Technical failure of endoscopic ultrasound-guided choledochoduodenostomy: multicenter study on rescue techniques, consequences, and risk factors

Abstract

Background We aimed to determine risk factors for technical failure of endoscopic ultrasound-guided choledochoduodenostomy (EUS-CDS), evaluate short- and long-term consequences, and identify salvage techniques in patients with biliary obstruction.

Methods This retrospective multicenter study of EUS-CDS (2018–2024) included technical failure, classified as type 1 (digestive flange mispositioned), type 2 (biliary flange mispositioned), type 3 (stent deployment failure), type 4 (catheter/lumen-apposing metal stent [LAMS] through the bile duct), and type 5 (others). Controls were successful EUS-CDS in the same center and period. The primary end point was risk factors for failure. Secondary end points were endoscopic rescue techniques and immediate- and long-term consequences.

Results Technical failure occurred in 7 % (95 %CI 5–9). Across 23 centers, 296 patients were analyzed (mean age 71 [SD 16] years, 53 % male), including 100 technical failures: type 1 (26 %), type 2 (41 %), type 3 (11 %), type 4 (6 %), type 5 (16 %). Risk factors for technical failure in multivariate analysis included CBD diameter ≤ 15 mm, duodenal stenosis, wired technique, and low operator experience (≤ 10 LAMS). Endoscopic salvage was successful in 77 %, with 53 % using a covered metal stent and 22 % using a new LAMS. Early failures were mild in 56 % of cases, but 12 % resulted in death within 30 days. Immediate endoscopic salvage reduced severe clinical adverse events (P < 0.001) and increased success rates (P < 0.001).

Conclusions EUS-CDS failures were not rare and were severe in half of cases. Recognizing risk factors, identifying failures during the procedure, and knowing endoscopic salvage methods are crucial.

 

Scroll to top

Introduction

Distal common bile duct (CBD) obstructions are increasingly common, with pancreatic adenocarcinoma being the leading cause [1]. These obstructions are often diagnosed due to jaundice, which significantly affects quality of life, impedes chemotherapy, and complicates surgeries, making biliary drainage crucial [1] [2] [3]. Endoscopic retrograde cholangiopancreatography (ERCP) remains the gold standard procedure for biliary drainage, with a technical success rate ranging from 76.3 % to 91.2 %, and a clinical success rate of between 90.8 % and 97.8 % [3] [4] [5]. However, ERCP fails in 8.8 % to 23.7 % of cases, requiring alternative methods such as endoscopic ultrasound (EUS)-guided biliary drainage, including choledochoduodenostomy (EUS-CDS) [6] [7] [8] [9].

The first EUS-CDS case using a plastic stent was reported in 2001 [10]. Initially considered complex for novice endoscopists, the technique became more accessible with the introduction of fully covered self-expandable metal stents (FC-SEMS) [11]. In 2016, Itoi and Binmoeller introduced the use of lumen-apposing metal stents (LAMS) for EUS-CDS, which simplified the procedure by eliminating the need for prior cystotomy [12] [13]. This innovation led to widespread adoption, with studies reporting high technical (88.5 %–98 %) and clinical (90 %–100 %) success rates [5] [7] [14] [15] [16] [17] [18] [19] [20] [21] [22]. European guidelines now recommend EUS-CDS with LAMS as a first-line treatment in expert centers, especially when ERCP is challenging [6] [23]. Additionally, EUS-CDS with LAMS has been shown to be faster than ERCP, typically requiring 10–14 minutes compared with 23–25 minutes [5] [22] [24]. However, the required technical skill for practitioners performing this procedure remains undefined [19] [25] [26].

Despite being considered straightforward, EUS-CDS has a technical failure rate of 2 %–11 % [5] [7] [14] [15] [16] [17] [18] [19] [20] [21] [22] [26] [27]. A recent review reported a 5.7 % failure rate in LAMS deployment, often attributed to improper positioning of the distal or proximal flange (41 % and 6 % of cases, respectively), with 45 % of cases having unspecified causes [26]. Most misdeployments (87 %) were corrected endoscopically, but some required radiological intervention (4.4 %), rendezvous technique (2.9 %), or surgery (2.9 %) [26]. Although some risk factors for these failures have been identified, there is no standardized classification for EUS-CDS technical failures, nor is there sufficient long-term data on their impact on patient outcomes.

The aims of our study were to identify the risk factors contributing to these technical failures, provide technical precautions for EUS-CDS, describe the endoscopic rescue tailored to different failure types, and evaluate their short- and long-term consequences.

 

Scroll to top

Methods

Patients and definitions

This was a retrospective, multicenter, European, case–control observational study. Adult patients who underwent EUS-CDS with LAMS between December 2018 and January 2024 were included. All patients who underwent EUS-CDS had an indication for CBD drainage (jaundice and bile duct obstruction). Depending on ESGE guidelines, patients either had or had not undergone an ERCP attempt prior to EUS-CDS [3].

Endoscopists from participating centers were contacted by email or phone, and data were collected in a secure, anonymized database by either center investigators or the principal investigator. The study received approval from the ethics committee (IRB:202201185-CHUMontpellier). At each center, patients who underwent EUS-CDS were identified either through systematic patient registry or via stent traceability logs, which record all stent placements in French endoscopy departments.

Cases were defined as patients with technical failure in EUS-CDS drainage, identified either immediately or within 48 hours post-procedure. Technical failure was defined as any difficulty during LAMS placement in EUS-CDS that prevented the stent from achieving the correct position (one flange in the CBD and one in the digestive tract), or where the procedure was hindered by significant bleeding or a stent manufacturing defect. The controls were consecutive patients who achieved technical success with EUS-CDS in the same center during the same period (temporal matching). Technical success of EUS-CDS was defined as the successful placement and functioning (bile duct issue) of a stent between the CBD and the digestive tract without immediate complication. For each case, two controls were selected: the procedures immediately before and after the case. LAMS size, placement method (freehand with direct puncture or over a guidewire), and the decision to place a pigtail stent through the LAMS were determined by the operator, following the practices of the participating center [24].

 

Scroll to top

Data collection

The following data were collected at each center: patient demographics and epidemiological characteristics, underlying causes of biliary obstruction, and specific details regarding the type of stent and placement techniques used. For the cases, additional information on the nature of the failure and the methods used to correct it was documented. Follow-up data were also collected for both cases and controls, including the number and nature of biliary reinterventions, the ability to initiate cancer treatment, survival status, and, when applicable, the cause of death.

 

Scroll to top

Classification of EUS-CDS failures

Based on literature data on EUS-guided gastroenterostomies [28] [29] and the investigators’ own experience, we established a classification for EUS-BD failures ([Fig. 1]).

Fig. 1 Classification of endoscopic ultrasound-guided choledochoduodenostomy failures. LAMS, lumen-apposing metal stent.

• Type 1: Stent only in the CBD; digestive flange mispositioned.

• Type 2: Stent only in the digestive tract; biliary flange mispositioned.

• Type 3: Stent deployment failure (inherent defect in the LAMS).

• Type 4: Catheter passes through the CBD; perforation of the CBD on both sides.

• Type 5: Others – 5a bleeding; 5b complex technical failures and misplacements (probable undetected error with delayed bile peritonitis, misposition in the main pancreatic duct, LAMS in the stomach and no flange in place, LAMS in the retroperitoneum and no flange in place).

 

Scroll to top

Study aims and end points

The primary aim was to identify risk factors contributing to EUS-CDS technical failure. Secondary aims were 1) to evaluate short- and long-term consequences of EUS-CDS technical failure, 2) to describe the effectiveness of endoscopic rescue techniques based on the types of technical failures, and 3) to determine factors associated with clinical success. The severity of short-term consequences was assessed using the AGREE score, which classifies adverse events occurring within 30 days post-drainage [30]. AGREE V corresponds to patient death. Furthermore, severe adverse events were defined as any complication requiring endoscopic reintervention, radiological intervention, surgery, admission to an intensive care unit, or resulting in death due to procedural complications. Minor adverse events were defined as those requiring only additional medical treatment, such as antibiotics. Biliary-related deaths included those linked to the biliary drainage procedure within 30 days, as well as deaths from biliary events unrelated to the procedure, such as cholangitis due to LAMS obstruction. The effectiveness of endoscopic rescue was defined by the absence of severe adverse events combined with clinical success. Clinical success was defined as a 50 % reduction in bilirubin levels by Day 7 or normalization by Day 30 [3] [22] [31].

 

Scroll to top

Statistical analyses

The incidence of technical failure was assessed using data from the five largest centers with prospective databases (Beaujon, Lille, Limoges, Nancy, and Rennes). The Clopper–Pearson method was used to calculate the 95 %CI. The overall analysis includes all cases and controls from the entire study population.

Population characteristics are described as percentages for qualitative variables and means (SD) for quantitative variables. Qualitative variables were compared using Fisher’s exact test or chi-squared test, while quantitative variables were analyzed using the Student’s t test in univariate analyses. Multivariate analysis was performed using stepwise logistic regression to evaluate the strength of associations for identified risk factors. To ensure independence among predictors, a collinearity test was performed before including variables in the multivariate model. Survival analysis was conducted using the Kaplan–Meier method, with patients censored at the time of biliary obstruction or death. A two-sided P value of < 0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Software version 18.3 software (SPSS Inc., Chicago, Illinois, USA).

 

Scroll to top

 

Scroll to top

Results

Population

A total of 23 digestive endoscopy centers (both public and private) participated in the study across France, Belgium, and Switzerland. In total, 296 patients were included: 100 cases and 196 controls. Four controls were missing from the expected number due to insufficient controls in the center (n = 2) and lack of data from consecutive controls (n = 2).

[Table 1] summarizes the population characteristics. Sex distribution was balanced, with 53 % of patients being men (n = 157), and the mean age at the time of drainage was 71 (SD 16) years. Regarding the etiology of biliary obstruction, malignant causes accounted for 95 % of cases (n = 280), with pancreatic adenocarcinoma being the most frequent (77 %, n = 227). Out of the 249 patients (84 %) with available follow-up, the mean follow-up duration was 214 (SD 30) days, and 163 (65 %) patients were deceased at the last update.

Table 1 Population characteristics and risk factors for endoscopic ultrasound-guided choledochoduodenostomy technical failure.

Characteristics	Patients (n = 296)[1]	Cases (n = 100)[1]	Controls (n = 196)[1]	P (univariate analysis)	P (multivariate analysis), OR (95 %CI)
Age, mean (SD), years	71 (16)	71 (12)	73.7 (11)	0.08
Male sex, n (%)	157 (53)	60 (60)	97 (49)	0.09
Malignant obstruction, n (%)	280 (95)	95 (95)	185 (94)	0.83
Pancreatic adenocarcinoma, n (%)	227 (77)	75 (75)	152 (78)	0.36
Stage of cancer disease, n (%)				0.27
I (or operable)[2]	50 (18)	11 (12)	39 (21)
II (or borderline)	46 (16)	16 (17)	30 (16)
III (or locally advanced)	85 (31)	29 (31)	56 (30)
IV (metastatic)	97 (35)	37 (40)	60 (32)
Planned care, n (%)				0.28
Chemotherapy	192 (68)	70 (70)	122 (62)
Surgery	40 (14)	10 (10)	30 (15)
Palliative care	52 (18)	15 (15)	37 (19)
Causes of ERCP failure, n (%)				0.59
Difficult ERCP or papillary invasion	159 (54)	53 (53)	106 (54)
ERCP not attempted	37 (12.5)	9 (9)	28 (14)
Duodenal stenosis	90 (30)	38 (38)	52 (27)
Ascites the day of EUS-CDS, n (%)	20	8 (8)	12 (6)	0.53
Bilirubin before drainage, mean (SD), mmol/L	231 (148)	217 (140)	241 (152)	0.28
Main bile duct diameter, mean (SD), mm	17.3 (4.0)	15.8 (3.3)	18 (4)	< 0.001
Main bile duct diameter ≤ 15 mm, n (%)	58 (30)	36 (36)	22 (11)	< 0.001	< 0.001 2.5 (1.46–4.27)
Operator experience, n (%)
≤ 5 EUS-CDS procedures	44 (22)	23 (23)	21 (11)	0.005
≤ 10 LAMS placements	52 (27)	28 (28)	24 (12)	< 0.001	0.03 2.10 (1.06–4.15)
Antral location (vs. bulb), n (%)	7 (4 %)	6 (6)	1 (1)	0.007	–[3]
Technique “Wired”, n (%)	49 (25)	23 (23)	15 (8)	< 0.001	0.008 2.86 (1.32–6.21)
Coaxial double-pigtail stent, n (%)	16 (8)	5 (5)	11 (6)	0.83
Duodenal stenosis at drainage, n (%)	90 (30)	38 (38)	52 (27)	0.04	0.02 1.97 (1.12–3.46)

ERCP, endoscopic retrograde cholangiopancreatography; EUS-CDS, endoscopic ultrasound-guided choledochoduodenostomy; LAMS, lumen-apposing metal stent; OR, odds ratio.

¹ The total number of patients varies across different characteristics due to missing data. Percentages are calculated based on the number of patients with available data for each variable rather than the full cohort (N = 296). This variability arises from differences in data collection across centers and the nonapplicability of certain parameters to all patients.

² The stage of cancer disease () pertains more specifically to pancreatic cancer.

³ Not included due to total number < 10.

 

Scroll to top

Technical failure rate

We estimated the incidence of technical failures to be 7 % (95 %CI 5–9), based on 43 failures among 618 EUS-CDS procedures across the five centers with prospective databases.

 

Scroll to top

Primary end point: risk factors for EUS-CDS failure

In univariate analysis, risk factors for EUS-CDS failure were a CBD diameter of ≤ 15 mm (P < 0.001), operators with experience of ≤ 10 LAMS drainage procedures (P = 0.0009), or ≤ 5 EUS-CDS procedures (P = 0.005), antral location of drainage (P = 0.007), and the presence of duodenal stenosis at the time of drainage (P = 0.04) ([Table 1]). The freehand technique was a protective factor compared with the wired technique (P < 0.001). In contrast, the presence of ascites at the time of drainage was not significantly associated with failure (P = 0.53).

In multivariate analysis, the independent risk factors for technical failure of EUS-CDS were CBD ≤ 15 mm (P < 0.001, OR 2.5 [95 %CI 1.46–4.27]), operator having performed ≤ 10 LAMS drainage procedures (P = 0.03, OR 2.10 [95 %CI 1.06–4.15]), duodenal stenosis at the time of drainage (P = 0.02, OR 1.97 [95 %CI 1.12–3.46]), and use of the wired technique (P = 0.008, OR 2.86 [95 %CI 1.32–6.21]) ([Table 1]).

 

Scroll to top

Endoscopic rescue and clinical consequences based on EUS-CDS failure type

From most to least frequent, the failures were type 2 (41 %, n = 41), followed by type 1 (26 %, n = 26), type 5 (16 %, n = 16), type 3 (11 %, n = 11), and finally type 4 (6 %, n = 6) among the 100 cases. [Fig. 2] illustrates the types of technical failures and the frequency of successful endoscopic rescue, defined as the absence of severe adverse events and achievement of clinical success.

Fig. 2 Types of technical failure with frequency of successful endoscopic salvage and clinical consequences. CBD, common bile duct.

Type 2 failures (n = 41 [41 %]; severe adverse events 35 %)

Endoscopic salvage was possible in 73 % of cases (n = 30). This was most frequently achieved by placing a guidewire through the LAMS catheter still in position within the CBD and deploying an FC-SEMS (n = 19; 46 %). In nine patients (22 %), a new EUS-CDS was performed immediately with another LAMS.

 

Scroll to top

Type 1 failures (n = 26 [26 %]; severe adverse events 41 %)

Endoscopic salvage was possible in 81 % of cases (n = 21), primarily by advancing a guidewire into the CBD through the LAMS catheter and placing an FC-SEMS and/or a pigtail stent (n = 19; 90 %).

 

Scroll to top

Type 3 failures (n = 11 [11 %]; severe adverse events 27 %)

Endoscopic salvage was possible in 91 % by deploying an FC-SEMS over a guidewire.

 

Scroll to top

Type 4 failures (n = 6 [6 %]; severe adverse events 66 %)

Endoscopic salvage was only possible in half of cases (n = 3). Successful endoscopic rescue involved either a new EUS-CDS with guidewire and FC-SEMS placement or repositioning the catheter with a new LAMS for guidewire insertion. In cases where immediate endoscopic rescue failed (n = 3), radiological drainage or emergency surgery was required, often resulting in worse outcomes.

 

Scroll to top

Type 5 failures (n = 16 [16 %]; severe adverse events 56 %)

Six patients experienced bleeding (type 5a) that obstructed the drainage procedure but was managed endoscopically using clips or a hemostatic stent. In two patients, no adverse event was noted during the procedure, but 24 to 48 hours later, they presented with clinical signs of sepsis revealing biliary peritonitis, despite the LAMS being in place. This was likely due to an undetected biliary lesion or a coagulation injury to the contralateral biliary wall. For six patients, the failure was either due to unsuccessful bile duct puncture/stent placement using the wired technique (n = 2) or the stent was not in place (neither flange in place: in the stomach (n = 2) or in the retroperitoneum (n = 2). Finally, the last two patients experienced LAMS misplacement in the main pancreatic duct. This misposition led to the death of one patient, while in the other, the stent was left in place and successful salvage gallbladder drainage was performed for biliary decompression.

No significant differences were found among the five failure types in terms of endoscopic salvage (P  = 0.57), severe adverse events (P  = 0.22), or clinical success (P  = 0.65).

 

Scroll to top

 

Scroll to top

Techniques and additional interventions for endoscopic salvage after failed EUS-CDS

The endoscopic salvage techniques employed varied significantly. In 53 % of cases (n = 53), the issue was promptly identified, and a guidewire was placed through the LAMS catheter still in position within the CBD, allowing for FC-SEMS deployment. In 22 % of technical failures (n = 22), the CBD was re-punctured with a new LAMS. 

Other endoscopic approaches were used, sometimes complementarily: 11 (11 %) ERCP, 8 (8 %) hepaticogastric drainage procedures, and 4 (4 %) cholecystogastric drainage procedures. Endoscopic rescue was immediately a technical success in 77 cases (77 %). However, additional interventions were necessary in some patients: 11 % needed radiological drainage, while 6 % required emergency surgery. Six patients did not receive immediate treatment due to spontaneously resolving hemorrhages, LAMS dislodgment before cystotome use (wired technique), or lack of further therapeutic options after a failed salvage attempt (n = 3). Fig. 1 s (see the online-only Supplementary material) illustrates an example of successful endoscopic salvage for a type 2 EUS-CDS failure.

 

Scroll to top

Early clinical outcomes following EUS-CDS technical failures

Among the 100 cases, 56 % (n = 56) of early adverse events (< 30 days) were non-severe, while 44 % (n = 44) were serious, with 12 % (n = 12) resulting in death. Among these, four deaths were attributed to limitations in care following EUS-CDS failure, while seven were directly related to drainage failure (e. g. biliary peritonitis, severe hemorrhage). One patient died from a pulmonary embolism after endoscopic salvage.

When comparing 30-day mortality rates, cases exhibited a significantly higher mortality rate of 12 % compared with 5 % (n = 10) in the control group (P = 0.03). Specifically, 10 patients in the control group died within 30 days, with causes including anesthesia complications (n = 2), septic shock from an obstructed biliary stent (n = 1), and cancer-related deterioration (n = 7).

Immediate successful endoscopic salvage was associated with significantly fewer severe adverse events and higher clinical success compared with cases requiring radiological, surgical, or delayed salvage (P < 0.001 and P < 0.001, respectively) ([Table 2]).

Table 2 Early clinical outcomes and AGREE classification following technical failures in endoscopic ultrasound-guided choledochoduodenostomy.

	Total EUS-CDS failure (n = 100)[1]	EUS-CDS with endoscopic technical rescue success (n = 77)[1]	EUS-CDS with endoscopic technical rescue failure (n = 23)[1]	P (univariate)
Morbidity, AGREE class, n (%)				NA[2]
I	43 (43)	41 (53)	2 (9)
II	13 (13)	13 (17)	0 (0)
IIIa	21 (21)	11 (14)	10 (43)
IIIb	5 (5)	3 (4)	2 (9)
Iva	6 (6)	2 (3)	4 (17)
V	12 (12)	7 (9)	5 (22)
Severe adverse events, n (%)	44 (44)	22 (29)	22 (96)	< 0.001
Mortality (30 days) or AGREE V, n (%)	12 (12)	7 (9)	5 (22)	0.13
Clinical success, n (%)	80 (80)	68 (91)	12 (55)	< 0.001
Effectiveness of endoscopic rescue, n (%)	54 (54)	53 (69)	1 (4)[3]	< 0.001

EUS-CDS, endoscopic ultrasound-guided choledochoduodenostomy; NA, not applicable.

¹ The total number of patients varies across different variables due to missing data. Percentages are calculated based on the number of patients with available data for each variable rather than the full cohort.

² Some groups had fewer than 5 patients, so the chi-squared test was not applicable.

³ This patient underwent a second endoscopic procedure 48 hours later (endoscopic retrograde cholangiopancreatography), which allowed for effective drainage of the bile ducts.

 

Scroll to top

Medium and long-term consequences of EUS-CDS failure

In univariate analysis, clinical success was significantly higher among controls (93 %) than among cases (81 %) (P = 0.001) ([Table 3]). No significant differences were observed in the initiation of chemotherapy (P = 0.30), potential cancer surgery (P = 0.60), or long-term biliary dysfunction (P = 0.60). Cholangitis rates over a mean follow-up of 214 (SD 30) days were comparable between the two groups. While overall mortality was comparable between groups (P = 0.70), the disparity in biliary-related deaths highlights the impact of technical failure in EUS-CDS (31 % vs. 6 %, P < 0.001). Fig. 2 s shows the survival curves.

Table 3 Medium- and long-term consequences of endoscopic ultrasound-guided choledochoduodenostomy failure.

	Cases (n = 100)	Controls (n = 196)	Univariate analysis P
Bilirubin 48 hours after drainage, mean (SD), mmol/L	160.5 (137)	140.5 (107)	0.24
Clinical success, n (%)	81 (81)	182 (93)	0.001
Chemotherapy started, n (%)	56 (56)	113 (58)	0.30
Surgery performed, n (%)	9 (9)	17 (9)	0.60
LAMS dysfunction, n (%)	16 (16)	33 (17)	0.91
Of which cholangitis, n (%)	8 (50)	22 (67)	0.39
Deaths, n (%) (N%)	55 (55)	108 (55)	0.70
Biliary-related deaths, n (%)	17 (31)	6 (6)	< 0.001

LAMS, lumen-apposing metal stent.

 

Scroll to top

 

Scroll to top

Discussion

To our knowledge, this study is the first to comprehensively detail the variety of failures in EUS-CDS with LAMS and their endoscopic salvage. Clinical consequences were diverse but significantly impacted 44 % of cases. Immediate endoscopic salvage, when feasible, significantly reduced the incidence of severe outcomes. Finally, we identified risk factors for failure, including CBD, the presence of duodenal stenosis, antral drainage location, the use of the wired technique, and operator experience.

Previous studies on EUS-CDS suggested the rarity (0 % to 5 %) of technical failure and the ease of EUS-CDS with LAMS [5] [14] [15] [16] [17]. In our cohort, the failure rate was estimated at 7 %, consistent with literature [5] [14] [15] [16] [17]. This indicates that such events may not be as rare as previously thought. Additionally, the 12 % 1-month mortality rate observed in our study is particularly concerning, especially given the global increase in the number of EUS-CDS procedures. In 2022, 1450 LAMSs (6 × 8 mm HOT-AXIOS; Boston Scientific, Marlborough, Massachusetts, USA) specifically designed for EUS-CDS were sold in France by Boston Scientific, representing a twofold increase compared with 2020 (data provided by Boston Scientific). Based on this volume, we estimate approximately 100 technical failures annually in France, resulting in 10 to 20 patient deaths per year. Nevertheless, our study included a highly fragile patient population. Indeed, the 30-day mortality was not zero in the control group, with deaths related to anesthesia but mostly due to the rapid progression of cancer. However, mortality was significantly higher in the cases and was directly linked to the drainage procedure.

The most common types of failures were types 1 and 2, where one of the two flanges was positioned incorrectly [26]. These failures are likely attributable to excessive traction force during LAMS deployment, an overly long distance between the duodenum and the CBD relative to the stent size (particularly when the wall-to-wall distance is measured under duodenal pressure, which can artificially shorten it), or abnormal rigidity or peristalsis that impairs proper tissue apposition ([Video 1]). In our study, no correlation was found between technical failures and the wall-to-wall distance, as it was considered normal by the endoscopists and lacked precise data.

Video 1 Successful endoscopic rescue of a type 2 technical failure of endoscopic ultrasound-guided choledochoduodenostomy (courtesy of Enrique Perez-Cuadrado-Robles, Department of Gastroenterology, Hopital européen Georges Pompidou [APHP], Paris, France).

Additionally, immediate endoscopic salvage was highly beneficial for patients. Biliary perforation requires immediate treatment, as postponing management in the hope of spontaneous resolution is not advisable [32]. For example, when a new puncture is necessary, the newly placed LAMS can effectively seal the initial puncture site, covering the biliary defect. This study highlighted various possible endoscopic salvage methods, the main ones being guiding a wire into the intrahepatic bile ducts and placing a new LAMS or FC-SEMS. However, these options were not always feasible, effective, or sufficient. The success of the salvage strategy largely depends on the timely recognition of the failure and whether the LAMS catheter remains in place within the CBD. An algorithm illustrating the endoscopic response based on the timing of failure recognition is provided in [Fig. 3].

Fig. 3 Management algorithm for endoscopic ultrasound-guided choledochoduodenostomy technical failure. CBD, common bile duct; CT, computed tomography; ERCP, endoscopic retrograde cholangiopancreatography; FC-SEMS, fully covered self-expandable metal stent; LAMS, lumen-apposing metal stent; MBD, main bile duct; NPO, nil per os.

Several risk factors should be considered before performing EUS-CDS to minimize failure rates. Duodenal stenosis poses a significant risk, likely due to the increased thickness and rigidity of the duodenal wall, which hinders proper tissue contact and increases the likelihood of stent dislodgment. Additionally, duodenal stenosis is associated with a higher incidence of long-term stent obstruction [5] [15] [22]. Consequently, EUS-CDS should generally be avoided in cases of duodenal stenosis, as it compromises both the safety and efficacy of the procedure. Similarly, a CBD diameter of ≤ 15 mm is a well-documented risk factor for obstruction and for EUS-CDS placement failure. To ensure adequate deployment of the first flange of the LAMS, an 18-mm space is theoretically required when using a 6 × 8 mm HOT-AXIOS stent. Therefore, the 15-mm threshold identified is consistent with findings from previous literature [5] [15] [22] [23] [26].

Although rarely used, the antral position should generally be avoided due to its association with a higher failure rate. Such positioning occurred either unintentionally (e. g. unintended retraction during puncture in an unstable position) or when it was the only location where the CBD diameter exceeded 15 mm. The antropyloric region, marked by significant peristalsis and increased thickness, presents additional challenges and is generally not suitable for this procedure [7] [8] [26].

The wired technique has proven to be problematic. Initial puncture with a needle followed by guidewire insertion introduces additional complexity to the procedure and may increase the likelihood of technical failure. In cases with minimally dilated CBD, experienced operators familiar with EUS-CDS using traditional FC-SEMS might be tempted to apply the wired technique with LAMS. However, our study found that using the wired technique was consistently associated with technical failure, regardless of CBD diameter. It is likely that inserting the guidewire prior to deploying the first flange of the LAMS shifts the catheter backward, reducing the space necessary for proper flange release [26].

Finally, the operator’s experience emerged as a risk factor for failure [25]. Although this may seem intuitive, our study is the first to emphasize the critical importance of mastering LAMS placement, recommending at least 10 prior placements as a minimum prerequisite, to reduce the risk of adverse events in EUS-CDS. While this method is promoted as easy with a rapid learning curve, it still requires extensive expertise, particularly in cases of technical failure. Even though the technique is theoretically straightforward, successful salvage strategies demand substantial experience in interventional endoscopy [26] [33].

At a time when ERCP is being challenged by new biliary drainage techniques, particularly EUS-CDS, this study provides a critical perspective on its simplicity and safety in clinical practice. While ERCP has been historically more widely available, it is more complex and has a failure rate of 15 %–20 %, even among experts managing tumor-related cases [8]. In contrast, EUS-CDS appears straightforward due to its rapidity and frequent success. However, it should be reserved for centers with the necessary technical expertise in managing potential failures, including interventional radiologists, surgeons, and intensivists. This caution is essential because EUS-CDS fails in about 7 % of cases, and when it does, we observed a 12 % mortality rate. In contrast, ERCP has a 15 %–20 % failure rate and can lead to pancreatitis in less than 5 % of cases involving malignant biliary obstruction, with a mortality rate likely below 15 % [3] [8]. Therefore, despite its apparent ease, EUS-CDS requires a skilled environment to ensure effective salvage and minimize adverse outcomes. Its role as a first-line intervention remains unclear and warrants further investigation.

The main limitation of the study is its retrospective and self-reported nature regarding the cases, except for centers with prospective databases. Additionally, due to the large number of centers, there was heterogeneity among patients and practices. However, the strength of this study lies in the substantial number of cases included for an event that remains underreported. The study design (case–control) is debatable, but given that this is a relatively rare event (7 % reported here), to collect data on all patients with successful procedures from the centers was not feasible. Instead, collecting control data consecutively to the cases (one before and one after) within the same center helped to limit potential temporal bias, especially regarding the operators’ experience. Furthermore, due to the inclusion of EUS-CDS procedures using only LAMS, the results of this study cannot be extended to EUS-CDS with other stents. Furthermore, with nearly all stents being 6 × 8 mm (fewer than 10 exceptions out of 296), differences in technical failure risk based on stent size could not be assessed. Despite these limitations, our study offers novel insights into the technical challenges of EUS-CDS with LAMS and highlights actionable factors to improve patient outcomes.

In conclusion, EUS-CDS failures, while uncommon, can lead to severe consequences, with approximately 12 % of patients experiencing fatal outcomes within 30 days. Recognizing risk factors such as a CBD diameter ≤ 15 mm or the presence of duodenal stenosis prior to the procedure, promptly identifying technical failures during the procedure, and mastering tailored endoscopic salvage techniques are critical for mitigating risks. Immediate endoscopic management has been shown to significantly reduce the severity of adverse events and improve clinical outcomes. These findings underscore the need for careful patient selection and a high level of technical expertise when performing EUS-CDS.

 

Scroll to top

 

Scroll to top

Conflict of Interests

M. Schaefer, E. Perez-Cuadrado-Robles, A. Belle, J.M. Gonzalez, A. Lemmers, R. Gérard, and J. Albouys are consultants for Boston Scientific (which markets the LAMS used for EUS-CDS). The remaining authors declare that they have no conflict of interest.

• References

• 1 Neuzillet C, Gaujoux S, Williet N. et al. Pancreatic cancer: French clinical practice guidelines for diagnosis, treatment and follow-up (SNFGE, FFCD, GERCOR, UNICANCER, SFCD, SFED, SFRO, ACHBT, AFC). Dig Liver Dis 2018; 50: 1257-1271
• 2 Pavlidis ET, Pavlidis TE. Pathophysiological consequences of obstructive jaundice and perioperative management. Hepatobiliary Pancreat Dis Int 2018; 17: 17-21
• 3 Dumonceau J-M, Tringali A, Papanikolaou I. et al. Endoscopic biliary stenting: indications, choice of stents, and results: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline – updated October 2017. Endoscopy 2018; 50: 910-930
• 4 Bang JY, Navaneethan U, Hasan M. et al. Stent placement by EUS or ERCP for primary biliary decompression in pancreatic cancer: a randomized trial (with videos). Gastrointest Endosc 2018; 88: 9-17
• 5 Teoh AYB, Napoleon B, Kunda R. et al. EUS-guided choledocho-duodenostomy using lumen apposing stent versus ERCP with covered metallic stents in patients with unresectable malignant distal biliary obstruction: a multicenter randomized controlled trial (DRA-MBO trial). Gastroenterology 2023; 165: 473-482.e2
• 6 van der Merwe SW, van Wanrooij RLJ, Bronswijk M. et al. Therapeutic endoscopic ultrasound: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2022; 54: 185-205
• 7 Amato A, Sinagra E, Celsa C. et al. Efficacy of lumen-apposing metal stents or self-expandable metal stents for endoscopic ultrasound-guided choledochoduodenostomy: a systematic review and meta-analysis. Endoscopy 2021; 53: 1037-1047
• 8 Holt BA, Hawes R, Hasan M. et al. Biliary drainage: role of EUS guidance. Gastrointest Endosc 2016; 83: 160-165
• 9 Debourdeau A, Daniel J, Caillo L. et al. Effectiveness of endoscopic ultrasound (EUS)‐guided choledochoduodenostomy vs. EUS‐guided gallbladder drainage for jaundice in patients with malignant distal biliary obstruction after failed endoscopic retrograde cholangiopancreatography: retrospective, multicenter study (GALLBLADEUS Study). Dig Endosc 2025; 37: 103-114
• 10 Giovannini M, Moutardier V, Pesenti C. et al. Endoscopic ultrasound-guided bilioduodenal anastomosis: a new technique for biliary drainage. Endoscopy 2001; 33: 898-900
• 11 Isayama H, Nakai Y, Kawakubo K. et al. Recent progress in endoscopic ultrasonography guided biliary intervention. Clin J Gastroenterol 2012; 5: 93-100
• 12 Itoi T, Binmoeller KF. EUS-guided choledochoduodenostomy by using a biflanged lumen-apposing metal stent. Gastrointest Endosc 2014; 79: 715
• 13 Walter D, Will U, Sanchez-Yague A. et al. A novel lumen-apposing metal stent for endoscopic ultrasound-guided drainage of pancreatic fluid collections: a prospective cohort study. Endoscopy 2015; 47: 63-67
• 14 Geyl S, Redelsperger B, Yzet C. et al. Risk factors for stent dysfunction during long-term follow-up after EUS-guided biliary drainage using lumen-apposing metal stents: a prospective study. Endosc Ultrasound 2023; 12: 237-244
• 15 Vanella G, Bronswijk M, Dell’Anna G. et al. Classification, risk factors, and management of lumen apposing metal stent dysfunction during follow-up of endoscopic ultrasound-guided choledochoduodenostomy: multicenter evaluation from the Leuven–Amsterdam–Milan Study Group. Dig Endosc 2023; 35: 377-388
• 16 Jacques J, Privat J, Pinard F. et al. Endoscopic ultrasound-guided choledochoduodenostomy with electrocautery-enhanced lumen-apposing stents: a retrospective analysis. Endoscopy 2019; 51: 540-547
• 17 Ginestet C, Sanglier F, Hummel V. et al. EUS-guided biliary drainage with electrocautery-enhanced lumen-apposing metal stent placement should replace PTBD after ERCP failure in patients with distal tumoral biliary obstruction: a large real-life study. Surg Endosc 2022; 36: 3365-3373
• 18 Krishnamoorthi R, Dasari CS, Thoguluva Chandrasekar V. et al. Effectiveness and safety of EUS-guided choledochoduodenostomy using lumen-apposing metal stents (LAMS): a systematic review and meta-analysis. Surg Endosc 2020; 34: 2866-2877
• 19 Albouys J, Guilmoteau T, Schaefer M. et al. How to prevent and treat biliary lumen-apposing metal stent dysfunction?. Dig Endosc 2024; 36: 492-494
• 20 Artifon ELA, Aparicio D, Paione JB. et al. Biliary drainage in patients with unresectable, malignant obstruction where ERCP fails: endoscopic ultrasonography-guided choledochoduodenostomy versus percutaneous drainage. J Clin Gastroenterol 2012; 46: 768
• 21 Han SY, Kim S-O, So H. et al. EUS-guided biliary drainage versus ERCP for first-line palliation of malignant distal biliary obstruction: a systematic review and meta-analysis. Sci Rep 2019; 9: 1-9
• 22 Chen Y-I, Sahai A, Donatelli G. et al. Endoscopic ultrasound-guided biliary drainage of first intent with a lumen-apposing metal stent vs endoscopic retrograde cholangiopancreatography in malignant distal biliary obstruction: a multicenter randomized controlled study (ELEMENT trial). Gastroenterology 2023; 165: 1249-1261.e5
• 23 Van Wanrooij RLJ, Bronswijk M, Kunda R. et al. Therapeutic endoscopic ultrasound: European Society of Gastrointestinal Endoscopy (ESGE) Technical Review. Endoscopy 2022; 54: 310-332
• 24 Dietrich CF, Braden B, Burmeister S. et al. How to perform EUS-guided biliary drainage. Endosc Ultrasound 2022; 11: 342-354
• 25 Biau DJ, Williams SM, Schlup MM. et al. Quantitative and individualized assessment of the learning curve using LC-CUSUM. Br J Surg 2008; 95: 925-929
• 26 Armellini E, Metelli F, Anderloni A. et al. Lumen-apposing-metal stent misdeployment in endoscopic ultrasound-guided drainages: a systematic review focusing on issues and rescue management. World J Gastroenterol 2023; 29: 3341-3361
• 27 Khoury T, Sbeit W, Fumex F. et al. Endoscopic ultrasound vs. ERCP-guided primary drainage of inoperable distal malignant biliary obstruction: systematic review and meta-analysis of randomized controlled trials. Endoscopy 2024; 56: 955-963
• 28 Vanella G, Dell’Anna G, Capurso G. et al. EUS-guided gastroenterostomy for management of malignant gastric outlet obstruction: a prospective cohort study with matched comparison with enteral stenting. Gastrointest Endosc 2023; 98: 337-347.e5
• 29 Ghandour B, Bejjani M, Irani SS. et al. Classification, outcomes, and management of misdeployed stents during EUS-guided gastroenterostomy. Gastrointest Endosc 2022; 95: 80-89
• 30 Nass KJ, Zwager LW, van der Vlugt M. et al. Novel classification for adverse events in GI endoscopy: the AGREE classification. Gastrointest Endosc 2022; 95: 1078-1085.e8
• 31 Fritzsche JA, Fockens P, Besselink MG. et al. Endoscopic ultrasound-guided choledochoduodenostomy using single-step lumen-apposing metal stents for primary drainage of malignant distal biliary obstruction (SCORPION-p): a prospective pilot study. Endoscopy 2024; 56: 47-52
• 32 Çavuşoğlu SD, Doğanay M, Birben B. et al. Management of bile duct injuries: a 6-year experience in a high volume referral center. Euroasian J Hepatogastroenterol 2020; 10: 22-26
• 33 Elmassry M, Haché B, Thongpiya J. et al. Endoscopic ultrasound–guided choledochoduodenostomy with lumen-apposing metal stent through duodenal stent, a success case, and a salvage case. ACG Case Rep J 2024; 11: e01315

Publication History

Received: 11 June 2024

Accepted after revision: 17 February 2025

Accepted Manuscript online:
17 February 2025

Article published online:
14 May 2025