How are you taking part in this consultation?

You will not be able to change how you comment later.

You must be signed in to answer questions

    The content on this page is not current guidance and is only for the purposes of the consultation process.

    Interventional procedure overview of endoscopic ultrasound-guided gallbladder drainage for acute cholecystitis when surgery is not an option

    Closed for comments This consultation ended on at   Request commenting lead permission

    Evidence summary

    Population and studies description

    This interventional procedures overview is based on more than 15,000 patients from 3 systematic reviews (Fabbri 2022; Mohan 2020; Podboy 2021), 1 randomised controlled trial (Teoh 2020), 2 non-randomised comparative studies (Siddiqui 2019; Teoh 2021), 1 registry (Teoh 2019), 1 case series (Torres Yuste 2019) and 1 case report (Kim 2019). About 1,000 patients had the procedure. There was some overlap in patients between the studies and overlap in papers between the systematic reviews. Some of the studies summarised separately were also included in at least 1 of the reviews. The randomised controlled trial by Teoh et al. (2020) and the registry study by Teoh et al. (2017) were included in the systematic reviews by Podboy et al. (2021) and Fabbri et al. (2022). The non-randomised comparative studies by Siddiqui et al. (2019) and Teoh et al. (2021) were included in Podboy et al. (2021).

    This is a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. This overview presents 9 studies as the key evidence in table 2 and table 3, and lists 68 other relevant studies in table 5. Because of the large body of evidence, case series with 10 or fewer patients have been excluded.

    Of the 72 studies included in the systematic review by Mohan et al. (2020), 7 were prospective and the rest were retrospective. Based on the Newcastle – Ottawa scale for cohort studies to assess bias, 37 studies were considered to be high quality, 33 studies were of medium quality, and 2 studies were considered to be of low quality. Of the 27 studies in Fabbri et al. (2022), 6 were prospective (including 1 randomised controlled trial) and the rest were retrospective. The authors described the overall quality of the included studies as sufficient. The only randomised controlled trial was at low risk of bias. Of the observational studies, 7 were high quality and the remaining studies were at high risk of bias. No study was at very high risk of bias. Of the 10 studies included in Podboy et al. (2021), 2 were randomised controlled trials and 8 were retrospective case cohort studies (1 matched 1:1 and 1 propensity score matched). The overall quality of evidence was evaluated using the Grading of Recommendations Assessment Development and Evaluation method. Of the retrospective studies, all 8 were considered very low quality, with studies lacking randomisation and allocation concealment. For the 2 randomised controlled trials, the level of evidence was considered moderate to low.

    The systematic reviews did not report where the individual studies were done. The other studies included patients from Hong Kong, US, Denmark, Spain, Korea, China, Australia, India, Thailand and Singapore. In the review by Podboy et al. (2021), the median follow up was 212.5 days for patients who had EUS-GBD. The longest median follow up was 24.4 months, but this was a retrospective case series with only 22 patients (Torres Yuste 2019).

    All studies included patients with acute cholecystitis for whom cholecystectomy was a high risk. The mean or median age ranged from 63 to 88 years. In the systematic review by Fabbri et al. (2022), 33% of patients had an underlying malignancy.

    Most of the comparative studies compared EUS-GBD with PT-GBD or ETP-GBD, but 1 compared it with LC (Teoh 2021).

    Table 2 presents study details.

    Figure 1 Flow chart of study selection

    Table 2 Study details

    Study no.

    First author, date

    country

    Patients (male: female)

    Age

    Study design

    Inclusion criteria

    Intervention

    Follow up

    1

    Fabbri C, 2022

    Country not reported for individual studies

    n=1,004 (56.3% male; 95% CI 51.3% to 61.2%)

    Weighted mean 74 years (95% CI 71.9 to 76.1)

    Systematic review and meta-analysis (27 studies); patients were enrolled between February 2009 and February 2020.

    A study population which included patients treated with transmural EUS-GBD through metal stents for acute cholecystitis; an almost complete description of the baseline patients' clinical features analysed; and a complete description of technical and clinical success rates, together with the description of post-procedural adverse events.

    Transmural EUS-GBD; 64% of procedures used a duodenal approach (n=699).

    Number of stents used: 654 LAMS, 205 ATSEMS, 145 SEMS or metal stent

    Not reported

    2

    Mohan B, 2020

    Country not reported for individual studies

    n=15,131 (61% male)

    The mean or median age ranged from 65 years to 85 years.

    Systematic review and comparative meta-analysis

    (72 studies)

    Studies that evaluated EUS-GBD, ETP-GBD, and PT-GBD in patients with acute cholecystitis who were considered high risk for immediate cholecystectomy because of age or underlying chronic disease processes such as malignancy were included.

    Exclusion criteria: studies on gallbladder drainage for malignant stricture of the bile ducts, studies with a sample size of fewer than 10 patients, studies conducted in a paediatric population (age less than 18 years), and studies not published in the English language.

    EUS-GBD (n=557), ETP-GBD (n=1,223) or PT-GBD (n=13,351).

    Not reported

    3

    Podboy A, 2021

    Country not reported for individual studies

    n=1,267

    EUS-GBD 51% male

    PT-GBD 58% male

    ETP-GBD 67% male

    Median 71.0 years (range 62 to 82.7) for EUS-GBD, 74.7 years (range 60.4 to 81.2) for PT-GBD and 64.1 years (range 60 to 77.9) for ETP-GBD

    Systematic review and network meta-analysis

    (10 studies)

    Comparative studies assessing the efficacy of either of 2 or all modalities (PT-GBD, EUS-GBD, and ETP-GBD) used for treating acute cholecystitis in patients at high risk for cholecystectomy. Studies were included if they reported all 3 primary outcomes of interest: (1) technical success, defined as the ability to access and drain the gallbladder by placement of a drainage tube or stent with immediate drainage of bile; (2) clinical success, defined as resolution of clinical symptoms or improvement in biochemical parameters; and (3) postprocedure adverse events.

    EUS-GBD (n=472), PT-GBD (n=493) and ETP-GBD (n=302).

    Wide variations in anatomic puncture location, stent size, and stent type were noted within the included studies.

    Median 212.5 days for EUS-GBD, 177.5 days for PT-GBD, and 150 days for ETP-GBD

    4

    Teoh A, 2020

    Hong Kong

    n=79

    (42:37)

    Mean 81.9 years (EUS-GBD);

    79.8 years (percutaneous)

    Randomised controlled trial

    Consecutive patients aged 18 or over with grade 2 or 3 acute calculous cholecystitis, at very high risk for cholecystectomy.

    Patients were deemed very high risk if they satisfied 1 or more of the following criteria: Age 80 or above, American Society of Anesthesiology grade 3 or above, age adjusted Charlson Comorbidity Index above 5 or Karnofsky score less than 50.

    EUS-GBD using LAMS (n=39) versus PT-GBD (n=40)

    1 year

    5

    Siddiqui A, 2019

    US, Denmark

    n=372

    (233:139)

    Mean 63 years

    Non-randomised comparative study (retrospective)

    Patients with acute cholecystitis who were not candidates for cholecystectomy. Only patients with at least 3 months follow up were included.

    EUS-GBD using LAMS (n=102), PT-GBD (n=146), ETP-GBD (n=124)

    Median 3 months (range 3 to 9)

    6

    Teoh A, 2021

    Hong Kong

    n=60

    (35:25)

    Mean 78 years (EUS-GBD); 76.4 years (LC)

    Non-randomised comparative study (propensity score analysis of a retrospective database)

    Patients with acute cholecystitis who had EUS-GBD or LC as a definitive management. In the EUS-GBD group, patients were at very high risk for cholecystectomy, satisfying 1 or more of the following criteria: age 80 years or above, ASA grade 3 or above, age adjusted Charlson comorbidity index more than 5, Karnofsky score less than 50. Elderly patients who refused surgery were also included.

    In the LC group, patients were surgically fit and presented with fewer than 7 days of pain.

    EUS-GBD within 6 to 8 hours of diagnosis (n=30) or LC within 24 hours of diagnosis (n=30).

    Device used for EUS-GBD: LAMS (AXIOS, Boston Scientific, US). An additional double-pigtail plastic stent was inserted through the LAMS if there was concern that the gallstones could block the stent.

    Mean 571.9 days (EUS-GBD); 264.4 days (LC), p=0.002

    7

    Teoh A, 2019

    Hong Kong, Spain, Denmark, US, Korea, China, Australia, India, Thailand, Singapore

    n=379 (199:180)

    Mean 73.6 years

    Registry (international retrospective)

    Consecutive patients who had EUS-GBD planned and attempted for symptomatic gallstones, acute cholecystitis or conversion of percutaneous cholecystostomy were included. Cholecystectomy was considered to be unsuitable if patients satisfied 1 or more of the following criteria: Age 80 years or above, ASA grade 3 or above, age adjusted Charlson score 4 or higher or Karnofsky score less than 50.

    EUS-GBD using a variety of LAMS and stents with antimigration designs. Single-step (including the HOT AXIOS stent, Boston Scientific) and multi-step (including the SPAXUS stent, Taewoong Medical, Korea; the BONA-AL stent, Standard Sci Tech Inc., Korea; and the Microtech stent, Nan Jing Co. Ltd., China) devices were used.

    Mean 433.6 days

    8

    Torres Yuste, 2019

    Spain

    n=22

    (14: 8)

    Median 88.3 years (IQR 82.6 to 92.7 years)

    Case series (retrospective)

    Patients who had EUS-GBD for acute cholecystitis. Cholecystectomy was dismissed in all cases because of the patients' advanced age and poor physical status (class 3 or higher on the ASA Physical Status classification). None of the patients improved after 24 to 72 hours of conservative management with intravenous antibiotics, fluid replacement and bowel rest.

    Patients with a stent indwell time less than 12 months were excluded.

    EUS-GBD using LAMS. The deployment technique depended on the type of stent used. With a conventional LAMS (AXIOS, Boston Scientific) serial dilation with cystotome followed by a biliary balloon was done before insertion of the stent under EUS and fluoroscopic guidance. In case of an electrocautery-enhanced (hot) LAMS (AXIOS, Boston Scientific) the stent was deployed directly over the guidewire.

    Median 24.4 months (IQR (18.2 to 42.4 months)

    9

    Kim J, 2019

    US

    n=2

    (0: 2)

    80 and 86 years

    Case reports

    Patients with severe retrograde reflux of gastric contents into the gallbladder after EUS-GBD.

    EUS-GBD

    Not reported

    Table 3 Study outcomes

    First author, date

    Efficacy outcomes

    Safety outcomes

    Fabbri C, 2022

    Pooled results (27studies, n=1,004 unless stated otherwise)

    • Technical success=98.0% (95% CI 96.3 to 99.3%; I2=23.6%)

    • Per-procedure clinical success=98.8% (95% CI 97.2 to 99.8%; I2=29.4%)

    • Overall clinical success=95.4% (95% CI 92.8 to 97.5%; I2=35.3%)

    • Cholecystitis recurrence=3.4% (95% CI 1.6 to 5.7%; 21 studies, n=965, I2=32.3%)

    • Hospital stay (days)=8.0 (95% CI 6.4 to 9.6; 11 studies, n=687, I2=91.3%)

    Technical success increased when studies reported an experience proxied to more than 10 cases per year (OR 2.84; 95% CI 1.06 to 7.59).

    The overall clinical success improved when the experience proxied more than 10 cases per year (OR 3.85; 95% CI 1.46 to 10.15). In addition, the use of anti-migrating devices (LAMs or ATSEMs) improved this outcome (OR 2.33; 95% CI 1.16 to 4.68).

    Multivariable meta-regression identified both as independent modifiers of the overall clinical success. The OR for experience was 3.52 (95% CI 1.33 to 9.33) and that of anti-migrating devices was 2.16 (95% CI 1.07 to 4.36). After meta-regression, the residual I2 was 0% (Egger's test: p=0.147).

    Pooled results (27 studies, n=1,004)

    Procedure-related adverse events=14.8% (95% CI 8.8 to 21.8%; I2=82.4%)

    • Stent malfunction or dislodgment=3.5% (95% CI 0.1 to 6.7%; I2=64.5%)

    • Procedure-related mortality=0.1% (95% CI 0.0 to 0.1%; I2=0%)

    • 30-day all-cause mortality=2.9% (95% CI 0.5 to 6.6%; I2=74.1%)

    • Bile leak or peritonitis=1.5% (95% CI 0.1 to 3.9%; I2=59.5%)

    • Bleeding=0.3% (95% CI 0.1 to 1.2%; I2=10.3%)

    • Pneumoperitoneum=0.5% (95% CI 0.0 to 0.6%, I2=0%)

    Adverse events reduced when the experience proxied to more than 10 cases per year (OR 0.35; 95% CI 0.19 to 0.65). The use of anti-migrating devices also reduced adverse events (OR 0.30; 95% CI 0.17 to 0.52).

    Multivariable meta-regression identified the use of anti-migrating devices (OR 0.36; 95% CI 0.14 to 0.98) as the only modifier of procedure-related adverse events. After meta-regression, the residual I2 was still substantial (67.8%, Egger's test: p=0.145).

    Mohan B, 2020

    Pooled rates of technical success

    • EUS-GBD=95.3% (95% CI 92.8 to 96.9, I2=0%; 14 studies, 557 patients)

    • ETP-GBD=83% (95% CI 80.1 to 85.5, I2=29%; 21 studies, 851 patients)

    • PT-GBD=98.7% (95% CI 98.0 to 99.1, I2=0%; 33 studies, 2,203 patients)

    p=0.001 for EUS-GBD versus ETP-GBD and PT-GBD

    Pooled rates of clinical success

    • EUS-GBD=96.7% (95% CI 94.0 to 98.2, I2=0%; 14 studies, 557 patients)

    • ETP-GBD=88.1% (95% CI 83.6 to 91.4, I2=50%; 22 studies, 1,223 patients)

    • PT-GBD=89.3% (95% CI 86.6 to 91.5, I2=84%; 38 studies, 11,800 patients)

    p=0.001 for EUS-GBD versus ETP-GBD and PT-GBD

    Pooled rates of recurrence

    • EUS-GBD=4.2% (95% CI 2.4 to 7.4, I2=0%; 14 studies, 557 patients)

    • ETP-GBD=4.6% (95% CI 2.8 to 7.4, I2=53%; 22 studies, 1,223 patients)

    • PT-GBD=10.8% (95% CI 8.3 to 13.9, I2=76%; 37 studies, 3,677 patients)

    p=0.99 for EUS-GBD versus ETP-GBD and p=0.001 for EUS-GBD versus PT-GBD

    Pooled rates of adverse events

    • EUS-GBD=12.4% (95% CI 6.9 to 21.1, I2=6%; 13 studies, 546 patients)

    • ETP-GBD=9.6% (95% CI 5.9 to 15.3, I2=27%; 21 studies, 1,209 patients)

    • PT-GBD=15.1% (95% CI 11.1 to 20.3, I2=95%; 39 studies, 11,997 patients)

    p=0.32 for EUS-GBD versus ETP-GBD and p=0.56 for EUS-GBD versus PT-GBD

    Pooled rates of bleeding

    • EUS-GBD=4.3% (95% CI 2.7 to 6.8, I2=0%; 13 studies, 546 patients), p=0.02

    • ETP-GBD=1.9% (95% CI 1.1 to 3.1, I2=0%; 21 studies, 1,209 patients)

    • PT-GBD=2% (95% CI 1.5 to 2.7, I2=0%; 37 studies, 3,597 patients)

    Pooled rates of perforation

    • EUS-GBD=3.7% (95% CI 2.3 to 6, I2=0%); 13 studies, 546 patients), p=0.04

    • ETP-GBD=2% (95% CI 1.2 to 3.2, I2=0%; 21 studies, 1,209 patients)

    • PT-GBD=2% (95% CI 1.4 to 2.9, I2=0%; 36 studies, 3,524 patients)

    Pooled rates of bile leak or bile peritonitis

    • EUS-GBD=2.9% (95% CI 1.6 to 5.1, I2=0%; 13 studies, 546 patients)

    • ETP-GBD=1.4% (95% CI 0.8 to 2.5, I2=0%); 21 studies, 1,209 patients)

    • PT-GBD=2.7% (95% CI 2.1 to 3.5, I2=0%; 37 studies, 3,597 patients)

    Pooled rates of pancreatitis

    • EUS-GBD=1.4% (95% CI 0.7 to 3.1, I2=0%; 13 studies, 546 patients)

    • ETP-GBD=5.1% (95% CI 3.5 to 7.3, I2=17%; 21 studies, 1,209 patients), p=0.003

    • PT-GBD=1.1% (95% CI 0.7 to 1.7, I2=0%; 36 studies, 3,524 patients)

    Pooled rates of stent occlusion

    • EUS-GBD= 2.6% (95% CI 1.2 to 5.6, I2=0%; 12 studies, 506 patients)

    • ETP-GBD=1.8% (95% CI 0.9 to 3.6, I2=0%; 20 studies, 1,171 patients)

    • PT-GBD=1.8% (95% CI 1.1 to 2.8, I2=56%; 36 studies, 3,524 patients)

    Pooled rates of stent migration

    • EUS-GBD=2.7% (95% CI 1.3 to 5.4, I2=0%; 13 studies, 546 patients)

    • ETP-GBD=2.2% (95% CI 1.2 to 3.9, I2=0%; 20 studies, 1,171 patients)

    • PT-GBD=7.4% (95% CI 5.5 to 10, I2=79; 38 studies, 3,977 patients), p=0.01

    Pooled rates of mortality

    • EUS-GBD=26% (95% CI 16.7 to 38.1, I2=86%); 9 studies, 398 patients), p=0.001

    • ETP-GBD= 16.6% (95% CI 10.5 to 25.2, I2=77%; 13 studies, 884 patients)

    • PT-GBD=11.2% (95% CI 8.7 to 14.1, I2=83%; 37 studies, 3,597 patients)

    Podboy A, 2021

    Technical success – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=1.04 (95% CI 1.01 to 1.09, 7 studies, 782 patients)

    • ETP-GBD vs EUS-GBD, RR= 0.81 (95% CI 0.43 to 0.99, 3 studies, 589 patients)

    • ETP-GBD vs PT-GBD, RR=0.78 (95% CI 0.42 to 0.96; 1 study, 270 patients)

    There was no statistically significant heterogeneity between the pooled studies.

    On the network ranking estimate PT-GBD was noted to be ranked most likely to result in technical success followed by EUS-GBD and ETP-GBD (EUS-GBD vs PT-GBD vs ETP-GBD: 2.00 vs 1.02 vs 2.98).

    Clinical success – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=0.99 (95% CI 0.87 to 1.05, I2=40.3%; 7 studies, 778 patients)

    • ETP-GBD vs EUS-GBD, RR=0.75 (95% CI 0.38 to 0.97, I2=; 4 studies, 402 patients)

    • ETP-GBD vs PT-GBD, RR=0.76 (95% CI 0.40 to 0.98; 1 study, 71 patients)

    There was no statistically significant heterogeneity between the pooled studies.

    The network ranking estimates closely preferred EUS-GBD over PT-GBD and ETP-GBD (EUS-GBD vs PT-GBD vs ETPGBD:1.48 vs 1.55 vs 2.98).

    Recurrent cholecystitis – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=1.96 (95% CI 0.75 to 4.09; 6 studies, 708 patients)

    • ETP-GBD vs EUS-GBD, RR=3.72 (95% CI 1.39 to 7.54; 3 studies, 443 patients)

    • ETP-GBD vs PT-GBD, RR=2.20 (95% CI 0.66 to 5.37; 1 study, 270 patients)

    There was no statistically significant heterogeneity between the pooled studies.

    Ranking estimates preferred EUS-GBD over other modalities (EUS-GBD vs PT-GBD vs ETP-GBD: 1.089 vs 2.02 vs 2.891).

    Need for reintervention – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=3.07 (95% CI 1.52 to 2.94; 4 studies, 394 patients)

    • ETP-GBD vs EUS-GBD, RR=0.77 (95% CI 0.23 to 0.68; 2 studies, 216 patients)

    • ETP-GBD vs PT-GBD, RR=0.26 (95% CI 0.08 to 0.23; 1 study, 270 patients)

    There was no statistically significant heterogeneity between the pooled studies.

    Ranking estimates preferred ETP-GBD (EUS-GBD vs PT-GBD vs ETP-GBD: 1.81 vs 2.99 vs 1.199).

    Adverse events – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=1.02 (95% CI 0.42 to 1.91; 6 studies, 702 patients)

    • ETP-GBD vs EUS-GBD, RR=1.16 (95% CI 0.42 to 2.38; 3 studies, 498 patients)

    • ETP-GBD vs PT-GBD, RR=1.26 (95% CI 0.41 to 3.06; 2 studies, 341 patients)

    There was statistically significant heterogeneity between the pooled studies comparing EUS-GBD and PT-GBD (p=0.013, I2=65.3%) but not for the remaining pooled categories.

    The network ranking estimate was lowest for PT-GBD followed by EUS-GBD (EUS-GBD vs PT-GBD vs ETP-GBD: 1.96 vs 1.88 vs 2.16), although there was wide variability.

    Unplanned admissions – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=4.54 (95% CI 1.52 to 8.92; 4 studies, 658 patients)

    • ETP-GBD vs EUS-GBD, RR=1.32 (95% CI 0.06 to 5.83; 1 study, 226 patients)

    • ETP-GBD vs PT-GBD, RR=0.31 (95% CI 0.02 to 1.31; 1 study, 270 patients)

    There was statistically significant heterogeneity between EUS-GBD and PT-GBD (p=0.029, I2=66.9%).

    The network rating preferred ETP-GBD over other modalities (EUS-GBD vs PT-GBD vs ETP-GBD: 1.582 vs 2.944 vs 1.474).

    Disease-specific mortality – network meta-analysis:

    • PT-GBD vs EUS-GBD, RR=0.83 (95% CI 0.19 to 1.81; 3 studies, 287 patients)

    • ETP-GBD vs EUS-GBD, RR=0.47 (95% CI 0.03 to 1.80; 1 study, 71 patients)

    • ETP-GBD vs PT-GBD, RR=0.81 (0.04 to 3.93; 0 studies, 0 patients)

    There was no statistically significant heterogeneity between the pooled studies.

    Ranking estimate EUS-GBD vs PT-GBD vs ETP-GBD: 2.62 vs 2.09 vs 1.29.

    Teoh A, 2020

    Technical success:

    • EUS-GBD=97.4% (38/39)

    • PT-GBD=100% (40/40), p=0.494

    Clinical success:

    • EUS-GBD=92.3% (36/39)

    • PT-GBD=92.5% (37/40), p=1.00

    Analgesic requirements (total paracetamol in mg); mean (SD)

    • EUS-GBD=3,345 (5,663)

    • PT-GBD=5,165 (5,068), p=0.034

    Recurrent acute cholecystitis at 1 year:

    • EUS-GBD=2.6% (1/39)

    • PT-GBD=20% (8/40), p=0.029

    Reinterventions after 30 days:

    • EUS-GBD=2.6% (1/39)

    • PT-GBD=30% (12/40), p=0.001

    30-day adverse events (p=0.001):

    • EUS-GBD=12.8% (5/39); blocked stent (n=2), perforation (n=1), atrial fibrillation (n=1), pneumonia (n=3)

    • PT-GBD=47.5% (19/40); tube dislodgement (n=15), multiorgan failure (n=3), pericholecystic collection (n=1), acute myocardial infarction (n=1), atrial fibrillation (n=1), pneumonia (n=1), bleeding (n=1), decompensated liver cirrhosis (n=1), urinary tract infection (n=1)

    30-day mortality (p=1):

    • EUS-GBD=7.7% (3/39)

    • PT-GBD=10% (4/40)

    Adverse events at 1 year (including recurrent acute cholecystitis), p<0.001:

    • EUS-GBD=25.6% (10/39); 5 within 30 days listed above plus 1 recurrent acute cholecystitis, 1 blocked stent or tube and 3 common bile duct stones needing ERCP

    • PT-GBD=77.5% (31/40); 19 within 30 days listed above plus 8 recurrent acute cholecystitis, 18 tube dislodgement, 2 blocked stent or tube and 1 common bile duct stones needing ERCP

    Unplanned admissions (p=0.002):

    • EUS-GBD=15.4% (6/39)

    • PT-GBD=50% (20/40)

    Siddiqui A, 2019

    Technical success:

    • EUS-GBD=94% (96/102)

    • PT-GBD=98% (143/146)

    • ETP-GBD=88% (109/124), p=0.003

    Clinical success:

    • EUS-GBD=90% (92/102)

    • PT-GBD=97% (141/146)

    • ETP-GBD=80% (99/124), p<0.001

    After controlling for age, gender, pathology, number of sessions, and technical success, only the number of sessions variable (1 or more than 1 session) was a statistically significant predictor of clinical resolution (OR=0.036, 95% CI 0.004 to 0. 353, p=0.0043).

    Need for additional surgical intervention:

    • EUS-GBD=0% (0/102)

    • PT-GBD=49.7% (73/146)

    • ETP-GBD=11.4% (14/124), p<0.001

    Surgical removal of the gallbladder was eventually

    needed in 24% of patients either for recurrent acute

    cholecystitis not amenable to non-surgical therapy or when the patient had improved medically. The timing of surgery was between 1 and 4 months after the index intervention.

    In patients that did not need a cholecystectomy, those

    that had ETP-GBD had a statistically significantly lower clinical success (78%) for resolution of acute cholecystitis compared to those that had PT-GBD (94%) or EUS-GBD with LAMS (92%); p=0.002.

    Total procedural adverse events (p=0.07):

    • EUS-GBD=11.8% (12/102); 2 perforation, 5 self-limited bleeding, 1 infection, 2 bile leaks, 2 self-limited abdominal pain needing observation

    • PT-GBD=4.1% (6/146); 3 self-limited bleeding, 1 infection, 2 bile leaks

    • ETP-GBD=7.2% (9/124); 2 self-limited bleeding, 3 pancreatitis, 4 self-limited abdominal pain needing observation

    There was no statistically significant difference in the severity of procedural adverse events between the 3 groups (p=0.85).

    Patients with calculous pathology were 2 times more likely to have an adverse event (OR=1.9, 95% CI 1.04 to 3.57, p=0.04), and patients who had more than 1 procedure session were almost 3 times more likely to have an adverse event (OR=2.7, 95% CI 1.23 to 6.07, p=0.0138).

    Long-term (3 to 9 month follow up) adverse events (p<0.001):

    • EUS-GBD=1.9% (2/102); 1 catheter or stent dislodgment, 1 infection

    • PT-GBD=19.8% (29/146); 11 catheter or stent dislodgment, 2 pain, 4 catheter or stent occlusion, 5 cellulitis, 5 infection, 2 abscess

    • ETP-GBD=4.8% (6/124); 2 catheter or stent dislodgment, 4 catheter or stent occlusion

    Unplanned hospital admissions (p<0.001):

    • EUS-GBD=4% (4/102)

    • PT-GBD=19.8% (29/146)

    • ETP-GBD=3.2% (4/124)

    Teoh A, 2021

    Technical success:

    • EUS-GBD=100% (30/30)

    • LC=100% (30/30)

    Clinical success:

    • EUS-GBD=93.3% (28/30)

    • LC=100% (30/30)

    Mean length of hospital stay (SD):

    • EUS-GBD=6.8 (8.1)

    • LC=5.5 (2.7)

    Recurrent cholecystitis:

    • EUS-GBD=3.3% (1/30)

    • LC=0% (0/30)

    Recurrent biliary events:

    • EUS-GBD=10.0% (3/30)

    • LC=10.0% (3/30)

    All were because of the presence of common bile duct stones, which were not suspected on preprocedural imaging. These events all resulted in unplanned admissions, and the stones were removed by ERCP.

    Reinterventions:

    • EUS-GBD=13.3% (4/30)

    • LC=10.0% (3/30)

    30-day adverse events:

    • EUS-GBD=13.3% (4/30); 2 patients died (described below), 1 patient had upper gastrointestinal bleeding needing endoscopic haemostasis, and 1 had a blocked stent because of gallstones needing endoscopic insertion of an additional double-pigtail stent through the LAMS.

    • LC=13.3% (4/30); 1 patient had intraabdominal collections needing percutaneous drainage, 1 had multiorgan failure, 1 had upper gastrointestinal bleeding needing endoscopic haemostasis, and 1 patient had a chest infection.

    30-day mortality:

    • EUS-GBD=6.7% (2/30); 1 aspiration pneumonia and 1 uncontrolled sepsis

    • LC=0% (0/30)

    Unplanned readmissions:

    • EUS-GBD=10.0% (3/30)

    • LC=10.0% (3/30)

    Teoh A, 2019

    Technical success=95.3% of patients

    Clinical success=90.8% of patients

    Recurrent cholecystitis=2.4% (9/379)

    Unplanned procedural events=9.2% of patients (defined as any deviations of the procedure from the planned procedural steps)

    Endoscopists who had done fewer than 25 of the procedures had significantly more procedures that were longer than 30 minutes (p=0.006), more unplanned procedural events (p=0.012) and more 30-day adverse events (p=0.031).

    30-day adverse event rate=15.3% (58/379)

    30-day mortality rate=9.2%

    Procedure-related adverse events, n=20

    • Stent obstruction, n=3

    • Stent migration, n=3

    • Bile leak, n=3

    • Duodenal perforation, n=2

    • Pneumoperitoneum, n=2

    • Gastric outlet obstruction from stent, n=1

    • Bleeding from puncture site, n=1

    • Infected abdominal collection, n=1

    • Bleeding in gallbladder, n=1

    • Bleeding from cystic artery, n=1

    • Infected ascites, n=1

    In multivariate analyses, clinical failure (p=0.014; RR 8.69, 95% CI 1.56 to 48.47) and procedure done by endoscopist with experience of fewer than 25 procedures (p=0.002; RR 4.68, 95%CI 1.79 to 12.26) were statistically significant predictors of 30-day adverse events. The only statistically significant predictor of 30-day mortality was the presence of 30-day adverse event (p<0.001; RR 103, 95%CI 11.24 to 944.04).

    Torres Yuste, 2019

    Gallstone-related hospital admissions during follow up=4.5% (1/22)

    During follow up 12 patients (54.5%) visited the emergency room 34 times with a median of 1 visit per patient (IQR 0 to 3; range 0 to 7).

    There were 36 hospital admissions during follow up, with a median of 1 admission per patient (IQR 0 to 3; range 0 to 9).

    63.6% (14/22) of patients died during follow up: 1 of them of pancreatic cancer progression, while the rest died from non-biliary causes.

    Long-term endoscopic follow up was available in 3 patients: 1 patient had a patent stent at 31 months, 1 patient had an obstructed stent because of overgrowth at 42 months after stent deployment and 1 patient had buried-stent syndrome but with a patent fistula at 51 months follow up.

    There were no LAMS-related adverse events after the first year of follow up.

    Kim J, 2019

    Both procedures were successful initially.

    Gastric reflux into the gallbladder

    Patient 1: frail 86-year-old with end-stage renal disease on haemodialysis, diabetes mellitus, hypertension, congestive heart failure, obstructive apnoea, and coronary artery disease who had had coronary artery stenting. She was admitted weeks after EUS-GBD with intractable right upper quadrant pain. A CT scan confirmed gastric reflux into the gallbladder. This was eventually managed with repeat endoscopic drainage on an outpatient basis, though the patient did have another hospitalisation for severe antibiotic-associated colitis.

    Patient 2: an 80-year-old woman with diabetes mellitus, hypertension, coronary artery disease, and hypothyroidism who developed a stent occlusion with recurrent cholecystitis needing a repeat endoscopic procedure, an infected biloma needing an intraperitoneal drain, and, ultimately, extraluminal free air. The patient eventually had an open cholecystectomy, partial liver resection, gastric antrectomy, and Billroth II gastrojejunostomy.

    Procedure technique

    Most of the procedures were done using LAMS. In the systematic review of 27 studies by Fabbri et al. (2022), 64% of procedures used a duodenal approach and LAMS were the most common type of stent used (n=654). Details of procedure technique were not reported in the systematic review by Podboy et al. (2021) but the authors noted there were wide variations in anatomic puncture location, stent size and stent type within the included studies.

    Efficacy

    Technical success

    Technical success of EUS-GBD was reported in 6 studies and ranged from 95% to 100%. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled rate of technical success was 98% (95% CI 96% to 99%, I2=24%; Fabbri 2022). In the systematic review of 72 studies comparing EUS-GBD, ETP-GBD and PT-GBD, the pooled rates of technical success were 95% for EUS-GBD (95% CI 93% to 97%, I2=0%; 14 studies, n=557), 83% for ETP-GBD (95% CI 80% to 86%, I2=29%; 21 studies, n=851) and 99% for PT-GBD (95% CI 98% to 99%, I2=0%; 33 studies, n=2,203; Mohan 2020). In the network meta-analysis of 10 comparative studies, PT-GBD was noted to be ranked most likely to result in technical success followed by EUS-GBD and ETP-GBD (Podboy 2021).

    In the randomised controlled trial of 79 patients, technical success was 97% (38/39) for EUS-GBD and 100% (40/40) for PT-GBD (p=0.494, Teoh 2020). In the non-randomised comparative study of 372 patients, technical success was 94% (96/102) for EUS-GBD, 98% (143/146) for PT-GBD and 88% (109/124) for ETP-GBD (p=0.003; Siddiqui 2019). In the non-randomised comparative study of 60 patients, technical success was 100% (30/30) for both EUS-GBD and LC (Teoh 2021). In the registry study of 379 patients, technical success was 95% (Teoh 2019).

    Clinical success

    Clinical success of EUS-GBD was reported in 6 studies and ranged from 90% to 97%. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled rate of overall clinical success was 95% (95% CI 93% to 98%, I2=35%; Fabbri 2022). In the systematic review of 72 studies, the pooled rates of clinical success were 97% for EUS-GBD (95% CI 94% to 98%, I2=0%; 14 studies, n=557), 88% for ETP-GBD (95% CI 84% to 91%, I2=50%; 22 studies, n=1,223) and 89% for PT-GBD (95% CI 87% to 92%, I2=84%; 38 studies, n=11,800; Mohan 2020). In the network meta-analysis of 10 comparative studies, the network ranking estimates closely preferred EUS-GBD for clinical success over PT-GBD and ETP-GBD (Podboy 2021).

    In the randomised controlled trial of 79 patients, clinical success was 92% (36/39) for EUS-GBD and 93% (37/40) for PT-GBD at 1 year follow up (p=1.00, Teoh 2020). In the non-randomised comparative study of 372 patients, clinical success was 90% (92/102) for EUS-GBD, 97% (141/146) for PT-GBD and 80% (99/124) for ETP-GBD (p<0.001) with a median follow up of 3 months (Siddiqui 2019). In the non-randomised comparative study of 60 patients, clinical success was 93% (28/30) for EUS-GBD and 100% (30/30) for LC, with mean follow up periods of 572 days and 264 days respectively (p=0.002; Teoh 2021). In the registry study of 379 patients with a mean follow up of 434 days, clinical success was 91% (Teoh 2019).

    In the case series of 22 patients with a median follow up of 24 months, long-term endoscopic follow up was available in 3 patients (Torres Yuste 2019). Of these 3 patients, 1 had a patent stent at 31 months, 1 had an obstructed stent because of overgrowth at 42 months after stent deployment and 1 had buried-stent syndrome but with a patent fistula at 51 months follow up.

    Recurrence of cholecystitis

    Recurrence rates after EUS-GBD were reported in 5 studies and ranged from 2% to 4%. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled recurrence rate was 3% (95% CI 2% to 6%, I2=32%; 21 studies, n=965; Fabbri 2022). In the systematic review of 72 studies, the pooled recurrence rates were 4% for EUS-GBD (95% CI 2% to 7%, I2=0%; 14 studies, n=557), 5% for ETP-GBD (95% CI 3% to 7%, I2=53%; 22 studies, n=1,223) and 11% for PT-GBD (95% CI 8% to 14%, I2=76%; 37 studies, n=3,677; Mohan 2020). In the network meta-analysis of 10 comparative studies, the network ranking estimates preferred EUS-GBD for recurrence rates over PT-GBD and ETP-GBD (Podboy 2021).

    In the randomised controlled trial of 79 patients, rate of recurrence was 3% (1/39) for EUS-GBD and 20% (8/40) for PT-GBD at 1 year follow up (p=0.029, Teoh 2020). In the non-randomised comparative study of 60 patients, rate of recurrence was 3% (1/30) for EUS-GBD and 0% (0/30) for LC, with mean follow up periods of 572 days and 264 days respectively (Teoh 2021). In the registry study of 379 patients with a mean follow up of 434 days, rate of recurrence was 2% (9/379; Teoh 2019).

    Recurrence of biliary events

    In the randomised controlled trial of 79 patients, common bile duct stones needing ERCP was reported in 8% (3/39) of patients who had EUS-GBD and 3% (1/40) of patients who had PT-GBD at 1 year follow up (Teoh 2020). In the non-randomised comparative study of 60 patients, recurrent biliary events were reported for 10% (3/30) of patients both in the EUS-GBD and the LC group, with mean follow up periods of 572 days and 264 days respectively (Teoh 2021). In the case series of 22 patients, there was 1 gallstone-related hospital admission during follow up (median 24 months, Torres Yuste 2019).

    Reintervention

    The rate of reinterventions was reported in 3 studies. In the non-randomised comparative study of 372 patients, a need for additional surgical intervention was reported for 0% (0/102) of patients who had EUS-GBD, 50% (73/146) of patients who had PT-GBD and 11% (14/124) of patients who had ETP-GBD (p<0.001) with a median follow up of 3 months (Siddiqui 2019).

    In the randomised controlled trial of 79 patients, reinterventions after 30 days were reported for 3% (1/39) of patients who had EUS-GBD and 30% (12/40) of patients who had PT-GBD (p=0.001, Teoh 2020). In the non-randomised comparative study of 60 patients, the reintervention rate was 13% (4/30) for EUS-GBD and 10% (3/30) for LC, with mean follow up periods of 572 days and 264 days respectively (Teoh 2021).

    Hospital admissions

    The rate of unplanned hospital admissions after the procedure was reported in 3 studies. In the randomised controlled trial of 79 patients, the rate of unplanned hospital admissions was 15% (6/39) of patients who had EUS-GBD and 50% (20/40) of patients who had PT-GBD (p=0.002, Teoh 2020). In the non-randomised comparative study of 372 patients, the rate of unplanned hospital admissions was 4% (4/102) of patients who had EUS-GBD, 20% (29/146) of patients who had PT-GBD and 3% (4/124) of patients who had ETP-GBD (p<0.001) with a median follow up of 3 months (Siddiqui 2019). In the non-randomised comparative study of 60 patients, the rate of unplanned readmissions was 10% (3/30) for EUS-GBD and 10% (3/30) for LC, with mean follow up periods of 572 days and 264 days respectively (Teoh 2021).

    Safety

    Mortality

    Mortality was reported in 5 studies. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled procedure-related mortality was <1% (95% CI 0% to 0.1%, I2=0%) and 30-day all-cause mortality was 3% (95% CI 0.5% to 7%, I2=74%; Fabbri 2022). In the systematic review of 72 studies, the pooled mortality rates were 26% for EUS-GBD (95% CI 17% to 38%, I2=86%; 9 studies, n=398), 17% for ETP-GBD (95% CI 11% to 25%, I2=77%; 13 studies, n=884) and 11% for PT-GBD (95% CI 9% to 14%, I2=83%; 37 studies, n=3,597; Mohan 2020).

    In the randomised controlled trial of 79 patients, 30-day mortality was 8% (3/39) for EUS-GBD and 10% (4/40) of patients who had PT-GBD (p=1, Teoh 2020). In the non-randomised comparative study of 60 patients, the 30-day mortality was 7% (2/30) for EUS-GBD and 0% (0/30) for LC (Teoh 2021). In the registry study of 379 patients, 30-day mortality was 9% (Teoh 2019).

    Stent malfunction, dislodgment or migration

    Stent malfunction, dislodgment or migration was reported in 4 studies. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled rate of stent malfunction of dislodgment was 4% (95% CI 0.1% to 7%, I2=64.5%; Fabbri 2022). In the systematic review of 72 studies, the pooled rates of stent migration were 3% for EUS-GBD (95% CI 1% to 5%, I2=0%; 13 studies, n=546), 2% for ETP-GBD (95% CI 1% to 4%, I2=0%; 20 studies, n=1,171) and 7% for PT-GBD (95% CI 6% to 10%, I2=79%; 38 studies, n=3,997; Mohan 2020).

    Catheter or stent dislodgment during follow up were reported in 1% (1/102) of patients who had EUS-GBD, 8% (11/146) of patients who had PT-GBD and 2% (2/124) of patients who had ETP-GBD in the non-randomised comparative study of 372 patients (Siddiqui 2019). Stent migration was reported in 1% (3/379) of patients and gastric outlet obstruction from the stent was reported in 1 patient in the registry study of 379 patients (Teoh 2019).

    Stent occlusion

    Stent occlusion was reported in 5 studies. In the systematic review of 72 studies, the pooled rates of stent occlusion were 3% for EUS-GBD (95% CI 1% to 6%, I2=0%; 12 studies, n=506), 2% for ETP-GBD (95% CI 1% to 4%, I2=0%; 20 studies, n=1,171) and 2% for PT-GBD (95% CI 1% to 3%, I2=56%; 36 studies, n=3,524; Mohan 2020).

    A blocked stent within 30 days was reported in 5% (2/39) of patients who had EUS-GBD in the randomised controlled trial of 79 patients. At 1 year, a blocked stent was reported in 1 additional patient who had EUS-GBD and 5% (2/40) of patients who had PT-GBD (Teoh 2020). Catheter or stent occlusion during follow up were reported in 0% (0/102) of patients who had EUS-GBD, 3% (4/146) of patients who had PT-GBD and 3% (4/124) of patients who had ETP-GBD in the non-randomised comparative study of 372 patients (Siddiqui 2019). A blocked stent because of gallstones was reported in 1 patient who had EUS-GBD in the non-randomised comparative study of 60 patients (Teoh 2021). Stent obstruction was reported in 1% (3/379) of patients in the registry study of 379 patients (Teoh 2019). An obstructed stent because of overgrowth at 42 months after stent deployment was reported in 1 patient in the case series of 22 patients (Torres Yuste 2019).

    Bile leak or peritonitis

    Bile leak or peritonitis was reported in 4 studies. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled rate of bile leak or peritonitis was 2% (95% CI 0.1% to 4%, I2=59.5%; Fabbri 2022). In the systematic review of 72 studies, the pooled rates of bile leak or bile peritonitis were 3% for EUS-GBD (95% CI 2% to 5%, I2=0%; 13 studies, n=546), 1% for ETP-GBD (95% CI 1% to 3%, I2=0%; 21 studies, n=1,209) and 3% for PT-GBD (95% CI 2% to 4%, I2=0%; 37 studies, n=3,597; Mohan 2020).

    Procedural bile leaks were reported in 2% (2/102) of patients who had EUS-GBD and 1% (2/146) of patients who had PT-GBD in the non-randomised comparative study of 372 patients (Siddiqui 2019). Bile leak was reported in 1% (3/379) of patients in the registry study of 379 patients (Teoh 2019).

    Bleeding

    Bleeding was reported in 5 studies. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled rate of bleeding was <1% (95% CI 0.1% to 1%, I2=10.3%; Fabbri 2022). In the systematic review of 72 studies, the pooled rates of bleeding were 4% for EUS-GBD (95% CI 3% to 7%, I2=0%; 13 studies, n=546), 2% for ETP-GBD (95% CI 1% to 3%, I2=0%; 21 studies, n=1,209) and 2% for PT-GBD (95% CI 2% to 3%, I2=0%; 37 studies, n=3,597; Mohan 2020).

    Procedural self-limited bleeding was reported in 5% (5/102) of patients who had EUS-GBD, 2% (3/146) of patients who had PT-GBD and 2% (2/124) of patients in the non-randomised comparative study of 372 patients (Siddiqui 2019). Upper gastrointestinal bleeding needing endoscopic haemostasis was reported in 1 patient who had EUS-GBD and 1 patient who had LC in the non-randomised comparative study of 60 patients (Teoh 2021).

    Bleeding from the puncture site, bleeding in the gallbladder and bleeding from the cystic artery were each reported in 1 patient in the registry study of 379 patients (Teoh 2019).

    Perforation

    Perforation was reported as an adverse event in 4 studies. In the systematic review of 72 studies, the pooled rates of perforation were 4% for EUS-GBD (95% CI 2% to 6%, I2=0%; 13 studies, n=546), 2% for ETP-GBD (95% CI 1% to 3%, I2=0%; 21 studies, n=1,209) and 2% for PT-GBD (95% CI 1% to 3%, I2=0%; 36 studies, n=3,524; Mohan 2020).

    Perforation was reported in 1 patient who had EUS-GBD in the randomised controlled trial of 79 patients (Teoh 2020). Perforation was reported in 2% (2/102) of patients who had EUS-GBD in the non-randomised comparative study of 372 patients (Siddiqui 2019). Duodenal perforation was reported in <1% (2/379) of patients in the registry study of 379 patients (Teoh 2019).

    Pneumoperitoneum

    Pneumoperitoneum was reported in 2 studies. In the systematic review of 27 studies on EUS-GBD (n=1,004), the pooled rate of pneumoperitoneum was <1% (95% CI 0% to 1%, I2=0%; Fabbri 2022). Pneumoperitoneum was reported in <1% (2/379) of patients in the registry study of 379 patients (Teoh 2019).

    Pancreatitis

    In the systematic review of 72 studies, the pooled rates of pancreatitis were 1% for EUS-GBD (95% CI 1% to 3%, I2=0%; 13 studies, n=546), 5% for ETP-GBD (95% CI 4% to 7%, I2=17%; 21 studies, n=1,209) and 1% for PT-GBD (95% CI 1% to 2%, I2=0%; 36 studies, n=3,524; Mohan 2020).

    Infection

    Periprocedural infection was reported in 1 patient who had EUS-GBD and 1 who had PT-GBD in the non-randomised comparative study of 372 patients (Siddiqui 2019). In the same study, infection during the follow up period (3 to 9 months) was reported in 1% (1/102) of patients who had EUS-GBD and 3% (5/146) of patients who had PT-GBD. Infected abdominal collection and infected ascites were each reported in 1 patient in the registry study of 379 patients (Teoh 2019).

    Other

    Gastric reflux into the gallbladder after EUS-GBD was described in 2 patients as case reports (Kim 2019). In 1 patient, this was managed with repeat endoscopic drainage. The other patient eventually had an open cholecystectomy, partial liver resection, gastric antrectomy and Billroth II gastrojejunostomy.

    Anecdotal and theoretical adverse events

    Expert advice was sought from consultants who have been nominated or ratified by their professional Society or Royal College. They were asked if they knew of any other adverse events for this procedure that they had heard about (anecdotal), which were not reported in the literature. They were also asked if they thought there were other adverse events that might possibly occur, even if they have never happened (theoretical).

    They did not list any anecdotal adverse events.

    They listed the following theoretical adverse events:

    • One theoretical adverse event is around eventual progression to surgery. This technique has been used to "bridge" an unfit patient to surgery but the effect of these stents on eventual surgery is not clear.

    • The procedure involves creating a connection between the gallbladder and stomach/duodenum (a cholecystoduodenal fistula). This would likely lead to more adhesions around the gallbladder and the bowel and makes any future surgical cholecystectomy or other abdominal surgery technically more difficult. The potential risk of complications from future abdominal operation could be higher as compared to percutaneous cholecystostomy.

    Three professional expert questionnaires for this procedure were submitted. Find full details of what the professional experts said about the procedure in the specialist advice questionnaires for this procedure.

    No professional expert questionnaires were submitted.

    Validity and generalisability

    • Two randomised controlled trials on EUS-GBD for acute cholecystitis were identified, the most recent of which is included in the key evidence tables and the other (Jang 2012) is included in table 5. Most of the evidence is from retrospective observational studies.

    • There was evidence from Europe, North America, Asia and Australasia, but no studies based in the UK were identified for inclusion in the key evidence.

    • Most of the procedures were done by experienced endoscopists in tertiary referral centres. One study reported that unplanned procedural events were statistically significantly more common when the procedure was done by endoscopists with experience of fewer than 25 procedures (p=0.033; Teoh 2019).

    • In the systematic review by Fabbri et al. (2022), 8 of the 27 studies reported 10 or more cases per year per centre.

    • None of the studies reported quality of life outcomes.

    • The longest mean or median follow up was 24 months, but this was in a small retrospective case series with only 22 patients (Torres Yuste 2019).

    • There were variations in technique and stents used within and between studies. In 1 systematic review and meta-analysis, the use of anti-migrating devices was shown to reduce the rate of procedure-related adverse events and increase the overall clinical success (Fabbri 2022).

    • In 1 systematic review and meta-analysis, the pooled mortality rate was higher for EUS-GBD than for ETP-GBD and PT-GBD. This result should be interpreted with caution because there was considerable heterogeneity and not all studies reported mortality data. The authors suggested that a possible explanation for the higher all-cause mortality was that most of the EUS-GBD studies were done in patients for whom the overall survival was low to begin with and the studies adequately followed up their patients to report a mortality event, unlike the ETP-GBD and PT-GBD studies (Mohan 2020).

    • In the non-randomised study comparing EUS-GBD with LC, the follow up period was statistically significantly longer in the EUS-GBD group (Teoh 2021).

    • None of the papers included in tables 2 and 3 reported that the study was funded by a company. Declarations of interest were reported by at least 1 author in all the papers, except the letter to the editor that described adverse safety events (Kim 2019).

    • Ongoing trial

      • Efficacy and Safety of Lumen Apposing Metal Stents: a Retrospective Multicentre Study (NCT03903523); Italy; observational cohort study; n=500; estimated end date December 2023