Introduction
Distal malignant biliary obstruction (DMBO) is a frequent condition associated with
biliopancreatic cancers. While endoscopic retrograde cholangiopancreatography (ERCP)
is the current gold standard for biliary drainage [1 ], patients with DMBO may face a higher rate of difficult biliary cannulation (DBC),
increasing the risk of both procedural failure, and adverse events (AEs) [2 ]. When standard cannulation fails, advanced cannulation techniques, such as the double-guidewire
technique, needle-knife precut papillotomy or fistulotomy, and transpancreatic sphincterotomy,
should be considered as alternative methods to achieve bile duct access [3 ]
[4 ].
Endoscopic ultrasound-guided biliary drainage (EUS-BD) can be considered an alternative
approach in this context because of its favorable profile in terms of technical success,
procedural times, and safety [5 ]. Moreover, since dedicated single-step cautery-enhanced lumen-apposing metal stents
(EC-LAMSs) have become available for the creation of an EUS-guided transmural choledochoduodenostomy
(EUS-CDS), the complexity of the procedure has been reduced, with less need for device
exchanges [6 ]
[7 ]. Consequently, in several studies, EUS-BD has been reported increasingly earlier
in the work-flow, to the point of being proposed as a primary drainage option for
patients with DMBO [8 ]
[9 ].
The aim of our study was to investigate how the implementation of EUS-CDS with a LAMS
impacted the management and outcomes of patients with DMBO in a real-life setting,
with a specific focus on patients with DBC.
Methods
This was an international multicenter retrospective study performed in four European
tertiary institutions. The start of the study period varied among the different centers,
based on the availability of the LAMS (Table 1s , see online-only Supplementary material). Data were recorded up to December 2023.
Consecutive patients aged ≥18 years, with a dilated extrahepatic bile duct (>12 mm)
on computed tomography, magnetic resonance imaging, or EUS, undergoing ERCP for DMBO
were considered for inclusion. The full inclusion and exclusion criteria are detailed
in [Table 1 ].
Table 1 The full study inclusion and exclusion criteria.
Inclusion criteria
Exclusion criteria
CBD, common bile duct; ERCP, endoscopic retrograde cholangiopancreatography; EUS-CDS,
endoscopic ultrasound-guided choledochoduodenostomy.
Age ≥18 years
Prior ERCP attempts
Informed consent given
Treated with EUS-CDS as the primary approach
Histologically or radiologically confirmed distal malignant bile duct neoplasia (2
cm away from the portal hilum)
Surgically altered anatomy with prior Billroth II or Roux-en-Y reconstruction
Radiologically confirmed distal CBD obstruction, with dilated CBD ≥12 mm in diameter
Tumor involvement of the pylorus or duodenum, either preventing the papilla being
reached or causing gastric outlet obstruction symptoms
Native papilla
History of bleeding disorder or use of antithrombotic agent.
Adequate follow-up (>3 months) not available
The methods of our study were based on the Strengthening the Reporting of Observational
Studies in Epidemiology (STROBE) recommendations [10 ]. Institutional review board approval was obtained. All authors had access to the
study data, and reviewed and approved the final manuscript.
Endoscopic procedure
All procedures were performed with the patient under deep sedation or general anesthesia,
and in the left-lateral or prone position. Once the papilla had been reached, a standard
cannulation technique using a sphincterotome and guidewire was firstly attempted.
If DBC was identified, defined as any situation in which deep cannulation was not
achieved with the standard cannulation technique, advanced techniques such as the
double-guidewire technique, pancreatic duct stenting, needle-knife precut papillotomy
or fistulotomy, transpancreatic sphincterotomy, or an EUS-guided approach were used
at the discretion of operator (Fig. 1s ). EUS-BD was performed either before or after other advanced cannulation techniques
had been attempted; we defined “early” EUS-BD when it was attempted before other advanced
cannulation techniques.
After deep cannulation of the CBD and appropriate cholangiography had been achieved,
a self-expandable metal stent (SEMS) was placed. The opportunity to proceed with sphincterotomy
prior to stent placement, as well as the stent type (covered/uncovered) and length,
was left to the discretion of the endoscopist.
For the EUS-guided approach, the CBD was identified using a linear echoendoscope and
punctured either directly by the delivery system of the EC-LAMS (Hot AXIOS stent;
Boston Scientific Medical, Marlborough, Massachusetts, USA), or with a 19-gauge needle.
In patients in whom the bile duct was punctured by the needle, bile was aspirated
and contrast was injected to delineate the length and site of the stricture. A 0.025-inch
or 0.035-inch guidewire was then advanced into the CBD before delivery system insertion.
A 6 × 8-mm or 8 × 8-mm LAMS was used. The possible placement of coaxial double-pigtail
plastic stents was at discretion of the operator. No EUS-guided strategies, other
than EUS-CDS with the EC-LAMS, were used to manage the study condition – namely, DMBO
with a dilated CBD – during the study period.
Outcomes and statistical analysis
The rates of both technical and clinical success were assessed as the effectiveness
outcomes. Technical success was defined as the ability to access and drain the CBD
by placement of a stent. Clinical success was defined as a >50% drop in bilirubin
levels within 15 days. The rate of procedure-related AEs, those experienced within
30 days of the procedure and defined according to the American Society for Gastrointestinal
Endoscopy (ASGE) lexicon [11 ], was reported as the safety outcome.
Data were reported as the mean (SD) for continuous variables, and counts and percentages
for categorical variables.
Subgroup analyses based on drainage technique (standard cannulation, “early” EUS-BD,
advanced cannulation techniques, “late” EUS-BD) were performed for both the effectiveness
and safety outcomes. Comparisons were made by chi-squared test or Fisher’s exact test
for categorical data, the Mann–Whitney U test or independent Student's t test were used for non-normally distributed and normally distributed continuous data,
respectively.
The risk factors potentially associated with AEs through a cause–effect relationship
were investigated by univariable and multivariable analysis, with potential factors
being related to the patient (age, sex, body mass index) and the procedure (advanced
cannulation techniques [double-guidewire technique, pancreatic duct stenting, needle-knife
precut papillotomy or fistulotomy, transpancreatic sphincterotomy], and EUS-guided
approach). The strength of association was expressed as the relative risk (RR) with
95%CI. A logistic regression model was then constructed to identify factors that were
independently associated with AEs, with the strength of association expressed as the
adjusted RR (aRR) with 95%CI. A P value <0.05 was considered statistically significant. Statistical analyses were performed
using STATA statistical software (version 18; Texas, USA).
Results
A total of 1016 patients (71.0% men; mean [SD] age, 72.4 [8.1] years) were included
over the study period ([Fig. 1 ]). The most common etiology of DMBO was pancreatic cancer (n = 787; 77.5%), followed
by extrahepatic cholangiocarcinoma (n = 98; 9.6%) and ampullary cancer (n = 74; 7.3%).
The baseline and demographic characteristics of the included patients are summarized
in [Table 2 ].
Fig. 1 Study flow of patient inclusion and exclusion, and eventual management.
Table 2 Baseline characteristics of the 1016 patients included in the study.
Characteristic
BMI, body mass index; CBD, common bile duct.
Age, mean (SD), years
72.4 (8.1)
Sex, male, n (%)
721 (71.0)
BMI, n (%), kg/m2
119 (11.7)
646 (63.6)
251 (24.7)
Primary tumor, n (%)
787 (77.5)
98 (9.6)
74 (7.3)
4 (0.4)
53 (5.2)
CBD diameter, mean (SD), mm
16.1 (4.0)
Total bilirubin, mean (SD), mg/dL
11.5 (7.2)
In 492 patients (48.4%), the CBD cannulation and drainage was successfully achieved
by standard cannulation. In 524 patients (51.6%), the procedure matched the definition
of DBC. In 472 patients (46.5%), an advanced cannulation technique was attempted first,
with a needle-knife precut being the most frequent choice of technique (n = 251; 24.7%),
followed by transpancreatic sphincterotomy (n = 76; 7.5%), the double-guidewire technique
(n = 68; 6.7%), and pancreatic stenting (n = 5; 0.5%). In 72 patients (7.1%), more
than one advanced rescue strategy was attempted. The pancreatic stent was left in
place prophylactically in 129 patients (12.7%).
In 106 patients (10.4%), an EUS-BD was performed, with 52 of these being performed
as the first option after standard cannulation failure (early EUS-BD). Procedural
features are detailed in [Table 3 ].
Table 3 Procedural findings and techniques used.
Feature
Number of patients
Morphology of the papilla
612
192
123
59
25
Duodenal invasion
202
Difficult biliary cannulation
524
Advanced cannulation technique used
472
131
120
76
68
5
72
Prophylactic pancreatic stent placement
129
Effectiveness outcomes
Among the 1016 included patients, there was one failure of biliary drainage, which
was due to an intraprocedural major bleeding event after precut papillotomy, resulting
in a 99.9% rate of technical success. Clinical success was achieved in 956 patients
(94.1%), with a mean (SD) follow-up of 8.1 (6.0) months. No differences in terms of
technical and clinical success were shown in the subgroup analysis based on drainage
technique (standard cannulation, “early” EUS-BD, advanced cannulation techniques,
or “late” EUS-BD) ([Table 4 ]).
Table 4 Outcomes of the included patients.
Standard ERCP cannulation (n = 492)
Difficult biliary cannulation
P value
“Early” EUS-BD
Advanced ERCP cannulation (n = 418)
“Late” EUS-BD
ERCP, endoscopic retrograde cholangiopancreatography; EUS-BD, endoscopic ultrasound-guided
biliary drainage.
Technical success, n %
492 (100)
52 (100)
417 (99.8)
54 (100)
–
Clinical success, n %
470 (95.6)
48 (92.3)
386 (92.3)
52 (96.3)
0.18
Adverse events, n (%)
59 (12.0)
4 (7.7)
89 (21.3)
14 (25.9)
<0.001
24
2
42
5
22
1
22
4
3
0
10
3
5
1
7
0
0
0
2
2
5
0
6
0
Safety outcomes
Procedure-related AEs were experienced by 167 patients (16.4%). Cholangitis (n = 73;
7.2%) and pancreatitis (n = 50; 4.9%) were the most frequently reported AEs; bleeding,
cholecystitis, and perforation occurred in 16 (1.6%), 13 (1.3%), and four patients
(0.4%), respectively. Patients with DBC showed a higher risk for AEs (107/524; 20.4%)
compared with those treated by standard cannulation (59/492; 12.0%; RR 1.7, 95%CI
1.3–2.3; P = 0.003).
Patients undergoing “early” EUS-BD showed a risk of AEs comparable with those managed
with standard cannulation (7.7% vs. 12.0%, respectively; RR 0.6, 95%CI 0.2–1.7; P = 0.38) on subgroup analysis based on drainage technique. In contrast, patients requiring
either advanced cannulation techniques (21.3% vs. 12.0%; RR: 1.8, 95%CI 1.3–2.4; P = 0.001) or “late” EUS-BD (25.9% vs. 12.0%; RR 2.2, 95%CI 1.3–3.6; P = 0.001) were found to have a higher risk of procedure-related AEs compared with
those undergoing standard cannulation ([Fig. 2 ]).
Fig. 2 Comparison of adverse events among the different subgroups.
On investigation of the predictive factors for AEs by univariable and multivariable
analysis, an attempt at any advanced cannulation technique was independently associated
with the occurrence of AEs (P = 0.001) (Table 2s ). The need for EUS-BD showed no impact on increasing risk of AEs.
Focusing on the risk of the different AEs, the use of any advanced cannulation technique
was associated with a statistically significantly higher risk of cholangitis (RR 2.1,
95%CI 1.3–3.3; P = 0.002), and bleeding (RR 5.0, 95%CI 1.4–17.4; P = 0.01). Additionally, no perforations were reported when advanced cannulation techniques
were avoided, compared with four perforations that were reported when there was a
need for any advanced cannulation technique (P = 0.12). No statistically significant differences were found for risk of pancreatitis
(RR 1.3, 95%CI 0.8–2.3; P = 0.34) and cholecystitis (RR 1.6, 95%CI 05–5.1; P = 0.41). A detailed breakdown of the procedure-related AE rates associated with the
different advanced techniques used is given in Table 3s.
Discussion
Our multicenter experience shows that, since the introduction of EUS-BD among the
rescue strategies, adequate biliary drainage was obtained in almost all patients with
DMBO with a dilated biliary tree (CBD >12 mm), despite the risk of DBC. Furthermore,
although the risk of AEs is higher in this subgroup of patients, this appears to be
mainly related to the advanced cannulation techniques, while early EUS-BD showed a
safety profile comparable with that of the standard cannulation technique.
The integration of EUS and ERCP has already suggested its potential for DBC in retrospective
series comparing advanced cannulation techniques and EUS-guided biliary access, mainly
through EUS-guided rendezvous [12 ]
[13 ]. Moreover, a recent randomized controlled trial (RCT) [14 ] confirmed that EUS-guided rendezvous is a viable option for DBC during ERCP performed
for benign conditions. Focusing on malignant conditions, as in our analysis, may a
have a more relevant clinical impact for several reasons.
First of all, the goal of achieving biliary drainage during the index procedure is
of paramount importance for patients with DMBO, considering that most of them are
not candidates for upfront surgical management [15 ], meaning they require chemotherapy as soon as possible. The possibility of using
biliary drainage routes other than through the papilla allows us to overcome the paradigmatic
equivalence between "DBC" and "difficult biliary access." As a matter of fact, EUS-BD
may minimize the impact of the various factors that might make transpapillary access
challenging [2 ]. We may find a clue to the growing impact of EUS-BD in clinical practice by highlighting
the increasing use of EUS as a rescue strategy after ERCP failure. While in 2016,
Holt et al. [16 ] described its use as very limited, more recent data show a more dynamic integration
with ERCP, with up to 9.1% of cases in treatment-naïve patients requiring EUS-BD in
2021 [17 ]. This is indeed comparable with the 10.4% EUS-BD rate reported in our cohort, and
demonstrates a steady and diffuse shift away from the concept of cannulating the papilla
at any (patients’) cost, considering the opportunity of a viable alternative.
Dedicated single-step EC-LAMSs have simplified EUS-CDS, making it an efficient alternative
to ERCP owing to its favorable procedural time and high technical success rate. This
endoscopist-friendly profile is probably the main reason explaining the rapid diffusion
of the technique; however, while its safety is encouraging, the clinical effectiveness
of EUS-CDS, though promising, still requires long-term validation. In contrast, ERCP
has an established track record in the management of DMBO.
Secondly, our results are even more relevant when looking at the possible impact on
safety. Of note, advanced cannulation techniques have already been shown to allow
cannulation in most cases of DBC. On the flip side, the cost in terms of procedure-related
AEs may be relevant. With this in mind, adopting EUS-BD as one strategy among the
others, not relegating it as the option to be used when all the others have failed,
was shown to maximize the benefit of this approach. After all, we had previously learned
about precut [18 ], and the opportunity of its early adoption to minimize the manipulation of the papilla
in order to reduce the risk of pancreatitis. From our results, when EUS-BD is performed
after other advanced approaches have already been attempted, it may carry their risk
of procedure-related AEs. In contrast, the early adoption of an EUS-guided approach
showed potential to be considered the safest option in selected cases, with a reported
risk of procedure-related AEs comparable with ERCP performed with standard cannulation
techniques. In particular, in our series, the use of any advanced cannulation technique
increased the risk of cholangitis, bleeding, and perforation.
As previously underlined, time is a crucial factor for these patients and should be
regarded as an oncological outcome, as extended hospital stays due to procedure-related
AEs, such as post-ERCP pancreatitis, can delay oncological treatments and, occasionally,
hinder surgical options. Previous evidence has suggested that EUS-BD may have a role
as a feasible and safe option offering a bridge-to-surgery [19 ].
Furthermore, the promising findings of our analysis in term of both effectiveness
and safety are even more relevant when considering that they have an impact on the
vast majority of patients with jaundice due to DMBO, because they have been proven
to present a CBD dilatation that is permissive of EUS-CDS in most cases [20 ]. Although it was beyond the scope of our study, it is also important to acknowledge
that EUS-CDS with a LAMS is not the only EUS-guided tool in our armamentarium. Mastery
of interventional EUS offers a range of additional options (e.g. EUS-CDS with a SEMS,
the rendezvous technique, antegrade stenting, gallbladder drainage, hepaticogastrostomy,
etc.), which can be tailored to address various challenging scenarios (e.g. CBD <12
mm, duodenal infiltration, altered anatomy, etc.), while reserving the percutaneous
approach as a final option where there has been failure of the integrated endoscopic
approach.
Thirdly, our findings highlight the need to no longer view EUS and ERCP as separate
entities. This has clinically significant practical implications beyond the clinical
perspective, starting with the need for thorough preprocedural patient counseling
to obtain truly informed consent. In this regard, we believe it is crucial to emphasize
the importance of integrated training in biliopancreatic endoscopy. This approach
would equip endoscopists with the skills needed to master all of the relevant techniques
for managing challenging situations [21 ]
[22 ]. Furthermore, speculating on the possible consequences of such an approach, looking
beyond the training (and the performance) of the single endoscopist, an integrated
setting may possibly require fewer biliopancreatic endoscopists, but always with integrated
EUS–ERCP skills, to maximize the patient outcomes, while minimizing the risk of deskilling
in advanced techniques that might be expected over time.
The main strength of our analysis is the precise methodology delineating, through
definite exclusion criteria, a specific condition that endoscopists must frequently
deal with. This allowed us to limit the risk of confounders when summarizing our findings,
allowing us to provide real-life, but still reliable, data on the endoscopic management
of patients with DMBO and CBD dilatation permissive for EUS-CDS. For instance, for
the purpose of our analysis, we excluded patients with luminal involvement causing
gastric outlet obstruction symptoms, because of emerging evidence suggesting EUS-CDS
might be a suboptimal choice in this setting [23 ]
[24 ]. Secondly, the multicenter setting enhances the reproducibility of our findings.
Despite its strengths, our study has several limitations. The primary drawback is
its retrospective nature, which may introduce biases that could undermine our conclusions.
Nevertheless, our real-world experience demonstrates the safety of early EUS-BD in
patients with DBC, potentially guiding the design of future prospective comparative
studies. Such studies should place an emphasis on providing data that are particularly
difficult to obtain retrospectively, such as the reason for defining a “difficult”
biliary cannulation (according to specific criteria, i.e. the European Society of
Gastrointestinal Endoscopy [ESGE] criteria), and the grade of severity of the AEs.
As a matter of fact, we could not report such data because of the high risk of providing
unreliable data.
Furthermore, the disparity in the size of the subgroups limits our ability to make
statistical inferences about specific features associated with better outcomes for
either approach, as well as providing reliable sensitivity analysis focused on the
specific centers. Nonetheless, the similar numbers of patients included by the different
centers per year is reassuring with regard to the risk of behavioral heterogeneity
among the centers, while, given the scarcity of literature on this topic, our study
provides the first piece of evidence highlighting the potential benefit of the early
adoption of EUS-CDS, which will likely assist endoscopists in their decision-making
process when dealing with these challenging situations.
To identify the subgroup of patients who might benefit most from this approach, we
should focus on those who have, on one hand, a high risk of procedure-related AEs
owing either to prolonged procedural time or to the need for advanced cannulation
techniques and, on the other hand, a high likelihood of easy drainage being achieved
by EUS-CDS. Risk factors for ERCP-related AEs have been extensively studied. For example,
elderly, obese, and/or co-morbid patients have a higher risk of procedural complications
(such as cholangitis, bleeding, and perforation), as well as increased morbidity and
mortality owing to sedation-related issues. Additionally, young women are at a higher
risk of developing post-procedural pancreatitis. In contrast, a clear profile of patients
for whom an easy EUS-CDS procedure is expected has not yet been defined, although
CBD size may serve as a useful proxy of potential technical challenges. A recent RCT
[8 ] suggested that patients with a CBD diameter of <16 mm had a higher rate of procedure-related
AEs (10.7% vs. 5.3%), although this difference did not reach statistical significance.
Nonetheless, this finding suggests that smaller CBD diameters could make the introduction
and deployment of cautery-enhanced LAMSs more challenging.
In conclusion, despite patients with DMBO having a relevant risk of DBC, an EUS-CDS
with a LAMS represents a reproducible option to obtain adequate biliary drainage in
virtually all patients with a dilated CBD (>12 mm). In these patients, the early adoption
of EUS-BD, even before any other advanced cannulation technique has been attempted,
may minimize the risk of procedure-related AEs in selected cases. Prospective comparative
studies are needed to confirm our findings.
