GI and Hepatology Updates, August 2026: Five Developments That Change What You Do Monday Morning

A patient with compensated cirrhosis and a liver stiffness of 18 kPa asks whether she needs a beta-blocker before she has ever bled. A pancreatic cancer patient two lines into treatment has no good next option. A HER2-positive gastric cancer diagnosis lands on your desk and the first-line regimen you'd have reached for six months ago is no longer the strongest evidence-backed choice. August 2026 moved the needle on all three of these conversations, plus two more. Here's what actually changed and what to do differently as a result.

Baveno VIII redefines how early you intervene in portal hypertension

The Baveno VIII Consensus, published online in the Journal of Hepatology on 6 August and highlighted by EASL on 25 August, is not an incremental update. Its 272 statements span compensated advanced chronic liver disease (cACLD), clinically significant portal hypertension (CSPH), variceal bleeding, ascites, TIPS, recompensation, portal vein thrombosis, Budd–Chiari syndrome, and portosinusoidal vascular disorder - and for the first time, dedicated pediatric recommendations.

The conceptual shift that matters most in daily practice: non-invasive tests (NITs) are no longer framed purely as a way to rule in or rule out CSPH. Baveno VIII repositions elastography and companion markers as tools to identify patients at risk of a first decompensating event, which pulls preventive therapy earlier in the disease course rather than waiting for a variceal bleed or new ascites to trigger treatment. Thresholds for NITs, bleeding management, and TIPS timing were also refined. For a patient sitting at 18 kPa with no prior decompensation, this is the difference between "come back if something happens" and a structured conversation about non-selective beta-blockade now. Tools that quantify decompensation risk alongside liver stiffness - our MELD-Na interpretation guide covers the related question of when a compensated patient's trajectory is shifting - are worth having in the same workflow. For a broader comparison of scoring systems as this data gets folded into risk conversations, the MELD 3.0 calculator is a useful companion reference.

RAS finally becomes a druggable target in metastatic pancreatic cancer

On 26 August, the FDA approved daraxonrasib (Rasonque), a once-daily oral pan-RAS inhibitor, for adults with metastatic pancreatic adenocarcinoma after at least one prior systemic therapy, or for patients who cannot tolerate multiagent chemotherapy. This is not a niche approval restricted to a rare KRAS subtype - daraxonrasib inhibits both active mutant and wild-type RAS signaling, which is what makes it broadly applicable in a tumor type where roughly 90% of cases carry a KRAS mutation.

The phase 3 RASolute 302 trial, published in NEJM, enrolled 500 patients and reported a median overall survival of 13.2 months versus 6.7 months with chemotherapy (HR 0.40), median progression-free survival of 7.2 versus 3.6 months, and an objective response rate of 30% versus 11%. Those are not marginal numbers for pancreatic adenocarcinoma, a disease where second-line options have historically added weeks, not months. For a patient progressing after first-line FOLFIRINOX or gemcitabine/nab-paclitaxel, sequencing now needs to account for a genuine survival benefit rather than a palliative holding pattern.

HER2-positive gastroesophageal adenocarcinoma gets a new first-line standard

The FDA approved zanidatamab plus tislelizumab plus fluoropyrimidine/platinum chemotherapy on 25 August as first-line therapy for HER2-positive unresectable or metastatic gastric, gastroesophageal-junction, or esophageal adenocarcinoma (IHC 3+, or IHC 2+/ISH+). Zanidatamab plus chemotherapy alone, without tislelizumab, also received a separate indication for IHC 3+ disease.

In the phase 3 HERIZON-GEA-01 trial, also peer reviewed in NEJM, the triplet produced a median overall survival of 26.4 months versus 19.2 months with trastuzumab-based chemotherapy (HR 0.72), and median progression-free survival of 12.4 versus 8.1 months (HR 0.63). Trastuzumab-based regimens have been the HER2 backbone in upper-GI adenocarcinoma for over a decade; this dataset is the strongest challenge to that default so far. Practically, it raises the stakes on getting HER2 IHC and ISH testing right at diagnosis, since the treatment path now bifurcates more consequentially than it used to.

Ustekinumab moves into pediatric ulcerative colitis from age 2

The FDA approved ustekinumab for moderately-to-severely active ulcerative colitis in children as young as 2 years old on 28 August, extending an indication that already covered pediatric Crohn's disease earlier this year. It's now the first FDA-approved non-TNF monoclonal antibody available for both major forms of pediatric IBD in this age group. Rather than a large new placebo-controlled pediatric efficacy trial, the approval leaned on established adult UC efficacy data, pediatric Crohn's pharmacokinetic and safety data, and a 52-week study in 112 children with UC - an evidence-extrapolation approach that's becoming more common in pediatric IBD regulatory pathways.

For very young patients where anti-TNF therapy has failed or isn't appropriate, there's now an IL-12/23 pathway option rather than a gap in the algorithm. The EMA's CHMP issued a positive opinion for the same pediatric UC indication on 25 June, so this isn't isolated to US practice. As biologic sequencing in pediatric IBD gets more granular, it's also worth keeping an eye on mechanistically distinct subgroups - emerging work on anti-IL-10 autoantibody-driven IBD is a reminder that "refractory" disease sometimes has a specific, targetable driver rather than just needing the next biologic in line.

AGA says stop surveilling Barrett's and colorectal neoplasia by age alone

A new AGA Clinical Practice Update, published online 3 August in Clinical Gastroenterology and Hepatology and summarized by AGA on 24 August, makes a de-implementation argument: endoscopic surveillance in older adults should be driven by life expectancy, comorbidity burden, prior neoplasia or polyp history, procedural risk, and whether the patient could actually undergo treatment if cancer were found - not by age or an automatic recall interval.

For Barrett's esophagus, an inability to tolerate endoscopic therapy, surgery, or oncologic treatment should prompt stopping surveillance regardless of age. For colorectal screening and polyp surveillance, individualization should start after age 75, with colonoscopy risk explicitly weighed rather than assumed acceptable. This is nine Best Practice Advice statements from an expert review, not a formally graded guideline, so the recommendations leave room for judgment - but the underlying message is clear: "another surveillance scope because it's due" is no longer a sufficient rationale on its own, and documenting an explicit benefit-harm discussion is now the expected standard.

Bottom line for clinical practice

• Start using NIT-based risk stratification under Baveno VIII to identify cACLD patients at risk of first decompensation, not just to confirm CSPH after the fact.

• Reconsider second-line sequencing in metastatic pancreatic adenocarcinoma now that daraxonrasib offers a survival benefit (HR 0.40) broader than prior KRAS-subtype-restricted options.

• Confirm HER2 IHC/ISH status at diagnosis in gastroesophageal adenocarcinoma before defaulting to a trastuzumab-based first-line regimen.

• In pediatric UC unresponsive to or unsuitable for anti-TNF therapy, ustekinumab is now an approved option from age 2.

• Before continuing Barrett's or colorectal surveillance in an older adult, document an explicit discussion of life expectancy, comorbidity, and treatability - not just that the interval is due.

Keeping pace with five practice-changing developments in a single month isn't realistic to do from memory alone. Next time a case like this lands on your list, walk it through GastroAGI - it will reason through the current evidence and return a guideline-anchored answer in seconds.