Abbreviations
CP:
chronic pancreatitis
ERCP:
endoscopic retrograde cholangiopancreatography
ESGE:
European Society of Gastrointestinal Endoscopy
ESWL:
extracorporeal shockwave lithotripsy
FCSEMS:
fully covered self-expandable metal stent
LAMS:
lumen-apposing metal stent
MPD:
main pancreatic duct
MRCP:
magnetic resonance cholangiopancreatography
MRI:
magnetic resonance imaging
OR:
odds ratio
PFC:
pancreatic fluid collection
PPC:
pancreatic pseudocyst
RR:
relative risk
RCT:
randomized controlled trial
SEMS:
self-expandable metal stents
S-MRCP:
secretin-enhanced magnetic resonance cholangiopancreatography
This Guideline is an official statement of the European Society of Gastrointestinal
Endoscopy (ESGE). It addresses the indications for, techniques, and results of treatment
of chronic pancreatitis by extracorporeal shockwave lithotripsy and/or endoscopy.
1 Introduction
The Clinical Guideline on the endoscopic treatment of chronic pancreatitis (CP) published
in 2012 by the European Society of Gastrointestinal Endoscopy (ESGE) made recommendations
on the indications and modalities of treatment for CP [1]. New evidence has become available since then and is discussed in the present update,
and new recommendations are issued.
2 Methods
ESGE commissioned this Guideline and appointed a Guideline leader (J.M.D.) who invited
the listed authors to participate in the project development. The key questions were
prepared by the coordinating team (J.M.D., A.T., M.D.) and then approved by the other
members. The coordinating team formed task force subgroups, each with its own leader,
who was assigned key questions (see Appendix 1 s, online-only Supplementary Material).
Each task force performed a systematic literature search to prepare evidence-based
and well-balanced statements on their assigned key questions. The literature search
was performed using MEDLINE and Embase to identify new publications since January
2012 published in English. The Grading of Recommendations Assessment, Development
and Evaluation (GRADE) system was adopted to define the strength of recommendation
and the quality of evidence [2]. Each task force proposed statements on their assigned key questions which were
discussed during a meeting in Brussels, Belgium, in June 2017. Literature searches
were re-run in August 2018. This time-point should be the starting point in the search
for new evidence for future updates to this Guideline. In August 2018 a draft prepared
by J.M.D. was sent to all group members for review. The draft was also reviewed by
two members of the ESGE Governing Board, by external reviewers, and by the ESGE National
Societies and Individual Members. After agreement on a final version, the manuscript
was submitted to the journal Endoscopy for publication. All authors agreed on the final revised version.
This Guideline was issued in 2018 and will be considered for review in 2022, or sooner
if new and relevant evidence becomes available. Any updates to the Guideline in the
interim period will be noted on the ESGE website: https://www.esge.com/esge-guidelines.html.
3 Choice of treatment and initial work-up
3 Choice of treatment and initial work-up
ESGE suggests endoscopic therapy and/or extracorporeal shockwave lithotripsy (ESWL)
as the first-line therapy for painful uncomplicated chronic pancreatitis (CP) with
an obstructed main pancreatic duct (MPD) in the head/body of the pancreas. The clinical
response should be evaluated at 6 – 8 weeks; if it appears unsatisfactory, the patient’s
case should be discussed again in a multidisciplinary team and surgical options should
be considered.
Weak recommendation, low quality evidence.
The first step proposed to relieve pain in patients with uncomplicated CP includes
lifestyle modifications plus, in selected patients, endoscopic therapy and/or ESWL
[3]. If endoscopic therapy and/or ESWL provide no persistent pain relief or technically
fail, or if the patient is not a good candidate for endoscopic therapy and/or ESWL,
medical treatment including analgesics and adjunctive agents (e. g., pharmaceutical
agents aimed to relieve neuropathic pain) are proposed, with the final step being
early surgery for nonresponders. In a large prospective multicenter U.S. cohort (n = 521),
medical therapy, endoscopic therapy, and pancreatic surgery were performed in 69 %,
52 %, and 18 % of patients, respectively [4]. Similarly, in 33 series of CP patients treated with endoscopic therapy and/or ESWL,
surgery was performed during long-term follow-up in a minority of patients, less frequently
in those with stones as the main obstructing factor (117 of 1695 [6.9 %], 13 series,
Table 1 s, see Supplementary Material, online-only) as compared to those with strictures (157
of 1061 patients [14.8 %], 20 series, Table 2 s; P < 0.001).
Two trials have suggested that surgery was superior to endoscopic therapy and/or ESWL
for pain relief [5]
[6]
[7]. In the first trial [5], pain was absent after 5 years of follow-up in 15 % vs. 34 % of patients treated
with endoscopic therapy vs. surgery, respectively, showing that neither of these options
is entirely satisfactory. Furthermore, neither ESWL nor cumulative stenting were used
and the randomized design of the study is questionable. In the other trial [6]
[7], only 39 patients were included; all of them had advanced CP and most were opioid-dependent.
For these reasons, the results cannot be extrapolated to all patients with CP. A cost
– effectiveness model based on data of this randomized controlled trial (RCT) unsurprisingly
concluded that surgery was more effective and less costly than endoscopic therapy
in CP [8], but another RCT has shown that ESWL could provide satisfactory clinical results
at a relatively low cost in patients with obstructive stones in the main pancreatic
duct (MPD) (62 % of patients with no pain relapse at 4-year follow-up after ESWL)
[9]. Finally, a retrospective study (86 CP patients) reported similar pain relief 5.4
years after endoscopic therapy and/or ESWL vs. surgery, but surgery carried more complications
and higher costs [10].
Endoscopic therapy and/or ESWL aim to relieve an obstruction in the MPD. They are
proposed only to patients with marked ductal changes, mainly dilation, corresponding
to the most severe grade in the Cambridge classification of pancreatitis [11]. No recent publications have reported the results of endoscopic therapy in patients
with less severe changes [12]
[13]. In painless CP, endoscopic therapy and/or ESWL are not performed because the only
potential benefit (preserving the pancreatic function) is uncertain: a single prospective
nonrandomized comparative study examined this in 42 CP patients and found that the
mean value of the N-benzoyl-L-tyrosyl para-aminobenzoic acid test was higher at 5-year
follow-up after stenting vs. no stenting of a MPD stricture while no differences were
observed for overt diabetes [14]. These results have not been confirmed and in most long-term studies the pancreatic
function deteriorated during follow-up [15]
[16]
[17]
[18]
[19].
ESGE suggests, for the selection of patients for initial or continued endoscopic therapy
and/or ESWL, taking into consideration predictive factors associated with a good long-term
outcome. These include, at initial work-up, absence of MPD stricture, a short disease
duration, a short disease duration, non-severe pain, absence or cessation of cigarette
smoking and of alcohol intake, and, after initial treatment, complete removal of obstructive
pancreatic stones and resolution of pancreatic duct stricture with stenting.
Weak recommendation, low quality evidence.
During the pretherapeutic evaluation of a patient, factors associated with a good
long-term
clinical outcome may help to select patients for endoscopic therapy and/or ESWL. These
factors should be considered as orientative only as the differences in the proportions
of patients with long-term success for an individual factor are small. The factors
include absence of MPD stricture (see above) as well as short disease duration, non-severe
pain (including low dose use of narcotics), the absence or cessation of cigarette
smoking and of alcohol intake, cephalic location of stones, the absence of pancreas
divisum if MPD stenting is required, and steatorrhea (4, 4, 3, 2, 1, 1, and 1 studies,
respectively) (Table 3 s). Favorable prognostic factors related to endoscopic therapy and/or ESWL include
complete stone removal and MPD stricture resolution after stenting (2 and 1 studies,
respectively).
Patients with an MPD obstruction located only in the tail of the pancreas are not
considered candidates for ESWL and/or endoscopic therapy by some groups of authors
[20].
ESGE suggests performing a high quality pancreatic computed tomography (CT) scan and/or
magnetic resonance imaging with cholangiopancreatography to reasonably rule out pancreatic
cancer and to plan treatment in patients with chronic pancreatitis.
Weak recommendation, low quality evidence.
The risk of pancreatic cancer is increased in patients with CP, particularly in the
first years following diagnosis [21]. A meta-analysis (52 studies, 5399 patients) found that endoscopic ultrasonography
(EUS), CT scan, and magnetic resonance imaging (MRI) present similar diagnostic accuracies
for the diagnosis of pancreatic cancer [22]. Imaging methods of the pancreas are constantly refined and they are often used
in combination [23]
[24]. In the particular context of CP, MRI with diffusion-weighted imaging has shown
sensitivity and specificity for the diagnosis of malignancy of 86 % and 82 %, respectively,
in a meta-analysis [25], while EUS-guided sampling seems to be less sensitive according to a retrospective
and a prospective study (54 % and 74 % vs. 89 % and 91 % in the presence vs. the absence
of CP, respectively) [26]
[27]. The yield of EUS elastography and contrast-enhanced harmonic EUS as well as methods
to improve the accuracy of EUS-guided sampling are discussed in dedicated ESGE Guidelines
[28]
[29].
Non-contrast CT scan accurately delineates calcified stones in the pancreas and allows
measurement of stone density, a factor associated with the completeness of stone extraction
[30]. Contrast enhancement may help to locate stones relative to the ducts [31]
[32]. Magnetic resonance cholangiopancreatography (MRCP) identifies ductal abnormalities;
in two retrospective studies its diagnostic accuracy for ductal abnormalities was
73.2 % (41 children with CP) and 92.2 % (30 adults with CP) [33]
[34].
4 Pancreatic stone management
4 Pancreatic stone management
Pancreatic stones seem to arise as either direct and evenly calcified stones or as
radiolucent
protein plugs that may or may not become calcified during the course of the disease
[35]. The vast majority of pancreatic stones are calcified and radiopaque; their prevalence
increases with time to reach 50 % and 100 %, at 5 and 14 years after the onset of
the disease, approximately [36]. In a multicenter survey (879 CP patients with a mix of newly diagnosed and long-standing
disease), calcified pancreatic stones were detected in 62 % of patients; they were
more frequent in men, heavy drinkers (> 80 g/day), and heavy smokers ( ≥ 20 cigarettes/day)
[37]. Pancreatic stones in CP patients who undergo endoscopic therapy and/or ESWL are
solitary in 10 % – 62 % of patients; they are most frequently located in the pancreatic
head only, with a mean size of 10 mm, and they are associated with strictures in approximately
50 % of patients (Table 4 s).
Successful stone fragmentation following ESWL has been defined as stones broken into
fragments ≤ 2 or 3 mm, or by the demonstration at X-ray of decreased stone density,
increased stone surface, and heterogeneity of the stone which may fill the MPD and
adjacent side branches [38]. Ductal clearance has been defined as complete, partial, or unsuccessful if the
proportion of stones cleared was > 90 %, 50 % – 90 %, or < 50 %, respectively [39].
ESGE recommends ESWL for the clearance of radiopaque obstructive MPD stones larger
than 5 mm located in the head/body of the pancreas, and endoscopic retrograde cholangiopancreatography
(ERCP) for MPD stones that are radiolucent or smaller than 5 mm.
Strong recommendation, moderate quality evidence.
Endoscopy alone, using pancreatic sphincterotomy and a basket or a balloon, allows
stone extraction in a minority of CP patients: 9 % of 1041 patients in two retrospective
series [40]
[41] and 14 % of 1834 patients in a survey of 125 hospitals [42]. Failed stone extraction using these techniques is associated with stones > 10 mm,
diffuse location, stone impaction, and location upstream from a stricture [41]
[43]. Furthermore, pancreatic mechanical lithotripsy carries a complication rate threefold
higher compared with biliary mechanical lithotripsy according to a retrospective study
of 712 patients [44]. Complications in the 69 patients with pancreatic stones included trapped or broken
basket, traction wire fracture, and one pancreatic ductal leak which required surgery
[44]. In one of the abovementioned series, ESWL allowed the endoscopic extraction of
pancreatic stones in > 80 % of the patients after failed stone extraction at primary
endoscopy [40]. Similarly, a retrospective study (70 patients) found that performance of ESWL prior
to the endoscopic attempt at stone extraction was the only independent factor associated
with successful stone clearance [45]. Therefore, a primary endoscopic attempt at pancreatic stone extraction is reserved
to selected patients, based on a reasonable expectation of success or on technical
difficulty in performing ESWL as with radiolucent stones or stones < 5 mm that are
difficult to target using X-rays.
A meta-analysis (27 studies including 6 prospective ones, in total 3189 patients with
pancreatic stones > 5 mm) reported that pancreatic ESWL allowed complete/partial MPD
clearance in 70 %/22 % of patients, respectively, that pain was absent or mild-moderate
during the 2 years following treatment in 52.7 % and 33.4 % of patients, respectively,
and that quality of life improved after ESWL in 88.2 % of patients [39]. ERCP was combined with ESWL in most studies. Table 1 s summarizes the outcomes of ESWL alone or combined with endoscopic stone extraction.
Pain relapsed in 30 % – 50 % of patients during a follow-up of 1 – 14 years and surgery
was performed in 6.9 % of patients. Of note, the studies that reported the timing
of pain relapse showed that patients with no pain relapse at 2-year follow-up rarely
experience pain relapse thereafter [9]
[16]
[45], in particular if stone clearance has been complete [30]. Approximately half of patients with relapsing pain present with stone recurrence
[46].
ESGE suggests restricting the use of endoscopic therapy after ESWL to patients with
no spontaneous clearance of pancreatic stones after adequate fragmentation by ESWL.
Weak recommendation, moderate quality evidence.
The addition of endoscopic therapy to ESWL provided no additional benefit in two studies
that compared ESWL vs. ESWL systematically combined with endoscopic therapy [9]
[47]. An RCT (55 patients) of ESWL alone vs. ESWL combined with endoscopic therapy reported
similar decreases in MPD diameter and in number of pain episodes/year; patients who
had ESWL combined with endoscopic therapy had a longer hospital stay and higher treatment
costs [9]. Furthermore, a retrospective series (146 patients) found no differences in pain
resolution 6 months after ESWL alone vs. combined with endoscopic therapy; the criteria
for performing endoscopic therapy or not were not stated [47].
The first case series of ESWL alone for pancreatic stones was reported in 1996 from
Japan; it reported pain relief in 22 of 28 patients (79 %) at 44-month follow-up [48]. Three surveys of the treatment of pancreatic stones in Japanese hospitals during
5-year periods were reported in 2018 (125 hospitals, 1834 patients), 2013 (34 hospitals,
916 patients) and 2005 (11 hospitals, 555 patients) [41]
[42]
[49]. The rates of spontaneous stone clearance after ESWL were 15 %, 49 %, and 70 %,
respectively, and the proportions of patients who had endoscopic therapy after ESWL
were 81 %, 56 %, and 43 %, respectively. The inclusion of a greater number of less
specialized hospitals in the most recent survey might explain these differences [42]. In all of these studies, no differences in baseline characteristics of patients
who had ERCP alone or combined with ERCP were reported except for gender in one study
[47].
ESWL: technical factors, complications and contraindications
Pancreatic stone fragmentation is obtained after ESWL in approximately 90 % of patients
[50]; this may require multiple ESWL sessions (up to 8 in a large series with a high
rate of successful fragmentation) [20]. More shockwaves may be required for stones that are larger [51], multiple [52], or associated with a MPD stricture [53], while pancreatic stenting prior to ESWL seems to decrease the number of shockwaves
and of ESWL sessions required [51]. Multicenter surveys have suggested that stone fragmentation is less frequently
successful in low case volume centers while the role of the type of lithotripter has
been controversial [41]
[42]
[49].
After ESWL, endoscopic clearance of stone fragments has been more frequently successful
with solitary stones [17]
[20]
[30]
[45]
[53], stones located in the pancreatic head [20], stones with a density at CT scan of < 820.5 Hounsfield units [30], if a pancreatic stent had been inserted prior to ESWL [54]
[55], if secretin had been administered at the beginning of ESWL [55], and if ERCP was delayed by more than 2 days after ESWL [56]. Pancreatic pseudocysts (PPCs) did not affect stone clearance or adverse events
in a prospective series of 849 patients (59 with a PPC) [57].
The most frequent complication of ESWL is pancreatitis; it has been reported in 4.2 %
of the patients in a meta-analysis, but most of the included studies were retrospective
and did not allow the attribution of complications to either endoscopic therapy or
ESWL as both were performed in most patients [39]. In a prospective study (634 patients, 1470 ESWL sessions), transient adverse events
(asymptomatic hyperamylasemia, hematuria, gastrointestinal mucosal injury) and complications
were detected in 21.2 % and 6.7 % of the procedures, respectively [58]. Complications included pancreatitis, infection, steinstrass (acute stone incarceration
in the papilla), bleeding, and perforation; they were classified as moderate or severe
in 1.1 % of the cases. Skin erythema and tenderness in the region in contact with
the shockwave head were noted in most patients [58].
Contraindications to ESWL include non-correctable coagulation disorders, pregnancy,
and presence in the shockwave path of bone, calcified vessels, or lung tissue [59]. Specific precautions should be taken for patients with implantable defibrillators
and pacemakers [60].
ESGE suggests considering pancreatoscopy-guided lithotripsy when ESWL is not available
or for stones that were not fragmented after adequately performed ESWL.
Weak recommendation, low quality evidence.
Reports of intracorporeal lithotripsy using electrohydraulic or laser lithotripsy
under peroral pancreatoscopy are sparse. A systematic review (10 studies, 87 patients)
reported successful MPD clearance in 43 % – 100 % of patients [61]. Results may be biased as the reports included selected patients with anatomical
features thought to permit passage of the pancreatoscope to the target stone in a
stable position. The largest study reported complete and partial stone clearance in
24 (63 %) and 10 (26 %) of 38 patients, respectively, after a total of 280 endoscopic
therapy sessions, including 88 with pancreatoscopy; complications (post-ERCP pancreatitis
and one perforation) were reported for 20 procedures and the overall clinical success
rate was 74 % [62].
5 Pancreatic strictures
Since the previous publication of this Guideline no new definitions of the types of
MPD strictures in CP have been reported. Besides benign vs. malignant and single vs.
multiple, strictures may be classified as either non-dominant or dominant [63]. Dominant MPD strictures are defined by the presence of at least one of the following
characteristics: upstream MPD dilatation ≥ 6 mm in diameter, prevention of contrast
medium outflow alongside a 6-Fr catheter inserted upstream from the stricture, or
abdominal pain during continuous infusion of a nasopancreatic catheter inserted upstream
from the stricture with 1 L saline for 12 – 24 h.
Stent insertion across a dominant MPD stricture (or the most proximal [tail] one in
the case of multiple strictures) defines technical success. It aims to: (i) decompress
the MPD, thereby ameliorating pain, and (ii) persistently dilate the stricture(s).
Less frequent indications include facilitation of MPD stone clearance in association
with ESWL as detailed above, and to bypass an obstruction in the ventral duct by inserting
a stent through the minor papilla into the MPD [64]. A prospective non-randomized study showed in 42 patients with a dominant MPD stricture
that pain recurred less frequently in patients who had received a temporary pancreatic
stenting vs. those who had not (15 % vs. 50 % during a 5-year follow-up) [14]. Before stent dilation therapy is embarked upon, malignancy should be reasonably
excluded, for example by brush cytology and cross-sectional imaging (see Section 3).
Refractory MPD strictures are defined as symptomatic dominant strictures that persist
or relapse after 1 year of single pancreatic stent placement. A validated short-term
definition for clinical success is still lacking. For long-term evaluation, the absence
of pain during the year following stent removal still seems a reasonable and workable
definition.
ESGE suggests treating painful dominant MPD strictures with a single 10-Fr plastic
stent for one uninterrupted year if symptoms improve after initial successful MPD
drainage. The stent should be exchanged if necessary, based on symptoms or signs of
stent dysfunction at regular pancreas imaging at least every 6 months. ESGE suggests
consideration of surgery or multiple side-by-side plastic stents for symptomatic MPD
strictures persisting beyond 1 year after the initial single plastic stenting, following
multidisciplinary discussion.
Weak recommendation, low quality evidence.
Insertion of a single plastic stent has been used as the initial endoscopic therapy
for MPD
strictures (Table 2 s); these strictures were single in > 80 % of the patients [65]
[66], and some studies explicitly excluded patients with multiple strictures [67]. After temporary insertion of a single plastic stent in the MPD, stricture resolution
was achieved in 9 % [68] to 50 % [6] of 145 patients in five studies [6]
[67]
[68]
[69]
[70] but this is not required for long-term pain relief [67]. Long-term pain relief was reported in 67.5 % of 536 patients (95 % confidence interval
[CI] 51.5 % – 80.2 %) in a meta-analysis of 9 studies [71]. The follow-up duration after stent removal was not calculated but in most studies
it was ≥ 24 months, the period during which almost all pain relapses occur [6]
[14]
[66]
[70]
[72]
[73]
[74].
Refractory strictures may be treated by surgery, multiple side-by-side plastic stents
(Table 5 s), or self-expandable metal stents (SEMSs) (Table 6 s).
The temporary insertion of multiple side-by-side plastic stents in 48 patients yielded
stricture resolution and pain relief at 9.5-year follow-up in 89.5 % and 77.1 % of
the patients, respectively [75]
[76].
With respect to SEMSs, uncovered and partially covered types have provided disappointing
results [77] but temporary placement of a fully covered SEMS (FCSEMS) has provided pain improvement
in 85 % of patients according to a systematic review of four prospective series (total
61 patients) [78]. These studies were limited by a very short follow-up, and three more recent studies
(n = 41) have reported pain improvement in 37 % – 88 % of patients during a follow-up
of 3 – 4 years [79]
[80]
[81]. Pancreatic FCSEMS need further evaluation in the setting of clinical trials because
of potential complications as listed below.
Pancreatic stenting: technical factors and complications
Whether or not a pancreatic sphincterotomy should be performed before pancreatic stent
insertion has not been addressed in any study, but both methods have been reported
for the insertion of a single plastic stent as well as for a SEMS [18]
[65]
[79]
[82]
[83]
[84]
[85]. With respect to the performance of a biliary sphincterotomy prior to pancreatic
sphincterotomy, this should only be performed in selected cases, according to a small
RCT, mostly if biliary drainage is indicated or to facilitate access to the MPD [86].
In many but not all studies [51]
[54]
[55], pancreatic stenting was performed after stone fragmentation and removal. In prospective
series, technical success was reported in 92 % of attempted insertions of a first
stent [6]
[14]
[67]
[87]. The stenting duration averaged 10.6 months (range 3.2 – 23 months) in 18 series
totaling 811 patients [5]
[6]
[14]
[64]
[65]
[66]
[67]
[70]
[72]
[73]
[74]
[82]
[87]
[88]
[89]
[90]
[91]
[92].
Multiple stent designs have been proposed, including straight, S-shaped, and winged
stents, and stents with or without sideholes [93]
[94]. Few comparative studies have been reported; in a prospective study, stents with
large sideholes have been suggested to occlude less frequently compared to other types,
but only a minority of patients had CP [95]. With respect to stent diameter, CP patients treated with stents ≤ 8.5-Fr were 3.2
times more likely to be hospitalized for abdominal pain than those who had received
10-Fr stents in a retrospective study of 163 CP patients [96].
“On-demand” stent exchange consists of exchanging pancreatic stents when deemed necessary,
based on patient symptoms and/or additional investigations at 1 – 6-month intervals
(i. e., secretin-enhanced MRCP [S-MRCP] [66], abdominal ultrasound alone [68] or supplemented either with abdominal plain film [66] or with blood/urinary amylase measurements [69]). With this stent exchange policy, sepsis of pancreatic origin was reported in 15
(5.2 %) of 288 patients in four series [66]
[68]
[69]
[72] and surgery was required in two patients for pancreatic abscesses; this was reported
in the only series in which no additional investigations at regular intervals were
performed [72]. On the other hand, in 12 series (521 patients) with stent exchange scheduled at
shorter intervals, usually 3 months, septic complications have not been reported [5]
[14]
[65]
[67]
[70]
[73]
[74]
[88]
[90]
[92]
[97]
[98].
Compared with surgery, hospital stays and medical expenses were similar for patients
who had pancreatic stenting for less than 1 year (n = 19) but higher for those who
required longer pancreatic stenting (n = 15), in a retrospective study [97]. In that study, a single plastic stent was re-inserted if a stricture persisted
at pancreatography after stent removal within 3 months of the first ERCP.
With respect to FCSEMSs, stents of 6 – 10 mm in diameter have been used (Table 6 s); the mean stenting duration was 2 – 6 months and stents were removed uneventfully
in 108 (98 %) of 110 patients. (The stent-in-stent technique was used in the two remaining
patients and distal FCSEMS migration had occurred in 6 other patients.) Finally, in
a pilot study, a biodegradable non-covered self-expandable stent has provided clinical
success in 10 of 19 patients (53 %) who had no stricture resolution at least 6 months
after plastic stent insertion (median 10 months); adverse events were reported in
4 patients (21 %) [99].
Regarding complications with plastic stents, mild pancreatitis or worsening of pancreatic
pain
were most commonly reported at short term (average 6.2 %, range 4 % – 39 %) followed
by sepsis,
cholangitis, and post-sphincterotomy bleeding (average, 2.6 %, 2.3 %, and 1.5 %, respectively)
(Table 2 s). Severe pancreatitis has been rarely reported [73]. During follow-up, proximal and distal stent migration is reported in 2.7 % and
3.6 % of cases respectively, and bench tests using a column of water at a pressure
lower than that observed in patients with CP [89]
[100] have shown that almost all stents become obstructed at 3 months. However stent obstruction
does not correlate with symptoms [82]
[89]
[100]. Stent-induced ductal lesions were described in 18 % of patients (range 0 – 26 %)
and mortality was reported in 0.4 % (7/1620) (Table 2 s).
With SEMSs, stent migration (15 % – 46 %) and de novo strictures (16 % – 27 %) have
also been
reported and specific complications include severe pain (7 % – 20 %) leading to cholestasis
and FCSEMS removal (15 %) (Table 6 s).
ESGE recommends performance of endosonography-guided access and drainage of the MPD
only in tertiary centers after multidisciplinary discussion and preferably in a research
setting.
Strong recommendation, low quality evidence.
Potential indications for endosonography-guided access and drainage of the MPD include
patients with symptomatic MPD obstruction and failed conventional transpapillary drainage.
Briefly, the technique consists of puncturing the MPD through the gastric or duodenal
wall and advancing a guidewire into the MPD to proceed with transpapillary (rendezvous
technique) or transmural drainage using a plastic stent [50], or more recently a FCSEMS [101]. It is recognized as one of the most difficult techniques of EUS-guided therapy
[102].
Endosonography-guided access and drainage of the MPD has been reported in retrospective,
small, single-center studies [103]
[104]
[105]
[106]
[107] or larger multicenter studies (36 to 80 patients) [108]
[109]
[110] with a follow-up ranging from a few weeks up to 55 months (median 1 year). In all
these series, the annual inclusion rate per center was always below 4, illustrating
the rarity of the indications.
Immediate pain relief after successful endosonography-guided access and drainage of
the MPD has been reported in a majority of patients with obstructive CP (range 50 % – 100 %).
In the two series to date with available long-term follow-up, complete or major pain
relief was achieved in 70 % – 90 % of patients but the probability of remaining free
of pain dropped sharply over time [108]
[109].
Failed endosonography-guided access and drainage of the MPD occurs in approximately
10 % of cases and the incidence of moderate to severe complications also averages
10 % in the largest series, including severe pancreatitis, perforation, bleeding,
and hematoma [103]
[104]
[105]
[106]
[107]
[108]
[109]
[110]. No procedure-related mortality has been reported. Migration and occlusion of stents
necessitating endoscopic re-intervention frequently occur (20 % – 55 % of patients).
6 Pseudocyst management
Approximately one third of CP patients develop PPC during the course of their disease
[36]. PPCs should be differentiated from cystic neoplasms such as potentially malignant
mucinous neoplasms, particularly when they present for the first time.
Endoscopic therapy of PPCs consists of inserting a drain from the digestive lumen
into the PPC, through the digestive wall (“transmural drainage”), through the papilla
(“transpapillary drainage”), or using a combination of these routes. Transpapillary
PPC drainage is feasible only if the PPC communicates with the MPD, a situation detected
in approximately half of PPCs [111]. Technical and clinical success are usually defined, respectively, as the insertion
of at least one stent between the PPC and the digestive lumen (plus removal if indicated)
[112], and disappearance of symptoms with complete resolution of the PPC or a decrease
in size to less than 2 cm [113].
ESGE recommends treating CP-related pseudocysts if they are symptomatic (abdominal
pain, gastric outlet obstruction, early satiety, weight loss or jaundice) or present
with complications (infection, bleeding, rupture, or fistulization to adjacent hollow
structures).
Strong recommendation, low quality evidence.
Spontaneous regression of chronic PPCs is infrequent (0 to 27 %) and occurs most commonly
for PPCs smaller than 4 cm and/or located within the pancreas [114]
[115]. The indications for treatment listed above are commonly accepted. In asymptomatic
patients with a PPC compressing a major vessel, the risk – benefit ratio of any intervention
should be thoroughly analyzed; progressively enlarging collections are considered
a valid indication by some authors while others suggest that such patients be followed
until symptoms develop [116]
[117].
ESGE recommends endoscopic drainage over percutaneous or surgical treatment for uncomplicated
CP-related pseudocysts that are within endoscopic reach.
Strong recommendation, moderate quality evidence.
A meta-analysis of 7 retrospective studies (490 patients with various types of pancreatic
fluid collections [PFCs]) found that, compared with percutaneous drainage, endoscopic
drainage was associated with a higher clinical success rate, fewer re-interventions,
shorter hospital stay, and similar morbidity and recurrence rates [118]. Although percutaneous drainage has mostly been abandoned for the definitive treatment
of CP-related pseudocysts because it often results in an external fistula [119], it may be useful as an emergency measure (e. g., for infected PPC not accessible
to endoscopic drainage in a frail patient).
A meta-analysis (5 comparative studies including one RCT, 255 patients) found that,
compared with endoscopic therapy, surgery has a higher success rate (odds ratio [OR]
0.43, 95 %CI 0.20 – 0.95), but is associated with a longer length of hospital stay
and higher hospital costs as well as similar rates of morbidity (18.0 % vs. 11.5 %)
and recurrence (3.2 % vs. 3.1 %) [120]. A more recent multicenter prospective cohort study (71 patients) reported a similar
overall success rate and a shorter hospital stay for endoscopic therapy vs. surgery
[121].
ESGE suggests MRI with secretin-enhanced magnetic resonance cholangiopancreatography
(S-MRCP) for characterizing pancreatic fluid collections and the MPD anatomy before
endoscopic drainage of CP-related pseudocysts.
Weak recommendation, low quality evidence.
CT scan, MRI, and EUS allow the characterization of PFCs but the assessment of their
solid content is less precise with CT scan [122]
[123]
[124]; this is important only in subacute PFCs where necrotic debris may impede endoscopic
drainage. S-MRCP also allows diagnosis of MPD rupture. This has important consequences
for treatment planning: (i) in the absence of MPD rupture, endoscopic drainage can
be transmural only; (ii) if a partial MPD rupture is present, insertion of a stent
bridging the rupture (as opposed to below it) is associated with treatment success
[63]
[64]; and (iii) in the case of a complete MPD rupture (disconnected pancreatic duct syndrome),
removal of transmural stents is associated with PFC recurrence so that long-term indwelling
of transmural double-pigtail plastic stents should be considered [125]
[126]. Therefore, some centers perform imaging of the MPD by S-MRCP and/or ERCP prior
to drainage of and/or stent removal from PFCs.
Although ERCP is still considered to be the gold standard for the diagnosis of MPD
disruption, it presents limitations including an accuracy rate of approximately 75 %
and adverse events such as infection of a sterile PFC [127]
[128]. In small series, S-MRCP showed an accuracy of > 90 % for diagnosing MPD disruption
in patients with PFCs [123]
[129].
These imaging modalities have not been compared for the detection of pseudoaneurysms
close to pseudocysts, which is another important consideration when planning treatment.
ESGE suggests transpapillary drainage for small (< 50 mm) CP-related pseudocysts communicating
with the MPD in the head or body of the pancreas and transmural drainage for other
CP-related pseudocysts.
Weak recommendation, low quality evidence.
Compared with transmural drainage, transpapillary drainage provides similar success
with a
similar morbidity rate but fatal or surgical complications are less frequent (1/176
vs. 15/283; P = 0.007); however, transpapillary drainage as the only endoscopic therapy has been
performed for relatively smaller collections (generally ≤ 50 mm) than those managed
by transmural drainage alone or combined transpapillary and transmural drainage (Table 7 s). If transmural drainage is performed, the addition of transpapillary drainage seems
to add no benefit according to a meta-analysis of 9 non-randomized comparative studies
(7 including PPCs exclusively, 604 drainage procedures) [130]. No definitive conclusion can be drawn as the proportion of patients in whom a transpapillary
stent was inserted across as opposed to below a partial MPD disruption, a predictor
of success following transpapillary drainage [131]
[132], was not known. However, this factor may be of marginal importance as the insertion
of a stent across a partial MPD rupture succeeds in only 33 % – 67 % of the patients
[131]
[132].
ESGE recommends endosonography-guided over conventional access for the transmural
drainage of CP-related pseudocysts.
Strong recommendation, moderate quality evidence.
For the transmural drainage of PPCs, a systematic review (four studies, 229 patients)
found a higher technical success rate for EUS vs. conventional approach (relative
risk [RR] 12.38, 95 %CI 1.39 – 110.22) and no other significant differences (complications,
short and long-term clinical success) [133]. The difference in technical success was due to the presence of non-bulging collections
which account for approximately half of PFCs [111]; EUS guidance is the only option for transmural drainage in these cases.
ESGE suggests the use of double-pigtail plastic stents for the transmural drainage
of CP-related pseudocysts; a fully covered biliary SEMS can be considered if disconnected
pancreatic duct syndrome has been excluded and indwelling duration is expected to
be less than 6 weeks.
Weak recommendation, low quality evidence.
Plastic stents are generally used for the transmural drainage of PPCs. Three retrospective
studies examined the role of the number or diameter of plastic stents in a total of
307 patients; all studies included patients with various types of PFCs [134]
[135]
[136]. Double-pigtail stents of 7 – 10 Fr were used in the two most recent series as straight
stents may migrate and erode large vessels [135]. One study found that the insertion of a single stent was associated with failure
of endoscopic therapy (defined as severe procedure-related complication or need for
another treatment modality) [135] while two studies found no differences according to the number and diameter of plastic
stents [134]
[136].
Plastic stents and FCSEMSs have been compared for the transmural drainage of PPCs
in three meta-analyses [113]
[137]
[138]. The two most recent meta-analyses included comparative studies exclusively but
only approximately 10 % of patients had CP. These two meta-analyses reported: (i)
a similar success and a lower morbidity rate with FCSEMSs vs. plastic stents (OR 0.4,
95 %CI 0.21 – 0.73) (three studies, 301 patients) [138]; and (ii) a higher success rate with FCSEMSs vs. plastic stents (OR 5.35, 95 %CI
1.35 – 21.19) (morbidity analysis not reported) (two studies, 250 patients) [113]. Biliary FCSEMSs were used in most patients while lumen-apposing metal stents (LAMSs)
were used in 5 % [138] and 6 % of the patients [113]; in the studies that used a standard biliary FCSEMS, a double-pigtail plastic stent
was inserted through the FCSEMS to prevent its migration. The older meta-analysis
included non-comparative studies only and it found no differences between stents in
terms of success rates or morbidity [137].
A meta-analysis (6 retrospective studies, 504 patients) compared LAMSs with multiple
plastic stents for the treatment of PFCs but only 11 % of patients had a PPC; LAMSs
were associated with a higher clinical success rate (RR 2.70, 95 %CI 1.49 – 5.00)
and a lower morbidity rate (RR 0.39, 95 %CI, 0.18 – 0.84) [139]. A decision model analysis concluded that LAMSs were less cost-effective than plastic
stents [140].
ESGE recommends retrieval of transmural plastic stents at least 6 weeks after pancreatic
pseudocyst regression if MPD disruption has been excluded, and long-term indwelling
of transmural double-pigtail plastic stents in patients with disconnected pancreatic
duct syndrome.
Strong recommendation, low quality evidence.
Transmural plastic stents are generally removed at least 6 weeks after insertion as
a retrospective study showed that earlier plastic stent removal was associated with
treatment failure [135]. In an RCT (28 patients, 15 of whom had a CP-related PPC), PFCs recurred more frequently
in patients randomized to stent removal 2 months after drainage vs. no stent removal
(38 % vs 0); PFC recurrence tended to be associated with MPD rupture as identified
at S-MRCP (4/5 vs 2/9, P = 0.063) [126].
Disconnected pancreatic duct syndrome generally results from severe necrotizing pancreatitis
and has been discussed in a dedicated ESGE Guideline [141]. Retrospective studies have shown that long-term indwelling of double-pigtail transmural
plastic stents is effective, with PFC recurrence being uncommon and associated with
stent migration < 6 months after PFC resolution and MPD disruption at the pancreatic
head level [142]
[143].
With respect to LAMSs, an RCT reported LAMS-related severe adverse events in 6 of
12 patients (50 %), including bleeding, buried stent syndrome, and biliary stricture,
all occurring > 3 weeks after LAMS placement [144]
[145]. Stent-related morbidity dropped to levels similar to those observed with plastic
stents after the study protocol was changed to removal of LAMSs within 4 weeks. The
placement of a coaxial double-pigtail stent through the LAMS has also been proposed
to prevent delayed adverse events [146].
ESGE recommends the use of endosonographic guidance if the transmural route is selected
for draining CP-related pseudocysts in patients with portal hypertension. In the case
of arterial pseudoaneurysm close to a CP-related pseudocyst, ESGE recommends arterial
embolization prior to endoscopic drainage.
Strong recommendation, low quality evidence.
Extrahepatic portal hypertension develops during the course of CP in ≥ 15 % of patients
[147]. The only two series that reported the results of endoscopic drainage for PFCs in
patients with portal hypertension used EUS guidance; bleeding was reported in 1 of
26 patients (4 %) [148]
[149].
Pseudoaneurysms complicate the course of CP in 1 % – 10 % of patients, mostly those
with a PPC, and their rupture is associated with a high mortality [150]. Therefore, some authors recommend embolization of arterial pseudoaneurysms before
attempting endoscopic therapy of PPC close to pseudoaneurysms [151]. This strategy has not been tested but, in patients with bleeding pseudoaneurysms,
two retrospective series have reported a 94 % – 100 % mid-term success rate with arterial
embolization followed by endoscopic therapy of the PPC in a total of 40 patients [152]
[153].
7 Biliary strictures
ESGE suggests performance of an ERCP when a CP patient presents with a ≥ 4-week biliary
obstruction (jaundice, asymptomatic elevation of serum alkaline phosphatase [> 2 or
3 times the upper limit of normal values] and/or bilirubin) to achieve biliary decompression
by means of stent placement. If follow-up shows that the obstruction is caused by
a genuine fibrosis rather than transient inflammatory compression, endoscopic stent
treatment should be continued in order to dilate the stricture. After 1 year of unsuccessful
endotherapy, surgery should be considered.
Weak recommendation, low quality evidence.
Biliary strictures complicate the course of CP in 3 % – 23 % of patients, with studies
reporting a prevalence as high as 46 % [154]. Symptoms may be absent or include jaundice, cholangitis or choledocholithiasis.
Jaundice resolves spontaneously in 20 % – 50 % of patients within 1 month, because
of resolution of edema or of a PPC in the head of the pancreas but secondary biliary
cirrhosis is relatively frequent (7.3 % of 288 patients in a review of 11 studies)
[154]. Therefore, an asymptomatic elevation of serum alkaline phosphatase (> 2 or 3 times
the upper limit of normal values) and/or bilirubin for longer than 1 month are usually
accepted as an indication for bile duct drainage [155].
As underlined in Section 3, an underlying malignancy should be reasonably excluded.
A single retrospective study compared surgery vs. endoscopic therapy (multiple side-by-side
plastic stents or FCSEMS) for the treatment of CP-related biliary strictures in 39
patients [156]. Compared with surgery, endoscopic therapy presented a lower procedural morbidity
rate (21 % vs. 83 %) and a lower success rate at 2 years (15 % vs. 66 %). The success
rate was noticeably lower than in other studies (Table 8 s), including an RCT, maybe because incomplete stricture resolution at ERCP was considered
a failure. Outcomes were similar in patients who had surgery as a primary treatment
or following unsuccessful endoscopic therapy. The authors proposed to attempt endoscopic
therapy first in the absence of associated lesions (e. g., inflammatory cephalic mass),
and to evaluate its success after 12 months or three endoscopic procedures.
ESGE suggests the temporary insertion of multiple side-by-side plastic stents or of
a FCSEMS for treating CP-related benign biliary strictures.
Weak recommendation, moderate quality evidence.
The strategy of endoscopic therapy for benign biliary strictures is detailed in a
dedicated ESGE
Clinical Guideline [157]; it consists of temporarily dilating the stricture using multiple side-by-side plastic
stents or a FCSEMS (single plastic stents or uncovered SEMSs have long been abandoned
because of poor long-term results (Table 8 s) [158]. An RCT (60 CP patients) found that multiple plastic stents and covered SEMSs provided
similar success rates 2 years after stent removal (88.0 % vs. 90.9 %, respectively),
with similar treatment-related morbidity (23.3 % vs. 28.6 %, respectively) [159]. The stenting duration was 6 months in both groups. Various stenting durations have
not been compared in the literature (scheduled stenting durations with multiple plastic
stents and covered SEMSs have generally been for 1 year and for 6 – 12 months, respectively).
Short biliary strictures may respond better than longer ones to stenting, as suggested
by a small study (10 CP patients) [160].
ESGE recommends maintaining a registry of patients with biliary stents and recalling
them for stent removal or exchange.
Strong recommendation, low quality evidence.
Patient compliance with stent exchange may be poor, giving rise to potentially fatal
complications [161]
[162]. To prevent this, various recall systems have proven useful in pilot studies [163]
[164]. Removable FCSEMSs can result in better patient compliance since the number of ERCPs
is reduced to two. Of course, patient compliance with repeat interventions should
be ensured prior to endoscopic therapy and hepaticojejunostomy remains a valid option
for noncompliant patients or if the stricture does not respond to endoscopic therapy.
Disclaimer
The legal disclaimer for ESGE guidelines [165] applies to the current Guideline.