S1608 The Risk of Gastric Cancer (Overall and Cardia Only) With the Use of Proton Pump Inhibitors in Observational Studies: A Systematic Review and Meta-Analysis
Bibliographic record
Abstract
Introduction: There are a number of reports that raise safety concerns on the association between proton pump inhibitors (PPIs) and the risk of gastric cancer both overall as well as isolated gastric cardia. Methods: We performed an updated systematic review and meta-analysis of observational studies published through January 2022 that examined the association between > 6 months PPI use and the risk of gastric cancer overall and gastric cardia cancer specifically. The inclusion criteria were classification as either a case-control or cohort study, participants 18 years or older, long-term ( > 6 months) PPI use, and gastric cancer as a specified outcome. Excluded studies were case reports, interventional studies, conducted in the pediatric population and used animal models. Study quality was evaluated using the Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool. Pooled adjusted risk estimates were calculated using random effects model, and heterogeneity among studies was examined as I2 statistic. Results: We identified 15 publications with 16 studies fulfilling inclusion and exclusion criteria: 6 reported nested case-control studies, 1 case-control, 1 prospective cohort, 6 retrospective cohort, and 1 paper with both case-control and prospective cohort studies. Only 8 studies accounted for Helicobacter pylori. Most studies (5 of 8) that examined gastric cardia risks reported no significant association. Pooled unadjusted odds ratios from case control studies were 1.73 (95% CI: 1.43, 1.85) from the 5 overall gastric studies, and 1.26 (95% CI: 0.58, 2.72) from the 6 gastric cardia studies. Most (12 of 15) reported a statistically significant adjusted risk estimate of the association of PPI usage and overall gastric cancer risk, and 3 studies reported no association. Pooled unadjusted risk ratios from cohort studies were 2.57 (95% CI: 2.01, 3.28) from the 8 overall gastric cancer studies, and 1.61 (95% CI: 0.93-2.78) from the 3 gastric cardia studies. The I2 ranged between 23 and 88%. Most (15 of 16) had at least moderate risk of bias on the ROBINS-I tool with 1 study with a serious risk of bias (Table). Conclusion: Observational studies show inconsistent and mostly non-significant (individual or pooled) association between PPI and gastric cardia. For overall gastric, pooled results show a modest statistical association, but considerable heterogeneity among studies, study design limitations and at least moderate risk of bias do not support an etiological relationship between PPI and gastric cancer. Table 1. - Fifteen observational studies that examined the association between PPIs and risk of gastric cancer (GC) Paper (Year of Publication) Area Study design Gastric cancer (GC) site Patient source Adjusted risk estimate (95% CI) ROBINS-I Score Garcia Rodriguez et. al (2006) United Kingdom Nested case-control Overall (any gastric site)Gastric cardia General Practitioners Research Database Adjusted OR for Cardia: 1.06 (CI 0.57-2.00) Moderate risk Duan et. al (2008) United States Case-control Overall (any gastric site)Gastric cardia University of Southern California Cancer Surveillance Program Adjusted OR for Cardia with PPI use for < 1year: 0.65 (0.31-1.37)1-3 years: 0.70 (CI 0.32-1.49)> 3 years: 1.29 (CI 0.70-2.36) Moderate risk Tamim et. al (2008) Canada Nested case-control Overall (any gastric site)Gastric cardia Quebec health insurance plan databases Adjusted OR for overall GC: 1.46 (CI 1.22-1.74)Cardia: 0.58 (CI 0.26-1.32) Moderate risk Poulsen et. al (2009) Denmark Prospective cohort Overall (any gastric site) Danish Civil Registration System Adjusted IRR for overall GC: 1.2 (CI 0.8-2.0) Moderate risk Wennerstrom et. al (2017) Netherlands Nested case-control Overall (any gastric site)Gastric cardia Danish Civil Registration System Adjusted HR for Cardia: 2.51 (CI 2.26-2.79) Moderate risk Brusselaers et. al (2017) Sweden Retrospective cohort Overall (any gastric site)Gastric cardia The Swedish Cancer Registry Adjusted SIR for overall GC: 3.38 (CI 3.25-3.53)Cardia: 3.55 (CI 3.27-3.86) Moderate risk Lai et. al (2018) Taiwan Case-control Overall (any gastric site) Taiwan National Health Insurance Program Adjusted OR for overall GC: 2.00 (CI 1.36-2.95) Moderate risk Peng et. al (2018) Taiwan Nested case-control Overall (any gastric site)Gastric cardia National Health Insurance Research Database; Registry of Catastrophic Illness Adjusted OR for overall GC: 2.48 (CI 1.38-4.38)Cardia: 2.58 (CI 1.38-4.83) Moderate risk Cheung et. al (2018) Hong Kong Retrospective cohort Overall (any gastric site)Gastric cardia Clinical Data Analysis and Reporting System of Hong Kong Hospital Authority Adjusted HR for overall GC: 2.19 (CI 1.31-3.66)Cardia: 1.24 (CI 0.35-4.34) Moderate risk Niikura et. al (2018) Japan Retrospective cohort Overall (any gastric site) University Tokyo Hospital database Adjusted HR for overall GC: 3.61 (CI 1.49-8.77) Serious risk Brusselaers et. al (2019) Sweden Retrospective cohort Overall (any gastric site) The Swedish Cancer Registry Adjusted SIR for overall GC: 2.97 (CI 2.83-3.10) Moderate risk Lee et. al (2020) United States Nested case-control Overall (any gastric site)Gastric cardia Kaiser Permanente Northern California Adjusted OR for overall GC: 1.07 (CI 0.81-1.42)Cardia: 0.98 (CI 0.68-1.40) Moderate risk Liu et. al (2020) United Kingdom Nested case-control Overall (any gastric site) Primary Care Clinical Information Unit Adjusted OR for overall 1.49 (CI 1.24-1.80) Moderate risk Liu et. al (2020) United Kingdom Prospective cohort Overall (any gastric site) UK BioBank Adjusted HR for overall GC: 1.28 (CI 0.86-1.90)Cardia: 0.81 (CI 0.40-1.64) Moderate risk Ng et. al (2021) Hong Kong Retrospective cohort Overall (any gastric site) PCI Registry and Clinical Data and Analysis Reporting System Adjusted HR for overall GC: 2.06 (CI 1.01-4.18) Low risk Seo et. al (2021) South Korea Retrospective cohort Overall (any gastric site) National Health Insurance Service-National Sample Cohort (old version); Observational Medical Outcomes Partnership-Common Data Model Adjusted HR for overall GC: 2.37 (CI 1.56-3.68) Moderate risk
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.023 | 0.052 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.015 | 0.035 |
| Bibliometrics | 0.010 | 0.010 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.003 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".