This Week in Gastroenterology — Aug 6, 2026

Welcome to This Week in Gastroenterology. This week we're covering 10 notable papers spanning positioning of advanced therapies in inflammatory bowel disease, risk stratification and finite therapy in chronic liver disease, and the quality and safety of endoscopic practice — from artificial intelligence in the colonoscopy suite to managing gastrointestinal bleeding when transfusion isn't an option. Let's dive in.

We'll start with inflammatory bowel disease, where the hardest question in clinic is what to reach for after the first advanced therapy fails. In Clinical Gastroenterology and Hepatology, the TRIDENT-UC study pooled 312 patients with ulcerative colitis across nine centres who had already failed at least one advanced therapy, and compared upadacitinib, tofacitinib and ustekinumab using propensity-score weighting [1]. At week sixteen, steroid-free clinical remission was achieved by roughly forty-eight percent of the upadacitinib group compared with about twenty-seven percent for each of the other two agents, and by one year three quarters of upadacitinib patients were in steroid-free remission versus a little over half in the comparator arms. The endoscopic remission gap was wider still — upadacitinib patients had several-fold higher odds of endoscopic remission at both time points. This is retrospective and unrandomised, so channelling bias can't be excluded, but the magnitude and consistency across biochemical and endoscopic endpoints reinforce what the head-to-head trial data have hinted at: in bio-exposed ulcerative colitis, upadacitinib deserves to sit near the front of the queue when there is no contraindication to a JAK inhibitor. Alongside that, Inflammatory Bowel Diseases tackled a related sequencing question after surgery [9]. Among 635 biologic-exposed Crohn's patients undergoing ileocolonic resection, giving no postoperative prophylaxis carried about a third higher risk of composite endoscopic or radiographic recurrence compared with reusing the same biologic class the patient had been on before. Switching to a new class was no better than reusing the old one, and in patients undergoing their first resection, or previously exposed only to anti-TNF agents, reutilisation actually produced longer recurrence-free survival. The practical message is that a preoperative biologic failure is not automatically a reason to abandon that class after surgery — the drug that failed to control luminal disease may still work perfectly well in the very different setting of preventing recurrence at a fresh anastomosis.

Turning to hepatology, and a genuine randomised trial addressing finite therapy in hepatitis B. Hepatology published the NUC-B trial, in which 156 non-cirrhotic, HBeAg-negative patients on nucleos(t)ide analogues were randomised either to stopping the drug alone, or to stopping it followed by sixteen weeks of pegylated interferon started four weeks later [2]. At three years, surface antigen loss occurred in three percent of those who simply stopped, compared with fourteen percent of those who received adjuvant interferon — a statistically significant difference, though the confidence interval is wide and the trial fell well short of its 240-patient recruitment target. Just as important for safety, exaggerated flares occurred in nearly twenty-eight percent of the withdrawal-only group but only about thirteen percent of the interferon group, and slightly fewer interferon patients had to restart antiviral therapy. So interferon here isn't just boosting functional cure — it appears to blunt the hazardous rebound that makes stopping therapy frightening. It is a small trial and interferon remains poorly tolerated, but for carefully selected patients pursuing finite therapy, this changes the calculus.

Staying with the liver, two papers speak to how much we can trust non-invasive tests. Gastroenterology reported the prospective FICUS cohort of 538 patients with uncomplicated primary sclerosing cholangitis followed for a median of five years with annual transient elastography [4]. Baseline stiffness stratified outcomes sharply: five-year transplant-free survival was about ninety-four percent below ten kilopascals, seventy-eight percent between ten and fifteen, and only forty-six percent above fifteen kilopascals. Crucially, the trajectory mattered too — patients with a significantly rising stiffness slope had roughly triple the risk of death or transplantation, and each one kilopascal per year of increase carried an eighteen percent higher risk. That supports serial elastography as a risk-stratification tool in primary sclerosing cholangitis and as a candidate surrogate endpoint in trials. But the Journal of Hepatology adds an important caveat about what counts as a real change [6]. Using repeat measurements from the screening pipeline of a phase two trial in biopsy-confirmed at-risk MASLD across 198 centres, the investigators quantified measurement variation. Liver stiffness by transient elastography had a within-subject coefficient of variation of about twenty percent, translating to a reproducibility coefficient in the range of fifty-five to sixty-four percent — meaning a single stiffness value would need to change by roughly half again before you could confidently call it real. Controlled attenuation parameter was more stable, and the enhanced liver fibrosis score was the most reproducible of all, with a reproducibility coefficient under ten percent. Read together, these two papers say: trust the direction of travel in elastography over years, as FICUS shows, but be very cautious about reacting to a single kilopascal shift between two visits.

Our third theme is endoscopic practice — efficacy, safety and quality. Gastrointestinal Endoscopy published a Bayesian network meta-analysis of six randomised trials and ten causally adjusted cohort studies, totalling nearly 1,600 patients with malignant gastric outlet obstruction, comparing endoscopic ultrasound-guided gastroenterostomy, enteral stenting and surgical gastrojejunostomy [5]. Endoscopic ultrasound-guided gastroenterostomy reduced re-intervention dramatically — roughly an eighty percent relative reduction versus enteral stenting and seventy percent versus surgery — with about twenty percent higher clinical success than stenting, six fewer days in hospital and faster return to oral intake compared with surgery. Technical success did not differ in direct pairwise comparisons, though the network favoured surgery on that metric, and the authors are careful to note these data come largely from expert centres. Still, where the expertise exists, this is now the default option to discuss with patients. From the same journal comes a more cautionary safety study in primary sclerosing cholangitis [7]. A registry-based comparison of eighty procedures under a new single-shot antibiotic prophylaxis protocol against a hundred historical procedures with multi-day antibiotics found no significant difference in the overall rate of post-ERCP cholangitis, even when dilation was performed, and antibiotic exposure dropped from a median of three and a half days to a third of a day. But among those who did develop cholangitis, the single-shot cohort had a markedly higher proportion of moderate-to-severe episodes — about forty-six percent versus nine percent. Dominant strictures and dilation each roughly tripled cholangitis risk. So antibiotic stewardship here looks feasible, but with a signal worth watching, particularly in patients undergoing dilation.

On quality measurement, the American Journal of Gastroenterology described an artificial intelligence tool applied to over 18,000 recorded colonoscopies from 55 attendings [10]. The automated measures of withdrawal time correlated almost perfectly with manual documentation, insertion time correlated moderately, and automated polyps-per-colonoscopy correlated well with both adenoma and serrated detection rates. It also accurately captured polypectomy technique, including a cold snare rate that varied from sixty-three to ninety-five percent across operators. This makes continuous, low-burden quality auditing plausible rather than aspirational. And Gut published a sobering national analysis of nearly 4,700 mismatch repair carriers in the English NHS [3]. Surveillance colonoscopy at intervals of three years or less was associated with lower colorectal cancer-specific and all-cause mortality, but with no reduction in overall cancer incidence and no strong evidence of stage shift; surveillance every two years or less was actually associated with higher total incidence, driven by early-stage cancers without a fall in late-stage disease. The authors raise overdiagnosis, spectrum bias and short follow-up as explanations, and selection bias cannot be excluded in a non-randomised design. It's a reminder that shorter intervals are not automatically better in Lynch syndrome.

Finally, a practical question in acute bleeding. The American Journal of Gastroenterology reported a nine-hospital propensity-matched cohort of 131 Jehovah's Witness patients hospitalised with gastrointestinal bleeding who declined transfusion, matched to 131 controls [8]. Thirty-day mortality did not differ significantly between the groups, and there were no differences in intensive care admission, vasopressor use, length of stay or readmission for rebleeding. Both groups reached the same median nadir haemoglobin of 6.6 grams per decilitre. Patients declining transfusion were more likely to need intermediate care admission and received more bloodless medicine consultation and anaemia-directed therapy. This is retrospective and modest in size, but it should reassure teams that a restrictive, transfusion-free pathway with intensive supportive management is not inherently catastrophic.

If you only have time for one paper this week, make it the TRIDENT-UC study in Clinical Gastroenterology and Hepatology [1]. It gives you the real-world comparative data you need for the decision you make most often in the ulcerative colitis clinic — what to do after the first advanced therapy fails.

Here are the key takeaways from this week in Gastroenterology. In bio-exposed ulcerative colitis, upadacitinib outperformed both tofacitinib and ustekinumab on steroid-free and endoscopic remission in real-world practice. After ileocolonic resection in Crohn's disease, reusing a previously failed biologic class is a legitimate prophylactic strategy, and doing nothing is the worst option. Adding a short course of pegylated interferon after stopping nucleos(t)ide analogues in HBeAg-negative hepatitis B raised surface antigen loss and halved exaggerated flares. Transient elastography, both the absolute value and the slope over time, stratifies outcomes powerfully in primary sclerosing cholangitis — but remember that its short-term measurement variability is around twenty percent, so don't over-interpret a single change. And endoscopic ultrasound-guided gastroenterostomy now looks superior to both stenting and surgery for malignant gastric outlet obstruction in expert hands.

That's your roundup for This Week in Gastroenterology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.