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    The content on this page is not current guidance and is only for the purposes of the consultation process.

    Evidence summary

    Population and studies description

    This overview is based on 2,644 patients from 1 systematic review and meta-analysis, 1 RCT, 1 non-randomised comparative study, 5 case series and 1 case report. Of these 2,644 patients, 2,533 patients had the procedure. 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 78 other relevant studies in table 5. Due to the large body of evidence for this procedure, studies that met the criteria but had less than 50 patients were not included in table 5.

    The studies were done in Italy, the UK, Ireland, the USA, France, Ecuador, Spain, India, Brazil, South Korea, Germany and Japan. The systematic review was done in the USA but locations of included studies were not reported. Of the case series reported, 4 were retrospective and 1 was prospective, and the single non-randomised comparative study included was prospective.

    The study from Fugazza (2022) was a multicentre analysis of 256 people with jaundice caused by DMBO who had EUS-CDS after ERCP failure. The case series from On (2022) was a multicentre review of EUS-CDS in 120 consecutive patients with DMBO. An additional case series of 205 patients who had EUS‑HGS looked at the procedure specifically in the context of malignant biliary obstruction (Marx 2022). A case report highlighted an adverse event which occurred after EUS-HGS in a person with pancreatic adenocarcinoma (Takahashi 2022).

    There was also 1 RCT of 125 people which compared EUS-BD and ERCP safety and efficacy in people with unresectable malignant distal biliary obstruction (Paik 2018).

    The systematic review from Dhindsa (2022) included 1,437 patients who had EUS-BD across 23 studies and did not exclude studies on the basis of benign or malignant biliary obstruction.

    The prospective non-randomised comparative study of 182 people (Tyberg 2022) comparing EUS-CDS and EUS-HGS, which recruited people from the USA, Brazil, India, Ecuador, France and Spain, did not specifically exclude cases of benign biliary obstruction. But, in the final analysis 91% (165/182) of included people had malignant biliary obstruction as an indication.

    Two studies looked at people with biliary obstruction because of benign causes. The case series of 215 people from Hathorn (2022) was a retrospective single-centre study in the USA which compared EUS-guided transhepatic biliary drainage outcomes in populations with benign and malignant causes of biliary obstruction. The case series of 103 people from Fuldner (2021) was a single-centre study which also examined the efficacy and safety of EUS-BD after failed ERCP in people with benign biliary obstruction.

    Follow-up of all key studies ranged from a median of 70 days to a mean of 257 days; however, study follow-up was not reported in the systematic review and meta-analysis or the case report. The mean age of patients ranged from 62.9 years to 73.4 years.

    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 (years)

    Study design

    Inclusion criteria

    Intervention

    Follow up

    1

    Dhindsa 2020

    (SR+MA done in USA, countries not reported for included studies)

    1,437 (774:663)

    Mean 67

    Systematic review and meta-analysis

    Cohort studies looking at EUS-BD with information on adverse events, subcategories of adverse events, technical success rate, and clinical

    success rate.

    EUS-BD (transgastric 29.5%, transduodenal 34.4%, other techniques 19.8%)

    Not reported

    2

    Fugazza 2022

    Italy

    256 (141:115)

    Mean 73.4

    Case series

    Patients with jaundice caused by DMBO after ERCP failure.

    EUS-CDS

    Mean 151±162 days

    3

    On 2022

    UK and Ireland

    120 (66:54)

    Mean 73

    Case series

    Patients with DMBO

    EUS-CDS

    Median 70 days (range 3 to 869)

    4

    Hathorn 2022

    USA

    215 (101:114)

    (130 malignant, 85 benign)

    Mean 62.9

    Case series

    Patients who had EUS-guided transhepatic biliary drainage

    EUS-HGS

    Mean 257.31 ± 308.11 days

    (124.53± 229.86 days for benign disease, 457.27 ± 466.31 days for malignant disease)

    5

    Marx 2022

    France

    205 (104:101)

    Mean 68

    Case series

    Patients that had EUS-HGS between 2002 and 2018 for malignant biliary obstruction

    EUS-HGS

    Median 6.4 months (95% CI 4.1 to 10 months)

    6

    Tyberg 2022

    USA, France, Ecuador, Spain, India, Brazil

    182 (93:89) (95 EUS-HGS versus 87 EUS-CDS)

    Mean 70

    Non-randomised comparative study

    Patients with biliary obstruction with prior failed ERCP who had EUS-HGS or EUS-CCD; at least 6-month follow-up data or death within 6 months.

    EUS-HGS and EUS-CDS

    6 months or until death

    7

    Paik 2018

    South Korea

    125 (67:58)

    (64 EUS-BD versus 61 ERCP)

    Mean 66.8

    RCT

    Presence of unresectable malignant distal biliary obstruction (>2 cm distal to the hilum) with patholologic or radiologic diagnosis prior to endoscopic intervention; >18 years old; a Karnofsky index of 30% or greater, no serious or uncontrolled coexisting medical illness.

    EUS-BD (EUS-CDS 50%, EUS-HGS 50%) and ERCP

    Median 155 days (IQR 100 to 234 days)

    8

    Fuldner 2021

    Germany

    103

    66

    Case series

    Patients with benign biliary obstruction and previously failed ERCP.

    EUS-BD (transgastric 70%, transduodenal 16%, transjejuneal 14%)

    Not reported

    9

    Takahashi 2022

    Japan

    1 (female)

    77

    Case report

    -

    EUS-HGS

    Not reported

    Table 3 Study outcomes

    First author, date

    Efficacy outcomes

    Safety outcomes

    Dhindsa, 2020

    • Pooled rate of technical success: 91.5% (95% CI 87.7% to 94.2%; n = 23 studies; I2= 76.5%)

    • Pooled rate of clinical success: 87% (95% CI 82.3% to 90.6%; n = 21 studies; I2 = 72.4%)

    • Pooled rate of reintervention: 6.5% (95% CI 3.8% to 10.8%; n = 19 studies; I2 = 69.3%

    • Pooled rate of all adverse events: 17.9% (95% CI 14.3% to 22.2%; n=23 studies; I2 = 69.1%)

    • Pooled rate of bile leak: 4.1% (95% CI 2.7% to 6.2%; n=23 studies; I2 = 46.7%)

    • Pooled rate of bleeding: 3.5% (95% CI 2.3% to 5.3%, n=23 studies; I2 = 41.8%)

    • Pooled rate of pneumoperitoneum: 3.3% (95% CI 2.3% to 4.6%; n=23 studies; I2 = 2%)

    • Pooled rate of stent migration: 3.9% (95% CI 2.5% to 6.2%; n=22 studies; I2 = 43.5%)

    • Pooled rate of infection: 3.8% (95% CI 2.8% to 5.1%; n=22 studies; I2 = 0%)

    Fugazza, 2022

    Technical success of EUS-CDS: 93.3% (239/256)

    • Rate of reintervention (endoscopic or radiologic): 9.2% (22/239)

    Clinical success

    • Clinical success (decrease in bilirubin level of at least 50% within 2 weeks of procedure): 96.2% (230/239)

    • Mean percentage decrease in bilirubin levels: 72% (4.11 mg/dL ± 3.96 mg/dL at 2 weeks from baseline 14.7 mg/dL ± 7.11 mg/dL (p<0.001).

    • Death (from underlying disease): 56.1% (134/239)

    Adverse events

    • Overall: 27 events, 10.5% of patients (25/239)

    • Bleeding: 1.7% (4/239) – 3 moderate, 1 severe

    • Stent migration: 0.8% (2/239) – moderate

    • Infection: 2.1% (5/239) – 3 mild, 1 moderate, 1 severe

    • Stent occlusion: 6.7% (16/239) -15 moderate, 1 severe

    On, 2022

    Technical success of EUS-CDS: 90.8% (109/120)

    • Reintervention: 8.3% (9/109)

    Clinical success

    • Clinical success (decrease in bilirubin level of at least 50% within 2 weeks of procedure): 94.8% (92/97).

    • Clinical success could not be determined in 12 patients because of a lack of laboratory values within the stipulated time frame (n=5), loss to follow-up (n=3), and death because of underlying disseminated disease (n=4).

    • Serum bilirubin levels: Median serum bilirubin 2 weeks after procedure 76.5 micromoles/litre (range 8 to 352) compared with 222.5 micromoles/litre (range 45 to 687) at baseline (p<0.001)

    Adverse events

    Overall: 17.5% (21/120)

    Severe adverse events

    • Duodenal perforation: 1.7% (2/120)

    Moderate adverse events

    • Stent migration: 1.7% (2/120)

    • Stent occlusion (no cholangitis): 0.8% (1/120)

    • Combined cholangitis and stent occlusion: 5% (6/120)

    • Bile leak: 1.7% (2/120)

    Mild adverse events

    • Pneumoperitoneum: 1.7% (2/120)

    • Retroperitoneal air: 0.8% (1/120)

    • Cholangitis (no stent occlusion): 6.6% (5/120)

    Hathorn, 2022

    • Technical success (all patients): 94.9% (204/215)

    • Technical success (malignant patients): 96.9% (126/130)

    • Technical success (benign patients): 91.8% (78/85)

    Clinical success (defined by appropriate reduction in serum bilirubin; exact threshold not reported)

    • Clinical success (all patients): 87.3% (89/102)

    • Clinical success (malignant patients): 86% (63/72)

    • Clinical success (benign patients): 90% (27/30)

    Rate of reintervention

    In patients with malignant disease in whom at least a 6-month follow-up was available, 17.4% (22/126) of patients required reintervention.

    Survival

    • 66 deaths in malignant group due to progressive disease) – mean time to death 142 ± 238 days)

    • 8 deaths in benign group (unrelated to procedure)

    Adverse events

    Overall (all patients): 18.6% (40/215)

    Overall (benign patients): 22.3% (19/85)

    Overall (malignant patients): 16.1% (21/130)

    Severe adverse events (>7 days spent in hospital post procedure or because of procedure related readmit, or technical issue

    resulting in procedural termination or surgery):

    • Perforation: 0.5% (1/215; benign)

    • Septic shock: 1.9% (4/215; all malignant)

    • Haemodynamic instability: 0.5% (1/215; malignant)

    Moderate adverse events (required additional monitoring and length of stay <7 days):

    • Moderate abdominal pain (prolonged hospital stay: 0.5% (1/215; benign)

    • Haematemesis (unrelated to stent): 0.5% (1/215; benign)

    • Respiratory distress: 0.5% (1/215; benign)

    • Cholangitis: 2.8% (6/215; benign)

    • Sepsis (secondary to cholangitis): 2.3% (5/215; all malignant)

    • Perihepatic abscess or biloma: 1.3% (3/215; all malignant)

    • Bacteraemia: 0.5% (1/215; malignant)

    • Fungaemia: 0.5% (1/215; malignant)

    • Bile leak: 0.5% (1/215; malignant)

    • Mallory-Weiss tear (triggered by postprocedural nausea and vomiting): 0.5% (1/215; malignant)

    Mild adverse events (no additional hospital monitoring or procedure required)

    • Mild postprocedural pain: 2.3% (5/215; 3 benign, 2 malignant)

    • Bacteraemia: 1.3% (3/215; all benign)

    • Biloma (requiring antibiotics): 0.9% (2/215; all benign)

    • Fever (requiring antibiotics): 0.9% (2/215; 1 benign, 1 malignant)

    • Minor bleeding: 0.5% (1/215; malignant)

    Marx, 2022

    Clinical success

    Proportion of patients with >50% decrease in plasma bilirubin 1 month after procedure or improvement in the patient's health condition allowing discharge of the patient: 93% (143/153)

    (22 patients lost to follow-up, 30 patients died before bilirubin data could be collected)

    Survival

    Median survival 5.3 months (95% CI 2.9 to 7.5 months).

    All mortality within 30 days: 15% (30/205)

    Reintervention

    • Recurrent biliary obstruction (requiring reintervention): 36% (47/130)

    • Mean number of reinterventions per patient for stent revision: 1.7

    Stent patency: mean 153±433 days

    Procedure related deaths: 5% (10/205)

    Causes of death:

    • Biliary leakage n=1

    • Severe sepsis n=1

    • Stent obstruction n=1,

    • Stent migration n=3

    • Acute kidney injury, n=1;

    • Acute pulmonary oedema, n=1

    • Combination of minor complications causing progressive deterioration of the patient's health condition leading to death, n=2)

    Early complications: 46 across 37/205 patients (18%)

    • Stent migration: 9.3% (19/205)

    • Stent occlusion: 3.9% (8/205)

    • Cholangitis: 2.4% (5/205)

    • Bile leakage: 2.9 (6/205)

    • Pneumoperitoneum: 2.4% (5/205)

    • Bleeding: 1.5% (3/205)

    Late stent migration (>30 days after EUS-HGS): 0.7% (1/130)

    Tyberg, 2022

    Technical success

    • Technical success EUS-HGS: 92% (87/95)

    • Technical success EUS-CDS: 92% (80/87)

    Clinical success

    Clinical success (defined as relief of cholangitis, jaundice, and/or pruritus with at least a 30% decrease in pre-treatment bilirubin level within a week after placement or normalisation of bilirubin within 30 days):

    • EUS-HGS: 86% (75/87)

    • EUS-CDS: 100% (80//80)

    Stent patency and reintervention

    • Rate of re-stenting EUS-HGS: 29% (25/87)

    • Rate of re-stenting EUS-CDS: 10% (8/80)

    p value for difference =0.01

    EUS-CDS was 4.5 times more likely than EUS-HGS to achieve longer stent patency (>30 days) or manage obstruction. OR 4.5 (95% CI 1.15 to 17.65, p=0.0302)

    Survival

    • 6 months survival (all patients): 42% (77/182)

    Adverse events

    All adverse events EUS-HGS: 21% (20/95)

    All adverse events EUS-CDS: 30% (26/87)

    (p value for difference=0.17)

    • Biloma: 1% (1/95) EUS-HGS

    • Cholangitis: 2.1% (2/95) EUS-HGS, 1.1% (1/87) EUS-CDS

    • Bleeding: 6.3% (6/95) EUS-HGS, 3.4% (3/87) EUS-CDS

    • Peritonitis: 1% (1/95) EUS-HGS

    • Perforation: 2.1% (2/95) EUS-HGS, 2.3% (2/87) EUS-CDS

    • Migration: 1% (1/95) EUS-HGS, 1.1% (1/87) EUS-CDS

    • Infection: 2.1% (2/95) EUS-HGS, 3.4% (3/87) EUS-CDS

    • Other events (events not specified): 5.3% (5/95) EUS-HGS, 11.5% (10/87) EUS-CDS

    Paik 2018

    Technical success

    • Technical success all EUS-BD: 93.8% (60/64)

    • Technical success EUS-CDS: 90.6% (29/32)

    • Technical success EUS-HGS 96.9% (31/32)

    • Technical success ERCP: 90.2% (55/61)

    • Difference 3.6% (95% CI -4.4% to 11.6%)

    p value for noninferiority between EUS-BD and ERCP = 0.003

    • Length of hospital stay ERCP: median 5 days (IQR 4 to 6 days)

    • Length of hospital stay EUS-BD: median 4 days (IQR 3 to 5 days)

    • p value for significance in hospital length of stay = 0.03

    Clinical success (defined as reduction in bilirubin to 50% less than pre-treatment levels within 1 week or less than 25% of pre-treatment levels within 4 weeks):

    • Clinical success EUS-BD: 90% (54/60)

    • Clinical success ERCP: 94.5% (52/55)

    p value for difference = 0.49

    Rate of reintervention

    • Rate of reintervention (EUS-BD, ITT analysis): 15.6% (10/64)

    • Rate of reintervention (ERCP, ITT analysis): 42.6% (26/61)

    p value for difference = 0.001

    Stent patency

    • Stent patency EUS-BD (6 months, Kaplan-Meier estimate): 85.1% (n=15)

    • Stent patency ERCP (6 months, Kaplan-Meier estimate): 48.9% (n=20)

    p value for significance = 0.001

    • Mean stent patency time: 208 days for EUS-BD, 165 days for ERCP

    Survival

    • Deaths due to disease progression, 6 months (ITT analysis): 71.9% (46/64) for EUS-BD, 83.6% (51/61) for ERCP

    • Median overall survival (Kaplan–Meier estimate): 178 days EUS-BD versus 144 days ERCP, p=0.92

    Quality of life (12 weeks after procedure, assessed using EORTC QLQ-C30 – scored 1 to 100)

    • Change in global functioning score: +4.17 for EUS-BD versus -9.03 ERCP, p=0.001

    • Change in emotional functioning score: +1.62 for EUS-BD versus -9.72 ERCP, p=0.001

    • Change in cognitive functioning score: +0.93 for EUS-BD versus -11.11 ERCP, p=0.003

    • Change in fatigue score: -3.40 for EUS-BD versus +8.02 ERCP, p=0.02

    • Change in pain score: -17.59 for EUS-BD versus +4.63 ERCP, p=0.01

    • Change in financial difficulties score: +2.78 for EUS-BD versus 18.52 ERCP, p=0.01

    Adverse events

    Early adverse events (<2 weeks, procedure-related): 6.3% (4/64) for EUS-BD, 19.7% (12/61) for ERCP (p=0.03)

    Late adverse events 4.7% (3/64) for EUS-BD versus 19.7% (12/61) for ERCP (p=0.03)

    EUS-BD adverse events:

    • Self-limited pneumoperitoneum: 3.1% (2/64)

    • Bile peritonitis: 1.6% (1/64)

    • Acute cholangitis: 6.3% (4/64)

    ERCP adverse events:

    • Acute pancreatitis: 14.8% (9/61)

    • Acute cholecystitis: 8.2% (5/61)

    • Acute cholangitis: 9.8% (6/61)

    • Stent occlusion: 4.9% (3/61)

    • Stent migration: 1.6% (1/61)

    Fuldner 2021

    Technical success

    Technical drainage success rate: 96% (99/103)

    Stent placement rate 86% (85/99)

    2/103 patients included did not need intervention, 17/103 patients had stone extraction without stent placement

    • Rate of reintervention: 18.8% (19/103)

    Overall complication rate: 25.2% (26/103)

    Complications:

    • Post-interventional pain: 2% (2/103)

    • Biloma/liver abscess: 3% (3/103)

    • Bleeding: 6% (6/103)

    • Perforation: 1% (1/103)

    • Bile leakage: 3% (3/103)

    • Stent dislocation: 11% (11/103)

    Takahashi 2022

    None reported.

    Patient displayed fever and cognitive dysfunction 73 days after EUS-HGS. Abdominal CE-CT scan showed large infected biloma and liver abscess in the left hepatic lobe, and CT imaging showed brain abscess. Successfully treated with antibiotics and additional stenting, and patient discharged from hospital 48 days after admission.

    Suspected cause: Bile leakage from puncture tract of the EUS-HGS owing to dysfunction of the HGS stent, and biloma was formed which then became infected (causing liver and brain abscess).

    Procedure technique

    Of the 9 studies included, 8 specified the exact technique used; the systematic review specified the proportion of interventions that were transgastric, transduodenal or used another route, but was not more specific (Dhindsa 2020).

    Of the remaining 8 studies, 6 studies used EUS-HGS, and 5 studies used EUS‑CDS (3 studies used both techniques). The case series of 103 people used multiple stenting techniques in addition to EUS-HGS and EUS-CDS which included EUS-RV and antegrade internal drainage (Fuldner 2021).

    A variety of plastic and metal stents were used in the key evidence papers, but the most commonly named stents were the HotAXIOS stents (Boston Scientific), and the Hanarostent (MiTech), which were each named in 2 of the 9 studies.

    Efficacy

    Technical success

    Technical success was reported in 7 studies. In the systematic review and meta-analysis of 1,437 people, the pooled rate of technical success (defined as successful stent deployment) was 92% (95% CI 87.7% to 94.2%; n=23 studies; I2= 76.5%; Dhindsa 2020).

    In the case series of 256 people with jaundice caused by DMBO, the rate of technical success (defined as successful stent placement) of EUS-CDS was 93% (239/256; Fugazza 2022).

    In the case series of 120 people with DMBO, the technical success rate of EUS‑CDS (defined as successful stent placement) was 91% (109/120; On 2022).

    In the case series of 215 people, overall technical success rate of transhepatic EUS-BD was 95% (204/215). When split by indication, reported technical success in malignant patients was 97% (126/130) and technical success in benign patients was 92% (78/85; Hathorn 2022).

    In the non-randomised comparative study of 182 people, technical success was 92% for both EUS-CDS (80/87) and EUS-HGS (87/95; Tyberg 2022).

    In the RCT of 125 people, technical success of EUS-BD was 94% (60/64) and was comparable to ERCP technical success of 90% (55/61); difference 3.6% (95% CI -4.4% to 11.6%, p=0.003 for noninferiority between procedures; Paik 2018).

    In the case series of 103 people with benign biliary obstruction, the technical drainage success rate was 96% (99/103) and stent placement rate was 86% (85/99; Fuldner 2021).

    Clinical success

    Clinical success, largely defined by reduction in bilirubin serum levels, was reported in 7 studies. In the systematic review and meta-analysis of 1,437 people, the pooled rate of clinical success (defined in most included studies as a reduction in serum total bilirubin by 50% at 1 week or <3 mg/dL at 2 weeks after the procedure) was 87% (95% CI 82.3% to 90.6%; n=21 studies; I2= 72.4%; Dhindsa 2020).

    In the case series of 256 people with jaundice caused by DMBO, the proportion of people with a decrease in bilirubin levels of at least 50% within 2 weeks of EUS-CDS was 96% (230/239). The mean percentage decrease in bilirubin levels decreased by 72% to 4.11 mg/dL ± 3.96 mg/dL 2 weeks after the procedure from an initial baseline of 14.7 mg/dL ± 7.11 mg/dL (p<0.001; Fugazza 2022).

    In the case series of 120 people, decrease in bilirubin levels of at least 50% within 2 weeks of procedure was 95% (92/97). The median serum bilirubin levels 2 weeks after procedure were 76.5 micromoles/litre (range 8 to 352) compared with 222.5 micromoles/litre (range 45 to 687) at baseline (p<0.001; On 2022).

    In the case series of 215 people, clinical success (defined by an appropriate reduction in serum bilirubin – exact threshold not reported) was 87% (89/102) across all patients in which this outcome data was available. When split by indication, clinical success in malignant patients was 86% (63.72) and clinical success in benign patients was 90% (27/30; Hathorn 2022).

    In the case series of 205 people, the proportion of people with either >50% decrease in plasma bilirubin 1 month after procedure or sufficient improvement in health for discharge was 93% (143/153) in patients with available data (Marx 2022).

    In the non-randomised comparative study of 182 people, clinical success (defined as relief of cholangitis, jaundice, or pruritus with at least a 30% decrease in pre-treatment bilirubin level within a week of placement or normalisation of bilirubin within 30 days) was 86% (75/87) in EUS-HGS and 100% in EUS-CDS (80/80; Tyberg 2022).

    In the RCT of 125 people, clinical success of EUS-BD was 90% (54/60); compared to ERCP which had clinical success of 95% (52/55), there was no statistically significant difference (p=0.49; Paik 2018).

    Rate of reintervention

    Rates of reintervention were reported in 8 studies. In the systematic review and meta-analysis of 1,437 people, the pooled rate of reintervention was 7% (95% CI 3.8% to 10.8%; n=19 studies; I2= 69.3%; Dhindsa 2020).

    The rate of reintervention was 9% (22/239) in the case series of 256 people with DMBO (Fugazza 2022), 8% (9/109) in the case series of 120 people with DMBO (On 2022), and 19% (19/103) in the case series of 103 people with benign biliary obstruction (Fuldner 2021).

    In the case series of 215 people, 17% (22/126) of patients with malignant disease with 6-month follow-up required reintervention (Hathorn 2022).

    In the case series of 205 people, 36% (47/130) of people experienced recurrent biliary obstruction requiring reintervention, with a mean number of 1.7 reinterventions per patient (Marx 2022).

    In the non-randomised comparative study of 182 people, the rate of re-stenting was 29% (25/87) for EUS-HGS and 10% (8/80) for EUS-CDS; this difference in rate of re reintervention between EUS-BD techniques was statistically significant (p=0.01; Tyberg 2022).

    In the RCT of 125 people, there was a statistically significant lower rate of rate of reintervention for EUS-BD in ITT analysis compared to ERCP (16% [10/64] versus 43% [26/61], p=0.001; Paik 2018).

    Stent patency

    Stent patency was reported in 3 studies. In the case series of 205 people, mean stent patency duration was 153±433 days (Marx 2022).

    In a non-randomised comparative study of 182 people, EUS-CDS was 4.5 times more likely than EUS-HGS to achieve longer stent patency (>30 days) or manage obstruction (OR 4.5 [95% CI 1.15 to 17.65], p=0.0302; Tyberg 2022).

    In an RCT of 125 people, stent patency at 6 months according to Kaplan–Meier estimation was 85% for EUS-BD and 49% for ERCP, with a statistically significant difference in stent patency between the 2 techniques (p=0.001; Paik 2018).

    Survival

    Survival and mortality was reported in 4 studies. In the case series of 215 people, 51% (66/130) of people with malignant biliary obstruction died within a mean follow-up period of 457 days because of progressive disease, and 9% (8/85) people with benign disease died during a mean follow-up of 125 days due to causes unrelated to the procedure (Hathorn 2022).

    In the case series of 205 people with malignant biliary obstruction, the median survival was 5.3 months (95% CI 2.9 to 7.5 months), and 15% (30/205) patients died within 30 days of the procedure (Marx 2022).

    In the non-randomised comparative study of 182 people, 42% (77/182) survived for at least 6 months following the procedure (Tyberg 2022).

    In the RCT of 125 people, the median overall survival according to Kaplan–Meier estimation was 178 days for EUS-BD and 144 days for ERCP, with no statistically significant difference between the two (p=0.92). In the ITT analysis, the percentage of deaths due to disease progression was 72% (46/64) for EUS-BD and 84% (51/61) in ERCP (Paik 2018).

    Improvement in quality of life

    Improvement in quality of life was reported in 1 study. In the RCT of 125 people, quality of life at 12 weeks was assessed using the EORTC QLQ-C30 (scored 1 to 100 for each parameter); there were statistically significant differences between EUS-BD and ERCP in mean change in global functioning (+4.17 EUS-BD versus ‑9.03 ERCP, p=0.001), mean change in emotional functioning (+1.62 EUS-BD versus -9.72 ERCP, p=0.001) and mean change in cognitive functioning (+0.93 for EUS-BD versus -11.11 ERCP, p=0.003), where higher scores in these domains indicate improved quality of life.

    In the same study, there were also statistically significant differences in EORTC QLQ-C30 (scored 1 to 100 for each parameter) between EUS-BD and ERCP in mean change in fatigue score (-3.40 for EUS-BD versus +8.02 ERCP, p=0.02), mean change in pain score (-17.59 for EUS-BD versus +4.63 ERCP, p=0.01) and mean change in financial difficulties (+2.78 for EUS-BD versus 18.52 ERCP, p=0.01), where higher scores in these domains indicate worse symptoms or outcomes (Paik 2018).

    Safety

    Procedural death

    Procedure-related death was reported in 1 study. In the case series of 205 people with malignant biliary obstruction, 5% (10/205) of people died of procedure related causes; in 1 person this was due to biliary leakage, in 1 person this was due to severe sepsis, in 1 person this was due to stent obstruction, in 3 people this was due to stent migration, in 1 person this was due to acute kidney injury, in 1 person this was due to acute pulmonary oedema, and in 2 people this was due to a combination of minor complications (Marx 2022).

    Perforation

    Perforation was reported in 4 studies. Duodenal perforation was reported in 2% (2/120) of people in the case series of 120 people (On 2022). Perforation was also reported in 1 person in the case series of 215 people (Hathorn 2022), in 1 person in the case series of 103 people (Fuldner 2021), and in 2% of people in both EUS-HGS (2/95) and EUS-CDS (3/87) in the non-randomised comparative study of 182 people (Tyberg 2022).

    Sepsis

    Sepsis was reported in 2 studies. In the case series of 215 people, 2% (5/215) of people had sepsis secondary to cholangitis and 2% (4/215) of people had septic shock (Hathorn 2022). In the case series of 205 people, there was one procedural related death from severe sepsis (Marx 2022).

    Biloma

    Biloma was reported in 3 studies and a case report. The case report outlined the case of a woman with formation of biloma which then became infected, leading to liver and brain abscesses. This was successfully treated with antibiotics and additional stenting (Takahashi 2022).

    Biloma or abscess was reported in 2% (5/215) of people in the case series of 215 people (Hathorn 2022), and biloma was reported in 1 person who had EUS-HGS in the non-randomised comparative study of 182 people (Tyberg 2022).

    In the case series of 103 people, 3% (3/103) of people had either biloma or a liver abscess (Fuldner 2021).

    Infection

    Infection was reported in 4 studies. In the systematic review of 1,437 people, the pooled rate of infection was 4% (95% CI 2.8% to 5.1%, I2 = 0%; Dhindsa 2020)

    Infection was also reported in 2% of people in the in the case series of 256 people (Fugazza 2022), and in the non-randomised comparative study of 182 people, infection was reported in 2% (2/95) of patients who had EUS-HGS and 3% (3/87) of patients who had EUS-CDS (Tyberg 2022).

    In the case series of 215 people, bacteraemia was reported in 2% (4/215) of people (3 cases in the benign group and 1 in the malignant group), and fungaemia was reported in 1 person. Fever requiring antibiotics was reported in <1% (2/215) of people (Hathorn 2022).

    Stent migration

    Stent migration was reported in 7 studies. In the systematic review of 1,437 people, the pooled rate of stent migration was 4% (95% CI 2.5% to 6.2%, I2 = 43.5%; Dhindsa 2020).

    Stent migration was also reported in <1% of people (2/239) in the case series of 256 people (Fugazza 2022), in 1 person without combined cholangitis and 5% (6/120) of people with combined cholangitis in the case series of 120 people (On 2022), in 10% (20/205) of people (with 3 subsequent deaths) in the case series of 205 people (Marx 2022), in 1 person in the RCT of 125 people (Paik 2018) and in 11% (11/103) of people in the case series of 103 people (Fuldner 2021).

    Stent migration was reported in 1 person in the EUS-HGS group and 1 person in the EUS-CDS group in the non-randomised comparative study of 182 people (Tyberg 2022).

    Stent occlusion

    Stent occlusion was reported in 3 studies. Stent occlusion was reported in 7% (16/239) of people in the case series of 256 people (Fugazza 2022), in 6% (7/120, 6 with combined cholangitis, 1 without cholangitis) of people in the case series of 120 people (On 2022), in 4% (8/205) of people in the case series of 205 people (with 1 subsequent death; Marx 2022).

    Cholangitis

    Cholangitis without accompanying stent occlusion was reported in 4 studies.

    Cholangitis was reported in 7% (5/120) of people in the case series of 120 people (On 2022), 3% (6/215) of people in the case series of 215 people (Hathorn 2022), 2% (5/205) of people in the case series of 205 people (Marx 2022), and in 6% (4/64) of people in the RCT of 125 people (Paik 2018).

    In the non-randomised comparative study of 182 people, cholangitis was reported in 2% (2/95) of people who had EUS-HGS, and in 1% (1/87) of people who had EUS-CDS (Tyberg 2022).

    Bile leak

    Bile leak was reported in 6 studies, including the case report of biloma described above (Takahashi 2022). In the systematic review of 1,437 people, the pooled rate of bile leak was 4% (95% CI 2.7% to 6.2%, I2 = 46.7%; Dhindsa 2020).

    Bile leak was also reported in 2% (2/120) of people in the case series of 120 people (On 2022), in 1 person in the case series of 215 people (Hathorn 2022), in 3% (6/205) of people in the case series of 205 people (1 of which led to subsequent death; Marx 2022) and in 3% (3/103) of people in the case series of 103 people (Fuldner 2021).

    Bleeding

    Bleeding was reported in 5 studies. In the systematic review of 1,437 people, the pooled rate of bleeding was 4% (95% CI 2.3% to 5.3%, I2 = 41.8%; Dhindsa 2020).

    Bleeding was also reported in 2% (4/239) of people in the case series of 256 people (3 moderate cases, 1 severe case; Fugazza 2022), in 1 person in the case series of 215 people (Hathorn 2022), in 2% (3/205) of people in the case series of 205 people (Marx 2022), and 6% (6/103) of people in the case series of 103 people (Fuldner 2021).

    In the non-randomised comparative study of 182 people, bleeding was reported in 6% (6/95) of people who had EUS-HGS and 3% (3/87) of people who had EUS-CDS (Tyberg 2022).

    The case series of 215 people also reported haematemesis in 1 person (Hathorn 2022).

    Peritonitis

    Peritonitis was reported in 2 studies; in 1 person who had EUS-HGS in the non-randomised comparative study of 182 people (Tyberg 2022), and in 1 person in the RCT of 125 people (Paik 2018).

    Pneumoperitoneum

    Pneumoperitoneum was reported in 4 studies. In the systematic review of 1,437 people, the pooled rate of pneumoperitoneum was 3% (95% CI 2.3% to 4.6%, I2 = 2%; Dhindsa 2020).

    Pneumoperitoneum was also reported in 2% (2/120) of people in the case series of 120 people (On 2022) and in 2% (5/205) of people in the case series of 205 people (Marx 2022). Self-limited pneumoperitoneum was also reported in 3% (2/64) of people in the RCT of 125 people (Paik 2018).

    One person in the case series of 120 people experienced retroperitoneal air as a result of the procedure (On 2022).

    Pain

    Pain was reported in 2 studies; in 3% (6/215) of people in the case series of 215 people (Hathorn 2022), and 2% (2/103) of people in the case series of 103 people (Fuldner 2021).

    Other adverse events

    Other adverse events were reported in 2 studies. In the case series of 215 people, 1 person reported haemodynamic instability,1 person reported respiratory distress and 1 person reported a Mallory-Weiss tear as a result of postprocedural nausea and vomiting (Hathorn 2022).

    In the non-randomised comparative study of 182 people the rate of non-specified adverse events was 5% (5/95) in the EUS-HGS group and 12% (10/87) in the EUS-CDS group (Tyberg 2022).

    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).

    For this procedure, professional experts did not describe any additional anecdotal or theoretical adverse events.

    Four 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.

    Validity and generalisability