Abbreviations
CT:
computed tomography
DGW:
double guidewire
EPBD:
endoscopic papillary balloon dilation
ESGE:
European Society of Gastrointestinal Endoscopy
ERCP:
endoscopic retrograde cholangiopancreatography
NSAID:
nonsteroidal anti-inflammatory drug
PEP:
post-ERCP pancreatitis
PGW:
pancreatic guidewire
RCT:
randomized controlled trial
SOD:
sphincter of Oddi dysfunction
SOM:
sphincter of Oddi manometry
ULN:
upper limit of normal
1. Introduction
The Guideline on prophylaxis of post-ERCP pancreatitis (PEP) issued by the European
Society of Gastrointestinal Endoscopy (ESGE) in 2010 aimed to provide a qualified
basis for gastrointestinal endoscopists to take measures to minimize the incidence
and severity of PEP [1]. Shortly before the publication of the ESGE Guideline, nonsteroidal anti-inflammatory
drugs (NSAIDs) were reportedly rarely used in clinical practice for prevention of
PEP (16 % of respondents to a survey performed in June 2009), and this was attributed
by survey participants to the lack of sufficient data [2]. Similarly, in an Austrian nationwide ERCP survey, PEP prophylaxis was administered
in only 4.0 % of patients in 2010 and in 7.0 % of patients in 2011 [3]. A more recent survey from the UK found that the proportion of endoscopists using
NSAIDs had increased to 34.6 % in 2012 [4]. Obviously, prophylactic measures against PEP are still greatly underused in daily
clinical practice. At the same time, PEP is still the most frequent and severe complication
encountered following ERCP.
New evidence that has become available since the publication of the ESGE Guideline
in 2010 is discussed in the present update and new recommendations are issued.
2. Methods
ESGE commissioned this update of the Guideline on prevention of PEP. Methods similar
to those used in the previous Guideline were applied [1]. A literature search of PubMed/MEDLINE, a search using the Cochrane Library, Embase,
and the internet was performed to identify publications since 2009 on this topic.
The search focused on fully published prospective studies, particularly randomized
controlled trials (RCTs) and meta-analyses. Retrospective analyses and pilot studies
were also included if they addressed topics not covered in the prospective studies.
Thereafter, the commissioned authors met once and subsequently developed the updated
Guideline. The Guideline committee chairs (C.K., J.M.D.) worked with the subgroup
leaders (C.K., T.M., A.A., T.M., P.T., T.B., J.M.D.) who developed draft proposals
that were distributed and reviewed electronically. In May 2014, a draft prepared by
C.K. and J.M.D. was sent to all group members. After agreement on a final version,
the manuscript was sent to all individual ESGE members and individual ESGE member
societies, and was reviewed by two experts selected by the ESGE Governing Board. After
incorporation of comments, the manuscript was then sent to the journal Endoscopy for publication. The final wording of the Guideline document was agreed by all commissioned
authors.
3. Definitions
None.
Two definitions of PEP may currently be used, neither of these being ideal in the
setting of PEP: the consensus definition and grading of severity of PEP according
to Cotton et al. and the more recent revised Atlanta international consensus definition
and classification of acute pancreatitis.
Background:
The consensus definition and grading of severity of PEP developed by Cotton et al.
has been used for > 20 years [5]. It has allowed standardized reporting of the incidence and severity of PEP. Post-ERCP
pancreatitis was originally defined as “clinical pancreatitis with amylase at least
three times the upper limit of normal at more than 24 hours after the procedure, requiring
hospital admission or a prolongation of planned admission.” Various modifications
were introduced by Freeman et al. who proposed using lipase as a possible alternative
to amylase and defining clinical pancreatitis as “new or worsened abdominal pain,”
hence taking into account patients who undergo ERCP in the setting of acute pancreatitis
or of a flare of chronic pancreatitis [6]. The grading system for the severity of PEP by the consensus definition is not ideal
as it is mainly based on the length of hospitalization.
New information since 2009:
The Atlanta classification of acute pancreatitis was updated in 2012 [7]. Although this classification provides clear definitions to classify acute pancreatitis,
its limitations include the fact that it was not primarily developed to define PEP.
Also, the benefit of a contrast-enhanced computed tomography (CT) scan has not been
demonstrated in the setting of PEP (contrast-enhanced CT scan is required if abdominal
pain suggests strongly that acute pancreatitis is present, but the serum amylase and/or
lipase activity is less than three times the upper limit of normal [ULN]). According
to this classification, the diagnosis of PEP requires two of the three following criteria:
(i) abdominal pain consistent with acute pancreatitis (acute onset of a persistent,
severe, epigastric pain often radiating to the back); (ii) serum lipase or amylase
activity at least three times greater than the ULN; and (iii) characteristic findings
of acute pancreatitis on contrast-enhanced CT and, less commonly, magnetic resonance
imaging or transabdominal ultrasonography. This classification defines three degrees
of severity based on the presence or absence of organ failure (plus its duration)
and of local or systemic complications.
A prospective study has shown the two definitions presented above to be poorly correlated
[8].
The Pancreatitis Across Nations Clinical Research and Education Alliance (PANCREA)
has defined four degrees of severity for pancreatitis, based on the presence or absence
of complications, both local (necrosis of the pancreas and/or peripancreatic tissue)
and systemic (cardiovascular, renal, or respiratory organ failure) [9].
4. Incidence, risk factors, and severity of PEP
4. Incidence, risk factors, and severity of PEP
4.1. Incidence
Pancreatitis is the most frequent complication after ERCP with an incidence of 3.5 %
in unselected patients; it is of mild or moderate severity in approximately 90 % of
cases.
No changes.
Background:
Data on incidence rate and severity of PEP were mainly based on a systematic review
of 21 prospective studies involving more than 16 000 patients [10]. Post-ERCP pancreatitis was found to be the most frequent complication following
ERCP, with an incidence of 3.47 % (95 % confidence interval [95 %CI] 3.19 % – 3.75 %).
Based upon data from studies that have included unselected patients, PEP is mild,
moderate, and severe in 45 %, 44 %, and 11 % of cases, respectively. Death occurs
in 3 % of cases of PEP (95 %CI 1.65 % – 4.51 %) [10].
New information since 2009:
Few new data have become available and these have yielded similar results. In an ERCP
benchmarking program, PEP was reported to occur in 4.2 % of 13 513 unselected procedures
[3]. In a retrospective study of 886 procedures, 39 patients (4.4 %) were diagnosed
with pancreatitis, of mild moderate, and severe type in 69 %, 23 %, and 8 %, respectively
[11].
4.2. Risk stratification
4.2.1 Hospital and endoscopist volume for ERCP
There is no evidence that hospital ERCP volume has an influence on the incidence of
PEP; data about a potential relationship between PEP incidence and endoscopist case
volume are conflicting. Low annual case volumes, of endoscopists and centers, are
associated with higher ERCP failure rates (Evidence level 2 + ).
No changes.
Background:
Factors that may affect the outcome of ERCP that are specifically related to hospital
procedure volume include availability of equipment and adequacy of anesthesia, endoscopic
and radiologic support, and nursing assistance. The number of ERCPs performed in many
centers is not as high as commonly believed: in three large (regional or national)
studies, the median annual number of ERCPs was between 49 and 235 [12]
[13]
[14]. In one large study, the median annual number of ERCPs per endoscopist was 111 and
40 % of endoscopists performed fewer than 50 sphincterotomies/year [15].
Multivariate analyses from two prospective audits performed in England and Italy (66
and 9 centers, respectively) found no significant association between annual hospital
volume of ERCPs and incidence of PEP [16]
[17].
New information since 2009:
A prospective Swedish study of 12 718 procedures showed no significant difference
among PEP rates in centers with low (< 100 ERCPs/year), medium, and high ( > 500 ERCPs/year)
volumes [18]. In a prospective multicenter study that included 3635 ERCPs in 11 high volume ( > 200
ERCPs/year) and 10 low volume centers (median of 275 and 45 ERCPs/year, respectively),
there was no significant difference in the incidence of PEP (3.9 % vs. 3.1 %) [19]. However, these results were confounded because a higher proportion of patients
at high risk of PEP was treated in high volume centers. In this study, the PEP rates
did not differ significantly between expert and nonexpert operators (3.8 % vs. 5.5 %,
respectively; P = 0.34).
4.2.2. Patient- and procedure-related risk factors for PEP
Independent patient-related and procedure-related risk factors for PEP are listed
in [Table 1]. Risk factors synergistically increase the risk of PEP (Evidence level 1 + ).
Table 1
Independent risk factors for post-endoscopic retrograde cholangiopancreatography (post-ERCP)
pancreatitis (PEP).[1]
|
Adjusted odds ratios (95 % confidence intervals in parentheses except where indicated
otherwise)
|
Pooled incidence of PEP in patients with vs. those without risk factor
|
|
Patient-related risk factors
|
|
|
|
Definite risk factors
|
|
|
|
Suspected sphincter of Oddi dysfunction (SOD)
|
1.91 (1.37 – 2.65)
|
8.6 % vs. 2.5 %
|
|
Female gender
|
3.5 (1.1 – 10.6)
|
4.0 % vs. 2.1 %*
|
|
Previous pancreatitis
|
2.46 (1.93 – 3.12)
|
6.7 % vs. 3.8 %
|
|
Likely risk factors
|
|
|
|
Previous PEP
|
8.7 (3.2 – 23.86)
|
30 % vs. 3.5 %
|
|
Younger age
|
Range of odds ratios: 1.09 – 2.87
|
6.2 % vs. 2.6 %
|
|
Nondilated extrahepatic bile ducts
|
|
3.8 % vs. 2.3 %
|
|
Absence of chronic pancreatitis
|
1.87 (1.00 – 3.48)
|
4.0 % vs. 3.1 %
|
|
Normal serum bilirubin
|
1.89 (1.22 – 2.93)
|
4.15 % vs. 1.43 %
|
|
Procedure-related risk factors
|
|
|
|
Definite risk factors
|
|
|
|
Cannulation attempts duration > 10 minutes[2]
|
1.76 (1.13 – 2.74)
|
3.8 % vs. 10.8 %
|
|
Pancreatic guidewire passages > 1
|
2.77 (1.79 – 4.30)
|
2.9 % vs. 9.5 %
|
|
Pancreatic injection
|
2.2 (1.60 – 3.01)
|
3.3 % vs. 1.7 %
|
|
Likely risk factors
|
|
|
|
Precut sphincterotomy[3]
|
2.3 (1.4 – 3.7)
|
5.3 % vs. 3.1 %
|
|
Pancreatic sphincterotomy
|
3.07 (1.64 – 5.75)
|
2.6 % vs. 2.3 %
|
|
Biliary balloon sphincter dilation
|
4.51 (1.51 – 13.46)
|
9.3 % vs. 2.6 %
|
|
Failure to clear bile duct stones
|
3.35 (1.33 – 9.10)
|
1.7 % vs. 1.6 %
|
|
Intraductal ultrasound (IDUS)[4]
|
2.41 (1.33 – 4.39)
|
8.37 % vs. 2.76 %
|
1 For definite or likely risk factors, adjusted odds ratios are reproduced either from
Masci et al. [20] or from included studies that identified the characteristic as an
independent risk factor. Pooled incidences were calculated using figures available
in all of the included studies that provided sufficient data for calculation [6, 16, 17, 19, 21, 22, 25 – 27, 115, 116, 177].
(See text for details about included studies)
2 Cannulation attempts of duration > 5 minutes may already increase the incidence of
PEP as shown by Halttunen et al. (11.7 % vs. 2.7 % for cannulation attempts ≥ 5 minutes
vs. < 5 minutes, respectively) [115].
3 Evidence is growing that precut sphincterotomy is not a definite risk factor for
PEP by itself, the increased risk of PEP being related to cannulation efforts that
preceded precut [28].
4 Based on one retrospective study [29].
Risk factors for PEP, in particular those related to the procedure (cannulation attempts > 10
minutes and pancreatic guidewire passages > 1) have been updated in [Table 1]. Risk factors synergistically increase the risk of PEP.
Background:
Independent risk factors for PEP were presented in a table based on data from a meta-analysis
[20] plus those from five prospective, multicenter studies that analyzed potential risk
factors for PEP using multivariate analysis [6]
[16]
[17]
[21]
[22]. The list was not exhaustive because not all potential risk factors had been analyzed.
For example, ampullectomy is generally considered to be a definitive risk factor for
PEP on the basis of several small prospective studies [23]
[24].
As risk factors for PEP were shown to be independent by multivariate analysis, they
might have a cumulative effect. Freeman et al. calculated the adjusted odds ratio
(OR) for various combinations of risk factors by using data prospectively collected
from about 2000 ERCPs: the highest risk of PEP (42 %) was found for female patients
with a normal serum bilirubin level, suspected sphincter of Oddi dysfunction (SOD),
and difficult biliary cannulation [21].
New information since 2009:
Suspected or known dysfunction of the sphincter of Oddi, female gender, younger age,
and previous history of pancreatitis are well-known independent risk factors for PEP.
A recent prospective study confirmed these findings [25]. In a large ERCP series that included 11 497 procedures, SOD was confirmed as an
independent risk factor [26]. Recently, in a Swedish case – control study of 12 718 ERCP procedures, independent
risk factors were young age, female gender, prolonged procedure time, and elective
ERCP; whereas rendezvous procedures reduced the risk of PEP [18]. Another prospective multicenter study confirmed independent procedure- and patient-related
risk factors for PEP (> 10 attempts to cannulate the papilla of Vater, OR 14.9; previous
PEP, OR 8.7; precut, OR 3.1; pancreatic duct cannulation, OR 2.1) [19]. Some of these predictors were confirmed in two other studies [25]
[27]. However, needle-knife sphincterotomy was found not to be an independent risk factor
for PEP [3]
[25]
[28].
The role of intraductal ultrasound as a risk factor for PEP remains unclear but this
factor was identified by multivariate analysis (OR 2.41, P = 0.004) in a retrospective analysis of 2364 ERCP procedures [29].
No new data have become available regarding the synergistic effect of risk factors
for PEP.
4.3. Prediction of PEP
Serum amylase values less than 1.5 times the ULN, obtained at 2 – 4 hours post-ERCP,
almost exclude PEP; values more than 3 or 5 times the ULN at 4 – 6 hours post-ERCP
have increasing positive predictive values for PEP (Evidence level 2 + ). It is recommended
that serum amylase be determined in patients to be discharged on the day of ERCP;
patients with amylase values less than 1.5 times the ULN can be discharged without
concern about risk of PEP (Recommendation grade B).
Serum amylase or lipase values less than 1.5 and 4 times the ULN, respectively, obtained
at 2 – 4 hours post-ERCP have a very high negative predictive value for PEP (Evidence
level 2 +). ESGE suggests testing serum amylase or lipase 2 – 6 hours after ERCP in
patients presenting with pain and who are to be discharged on the day of ERCP; patients
with amylase or lipase values less than 1.5 and 4 times the ULN, respectively, can
be discharged without concern about risk of PEP (Recommendation grade B).
Background:
The recommendations were based on five studies that reported similar predictive values
based on serum amylase levels obtained 2 to 6 hours following ERCP. In one study of
PEP [30], lipase values at a cutoff of 4 times the ULN had a negative and positive predictive
value for PEP of 99 % and 15 %, respectively.
New information since 2009:
Two studies confirmed previous findings. A prospective study from Brazil that included
300 patients showed that serum hyperamylasemia < 1.5 times the ULN at 4 hours and
< 2 times the ULN at 12 hours had a negative predictive value of 94 % for the development
of PEP [8]. Serum hyperamylasemia following ERCP had a poor positive predictive value for PEP.
A retrospective study investigated, in addition to the 4-hour post-ERCP serum amylase
level, the impact of having a pancreatogram in predicting PEP among 886 ERCPs [11]; the negative predictive value of serum amylase < 2.5 times ULN for moderate or
severe PEP was 99.2 % and 100 %, in patients who, respectively, did and did not have
a pancreatogram.
5. Pharmacologic agents available for PEP prophylaxis
5. Pharmacologic agents available for PEP prophylaxis
5.1. Introduction
Post-ERCP pancreatitis appears unavoidable even in the hands of expert endoscopists.
Consequently, attempts to reduce the rate of this complication by pharmacological
intervention should be pursued. While a few medications have proven effective for
preventing PEP, we acknowledge that multiple factors in addition to efficacy influence
the decision to issue a clinical recommendation. In particular, the magnitude of benefit,
as expressed by the number needed to treat (NNT), the robustness and consistency of
supporting RCTs, the safety profile of the medication, its ease of administration,
availability, and cost were considered in issuing these recommendations.
5.2. Drugs with proven efficacy
5.2.1. Nonsteroidal anti-inflammatory drugs (NSAIDs)
NSAIDs reduce the incidence of PEP; effective PEP prophylaxis has only been demonstrated
using 100 mg of diclofenac or indomethacin administered rectally (Evidence level 1 ++).
Routine rectal administration of 100 mg of diclofenac or indomethacin immediately
before or after ERCP is recommended (Recommendation grade A).
NSAIDs reduce the incidence of PEP in patients at high as well as low risk for PEP;
effective PEP prophylaxis has only been demonstrated using diclofenac or indomethacin
administered rectally (Evidence level 1 ++ ). ESGE recommends routine rectal administration
of 100 mg of diclofenac or indomethacin immediately before or after ERCP in all patients
without contraindication (Recommendation grade A).
Background:
Three different meta-analyses pooled data from four RCTs that compared rectally administered
diclofenac or indomethacin at a dose of 100 mg vs. placebo [31]
[32]
[33]. Two RCTs evaluated the effect of rectal administration of 100 mg diclofenac immediately
after the procedure, while the other two evaluated rectal administration of 100 mg
indomethacin immediately before the procedure. Both schedules showed similar results.
Patients who were considered to be at high risk for PEP were included in two studies.
Overall, PEP occurred in 4.4 % patients in the treatment groups vs. 12.5 % patients
in the placebo groups with an estimated pooled relative risk (RR) of 0.36 (95 %CI
0.22 – 0.60), and an NNT to prevent one episode of PEP of 15. The administration of
NSAIDs was associated with a similar decrease in the incidence of PEP regardless of
risk [33]. No adverse events attributable to NSAIDs were reported.
New information since 2009:
Despite previous meta-analytical results, NSAIDs were not commonly used by endoscopists
in clinical practice: only 16 % of respondents to a survey performed in June 2009,
shortly before publication of the ESGE Guideline, used NSAIDs for PEP prophylaxis
and this low figure was attributed to the lack of sufficient data [2]. A more recent survey (June 2012) found that the proportion of endoscopists using
NSAIDs has increased to 35 % [4].
In a multicenter RCT [34], patients at high risk for PEP received either a single dose of rectal indomethacin
or of placebo immediately after ERCP. A total of 602 patients were enrolled, including
82 % with a clinical suspicion of SOD and 82 % who received prophylactic pancreatic
stenting based on the high risk of PEP. Post-ERCP pancreatitis developed in 9.2 %
vs. 16.9 % of patients in the indomethacin vs. placebo group, respectively (P = 0.005). Of note, moderate-to-severe PEP was less frequent in the indomethacin vs.
placebo group (4.4 % vs. 8.8 %, respectively; P = 0.03). The benefit of 100 mg rectal indomethacin was also confirmed in an RCT that
included 228 patients but the results were not significantly different from placebo,
likely owing to the small sample size [35]. A lower dosage of diclofenac, either 50 mg or 25 mg in patients weighing ≥ 50 kg
or < 50 kg, respectively, was compared with placebo in a small RCT from Japan: the
PEP incidence was 3.9 % vs. 18.9 % in the diclofenac vs. the control group, respectively
(P = 0.017) [36]. A small RCT of 80 patients demonstrated a nonsignificant trend toward benefit associated
with the combination of intramuscular diclofenac and intensive intravenous fluid administration
[37]. On the other hand, it seems that selective cyclo-oxygenase-2 inhibitor (coxib)
administration is ineffective, as shown by an RCT that included 371 patients where
intravenous valdecoxib or glyceryl trinitrate (GTN) patch were compared with placebo:
the incidence of PEP was similar in all treatment groups (10 %, P = 0.986) [38].
Six meta-analyses published between 2009 and 2014 ([Table 2]) compared NSAIDs vs. placebo administration for prevention of PEP, and all of them
concordantly showed the benefit of NSAIDs in preventing either mild or moderate/severe
PEP. The NNT figures, reported in the majority of these studies, have varied from
11 to 34 [31]
[39]
[40]
[41]
[42]
[43].
Table 2
Meta-analyses published in 2009 or later that assessed the effect of NSAIDs on post-endoscopic
retrograde cholangiopancreatography (post-ERCP) pancreatitis (PEP).[1]
|
First author,
year
|
Studies, n
(Patients, n)
|
PEP incidence
|
OR
(95 %CI)
|
NNT
|
Comments
|
|
NSAIDs
|
Placebo
|
|
Dai,
2009 [31]
|
6
(1300)
|
8.9 %
|
16.8 %
|
0.46
(0.32 – 0.65)
|
n.a.
|
Inclusion of placebo-controlled RCTs only (diclofenac or indomethacin administered
orally or rectally). Different ERCP techniques and criteria for pancreatitis, low
number of high risk patients
|
|
Ding,
2012 [39]
|
10
(2269)
|
8.0 %
|
13.9 %
|
0.57
(0.38 – 0.86)[1]
|
17
|
Different NSAIDs (diclofenac, indomethacin, valdecoxib) and different administration
routes (oral, rectal, intraduodenal, intramuscular and intravenous) and ERCP techniques;
moderate and severe PEP also reduced.
Moderate/severe PEP: RR 0.46 (0.28 – 0.38); NNT, 34
|
|
Yaghoobi,
2013 [40]
|
4
(1470)
|
5.1 %
|
10.3 %
|
0.49
(0.34 – 0.71)
|
20
|
Only rectal NSAIDs;
significant decrease of PEP incidence in both high risk and low risk patients, moderate
and severe PEP also reduced.
Moderate/severe PEP: RR 0.45 (0.24 – 0.83)
|
|
Sun,
2013 [41]
|
7
(1846)
|
6.4 %
|
16.0 %
|
0.45
(0.34 – 0.61)[1]
|
n.a.
|
Only rectal NSAIDs;
both diclofenac and indomethacin effective, significant in both high risk and average-risk
cohorts, moderate and severe PEP reduced.
Mild PEP: RR 0.54 (0.35 – 0.83)
Moderate/severe PEP: RR 0.39 (0.22 – 0.70)
|
|
Yuhara,
2014 [42]
|
9
(1981)
|
7.8 %
|
16.0 %
|
0.55
(0.43 – 0.72)*
|
n.a.
|
Inclusion of RCTs testing rectal, oral, intramuscular and intravenous NSAIDs (diclofenac,
indomethacin). No significant effect on PEP with intramuscular administration.
|
|
Sethi,
2014 [43]
|
7
(2133)
|
6.6 %
|
15.1 %
|
0.44
(0.34 – 0.57)[1]
|
11
|
Only rectal NSAIDs;
no difference between diclofenac and indomethacin, no difference between timing (pre-
and post-ERCP).
Relative risk in low risk and high risk patients, respectively: 0.42 (0.26 – 0.66)
and 0.45 (0.31 – 0.65).
|
NSAIDs, nonsteroidal anti-inflammatory drugs; OR, odds ratio; CI, confidence interval;
NNT, number needed to treat; n.a., not available; RCT, randomized controlled trial;
RR, relative risk.
1 Risk reduction.
In a post hoc analysis of an RCT of NSAIDs vs. placebo for PEP prophylaxis, administration
of rectal NSAIDs alone was more effective and less costly than prophylactic pancreatic
stent placement alone or combined with rectal NSAIDs [44]. However, the authors cautioned that these findings should not change current clinical
practice because their post hoc observational study did not produce the same quality
of evidence as their RCT, which supported the use of indomethacin in addition to prophylactic
pancreatic stenting in high risk cases [34].
5.3. Possibly effective drugs
5.3.1. Somatostatin and octreotide
Based on an ad hoc meta-analysis of results from 10 high quality RCTs, somatostatin
proved to be ineffective in preventing PEP (Evidence level 1 ++). We do not recommend
universal administration of prophylactic somatostatin in average-risk patients undergoing
ERCP (Recommendation grade A). Administration of somatostatin might be more efficacious
using specific dose schedules, but caution is needed when interpreting the results
of subgroup analyses as they often exaggerate differences between treatments in RCTs.
Octreotide administration did not affect the overall incidence of PEP when data from
eight high quality trials were pooled (Evidence level 1 ++). Prophylaxis with octreotide
is not recommended (Recommendation grade A). In future studies the efficacy of prophylactic
administration of octreotide should be evaluated using a dose greater than or equal
to 0.5 mg.
Some meta-analytical results seem to support the benefit of somatostatin and octreotide
for averting PEP but their clinical use cannot be recommended except in well selected
cases, owing to discordant data from different routes or dosages and the excessively
high NNT values (Recommendation grade A).
Background:
Three meta-analyses assessed the effect of somatostatin/octreotide for PEP prophylaxis
before the 2010 ESGE Guideline publication [1]; an updated meta-analysis of high quality RCTs (Jadad score > 3) was performed for
writing the ESGE Guideline. Respectively, 10 and 8 RCTs were included for assessing
somatostatin and octreotide. Overall, the updated meta-analysis found no significant
reduction of PEP incidence with somatostatin vs. placebo (OR 0.57; 95 %CI 0.32 – 1.03)
or with octreotide vs. placebo (OR 0.73; 95 % CI, 0.41 – 1.30). However, subgroup
analyses suggested that some dosages or administration schedules might be associated
with a protective effect against PEP.
New information since 2009:
A meta-analysis that compared octreotide vs. placebo (18 RCTs, 3171 patients) found
no significant difference in PEP incidence (OR 0.77; 95 %CI 0.56 – 1.05) [45]. However, at post hoc subgroup analysis, dosage seemed to have an impact: when the
compound was given at a dose ≥ 0.5 mg (6 RCTs; 1470 patients) the OR for developing
PEP dropped to 0.45 (95 %CI 0.28 – 0.73; NNT = 25), whereas the agent proved ineffective
at a dosage < 0.5 mg. Other subgroup analyses of octreotide administration that assessed
the route (intravenous vs. subcutaneous) and the schedule (before or after ERCP) were
inconclusive.
A second meta-analysis assessed the effect of somatostatin (10 RCTs) and of octreotide
(7 RCTs) in a total of 3818 patients [46]. Overall, somatostatin reduced the risk of PEP (RR 0.52; 95 %CI 0.30 – 0.90), while
octreotide was not effective (RR 0.86; 95 %CI 0.45 – 1.63). Subgroup analyses suggested
that higher doses of somatostatin (3 mg given as an infusion over 12 hours) or lower
doses (250 µg) given as bolus injection may be most efficacious, especially for the
subgroup of patients at relatively higher risk for PEP, i. e., those undergoing pancreatic
duct injection (OR 0.35; 95 %CI 0.15 – 0.82) and biliary sphincterotomy (OR 0.33;
95 %CI 0.16 – 0.70). With regard to octreotide, subgroup analysis showed a protective
effect when administered at high dose (OR 0.42; 95 %CI 0.20 – 0.90).
More recently, two RCTs compared the effect of somatostatin vs. placebo for PEP prophylaxis.
In one RCT, the efficacy of somatostatin alone could not be assessed because the treatment
group received somatostatin plus diclofenac [47]; the other RCT found that high dose somatostatin was associated with a reduction
of post-ERCP hyperamylasemia, but not incidence of PEP [48].
5.3.2. Protease inhibitors
Prophylaxis with gabexate or ulinastatin does not reduce the incidence of PEP (Evidence
1 ++). Neither drug is recommended for prophylaxis of PEP (Recommendation grade A).
Some meta-analytical results seem to support the benefit of gabexate and ulinastatin
at high doses for averting PEP but their clinical use cannot be recommended owing
to the discordance of the data. A novel protease inhibitor, nafamostat, is likely
effective for preventing PEP in patients at low risk of PEP but not in high risk patients.
Background:
Gabexate for PEP prophylaxis has been evaluated in six high quality RCTs [49]
[50]
[51]
[52]
[53]
[54]; when the results were pooled, no significant difference was found between the control
and treatment groups. Ulinastatin for PEP prophylaxis has been compared with placebo
(two RCTs) and with gabexate (two RCTs), with contradictory results [55]
[56]
[57]
[58]. Nafamostat is a novel protease inhibitor that inhibits trypsin, a proteolytic enzyme
considered to play an initial role in the pathogenesis of pancreatitis; compared with
gabexate its half-life is 20 times longer and its potency 10 to 100 times greater.
New information since 2009:
Four meta-analyses assessed the efficacy of protease inhibitors for PEP prophylaxis
[42]
[59]
[60]
[61]. Subgroup analysis of 8 high quality (Jadad score > 3) RCTs on gabexate administration
showed that the agent was not associated with a decreased risk of PEP (RR 0.64; 95 %CI
0.36 – 1.13) and there was great heterogeneity among the studies. Subgroup analysis
of 6 ulinastatin RCTs showed the agent was not associated with a decreased risk of
PEP in either high quality studies (RR 0.65; 95 %CI 0.33 – 1.30) or low quality studies
(RR 0.75; 95 %CI 0.49 – 1.16).
The question of drug dosage was addressed in one meta-analysis: when ulinastatin was
administered at sufficient doses (≥ 150 000 units), it averted PEP (OR 0.39; 95 %CI
0.19 – 0.81; NNT 6); in a similar fashion, gabexate was effective at either slow infusion
of high dose (≥ 150 000 units) (OR 0.44; 95 %CI 0.25 – 0.79; NNT 7) or rapid infusion
of low dose (OR 0.37; 95 %CI 0.20 – 0.69; NNT 6) [61].
The benefit of nafamostat has been assessed in a meta-analysis that pooled data from
5 RCTs and 2678 patients [42]. This agent lowered the incidence of PEP (RR 0.47; 95 %CI 0.33 – 0.67). All three
RCTs available as full text found nafamostat to be effective in low risk patients
but not in high risk patients [62]
[63]
[64]. When the results observed for high risk patients in the three RCTs were pooled,
PEP incidence was not statistically different in the treatment vs. the placebo group
(8.9 % [32 /358] vs. 13.1 % [37 /283], respectively; P = 0.12).
5.3.3. Drugs influencing sphincter of Oddi pressure
Nitroglycerin reduces the incidence of PEP; however, when administered transdermally,
it is ineffective (Evidence grade 1 ++). Side effects such as transient hypotension
and headache may occur. We do not recommend the routine use of nitroglycerin for prophylaxis
of PEP (Recommendation grade A).
There is no evidence that … drugs reducing sphincter of Oddi pressure (other than
nitroglycerin) [namely, botulinum toxin, epinephrine, lidocaine, and nifedipine] …
reduce the incidence of PEP … [cf. section 5.4 below].
Glyceryl trinitrate (GTN) may be effective in preventing PEP when administered sublingually.
Topical epinephrine may be effective to prevent PEP in purely diagnostic ERCP. ESGE
does not recommend the routine use of GTN or of epinephrine for PEP prophylaxis. No
changes concerning botulinum toxin, lidocaine, and nifedipine.
Background:
The influence of GTN on the incidence of PEP was evaluated in two meta-analyses that
pooled data from 5 RCTs involving 1662 patients [65]
[66]. Both meta-analyses showed an overall significant reduction of PEP with an RR of
0.61 (95 %CI 0.44 – 0.86) and an NNT of 26. In the majority of the patients, GTN was
administered transdermally. In a subgroup analysis, transdermal GTN failed to show
a significant reduction of PEP (RR 0.66; 95 %CI 0.43 – 1.01).
Botulinum toxin [67], epinephrine [68], lidocaine [69], and nifedipine [70]
[71], were not found to prevent PEP in the corresponding RCTs.
New information since 2009:
A meta-analysis that pooled data from 8 RCTs (1920 patients) found that GTN decreases
PEP incidence compared with placebo (5.9 % vs. 9.8 %, respectively; P = 0.002) [72]. A more recent meta-analysis extended the observation to 12 RCTs (2649 patients)
and found again that GTN reduces the overall incidence of PEP (RR 0.67; 95 %CI 0.52 – 0.87)
but was, however, ineffective in lowering the incidence of moderate to severe PEP
(RR 0.70; 95 %CI 0.42 – 1.15) [73]. The route of GTN administration may influence its effectiveness. Subgroup analyses
revealed that sublingual administration of GTN was more effective than transdermal
and topical administration (RR 0.47; 95 %CI 0.28 – 0.78).
The prophylactic merit of topical epinephrine (0.02 %, 20 mL sprayed on the papilla)
was compared with placebo in diagnostic-only ERCP in two RCTs [68]
[74]. Matsushita et al. originally evaluated spraying of epinephrine onto the papilla
in 370 patients undergoing diagnostic ERCP [68]; the incidence of PEP was 0 % in the epinephrine group vs. 2.2 % in the control
group (not significant). In the second study, Xu et al. randomized 941 patients to
topical epinephrine vs. placebo; the rates of PEP were 1.95 % vs. 6.45 %, respectively
(P = 0.0086) [74]. When the results of the two studies were pooled in a meta-analysis, topical epinephrine
proved efficacious in reducing PEP (OR 0.25; 95 %CI 0.06 – 0.65; NNT 15) [75]. However, in the two RCTs of epinephrine, only patients with purely diagnostic ERCP
were included, no guidewire was initially used, cannulation times were very long,
and the definition of PEP was not standard. Based on these shortcomings, routine use
of topical epinephrine cannot be recommended for PEP prophylaxis.
5.3.4. Antibiotics
Ceftazidime reduced the incidence of PEP in a single study (Evidence grade 1 – ).
Further data are needed before recommending ceftazidime for the prophylaxis of PEP
(Recommendation grade C).
Antibiotics have not been proven effective in PEP prophylaxis; further data are needed
(Recommendation grade C).
Background:
In an RCT that tested ceftazidime for prophylaxis of PEP, the incidence of PEP was
lower in the treatment than in the control group (2.6 % vs. 9.4 %, respectively; P = 0.009) [76]. This study was of low methodological quality owing to unclear allocation concealment.
New information since 2009:
A network meta-analysis ranked antibiotics in fourth position among 16 drugs for the
efficacy of PEP prophylaxis (OR 0.46; 95 %CI 0.15 – 1.07; NNT 21) [75]. However, the difference observed between antibiotics and placebo was not statistically
significant, only 254 patients were included in the treatment arms of four RCTs [76]
[77]
[78]
[79] included in the meta-analysis (not 1082 as stated in the meta-analysis), and two
of these RCTs did not have PEP prophylaxis as their primary study endpoint [76]
[78].
5.3.5. Intensive hydration
No statement.
In a pilot study, intensive hydration seemed to effectively prevent PEP. Large-scale
RCTs to establish an evidence-based approach to intensive hydration are needed.
New information since 2009:
Based on a pilot study in 62 patients, intensive hydration in the periprocedural period
with intravenous lactated Ringer’s solution appears to reduce PEP incidence [80]. None of the patients who received aggressive hydration developed PEP, compared
with 17 % of patients who received standard hydration (P = 0.016). No patients had evidence of volume overload. Two observational studies
support this strategy of hydration for attenuating the severity of PEP [81]
[82].
5.4. Drugs proven ineffective
There is no evidence that glucocorticoids, drugs reducing sphincter of Oddi pressure
(other than nitroglycerin), antioxidants, heparin, interleukin-10, or some anti-inflammatory
drugs (other than diclofenac and indomethacin), such as pentoxifylline, semapimod,
and the recombinant platelet-activating factor acetylhydrolase, reduce the incidence
of PEP (Evidence levels from 1 – to 1 ++). None of these drugs is recommended for
PEP prophylaxis (Recommendation grade A).
No change except for GTN and epinephrine (cf. section 5.3.3 above).
Background:
The efficacy of glucocorticoids for PEP prophylaxis was evaluated in two meta-analyses
that included 6 RCTs [83]
[84]; the incidence of PEP was not significantly different in the glucocorticoids vs.
the control group (11.8 % vs. 10.6 %, respectively). Three RCTs that evaluated interleukin-10
have yielded contradictory results [85]
[86]
[87]. Subcutaneous heparin was not found to reduce PEP incidence in 2 RCTs [88]
[89]. Drugs potentially reducing the sphincter of Oddi pressure (other than GTN and topical
epinephrine), including botulinum toxin, lidocaine, and nifedipine, have not been
found to reduce PEP incidence in RCTs [67]
[68]
[90]
[91]
[92]. Three antioxidant agents, namely allopurinol, N-acetylcysteine, and natural beta-carotene
were not found to be effective in preventing PEP in 5 RCTs and one meta-analysis [93]
[94]
[95]
[96]
[97]
[98].
New information since 2009:
Regarding heparin, a meta-analysis that analyzed four clinical trials (including 3
RCTs) with a total of 1438 patients showed no beneficial effect of prophylactic heparin
for the prevention of PEP [99]. Regarding antioxidant supplementation, a meta-analysis of 11 RCTs (total, 3010
patients) showed no beneficial effect on the incidence and the severity of PEP [100]. Regarding allopurinol, an RCT found no effect for the prevention of PEP [101]. No new data are available for glucocorticoids, interleukin-10, pentoxifylline,
semapimod, and the recombinant platelet-activating factor acetylhydrolase.
6. Pancreatic stent placement for PEP prophylaxis
6. Pancreatic stent placement for PEP prophylaxis
Prophylactic pancreatic stent placement is recommended to prevent PEP in patients
who are at high risk for development of PEP. Short, 5-Fr diameter, plastic pancreatic
stents are currently recommended. Passage of the stent from the pancreatic duct should
be evaluated within 5 to 10 days of placement and retained stents should be promptly
removed endoscopically (Evidence level 1 + ; Recommendation grade A).
Prophylactic pancreatic stenting decreases the risk of PEP in high risk and mixed-case
groups; it nearly eliminates the risk of severe PEP. 5-Fr pancreatic stents are more
efficacious than 3-Fr stents in preventing PEP. ESGE recommends the placement of 5-Fr
pancreatic stents in cases at high risk of PEP. Passage of the stent from the pancreatic
duct should be evaluated within 5 to 10 days of placement and retained stents should
be promptly removed endoscopically (Evidence level 1 + ; Recommendation grade A).
Background:
Two meta-analyses have demonstrated that, in patients at high risk of PEP, prophylactic
pancreatic stent placement significantly reduces the incidence of PEP [102]
[103]. The OR was 0.44 (95 %CI 0.24 – 0.81), with an absolute risk reduction of 12.0 %
(95 %CI 3.0 – 21.0). A multicenter RCT (201 patients) showed a decreased incidence
of PEP when prophylactic pancreatic stent placement was performed, regardless of the
concomitant occurrence of other known risk factors for PEP (PEP incidence in the stent
vs. no-stent group, 3.2 % vs. 13.6 %, respectively; P = 0.019) [104]. In these studies, the risk of severe PEP was nearly eliminated following successful
placement of a prophylactic pancreatic stent.
Pancreatic stent placement was shown to be cost-effective only in patients at high
risk for PEP [105].
New information since 2009:
A meta-analysis of 14 RCTs with a total of 1541 patients showed a significant reduction
in the incidence and the severity of PEP when prophylactic pancreatic stenting was
used [106]. In addition, subgroup analysis showed that pancreatic stenting reduced the risk
of PEP in high risk and mixed-case groups. In a network meta-analysis, prophylactic
pancreatic stenting alone was shown to be less effective than NSAIDs alone, and the
combination of NSAIDs with prophylactic pancreatic stenting did not further reduce
the risk of PEP [107].
The ideal stent characteristics for PEP prophylaxis and the optimal duration of stent
placement are not definitively known. However, a network meta-analysis showed that
the probabilities of 5-Fr and 3-Fr stents being ranked as the most efficacious for
the prevention of PEP were 96.8 % vs. 3.1 %, respectively, with 5-Fr single-pigtail,
unflanged stents and 5-Fr straight, flanged stents producing similar results [108]. Furthermore, placement of 5-Fr stents requires fewer guidewires and is easier than
that of 3-Fr stents [109]
[110]. It is believed that stents need to remain in place for a minimum of 12 – 24 hours
to provide benefit, since removal at the end of ERCP negates the protection from PEP
[111] and early outward migration may also result in PEP [112]. Adverse events related to attempted prophylactic pancreatic stenting include PEP,
stent-induced pancreatic ductal damage, and inward migration [113]. Removal of proximally migrated small-diameter stents can be technically challenging,
if not impossible [114].
7. ERCP techniques
7.1. General considerations
7.1.1. Patient position
There is no evidence that the incidence of PEP is influenced by patient position during
ERCP (Evidence level 2 + + ). Therefore, no recommendation is made regarding patient
position.
No changes.
Background:
Comparisons in RCTs of different patient positions during ERCP showed no significant
difference in PEP incidence.
New information since 2009:
None.
7.1.2. Cannulation attempts
Trauma resulting from repeated attempts at biliary cannulation has been proven to
be a risk factor for the development of PEP (Evidence level 2 + + ). The number of
cannulation attempts should be minimized (Recommendation grade B).
ESGE recommends keeping the number of cannulation attempts as low as possible (Recommendation
grade B).
Background:
The risk of PEP is higher after multiple attempts at duct cannulation [6]
[21]
[115]
[116]. The rendezvous technique allows cannulation trauma to be minimized, and has been
used in most studies that evaluated intraoperative endoscopic sphincterotomy.
New information since 2009:
A meta-analysis assessed 5 RCTs that compared preoperative vs. intraoperative endoscopic
sphincterotomy [117]; PEP was less frequent with the latter approach while the incidence of other ERCP-related
complications was similar with both techniques. A nationwide case – control study
from Sweden confirmed that the rendezvous technique was associated with a reduction
in the risk of PEP compared with standard cannulation techniques, from 3.6 % to 2.2 %
(OR 0.5; 95 %CI 0.2 – 0.9; P = 0.02) [18].
7.1.3. Contrast medium
Injection of contrast medium into the pancreatic duct is an independent predictor
of PEP (Evidence level 1 + ). If pancreatic duct injection occurs incidentally or
is required, the number of injections and volume of contrast medium injected into
the pancreatic duct should be kept as low as possible (Recommendation grade B). Compared
with traditional, high-osmolality contrast agents, low-osmolality contrast agents
are costlier but are not associated with reduction in the rates of PEP (Evidence level
1 – ). The routine use of these agents for ERCP is not recommended (Recommendation
grade B).
No changes.
Background:
In a large meta-analysis, pancreatic duct injection was found to be an independent
predictor of PEP (RR 2.2; 95 %CI 1.60 – 3.01) [20]. In a retrospective study that included more than 14 000 ERCPs, the extent of pancreatic
duct injection (head-only vs. head and body vs. injection to the tail) was independently
associated with PEP [118]. The hypothesis that low-osmolality contrast agents would be less harmful than high-osmolality
contrast agents because of less important fluid shifts in the pancreas was invalidated
in a meta-analysis of 13 RCTs that involved 3381 patients [119].
New information since 2009:
One article examined the role of low-osmolality contrast agents but it does not warrant
any change in the Guideline [120].
7.1.4. Carbon dioxide
Use of carbon dioxide (CO2) as a replacement for air for luminal insufflation during
ERCP does not influence the incidence of PEP but decreases the incidence and severity
of post-procedural abdominal pain (Evidence level 1 + ). Carbon dioxide is recommended
for insufflation, and might be particularly useful for outpatient ERCPs, to reduce
post-procedural abdominal pain and to avoid confusion with PEP (Recommendation grade
B).
No changes.
Background:
Because of its higher solubility in water compared with nitrogen and oxygen, the main
components of air, carbon dioxide is cleared from the bowel following endoscopy much
faster than air (through the bloodstream and respiration).
New information since 2009:
Three meta-analyses, which included between 5 and 7 RCTs (between 446 and 818 patients),
compared carbon dioxide vs. air for gut distension during ERCP exclusively [121]
[122]
[123]. All these meta-analyses found that the use of carbon dioxide reduces post-ERCP
abdominal pain without change in other complication rates or in procedure duration.
A survey showed that the use of carbon dioxide during endoscopy is uncommon; this
may be related to implementation costs and unawareness by endoscopists of the advantages
of carbon dioxide [124].
7.1.5. Cannulation techniques
For deep biliary cannulation, the wire-guided technique reduces the risk of PEP and
increases the success rate of primary cannulation when compared with the standard
contrast-assisted method (Evidence level 1 + + ). The wire-guided technique is recommended
for deep biliary cannulation (Recommendation grade A).
No changes.
Background:
The wire-guided biliary cannulation technique entails passage of a guidewire inserted
through a catheter (most often a hydrophilic guidewire inserted into a sphincterotome)
for deeply cannulating the bile duct. Two meta-analyses published in 2009 showed that,
in RCTs, the incidence of PEP was significantly lower with the wire-guided as compared
with the standard contrast-assisted cannulation technique [125]
[126].
New information since 2009:
Five comparative studies and a meta-analysis comparing the wire-guided vs. the standard
contrast-assisted method for selective biliary cannulation were published between
2009 and 2013 [127]
[128]
[129]
[130]
[131]
[132]. Four studies [128]
[129]
[130]
[131], two of which were RCTs [128]
[131], did not confirm the results of previous meta-analyses that showed a lower risk
of PEP with the wire-guided method. In most studies, the wire-guided method shortened
cannulation and fluoroscopy times. However, in a recent meta-analysis that extended
the analysis to 12 RCTs (3450 patients), the wire-guided method significantly lowered
the incidence of PEP compared with the contrast-assisted method (RR 0.51; 95 %CI,
0.32 – 0.82) [132]. In addition, the wire-guided cannulation technique was associated with greater
primary cannulation success (RR 1.07; 95 %CI 1.00 – 1.15), fewer precut sphincterotomies
(RR 0.75; 95 %CI 0.60 – 0.95), and no increase in other ERCP-related complications.
7.1.6. Electrosurgical current
The incidence of post-sphincterotomy pancreatitis is not influenced by the type of
electrosurgical current used (whether pure-cut or blended) (Evidence level 1 + ).
Blended current is recommended for biliary sphincterotomy, particularly in patients
at high risk of bleeding (Recommendation grade A).
No changes.
Background:
As pure-cut current produces less edema than blended current [133], it was hypothesized that it might reduce the incidence of PEP after biliary sphincterotomy.
A meta-analysis of four RCTs that included 804 patients found no significant difference
in the incidence of PEP following the use of pure vs. blended current [134]. However, the incidence of bleeding was significantly higher when pure-cut current
was used.
New information since 2009:
No new evidence has become available.
7.2. Effect of difficult biliary cannulation
7.2.1 Definition
None.
ESGE recommends that future studies define difficult biliary cannulation in an intact
papilla as any of the following: cannulation attempts of duration > 5 minutes, > 5
attempts, or 2 pancreatic guidewire passages.
Background:
Many different definitions of “difficult” biliary cannulation have been used, which
make comparisons between studies impractical.
New information since 2009:
In an effort to standardize this definition, Halttunen et al. prospectively collected
data on 907 biliary cannulations attempted by experienced endoscopists at 10 centers
[115]. The authors found the incidence of PEP progressively increased with various factors
perceived as causing difficulty in cannulation. Any of the following factors was associated
with a PEP incidence of > 10 % during wire-guided cannulation of a native papilla:
cannulation attempts of duration > 5 minutes, > 5 attempts, or 2 pancreatic guidewire
passages. The latter was also noted in another prospective study [116].
For difficult cannulation, commonly used options include persistent attempts at cannulation
using standard methods, pancreatic guidewire placement (with biliary cannulation attempted
either using a guidewire, the so-called “double guidewire” (DGW) technique, or using
contrast medium injection), precut of various types, repeat attempts at ERCP 24 – 48
hours later, and patient referral to another endoscopist.
7.2.2. Pancreatic guidewire-assisted technique
Data about the usefulness and safety of pancreatic guide wire placement to facilitate
biliary cannulation in difficult cases are conflicting. Prophylactic pancreatic stent
placement decreases the incidence of PEP with this technique (Evidence level 2 +).
Pancreatic guide wire assistance may facilitate biliary cannulation mostly in the
case of inadvertent but repeated cannulation of the pancreatic duct; if this method
is used, a pancreatic stent should be placed for PEP prophylaxis (Recommendation grade
B).
In cases of difficult biliary cannulation, pancreatic guidewire (PGW) placement allows
biliary cannulation in a proportion of cases similar to persistence in attempting
cannulation with standard cannulation techniques (or precut if it is used as a backup
technique), but the risk of PEP is likely higher. In such circumstances, PEP is effectively
prevented by prophylactic pancreatic stenting (Evidence level 1 –). ESGE suggests
restricting the use of a PGW as a backup technique to cases with repeated inadvertent
cannulation of the pancreatic duct; if this method is used, deep biliary cannulation
should be attempted using a guidewire rather than the contrast-assisted method and
a prophylactic pancreatic stent should be placed (Recommendation grade B).
Background:
In the PGW-assisted technique, a guidewire is inserted in the main pancreatic duct
to facilitate deep biliary cannulation by straightening the papillary anatomy and
to prevent repeated cannulation of the pancreatic duct [135]
[136]. In two RCTs, compared with persistence in applying the standard cannulation technique,
the PGW technique yielded overall similarly low success rates for biliary cannulation
(means, 57 % vs. 56 %, respectively) and a nonsignificantly higher incidence of PEP
(means, 14 % vs. 6 %, respectively) [137]
[138]. Discordances between these RCTs in terms of cannulation success and of PEP incidence
may be related to differences in the inclusion criteria (i. e., difficult biliary
cannulation alone or combined with repeated unintended pancreatic cannulation) and
in the use of prophylactic pancreatic stenting.
New information since 2009:
Seven new studies are summarized in [Table 3] [139]
[140]
[141]
[142]
[143]
[144]
[145]. In two RCTs that compared the PGW vs. the precut techniques [142]
[143], success rates of biliary cannulation were similar but, in one of the RCTs [143], the PGW technique was plagued by a higher incidence of PEP (38 % vs. 11 %, respectively;
P = 0.01). Prophylactic pancreatic stenting was not used in any of the RCTs mentioned
above. Another RCT showed that prophylactic pancreatic stenting significantly decreased
the incidence of PEP after the PGW technique had been used [146]. In a retrospective study that included 146 patients, prophylactic pancreatic stenting
was always attempted after the PGW technique had been used, and failed prophylactic
pancreatic stenting was the only independent predictor of PEP [144]. In another retrospective study that involved 142 patients, the incidence of PEP
decreased after the authors changed their cannulation technique following PGW placement,
from contrast-assisted to guidewire-assisted biliary cannulation [147].
Table 3
New studies evaluating pancreatic guidewire (PGW) placement for biliary cannulation
in the case of difficult biliary cannulation.
|
First author,
year
|
Design
Patients, n
|
Additional inclusion criteria
|
Pancreatic stenting
|
Successful biliary cannulation rate
|
PEP rate
|
Note
|
|
Xinopoulos,
2011 [139]
|
Retrospective
112
|
Repeated unintentional pancreatic duct cannulation
|
No
|
44 % (vs. 81 % with precut in patients who had no repeated unintentional pancreatic
duct cannulation)
|
6.1 % (similar to group with easy biliary cannulation [5.3 %] and to group with precut
in place of PGW [7.5 %])
|
Comparative study
|
|
Grönroos,
2011 [140]
|
Prospective
50
|
Repeated unintentional pancreatic duct cannulation
|
No
|
66 %
|
2 %
|
Authors advocate to shift to other technique if DGW is cumbersome
|
|
Belverde,
2012 [141]
|
Retrospective
121
|
Selective pancreatic duct cannulation achieved
|
No
|
97 %
|
2.6 %
|
|
|
Angsuwatcharakon,
2012 [142]
|
RCT
44
|
No
|
No
|
PGW vs. precut:
74 % vs. 81 % (n.s.)
|
PGW vs. precut:
17.4 % vs. 9.5 %
|
Shorter cannulation time with the PGW vs. precut technique
(172 s vs. 394 s (P < 0.001)
|
|
Yoo,
2013 [143]
|
RCT
71
|
Selective pancreatic duct cannulation achieved
|
No
|
DGW vs. precut:
91.2 % vs. 91.9 % (n.s.)
|
DGW vs. precut:
38 % vs. 11 % (P = 0.01)
|
|
|
Ito,
2013 [144]*
|
Retrospective
146
|
n.a.
|
Attempted in all
|
PGW: 70 %;
if failure with PGW, DGW was successful in 72 %
|
PGW, 8 %;
DGW, 4 %
|
Independent predictor of PEP: failed pancreatic stenting (OR 8.3; 95 %CI 2.3 – 30)
|
|
Tanaka,
2013 [145]
[1]
|
Retrospective
79
|
Opacification of the pancreatic duct
|
40 /79
|
82 % (PGW) vs. 83 % (DGW)
|
11 % (PGW) vs. 7 % (DGW)
|
|
PEP, post-endoscopic retrograde cholangiopancreatography (post-ERCP) pancreatitis;
PGW, pancreatic guidewire; DGW, double guidewire; RCT, randomized controlled trial;
n.s., not significant; n.a., not available; OR, odds ratio; CI, confidence interval
1 Biliary cannulation attempted by searching the bile duct through contrast medium
injection with a guidewire in the main pancreatic duct (PGW technique), as opposed
to searching through manipulation of a guidewire inserted in a catheter (DGW technique).
7.2.3. Precut biliary sphincterotomy
Various techniques of precut biliary sphincterotomy have been described; the fistulotomy
technique may present a lower incidence of PEP than standard needle-knife sphincterotomy
but further RCTs are required to determine which technique is safer and more effective,
based upon the papillary anatomy. There is no evidence that the success and complication
rates of biliary precut are affected by the level of endoscopist experience in this
technique but published data only report on the experience of one endoscopist (Evidence
level 2 – ). Prolonged cannulation attempts using standard techniques may impart a
risk for PEP greater than the precut sphincterotomy itself (Evidence level 2 +). Precut
sphincterotomy should be performed by endoscopists with expertise in standard cannulation
techniques (Recommendation grade D). The decision to perform precut biliary sphincterotomy,
the timing, and the technique, are based on anatomic findings, endoscopist preference,
and procedural indication (Recommendation grade C).
In cases of difficult biliary cannulation, early precut is associated with a lower
PEP incidence than persistent attempts using the standard approach but the overall
success and complication rates are similar with both approaches. Needle-knife fistulotomy
seems to be associated with fewer complications, including PEP, compared with other
precut techniques.
ESGE suggests that needle-knife fistulotomy should be the preferred precut technique
in patients with a bile duct dilated down to the papilla. Conventional precut and
transpancreatic sphincterotomy present similar success and complication rates; if
conventional precut is elected and pancreatic cannulation is easily obtained, ESGE
suggests attempting to place a small-diameter (3-Fr or 5-Fr) pancreatic stent to guide
the cut and leaving the pancreatic stent in place at the end of ERCP for a minimum
of 12 – 24 hours (Recommendation grade B).
Background:
Precut biliary sphincterotomy can be done by different techniques: needle-knife precut
starting at the orifice of the papilla (conventional precut), needle-knife above the
orifice (fistulotomy), transpancreatic sphincterotomy (septotomy) and needle-knife
over a pancreatic stent. Compared with prolonged attempts at biliary cannulation using
standard techniques, the use of precut sphincterotomy has long been thought to increase
the success of biliary cannulation and the incidence of PEP [6]
[20].
New information since 2009:
Two meta-analyses investigated the issue of timing of the precut procedure in 6 RCTs
[148]
[149]. Aside from the definition of early precut, differences between studies included
the technique of precut and the randomization ratio (from 1:1 to 1:3). Endoscopists
experienced in precut techniques performed all procedures. The overall incidence of
PEP was lower in patients allocated to early precut than to persistence in the use
of standard techniques (2.5 % vs. 5.3 %; OR 0.47; 95 %CI 0.24 – 0.91) but the overall
cannulation and complication rates were similar in patients allocated to early precut
or to persistence in the use of standard cannulation techniques.
In a retrospective study comparing different precut techniques, the risk of PEP was
significantly lower after fistulotomy (2.6 %) compared with conventional precut (20.9 %)
and pancreatic septotomy (22.4 %); the overall complication rate was also significantly
lower after fistulotomy vs. other precut techniques [150]. In another retrospective study of needle-knife fistulotomy performed in 204 patients,
complications (mostly PEP) progressively increased with decreasing common bile duct
(CBD) diameter, up to 14 % in patients with a CBD diameter < 4 mm [151]. Two retrospective studies compared transpancreatic septotomy and needle-knife precut
(with pancreatic stenting or not) for biliary access [152]
[153]; they found no significant differences in terms of PEP, overall complication and
overall success rates.
Needle-knife precut assisted by pancreatic stenting has been proposed for reducing
PEP. One RCT found that, compared with stent removal at the end of the procedure,
leaving the stent for 7 – 10 days reduces the incidence and severity of PEP (P < 0.05 for both comparisons) [111].
7.3. Specific therapeutic techniques
7.3.1. Balloon dilation as a substitute for endoscopic sphincterotomy
Compared with endoscopic sphincterotomy, endoscopic papillary balloon dilation (EPBD)
using small-caliber balloons (≤ 10 mm) is associated with a significantly higher incidence
of PEP and significantly less bleeding (Evidence level 1 ++). EPBD is not recommended
as an alternative to sphincterotomy in routine ERCP but may be useful in patients
with coagulopathy and altered anatomy (e. g. Billroth II) (Recommendation grade A).
If balloon dilation is performed in young patients, the placement of a prophylactic
pancreatic stent should be strongly considered (Evidence level 4; Recommendation grade
D).
Compared with endoscopic sphincterotomy, endoscopic papillary balloon dilation (EPBD)
using small-caliber balloons (≤ 10 mm) is associated with more PEP, less bleeding,
and fewer late stone recurrences; the risk of PEP is inversely associated with the
duration of EPBD (Evidence level 1 ++). ESGE does not recommend endoscopic papillary
balloon dilation as an alternative to sphincterotomy in routine ERCP, but it may be
advantageous in selected patients; if this technique is used, the duration of dilation
should be longer than 1 minute (Recommendation grade A).
Background:
The use of EPBD may be advantageous compared with endoscopic sphincterotomy, by decreasing
clinically significant bleeding in patients with coagulopathy, for preserving sphincter
of Oddi function in younger patients [154], and in patients with altered anatomy (Billroth II) where sphincterotomy is technically
difficult. In two meta-analyses, the use of EPBD resulted in a lower success rate
than endoscopic sphincterotomy for the initial removal of biliary stones, with a significantly
higher incidence of PEP and significantly lower incidence of bleeding [155]
[156].
New information since 2009:
A meta-analysis of 10 high quality RCTs (1451 patients) that compared EPBD vs. endoscopic
sphincterotomy for biliary stone extraction found similar success rates of biliary
stone extraction with both methods but did not bring new findings with regard to complications
[157]; indeed it included fewer RCTs than the previously published Cochrane meta-analysis
[156].
A meta-analysis that included 12 RCTs (1649 patients) indicated that duration of EPBD
is inversely associated with the incidence of PEP: short (≤ 1 minute) EPBD was associated
with a higher risk of PEP (OR 3.87; 95 %CI 1.08 – 13.84) compared with endoscopic
sphincterotomy (4 RCTs), but long EPBD inflation time (> 1 minute) was not (OR 1.14;
95 %CI 0.56 – 2.35) (6 RCTs) [158]. A lower risk of PEP after long EPBD was also reported in the single RCT that compared
long (5-minute) vs. short (1-minute) EPBD (RR 0.32; 95 %CI 0.11 – 0.93) [159]. The authors suggested that the increased risk of PEP after short EPBD might be
related to inadequate balloon expansion resulting in a worsened compression of the
pancreatic duct.
A meta-analysis that included 3 RCTs (496 patients) with a follow-up longer than 1
year showed that stone recurrence was less frequent after EPBD vs. endoscopic sphincterotomy
(OR 0.48; 95 %CI 0.26 – 0.90) [160]. Another RCT that was not included in this meta-analysis and included 474 patients
also reported that, in patients with biliary stones ≤ 8 mm, overall late complications
and stone recurrence were less frequent after EPBD than after endoscopic sphincterotomy
(5.3 % vs. 17.3 %, P = 0.009; 4.4 % vs. 12.7 %; P = 0.048, respectively) (mean follow-up, 55 months); the difference was not significant
for patients with stones > 8 mm [161]. A retrospective cohort study with a median follow-up of 92 months also showed a
lower incidence of common bile duct stone recurrence after EPBD vs. endoscopic sphincterotomy
[162].
7.3.2. Large-balloon dilation for extraction of difficult biliary stones
Potential advantages of performing large-balloon dilation in addition to endoscopic
sphincterotomy for extraction of difficult biliary stones remain unclear (Evidence
level 3). Endoscopic sphincterotomy plus large-balloon dilation does not seem to increase
the risk of PEP and can avoid the need for mechanical lithotripsy in selected patients,
but not enough data are available to recommend routine use over biliary sphincterotomy
alone in conjunction to lithotripsy techniques (Recommendation grade D).
For the extraction of difficult biliary stones, endoscopic sphincterotomy plus large-balloon
dilation presents a risk of PEP similar to that of endoscopic sphincterotomy alone;
it presents a lower bleeding risk and, possibly, lower overall morbidity and requires
less use of mechanical lithotripsy. ESGE suggests performing endoscopic sphincterotomy
plus large-balloon dilation in place of endoscopic sphincterotomy alone for the extraction
of selected difficult biliary stones (Recommendation grade B).
Background:
In patients with a tapered distal bile duct or large biliary stones, endoscopic sphincterotomy
followed by biliary dilation using a large-diameter (12 – 20 mm) balloon has been
proposed to facilitate stone extraction [163]. This technique has been associated with high success rates for stone extraction
without the need for mechanical lithotripsy and with acceptable complication rates
in case series [164]
[165]
[166]
[167]; the single RCT that had compared this method with endoscopic sphincterotomy at
the time of publication of the first Guideline had found no differences in rates of
successful stone clearance, need for mechanical lithotripsy, and complications [168].
New information since 2009:
A meta-analysis of 7 RCTs that included 790 patients found that, compared with endoscopic
sphincterotomy alone, endoscopic sphincterotomy followed by large balloon dilation
was associated with similar PEP rates but significantly fewer overall complications
and significantly less bleeding [169]. However in this meta-analysis, studies which, according to their authors, were
retrospective, were considered to be RCTs [170]. Two other meta-analyses that included 3 RCTs and 4 – 6 retrospective studies (including
between 90 and 1295 patients) found similar incidences of PEP but significantly lower
bleeding risk with endoscopic sphincterotomy plus large balloon dilation vs. endoscopic
sphincterotomy [171]
[172]. Additional differences in favor of endoscopic sphincterotomy plus large balloon
dilation were mostly attributable to the results of retrospective studies; they included
a lower use of mechanical lithotripsy and a lower overall complication rate [171]
[172].
7.3.3. Sphincter of Oddi manometry
In patients undergoing pancreatic sphincter of Oddi manometry, use of the standard
perfusion catheter without an aspiration port has been shown to increase the risk
of PEP compared with modified water perfusion catheters (Evidence level 2 ++). Pancreatic
sphincter of Oddi manometry should be done using a modified triple-lumen perfusion
catheter with simultaneous aspiration or a microtransducer catheter (non-water-perfused)
(Recommendation grade B).
ESGE recommends that all patients undergoing ERCP for known or suspected sphincter
of Oddi dysfunction (SOD) receive rectal NSAIDs combined with pancreatic stenting.
Pancreatic sphincter of Oddi manometry should be done using a modified triple-lumen
perfusion catheter with simultaneous aspiration or a microtransducer catheter (non-water-perfused)
(Recommendation grade B).
Background:
It is well documented that patients undergoing ERCP for known or suspected SOD are
at high risk of PEP, regardless of whether sphincter of Oddi manometry (SOM) is performed
[173]. Biliary SOM does not appear to increase the risk of PEP [173]; pancreatic SOM is associated with an increased risk of PEP depending on the technique
used [174]. When water-perfused catheters are used to measure pancreatic sphincter pressure,
continuous aspiration of fluid from the main pancreatic duct through one of the three
manometry ports prevents overfilling of the duct [174], a known risk factor for PEP. A specialized aspiration manometry catheter is available
for this purpose, though one port of a standard triple lumen catheter can be used
for aspiration [174]. The use of solid-state manometry catheters decreases the risk of PEP as there is
no perfusion or filling of the pancreatic duct [175].
The results of SOM should not influence the decision of whether to institute measures
for prevention of PEP. As mentioned, all patients with known or suspected SOD are
at high risk for PEP and should receive PEP prophylaxis [175]. Indeed one study showed that pancreatic stent placement reduced PEP in patients
with an intact papilla undergoing SOM and when the manometry results were normal [176].
New information since 2009:
Recent data suggest that rectal NSAIDs are at least as effective as pancreatic stents
in patients with SOD [34]
[44]
[107]. However, the most evidence-based approach for preventing PEP in high risk cases
remains the combination of rectal NSAIDs and prophylactic pancreatic stenting. In
the RCT mentioned above, patients who received rectal indomethacin and a pancreatic
stent (n = 247) had a PEP rate of 9.7 % compared with 16.1 % in those who received
a stent alone (n = 249) (P = 0.04) [34]. To date, there are no clinical trial data examining whether rectal NSAIDs are effective
when administered instead of prophylactic pancreatic stenting.
The data apply mainly to the initial SOM or in patients with SOD and intact papillae.
One exception to routine PEP prophylaxis might be in patients with SOD who have had
dual sphincterotomy and continued pain who undergo repeat ERCP with re-evaluation.
If both sphincters are found to be patent and/or only the biliary sphincter is studied,
one might consider not placing a pancreatic stent. However, the risk of administration
of one dose of rectal NSAIDS is so small that its use outweighs the risk of avoiding
them.
8. Selection of measures for PEP prophylaxis
8. Selection of measures for PEP prophylaxis
For low-risk ERCPs, periprocedural rectal administration of nonsteroidal anti-inflammatory
drugs (NSAIDs) is recommended. For high risk ERCPs, prophylactic pancreatic stent
placement should be strongly considered (Evidence level 1 + ; Recommendation grade
A).
ESGE recommends routine rectal administration of 100 mg of diclofenac or indomethacin
immediately before or after ERCP in all patients without contraindication. In addition
to this, in cases at high risk for PEP, the placement of a 5-Fr prophylactic pancreatic
stent should be strongly considered. Sublingually administered GTN or 250 µg somatostatin
given in bolus injection might be considered as an option in high risk cases if NSAIDs
are contraindicated and if prophylactic pancreatic stenting is not possible or successful
(Recommendation grade A).
The following conditions are considered to represent high risk for PEP: endoscopic
ampullectomy, known or suspected SOD, pancreatic sphincterotomy, precut biliary sphincterotomy,
pancreatic guidewire-assisted biliary cannulation, endoscopic balloon sphincteroplasty,
and presence of more than three of the risk factors listed in [Table 1] (definite or likely). Procedures and patient conditions that do not fulfill these
criteria are considered to represent low risk for PEP.
