Population-based cohort study of modifiable risk factors for dementia and Alzheimer disease mortality in Spain

Abstract

Although many modifiable factors are linked to dementia risk, their association with dementia and Alzheimer disease (AD) mortality remains unclear. We aimed to identify late-life modifiable factors associated with these outcomes and assess sex differences. In the population-based Neurological Disorders in Central Spain (NEDICES) cohort, we followed 3085 older adults enrolled in 1994–1995 through 2017. We examined 13 modifiable factors measured at study entry. Dementia and AD deaths were defined by the underlying cause recorded on death certificates. Fine–Gray competing risk models accounted for death from other causes and generated subdistribution hazard ratio (SHR) estimates and adjusted population attributable fractions (PAFs). During follow-up, there were 182 dementia-coded deaths (including 105 AD-coded deaths) and 2423 deaths from other causes. Here we show that depressive symptoms are associated with higher dementia-coded mortality in women (SHR 1.52, 95% confidence interval [CI] 1.00–2.30; PAF 13.6%, 95% CI 0.0–28.4) and higher AD-coded mortality overall (SHR 1.76, 95% CI 1.11–2.79; PAF 15.4%, 95% CI 2.6–29.9), with a stronger association in women (SHR 2.23, 95% CI 1.28–3.88; PAF 27.2%, 95% CI 7.8–46.7). Several exposures show inverse associations (e.g., smoking, obesity, and arterial hypertension in men), consistent with competing mortality and late-life reverse causation. Combined positive PAFs are 10.8% for dementia-coded mortality and 18.2% for AD-coded mortality overall, reaching 41.1% in women for AD-coded mortality. Late-life depressive symptoms, particularly among women, emerged as the most prominent modifiable correlate of dementia- and AD-coded mortality, supporting improved recognition and management of depressive symptoms in older adults.