(A-112) Direct Oral Anticoagulants and Incidence of Dementia in Patients with Nonvalvular Atrial Fibrillation

Background: In patients with atrial fibrillation (AF), oral anticoagulation is essential for stroke prevention but may also prevent cognitive decline and dementia. Moderate evidence suggests that direct oral anticoagulants (DOACs) may reduce the risk of dementia compared to warfarin. However, it remains unclear whether DOACs have similar effectiveness against dementia.

Objectives: The primary objective was to compare the risk of dementia among AF patients initiating different DOACs, with head-to-head comparisons between apixaban, edoxaban, and rivaroxaban.

Methods: Using data from the Clinical Practice Research Datalink (CPRD), a large UK primary care database, we formed three DOAC new-user cohorts among patients aged ≥50 years newly diagnosed with AF. The first two cohorts compared edoxaban initiators to apixaban and rivaroxaban initiators, respectively, between 2015 (the year of edoxaban approval in the UK) and 2023. The third cohort compared apixaban and rivaroxaban initiators between 2013 (since apixaban was approved at the end of 2012) and 2023. Patients were considered exposed starting six months after their first DOAC prescription for latency considerations and censored upon switching to another oral anticoagulant or treatment discontinuation (with six-month grace period). The primary outcome was the first recorded dementia diagnosis during follow-up. Propensity score fine stratification and weighting was used to control for confounding. We fit Cox proportional hazards models to estimate weighted hazard ratios (HR) of dementia for each head-to head DOAC comparison. In a sensitivity analysis, patients were considered exposed to their initial DOAC until the end of follow-up, regardless of treatment termination or switching, to assess potential effects beyond treatment termination, analogous to an intention-to-treat (ITT) definition.

Results: The three cohorts included up to 131,663 users of apixaban, 48,405 users of edoxaban and 70,038 users of rivaroxaban with AF. DOACs were associated with similar risks of dementia, across the three comparative groups: edoxaban vs. rivaroxaban (HR, 1.05; 95% CI, 0.96-1.16), edoxaban vs. apixaban (HR, 1.06; 95% CI, 0.98-1.15), and apixaban vs. rivaroxaban (HR, 1.01; 95% CI, 0.96-1.06). Results from the sensitivity analysis using an ITT approach were generally consistent with those of the primary analysis.

Conclusions: In patients with AF, no meaningful differences in dementia risk were observed across DOACs in the primary analyses, suggesting that the choice of DOAC does not influence dementia risk.