Case-control studyindependent cross-checkfull text checked★ key study I-PPI-107 in-depth review
PPI-cancer link largely reflects early cancer symptoms
Association between proton pump inhibitor use and upper gastrointestinal cancer: A matched case-control study accounting for reverse causation and confounding by indication.
Overview
Sawaid IO et al.. PLoS Med 2026 · PMID 41493925 · DOI 10.1371/journal.pmed.1004842
People often start acid blockers because of symptoms of a still-undiagnosed cancer; after accounting for this, long-term PPI use was not linked to more upper GI cancer.
Does the study support the specific claim? No · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Re-examine PPIs and upper GI cancer.
- How was the study done?
Matched case-control, 875 cases.
- How many people, animals or samples were studied?
Israeli health records, incl. Arab citizens.
- What was actually found?
Strong recent-use link vanished for earlier use.
- How large was the observed effect?
aOR 4.01 recent esomeprazole; 0.62 remote omeprazole.
- What do the results mean in everyday life?
Alarm symptoms should be investigated, not blamed on the medicine.
- What are the study's limitations?
Observational; OTC use missed.
- What does it say about Elawady's specific claim?
Contradicts 'PPIs cause cancer'.
- Do the results apply to Germany and Europe?
Moderately applicable.
- What can explicitly not be concluded from this study?
Cannot exclude effects after >10 years.
Scientific in-depth review
1. Full reference
- Authors
- Sawaid IO et al.
- Journal
- PLoS Med
- Year
- 2026
- Identifier
- PMID 41493925 DOI 10.1371/journal.pmed.1004842
- Verification
- verified – PubMed efetch 2026-10-26 + Crossref 2026-10-26
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Internally funded by Leumit Health Services; one author supported by NIH intramural programme; no competing interests (full text)
2. Research question
Is PPI use associated with upper GI cancer after accounting for reverse causation and confounding by indication?
3. Study design and population
- Study design
- Case-control study – Matched case-control study (EHR, Israel) with time-window exposure modelling
- Population
Members of Leumit Health Services, Israel, incl. Arab sector (12% of cases)
- Country
Israel
- Period
Diagnoses 2003-2024
- Sample size
875 cases, 8750 controls
- Inclusion and exclusion criteria
Incident oesophageal, gastric or duodenal cancer; controls cancer-free, matched 1:10 on age, sex, ethnic sector, SES, enrolment year
4. Methods and endpoints
- Exposure / intervention
PPI purchases in windows 0-6 months, 6-12 months, 1-3 years, 3-10 years before diagnosis
- Comparator
No PPI in window
- Endpoints
Upper GI cancer
- Statistical methods
Conditional logistic regression adjusted for smoking, BMI, alcohol, H. pylori, healthcare use and symptom diagnoses
5. Actual results
- Results
Without adjustment for symptom diagnoses: esomeprazole aOR 4.01 (3.20-5.03), omeprazole aOR 2.38 (1.99-2.85). Associations diminished for exposure >1 year before diagnosis; after excluding the final year and adjusting for symptom diagnoses no harmful association; remote use (>3 y) omeprazole aOR 0.62 (0.51-0.75); similar in gastric-only subgroup (701 cases).
- Absolute risks
Not reported
- Authors' conclusion
Apparent harmful associations were concentrated in months before diagnosis and disappeared after accounting for diagnostic context.
6. Strengths
- Explicit handling of protopathic bias
- Arab population included
- Code publicly available
- Full text read
7. Weaknesses
- Observational; residual confounding
- Over-the-counter PPI not captured
- About 10 years of purchase data only
8. Risk of bias
- Orienting tool
- ROBINS-E
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Better bias control than earlier studies; residual confounding possible.
9. Assessment of the statistical interpretation
Pattern (high recent OR, null or protective remote OR) is typical of protopathic bias (stats/C-PPI-09_cancer.py). The 'protective' remote association should not be read as causal either.
10. Relevance to the specific claim
Shows how cancer-PPI associations arise from early cancer symptoms being treated with PPIs; contradicts PPIs as a cause of cancer.
11. Independent contrary findings
Abrahami 2022 found HR 1.45 vs H2RA after 5-year median follow-up (I-PPI-108).
12. Transferability to Germany and Europe
Moderate.
13. Overall methodological judgement
Important recent counterweight to the gastric cancer signal.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Effects of >10-year use.