Barry Marshall drank H. pylori in 1984 to prove it causes ulcers. He got the Nobel Prize in 2005. Now he and colleagues argue in this month's NEJM review that the same bacteria — responsible for 89% of non-cardia gastric cancers — should be treated to prevent cancer, not just ulcers.
1. Retire empiric clarithromycin triple.
Surveillance from the vonoprazan program: 22.2% clarithromycin resistance, 69.2% metronidazole, 1.2% amoxicillin in US/EU isolates. Dual clarithromycin-metronidazole resistance 11.7%. Worldwide increase in macrolide resistance now drives failure of standard therapies.
ACG 2024, endorsed by NEJM: do not use clarithromycin triple or levofloxacin triple empirically. Use only if susceptible.
2. Use the new first-line toolbox.
When susceptibility unknown (which is most cases):
a) Bismuth quadruple x 14d — PPI + bismuth subcitrate + tetracycline 500mg qid + metronidazole. Still achieves 85-90% despite metronidazole resistance due to bismuth effect. Remains preferred in ACG.
b) Rifabutin triple x 14d — PPI + amoxicillin 1g BID/TID + rifabutin 150mg BID. Resistance rare. Achieves ∼80% as 3rd-line, >90% as 4th-line when combined as VAR (vonoprazan + amoxicillin + rifabutin) in Japanese data.
c) Vonoprazan dual / triple — Vonoprazan 20mg BID is a PCAB, not a PPI. Potent, CYP2C19-independent acid suppression. FDA approved 2023. Dual (vonoprazan + amoxicillin 1g TID) and triple (add clarithromycin) both listed as acceptable empiric alternatives. Japanese trials: vonoprazan triple 72.8% → 87.9% vs PPI triple.
Practical point: Order molecular resistance PCR where available — stool or gastric biopsy for 23S rRNA (clarithromycin), gyrA (levofloxacin). It replaces culture for most decisions.
3. Think screen-and-treat for gastric cancer prevention.
IARC Working Group Report published in NEJM March 2026 summarized 36% reduction in gastric cancer incidence and 22% reduction in mortality with population eradication.
IARC now supports population-based screen-and-treat in high-risk populations, starting age ∼45, using active infection tests (13C-UBT or stool antigen, not serology). Japan's experience since 2013 insurance expansion supports feasibility.
Key caveat in the review: eradicate before extensive atrophy / intestinal metaplasia. Beyond that, risk is reduced but not abolished. Balance with stewardship and potential microbiome trade-offs (inverse association with esophageal adenocarcinoma, asthma).
Bottom line for your order set: Delete clarithromycin triple as default. Default to bismuth quadruple x 14d, or if penicillin-allergic patient can take it, consider rifabutin triple or vonoprazan dual. Document allergy status, prior macrolide/fluoroquinolone exposure, and send resistance PCR if available.
References
Li H, Megraud F, Marshall BJ. Helicobacter pylori Infection — Advances, Challenges, and the Road Ahead. N Engl J Med 2026;395:1621-1633. https://nej.md/4Bj3O58
IARC Working Group Report on H. pylori screen-and-treat for gastric cancer prevention (12 March 2026). https://www.iarc.who.int/news-events/iarc-working-group-report-on-helicobacter-pylori-screen-and-treat-strategies-for-gastric-cancer-prevention/
Screen-and-treat cuts gastric cancer incidence by 36% and mortality by 22% — IARC/NEJM summary. https://ureabreathtests.com/blogs/news/13c-urea-breath-test-vs-14c-stool-antigen-serology-a-2026-procurement-guide-for-h-pylori-screening-programs and https://oncodaily.com/blog/nieves-martinez-lago-new-nejm-sounding-board---gastric-cancer-prevention
ACG 2024 guideline shift: recommendation to use non-clarithromycin regimens; rifabutin triple and vonoprazan-amoxicillin dual as preferred empiric alternatives. https://tmedweb.tulane.edu/content/treatment_h_pylori_induced_ulcers
Rising macrolide resistance accounting for failure of standard therapies. https://bpgweb.azurewebsites.net/article/S0000-0000(25)00123-X/fulltext
Vonoprazan triple increased eradication 72.8% → 87.9% vs PPI. https://ichgcp.net/clinical-trials-registry/NCT04527055