Keywords
Pancreatobiliary (ERCP/PTCD) - Biliary tract - Endoscopic ultrasonography - Intervention
EUS
Introduction
Management of malignant biliary obstruction (MBO) is of paramount importance because
obstructive jaundice dramatically decreases the ability to administer systemic chemotherapy
to patients with unresectable disease. MBO reduces quality of life and increases the
risk of morbidity and mortality [1]. Currently, endoscopic retrograde cholangiopancreatography (ERCP) with transpapillary
placement of self-expanding metal stents is considered the gold‑standard technique
to achieve biliary drainage in distal MBO [2]
[3], enabling a high success rate, ranging from 86% to 99% when considering all indications,
and with an acceptable safety profile. However, in the setting of malignant disease,
a lower rate of success might be expected, with a higher need for advanced cannulation
techniques (i.e., needle knife pre-cut) with their associated adverse events (AEs)
[4]
[5]
[6]
[7].
In case of ERCP failure, a percutaneous approach is traditionally considered the main
secondary option. However, since first reported by Giovannini et al. [8], interventional endoscopic ultrasonography (EUS) has been demonstrated to be a viable
option for palliation of MBO. In particular, EUS-guided choledochoduodenostomy (EUS-CDS)
has emerged as an alternative treatment modality by providing internal biliary drainage
in patients with distal common bile duct (CBD) obstruction and upstream biliary system
dilation. Furthermore, the development of a lumen-apposing metal stent (LAMS) with
cautery-enhanced delivery system allows simple one-step puncture and stent delivery,
greatly shortening procedure duration. This rapidly spreading approach, initially
performed only as an alternative to percutaneous drainage for cases of ERCP failure,
is now challenging ERCP as the primary approach for relief of MBO [9]
[10]
[11]
[12]. Despite promising data, the major concern is mainly related to the long experience
with ERCP, which has allowed recognition of factors that affect the risk of technical/clinical
success, and development of AEs.
Initial experience with EUS-CDS proved its feasibility, efficacy, and safety and more
recently the volume of published data has grown exponentially, offering us the opportunity
for a deeper insight with this approach.
The primary objective of this meta-analysis was to quantitatively assess patient-
and procedure-related factors potentially influencing the outcomes of EUS-CDS using
LAMS in the management of distal MBO.
Methods
The methods for our analysis and inclusion criteria were based on Preferred Reporting
Items for Systematic Reviews and Meta-Analyses recommendations [13]. Data sources and search strategy, selection process, data extraction, and quality
assessment are reported in Appendix 1.
Selection criteria
For the purpose of this systematic review, we considered all clinical studies that
included patients with distal MBO who underwent EUS-CDS using LAMS after failed ERCP.
Small case series including < 10 patients, non-endoscopic studies, review articles,
and animal models were excluded. Data where EUS-CDS was performed as primary therapy
rather than ERCP were also excluded.
Outcomes assessed
The primary outcome was clinical success. We followed the definitions of clinical
success as defined by individual studies (Supplementary Table 1). Secondary outcomes
were technical success, defined as successful LAMS deployment under EUS guidance with
consequent biliary drainage, and AE rate. AEs were defined as any procedural/stent-related
event including abdominal pain, fever, perforation, bleeding, bile leak, jaundice,
stent obstruction and cholangitis. Severity of AEs was graded according to the American
Society for Gastrointestinal Endoscopy lexicon [14]. AEs were grouped into three categories: immediate, early, and late, based on the
timing of their onset and considered as per individual study definitions.
Statistical analysis
Study outcomes were pooled through a random-effects model based on DerSimonian and
Laird test [15], and results were expressed as rates and 95% confidence intervals (CIs). Comparisons
between different LAMS diameters, namely 8 × 8 mm vs 6 × 8 mm, were based on a random-effects
model and results were expressed as odds ratios (ORs) and 95% CIs.
The presence of heterogeneity was calculated through I2 tests I2; values of 0% to 40%, 30% to 60%, 50% to 90%, and 75% to 100% were indicated as low,
moderate, substantial, and considerable heterogeneity, respectively [16]. Any potential publication bias was verified through visual assessment of funnel
plots.
Multivariate meta-regression was performed to analyze the correlation between baseline
age, sex, and CBD diameter and clinical success rate.
A sensitivity analysis based on the timing of AE occurrence (immediate, early, or
late) was performed.
All statistical analyses were conducted using RevMan (version 5.0 for Windows; the
Cochrane Collaboration, Oxford, UK), Jamovi 1.6, R 4.0 software, and R 3.0.2 (R Foundation
for Statistical Computing, Vienna, Austria).
For all calculations, P < 0.05 was considered statistically significant.
Results
Studies
As shown in [Fig. 1], 2788 studies were initially identified. After exclusion of articles not fulfilling
the inclusion criteria, 12 studies [17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28] with 845 patients were included in the meta-analysis, of which 10 were non-comparative,
single-cohort, retrospective studies. Study characteristics are reported in [Table 1].
Fig. 1 PRISMA flow chart. Shamseer L, Moher D, Clarke M et al. Preferred reporting items for systematic review and meta-analysis protocols
(prisma-p) 2015: Elaboration and explanation. BMJ 2015.
Table 1 Study characteristics.
|
Study
|
Location and Year
|
No. centers involved
|
Study design
|
LAMS delivery system
|
Total patients (no.)
|
Gender, Male
%
|
Age
(median; range/mean ± SD)
|
CBD diameter
(median; range/mean ± SD)
|
Technical success
%
|
Clinical success
%
|
Study
|
Location and Year
|
|
Kunda
|
Denmark, 2016
|
7
|
Retrospective study
|
Cold & Hot Axios, Boston Scientific Corp.
|
57
|
54.4
|
73;49–93
|
17.9; 8–35
|
98.2
|
94.7
|
Kunda
|
Denmark, 2016
|
|
Tsuchiya
|
Japan and Hong Kong, 2017
|
5
|
Prospective Cohort study
|
Hot Axios, Boston Scientific Corp.
|
19
|
63.2
|
70.6 ± 13.9
|
17.3 ± 5.5
|
100.0
|
94.7
|
Tsuchiya
|
Japan and Hong Kong, 2017
|
|
Jacques
|
France, 2018
|
10
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
52
|
48.1
|
78;61–92
|
Not Reported
|
88.5
|
98.1
|
Jacques
|
France, 2018
|
|
El Chafic
|
USA, 2019
|
6
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
67
|
55.2
|
68.8 ± 11.8
|
17.6 ± 3.6
|
95.5
|
100.0
|
El Chafic
|
USA, 2019
|
|
Anderloni
|
Italy, 2019
|
1
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
46
|
52.2
|
73.1 ± 12.6
|
17.26 ± 3.34
|
93.5
|
97.7
|
Anderloni
|
Italy, 2019
|
|
Yung-Lun Chin
|
New Zealand, 2020
|
1
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
60
|
56.7
|
76;52–90
|
17 ± 10
|
100.0
|
85.0
|
Yung-Lun Chin
|
New Zealand, 2020
|
|
Jacques
|
France, 2020
|
7
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
70
|
54.3
|
75 ± 11
|
17.7 ± 5
|
98.6
|
98.6
|
Jacques
|
France, 2020
|
|
Garcia-Sumalla
|
Spain, 2021
|
3
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
22
|
24.4
|
75.3 ± 12.1
|
17.5; 9–27
|
95.5
|
77.3
|
Garcia-Sumalla
|
Spain, 2021
|
|
Bun Teoh
|
Korea, China, Thailand, Hong Kong, 2021
|
5
|
Prospective cohort study
|
S-LAMS (Niti-S Spaux, Taewoong Medical, Gyeonggi-do, Korea)
|
26
|
46.2
|
64.1 ± 13.2
|
Not reported
|
88.5
|
88.5
|
Bun Teoh
|
Korea, China, Thailand, Hong Kong, 2021
|
|
Fugazza
|
Italy, 2022
|
23
|
Retrospective study
|
Hot Axios, Boston Scientific Corp. and Nagi Stent
|
256
|
55.1
|
73.9 ± 12.6
|
17.3 ± 3.9
|
93.4
|
96.2
|
Fugazza
|
Italy, 2022
|
|
Ginestet
|
France, 2022
|
1
|
Retrospective study
|
Hot Axios, Boston Scientific Corp.
|
50
|
60.0
|
76.5 ± 0
|
Not reported
|
98.0
|
90.0
|
Ginestet
|
France, 2022
|
We used the Newcastle-Ottawa Scale for nonrandomized studies [29] to assess methodology quality, which gave a mean score of 4.8 (range 4–5) (Supplementary
Table 2, Appendix 1).
Technical success
The pooled technical success rate was 96% (95%CI 94%-98%), with moderate evidence
of heterogeneity (I2 = 52.29%; [Fig. 2]). No evidence of publication bias was detected through visual inspection of the
funnel plot (Supplementary Fig. 1, Appendix 1).
Fig. 2
a Technical and b clinical success rates. Forest plot.
Comparison rates for technical success using 8 × 8 mm vs 6 × 8 mm stents are reported
in Supplementary Fig. 2, Appendix 1. Based on four studies, no difference in terms
of technical success rates was observed between the two diameters (OR 1.21, 95%CI
0.51–2.85), with no evidence of heterogeneity (I2 = 0%).
Clinical success
As depicted in [Fig. 2], the pooled clinical success rate was 96% (95%CI 95%-98%), with evidence of mild
heterogeneity (I2 = 23.78%). No evidence of publication bias was detected through visual inspection
of the funnel plot (Supplementary Fig. 3, Appendix 1).
Meta-regression analysis did not find a significant association between some baseline
characteristics—namely sex, age, and CBD diameter—and clinical success rate (P = 0.45, P = 0.20, and P = 0.53, respectively; Supplementary Fig. 4, Appendix 1). Moreover, no difference
between 8 × 8 mm and 6 × 8 mm stents was observed in terms of clinical success (OR
1.42, 95%CI 0.58–3.46), with mild evidence of heterogeneity (I2 = 7%; Supplementary Fig. 5, Appendix 1)
AE rate
As reported in [Fig. 3], the pooled overall AE rate was 12% (95%CI 8%-16%), with evidence of substantial
heterogeneity (I2 = 71.62%). No evidence of publication bias was detected based on the funnel plot (Supplementary
Fig. 6, Appendix 1). Specifically, early, immediate, and late AE rates were 3% (95%CI
1%-4%), 2% (95%CI 1%-2%), and 4% (95%CI 2%-7%), respectively (Supplementary Fig. 7,
Appendix 1). This sensitivity analysis based on timing of AEs occurrence led to a
consistent decrease in heterogeneity of the estimates (I2 = 56%, 0%, and 74%; respectively).
Fig. 3 Adverse event rate. Forest plot.
Based on four studies, the AE rate was significantly lower with 8 × 8 mm as compared
with 6 × 8 mm stents (OR 0.59, 95%CI 0.35–0.99), with no evidence of heterogeneity
(I2 = 0%; [Fig. 4]).
Fig. 4 Adverse event and stent size.
The detailed list of AEs observed in the included studies is reported in Supplementary
Table 2, Appendix 1.
Discussion
According to our analysis EUS-CDS with LAMS is confirmed as a feasible option for
the management of distal MBO, with a favorable benefit/risk ratio due to a very high
rate of technical and clinical success (96%), assessed in almost 1000 procedures,
and relatively low rate of AEs (12%). Stent size was shown to be the main factor affecting
the risk of AEs.
Data from this meta-analysis may have a significant impact on clinical practice because
of the ongoing questions related to EUS-guided strategies for biliary drainage. In
this regard, the most concerning aspect is safety, especially considering the relatively
limited experience with EUS-CDS using small-diameter LAMS (< 10 mm), especially when
compared with the experience of ERCP for MBO. We herein showed that EUS-CDS is associated
with an adequate safety profile, with no reported fatal events. In fact, the 12% AE
rate seems comparable to the 5% to 15% AE rate reported for ERCP [3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30].
The impact of stent size on procedure outcome, even if not unexpected, was not obvious.
While a larger stent diameter theoretically provides better bile flow and a decreased
risk of stent occlusion, it may also theoretically allow food and debris to enter
with a resultant increased risk of cholangitis. From a technical point of view, placement
of a 6 × 8 mm LAMS might lead to a reduced risk of stent misdeployment because the
minimum space required for deployment is lower than the 8 × 8 mm LAMS. However, the
greater radial force of the 8 × 8 mm stent may decrease the risk of stent dislodgement.
The lower AE rate reported with the 8 × 8 mm LAMS over the 6-mm LAMS for CDS is an
important finding which has not been previously recognized, and only pooling data
from individual studies allowed us to reach the statistical power to show such association.
This is even more relevant considering that we were able to exclude the possible confounding
influence of CBD size. As a matter of fact, a more dilated CBD is considered an easier
target for biliary tract access, and in light of the more frequent choice of an 8
× 8mm LAMS in such cases, this could have biased our results. In our analysis, CBD
size did not influence technical or clinical success or risk of AEs, highlighting
the role of stent size as an independent factor. This suggests it is preferable to
use the 8 × 8 mm LAMS as opposed to the 6 × 8 LAMS when ductal diameter allows.
The second main result of our analysis is the efficacy profile, with a very high rate
of technical and clinical success (96%). If the definition of technical success is
homogeneous across the included studies, it can be argued that different definitions
of clinical success might affect the results. However, the various definitions of
clinical success remained consistent with the definitions suggested by the European
Society of Gastrointestinal Endoscopy [2], and the only moderate heterogeneity of our analysis downgraded this risk.
One of the main strengths of our analysis is the clinical setting of the included
studies. As a matter of fact, we only included data from EUS-CDS when used for demanding
cases of ERCP failure. Such a high rate of technical success, theoretically achievable
during the same ERCP session, gives the endoscopist the opportunity to successfully
relieve distal MBO in nearly 100% of cases during one endoscopic session. This prevents
the need for rescheduling procedures, allowing rapid improvement of quality of life
and institution of systemic chemotherapy [1]. We believe that procedural informed consent for EUS-CDS (or EUS-guided biliary
drainage) should be obtained at the time of consent for ERCP in cases of distal MBO
[2]
[31]
[32].
Our analysis has some limitations. First, the majority of the studies included are
retrospective, and this may introduce an element of selection bias. This, along with
the paucity of comparative studies available, demands some caution in interpretation
of the results. Second, considering that all the included patients had failed ERCP
before EUS-guided CDS was attempted, we cannot exclude some carryover effect causing
an overestimation of the AE risk. This means there may be a lower AE rate than seen
in this analysis. Despite the lack of head-to-head comparison with ERCP, our study
may be informative for designing comparative trials. Third, other important technical
points such as drainage technique (i.e. free-hand vs. wire-guided techniques), or
the possible use of larger stents (i.e. 10 × 10 mm) have not been homogeneously reported
across the included studies, and we are not able to make any evidence-based suggestion
about such technical points. Finally, the moderate-to-substantial level of heterogeneity
found in several analysis, coupled with differences across the included studies in
term of design, center numbers, and sample size may have affected the interpretability
of the results. Furthermore, we performed different analyses specifically to investigate
the reasons why heterogeneity (i.e. meta-regressions, subgroup analysis) partially
mitigated the issue. However, the extremely high rate of both technical and clinical
success and the paucity of failure events may have down-powered our meta-regression
analysis and a future analysis will be needed to highlight any relevant factors affecting
efficacy outcomes. On the other hand, the low level of heterogeneity found in the
subgroup analysis based on stent size (I2 = 0%) provides reassurance about the reliability of data for safety outcomes.
Conclusions
In conclusion, our analysis confirmed that EUS-CDS is a safe and effective option
for patients with distal MBO in whom ERCP has failed. The selection of appropriate
stent size seems important for achieving optimal outcomes.
