Gastroenterology & Hepatology
Endoscopy within six hours did not improve outcomes in non-variceal upper GI bleeding
A meta-analysis of 10,785 patients supports resuscitation first and endoscopy within 24 hours; a European consensus sets out care for symptoms after oesophago-gastric cancer surgery; and one in thirty hospital nurses carries hepatitis B.
First surveillance colonoscopy at 5 years after high-risk adenoma removal was non-inferior to 3 years
EPoS II's interim analysis supports spacing out surveillance after high-risk adenomas; half of strictly immune-tolerant hepatitis B patients became immune-active within five years; and even one Z allele raised the risk of major liver outcomes.
Lactulose cut falls in cirrhosis patients who had never had overt encephalopathy
LiveSMART found lactulose, with or without tele-delivered tai chi, reduced injurious and non-injurious falls in cirrhosis with portal hypertension; a multi-target blood test found three times as many early liver cancers as ultrasound but missed its specificity target; and family history raised colorectal cancer risk in IBD by the same absolute amount as in everyone else.
Non-enteric-coated pancreatic enzymes did not relieve pain in chronic pancreatitis
A double-blind trial of non-enteric-coated enzyme therapy found no pain benefit over placebo at 12 or 24 weeks. Also: a 30% fall in liver stiffness is only a modest marker of fibrosis regression in MASH, rebamipide as gastroprotection on DOACs, and a generic omeprazole supplement.
Cold snare resection of large colorectal polyps cut major complications to a third, but doubled residual adenoma
An individual patient data analysis of four trials sets out which large polyps suit cold resection. Also: AGA practice advice on MetALD, an international consensus on colorectal cancer risk in MASLD, and Rome V confirms that four in ten adults meet criteria for a gut-brain disorder.
Spleen stiffness of 40 kPa or less can spare endoscopy in non-cirrhotic portal vein thrombosis
Across 346 patients with chronic PVT from 16 VALDIG centres, spleen stiffness ≤40 kPa missed 3–5% of high-risk varices and could spare about four in ten screening endoscopies. Also: alcohol and cardiometabolic risk in cirrhosis, seven-day fexuprazan triple therapy for H. pylori, DPYD variants beyond the standard panel, and an FDA supplement for an adalimumab biosimilar.
Barrett's neoplasia: nine in ten Dutch patients had never had an endoscopy before diagnosis
A nationwide cohort shows surveillance finds only a sliver of Barrett-related cancer. Also: cardiovascular risk with IBD advanced therapies, the gut microbiome and cirrhosis admissions, a premixed IV pantoprazole recall, and IV PPI dosing after haemostasis.
MASH has two licensed drugs and an expert framework for choosing between them
An international panel sets out how to pick, monitor and switch between resmetirom and semaglutide in F2-F3 fibrosis, individual patient data from eight trials put numbers on peg-interferon add-on in hepatitis B, and a synthesis of 8,993 patients quantifies what happens when a GLP-1 is stopped.
One to two in five ampullary adenomas are already cancer, and six biopsies is the minimum
The AGA sets out how to assess and resect ampullary lesions; bulevirtide holds for 144 weeks and mostly relapses when stopped; a Cochrane review finds the thymosin-α1 evidence too weak to act on; and a global survey puts paediatric gut-brain disorders at 28%.
Half of 'immune tolerant' hepatitis B is not tolerant for five years
A 951-patient international cohort followed 13 years finds 51% transition to immune active disease within five years, with high-normal ALT tripling the risk; a blood test outperforms ultrasound for small hepatocellular carcinoma; and the IBD-with-IBS overlap finally gets a name.
HBsAg loss may not mean what the trials assume, and lactulose cut falls in cirrhosis
Integrated HBV DNA produces surface antigen indistinguishable from the cccDNA-derived kind, which undermines HBsAg as a functional-cure endpoint; and in 230 patients with portal hypertension, lactulose reduced injurious falls from 12% to 4%.
Pancreatic enzymes did nothing for pain in chronic pancreatitis, and a 30% fall in liver stiffness is a weaker signal than the guideline implies
An Indian randomised trial closes the question of whether non-enteric-coated enzyme replacement relieves pancreatic pain; a 30% decline in transient elastography detects fibrosis regression in MASH with an AUC of 0.68; 48 trials find no excess gastrointestinal cancer with SGLT2 inhibitors; and surveillance colonoscopy at five years holds against three.
Large colorectal polyps: cold snare had about 70% fewer major complications but twice the residual adenoma
A patient-level analysis of four trials defines where cold resection fits, experts call for earlier colorectal screening in MASLD, and even heterozygous alpha-1 antitrypsin carriers had more serious liver outcomes.
Spleen stiffness can spare endoscopy in non-cirrhotic portal vein thrombosis
A 40 kPa cut-off missed about 5% of high-risk varices and avoided four in ten endoscopies. Also: how alcohol and metabolic risk combine in cirrhosis, machine perfusion for liver grafts, seven-day P-CAB therapy for H. pylori, and an FDA supplement for an adalimumab biosimilar.
Cardiovascular risk of advanced IBD therapies: 43 studies, and no clear answer either way
Randomised and observational evidence point in opposite directions for anti-TNF agents, and every class estimate has an interval wide enough to contain both benefit and harm. Plus a Class II recall of premixed intravenous pantoprazole.
A framework for the two MASH drugs, and functional cure of hepatitis B turns entirely on where the antigen starts
An international panel consolidates phase 3 data and FDA labels for resmetirom and semaglutide into one care pathway for F2-F3 MASH. Separately, pooled individual data show peg-interferon add-on clears HBsAg in 37.7% below 100 IU/mL and 2.3% above 1000.
Thirty-five years of thymosin trials in hepatitis B, and Cochrane cannot tell you whether it works
Ten randomised trials, very low certainty on every outcome but one; the AGA sets out the endoscopic ground rules for ampullary adenomas; and antibiotic remission in pouchitis turns out to be a six-week suppression of exotoxin genes.
Why the gut leaks in cirrhosis, and why anxiety in IBD is not just a reaction to having IBD
IgA bound to liver macrophages in alcohol-associated disease, the vascular barrier that fails before bacteria translocate, ferroptosis as the initiating event in metabolic liver injury, and the neuropsychiatric comorbidity that belongs in the IBD clinic.
Misfolded protein in routine gut biopsies predicted dementia seven years out
Archival biopsies from unexplained GI symptoms read for neurodegenerative proteinopathy, bile acid patterns across 5,056 IBD patients, the FDA's own view of drug-induced liver injury, and every segment of the gut in scleroderma.
The 30% elastography rule is weaker than the guideline implies
A fall of 30% in liver stiffness identifies fibrosis regression in MASH only modestly well; SGLT2 inhibitors look clean for gastrointestinal cancer over short follow-up; and a Swedish register argues surveillance in IBD should follow the number of affected relatives, not their age.
Two in five adults meet criteria for a gut-brain disorder, and the number has not moved
Rome V finds the same 41% prevalence as Rome IV in 28,771 adults across 15 countries; bulevirtide holds hepatitis D for 144 weeks but mostly relapses when stopped; and a consensus panel argues MASLD should change how early you screen the colon.
Liver cancer surveillance should be aimed at who benefits, not just who is at risk, and patient navigation beats every other way of raising screening uptake
An expert opinion argues abbreviated MRI and biomarker panels need benefit stratification rather than risk stratification, a network meta-analysis of 76 trials ranks colorectal screening interventions, and gene therapy cuts cornstarch dependence in glycogen storage disease type Ia.
Immunotherapy added to TACE delays progression in liver cancer, and a blood test outperforms surveillance ultrasound
A phase 3 trial moves progression-free survival in embolisation-eligible hepatocellular carcinoma without yet moving overall survival, Z-allele heterozygotes turn out to carry real liver risk, and hepatitis B carries a higher relative cancer risk than smoking does for the lung.
Naltrexone tripled abstinence in compensated alcohol-associated cirrhosis
An Indian randomised trial gave 100 patients with compensated alcohol-related cirrhosis naltrexone 50 mg or placebo for 12 weeks: abstinence at 12 weeks was 64% versus 22%, with no hepatic decompensation attributable to the drug. Plus spleen stiffness spares endoscopy in portal vein thrombosis, and the ASGE rewrites lower GI bleeding.
Naltrexone tripled abstinence in compensated alcohol-related cirrhosis
An Indian trial finds naltrexone safe and effective in compensated alcohol-associated cirrhosis, spleen stiffness can spare four in ten screening endoscopies in portal vein thrombosis, and acetate buffering adds nothing to lactated Ringer's for post-ERCP pancreatitis.