincterotomy

14. Varadarajulu S, Wilcox CM. Randomized trial comparing needle-

knife and pull-sphincterotome techniques for pancreatic sphinc-

terotomy in high-risk patients. Gastrointest Endosc 2006;64:

716-22.

15. Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy

complications and their management: an attempt at consensus.

Gastrointest Endosc 1991;37:383-93.

16. Elta GH, Barnett JL, Wille RT, et al. Pure cut electrocautery current for

sphincterotomy causes less post-procedure pancreatitis than blended

current. Gastrointest Endosc 1998;47:149-53.

17. Loperfido S, Angelini G, Benedetti G, et al. Major early complications

from diagnostic and therapeutic ERCP: a prospective multicenter

study. Gastrointest Endosc 1998;48:1-10.

18. Macintosh DG, Love J, Abraham NS. Endoscopic sphincterotomy by

using pure-cut electrosurgical current and the risk of post-ERCP

pancreatitis: a prospective randomized trial. Gastrointest Endosc

2004;60:551-6.

www.giejournal.org

19. Verma D, Kapadia A, Adler DG. Pure versus mixed electrosurgical

current for endoscopic biliary sphincterotomy: a meta-analysis of

adverse outcomes. Gastrointest Endosc 2007;66:283-90.

20. Rashdan A, Fogel EL, McHenry L Jr., et al. Improved stent characteris-

tics for prophylaxis of post-ERCP pancreatitis. Clin Gastroenterol

Hepatol 2004;2:322-9.

Received January 4, 2008. Accepted October 11, 2008.

Current affiliations: Medical University of South Carolina, Charleston,

South Carolina, USA.

Reprint requests: Christopher Lawrence, MD, Medical University of South

Carolina, 96 Jonathan Lucas St, Suite 210, POB 250327, Charleston, SC 29425.

If you want to chat with an author of this article, you may contact him at

lawrench@musc.edu.

Volume 69, No. 7 : 2009 GASTROINTESTINAL ENDOSCOPY 1275