Introduction
Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CDS) is a relatively new
technique that allows the endoscopist to create a biliodigestive anastomosis. As the
tumor is bypassed with EUS-CDS, the effort required to obtain biliary access is more
straightforward compared with endoscopic retrograde cholangiopancreatography (ERCP).
Moreover, EUS-CDS obviates the need for manipulation of the papilla in order to gain
biliary access, and the stent does not cause acute obstruction of the pancreatic duct,
thereby precluding the risk of post-procedural pancreatitis.
EUS-CDS has already been shown to be superior to percutaneous approaches in patients
with distal malignant biliary obstruction (MBO) [1 ]
[2 ]
[3 ]. Based on these promising results, EUS-CDS is now also being compared with ERCP.
Current prospective studies, however, all used biliary self-expandable metal stents,
whereas electrocautery-enhanced lumen-apposing metal stents (LAMSs) would simplify
the procedure [1 ]
[2 ]
[4 ]
[5 ].
Therefore, the aim of this prospective pilot study (SCORPION-p) was to assess the
safety and feasibility of EUS-CDS using LAMS as the primary drainage strategy in patients
with distal MBO.
Methods
Study design
Consecutive patients were screened for eligibility between October 2021 and June 2022 at
Amsterdam UMC. Patients with a distal MBO confirmed by histology or cytology (including
rapid onsite evaluation strongly suggestive of malignancy) and who had an indication
for biliary drainage were considered eligible. The main exclusion criteria were surgically
altered anatomy, cancer extending into the antrum or proximal duodenum, extensive
liver metastases, World Health Organization performance score of 4, uncorrectable
coagulopathy, or clinically relevant gastric outlet obstruction (GOO). The study was
approved by the medical ethics committee of Amsterdam UMC. All patients provided written
informed consent before inclusion. An independent monitor performed clinical trial
monitoring.
Study procedures
All patients received a single dose of prophylactic broad-spectrum intravenous antibiotics
in line with European Society of Gastrointestinal Endoscopy guideline recommendations
[6 ]. Anticoagulants were stopped if applicable (i. e. an international normalized ratio
of < 1.5 was permitted). Antiplatelet monotherapy was allowed; in cases of dual antiplatelet
therapy, one of the two drugs needed to be discontinued 5 days prior to the procedure
and was restarted 24 hours post-procedurally.
The procedure was performed using a linear ultrasound endoscope (Olympus GF-UCT180;
Olympus Tokyo, Japan) with the patient in the left lateral or prone position. For
cases without a previous tissue diagnosis, a fine-needle biopsy and/or fine-needle
aspiration was performed to confirm malignant obstruction. The common bile duct (CBD)
was identified proximally to the level of the tumor obstruction and at least 2 cm
below the hilum. Subsequently, the origin of the cystic duct from the CBD was visualized.
Care was taken to avoid intervening blood vessels. To allow safe stent deployment,
the minimum bile duct diameter at the puncture site was set at 12 mm given that all
procedures were performed by experts in LAMS placement [6 ]
[7 ]. If the diameter was < 12 mm, a standard ERCP was performed. EUS-CDS was performed
using the “free-hand technique,” meaning that the electrocautery-enhanced LAMS was
directly introduced into the bile duct using pure cutting current (100 W). The Hot
AXIOS stent (Boston Scientific, Marlborough, Massachusetts, USA), 6 × 8 mm, was used.
In cases where the LAMS catheter could not be advanced deep enough into the bile duct,
a guidewire was advanced toward the hilum to redirect the catheter and facilitate
further advancement. In bile ducts with a small diameter, the distal flange was deployed
in a stepwise manner. The biliary system was visualized following LAMS placement by
contrast injection via a diagnostic catheter through the LAMS, in order to confirm
adequate stent position and exclude contrast leakage. The procedure is illustrated
in [Fig. 1 ]. Three gastroenterologists (P.F., R.P.V., R.L.J.W.), experienced in both EUS and
ERCP, performed all study procedures, with two of them being present in the endoscopy
suite during each procedure.
Fig. 1 Endoscopic ultrasound-guided choledochoduodenostomy with electrocautery-enhanced
lumen-apposing metal stent (LAMS) placement and confirmation of technical success
by cholangiogram. a Sonographic identification of the common bile duct (CBD) proximal to the tumor. A
target site where the CBD was ≥ 12 mm, with no intervening vessels or ascites, was
identified. Using pure cutting current, the LAMS was introduced into the bile duct
using the free-hand technique. b The distal flange was deployed in the CBD under endosonographic guidance. c The proximal flange was subsequently deployed under endoscopic guidance, and resulted
in immediate bile flow from the LAMS. d Cholangiogram via the LAMS confirmed adequate stent position.
Follow-up was performed after 2 weeks, 4 weeks, 3 months, and 6 months.
Outcome
The primary outcome was technical success. Secondary outcomes were: 1) clinical success,
defined as at least 50 % decrease of bilirubin and/or relief of symptoms without the
need for reintervention within 30 days; 2) procedure time, measured from introduction
of the endoscope until visual flow of bile through the LAMS; when fine-needle biopsy
or fine-needle aspiration was required, time was started after completion of this
procedure; 3) (serious) adverse events (AEs) within ≤ 30 days after the procedure.
Periprocedural AEs were events that occurred during the procedure. Severity of AEs
was graded according to the AGREE classification [8 ]. Stent dysfunction was defined as recurrent jaundice (conjugated bilirubin ≥ 35 µmol/L
[2.0 mg/dL]) after initial clinical success, persistent jaundice and dilatation of
the bile ducts, or cholangitis. The reason for stent dysfunction was classified according
to the Leuven-Amsterdam-Milan Study Group classification of EUS-CDS dysfunction [9 ]. Time to recurrent biliary obstruction was calculated from the moment of stent insertion
until stent dysfunction. Reinterventions in cases of stent dysfunction were reported.
Dysfunction-free survival was defined as the number of days after EUS-CDS until death
without experiencing stent dysfunction.
Statistical analysis
Descriptive statistics were used to report proportions and characteristics of the
results using R version 4.0.1 (R Foundation for Statistical Computing, Vienna, Austria).
Categorical variables were expressed as absolute and relative frequencies, and 95 %CIs
were constructed using the exact binomial distribution approximation. Continuous data
were presented as medians and interquartile ranges (IQRs). (Dysfunction-free) survival
was estimated using Kaplan–Meier survival analysis; as all patients who were still
undergoing follow-up were censored at 6 months, only a point estimate without 95 %CI
was provided.
Results
Baseline characteristics
A total of 30 patients were eligible and signed the informed consent. Eight patients
were excluded because rapid onsite evaluation could not confirm malignancy (n = 3),
the CBD diameter was < 12 mm (n = 2), or because there was no safe target site to
perform the procedure, either because the tumor was too close to the hilum (n = 2)
or because of ascites (n = 1). Finally, 22 consecutive patients with distal MBO were
enrolled. The full screening and selection process is depicted in Fig. 1 s in the online-only Supplementary material. Baseline characteristics of the included
patients are summarized in [Table 1 ].
Table 1
Baseline characteristics.
Characteristics
n = 22
Male sex, n (%)
7 (32)
Age, median (IQR), years
69.5 (64.0–75.3)
BMI, median (IQR), kg/m2
24.7 (23.7–26.1)
Type of tumor, n (%)
20 (91)
1 (5)
1 (5)
WHO performance score at inclusion, n (%)
6 (27)
12 (55)
2 (9)
2 (9)
Use of anticoagulant drugs, n (%)
7 (32)
Tumor stage at inclusion, n (%)
10 (46)
6 (27)
6 (27)
Serum total bilirubin, median (IQR), µmol/L
225 (130.75–335.25)
Diameter of CBD on EUS, median (IQR), mm
16.5 (13.25–20.75)
Concomitant chemotherapy at inclusion, n (%)
2 (9)
Cholecystectomy prior to intervention, n (%)
4 (18)
BMI, body mass index; IQR, interquartile range; WHO, World Health Organization; CBD,
common bile duct; EUS, endoscopic ultrasound.
Technical success
Immediate technical success was achieved in 18/22 patients. In two additional patients,
the distal flange was initially inadequately deployed in the bile duct wall, leading
to minor bile spill, which was immediately resolved after manipulation (n = 1) or
replacement with a second LAMS (n = 1), without clinical consequences. Therefore,
the overall technical success rate was 91 % (20/22; 95 %CI 71 %–99 %).
In two patients the procedure was unsuccessful. In one patient, the stent was unintentionally
placed in the cystic duct. In the other patient, the distal flange was deployed outside
the bile duct wall. An ERCP with closure of the defect in the duodenum using a through-the-scope
clip was performed in the same procedure and the patient recovered uneventfully. In
one patient, a double-pigtail stent was placed through the LAMS to prevent stent obstruction
by blood clots after self-limiting intraprocedural bleeding. Median procedure time
was 11 minutes (IQR 7–16).
Clinical success
Clinical success was achieved in 19/22 patients (86 %; 95 %CI 65 %–97 %). The patient
in whom the stent was unintentionally placed in the cystic duct underwent a second
successful EUS-CDS procedure due to inadequate biliary drainage. The other two patients
underwent successful placement of an additional double-pigtail stent to achieve adequate
biliary drainage because of early cholangitis (n = 1) or suspected stent obstruction
(n = 1).
Adverse events
Besides the perforation described above, which was treated endoscopically, no periprocedural
AEs occurred.
Eight patients (36 %) experienced a possible related AE ≤ 30 days after the procedure.
Two AEs were unrelated to stent dysfunction: one patient had mild intermittent abdominal
pain, which resolved after placement of a double-pigtail stent through the LAMS, and
one patient developed rhabdomyolysis and kidney failure < 2 weeks after the procedure,
which completely resolved and had unknown relation to the procedure. Six patients
developed cholangitis due to stent dysfunction. None of the patients developed pancreatitis
or delayed bleeding ([Table 2 ]).
Table 2
Adverse events ≤ 30 days including grading.
Adverse events
n = 22
Adverse events ≤ 30 days, n (%)
1 (5)
0 (0)
0 (0)
6 (27)
2 (9)
1 (5)
1 (5)
Severity of adverse events ≤ 30 days, n (%)[1 ]
3 (14)
5 (23)
1 (5)
30-day mortality, n (%)
1 (5)
1 According to the AGREE classification [8].
One patient died ≤ 30 days from fulminant disease progression, which was considered
unlikely to be related to the procedure ([Table 2 ]).
Stent dysfunction (n = 20)
A total of 11/20 patients with a technically successful procedure (55 %) had experienced
stent dysfunction during the 6-month follow-up, presenting with either cholangitis
(n = 10) or jaundice (n = 1). Stent dysfunction occurred after a median of 6 days
(IQR 5–87.5). Median estimated dysfunction-free survival was 140 days. Reason and
grading of stent dysfunction is shown in Table 1 s .
In two patients, cholangitis was treated successfully with antibiotics, but in nine
patients reintervention was required. Overall, endoscopic reinterventions were successful
in 8/9 patients (89 %). In patients who developed GOO due to disease progression (n = 3),
concomitant surgical (n = 1) or endoscopic (n = 1) gastroenterostomy was performed,
or the condition was left untreated according to the patient’s wishes (n = 1). Reinterventions
are summarized in Fig. 2 s .
Follow-up (n = 20)
Median total follow-up was 149 days (IQR 62.5–180). Five patients underwent surgical
resection after a median of 34 days (IQR 23.5–49.75). Eight patients died after a
median of 80 days (IQR 71–157). The remaining seven patients were still undergoing
follow-up after 6 months. Estimated median survival was 172 days.
Discussion
This pilot study prospectively evaluated the use of EUS-CDS with LAMS as the primary
drainage strategy in patients with distal MBO. EUS-CDS showed high technical and clinical
success rates in combination with minimal periprocedural AEs. The high rate of stent
dysfunction (55 %), however, presents a challenge that first needs to be addressed
before the potential benefits of EUS-CDS with LAMS can be realized.
Technical and clinical success rates were comparable to previous studies performing
EUS-CDS with LAMS after unsuccessful ERCP, with ranges of 89 %–100 % and 82 %–100 %,
respectively [2 ]
[9 ]
[10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
[17 ]. The stent dysfunction rate in this study, however, was considerably higher compared
with the 6 %–37 % reported previously for EUS-CDS with LAMS [9 ]
[10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
[17 ]. This discrepancy may be partially explained by the fact that the majority of studies
on this topic were retrospective and may have underestimated the rate of stent dysfunction.
Second, a relatively strict, though clinically relevant, definition of stent dysfunction
was used in the current study, including cholangitis as well as persistent or recurrent
jaundice. Third, despite GOO being an exclusion criterion, three patients developed
GOO during the course of the disease, which may have contributed to the occurrence
of cholangitis [9 ]
[18 ]. Fourth, the use of LAMS with a relatively small diameter (6 × 8 mm) may have contributed,
as currently there is some evidence that stents with larger diameters may reduce the
risk of stent dysfunction [16 ]. Finally, in our study, double-pigtail stents were not routinely placed through
the LAMS, although recent data show that this may be beneficial [19 ]. On the other hand, the fact that five patients underwent surgical resection after
a median of 34 days could have led to an underestimation; however, considering stent
dysfunction occurred after a median of 6 days, this factor is expected to be of limited
influence. Data on surgical resection after EUS-CDS are still scarce; however, we
believe the available data show no reason to be reluctant to perform EUS-CDS in operable
patients while awaiting further studies in this specific patient group [20 ].
Although the rate of cholangitis due to stent dysfunction was high, the course of
the disease was generally mild. The vast majority of patients were successfully treated
with antibiotics and/or endoscopic reintervention. Stent dysfunction after ERCP with
self-expandable metal stents, though lower than with EUS-CDS in the current study,
is also substantial, with a range of 3 %–43 % [4 ]
[5 ]. However, with regard to other AEs, such as pancreatitis, cholecystitis, and delayed
bleeding, the safety profile of EUS-CDS seems to be superior to that of ERCP [21 ]. Moreover, periprocedural AEs of EUS-CDS in the current study were limited, and
were managed endoscopically in the same session without clinical implications.
EUS-CDS, using the current technique, is unable to fully replace ERCP, however, as
EUS-CDS was not feasible in 17 % of our patients. In 5/30 included patients, the CBD
diameter was too small (< 12 mm) or there was no safe target site at which to perform
the procedure, making the patient ineligible for EUS-CDS. Lack of feasibility was
mainly due to insufficient bile duct dilatation, which is in line with a recent study
on pre-procedural cross-sectional imaging that identified a sufficiently (> 12 mm)
dilated CBD in only 78.8 % of patients [7 ]. Furthermore, EUS-CDS should not be conducted in patients with GOO due to the high
risk of influx of gastric contents in this specific group. Thus, endoscopists should
be well trained in both EUS and ERCP in order to switch from EUS-CDS to ERCP when
indicated, as well as to adequately manage periprocedural AEs.
The findings of this study are limited by the small sample size and the lack of a
control group. Future studies should directly compare the overall impact of AEs and
stent dysfunction of either technique on clinical condition, quality of life, and
delay or annulment of treatment. However, in order to conduct such a trial, the EUS-CDS
procedure should first be further optimized to lower the risk of stent dysfunction.
In conclusion, the present study supports the safety and feasibility of EUS-CDS using
LAMS as the primary drainage strategy in patients with distal MBO. However, the high
incidence of stent dysfunction currently limits the use of EUS-CDS with LAMS as a
valid alternative to ERCP with self-expandable metal stents. Further studies on the
benefit of coaxial stent placement through the LAMS or alternative stent designs are
necessary to reduce the risk of stent dysfunction.
