Clinical Gastroenterology and Hepatology 2016;14:1521–1532

PERSPECTIVES IN CLINICAL GASTROENTEROLOGY ANDHEPATOLOGY

Prevention of Post–Endoscopic RetrogradeCholangiopancreatography Pancreatitis:Medications and Techniques

Andrew Y. Wang, Daniel S. Strand, and Vanessa M. Shami

Division of Gastroenterology and Hepatology, University of Virginia Health System, Charlottesville, Virginia

This article has an accompanying continuing medical education activity, also eligible for MOC credit, on page e135. Learning Objective–Uponcompletion of this activity, successful learners will be able to identify risk factors associated with post-ERCP pancreatitis and ways of minimizing itsoccurrence.

Abbreviations used in this paper: CCC, catheter cannulation followed bycontrast opacification; CI, confidence interval; ERCP, endoscopic retro-grade cholangiopancreatography; GW, guidewire; IV, intravenous; LPBD,large papillary balloon dilation; NK, needle knife; NNT, number-needed-to-treat; NSAID, nonsteroidal anti-inflammatory drug; NTG, nitroglycerin; OR,odds ratio; PD, pancreatic duct; PEP, post–endoscopic retrograde chol-angiopancreatography pancreatitis; RCT, randomized controlled trial; RR,relative risk; SOD, sphincter of Oddi dysfunction; SS, somatostatin; TPS,transpancreatic sphincterotomy or septotomy.

Most current article

© 2016 by the AGA Institute1542-3565/$36.00

http://dx.doi.org/10.1016/j.cgh.2016.05.026

Over the past 2 decades, it increasingly has been recog-nized that endoscopic retrograde cholangiopancreatog-raphy (ERCP) is the most predictable provocateur of acutepancreatitis, with an incidence of more than 15% in high-risk patients. For this reason, there has been considerableinterest in the effect of periprocedural drug administrationas well as different ERCP techniques on both the incidenceand severity of post-ERCP pancreatitis. Although manyagents and techniques have shown promise in small clin-ical studies, the majority of these have failed to yieldconsistent benefit in larger randomized patient groups.This review summarizes the data on medications and ERCPtechniques that have been studied for the prevention ofpost-ERCP pancreatitis.

Keywords: Pancreatitis; ERCP; Stents; Medications; Prevention.

The most common and dreaded complication ofendoscopic retrograde cholangiopancreatography

(ERCP) is pancreatitis, which has an incidence of morethan 15% in high-risk patients.1–3 Recognition of risk fac-tors for post-ERCP pancreatitis (PEP) is important forboth the interventional endoscopist, and for physicianswho refer patients for ERCP (Table 1). However strongthe indication for ERCP, recognition of medications thathelp to minimize the incidence of PEP (Table 2) andcognizance of techniques that both increase and decreasethe incidence of PEP are essential. In this review, weevaluate the medications and techniques that can influ-ence, and potentially prevent, PEP.

Pharmacologic Prevention

Anti-Inflammatory Drugs

Regardless of the inciting cause of acute pancreatitis,the subsequent elaboration of local and systemic proin-flammatory mediators is central to the development ofsystemic inflammatory response syndrome and serious

end-organ injury.4 This understanding presents a numberof suitable targets for pharmacoprevention in the scenarioin which the initial injury can be defined easily, as in PEP.

Nonsteroidal Anti-Inflammatory Drugs. Nonsteroidalanti-inflammatory drugs (NSAIDs) are potent inhibitorsof cyclooxygenase and phospholipase A2. Between 2003and 2008 there were 4 independently conducted smalltrials of either rectally administered diclofenac (2studies)5,6 or indomethacin (2 studies),7,8 which sug-gested the potential for benefit. A single study investi-gating the oral administration of diclofenac wasnegative.9 All 4 of the studies using the rectal route ofadministration were included in a meta-analysis, with atotal of 912 patients. This study found a relative risk(RR) of 0.36 (95% confidence interval [CI], 0.22–0.60) infavor of rectal NSAIDs for the prevention of PEP.10 Acomplementary meta-analysis of the same data sug-gested a reduction in both mild and moderate to severePEP.11 These data led the European Society for Gastro-intestinal Endoscopy to support the use of rectal NSAIDsfor the prevention of PEP.12

In 2012, Elmunzer et al13 published a multicenter,double-blind, placebo-controlled trial of rectal indo-methacin (2 suppositories of 50 mg given immediatelyafter ERCP) that included a total of 602 patients. Theresults of this study were highly significant and in favorof indomethacin because PEP developed in 16.9% ofcontrol patients and in only 4.4% of those who received

Table 1. Risk Factors for Post-ERCP Pancreatitis in Multivariate Analysis

Yesa Maybeb Noc

Female sex CBD stone absent Small CBD diameterYounger age Normal serum bilirubin Periampullary diverticulumSuspected SOD Chronic pancreatitis absent Pancreas divisumPrior post-ERCP pancreatitis Pancreatic brush cytology Allergy to contrast mediaRecurrent pancreatitis Pain during ERCP Prior failed ERCPPD injection Pancreatic acinarization Therapeutic vs diagnosticPancreatic sphincterotomy Low-volume endoscopist Intramural contrast injectionDifficult or failed cannulation Trainee participation Biliary sphincterotomyPrecut (access) sphincterotomy Balloon dilation (of intact biliary sphincter)d Sphincter of Oddi manometryFailed prophylactic PD stenting

NOTE. Adapted with permission from Freeman56 and updated.CBD, common bile duct.aSignificant by multivariable analysis in most studies or by meta-analysis.bSignificant by univariate analysis only in most studies, or by single multivariate analysis.cInconsistent significance by univariate analysis, not significant by multivariate analysis.dConflicting prospective randomized studies.

1522 Wang et al Clinical Gastroenterology and Hepatology Vol. 14, No. 11

indomethacin (P ¼ .005). In addition, moderate-to-severe pancreatitis was found to be almost twice ascommon in control patients (8.8% vs 4.4%; P ¼ .03).

Although unquestionably well designed, this studywas criticized for its inclusion of high-risk patients, manyof whom had suspected sphincter of Oddi dysfunction(SOD). More than half of the included patients hadsphincter hypertension at manometry before biliarysphincterotomy. In subgroup analysis, there was a trendtoward benefit for patients without SOD (8.5% vs 20.0%;P ¼ .11), although this was not statistically significant.The investigators argued that this finding was likely of

Table 2. Summary of Benefit of Various PharmacologicAgents for Prevention of Post-ERCP Pancreatitis

No benefit/ harmaMixed results/

unclearbShows consistent

benefitc

Aprepitant Allopurinol Rectal diclofenacAprotonin Gabexate mesilate Rectal indomethacinb-carotene Intravenous fluidsBotulinum toxin NafamostatCorticosteroids OctreotideHeparin SecretinInterleukin 10 SomatostatinM-acetyl cysteine UlinastatinNifedipine Intravenous glyceryl

trinitrateOral diclofenacPentoxifyllineRisperidoneSeleniumTopical epinephrineTopical lidocaineTransdermal nitroglycerin

aPublished trials showed no statistical advantage over placebo, or increasedincidence of PEP.bSome RCTs showed a benefit but other RCTs or meta-analyses did not.Insufficient published experience to define a clear role in PEP.cMultiple RCTs and/or meta-analyses showed a consistent and significantreduction in the incidence or severity of PEP.

clinical significance, and the failure to achieve signifi-cance was owing to the relatively small fraction of pa-tients without suspected SOD (only 107 of 602).

Since the publication of this landmark study,13 therehave been 2 additional meta-analyses performed thatincluded these data.14,15 Both showed that the use ofNSAIDs significantly decreased theoverall incidenceofPEP.

The remaining question, in our view, is to determinewhat place NSAIDs occupy in the framework of otherpreventative strategies for PEP, including pancreaticstent placement. This question was partly addressed byan exploratory post hoc analysis by Elmunzer et al,16

which suggested that NSAIDs may have been effica-cious regardless of the placement of a pancreatic duct(PD) stent. More recently, a network meta-analysispublished in Clinical Gastroenterology and Hepatologyreported that NSAIDs in fact may be more efficaciousthan PD stent placement, although this question remainsunsettled.17

It is notable that rectal NSAIDs are extremely inex-pensive. A cost-benefit analysis of indomethacin mono-therapy suggested an average per-patient savings of$1472 when compared with stent placement.16 In anenvironment increasingly conscious of the escalatingcost of health care, these savings warrant seriousconsideration.

It is our practice to administer a single 100-mg doseof indomethacin, via a rectal route, immediately afterERCP, to any patient who might have some increased riskfor PEP, unless there is an absolute contraindication,such as in patients with NSAID allergies.

Corticosteroids. Early studies using corticosteroids,which act principally through the inhibition of phos-pholipase A2, showed some success in reducing PEP.18

However, no benefit in the reduction of PEP wasobserved when larger randomized patient samples wereanalyzed.19,20 Other anti-inflammatory agents that havebeen unsuccessfully tested include pentoxifylline,21 re-combinant human interleukin 10,22 and low-dose

November 2016 Prevention of PEP 1523

unfractionated heparin.23 We do not routinely admin-ister corticosteroids to patients undergoing ERCP for thepurpose of reducing the risk of PEP.

Sphincter Relaxants

Using a mechanism that is conceptually analogous toPD placement, several agents aimed at promoting PDdrainage via papillary relaxation have been investigatedwith respect to PEP. These include secretin (which is alsoa regulatory hormone and is described in the followingsection), nitrates, nifedipine, botulinum toxin, lido-caine,24 and epinephrine.25

Nitrates. In a low-risk population of 186 subjects,2 mg of sublingual glyceryl trinitrate (nitroglycerin[NTG]) administered 5 minutes before ERCP initially wasshown to reduce the overall incidence of PEP signifi-cantly from 17.7% to 7.8%.26 In subsequent series, boththe transdermal27–29 and intravenous30 routes of NTGadministration have been evaluated with mixed results.

In an attempt to clarify the available data, a meta-analysis of 5 randomized controlled trials (RCTs) thatincluded 1660 patients was conducted. This publicationshowed a RR reduction of 0.61 (95% CI, 0.44–0.86) infavor of the administration of NTG. This analysis sug-gested a number-needed-to treat (NNT) of 26 to prevent1 episode of PEP. Unfortunately, this analysis was limitedby significant heterogeneity between individual studiesincluding the route of administration, drug dose, andtiming. Although overall in favor of nitrates, thepreferred transdermal method of delivery did not confera significant advantage over placebo. In addition, use ofnitrates also was associated with a significant increasedrisk of both hypotension (RR, 2.25) and headache (RR,3.64).31 A separate meta-analysis that included 8 RCTsand 1920 patients showed similar results.32

At the present time, the use of nitrates for preventionis not recommended. The combination of conflicting dataregarding their efficacy, as well as the potential forserious side effects (such as hypotension), makes itdifficult to recommend the routine use of nitrates for theprevention of PEP.33

Nifedipine, lidocaine, epinephrine. Systemic nifedi-pine, a dihydropyridine thought to promote papillaryrelaxation through its action on L-type calcium channels,was ineffective in 2 controlled trials.34,35 Similarly,neither lidocaine24 nor epinephrine25 have been shownto be effective in preventing PEP when sprayed topicallyon the major papilla.

Botulinum toxin. In a novel approach, two 25-U ali-quots (each of 0.25 ml) of botulinum toxin were injectedinto the pancreatic sphincter in 12 of 26 randomizedpatients with suspected SOD. All of these subjects hadincreased basal sphincter pressures and subsequentlyunderwent biliary sphincterotomy. This study failed toshow a statistical advantage of botulinum whencompared with sham injection (25% vs 43% respectiverates of PEP; P ¼ .34), and was terminated early because

of the unacceptably high rate of PEP in the control armreceiving saline injection.36 As such, we do not recom-mend using medications that might help with papillarysphincter relaxation to prevent PEP at this time.

Regulatory Hormones

Somatostatin/octreotide. The peptide hormone so-matostatin (SS), along with its synthetic analogue octreo-tide, have been investigated extensively for thechemoprophylaxis of PEP. Both of these compounds func-tion by inhibiting the exocrine secretion of the pancreasdirectly and through the suppression of the complemen-tary actors secretin and cholecystokinin.37 Despite thisphysiologic basis, clinical trials using both agents haveshown inconsistent and often conflicting results.

A meta-analysis, published in 2007, that included 16trials compared the administration of SS with controlsfor the incidence of pancreatitis, hyperamylasemia, andpain. Although pancreatitis occurred in 7.3% of controlscompared with 5.3% of patients treated with SS, thisdifference was not statistically significant.38 In 2008, asecond systematic review that included 15 trials and2621 patients also did not find any benefit in favor oftreatment with octreotide. Similar to the results with SS,pancreatitis occurred in 7.0% of controls compared with5.5% of treated patients, resulting in a nonsignificantodds ratio (OR) of 0.78 (95% CI, 0.57–1.08).39

Despite these earlier results, a more recent meta-analysis that included 3818 patients, examined 17 tri-als of SS or octreotide and suggested that these drugsmight prevent PEP when used in bolus fashion (SS) or insufficient dosages (octreotide). A subgroup analysis ofpatients in this study showed that these drugs were alsomore efficacious in cases of PD injection and biliarysphincterotomy.40 Large-scale studies need to be per-formed before determining the role of SS and octreotidein the prevention of PEP.

Secretin. Secretin is a biologically active peptide witha primary function of driving pancreatic bicarbonatesecretion. Secretin also promotes relaxation of thesphincter of Oddi.41 In a single-center RCT that included869 patients, intravenous (IV) administration of secretinshowed a significant reduction (8.7% vs 15.1%; P ¼.004) in the incidence of PEP when compared withcontrols who received the same volume of IV normalsaline. In subgroup analysis, this study showed thegreatest efficacy in PEP reduction among patients whounderwent biliary sphincterotomy (6 of 129 vs 32 of142; P < .001).42 Although compelling, these data werelimited to a single-center experience. Furthermore, boththe cost and historically limited availability of IV secretinare potential barriers to widespread implementation ofthis agent for PEP.

Please refer to the Supplementary Materials sectionfor additional information on the role of other hormones,protease inhibitors, and other medication therapies inpreventing PEP.

1524 Wang et al Clinical Gastroenterology and Hepatology Vol. 14, No. 11

Intravenous Fluid Administration

Given the long-standing interest in the chemopre-vention of PEP, it is surprising that the linchpin oftherapy for acute pancreatitis has received little atten-tion as a potential preventative (or ameliorative) strat-egy. DiMagno et al43 reported in a hypothesis-generatingstudy that greater periprocedural fluid volume was anindependent protective factor against moderate to severePEP. A study by Buxbaum et al44 included 62 patientsrandomized in a 2:1 fashion to either aggressive IV hy-dration with lactated Ringers (3.0 mL/kg/h during ERCPfollowed by a 20-mL/kg bolus and a 3-mg/kg/h infusionfor 8 hours postprocedure) or standard hydration (1.5mL/kg/h during and for 8 hours postprocedure).Patients in the aggressive hydration group received amedian volume of 3.8 L of lactated Ringer’s solutioncompared with 2.2 L in the standard hydrationgroup (P < .001). Risk factors for the development ofPEP were similar between patient groups. Notably, PEPoccurred only in the standard hydration group (17%)and did not occur in the aggressive hydration group (0%;P ¼ .016).

This study had several limitations. First, the samplesize was small and therefore it was underpowered andprone to type I/II errors. Second, the rate of PEP in thecontrol group was higher than expected (at 17%) and theexperimental arm had fewer than expected (0%) epi-sodes of pancreatitis. Despite these concerns, post-ERCPintravenous hydration remains a tempting target forfuture work on preventing PEP.

Endoscopic RetrogradeCholangiopancreatography Techniquesand Postprocedural Pancreatitis

As most experienced endoscopists know, medica-tions, sedative or otherwise, are not a substitute for goodendoscopic technique. In the following sections, the evi-dence behind ERCP best practices, as they relate tominimizing the risk of PEP, are discussed.

Use of Pancreatic Duct Stents as ProphylaxisAgainst Post–Endoscopic RetrogradeCholangiopancreatography Pancreatitis

The ability to place a small-caliber prophylactic PDstent is required for any endoscopist who engages inperforming ERCPs. As a general rule, if the pre-ERCP riskof pancreatitis is increased because of the patient’sclinical factors or procedural factors (Table 1), it gener-ally is recommended to attempt to place a prophylacticPD stent.

From numerous studies, it is clear that PD stentsreduce the risk and probably the severity of PEP.45–50

Mazaki et al51 conducted a meta-analysis that included

8 studies. In a pooled analysis, 336 patients had PD stentplacement and 344 patients formed the control group.PD stent placement was associated with a statisticallysignificant reduction in PEP (RR, 0.32; 95% CI,0.19–0.52; P < .001). Choudhary et al52 conducted aneven larger meta-analysis that included 8 RCTs (656subjects) and 10 nonrandomized studies (4904 sub-jects). Their meta-analysis of the RCTs showed thatprophylactic PD stents decreased the odds of PEP (OR,0.22; 95% CI, 0.12–0.38; P < .01). The absolute riskreduction was 13.3% (95% CI, 8.8%–17.8%). The NNTwas only 8. Similar findings also were noted from a meta-analysis of the nonrandomized studies.

However, Freeman et al53 showed in a multivariateanalysis of a study of 225 patients at increased risk forPEP, all of whom had attempted PD stent placement, thatunsuccessful pancreatic stent insertion was associatedwith the highest overall odds of PEP (OR, 16.1; 95% CI,1.3–200), which was even greater than that for patientswith SOD (OR, 3.2; 95% CI, 1.4–7.5) and prior PEP (OR,3.2; 95% CI, 1.4–7.1). A similar result was shown by Itoet al54 in a study of ERCP in patients with difficult biliarycannulation. In this study, on multivariate analysis, un-successful PD stenting was found to be the only riskfactor for PEP (OR, 8.3; 95% CI, 2.3–30). Choksi et al55

further corroborated the increased risk of PEP associ-ated with failed PD stenting in a post hoc analysis of datafrom high-risk patients randomized to receive rectalindomethacin or placebo.13 Failed PD stenting occurredin 7.8% of attempts in this study. Patients who did notreceive indomethacin and had failed PD stent placementhad a 34.7% rate of PEP, which was significantly greaterthan the PEP rates in patients in whom PD stenting wasnot attempted or in whom PD stenting was successful.Interestingly, patients who received indomethacin butalso had failed PD stent placement had a 5.3% rate ofPEP, which was not statistically different compared withthe 2 earlier-mentioned comparison groups. Overall,these studies underscore the need for proper trainingand experience in ERCP, including comfort in placementof prophylactic PD stents.

PD stents commonly are made of polyethylene, whichis sometimes mixed with vinyl or other softer mate-rials,56 and typically have side holes (Figure 1).57

Although the optimal size and length for a prophylacticPD stent have not been clearly determined, a meta-analysis suggested that 5F stents were superior to 3Fstents in preventing PEP.58 Generally, it is recommendedto use a small-caliber plastic stent (3F, 4F, or 5F) that iseither short (3- to 4-cm long and does not pass to thegenu) or long (stent ends in the pancreatic tail).59 Forendoscopists who preferentially use an 0.035” or 0.025”guidewire, 5F stents are used more commonly becauseplacement of a thinner stent usually requires opening asecond, smaller-caliber guidewire (typically an 0.018”wire), which adds cost to the procedure. We recommendthat when placing a PD stent as prophylaxis against PEPto use a 5F or smaller plastic stent that is short (�4 cm in

Figure 1. Commercially available pancreatic duct stents.Each stent is oriented from the duodenal end (left) to theintrapancreatic portion (right). Various stent calibers (3F, 4F,and 5F), lengths, and configurations (some with externalpigtails, some with internal flaps/flanges, and so forth) areshown. Most of these stents have side holes placed atstaggered points along the length of the stent. The leftcolumn shows an assortment of plastic (polyethylene)pancreatic stents (Zimmon and Geenen; Cook Medical,Winston-Salem, NC). The middle column shows a variety ofsoft plastic (proprietary material) pancreatic stents (FreemanPancreatic Flexi-Stent; Hobbs Medical, Stafford Springs, CT).The right column shows an array of soft plastic (combinationof polyethylene and vinyl) pancreatic stents (Advanix; BostonScientific, Natick, MA). In general, 5F stents may be passedover 0.035” and thinner guidewires, 4F stents may be passedover 0.025” and thinner guidewires (although some newer 4Fstents can be placed over an 0.035” guidewire), and 3F stentsrequire 0.018” (for most vendors) or 0.021” guidewires, whichmay be more difficult to use.

November 2016 Prevention of PEP 1525

length, so that it does not cross the genu) and reliablywill migrate out of the pancreatic duct within 2 weeks.

Although the data for PD stenting are compelling, it isinteresting that a post hoc analysis by Elmunzer et al16

showed that rectal indomethacin alone appeared to bemore effective for preventing PEP than no prophylaxis,prophylactic PD stenting alone, and the combination ofindomethacin and prophylactic PD stent placement.Although future RCTs to address this question formally areof utmost importance, it is our current recommendationthat in patients at increased risk for PEP, even if PDstenting is successful, strong consideration should be givento concomitant administration of rectal indomethacin.

Guidewire Cannulation Versus CatheterCannulation and Contrast Opacification

The target for bile duct and main PD cannulation issuccess in more than 85% of ERCP procedures, as setforth by the American College of Gastroenterology andthe American Society for Gastrointestinal EndoscopyTaskforce on Quality in Endoscopy.60 These thresholdswere validated by a meta-analysis that found a bile ductcannulation success rate of 89.3% (95% CI, 0.866–0.919)and a pancreatic duct cannulation success rate of 85.0%(95% CI, 0.813–0.886).61

With respect to selective cannulation of the bile duct,there have been at least 5 high-quality RCTs investigatingguidewire (GW) cannulation vs catheter cannulationfollowed by contrast opacification (CCC) (Table 3). In thepast 5 years, 6 RCTs and 1 large prospective studyinvestigated the approaches of GW vs CCC for biliaryaccess. Two studies62,63 showed statistically significantimprovement in the rate of biliary access by using theGW approach, and a single study reported reduced ratesof PEP using the GW approach.48 However, these studieswere heterogeneous and some of the studies includedtrainees who performed ERCP and other studies did notuse prophylactic PD stenting as part of the ERCP pro-cedures. To better address this issue, Tse et al64 con-ducted a Cochrane meta-analysis comparing these 2approaches and included 12 studies in the main analysisinvestigating PEP rates. In aggregate, there were 1784patients in the GW-assisted cannulation group and 1666patients included in the CCC group. Despite significantheterogeneity, unweighted pooled rates of PEP were3.5% for the GW group and 6.7% for the CCC group. TheGW-assisted cannulation technique significantly reducedPEP compared with the contrast-assisted cannulationtechnique based on an intention-to-treat analysis(RR, 0.51; 95% CI, 0.32–0.82; P ¼ .005). With respect topreventing PEP, the NNT was 31 for the GW group. TheGW group also had higher primary cannulation rates (RR,1.07; 95% CI, 1.00–1.15; P ¼ .05). With respect toimproved cannulation success, the NNT was 18 whenusing the GW technique.

Based on these aggregate data, most expert endo-scopists recommend GW cannulation when performingERCP. There is no good evidence to suggest that onecaliber or type of wire is superior to the others in termsof reducing PEP risk.65,66

Use of Pancreatic Duct Access toFacilitate Difficult Biliary Access

In instances with a floppy papilla or an angulatedintrapapillary bile duct, passage of a guidewire can leadto inadvertent pancreatic access, and selective biliaryaccess might not be possible using standard cannulationtechniques. In such situations there are several advancedtechniques that use a PD wire to facilitate selectivebiliary cannulation. First, the double-GW technique: thePD wire is left in place, and a second guidewire then ispassed via a sphincterotome or catheter and alongsidethe pancreatic wire, which allows for fluoroscopicreference, straightens the approach to the bile duct, andoccupies space in the PD, thereby facilitating biliarycannulation67,68 (Figure 2). Second, PD stent placementfollowed by attempted biliary cannulation, which if notsuccessful then a precut needle-knife (NK) biliary pap-illotomy can be performed.68 Third, transpancreaticsphincterotomy or septotomy (TPS) for biliaryaccess,67,69–71 which can be followed by double-GWbiliary cannulation and then PD stenting or,

Table 3. Prospective Studies that Compared Guidewire Cannulation to Catheter Cannulation Followed by Contrast Opacification for Biliary Access

Study Year CountrySites,n

Types ofguidewires

Cannulationdevice

Studydesign

Totalpatients, n

Biliarycannulationsuccess

Pvalue

PEP

Pvalue

PD stentsused

ERCP bytrainees

GW(%)

CCC(%)

GW(%)

CCC(%)

Lella et al90 2004 Italy 1 0.035” TeflonTracer (Cook)

Sphincterotome RCT 400 98.5 97.5 NS 0 4.1 <.01 Not specified N

Artifon et al91 2007 Brazil 3 0.035” Teflon(Cook)

Sphincterotome RCT 300 96.7 94.1 .78 8.6 16.6 .037 N N

Katsinelos et al62 2008 Greece 2 0.035” Jagwire(BostonScientific)

Catheter RCT 332 81.4 53.9 <.001 5.4 7.9 .37 Y N

Bailey et al63 2008 Australia 1 0.035” Jagwire(BostonScientific)

Sphincterotome RCT 413 81.4 73.9 .03 7.9 6.2 .48 Y Y

Lee et al48 2009 Korea 1 0.035” Jagwire(BostonScientific)

Sphincterotome RCT 300 98.7 97 >.5 2 11.3 .001 Y N

Nambu et al92 2011 Japan 1 0.035” AngledJagwire(BostonScientific)

Sphincterotome(for GW) vscatheter

RCT 172 77.9 73.8 NS 2.3 6 .28 Y Y

Mariani et al93 2012 Italy 1 0.035” TracerMetro/Hybrid(Cook)

Sphincterotome Prospective 1249 83.7 85.5 .46 5.2 4.4 .6 Y Y

Kawakami et al94 2012 Japan 15 0.035” Jagwire(BostonScientific)

Sphincterotomeor catheter

RCT 400 70.6 70.4 NS 4.0% 3.0% NS Y Y

Kobayashi et al95 2013 Japan 9 0.035” or 0.025”Teflon

Endoscopist’schoice

RCT 322 83 87 .4 6.1% 6.3% .95 Y Not specified

NOTE. The manufacturers of the guidewires were as follows: Cook Medical, Winston-Salem, NC; Boston Scientific, Natick, MA. Definitions for biliary cannulation success vary among studies. Statistically significant entries areshown in bold.

1526Wang

etal

Clinical

Gastroenterology

andHepatology

Vol.

14,No.

11

Figure 2. Attempted selective biliary cannulation using a sphincterotome and guidewire resulted in PD access. (A and B) ThePD guidewire was left in place, and the sphincterotome was loaded with a second guidewire to re-attempt biliary access. ThePD guidewire straightened the ampulla and biliary approach and also served as a fluoroscopic reference. (C) Biliary accesswas obtained and (D) cholangiography was performed. A biliary sphincterotomy was effected with the pancreatic guidewire inplace, (E and F) after which a 5F, 4-cm, single-pigtail, plastic stent was placed into the main PD via the duct of Wirsung.Retrieval balloon sweep in this patient with an abnormal intraoperative cholangiogram was performed. Indomethacin 100 mgwas given per rectum at the conclusion of the ERCP. The PD stent was left in place at the conclusion of this procedure and hadfallen out at follow-up abdominal radiograph about 2 weeks later (not shown). The patient did not experience PEP.

November 2016 Prevention of PEP 1527

alternatively, by first placing a PD stent and thenattempting biliary cannulation.

Rates of successful biliary access by using each of theseadvanced techniques are high, typically in the 67%–94%range.67,68,72 More recently, several studies73–75 haveshown that precut TPS for biliary access is not associatedwith higher rates of pancreatitis as compared with precutNK papillotomy. Wang et al73 reported a pancreatitisrate of 11.4% in a TPS group compared with 11.8% in aNK papillotomy group. Catalano et al74 found a lowerpancreatitis rate with TPS (3.4%) as compared with NKpapillotomy (11.8%), but this difference was not sta-tistically significant.75 Finally, a RCT by Yoo et al67 thatcompared TPS with the double-GW technique for biliaryaccess in patients with difficult biliary cannulation sur-prisingly found a significantly higher rate of PEP in thedouble-GW group (38.2%) as compared with the TPSgroup (10.8%; P < .011); however, PD stents were notused in this study. In published studies, PEP rates rangefrom 2.4% to 12.5% for TPS.67,70,71,74–76 Although the use

of PD stenting was heterogeneous in these studies, we dorecommend placing a prophylactic PD stent after TPS.

In summary, in instances in which attempts at se-lective biliary access result in repeated PD wire cannu-lation, we recommend leaving the PD wire in place andusing a second wire to attempt biliary access. If notsuccessful, either PD stenting followed by re-attemptedbiliary access and then NK precutting over the PD stentor precut TPS followed by PD stenting can be used insituations in which biliary access is imperative.

Early Precut Biliary Papillotomy

Precut biliary papillotomy using a NK can be per-formed freehand or over a PD stent in the case of inad-vertent PD cannulation (Figure 3), as previouslydescribed. In earlier prospective nonrandomizedstudies,2,77 precutting was associated with an increasedrisk of pancreatitis. This increased risk of PEP has been

Figure 3. A patient with borderline resectable pancreas adenocarcinoma and obstructive jaundice presented for ERCP andbiliary decompression to facilitate neoadjuvant chemoradiation. (A) The ampulla was prominent and elongated. (B) Neither thebile duct nor the pancreatic duct could be cannulated using a sphincterotome and guidewire. (C) A freehand needle-knifebiliary papillotomy was performed for biliary access. (D) Further unroofing of the papilla uncovered the bile duct orifice(arrow). (E) After biliary access, a completion biliary sphincterotomy was effected. (F) After sphincterotomy and dilation of abiliary stricture, copious drainage of dark bile was seen. The pancreas was not entered or opacified during this procedure, andPD stent placement was not attempted, but rectal indomethacin was given. PEP did not occur.

1528 Wang et al Clinical Gastroenterology and Hepatology Vol. 14, No. 11

challenged by a meta-analysis conducted by Cennamoet al78 in 2010 that included 6 RCTs (996 patients) thatwere designed to specifically assess the timing of precutpapillotomies, which were performed in tertiary referralcenters by skilled pancreaticobiliary endoscopists.Overall rates of cannulation were 90% in both the earlyprecut and the standard approach groups. PEP occurredin only 2.5% of patients randomized to early precut, butin 5.3% of patients from the persistent standardapproach group (OR, 0.47; 95% CI, 0.24–0.91). Theseresults were supported by another meta-analysis pub-lished that same year by Gong et al.79

In 2013, Swan et al80 randomized 73 patients withdifficult biliary cannulation to either NK papillotomy orto continued standard attempted cannulation. The mainoutcome measures were PEP and successful biliarycannulation. There was no difference in eventual can-nulation success between the groups. However, 65% ofthe patients assigned to the continued standard cannu-lation group required cross-over to the NK papillotomy

group. There was no significant difference in the devel-opment of PEP among patients in the early NK papil-lotomy group (20.5%) vs the standard cannulation group(17.6%; P ¼ 1.0). A 5F, plastic, single-pigtail stent wasplaced before NK papillotomy if the PD had been can-nulated at least twice, but PD stenting was not pursueddespite NK papillotomy if the PD was not accessed. Fifty-nine percent of patients in the NK papillotomy armreceived PD stents.

In some instances, neither biliary nor PD access willbe possible, thus necessitating NK papillotomy. However,Cha et al50 found in cases in which PD access waspossible, that placing a PD stent to facilitate NK precutpapillotomy and leaving the PD stent in place at theconclusion of the procedure reduced the frequency andseverity of PEP.

These data suggest that when biliary access is notachieved by conventional ERCP techniques after areasonable number or duration of attempts (althougharbitrary, at least 5 attempts or 5–10 minutes) that early

November 2016 Prevention of PEP 1529

precutting is a logical next step that should not beassociated with increased risk of PEP.

Pancreatic Sphincterotomy

As mentioned earlier, pancreatic sphincterotomy is arisk factor for PEP.53,81 Lawrence et al82 retrospectivelyexamined the rates of PEP in patients undergoing NK andpull-type (traction) pancreatic sphincterotomy of themajor papilla over a 13-year period. A total of 6.4% ofpatients who received NK pancreatic sphincterotomysuffered PEP as compared with 7.8% of patients in thepull-type sphincterotomy group, which was not a sta-tistically significant difference. Although data are limited,freehand NK pancreatic sphincterotomy over a PD stentor traction pancreatic sphincterotomy with a wire in themain PD followed by PD stenting both are acceptable.

Large Papillary Balloon Dilation forRemoval of Choledocholithiasis

Large papillary balloon dilation (LPBD) (from 12 to20 mm) of the bile duct for removal of large bile ductstones can be performed without biliary sphincter-otomy, or after a modest or complete biliary sphinc-terotomy. In a meta-analysis by Baron and Harewood,83

PEP occurred more commonly in the papillary balloondilation group (7.4%) as compared with a biliarysphincterotomy group (4.3%; P ¼ .05), althoughbleeding occurred less frequently in the balloon dilationgroup. In fact, a RCT of papillary balloon dilation (to �8mm) vs biliary sphincterotomy was terminated earlybecause of a significantly increased risk of pancreatitis,which in 2 cases led to death.84 However, LPBD afterbiliary sphincterotomy has been shown to be quiteefficacious at removing large bile duct stones with lowrates of pancreatitis (2.2%).85 As such, common practiceis to perform at least a modest biliary sphincterotomybefore LPBD for stone removal. Despite this moreWestern approach to LPBD, several Asian studies havenot shown increased PEP rates with LPBD without priorbiliary sphincterotomy.86–89

LPBD for removal of choledocholithiasis has becomean important technique in the armamentarium of thetherapeutic biliary endoscopist. At this time, we favor amodest biliary sphincterotomy before LPBD. If the PDwas not accessed and there are no other risk factors forPEP, we do not routinely place PD stents after biliarysphincterotomy and LPBD. However, in patients withclinically relevant choledocholithiasis and untreatablecoagulopathy or bleeding diatheses, which makessphincterotomy relatively contraindicated owing to anincreased risk of bleeding, LPBD over a prolonged dila-tion interval with or without PD stenting might bereasonable for bile duct stone clearance.

Please refer to the Supplementary Materials sectionfor additional information on the use of contrast

media, PD injection, and deep-enteroscopy–assistedERCP pertaining to the prevention of PEP.

Summary and Recommendations

In summary, the strongest data concerning the use ofpharmacotherapy in preventing PEP is for the use of rectalindomethacin in patients at increased risk for PEP. Theeffectiveness of using rectal indomethacin in average-riskpatients remains less clear; however, its low cost, ease ofadministration, and favorable side-effect profile make it areasonable preventative medication in this population aswell. In terms of ERCP techniques, placement of a pro-phylactic PD stent when the PD has been manipulated oropacified or in patients at increased risk for PEP is apractice that is supported by numerous high-qualitystudies. Future studies are needed to directly comparethe rates of PEP in patients who receive rectal indo-methacin with those who receive a PD stent. Furthermore,prospective trials also are needed to determine if pro-phylactic PD stenting combined with rectal indomethacinmight be superior to either therapy alone.

Supplementary Material

Note: To access the supplementary material accom-panying this article, visit the online version of ClinicalGastroenterology and Hepatology at www.cghjournal.org,and at http://dx.doi.10.1016/j.cgh.2016.05.026.

References

1. Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincter-

otomy complications and their management: an attempt atconsensus. Gastrointest Endosc 1991;37:383–393.

2. Freeman ML, Nelson DB, Sherman S, et al. Complications ofendoscopic biliary sphincterotomy. N Engl J Med 1996;335:909–918.

3. Cheon YK, Cho KB, Watkins JL, et al. Frequency and severity ofpost-ERCP pancreatitis correlated with extent of pancreaticductal opacification. Gastrointest Endosc 2007;65:385–393.

4. Gross V, Leser HG, Heinisch A, et al. Inflammatory mediatorsand cytokines–new aspects of the pathophysiology andassessment of severity of acute pancreatitis? Hepatogas-troenterology 1993;40:522–530.

5. Murray B, Carter R, Imrie C, et al. Diclofenac reduces the inci-dence of acute pancreatitis after endoscopic retrograde chol-angiopancreatography. Gastroenterology 2003;124:1786–1791.

6. Khoshbaten M, Khorram H, Madad L, et al. Role of diclofenac inreducing post-endoscopic retrograde cholangiopancreatog-raphy pancreatitis. J Gastroenterol Hepatol 2008;23:e11–e16.

7. Sotoudehmanesh R, Khatibian M, Kolahdoozan S, et al. Indo-methacin may reduce the incidence and severity of acutepancreatitis after ERCP. Am J Gastroenterol 2007;102:978–983.

8. Montano Loza A, Rodriguez Lomeli X, Garcia Correa JE, et al.[Effect of the administration of rectal indomethacin on amylaseserum levels after endoscopic retrograde cholangiopancreatog-raphy, and its impact on the development of secondarypancreatitis episodes]. Rev Esp Enferm Dig 2007;99:330–336.

1530 Wang et al Clinical Gastroenterology and Hepatology Vol. 14, No. 11

9. Cheon YK, Cho KB, Watkins JL, et al. Efficacy of diclofenac inthe prevention of post-ERCP pancreatitis in predominantly high-risk patients: a randomized double-blind prospective trial.Gastrointest Endosc 2007;66:1126–1132.

10. Elmunzer BJ, Waljee AK, Elta GH, et al. A meta-analysis of rectalNSAIDs in the prevention of post-ERCP pancreatitis. Gut 2008;57:1262–1267.

11. Zheng MH, Xia HH, Chen YP. Rectal administration of NSAIDs inthe prevention of post-ERCP pancreatitis: a complementarymeta-analysis. Gut 2008;57:1632–1633.

12. Dumonceau JM, Andriulli A, Deviere J, et al. European Society ofGastrointestinal Endoscopy (ESGE) guideline: prophylaxis ofpost-ERCP pancreatitis. Endoscopy 2010;42:503–515.

13. Elmunzer BJ, Scheiman JM, Lehman GA, et al. A randomizedtrial of rectal indomethacin to prevent post-ERCP pancreatitis.N Engl J Med 2012;366:1414–1422.

14. Ding X, Chen M, Huang S, et al. Nonsteroidal anti-inflammatorydrugs for prevention of post-ERCP pancreatitis: ameta-analysis. Gastrointest Endosc 2012;76:1152–1159.

15. Yuhara H, Ogawa M, Kawaguchi Y, et al. Pharmacologicprophylaxis of post-endoscopic retrograde cholangiopancreatog-raphy pancreatitis: protease inhibitors and NSAIDs in a meta-analysis. J Gastroenterol 2014;49:388–399.

16. Elmunzer BJ, Higgins PD, Saini SD, et al. Does rectal indo-methacin eliminate the need for prophylactic pancreatic stentplacement in patients undergoing high-risk ERCP? Post hocefficacy and cost-benefit analyses using prospective clinical trialdata. Am J Gastroenterol 2013;108:410–415.

17. Akbar A, Abu Dayyeh BK, Baron TH, et al. Rectal nonsteroidalanti-inflammatory drugs are superior to pancreatic duct stents inpreventing pancreatitis after endoscopic retrograde chol-angiopancreatography: a network meta-analysis. Clin Gastro-enterol Hepatol 2013;11:778–783.

18. Weiner GR, Geenen JE, Hogan WJ, et al. Use of corticosteroidsin the prevention of post-ERCP pancreatitis. GastrointestEndosc 1995;42:579–583.

19. Sherman S, Blaut U, Watkins JL, et al. Does prophylacticadministration of corticosteroid reduce the risk and severity ofpost-ERCP pancreatitis: a randomized, prospective, multicenterstudy. Gastrointest Endosc 2003;58:23–29.

20. Dumot JA, Conwell DL, O’Connor JB, et al. Pretreatment withmethylprednisolone to prevent ERCP-induced pancreatitis: arandomized, multicenter, placebo-controlled clinical trial. Am JGastroenterol 1998;93:61–65.

21. Kapetanos D, Kokozidis G, Christodoulou D, et al. A randomizedcontrolled trial of pentoxifylline for the prevention of post-ERCPpancreatitis. Gastrointest Endosc 2007;66:513–518.

22. Sherman S, Cheng CL, Costamagna G, et al. Efficacy of re-combinant human interleukin-10 in prevention of post-endoscopic retrograde cholangiopancreatography pancreatitisin subjects with increased risk. Pancreas 2009;38:267–274.

23. Barkay O, Niv E, Santo E, et al. Low-dose heparin for the pre-vention of post-ERCP pancreatitis: a randomized placebo-controlled trial. Surg Endosc 2008;22:1971–1976.

24. Schwartz JJ, Lew RJ, Ahmad NA, et al. The effect of lidocainesprayed on the major duodenal papilla on the frequency of post-ERCP pancreatitis. Gastrointest Endosc 2004;59:179–184.

25. Xu LH, Qian JB, Gu LG, et al. Prevention of post-endoscopicretrograde cholangiopancreatography pancreatitis by epineph-rine sprayed on the papilla. J Gastroenterol Hepatol 2011;26:1139–1144.

26. Sudhindran S, Bromwich E, Edwards PR. Prospective random-ized double-blind placebo-controlled trial of glyceryl trinitrate inendoscopic retrograde cholangiopancreatography-inducedpancreatitis. Br J Surg 2001;88:1178–1182.

27. Moreto M, Zaballa M, Casado I, et al. Transdermal glyceryl tri-nitrate for prevention of post-ERCP pancreatitis: a randomizeddouble-blind trial. Gastrointest Endosc 2003;57:1–7.

28. Kaffes AJ, Bourke MJ, Ding S, et al. A prospective, randomized,placebo-controlled trial of transdermal glyceryl trinitrate inERCP: effects on technical success and post-ERCP pancrea-titis. Gastrointest Endosc 2006;64:351–357.

29. Nojgaard C, Hornum M, Elkjaer M, et al. Does glyceryl nitrateprevent post-ERCP pancreatitis? A prospective, randomized,double-blind, placebo-controlled multicenter trial. GastrointestEndosc 2009;69:e31–e37.

30. Beauchant M, Ingrand P, Favriel JM, et al. Intravenous nitro-glycerin for prevention of pancreatitis after therapeuticendoscopic retrograde cholangiography: a randomized,double-blind, placebo-controlled multicenter trial. Endoscopy2008;40:631–636.

31. Bang UC, Nojgaard C, Andersen PK, et al. Meta-analysis:nitroglycerin for prevention of post-ERCP pancreatitis. AlimentPharmacol Ther 2009;29:1078–1085.

32. Bai Y, Xu C, Yang X, et al. Glyceryl trinitrate for prevention ofpancreatitis after endoscopic retrograde cholangiopancreatog-raphy: a meta-analysis of randomized, double-blind, placebo-controlled trials. Endoscopy 2009;41:690–695.

33. Anderson MA, Fisher L, Jain R, et al. Complications of ERCP.Gastrointest Endosc 2012;75:467–473.

34. Prat F, Amaris J, Ducot B, et al. Nifedipine for prevention ofpost-ERCP pancreatitis: a prospective, double-blind random-ized study. Gastrointest Endosc 2002;56:202–208.

35. Sand J, Nordback I. Prospective randomized trial of the effect ofnifedipine on pancreatic irritation after endoscopic retrogradecholangiopancreatography. Digestion 1993;54:105–111.

36. Gorelick A, Barnett J, Chey W, et al. Botulinum toxin injectionafter biliary sphincterotomy. Endoscopy 2004;36:170–173.

37. Foster E, Leung J. Pharmacotherapy for the prevention of post-ERCP pancreatitis. Am J Gastroenterol 2007;102:52–55.

38. Andriulli A, Leandro G, Federici T, et al. Prophylactic adminis-tration of somatostatin or gabexate does not prevent pancrea-titis after ERCP: an updated meta-analysis. Gastrointest Endosc2007;65:624–632.

39. Bai Y, Gao J, Zou DW, et al. Prophylactic octreotide adminis-tration does not prevent post-endoscopic retrograde chol-angiopancreatography pancreatitis: a meta-analysis ofrandomized controlled trials. Pancreas 2008;37:241–246.

40. Omata F, Deshpande G, Tokuda Y, et al. Meta-analysis: so-matostatin or its long-acting analogue, octreotide, for prophy-laxis against post-ERCP pancreatitis. J Gastroenterol 2010;45:885–895.

41. Sherman S, Lehman GA. Sphincter of Oddi dysfunction: diag-nosis and treatment. JOP 2001;2:382–400.

42. Jowell PS, Branch MS, Fein SH, et al. Intravenous syntheticsecretin reduces the incidence of pancreatitis induced byendoscopic retrograde cholangiopancreatography. Pancreas2011;40:533–539.

43. DiMagno MJ, Wamsteker EJ, Maratt J, et al. Do largerperiprocedural fluid volumes reduce the severity of post-endoscopic retrograde cholangiopancreatography pancrea-titis? Pancreas 2014;43:642–647.

November 2016 Prevention of PEP 1531

44. Buxbaum J, Yan A, Yeh K, et al. Aggressive hydration withlactated Ringer’s solution reduces pancreatitis after endoscopicretrograde cholangiopancreatography. Clin Gastroenterol Hep-atol 2014;12:303–307.

45. Sherman S, Gottlieb K, Uzer MF, et al. Effects of meperidine onthe pancreatic and biliary sphincter. Gastrointest Endosc 1996;44:239–242.

46. Tarnasky PR, Palesch YY, Cunningham JT, et al. Pancreaticstenting prevents pancreatitis after biliary sphincterotomy inpatients with sphincter of Oddi dysfunction. Gastroenterology1998;115:1518–1524.

47. Fazel A, Quadri A, Catalano MF, et al. Does a pancreatic ductstent prevent post-ERCP pancreatitis? A prospective random-ized study. Gastrointest Endosc 2003;57:291–294.

48. Lee TH, Park do H, Park JY, et al. Can wire-guided cannulationprevent post-ERCP pancreatitis? A prospective randomizedtrial. Gastrointest Endosc 2009;69:444–449.

49. Sofuni A, Maguchi H, Mukai T, et al. Endoscopic pancreatic ductstents reduce the incidence of post-endoscopic retrogradecholangiopancreatography pancreatitis in high-risk patients.Clin Gastroenterol Hepatol 2011;9:851–858; quiz e110.

50. Cha SW, Leung WD, Lehman GA, et al. Does leaving a mainpancreatic duct stent in place reduce the incidence of precutbiliary sphincterotomy-associated pancreatitis? A randomized,prospective study. Gastrointest Endosc 2013;77:209–216.

51. Mazaki T, Masuda H, Takayama T. Prophylactic pancreatic stentplacement and post-ERCP pancreatitis: a systematic review andmeta-analysis. Endoscopy 2010;42:842–853.

52. Choudhary A, Bechtold ML, Arif M, et al. Pancreatic stentsfor prophylaxis against post-ERCP pancreatitis: a meta-analysis and systematic review. Gastrointest Endosc 2011;73:275–282.

53. Freeman ML, DiSario JA, Nelson DB, et al. Risk factors for post-ERCP pancreatitis: a prospective, multicenter study. Gastro-intest Endosc 2001;54:425–434.

54. Ito K, Horaguchi J, Fujita N, et al. Clinical usefulness of double-guidewire technique for difficult biliary cannulation in endo-scopic retrograde cholangiopancreatography. Dig Endosc 2014;26:442–449.

55. Choksi NS, Fogel EL, Cote GA, et al. The risk of post-ERCPpancreatitis and the protective effect of rectal indomethacin incases of attempted but unsuccessful prophylactic pancreaticstent placement. Gastrointest Endosc 2015;81:150–155.

56. Freeman ML. Pancreatic stents for prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis.Clin Gastroenterol Hepatol 2007;5:1354–1365.

57. Buscaglia JM, DiMaio CJ, Pollack MJ, et al. Are large side holesassociated with reduced rates of pancreatic stent occlusion?Results of a prospective study. JOP 2009;10:496–500.

58. Afghani E, Akshintala VS, Khashab MA, et al. 5-Fr vs. 3-Frpancreatic stents for the prevention of post-ERCP pancreatitisin high-risk patients: a systematic review and network meta-analysis. Endoscopy 2014;46:573–580.

59. Rashdan A, Fogel EL, McHenry L Jr, et al. Improved stentcharacteristics for prophylaxis of post-ERCP pancreatitis. ClinGastroenterol Hepatol 2004;2:322–329.

60. Baron TH, Petersen BT, Mergener K, et al. Quality indicators forendoscopic retrograde cholangiopancreatography. Am J Gas-troenterol 2006;101:892–897.

61. Debenedet AT, Elmunzer BJ, McCarthy ST, et al. Intraproceduralquality in endoscopic retrograde cholangiopancreatography: a

meta-analysis. Am J Gastroenterol 2013;108:1696–1704; quiz1705.

62. Katsinelos P, Paroutoglou G, Kountouras J, et al. A comparativestudy of standard ERCP catheter and hydrophilic guide wire inthe selective cannulation of the common bile duct. Endoscopy2008;40:302–307.

63. Bailey AA, Bourke MJ, Williams SJ, et al. A prospective ran-domized trial of cannulation technique in ERCP: effects ontechnical success and post-ERCP pancreatitis. Endoscopy2008;40:296–301.

64. Tse F, Yuan Y, Moayyedi P, et al. Guidewire-assisted can-nulation of the common bile duct for the prevention of post-endoscopic retrograde cholangiopancreatography (ERCP)pancreatitis. Cochrane Database Syst Rev 2012;12:CD009662.

65. Halttunen J, Kylanpaa L. A prospective randomized study of thinversus regular-sized guide wire in wire-guided cannulation. SurgEndosc 2013;27:1662–1667.

66. Vihervaara H, Gronroos JM, Koivisto M, et al. Angled- orstraight-tipped hydrophilic guidewire in biliary cannulation: aprospective, randomized, controlled trial. Surg Endosc 2013;27:1281–1286.

67. Yoo YW, Cha SW, Lee WC, et al. Double guidewire technique vstranspancreatic precut sphincterotomy in difficult biliary can-nulation. World J Gastroenterol 2013;19:108–114.

68. Coté GA, Mullady DK, Jonnalagadda SS, et al. Use of apancreatic duct stent or guidewire facilitates bile duct accesswith low rates of precut sphincterotomy: a randomized clinicaltrial. Dig Dis Sci 2012;57:3271–3278.

69. Goff JS. Common bile duct pre-cut sphincterotomy: trans-pancreatic sphincter approach. Gastrointest Endosc 1995;41:502–505.

70. Goff JS. Long-term experience with the transpancreaticsphincter pre-cut approach to biliary sphincterotomy. Gastro-intest Endosc 1999;50:642–645.

71. Kahaleh M, Tokar J, Mullick T, et al. Prospective evaluation ofpancreatic sphincterotomy as a precut technique for biliarycannulation. Clin Gastroenterol Hepatol 2004;2:971–977.

72. Ito K, Fujita N, Noda Y, et al. Can pancreatic duct stentingprevent post-ERCP pancreatitis in patients who undergopancreatic duct guidewire placement for achieving selectivebiliary cannulation? A prospective randomized controlled trial.J Gastroenterol 2010;45:1183–1191.

73. Wang P, Zhang W, Liu F, et al. Success and complication ratesof two precut techniques, transpancreatic sphincterotomy andneedle-knife sphincterotomy for bile duct cannulation.J Gastrointest Surg 2010;14:697–704.

74. Catalano MF, Linder JD, Geenen JE. Endoscopic trans-pancreatic papillary septotomy for inaccessible obstructed bileducts: comparison with standard pre-cut papillotomy. Gastro-intest Endosc 2004;60:557–561.

75. Chan CH, Brennan FN, Zimmerman MJ, et al. Wire assistedtranspancreatic septotomy, needle knife precut or both fordifficult biliary access. J Gastroenterol Hepatol 2012;27:1293–1297.

76. Akashi R, Kiyozumi T, Jinnouchi K, et al. Pancreatic sphincterprecutting to gain selective access to the common bile duct: aseries of 172 patients. Endoscopy 2004;36:405–410.

77. Wang P, Li ZS, Liu F, et al. Risk factors for ERCP-relatedcomplications: a prospective multicenter study. Am J Gastro-enterol 2009;104:31–40.

1532 Wang et al Clinical Gastroenterology and Hepatology Vol. 14, No. 11

78. Cennamo V, Fuccio L, Zagari RM, et al. Can early precutimplementation reduce endoscopic retrogradecholangiopancreatography-related complication risk? Meta-analysis of randomized controlled trials. Endoscopy 2010;42:381–388.

79. Gong B, Hao L, Bie L, et al. Does precut technique improveselective bile duct cannulation or increase post-ERCP pancre-atitis rate? A meta-analysis of randomized controlled trials. SurgEndosc 2010;24:2670–2680.

80. Swan MP, Alexander S, Moss A, et al. Needle knife sphincter-otomy does not increase the risk of pancreatitis in patients withdifficult biliary cannulation. Clin Gastroenterol Hepatol 2013;11:430–436 e1.

81. Wilcox CM, Phadnis M, Varadarajulu S. Biliary stent placementis associated with post-ERCP pancreatitis. Gastrointest Endosc2010;72:546–550.

82. Lawrence C, Romagnuolo J, Cotton PB, et al. Post-ERCPpancreatitis rates do not differ between needle-knife andpull-type pancreatic sphincterotomy techniques: a multi-endoscopist 13-year experience. Gastrointest Endosc 2009;69:1271–1275.

83. Baron TH, Harewood GC. Endoscopic balloon dilation of thebiliary sphincter compared to endoscopic biliary sphincterotomyfor removal of common bile duct stones during ERCP: a meta-analysis of randomized, controlled trials. Am J Gastroenterol2004;99:1455–1460.

84. Disario JA, Freeman ML, Bjorkman DJ, et al. Endoscopicballoon dilation compared with sphincterotomy for extraction ofbile duct stones. Gastroenterology 2004;127:1291–1299.

85. Stefanidis G, Viazis N, Pleskow D, et al. Large balloon dilationvs. mechanical lithotripsy for the management of large bile ductstones: a prospective randomized study. Am J Gastroenterol2011;106:278–285.

86. Feng Y, Zhu H, Chen X, et al. Comparison of endoscopicpapillary large balloon dilation and endoscopic sphincterotomyfor retrieval of choledocholithiasis: a meta-analysis of random-ized controlled trials. J Gastroenterol 2012;47:655–663.

87. Liao WC, Lee CT, Chang CY, et al. Randomized trial of 1-minuteversus 5-minute endoscopic balloon dilation for extraction ofbile duct stones. Gastrointest Endosc 2010;72:1154–1162.

88. Kogure H, Tsujino T, Isayama H, et al. Short- and long-termoutcomes of endoscopic papillary large balloon dilation with orwithout sphincterotomy for removal of large bile duct stones.Scand J Gastroenterol 2014;49:121–128.

89. Chan HH, Lai KH, Lin CK, et al. Endoscopic papillary largeballoon dilation alone without sphincterotomy for the treatmentof large common bile duct stones. BMC Gastroenterol 2011;11:69.

90. Lella F, Bagnolo F, Colombo E, et al. A simple way of avoidingpost-ERCP pancreatitis. Gastrointest Endosc 2004;59:830–834.

91. Artifon EL, Sakai P, Cunha JE, et al. Guidewire cannulation re-duces risk of post-ERCP pancreatitis and facilitates bile ductcannulation. Am J Gastroenterol 2007;102:2147–2153.

92. Nambu T, Ukita T, Shigoka H, et al. Wire-guided selectivecannulation of the bile duct with a sphincterotome: a prospec-tive randomized comparative study with the standard method.Scand J Gastroenterol 2011;46:109–115.

93. Mariani A, Giussani A, Di Leo M, et al. Guidewire biliary can-nulation does not reduce post-ERCP pancreatitis comparedwith the contrast injection technique in low-risk and high-riskpatients. Gastrointest Endosc 2012;75:339–346.

94. Kawakami H, Maguchi H, Mukai T, et al. A multicenter, pro-spective, randomized study of selective bile duct cannulationperformed by multiple endoscopists: the BIDMEN study. Gas-trointest Endosc 2012;75:362–372, 72 e1.

95. Kobayashi G, Fujita N, Imaizumi K, et al. Wire-guided biliarycannulation technique does not reduce the risk of post-ERCPpancreatitis: multicenter randomized controlled trial. DigEndosc 2013;25:295–302.

Reprint requestsAddress requests for reprints to: Andrew Y. Wang, MD, AGAF, FACG, FASGE,Division of Gastroenterology and Hepatology, University of Virginia HealthSystem, Box 800708, Charlottesville, Virginia 22908. e-mail: ayw7d@virginia.edu; fax: (434) 244-7590.

Conflicts of interestThis author discloses the following: Andrew Wang receives research supportfrom Cook Medical on the topic of metal biliary stents. The remaining authorsdisclose no conflicts.

November 2016 Prevention of PEP 1532.e1

Supplementary Materials

Other Hormones

Other biologically active hormones that have beenused experimentally to prevent PEP include both calci-tonin and glucagon. Both agents were selected on thebasis of their ability to inhibit pancreatic secretion,1,2

however, neither has been proven to be effective.3,4

Overall, we do not presently advocate the use of so-matostatin, octreotide, secretin, or other regulatoryhormones for the purpose of reducing the risk of PEP.

Protease Inhibitors

The premature or inappropriate activation of tryp-sinogen has long been considered one of the incitingevents in the development of acute pancreatitis. There-fore, the administration of an antiprotease to prevent theconversion of trypsinogen to trypsin during ERCP isconceptually attractive.5 There are presently 3 low-molecular-weight antiproteases available with efficacyagainst trypsin and other serine proteases: gabexate,ulinastatin, and nafamostat.

Gabexate first was synthesized in 1977 and has beenused extensively in Japan and China for the managementof severe acute pancreatitis.6 Although previous RCTshave shown the effectiveness of gabexate in the pre-vention of PEP,7–9 other studies10–12 and subsequentdata from meta-analyses13,14 have failed to show aconsistent benefit.

Ulinastatin, a derivative of gabexate with a longerhalf-life, also showed promise in a multicenter Japanesetrial including 406 patients undergoing ERCP. In thistrial, the incidence of PEP was reduced significantly inpatients receiving 150,000 U of IV ulinastatin 10 minutesbefore ERCP (2.9% vs 7.4%; P ¼ .041).15 Despite thisinitial enthusiasm, subsequent trials with both similarand alternative dosing schedules have failed to show aconsistent benefit.16,17

Nafamostat mesilate, an alternative agent, also hasbeen shown to reduce the incidence of PEP comparedwith placebo when infused 1 hour before ERCP andcontinued afterward.18 In a comparatively large RCT thatincluded 608 patients, there was a marked reduction inPEP among patients who received nafamostat (13.0% vs4.0%; P < .0001). A subgroup analysis from thatstudy, however, showed that patients who were at highrisk for PEP did not receive a benefit (14.6% vs 5.9%;P ¼ .108).19

A meta-analysis of trials involving all 3 protease in-hibitors (along with nonsteroidal anti-inflammatorydrugs) was published in 2013. This systematic reviewincluded 5 trials of gabexate, 6 trials of ulinastatin, and 5trials of nafamostat from 1996 through 2012. This studyincluded a total of 8211 patients involved in trials ofantiproteases and suggested a reduction in PEP risk onlywith nafamostat (RR, 0.41; 95% CI, 0.28–0.59), but not

with gabexate (RR, 0.64; 95% CI, 0.36–1.13), or ulinas-tatin (RR, 0.65; 95% CI, 0.33–1.30).20

Despite these data, the use of protease inhibitors forPEP prophylaxis remains extremely limited outside ofAsia. Given the earlier-described data and the low costand relative simplicity of alternative options (eg, rectalindomethacin), we do not endorse the routine use ofprotease inhibitors for the prevention of PEP.

Other Medication Therapies

Serotonin (5-hydroxytryptamine) activates pancreaticenzyme secretion, and 5-hydroxytryptamine is associ-ated with the development and aggravation of acutepancreatitis.21 Risperidone is a 5-hydroxytryptamine 2Aantagonist that has been hypothesized to have the po-tential to reduce the risk of PEP. Risperidone has beenstudied in 2 RCTs, alone21 or in combination with uli-nastatin.22 Unfortunately, in both studies risperidonewas not found to reduce the rates of PEP significantly.

Antioxidants also have been studied to test if theymight have a role in preventing PEP. However, a meta-analysis of 3010 patients found that antioxidantsupplementation showed no beneficial effect on theincidence and the severity of PEP.23

Type of Contrast Medium

The type of contrast injected into the PD also hasbeen implicated in altering the risk of PEP, with low-osmolality, nonionic contrast agents such as iohexol be-ing preferred. However, Sherman et al24 compared ratesof PEP in patients undergoing ERCP using low-osmolality, nonionic, iohexol (672 mOsm/kg H2O,Omnipaqe 300; GE Healthcare, Wauwatosa, WI) or high-osmolality, ionic, diatrizoate sodium (1515 mOsm/kgH2O, Hypaque 50%; Amersham Health/GE Healthcare).They conducted an appropriately powered, double-blind,RCT that enrolled 690 patients but found no difference inthe rates of PEP between patients who had ERCP per-formed using iohexol (7.5%) vs diatrizoate sodium(7.2%; P > .05). These investigators concluded thatosmolality and ionicity do not appear to be major factorsin provoking pancreatitis. Presently, there are nocompelling data that suggest the type of contrastmedium affects the occurrence of PEP.

Pancreatic Duct Injection

During ERCP, the pancreas can be exposed to severalfactors (mechanical, chemical, hydrostatic, enzymatic,microbiologic, allergic, and thermal) that may contributeto PEP.25 When considering endoscopic retrograde pan-creatography, likely factors that increase the risk of PEPinclude opacification of the entire PD, acinarization(owing to overly forceful injection of contrast), andpossibly the number of PD injections.26

6. Ino Y, Arita Y, Akashi T, et al. Continuous regional arterial infu-sion therapy with gabexate mesilate for severe acute pancrea-titis. World J Gastroenterol 2008;14:6382–6387.

7. Cavallini G, Tittobello A, Frulloni L, et al. Gabexate for the pre-vention of pancreatic damage related to endoscopicretrograde cholangiopancreatography. Gabexate in digestiveendoscopy–Italian Group. N Engl J Med 1996;335:919–923.

8. Manes G, Ardizzone S, Lombardi G, et al. Efficacy of post-procedure administration of gabexate mesylate in the preventionof post-ERCP pancreatitis: a randomized, controlled, multi-center study. Gastrointest Endosc 2007;65:982–987.

9. Xiong GS, Wu SM, Zhang XW, et al. Clinical trial of gabexate inthe prophylaxis of post-endoscopic retrograde chol-angiopancreatography pancreatitis. Braz J Med Biol Res 2006;39:85–90.

10. Andriulli A, Leandro G, Federici T, et al. Prophylactic adminis-tration of somatostatin or gabexate does not prevent pancrea-titis after ERCP: an updated meta-analysis. Gastrointest Endosc2007;65:624–632.

11. Andriulli A, Solmi L, Loperfido S, et al. Prophylaxis of ERCP-related pancreatitis: a randomized, controlled trial of somato-statin and gabexate mesylate. Clin Gastroenterol Hepatol 2004;2:713–718.

12. Masci E, Cavallini G, Mariani A, et al. Comparison of two dosingregimens of gabexate in the prophylaxis of post-ERCPpancreatitis. Am J Gastroenterol 2003;98:2182–2186.

13. Andriulli A, Leandro G, Clemente R, et al. Meta-analysis of so-matostatin, octreotide and gabexate mesilate in the therapy ofacute pancreatitis. Aliment Pharmacol Ther 1998;12:237–245.

14. Seta T, Noguchi Y. Protease inhibitors for preventing compli-cations associated with ERCP: an updated meta-analysis.Gastrointest Endosc 2011;73:700–706 e1–2.

15. Tsujino T, Komatsu Y, Isayama H, et al. Ulinastatin for pancre-atitis after endoscopic retrograde cholangiopancreatography: arandomized, controlled trial. Clin Gastroenterol Hepatol 2005;3:376–383.

16. Ueki T, Otani K, Kawamoto K, et al. Comparison between uli-nastatin and gabexate mesylate for the prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis:a prospective, randomized trial. J Gastroenterol 2007;42:161–167.

17. Fujishiro K, Heaney CA. Justice at work, job stress, andemployee health. Health Educ Behav 2009;36:487–504.

18. Yoo KS, Huh KR, Kim YJ, et al. Nafamostat mesilate for pre-

1532.e2 Wang et al Clinical Gastroenterology and Hepatology Vol. 14, No. 11

Freeman et al27 found in a prospective multicenterstudy on the risk factors of PEP that 1 or more contrastinjections into the PD was a risk factor on adjustedmultivariate analysis for PEP, with an OR of 2.7. In aprospective RCT comparing cannulation methods, Baileyet al28 found on multivariate analysis that completefilling of the PD with contrast agent (OR, 3.5; P ¼ .02)was associated independently with PEP.

In general, complete filling of the PD should beavoided unless indicated for a specific diagnostic ortherapeutic reason, so as to reduce the risk of PEP.Furthermore, minimizing the number of PD injectionsand the amount of contrast injected should help reducethe risk of PEP.

Deep-Enteroscopy–Assisted EndoscopicRetrograde Cholangiopancreatography

Single-balloon, double-balloon, and spiral entero-scopy platforms have been available for several yearsand enable ERCP in patients with altered gastroduo-denal anatomy.29 The rate of pancreatitis after balloon-assisted enteroscopy without ERCP is less than 1%.30,31

In patients with an intact major papilla and withoutpancreas divisum, the rate of PEP afterballoon–enteroscopy assisted ERCP ranges from 2.3% to12.5%,32–35 which overall appears to be higher than forconventional ERCP. This increased risk of PEP istypically due to difficulty in selectively accessing theduct of interest using a forward-viewing enteroscopewithout an elevator.

In patients with Roux-en-Y gastric bypass anatomywho have particularly long Roux and afferent bil-iopancreatic limbs,36 in those in whom complex orrepeated pancreaticobiliary interventions are anticipated,or in patients who might be at increased risk for PEP,consideration should be given to performing ERCP usinga duodenoscope passed through a gastrostomy placed inthe gastric remnant. Typically, transgastric access to theexcluded stomach is accomplished via laparoscopy, butendoscopic techniques have been described.37

Supplementary References1. Odes HS, Barbezat GO, Clain JE, et al. The effect of calcitonin

on secretin-stimulated pancreatic secretion in man. S Afr Med J1978;53:201–203.

2. Schapiro H, Ludewig RM. The effect of glucagon on theexocrine pancreas. A review. Am J Gastroenterol 1978;70:274–281.

3. Silvis SE, Vennes JA. The role of glucagon in endoscopiccholangiopancreatography. Gastrointest Endosc 1975;21:162–163.

4. Brandes JW, Scheffer B, Lorenz-Meyer H, et al. ERCP: com-plications and prophylaxis a controlled study. Endoscopy 1981;13:27–30.

5. Testoni PA. Pharmacological prevention of post-ERCP pancre-atitis: the facts and the fiction. JOP 2004;5:171–178.

vention of post-endoscopic retrograde cholangiopancreatog-raphy pancreatitis: a prospective, randomized, double-blind,controlled trial. Pancreas 2011;40:181–186.

19. Park KT, Kang DH, Choi CW, et al. Is high-dose nafamostatmesilate effective for the prevention of post-ERCP pancrea-titis, especially in high-risk patients? Pancreas 2011;40:1215–1219.

20. Yuhara H, Ogawa M, Kawaguchi Y, et al. Pharmacologicprophylaxis of post-endoscopic retrograde cholangiopancreatog-raphy pancreatitis: protease inhibitors and NSAIDs in a meta-analysis. J Gastroenterol 2014;49:388–399.

21. Uchino R, Isayama H, Tsujino T, et al. Results of the Tokyo trialof prevention of post-ERCP pancreatitis with risperidone-2: amulticenter, randomized, placebo-controlled, double-blind clin-ical trial. Gastrointest Endosc 2013;78:842–850.

22. Tsujino T, Isayama H, Nakai Y, et al. The results of the Tokyo trialof prevention of post-ERCP pancreatitis with risperidone

(Tokyo P3R): a multicenter, randomized, phase II, non-placebo-controlled trial. J Gastroenterol 2013;48:982–988.

23. Gu WJ, Wei CY, Yin RX. Antioxidant supplementation for theprevention of post-endoscopic retrograde chol-angiopancreatography pancreatitis: a meta-analysis of ran-domized controlled trials. Nutr J 2013;12:23.

24. Sherman S, Hawes RH, Rathgaber SW, et al. Post-ERCPpancreatitis: randomized, prospective study comparing a low-and high-osmolality contrast agent. Gastrointest Endosc 1994;40:422–427.

25. Sherman S, Lehman GA. ERCP- and endoscopicsphincterotomy-induced pancreatitis. Pancreas 1991;6:350–367.

26. Freeman ML. Pancreatic stents for prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis.Clin Gastroenterol Hepatol 2007;5:1354–1365.

27. Freeman ML, DiSario JA, Nelson DB, et al. Risk factors for post-ERCP pancreatitis: a prospective, multicenter study. Gastro-intest Endosc 2001;54:425–434.

28. Bailey AA, Bourke MJ, Williams SJ, et al. A prospective ran-domized trial of cannulation technique in ERCP: effects ontechnical success and post-ERCP pancreatitis. Endoscopy2008;40:296–301.

29. Cosgrove ND, Wang AY. Endoscopic approaches to biliaryintervention in patients with surgically altered gastroduodenalanatomy. World J Surg Proc 2014;4:1–8.

30. Aktas H, de Ridder L, Haringsma J, et al. Complications ofsingle-balloon enteroscopy: a prospective evaluation of 166procedures. Endoscopy 2010;42:365–368.

31. Kopacova M, Tacheci I, Rejchrt S, et al. Double balloonenteroscopy and acute pancreatitis. World J Gastroenterol2010;16:2331–2340.

32. Raithel M, Dormann H, Naegel A, et al. Double-balloon-enteroscopy-based endoscopic retrograde chol-angiopancreatography in post-surgical patients. World JGastroenterol 2011;17:2302–2314.

33. Siddiqui AA, Chaaya A, Shelton C, et al. Utility of the shortdouble-balloon enteroscope to perform pancreaticobiliary in-terventions in patients with surgically altered anatomy in a USmulticenter study. Dig Dis Sci 2013;58:858–864.

34. Wang AY, Sauer BG, Behm BW, et al. Single-balloon entero-scopy effectively enables diagnostic and therapeutic retrogradecholangiography in patients with surgically altered anatomy.Gastrointest Endosc 2010;71:641–649.

35. Yamauchi H, Kida M, Okuwaki K, et al. Short-type single balloonenteroscope for endoscopic retrograde cholangiopancreatog-raphy with altered gastrointestinal anatomy. World J Gastro-enterol 2013;19:1728–1735.

36. Schreiner MA, Chang L, Gluck M, et al. Laparoscopy-assistedversus balloon enteroscopy-assisted ERCP in bariatric post-Roux-en-Y gastric bypass patients. Gastrointest Endosc 2012;75:748–756.

37. Law R, Wong Kee Song LM, Petersen BT, et al. Single-session ERCP in patients with previous Roux-en-Y gastricbypass using percutaneous-assisted transprostheticendoscopic therapy: a case series. Endoscopy 2013;45:671–675.

November 2016 Prevention of PEP 1532.e3