Aspirin and Risk of Dementia in Patients with Late-Onset Depression: A Population-Based Cohort Study

BACKGROUND: Late onset depression (LOD) often occurs in the context of vascular disease and may be associated with risk of dementia. Aspirin is widely used to reduce the risk of cardiovascular disease and stroke. However, its role in patients with LOD and risk of dementia remains inconclusive. Materials and Methods. A population-based study was conducted using data from National Health Insurance of Taiwan during 1996–2009. Patients fulfil diagnostic criteria for LOD with or without subsequent dementia (incident dementia) and among whom users of aspirin (75 mg daily for at least 6 months) were identified. The time-dependent Cox proportional hazards model was applied for multivariate analyses. Propensity scores with the one-to-one nearest-neighbor matching model were used to select matching patients. Cumulative incidence of incident dementia after diagnosis of LOD was calculated by Kaplan–Meier Method. RESULTS: A total of 6028 (13.4%) and 40,411 (86.6%) patients were defined as, with and without diagnosis of LOD, among whom 2,424 (41.9%) were aspirin users. Patients with LOD had more comorbidities such as cardiovascular diseases, diabetes, and hypertension comparing to those without LOD. Among patients with LOD, aspirin users had lower incidence of subsequent incident dementia than non-users (Hazard Ratio = 0.734, 95% CI 0.641–0.841, p < 0.001). After matching aspirin users with non-users by propensity scores-matching method, the cumulative incidence of incident dementia was significantly lower in aspirin users of LOD patients (p < 0.001). After matching aspirin users with non-users by propensity scores-matching method, the cumulative incidence of incident dementia was significantly lower in aspirin users of LOD patients ( CONCLUSIONS: Aspirin may be associated with a lower risk of incident dementia in patients with LOD. This beneficial effect of aspirin in LOD patients needs validation in prospective clinical trials and our results should be interpreted with caution.