Disproportionality analysis of adverse event reporting with aspirin, statins, and their co-exposure in FAERS, 2004-2025.

OBJECTIVE

To characterize and compare adverse-event reporting patterns for aspirin without statins, statins without aspirin, and aspirin-statin co-exposure in the FDA Adverse Event Reporting System (FAERS), and to explore statistical reporting-interaction signals.

METHODS

We analyzed FAERS reports from 2004 Q1 through 2025 Q1. Active-comparator disproportionality analyses compared aspirin-statin co-exposure with aspirin-without-statin and statin-without-aspirin reports using the reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian confidence propagation neural network (BCPNN), and empirical Bayesian geometric mean (EBGM), followed by Benjamini-Hochberg false discovery rate (BH-FDR) correction. Candidate preferred terms (PTs) common to both active-comparator analyses underwent adjusted statistical interaction analyses using four exposure groups, including a neither-drug reference. Primary models retained missing age, sex, and reporter type as explicit Unknown categories, and a complete-case sensitivity analysis excluded reports with missing adjustment covariates.

RESULTS

After deduplication, 18,932,599 unique FAERS reports were retained. The active-comparator analysis included 134,479 aspirin-without-statin reports, 202,565 statin-without-aspirin reports, and 70,806 aspirin-statin co-exposure reports. Among 9,332 and 9,736 PTs tested in the two active-comparator analyses, 119 and 254 PTs remained after prespecified signal-detection criteria and BH-FDR correction, respectively; 42 PTs were common to both comparisons. Adjusted interaction models were reliably estimable for 40 candidates, and 31 PTs met both additive and multiplicative reporting-interaction criteria. Missingness was substantial for age (42.33%), sex (12.12%), and reporter type (4.35%). The complete-case dataset retained 53.83% of reports, and 21 of the 31 primary both-scale-positive PTs remained jointly positive.

CONCLUSIONS

Aspirin-statin co-exposure showed distinct adverse-event reporting patterns in FAERS. The identified signals are exploratory statistical reporting interactions and do not establish increased incidence, clinical risk, causality, or pharmacological drug-drug interactions. Only 21 of the 31 primary both-scale-positive PTs remained jointly positive in the complete-case sensitivity analysis, reinforcing the need for cautious interpretation and validation in denominator-based pharmacoepidemiologic datasets.