EUS & ERCP Monthly

A free monthly literature review from the UK & Ireland EUS Society

Volume 1, Issue 4 · May 2026

Editor: Dr Venkat Mahesh, MRCP, FEBGH — Consultant Gastroenterologist · Education & Media Lead, UKIEUS

Welcome

This month we screened the May 2026 EUS and ERCP literature across six PubMed search domains and ranked the field using our five-domain editorial scoring framework. The result is five featured papers with detailed commentary and six noteworthy mentions. A standout dose-de-escalation trial leads, alongside practical questions on stent dwell time, double malignant obstruction, and the consolidating evidence for EUS-gastroenterostomy.

A note on our scoring. Our five-domain score is an editorial prioritisation tool, not a validated quality instrument. It is built on SORT and GRADE principles and is applied identically each month. Scores rank papers relative to one another within a single issue; they are not absolute measures of quality. All references are independently verified on PubMed before inclusion.

Top 5 Papers — May 2026

#1 Score 88/100 · Multicentre RCT (noninferiority) · ERCP / PEP prophylaxis

Half the standard dose held: 50 mg rectal indomethacin matched 100 mg for preventing post-ERCP pancreatitis in 2,562 patients

Qi Z-P, Wan R, Wan X-J, et al. 2,562 patients, four Chinese centres. Endoscopy 2026;58(7):689–698. DOI · PMID 42140208

Guideline tension — ASGE, ESGE and ACG all name 100 mg as the standard dose; this trial questions whether that specific figure is necessary.

In this single-blind, 1:1 randomised noninferiority trial, patients with a normal baseline amylase undergoing ERCP received either 50 mg or 100 mg rectal indomethacin before the procedure. Post-ERCP pancreatitis occurred in 8.1% (103/1,269) of the 50 mg group versus 7.7% (100/1,293) of the 100 mg group — a risk difference of 0.38% (95% CI −1.71 to 2.47), comfortably within the prespecified noninferiority margin.

The result is best read against the existing dose-response evidence rather than in isolation. The foundational Elmunzer trial established that 100 mg indomethacin reduces PEP; the RIDE trial then showed that escalating to 200 mg conferred no advantage (12% vs 15%, NS); and network meta-analysis has treated 100 mg as the reference prophylactic dose. Qi now closes the lower end of that curve, suggesting the effective dose may sit below the figure every guideline specifies.

Three caveats temper enthusiasm. First, this is a single-blind trial in an exclusively Chinese population, and the authors themselves reserve 100 mg for patients with prior pancreatitis. Second, noninferiority around an event rate of ~8% cannot exclude a clinically meaningful difference in higher-risk subgroups, who were under-represented. Third, the parallel question of optimal diclofenac dosing remains unsettled, supported only by retrospective data.

Bottom line for your practice: For UK practice, 100 mg remains the standard and there is no supply reason to change it — but this is the strongest evidence yet that the effective dose is lower than guidelines state, and it offers genuine reassurance where only 50 mg suppositories are available or where minimising NSAID exposure matters.

#2 Score 68/100 · Systematic review + meta-analysis · Therapeutic EUS / collection drainage

Leaving a LAMS in longer did not raise complications: 662 patients argue against the rigid four-week removal rule

López-Arce G, Ruiz-Romero D, Borjas-Almaguer OD, et al. 15 prospective studies, 662 patients. Endoscopy 2026 (in press). DOI · PMID 42161260

Guideline tension — ESGE issues a strong recommendation to retrieve LAMS within 4 weeks, but on its own admission from low-quality evidence.

The four-week rule has thin foundations. The ESGE necrotizing-pancreatitis guideline (Arvanitakis et al., 2018) recommends retrieval within 4 weeks but grades the evidence as low, and the AGA practice update similarly cautions against leaving metal stents long-term — both driven by an interim randomised signal of delayed bleeding and embedment rather than comparative dwell-time data.

This meta-analysis pooled 15 prospective studies (662 patients) and, using a time-event assessment, found no association between indwell duration and adverse events; the overall AE rate was 19% (most commonly stent obstruction and bleeding), with LAMS-associated mortality of 0.8%. It does not stand alone — a 2025 meta-analysis by Maida and colleagues (1,820 patients) likewise found comparable AE rates for early versus late removal (RR 1.03, 95% CI 0.47–2.28).

Two caveats keep this honest. First, these meta-analyses share much of the same observational primary literature, so their agreement reflects a common evidence pool rather than independent replication. Second, pooled observational data carry less weight than the controlled signal that prompted the original rule — so this challenges, rather than overturns, the four-week convention. What the data more confidently support is that timing is not the dominant driver of harm: collection type, larger stent diameter, and pancreatic-head location appear to matter more.

Bottom line for your practice: When a collection needs longer drainage — serial necrosectomy, or simple logistics — these data are reassurance that a clinically-driven removal schedule with imaging surveillance is defensible, rather than rushing to pull a working stent at an arbitrary four weeks.

#3 Score 68/100 · Multicentre RCT (underpowered) · ERCP / benign biliary stricture

Six months of a wider covered stent matched twelve months of a narrower one for chronic-pancreatitis biliary strictures — the first randomised test of dwell time

Lindström O, Udd M, Grönroos J, et al. 62 enrolled, 46 analysed, six Nordic centres. Surg Endosc 2026;40(6):5154–5162. DOI · PMID 42151467

Guideline-consistent — current ACG guidance treats CP-related strictures with a covered metal stent for at least 6 months; this trial probes whether longer adds anything.

Where prior randomised trials settled stent type, this one is the first to randomise stent duration. Patients with symptomatic CP-associated benign biliary strictures received either a 10 mm fully-covered metal stent for 12 months or a 12 mm stent for 6 months. Recurrent stricture occurred in 3/24 (13%) of the 12-month arm versus 4/22 (18%) of the 6-month arm (p = 0.694), and one stent in the 12-month group became impacted and required surgical hepaticojejunostomy.

The result sits against a well-established backdrop: the landmark Coté JAMA trial showed covered metal stents resolve mixed-aetiology strictures as well as multiple plastic stents with fewer ERCPs, and the CP-specific Ramchandani trial confirmed equivalence of a single FCSEMS to plastic stents over a 12-month course. Neither tested whether the metal stent can come out earlier.

Two limitations dominate. First, with only 46 patients analysed, the trial is badly underpowered to prove equivalence — a 13% versus 18% difference across 46 patients carries confidence intervals wide enough to hide a real effect, so “no difference” here means “no difference detected.” Second, the arms differ in calibre as well as duration, so a shorter dwell and a wider stent are confounded; the trial cannot tell you which mattered. The surgical-impaction event is a concrete reminder of why prolonged dwell is cautioned against.

Bottom line for your practice: This is a reasonable signal — not proof — that a wider covered stent for six months may spare patients the embedment risk of a longer dwell without sacrificing durability. It supports, rather than changes, the existing six-month minimum, and a larger trial isolating duration from calibre is still needed.

#4 Score 64/100 · Multicentre retrospective (11 centres) · Therapeutic EUS / double obstruction

When the bile duct and the stomach outlet are both blocked, pairing hepaticogastrostomy with gastroenterostomy gave the most durable drainage — 159 patients

Beuchard T, Ratone J-P, Schaefer M, et al. (ENDO-GOBO). 159 patients, 11 tertiary centres. Ther Adv Gastroenterol 2026;19:17562848261446493. DOI · PMID 42186581

No guideline anchor — concurrent malignant biliary and gastric-outlet obstruction has no consensus algorithm; this is among the largest datasets informing the choice.

Among 159 patients (two-thirds with pancreatic cancer) treated with one of four endoscopic combinations, the pairing of EUS-hepaticogastrostomy with EUS-gastroenterostomy gave the lowest stent dysfunction (12.0%) and the highest 12-month obstruction-free survival (84.0%), versus 42–52% for the other combinations.

Two honest qualifications govern how far this should travel. First, the headline rests on the secondary survival analysis, not the primary endpoint: the raw dysfunction comparison across the four arms was not statistically significant (p = 0.059), and one of the three favourable pairwise survival comparisons missed at p = 0.0512. Second, this is retrospective and the combination was chosen by anatomy and operator, not randomised, so selection confounding is unavoidable.

What lifts the finding above a single study is independent reproduction: the earlier multicentre CABRIOLET study, from a separate European group, found the same hierarchy — EUS-GE+HGS with near-zero dysfunction at one end and enteral-stent-plus-choledochoduodenostomy worst at the other (an independent predictor of failure, HR 5.6). The mechanism is intuitive: a choledochoduodenostomy drains into the very duodenum that is obstructed, whereas a hepaticogastrostomy drains to the stomach, independent of the downstream blockage. That durability is not free — severe adverse events were more frequent with hepaticogastrostomy (8–15%) than choledochoduodenostomy (0–1.6%, p = 0.014).

Bottom line for your practice: In specialist EUS centres, when both limbs need endoscopic drainage and durable palliation is the goal, EUS-HGS paired with EUS-GE is now supported by two independent multicentre cohorts — accepting that it trades higher durability for higher procedural morbidity, and that choledochoduodenostomy-plus-enteral-stenting is the least durable combination.

#5 Score 58/100 · Meta-analysis with trial sequential analysis · Therapeutic EUS / malignant GOO

EUS-gastroenterostomy beat surgery on complications and stenting on durability across 22 studies — now backed by trial sequential analysis

Menegat ALRS, Menegat BLRS, Dantas CR, et al. 22 studies. Surg Endosc 2026;40(6):4573–4597. DOI · PMID 42120567

Guideline-converging — society guidance already positions EUS-GE as an expert-centre alternative; this synthesis, and three 2025 RCTs, push it toward first choice.

This meta-analysis pooled 22 studies comparing EUS-gastroenterostomy with surgical gastrojejunostomy and with enteral stenting. Against surgery, EUS-GE showed fewer overall complications (RR 0.38), fewer major complications (RR 0.43), and far less reintervention (RR 0.15), while surgery retained higher technical success (RR 0.98); against enteral stenting, EUS-GE achieved higher clinical success (RR 1.11).

Two cautions are essential. First, the surgery comparison pools 22,211 patients — a figure that can only come from large administrative-database studies, so its precision is real but its causal weight is limited by selection: who gets EUS-GE versus surgery is decided by frailty, anatomy, and local expertise, not randomisation. Large n is not the same as high-quality evidence.

Second, and more importantly, the meta-analysis no longer stands as the primary evidence: by 2026 there are three concordant randomised trials. DRA-GOO showed EUS-GE cut six-month reintervention versus duodenal stenting (4% vs 29%; RR 0.15); ENDURO showed faster return to solid intake versus surgery (median 1 vs 3 days; HR 2.21) with non-inferior reintervention; and a third trial reported a superior composite outcome favouring EUS-GE. The distinctive contribution here is the trial sequential analysis, which formally tests whether the accumulated evidence is now firm rather than liable to reversal.

Bottom line for your practice: For malignant gastric outlet obstruction in centres with the expertise and devices, EUS-GE is now the preferred palliative option — a conclusion resting chiefly on three 2025 randomised trials, which this meta-analysis corroborates and stress-tests. Enteral stenting remains the pragmatic fallback where EUS-GE expertise is unavailable or prognosis is very short.

Also Published — Noteworthy Mentions

Six further May 2026 papers worth a place on your radar, each verified on PubMed.

1. Cholangioscopy-guided lithotripsy for difficult bile duct stones. Retrospective, n=132: technical/clinical success 98.5%/94.7%; stones >20 mm (OR 4.79) and no post-procedural drainage (OR 3.73) independently predicted cholangitis — a practical argument for draining after large-stone lithotripsy. Minami H, et al. J Hepatobiliary Pancreat Sci. DOI · PMID 42216491

2. EUS-guided gallbladder drainage in perforated cholecystitis. Retrospective, n=12 high-surgical-risk patients with gallbladder perforation — often considered a contraindication: 100% technical and 91.7% clinical success with electrocautery LAMS, one conservatively-managed pneumoperitoneum. Small but provocative. Godat S, et al. J Gastroenterol Hepatol. DOI · PMID 42128424

3. Long-term outcomes of EUS-guided gallbladder drainage in non-surgical candidates. Multicentre retrospective, n=139 (plastic stent): overall clinical success 94.6%, but cholecystitis recurred far more with malignant cystic-duct obstruction than benign disease (13.3% vs 2.5%; malignancy OR 5.92) — these patients need closer follow-up. Harada K, et al. J Clin Med. DOI · PMID 42194584

4. EUS-gastroenterostomy in surgically-altered anatomy. Tertiary-centre retrospective, n=10 post-surgical vs 63 naïve: comparable clinical success (90.0% vs 90.5%) and adverse events, but longer procedures (61 vs 45 min) — reassuring feasibility data for a technically harder setting. Bronswijk M, et al. Clin Endosc. DOI · PMID 42253008

5. EUS-hepaticogastrostomy as primary drainage for malignant hilar obstruction. Phase I/II prospective, n=20: 100% technical and 75% clinical success — but all four clinical failures were Bismuth type IV, where right-duct drainage was still needed. A safety-first signal that primary EUS-HGS suits lower-grade hilar disease. Ogata T, et al. DEN Open. DOI · PMID 42088734

6. Percutaneous core biopsy vs EUS-FNA for pancreatic lesions. Retrospective with propensity matching, n=229: percutaneous US-CNB out-performed EUS-FNA (98.4% vs 71.0% diagnostic yield). Read with caution — the EUS-FNA yield is atypically low for contemporary practice (modern FNB routinely exceeds 90%), so this likely reflects dated FNA technique rather than a real indictment of EUS. Rzayev F, et al. Ultrasound Q. DOI · PMID 42172108

About this newsletter

EUS & ERCP Monthly is a free educational literature review from the UK & Ireland EUS Society (UKIEUS), edited by Dr Venkat Mahesh. It is provided for educational use only and does not constitute clinical guidance; management decisions must integrate individual patient factors and local expertise. © UKIEUS 2026.

UK & Ireland EUS Society · www.ukieus.co.uk