Metformin-associated diarrhea in FAERS: a structured pharmacovigilance study with long-term temporal consistency assessment.

BACKGROUND

Metformin-associated diarrhea is well recognized, but spontaneous-report analyses remain sensitive to outcome definition, reporting background, and incomplete event-date data. This study evaluated whether the reporting imbalance persisted across prespecified definitions, comparator sets, and calendar periods.

METHODS

U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) reports from the first quarter of 2020 through the first quarter of 2026 were deduplicated and analyzed with metformin coded as a primary or secondary suspect drug. The primary outcome was the Medical Dictionary for Regulatory Activities preferred term DIARRHOEA. Sensitivity analyses used a broad loose-stool phenotype, a clean gastrointestinal event composite, alternative exposure roles, active comparators, co-medication exclusions, and stratification. Exact-date report-level event-date intervals among reports containing the strict-diarrhea preferred term were summarized descriptively. A separately processed dataset covering the first quarter of 2004 through the first quarter of 2026 assessed temporal consistency.

RESULTS

Among 9,112,609 reports, 60,452 involved metformin as a primary or secondary suspect drug and 4,868 included strict diarrhea. The primary reporting odds ratio (ROR) was 2.72 (95% confidence interval, 2.64-2.80). Restriction to non-metformin antidiabetic reports attenuated the ROR to 1.95 (1.89-2.01), showing sensitivity to the reporting background. Only 595 reports (12.2%) yielded a valid report-level event-date interval; the median interval was 14 days (interquartile range, 1-363). The long-term ROR was 2.76 (2.71-2.81), with marked annual variation.

CONCLUSION

Metformin showed a persistent but context-dependent diarrhea reporting imbalance in FAERS. Active-comparator attenuation limits causal and clinical-risk interpretation. The selected interval subset supports descriptive reporting only and cannot define diarrhea-specific onset or a clinical monitoring window.