Taiwan ran the experiment that most countries only argue about. On the Matsu Islands, six rounds of population-wide Helicobacter pylori screening and treatment between 2004 and 2018 reduced gastric cancer incidence by 53% against a historical control period, with a 95% confidence interval of 30% to 69%. The reduction in gastric cancer mortality over the same period was 25%, and it was not statistically significant.
That is the strongest evidence any country has produced on this question. It is also more complicated than the headline. The case for screening an island where two thirds of adults carried the bacterium is not the same as the case for testing one adult in Ohio.
Why is a stomach bacterium classified alongside tobacco?
Because "there is sufficient evidence in humans for the carcinogenicity of chronic infection with Helicobacter pylori."
That evaluation is from the International Agency for Research on Cancer, which reviewed H. pylori in Monographs Volume 100B. Its conclusion is IARC's top category. "Chronic infection with Helicobacter pylori is carcinogenic to humans (Group 1)."
The same monograph names the cancers, and is careful in the other direction on a third.
| Cancer | What IARC Volume 100B records |
|---|---|
| Non-cardia gastric carcinoma | Caused by chronic H. pylori infection |
| Low-grade B-cell MALT gastric lymphoma | Caused by chronic H. pylori infection |
| Esophageal adenocarcinoma | Evidence suggesting lack of carcinogenicity |
A Group 1 listing is a statement about causation, not about how large the risk is for any one carrier. Most people who carry H. pylori never develop gastric cancer.
What did the Taiwan mass eradication program actually measure?
Two papers in Gut report the Matsu program. They measured different things at different points in its life, and reading only one of them gives a distorted picture.
| Paper | Period covered | Main outcome measured | Result as published |
|---|---|---|---|
| Lee et al., Gut 2013 | Before 1995 to 2003, after 2004 to 2008 | Premalignant gastric lesions | Gastric atrophy incidence reduced 77.2%, 95% CI 72.3% to 81.2%. Reduction in intestinal metaplasia not significant. Gastric cancer incidence reduction 25%, rate ratio 0.753, 95% CI 0.372 to 1.524. |
| Chiang et al., Gut 2021 | Historical control 1995 to 2003, chemoprevention period to 2016 and 2018 | Gastric cancer incidence and mortality | Incidence reduced 53%, 95% CI 30% to 69%, p<0.001. Mortality reduced 25%, 95% CI -14% to 51%, p=0.18. |
The 2013 paper is the honest early answer. After four years the program had clearly reduced gastric atrophy. It had not yet shown a statistically significant reduction in cancer itself, and the authors said so, calling for long-term follow-up.
The 2021 paper is what long-term follow-up produced. Coverage reached 85.5% of the eligible population. H. pylori prevalence fell from 64.2% in 2004 to 15.0% in 2008, and reinfection stayed below 1% per person-year.
Gastric cancer incidence fell significantly. Mortality did not, and the authors wrote that a significant mortality reduction "is likely to be achieved with a longer follow-up period." That is a prediction, not a finding.
Does eradicating it carry any downside?
Yes, and both Matsu papers answer it with data rather than reassurance.
Treating a population-wide infection with acid suppression and antibiotics was followed by a measurable rise in reflux disease. The 2013 authors framed their own headline result as a gain in gastric atrophy purchased "at the expense of increased oesophagitis."
| What was measured | What the Matsu papers report |
|---|---|
| Peptic ulcer disease, 2013 paper | Fell 67.4% |
| Esophagitis after treatment, 2013 paper | Incidence 6% |
| Reflux esophagitis at endoscopy, 2021 paper | Around 25% through the full implementation period |
| Active peptic ulcer, 2021 paper | 17.3% in 2012 down to 3.0% in 2018 |
| Barrett's esophagus, histologically confirmed, 2021 paper | Stayed a rare diagnosis |
| Other digestive tract cancers, 2021 paper | No significant change in incidence |
| H. pylori antibiotic resistance rate, 2021 paper | No change |
The antibiotic worry did not materialize on Matsu. The 2021 authors attribute that to selecting effective regimens and retesting everyone who finished treatment.
That last clause matters. The resistance result belongs to a program that confirmed cure, not to one that treated and walked away.
Why does the United States not screen everyone for it?
The US Preventive Services Task Force has no recommendation on H. pylori screening, and no gastric cancer screening topic appears in its recommendation list.
Its H. pylori page records exactly how close it came. A draft research plan for screening for H. pylori infection was posted for public comment on November 17, 2022.
The Task Force then "determined that it will not move directly into the final research plan stage and evidence review at this time," citing the need to balance new topics against critical updates within a portfolio of more than 80. The topic is filed as Inactive.
That is not a verdict against testing. It is a statement that the United States has not yet done the evidence review, and the baseline is part of why.
| Measure | United States | Taiwan |
|---|---|---|
| Age-standardized stomach cancer incidence | 7.5 per 100,000 per year, SEER 2019 to 2023 | 9.29 per 100,000 in 2019 from the Taiwan Cancer Registry, down from 15.2 in 1999 |
| Lifetime risk of a stomach cancer diagnosis | About 0.8% | Not reported in the same form |
| Five-year relative survival | 39.8%, SEER 2016 to 2022 | Not reported in the same form |
| Adult H. pylori prevalence | Not measured by a single national screening survey | Age-standardized 32.3% of adults by breath test, 2019 to 2020 |
Those two national rates are closer than the mass-eradication story suggests. Matsu was not a national average. It is a locality where the gastric cancer rate is much higher than on Taiwan's main island and similar to high-risk coastal provinces of China. The program worked where the risk was concentrated.
The United States has its own concentration. The American College of Gastroenterology's 2025 guideline on gastric premalignant conditions states that gastric cancer incidence rates are two to thirteen times greater in non-White individuals, particularly early-generation immigrants from regions of high incidence, and that US five-year survival for gastric cancer is 36%.
The same guideline says there is insufficient evidence to recommend upper endoscopic screening for gastric cancer in US populations deemed high risk. Even for the highest-risk group in a low-incidence country, the endoscopic screening question is still open.
Who do the guidelines say should be tested?
All patients with current or past gastric or duodenal ulcers should be tested, and the guidelines name further groups alongside them.
The Houston Consensus Conference on testing for H. pylori in the United States produced those statements. The expert panel reached consensus on all six; the external validation group did not reach consensus on the last two.
- All patients with current or past gastric or duodenal ulcers should be tested, Grade 1A.
- All patients with uninvestigated dyspepsia should be tested, Grade 1A.
- Individuals with a family history of gastric cancer should be tested, Grade 1B.
- Patients who are first-generation immigrants from high-prevalence areas should be tested, Grade 1B.
- Family members living in the same household as someone with a proven active infection should be tested, Grade 1B.
- Individuals with a family history of peptic ulcer disease should be tested, Grade 1B.
Two further statements are weaker. Patients of Latino and African American racial or ethnic groups "may be considered" for testing because of high infection prevalence, Grade 2C. Patients taking proton pump inhibitors for more than a month were suggested for testing at Grade 2C, with the panel noting that long-term clinical data on adverse outcomes are still lacking.
The first statement in the whole document is the simplest. "We recommend that all patients with active H pylori infection be treated."
The 2024 ACG clinical guideline reaches the same operating position and adds what to treat with. When antibiotic susceptibility is unknown, bismuth quadruple therapy for 14 days is the preferred regimen for treatment-naive patients. Whether you fall into any of these groups is a conversation for you and your physician, not something an article can settle.
Which test, and why is serology the wrong one for confirming a cure?
Four tests are in routine use, and they do not answer the same question.
| Test | What it detects | Guideline position |
|---|---|---|
| Urea breath test | Active infection, using the urease enzyme H. pylori produces | Recommended for initial diagnosis and for confirming eradication, Grade 1A |
| Stool antigen immunoassay | Active infection | Recommended for initial diagnosis and for confirming eradication, Grade 1A |
| Biopsy during upper endoscopy | Active infection, plus the state of the stomach lining | If endoscopy is being performed, two biopsies each from the antrum and corpus should be obtained, Grade 1A |
| Serology, a blood antibody test | Past exposure, not current infection | "We recommend that serology not be utilized for detection of active H pylori infection", Grade 1A |
The serology problem is not academic. A study of more than 100 million people found serology was used in about 70% of first-time H. pylori tests, and only a minority of positive results had confirmatory testing.
Antibodies can persist after the bacterium is gone. A positive antibody result after treatment tells you almost nothing about whether the treatment worked.
Timing is part of the guideline too, and it is the clinician ordering the test who arranges it. The Houston Consensus sets out four statements on medication before an active-infection test.
| Medication | What the consensus statement says | Grade |
|---|---|---|
| Bismuth and antibiotics | Stopped at least 4 weeks before testing for active infection | 1C |
| Proton pump inhibitors | Discontinued at least 4 weeks before a breath test or stool antigen test | 1B |
| H2 blockers and antacids | May be used without affecting the accuracy of those two tests | 1A |
| PPIs not discontinued | A positive result still counts as positive. A negative one is repeated after an appropriate interval | 1C |
Read the last row before the second. The consensus does not treat a missed washout as a wasted test. It treats a negative result under those conditions as unresolved.
Nothing on that table is a direction to change any medication you take. Stopping or continuing a drug before a test is a decision for the physician ordering it.
The consensus is also blunt about follow-up. Every patient treated for H. pylori should receive post-treatment confirmation of eradication, using only a test for active infection, because current therapies fail in at least 20% of cases. Once eradication is confirmed, further testing is recommended against.
What does H. pylori testing cost in Taipei?
New Dawn Health lists the H Pylori Breath Test at $75, with a 15-minute appointment slot. For scale, the Centers for Medicare and Medicaid Services reimbursement for a urea breath test at the time of the Houston Consensus review was $91.89.
The Helicobacter pylori (H. pylori) Biopsy is a $25 add-on, and it applies only alongside the Upper Endoscopy & Colonoscopy at $499, because the tissue sample is taken during that procedure. That pairing is priced against US alternatives in our article on colonoscopy costs without insurance.
Planning matters more than price here. The four-week intervals in the table above are a real constraint on a short trip, and they are worth raising with the physician ordering the test before flights are booked.
Most screenings need a few weeks of notice, some need longer, and some can be arranged quickly. Confirm the slot before you book the flight, and if your dates are tight, ask, because arrangements are made case by case. The live booking calendar always shows the true earliest date. For the wider picture on seeing a gastroenterologist here, we have a separate guide.
The bottom line
H. pylori is a Group 1 carcinogen, and eradicating it across a high-prevalence island reduced gastric cancer incidence by a little over half in the best long-term study available. That is a real result and it is not a promise about your stomach.
The mortality benefit has still not reached statistical significance after more than a decade. The United States runs no population screening program for this and its task force has no recommendation, and the groups the guidelines actually name for testing are specific.
If you are in one of those groups, the test that answers the question is a breath test, a stool antigen test, or a biopsy, never a blood antibody test. Take the guideline list to your own physician and decide there.
Sources
- Gut, Chiang TH et al. 2021 — Mass eradication of Helicobacter pylori to reduce gastric cancer incidence and mortality: a long-term cohort study on Matsu Islands (accessed 2026-09-09)
- PubMed Central — Full text of the 2021 Matsu Islands cohort study (accessed 2026-09-09)
- Gut, Lee YC et al. 2013 — The benefit of mass eradication of Helicobacter pylori infection: a community-based study of gastric cancer prevention (accessed 2026-09-09)
- IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 100B — Helicobacter pylori (accessed 2026-09-09)
- Clinical Gastroenterology and Hepatology — Houston Consensus Conference on Testing for Helicobacter pylori Infection in the United States (accessed 2026-09-09)
- American Journal of Gastroenterology — ACG Clinical Guideline: Treatment of Helicobacter pylori Infection, 2024 (accessed 2026-09-09)
- American Journal of Gastroenterology — ACG Clinical Guideline: Diagnosis and Management of Gastric Premalignant Conditions, 2025 (accessed 2026-09-09)
- Helicobacter, Chen MJ et al. 2022 — Declining trends of prevalence of Helicobacter pylori infection and incidence of gastric cancer in Taiwan (accessed 2026-09-09)
- National Cancer Institute SEER — Cancer Stat Facts: Stomach Cancer (accessed 2026-09-09)
- US Preventive Services Task Force — Inactive: Helicobacter Pylori Infection: Screening (accessed 2026-09-09)
