Drug–Drug Interaction Signals for Apixaban‐Associated Haemorrhage in Atrial Fibrillation and Thromboembolic Disorders: A Disproportionality Analysis of the FAERS Database

Background Apixaban is commonly used in anticoagulation therapy. Its safety profile is associated with polypharmacy and potential drug–drug interactions (DDIs). Current understanding often focuses on canonical metabolic pathways, while other biological factors of haemorrhagic risk remain not fully described. Objectives This study investigated DDI signal drugs associated with haemorrhagic adverse events when apixaban is used with concomitant drugs. The study also explored potential pharmacological mechanisms. Methods A disproportionality and DDI signal analysis was performed on FDA Adverse Event Reporting System (FAERS) data from the first quarter of 2013 to the first quarter of 2025. DDI signals were identified using six different algorithms. A DDI signal met the analysis criteria if it was detected by all six algorithms including the Ω shrinkage method and the chi‐square test. Concomitant drugs associated with DDI signal drugs were classified using the DrugBank database to investigate pharmacological properties with a focus on metabolic enzymes and transporters. Results After the exclusion of established interactions, the analysis identified 107 medications and 49 preferred terms. Analysis at the level of enzymes and transporters indicated that noncanonical pathways exhibited risk intensities, including arylsulfatase A (ARSA, weighted mean Ω = 3.60), macrophage metalloelastase (MMP12, weighted mean Ω = 3.16), organic anion transporting polypeptide 2B1 (OATP2B1, weighted mean Ω = 2.82) and bile salt export pump (BSEP, weighted mean Ω = 2.62). DDI signal drugs associated with high‐intensity targets included acyclovir for duodenal ulcer haemorrhage (Ω = 2.90, 95% CI: 2.12–3.69), mesalazine for an increased tendency to bruise (Ω = 2.96, 95% CI: 1.69–4.22) and digitoxin for haematoma (Ω = 2.65, 95% CI: 1.83–3.46). Interaction signals were also recorded for morphine with injection site haematoma (Ω = 3.05, 95% CI: 1.90–4.21) and eltrombopag with internal haemorrhage (Ω = 3.11, 95% CI: 2.04–4.18). Conclusion Haemorrhagic risks in apixaban therapy are likely associated with a convergence of transport interference, potential impairment of haemostasis and indirect clinical factors, such as medication‐induced instability. These findings suggest that safety management may extend beyond canonical pathways to consider the possible impact of alternative transporters and functional stability. These results provide a preliminary basis for further clinical validation.