Introduction
The obesity epidemic continues to rise, with a doubling of global prevalence of obesity
from 6.4 % in 1980 to 12 % in 2008 driven by rising incidence in Asia [1]. Bariatric surgery has been shown to be more effective for weight loss than medical
therapy, with Roux-en-Y gastric bypass (RYGB) being considered the standard of care
over the past decade. While the rapid weight loss experienced after bariatric surgery
is desirable, it has been associated with changes in the composition of bile and the
subsequent development of gallstones [2]. This invariably leads to a proportion of patients developing choledocholithiasis
with complications ranging from asymptomatic elevations in liver enzymes to biliary
pancreatitis [3]. When such complications arise, endoscopic retrograde cholangiopancreatography (ERCP)
is often indicated.
Performance of ERCP in RYGB patients can be technically challenging for several reasons.
The Roux limb is intentionally created long to promote weight loss and typically exceeds
100 cm in length making the distance traversed by the endoscope significantly longer
than standard ERCP [4]. Furthermore, the native papilla is more challenging to cannulate as compared to
surgical bilio-enteric anastomosis due to the “upside-down” configuration and limited
availability of accessory instruments that are designed for long endoscopes [5]. Thus, the combination of the long enteral limb and native papilla in RYGB makes
for the most challenging ERCP of all post-surgical configurations. While several approaches
exist, laparoscopy-assisted ERCP (LA-ERCP) and enteroscopy-assisted ERCP (EA-ERCP)
are the most widely used modalities in RYGB patients [6].
LA-ERCP is performed by laparoscopically creating a gastrostomy through which a standard
duodenoscope can be advanced into the excluded stomach and duodenum [7]. Studies have shown this method to have high rates of success, however, it is resource
intensive and presents several technical risks and challenges [7]. This includes the logistical difficulties of coordinating surgeon, anesthetist
and gastroenterologist schedules [8] as well as a higher overall adverse event (AE) rate than standard ERCP due to the
laparoscopic approach [9].
EA-ERCP is performed utilizing overtube-based (single, double balloon or spiral) enteroscopy
where a special endoscope is passed orally through the Roux limb and the jejunostomy
up to the pancreaticobiliary limb to identify the papilla [10]. EA-ERCP has its limitations as well, including tortuosity of the endoscope trajectory,
unstable working platform, suboptimal accessory performance due to the small diameter
of the working channel and tangential view of the papilla [11].
While both LA-ERCP and EA-ERCP are considered safe and are widely used, their actual
success and AE rates have varied across studies. We conducted a meta-analysis comparing
success rates, procedural time and AEs of LA-ERCP and EA-ERCP in patients status post
RYGB with a native papilla.
Patients and methods
This meta-analysis was registered with the University Of York International Prospective
Register Of Systematic Reviews (PROSPERO, Registration number CRD42018114884). This
study was performed in accordance with the criteria established in the Preferred Reporting
Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
Search strategy and study selection
Studies were identified by performing a literature search of three electronic databases
(MEDLINE through PubMed, EMBASE and the Cochrane Library) with the last search performed
in October 2018. The detailed search strategy is outlined in supplementary [Table 1]. We attempted to identify additional studies by reviewing the reference list of
all included studies and manual search to retrieve other relevant articles that may
have been missed on the initial search strategy. Three investigators (F.A. and T.B.
and D.B.) screened all titles and abstracts for relevance to the study. The full text
of potentially eligible studies was subsequently reviewed by the three investigators
(F.A. and T.B. and D.B.). Disagreements were resolved by consensus or by consulting
with a third investigator (P.V.D.).
Table 1
Study characteristics for the enteroscopy-assisted endoscopic retrograde cholangiopancreatography
arm.
First author
|
Year
|
Country
|
Modality
|
Age (years)[1]
|
Male/Female
|
Papilla identification
|
Papilla cannulation
|
Therapeutic success
|
Procedure time (minutes)[1]
|
Ali
|
2018
|
USA
|
SE
|
22–75 (range)[2]
|
6/25[2]
|
24/28
|
22/22
|
22/22
|
189 (median)
|
Bukhari
|
2018
|
International
|
SBE/DBE
|
61.8 ± 11.5
|
12/18
|
21/30
|
18/30
|
NR
|
90.7 ± 34.9
|
Kashani
|
2018
|
USA
|
DBE
|
22–82 (range)
|
13/90
|
121/129
|
116/129
|
114/129
|
NR
|
De Koning
|
2016
|
Belgium
|
SBE/DBE
|
58 ± 2[2]
|
28/45[2]
|
14/24
|
14/24
|
14/24
|
NR
|
Trindade
|
2015
|
USA
|
SBE
|
28–80 (range)
|
NR
|
37/44
|
32/44
|
29/44
|
NR
|
Choi
|
2013
|
USA
|
DBE
|
56.1 ± 12.2
|
2/26
|
25/32
|
20/32
|
18/32
|
Mean: 101.2 range: (40–180)
|
Shah
|
2013
|
USA
|
SE/SBE/DBE
|
20–84 (range)[2]
|
36/93[2]
|
48/63
|
48/63
|
39/63
|
NR
|
Siddiqui
|
2013
|
USA
|
SBE
|
29–86 (range)
|
30/49
|
32/39
|
29/39
|
29/39
|
NR
|
Schreiner
|
2012
|
USA
|
SBE/DBE
|
53 (SD not reported)
|
1/31
|
23/32
|
19/32
|
19/32
|
106 (SD not reported)
|
Itoi
|
2011
|
Japan
|
SBE/DBE
|
55–88 (range)
|
12/3
|
15/15
|
15/15
|
15/15
|
NR
|
Saleem
|
2010
|
USA
|
SBE
|
NR
|
NR
|
7/15
|
7/15
|
NR
|
NR
|
Emmett
|
2007
|
USA
|
DBE
|
40–73 (range)[2]
|
7/7[2]
|
8/8
|
7/8
|
7/8
|
110 ± 37[2]
|
NR, not reported; SBE, single-balloon enteroscopy; DBE, double-balloon enteroscopy;
SE, spiral enteroscopy
1 Mean ± SD unless otherwise stated.
2 Numbers for overall study population, not reported for RYGB subgroup.
Inclusion and exclusion criteria
Inclusion criteria were: (1) retrospective or prospective, case series, case-control,
or cohort studies and clinical trials (including randomized clinical trials); (2)
studies involving patients who are status post RYGB requiring ERCP utilizing either
a LA or EA approach (single, double, “short” double balloon or spiral enteroscopy)
(3) studies reporting papilla identification rate, cannulation rate, therapeutic/diagnostic
success and procedural adverse events. Exclusion criteria were: (1) conference abstracts,
case reports and case series with less than 5 patients, (2) studies in languages other
than English (3) studies only involving patients with non-RYGB configurations (4)
reviews, commentaries, surveys, (5) and duplicate studies.
Data extraction
Data from each eligible study were extracted using a standardized data extraction
sheet. The extracted data included: (1) study authors, (2) year of publication, (3)
setting (location), (4) study period, (5) patient demographics (age, gender), (6)
number of patients/procedures, (7) indications for ERCP (8) ERCP approach (LA or EA)
(9) papilla identification rate, cannulation rate, therapeutic/diagnostic success
rate, (10) procedural adverse events and (11) procedural time.
Outcomes and definitions
The primary aim of this study was to conduct a meta-analysis comparing the papilla
identification rate, cannulation rate, therapeutic/diagnostic success rate of LA versus
EA-ERCP in patients who are status post RYGB with a native papilla. A secondary aim
was to compare the adverse event rates and procedural time associated with each modality.
Papilla identification was defined as successful visualization of the papilla of Vater
using the endoscope. Successful cannulation was defined as successful introduction
of a catheter into the desired duct. Therapeutic/diagnostic success was defined as
successful completion of the originally intended diagnostic or therapeutic indication
for ERCP as clinically indicated.
Assessment of methodologic quality
The quality of studies was assessed using the Newcastle Ottawa scale (NOS) [12]. Because the majority of included studies were case series, we utilized a modified
version of the NOS appropriate for our analysis. This tool removes from the NOS the
items that relate to comparability between two arms and retains items that assess
representation and selection of cases as well as ascertainment of exposure and outcome.
A point is assigned to each component of the modified scale, with the highest possible
score being 6/6. Studies were considered to be high quality if they scored 6/6, moderate
quality if they scored 5/6 and low quality if they scored 4/6 or less. The quality
of all studies was assessed by three investigators (F.A, T.B., D.B.). Egger’s regression
test was used to assess for publication bias.
Statistical analysis
Pooled rates were calculated utilizing a random effects model and the Freeman-Tukey
arcsine transformation was used [13]. The Cochran Q test and I2 were used to assess heterogeneity of included studies. I2 values < 25 %, 25 % to 50 % and > 50 % were considered to represent low, moderate,
and high heterogeneity, respectively. P < 0.05 was considered significant and all tests were two-tailed. The study was performed
in accordance with the PRISMA recommendations for reporting systematic reviews and
meta-analyses. Analysis was conducted using Stata, version 15 (Stata Corp, College
Station, Texas, United States).
Results
Search results
The flow diagram for study selection is depicted in [Fig.1]. Overall, 3859 studies were identified using our search strategy, of which 1615
were duplicates. Of the remaining 2244 studies after duplicate removal, 2134 were
excluded after screening titles and abstracts. Full text review was then performed
on 110 studies using the predefined inclusion and exclusion criteria, after which
26 studies were retained. Twenty two were case series (1 prospective, 21 retrospective)
[5]
[8]
[10]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31], two were retrospective cohort studies comparing balloon EA to LA-ERCP [6]
[32], and two were retrospective cohort studies comparing EA or LA-ERCP to other approaches
where data on laparoscopy or balloon enteroscopy was extracted and used for the pooled
analysis [33]
[34]. Studies were published between 2007 and 2018. Eight studies were multi-center studies,
two of which were conducted internationally. Eighteen studies were conducted in the
United States, four in Europe, one in Brazil, one in Japan. Data from the laparoscopy
arm were not used from two cohort studies [6]
[32] because the same data were included in the multicenter study by Abbas et al [8].
Fig. 1 PRISMA flow diagram. From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA
Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses:
The PRISMA Statement. PLoS Med 6(7): e1000097. doi:10.1371/journal.pmed1000097
Patient population and study characteristics
A total of 427 patients underwent 459 EA-ERCPs, and 882 patients underwent 886 LA-ERCPs.
Study characteristics are summarized in [Table 1] and [Table 2].
Table 2
Study characteristics for laparoscopy-assisted endoscopic retrograde cholangiopancreatography
arm.
First author
|
Year
|
Country
|
Age (years)[1]
|
Male/Female
|
Papilla identification
|
Papilla cannulation
|
Therapeutic success
|
Procedure time (minutes)[1]
[2]
|
Abbas
|
2018
|
International
|
51 (IQR, 43–61)
|
91/488
|
573/579
|
567/579
|
567/579
|
152 minutes (IQR, 109–210)
|
Kedia
|
2018
|
USA
|
55 (33–80)
|
7/36
|
42/43
|
42/43
|
42/43
|
184
|
Yancey
|
2018
|
USA
|
55.8 (29–67)
|
NR
|
15/16
|
15/16
|
15/16
|
NR
|
Frederiksen
|
2017
|
Denmark
|
Median: 46 (25–65)
|
4/24
|
31/31
|
31/31
|
31/31
|
NR
|
Lim
|
2017
|
USA
|
50.3 ± 9.8
|
0/35
|
35/35
|
35/35
|
35/35
|
NR
|
Bowman
|
2016
|
USA
|
48.5 (25–71)
|
4/11
|
16/16
|
16/16
|
16/16
|
NR
|
Paranandi
|
2016
|
UK
|
Median: 44
|
0/7
|
7/7
|
7/7
|
7/7
|
96
|
Grimes
|
2015
|
USA
|
48.5 (23–69)
|
36/2
|
36/38
|
36/38
|
NR
|
265
|
Snauwaert
|
2015
|
Belgium
|
Median: 57 (26–79)
|
5/18
|
23/23
|
23/23
|
23/23
|
NR
|
Falcao
|
2012
|
Brazil
|
35.3 (27–52)
|
4/19
|
23/23
|
23/23
|
23/23
|
92.69
|
Saleem
|
2012
|
USA
|
51 ± 12.58 (25–70)
|
3/12
|
15/15
|
15/15
|
15/15
|
45
|
Bertin
|
2011
|
USA
|
NR
|
NR
|
22/22
|
22/22
|
NR
|
236
|
Gutierrez
|
2009
|
USA
|
46 (27–72)
|
4/24
|
28/28
|
28/28
|
NR
|
200
|
Lopes
|
2009
|
USA
|
40 (19–55)
|
1/9
|
9/10
|
9/10
|
9/10
|
89
|
NR, not reported
1 Mean ± SD, (range) unless otherwise stated.
2 Numbers for overall procedure time (endoscopic retrograde cholangiopancreatography + laparoscopy).
Indications and adverse events
Detailed data on procedural indications were reported in nine of 12 EA-ERCP studies
and in 13 of 14 LA-ERCP studies. The most common procedural indication in the LA-ERCP
arm was choledocholithiasis in 48 % of cases (408/847), compared to 74 % (280/380)
in EA-ERCP. Adverse events were reported by 10/12 in the EA-ERCP arm and all studies
in the LA-ERCP arm. In the EA-ERCP arm, the most commonly reported AE was pancreatitis
in 5 % (23/459) of cases. Only one study [27] described the severity of pancreatitis where one of five pancreatitis cases was
considered to be severe. Small bowel perforation was uncommon and was reported in
1 % (6/459) of cases. Death was rare with only one case reported by Shah et al. [27] in the EA-ERCP arm where a patient developed an embolic stroke post-procedurally
and decision was made to withdraw care.
In the LA-ERCP arm, 12 of 14 studies classified AEs into either ERCP or laparoscopy-related.
The most common ERCP-related AE was pancreatitis, reported in 6 % (53/847) of cases.
Perforation was again uncommon and was reported in 1 % (10/847) of cases. The most
common laparoscopy-related adverse event was infection, reported in 5 % (44/847) of
cases, the majority of which were localized in nature. There were no reports of death,
however there was one reported case of tension pneumothorax in the study by Lopes
et al. [23] which was caused by an indwelling percutaneous transhepatic cholangiogram (PTC)
catheter crossing the diaphragm, however this was promptly recognized and managed
with chest tube insertion. [Table 3] and [Table 4] summarize the indications and AEs for both EA and LA-ERCP.
Table 3
Complications and procedural indications for enteroscopy-assisted endoscopic retrograde
cholangiopancreatography arm.
First author
|
Year
|
Indications (n)
|
Complications (n)
|
Ali
|
2018
|
Choledocholithiasis (14) Biliary stricture (8) Sphincter of Oddi dysfunction (5) Stent placement/removal (4) Pancreatitis (1) Type III choledochocele (1) Bile leak (1)
|
None
|
Bukhari
|
2018
|
Choledocholithiasis (30) Benign biliary stricture (5) Sphincter of Oddi dysfunction (2) Cholangitis (2)
|
Pancreatitis (1) Cholangitis (1) Small bowel perforation (1)
|
Kashani
|
2018
|
Sphincter of Oddi dysfunction (66) Choledocholithiasis (26) Pancreatitis (9) Biliary stricture (8) Bile leak (8) Cholangitis (6) Abnormal liver tests (5) Recurrent liver abscess (1)
|
Pancreatitis (10) Small bowel perforation (2) Cholangitis (1)
|
De Koning
|
2016
|
NR
|
NR
|
Trindade
|
2015
|
Choledocholithiasis (29) Cholangitis (10) Abnormal liver tests (9) Benign biliary stricture (4) Bile leak (4)
|
Pancreatitis (3)
|
Choi
|
2013
|
Choledocholithiasis (16) Sphincter of Oddi dysfunction (6) Biliary stricture (4) Bile leak (2)
|
Pancreatitis (1)
|
Shah
|
2013
|
Abnormal liver enzymes + dilated bile ducts (62) Dilated bile ducts on non-invasive imaging (21) Cholangitis (20) Abnormal liver enzymes (11) Pancreatitis (8) Other (7)
|
Mild pancreatitis (4) Severe pancreatitis (1) Bleeding (1) Abdominal pain leading to re-admission (3) Throat pain requiring physician contact (4) Perforation (2) Death (1)
|
Siddiqui
|
2013
|
Choledocholithiasis (48) Biliary stricture (18) Stent removal (5), Sphincter of Oddi dysfunction (3), Bile leak (3) Pancreatic stricture (2)
|
Abdominal pain (3) Pancreatitis (3) Post-procedural bleeding (1)
|
Schreiner
|
2012
|
“Preprocedure indications for ERCP included (1) dilation of the pancreaticobiliary
tree in the setting of laboratory abnormalities or clinical symptoms; (2) stones seen
on imaging; and/or (3) abdominal pain with abnormal laboratory test results suggesting
biliopancreatic origin.”
|
Pancreatitis (1)
|
Itoi
|
2011
|
Choledocholithiasis (15)
|
None
|
Saleem
|
2010
|
“Cholestasis, acute cholangitis, recurrent primary sclerosing cholangitis with strictures,
and choledocholithiasis.”
|
None
|
Emmett
|
2007
|
Repeat procedure (6) Recurrent pancreaticobiliary pain (5) Abnormal liver tests (4) Cholangitis (2) Chronic pancreatitis (2) Acute pancreatitis (1)
|
None
|
Complications and indications reported for overall study population when data on specific
RYGB patients are not reported in individual studies. NR, not reported.
Table 4
Complications and procedural indications for the laparoscopy-assisted assisted endoscopic
retrograde cholangiopancreatography arm.
First author
|
Year
|
Indications (n)
|
Complications (n)
|
Conversion to open (n)
|
Abbas
|
2018
|
Choledocholithiasis (254) Papillary stenosis (102) Dilated duct (75) Pancreatitis (56) Abnormal liver function tests (46) Bile duct stricture (20) Post cholecystectomy pain (10) Abdominal pain (9) Bile leak (7) Ampullary lesion (7) Biliary stent removal (3) Dilated pancreatic duct (3) Abnormal intraoperative cholangiogram (2) Pancreatic duct stone (1)
|
Laparoscopy-related Other postoperative infections (24) Laparoscopy-related bleeding (10) Gastric site leak (7) Gastric tube site infection (7) Postoperative respiratory adverse events (6) Postoperative cardiovascular adverse events (4) Laparoscopy-related perforation (3) Other laparoscopic related (11) ERCP-related Pancreatitis (43) Cholangitis (6) ERCP-related bleeding (3) ERCP-related perforation (2) Stent migration (1)
|
29
|
Kedia
|
2018
|
Choledocholithiasis (54) Papillary stenosis (5)
|
ERCP-related Perforation (2) Laparoscopy-related Intraperitoneal abscess (2) Wound dehiscence (1) Bleeding (1) Abdominal wall seroma (1) Cellulitis (1)
|
4
|
Yancey
|
2018
|
“Choledocholithiasis, cholangitis, and radiographic or clinical evidence of common
bile duct (CBD) obstruction.”
|
ERCP-related Necrotizing pancreatitis (1) Laparoscopy-related None
|
1
|
Frederiksen
|
2017
|
Choledocholithiasis (31)
|
ERCP-related Perforation (2) Pancreatitis (2) Laparoscopy-related Intraperitoneal abscess (3) Abdominal hematoma (3) Wound dehiscence (1)
|
2
|
Lim
|
2017
|
Sphincter of Oddi dysfunction (35) Choledocholithiasis (10) Biliary stricture (2) Pseudocyst (1) Cystic duct leak (1) Pancreatic leak (1)
|
ERCP-related Pancreatitis (3) Laparoscopy-related None
|
NR
|
Bowman
|
2016
|
Choledocholithiasis (5) Recurrent pancreatitis (3) Ampullary mass (1) Sphincter of Oddi dysfunction (1) Biliary stricture (1)
|
ERCP-related None Laparoscopy-related Abdominal abscess (1) Incisional hernia (1) Wound dehiscence (1)
|
1
|
Paranandi
|
2016
|
Choledocholithiasis (5) Papillary fibrosis (1) Retained biliary stent (1)
|
ERCP-related Pancreatitis (1) Laparoscopy-related Port-site infection (1)
|
0
|
Grimes
|
2015
|
Chronic abdominal pain/sphincter of Oddi dysfunction/pancreatic duct stenosis/chronic
pancreatitis (80) Choledocholithiasis (5)
|
ERCP-related Duodenal perforation (2) Laparoscopy related G-tube site infection (4) Posterior gastric wall injury (4) Persistent gastro-cutaneous fistula (2) Bleeding requiring transfusion (2) Pneumoperitoneum (2) Perforation (1) Abdominal wall hematoma (1)
|
1
|
Snauwaert
|
2015
|
Choledocholithiasis (16) Biliary pain (4) Jaundice (3)
|
None
|
2
|
Falcao
|
2012
|
Choledocholithiasis (14) Cholecystitis (6) Obstructive jaundice (3)
|
ERCP-related Pancreatitis (1) Laparoscopy related None
|
0
|
Saleem
|
2012
|
Sphincter of Oddi dysfunction (9) Choledocholithiasis (5) Recurrent acute pancreatitis (1)
|
None
|
0
|
Bertin
|
2011
|
Sphincter of Oddi dysfunction (18) Recurrent acute pancreatitis (4)
|
ERCP-related Perforation (1) Laparoscopy related Abdominal hematoma (1) Bile leak (1)
|
1
|
Gutierrez
|
2009
|
Sphincter of Oddi dysfunction (13) Pancreatitis (6) Choledocholithiasis (5) Cholangitis (3) Pancreatic mass evaluation (2) Gastrointestinal bleed (2) Bile leak (1)
|
ERCP-related Perforation (1) Laparoscopy-related Gastrostomy site leak (2) Gastrostomy site infection (1)
|
1
|
Lopes
|
2009
|
Choledocholithiasis (4) Biliary stricture (3) Sphincter of Oddi dysfunction (3)
|
ERCP-related Pancreatitis (2) Laparoscopy-related Tension pneumothorax (1)
|
0
|
Complications and indications reported for overall study population when data on specific
RYGB patients are not reported in individual studies.
NR, not reported.
Quality assessment
Risk of bias in the 26 studies was evaluated according to the modified Newcastle-Ottawa
assessment scale and is shown in Supplementary Table 2. Overall, 20 of 26 studies (77 %) were found to be of moderate to high quality and
six of 26 studies (33 %) were found to be low quality. Most quality issues were related
to a lack of adequate description of the characteristics and outcomes of the RYGB
cohort in studies that included patients with a broad variety of post-surgical anatomy.
It is important to note that majority of included studies were retrospective case
series, which inherently affects overall study quality.
Meta-analysis results
The pooled results of papilla identification, papilla cannulation and therapeutic
success rates are summarized in [Table 5].
Table 5
Summary of pooled outcomes for enteroscopy-assisted compared to laparoscopy-assisted
endoscopic retrograde cholangiopancreatography.
|
Papilla identification
|
Papilla cannulation
|
Therapeutic success
|
|
Pooled rate (%)
|
95 % CI
|
Pooled rate (%)
|
95 % CI
|
Pooled rate (%)
|
95 % CI
|
Enteroscopy-assisted ERCP
|
80.0
|
71.3–87.4
|
73.0
|
63.6–81.5
|
73.2
|
62.5–82.6
|
Single-balloon enteroscopy
|
78.5
|
56.6–94.1
|
75.3
|
53.4–91.9
|
77.2
|
48.9–96.1
|
Double-balloon enteroscopy
|
80.4
|
71.6–88.0
|
72.3
|
60.0–83.1
|
65.8
|
54.2–76.5
|
Spiral enteroscopy
|
78.9
|
65.8–89.5
|
89.4
|
51.3–98.8
|
85.5
|
34.1–97.3
|
Laparoscopy-assisted ERCP
|
98.5
|
97.6–99.2
|
97.8
|
96.7–98.7
|
97.9
|
96.7–98.7
|
CI, confidence interval; ERCP, endoscopic retrograde cholangiopancreatography.
Papilla identification
All studies in the EA arm and the LA arm reported papilla identification rates ([Fig. 2], [Fig. 3]). The pooled rate of papilla identification in LA-ERCP was 98.5 % (95 % confidence
interval [CI]: 97.6–99.2 %) with no heterogeneity identified in the pooled analysis
(I2 = 0.0 %). This was higher than the pooled rate of papilla identification in EA-ERCP
at 80.0 % (95 % CI: 71.3–87.4 %) with studies demonstrating a high degree of heterogeneity
(I2: 77.5 %). Among the EA-ERCP studies, four reported papilla identification rates utilizing
single-balloon enteroscopy with a pooled rate of 78.5 % (95 % CI: 56.6–94.1 %), 3
studies reported papilla identification rates utilizing double-balloon enteroscopy
with a pooled rate of 80.4 % (95 % CI: 71.6–88.0 %) and 2 studies reported papilla
identification rates utilizing spiral enteroscopy with a pooled rate of 78.9 % (95 %
CI: 65.8–89.5 %). There was no evidence of substantial publication bias based on visual
inspection of the funnel plot and Egger’s regression test (Supplementary Fig. 1a, 1b).
Fig. 2 Pooled papilla identification rate of enteroscopy-assisted ERCP arm.
Fig. 3 Pooled papilla identification rate of laparoscopy-assisted ERCP arm.
Papilla cannulation
All studies in the EA-ERCP arm and the LA-ERCP arm reported papilla cannulation rates
([Fig. 4], [Fig. 5]). The pooled rate of papilla cannulation LA-ERCP was 97.8 % (95 % confidence interval
[CI]: 96.7–98.7 %) with no heterogeneity identified in the pooled analysis (I2 = 0.0 %). This was higher than the pooled rate of papilla cannulation in EA-ERCP
at 73.0 % (95 % CI: 63.6–81.5 %) with studies demonstrating a high degree of heterogeneity
(I2: 77.4 %). Among EA-ERCP studies, four reported papilla cannulation rates utilizing
single-balloon enteroscopy with a pooled rate of 75.3 % (95 % CI: 53.4–91.9 %), three
studies reported papilla identification rates utilizing double-balloon enteroscopy
with a pooled rate of 72.3 % (95 % CI: 60.0–83.1 %) and two studies reported papilla
cannulation rates utilizing spiral enteroscopy with a pooled rate of 89.4 % (95 %
CI: 51.3–98.8 %). There was no evidence of substantial publication bias based on visual
inspection of the funnel plot and Egger’s regression test (Supplementary Fig. 2a, Supplementary Fig. 2b).
Fig. 4 Pooled papilla cannulation rate of enteroscopy-assisted ERCP arm.
Fig. 5 Pooled papilla cannulation rate of laparoscopy-assisted ERCP arm.
Therapeutic success
Ten studies in the EA-ERCP arm and 11 studies in the LA-ERCP arm reported therapeutic
success rates ([Fig. 6], [Fig. 7]). The pooled rate of therapeutic success in LA-ERCP was 97.9 % (95 % confidence
interval [CI]: 96.7–98.7 %) with no heterogeneity identified in the pooled analysis
(I2 = 0.0 %). This was higher than the pooled rate of therapeutic success in EA-ERCP
at 73.2 % (95 % CI: 62.5–82.6 %) with studies demonstrating a high degree of heterogeneity
(I2: 80.2 %). Among EA-ERCP studies, three studies reported therapeutic success rates
utilizing single-balloon enteroscopy with a pooled rate of 77.2 % (95 % CI: 48.9–96.1 %),
three studies reported therapeutic success rates utilizing double-balloon enteroscopy
with a pooled rate of 65.8 % (95 % CI: 54.2–76.5 %) and two studies reported therapeutic
success rates utilizing spiral enteroscopy with a pooled rate of 85.5 % (95 % CI:
34.1–97.3 %). There was no evidence of substantial publication bias based on visual
inspection of the funnel plot and Egger’s regression test (Supplementary Fig. 3a, Supplementary Fig. 3b).
Fig. 6 Pooled therapeutic success rate of enteroscopy-assisted ERCP arm.
Fig. 7 Pooled therapeutic success rate of laparoscopy-assisted ERCP arm.
Adverse events
Ten of 12 studies in the EA-ERCP arm and all studies in the LA-ERCP arm reported post-procedural
adverse events ([Fig. 8], [Fig. 9]). Overall AE rates for the LA arm were calculated as a composite of ERCP-related
adverse events, laparoscopy-related adverse events and conversion to open surgery.
The pooled rate of overall AEs in LA-ERCP was 19.0 % (95 % CI: 12.6–26.4 %) with studies
demonstrating a high degree of heterogeneity (I2: 74.1 %). This was higher than the pooled rate of adverse events in EA-ERCP at 6.5 %
(95 % CI: 3.9–9.6 %) with studies demonstrating low heterogeneity (I2: 16.2 %). Twelve of 14 LA-ERCP studies reported separate ERCP-related AEs, and the
pooled ERCP-related AE rate was 8 % (95 % CI: 5.4–10.9 %) with low heterogeneity (I2: 15.6 %). There was no evidence of substantial publication bias based on visual inspection
of the funnel plot (Supplementary Fig. 4a, Supplementary Fig. 4b).
Fig. 8 Pooled adverse event rate of enteroscopy-assisted ERCP arm.
Fig. 9 Pooled adverse event rate of laparoscopy-assisted ERCP arm.
Procedure duration
Four of 12 studies in the EA arm and nine of 14 studies in the LA arm reported procedural
time in minutes. Procedural time for the LA arm was calculated as a composite of laparoscopy
and ERCP time since only one study reported separate laparoscopy and ERCP times. Pooled
mean procedure time for LA-ERCP was 158.4 minutes (SD ± 20). This was higher than
the mean pooled procedure time for EA-ERCP at 100.5 minutes (SD ± 19.2),
Discussion
With the rise of the obesity epidemic and the popularity of bariatric surgery, patients
with RYGB requiring ERCP are increasingly encountered in clinical practice. While
several approaches exist, LA-ERCP and EA-ERCP are the most widely used modalities
in RYGB patients [6]. LA-ERCP is performed by laparoscopically creating a gastrostomy through which a
standard duodenoscope can be advanced into the excluded stomach and duodenum [7]. ERCP is then carried out in standard fashion using standard accessories. EA-ERCP
is performed utilizing overtube-based (single balloon, double balloon or spiral) enteroscopy,
where the endoscope/overtube combination is passed orally via the Roux limb. Once
the enteroenterostomy is reached, the pancreaticobiliary limb is accessed in retrograde
fashion in order to reach the papilla [10]. Once the papilla is identified, ERCP is carried out via the forward-viewing, 200-cm-long
enteroscope (therapeutic channel 2.8 mm) using dedicated “long” accessories. A short
version of the double-balloon enteroscope using standard accessories has been investigated
but has only recently become available in the United States.
Our meta-analysis suggests that LA-ERCP has significantly higher overall success rates
(therapeutic success 97.9 %; 95 % CI: 96.7–98.7 %) than EA-ERCP (therapeutic success
73.2 %; 95% CI: 62.5–82.6 %) at the expense of a higher adverse event rate and longer
procedural time. We find that all technical components of ERCP (papilla identification,
cannulation and therapeutic success) are more successful with LA-ERCP than EA-ERCP
([Fig. 1], [Fig. 2], [Fig. 3]). The higher papilla identification rate may be explained by the shorter distance
the endoscope must traverse to reach the papilla and use of a standard side-viewing
duodenoscope in correct orientation with LA-ERCP. Along the same lines, the higher
papilla cannulation rates may be explained by the combination of factors mentioned
previously as well as the availability of standard ERCP accessories for use compared
to EA-ERCP. Ultimately, the pooled therapeutic success rate with LA-ERCP was remarkably
high and is consistent with that of regular ERCP, highlighting that the main limitation
of ERCP in RYGB patients is the ability to reach the papilla and then having adequate
accessories for use. In contrast, EA-ERCP showed a more modest but heterogenous pooled
success rate, with individual studies reporting success rates ranging from 56–98 %.
Notably, this heterogeneity persisted in subgroup analyses separately assessing different
enteroscopy approaches and is unlikely to be attributed solely to the enteroscopy
modality utilized. We excluded case series with fewer than five patients to decrease
the effect variable operator experience may have on the pooled outcomes, however,
residual effects cannot be excluded and may also partially explain the noted heterogeneity.
In line with prior analyses, we found a higher overall AE rate with LA-ERCP. This
can mainly be attributed to infectious and bleeding AEs related to the laparoscopic
approach of the procedure rather than ERCP. This is supported by our finding that
pooled rates of ERCP-related AEs were similar between the two approaches. While many
reported laparoscopy-related AEs were self-limited, some were quite serious in nature
including bleeding requiring transfusion, intra-abdominal abscess formation and tension
pneumothorax. This supports an individualized approach that considers patient comorbidities
and characteristics when choosing the most appropriate modality for ERCP.
Expectedly, we also note a shorter mean procedural time with EA-ERCP compared to LA-ERCP.
This is readily explained by the additional time required for laparoscopic access
to the remnant stomach in LA-ERCP. While the time savings of using EA-ERCP may seem
attractive, particularly for busy endoscopy units, this must be weighed against the
potential for lower overall ERCP success rates compared to LA-ERCP. Notably, a failed
attempt at EA-ERCP may inevitably lead to additional interventions such as LA-ERCP
or percutaneous transhepatic biliary drainage, each with its associated cost, time,
and possible AEs. Interpreting pooled results of procedural time must be with caution
however; only four of 12 EA-ERCP studies reported procedural time with heterogenous
underlying enteroscopy modalities and operator experience.
Our study has several strengths. While the higher therapeutic success rate noted with
LA-ERCP (97.9 %) compared to EA-ERCP (73.2 %) is in line with other systematic reviews
on the topic, we attempted to address some of the limitations of other analyses. Recently,
Aiolfi et al. reported a pooled LA-ERCP success rate of 99 % in patients with RYGB
anatomy [35], however this was limited by the lack of a clear definition for “ERCP success.”
We utilized strict definitions and we calculated detailed pooled outcomes for papilla
identification, cannulation and therapeutic success, respectively. Ponte-Neto et al.
recently compared LA-ERCP to balloon-based ERCP, with similar findings to our analysis
[36], however, the power of the pooled rate of LA-ERCP success might have been limited
by lack of inclusion of the largest multi-center study to date by Abbas et al. which
reported outcomes of LA-ERCP in 567 patients from 34 centers [8]. Additionally, The Ponte-Neto analysis limited the enteroscopy arm to balloon-based
enteroscopy while we also include studies describing rotational spiral enteroscopy.
Finally, by focusing on patients with bariatric RYGB anatomy we aimed to reduce heterogeneity
attributed to variable Roux limb length and presence of bilio-enteric anastomoses
that may have affected other analyses that include patients with different anatomic
variations such as Billroth I, Billroth II, Roux-en-Y hepaticojejunostomy or pancreaticoduodenectomy
[37]
[38].
Our study has several limitations. The EA-ERCP arm included different enteroscopy
modalities including single-balloon, double-balloon and spiral enteroscopy, which
may contribute to the noted high degree of heterogeneity of our pooled outcomes, however,
we attempted to address this by performing subgroup analyses when data were available.
Aside from one prospective case series, the remaining included studies had a retrospective
design with the inherent limitations of the retrospective approach. This, however,
highlights the limitations of available literature rather than the individual analysis.
As noted above and inherent to the meta-analytic technique, not every study reported
all outcomes of interest and as such, not all studies were included in subgroup analyses
when this was the case. Finally, other emerging endoscopic approaches exist which
we did not include in our analysis such as endoscopic ultrasound-directed transgastric
ERCP (EDGE). Expertise in EDGE remains limited to select centers, but data suggests
success rates comparable to LA-ERCP [39].
Ultimately, choice of the optimal ERCP modality in patients with RYGB is dependent
on multiple factors including patient preference, indications for ERCP, clinical importance
of preserving the integrity of the RYGB, local expertise, and device availability.
Based on our current understanding and available data, we suggest the following approach.
LA-ERCP can be considered the preferred modality when a single ERCP is likely to address
the clinical problem (e. g. choledocholithiasis, papillary stenosis) or when cholecystectomy
is indicated thus allowing the ERCP and the cholecystectomy to be carried out in the
same session. EDGE may be considered when preserving the integrity of the RYGB is
of no clinical significance (e. g. pancreatic head mass likely to be cancer in need
of sampling and stenting) or when multiple ERCPs are anticipated (e. g. endoscopic
therapy for benign biliary stricture or chronic pancreatitis). Considering the significantly
lower success rates, EA-ERCP should be reserved for situations in which it is the
only available modality or for patients not willing to undergo LA-ERCP or EDGE.
Conclusion
In summary, this meta-analysis suggests that LA-ERCP should be considered a first-line
approach for ERCP in patients with RYGB due to its higher overall success rate compared
to EA-ERCP. However, LA-ERCP is associated with a higher burden of AEs and longer
procedural time. In the absence of high-quality comparative studies, the choice between
LA-ERCP and EA-ERCP must be made on a case-by-case basis that takes patient, facility,
and endoscopist characteristics into account.