This Week in Gastroenterology — Sep 1, 2026

Welcome to This Week in Gastroenterology. This week we're covering 10 notable papers spanning upper gastrointestinal cancer prevention and non-invasive risk stratification, therapeutic advances in liver and immune-mediated gastrointestinal disease, and safety questions around the drugs we prescribe every day. Let's dive in.

We'll start with gastric cancer prevention, where two papers published back-to-back speak directly to each other. In Clinical Gastroenterology and Hepatology, Grell and colleagues pooled 83 prevalence studies and 23 progression studies of gastric premalignant lesions, and the headline is that not all intestinal metaplasia carries the same threat [7]. Complete and incomplete intestinal metaplasia were about equally common, each affecting roughly one in ten patients undergoing endoscopy, but the progression rate to gastric cancer with incomplete metaplasia was about seven times higher — roughly twelve cases per thousand person-years compared with under two. By contrast, the classic distinction between antrum-limited and corpus-involving disease showed no significant difference in progression. On the dysplasia side, high-grade dysplasia was rare, affecting well under one percent of patients, but progressed at a rate roughly thirty times that of low-grade dysplasia. The authors' point is a practical one: histologic subtyping is cheap, informative, and badly underused. If your pathology reports are not distinguishing complete from incomplete metaplasia, that is a conversation worth having with your pathologists. The catch is substantial heterogeneity across every pooled estimate, so treat the absolute numbers as approximations rather than surveillance intervals. Complementing this, Gastrointestinal Endoscopy published a meta-analysis from Luu and colleagues on the Endoscopic Grading of Gastric Intestinal Metaplasia, or EGGIM, applied in real time with narrow-band or blue light imaging [8]. Across seven studies and just over 1,100 patients, an EGGIM score of five or higher identified extensive metaplasia with about ninety percent sensitivity and ninety percent specificity, and the area under the curve was 0.96. Performance looked broadly similar in Eastern and Western cohorts despite very different disease prevalence. Taken together, these two papers point toward a workflow where the endoscopist stages extent visually during the procedure and the pathologist refines risk by subtype afterwards.

Staying with non-invasive risk stratification, the American Journal of Gastroenterology published a meta-analysis from Canto and colleagues on the capsule sponge with Trefoil Factor 3 for Barrett's esophagus screening [9]. Six studies, just under 1,500 participants with reflux symptoms and no prior Barrett's diagnosis: pooled sensitivity was about 87 percent and specificity about 88 percent, with a negative predictive value of roughly 96 percent at a 20 percent disease prevalence. Specificity was higher in smokers and in those on acid suppression, which is a useful wrinkle — the test may be most efficient at avoiding unnecessary endoscopy precisely in the higher-risk group you were most tempted to scope. A parallel logic appears in the Journal of Hepatology, where Moga and colleagues asked whether spleen stiffness could spare screening endoscopy in chronic portal vein thrombosis without cirrhosis, a population in which guidelines currently mandate endoscopy for everyone [4]. Across derivation and validation cohorts totalling 346 patients, a spleen stiffness of 40 kilopascals or less by vibration-controlled transient elastography ruled out high-risk varices with a negative predictive value of about 96 percent, and would have avoided just over four in ten endoscopies, missing three to five percent of high-risk varices. Liver stiffness, notably, was not useful here — these livers are not stiff, so the cirrhosis cut-offs simply do not transfer. This is retrospective, so prospective confirmation is needed, but it is the first credible de-escalation strategy in this group.

On the therapeutic side, three papers extend our options in diseases where we have had little. The Journal of Hepatology reports two-year placebo-controlled and three-year open-label data from the ELATIVE trial of elafibranor in primary biliary cholangitis [2]. At week 104 in the double-blind period, roughly two thirds OF ELAFIBRANOR-TREATED PATIENTS achieved a biochemical response versus none on placebo, though alkaline phosphatase normalization was achieved by only about one in nine. In the open-label extension, biochemical response was sustained at around sixty percent at two years and sixty-five percent at three years, and normalization rates rose over time to about a quarter of continuously treated patients. In those with moderate-to-severe fatigue or pruritus at baseline, itch and fatigue scores improved numerically relative to placebo, and there were no unexpected safety findings through three and a half years of exposure. For second-line therapy after inadequate ursodiol response, this is reassuring durability data. In the American Journal of Gastroenterology, Kinoshita and colleagues report a phase 3 trial of cendakimab, an anti-interleukin-13 antibody, in 48 Japanese adults and adolescents with eosinophilic gastritis or duodenitis — conditions with no approved therapy anywhere [5]. At sixteen weeks, cendakimab significantly reduced peak gastrointestinal eosinophil counts versus placebo, with histologic benefit maintained to week 48. Symptom improvements were only numerical, which is the familiar disconnect in eosinophilic gastrointestinal disease, and the trial is small and single-country, so read this as proof of mechanism rather than a finished answer. Also in Clinical Gastroenterology and Hepatology, the USBIOS study from Ribaldone and colleagues addressed non-medical switching from reference ustekinumab to biosimilars in Crohn's disease [6]. Across 18 Italian centres, 337 stable, steroid-free patients switched and 125 continued reference drug; expanded clinical success at six months was about 92 percent in both arms, meeting the pre-specified non-inferiority margin, with biosimilar persistence near 94 percent and switch-back in barely one percent. The important caveat is that almost nobody in this cohort was on every-four-week dosing, so patients on intensified regimens remain unstudied.

Finally, two papers on the risks of the drugs themselves, plus a new guideline. In Clinical Gastroenterology and Hepatology, Forss and colleagues used nationwide Swedish registers from 2007 to 2024 to quantify serious infections in inflammatory bowel disease across therapy classes, with and without concomitant corticosteroids [3]. This is the cleanest absolute-risk data we have: in propensity-score matched comparisons within each therapy class, adding corticosteroids raised serious infection risk roughly two- to three-and-a-half-fold, with incidence rate differences of about four to eight extra serious infections per hundred person-years. The relative risk was highest in children and rose with cumulative steroid dose. The message is not new but the magnitude is: steroids, not the advanced therapy, are doing much of the infectious damage, and steroid sparing is an infection-prevention intervention. Separately, Annals of Internal Medicine published an updated systematic review from Moiz and colleagues covering 38 randomized trials and nearly 26,000 adults with overweight or obesity without diabetes on GLP-1 receptor agonists and co-agonists [10]. Placebo-subtracted weight loss ranged from about six percent with liraglutide up to about fifteen percent with semaglutide and nineteen percent with tirzepatide, with emerging multiagonists such as retatrutide and amycretin reaching the low twenties. For gastroenterologists the relevant number is that gastrointestinal adverse events occurred in about three quarters OF TREATED PATIENTS versus about forty percent on placebo, though discontinuation for adverse events stayed around eleven percent and serious events were rare. Heterogeneity precluded pooling, and safety reporting was inconsistent. And rounding out the week, Gastrointestinal Endoscopy published the American Society for Gastrointestinal Endoscopy GRADE-based guideline on endoscopy in acute lower gastrointestinal bleeding, addressing colonoscopy versus computed tomography angiography as the first modality, urgent versus non-urgent timing, prepped versus unprepped colonoscopy, and for diverticular bleeding specifically, band ligation versus clipping and direct versus indirect clip placement [1]. Worth downloading the recommendations table.

If you only have time for one paper this week, make it the Swedish register study on corticosteroids and serious infections in inflammatory bowel disease [3]. It puts hard absolute numbers on a risk we tend to underweight, and it should change how urgently you pursue steroid withdrawal, especially in children and in patients accumulating repeated courses.

Here are the key takeaways from this week in Gastroenterology. First, corticosteroids substantially amplify serious infection risk in inflammatory bowel disease regardless of which advanced therapy the patient is on, with the greatest relative effect in children and with cumulative dose. Second, gastric intestinal metaplasia should be risk-stratified: incomplete subtype progresses far faster than complete, and real-time EGGIM scoring with virtual chromoendoscopy reliably identifies extensive disease. Third, two non-invasive tests are ready for careful adoption — capsule sponge with Trefoil Factor 3 to triage Barrett's screening in reflux patients, and spleen stiffness at or below 40 kilopascals to spare variceal screening endoscopy in non-cirrhotic chronic portal vein thrombosis. Fourth, elafibranor shows durable biochemical benefit through three years in primary biliary cholangitis, and anti-interleukin-13 blockade improves histology in eosinophilic gastritis and duodenitis, though symptom gains were only numerical. And fifth, switching stable Crohn's patients from reference ustekinumab to a biosimilar appears safe at six months, with the caveat that intensified every-four-week dosing was barely represented.

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 it saves you time, a quick rating in your podcast app helps other physicians find the show.

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