Metformin and Helicobacter pylori Infection in Patients With Type 2 Diabetes

Patients with diabetes have an increased risk of gastric cancer (1,2), and metformin may protect against gastric cancer (3). This study investigated whether metformin might protect against Helicobacter pylori infection, an important risk factor of gastric cancer.

The reimbursement database of Taiwan’s National Health Insurance was used. Disease diagnoses were based on the ICD-9-CM. Diabetes was coded 250.XX, and H. pylori infection was defined by a diagnostic code of 041.86 plus receiving H. pylori eradication therapy (4). A propensity score (PS)-matched cohort (based on the Greedy 8→1 digit match algorithm) (3) comprised 16,060 metformin ever users and 16,060 metformin never users with new-onset diabetes, who were enrolled during 1999–2005 and followed up until the end of 2011. Patients with type 1 diabetes, cancer diagnoses, or H. pylori infections at entry or within 6 months of diabetes diagnosis, aged <25 or >75 years at entry, or with a follow-up duration <6 months were not included. PS was derived from covariates including age, sex, occupation, living region, hypertension, dyslipidemia, obesity, nephropathy, eye disease, stroke, ischemic heart disease, peripheral arterial disease, chronic obstructive pulmonary disease, tobacco abuse, alcohol-related diagnoses, heart failure, gingival and periodontal diseases, pneumonia, osteoporosis, tuberculosis infection, disease of pancreas, diseases of esophagus, stomach, and duodenum, Epstein Barr virus infection, hepatitis B virus infection, hepatitis C virus infection, and use of insulin, sulfonylureas, meglitinides, acarbose, rosiglitazone, pioglitazone, ACE inhibitors/angiotensin receptor blockers, calcium channel blockers, statins, fibrates, and aspirin. Hazard ratios (HRs) were estimated by Cox regression incorporated with the inverse probability of treatment weighting using PS (3). Standardized difference was calculated for each covariate, and a cutoff >10% may indicate potential confounding (5).

Results showed that the two groups of metformin ever and never users were well matched, with none of the covariates having a standardized difference >10%. The incidence of H. pylori infection and HRs by metformin exposure are shown in Table 1.

The findings suggested that metformin use was associated with a significantly reduced risk of H. pylori infection in a dose-response pattern.

This study has merits of avoidance of selection bias by use of a nationwide database that covers >99% of the Taiwan’s population. The use of medical records reduced potential bias related to self-reporting. Detection bias resulting from varying socioeconomic status could also be significantly reduced because the drug cost sharing is low in this health care system and can always be waived in patients with low income, veterans, and those who received prescription refills for chronic diseases.

It is recognized that the potential effects of unmeasured confounders could not be assessed. These might include biochemistry, immune profiles, education levels, household conditions, nutritional status, dietary patterns, anthropometric factors, lifestyle, smoking, alcohol use, family history, and genetic parameters.

Although the mechanisms remain unknown, the findings gave rationale to conduct clinical trials to investigate such an effect. Because metformin is cheap and safe without a risk of hypoglycemia when used alone, its usefulness for the prevention of H. pylori infection with an ultimate goal to reduce gastric cancer is worthy of in-depth investigation.