Clinically relevant anxiety and risk of Alzheimer's disease in an elderly community sample: 4.5 years of follow-up.

OBJECTIVES: To investigate whether clinically relevant anxiety increased the risk for developing Alzheimer's disease (AD) while controlling for the presence of depression and other confounders; and to report the population attributable fraction (PAF) associated with anxiety disorder. METHOD: We used data from the longitudinal, community-based Zaragoza Dementia and Depression (ZARADEMP) study. A random sample of 4057 dementia-free community dwellers aged ≥55 years were followed for 4.5 years. The Geriatric Mental State-Automated Geriatric Examination for Computer Assisted Taxonomy package was used for the diagnosis of clinically significant cases and subcases of anxiety; and AD was diagnosed by a panel of research psychiatrists according to DSM-IV criteria. Multivariate survival analysis with competing risk regression model was performed. RESULTS: We observed a significant association between anxiety cases at baseline and AD risk in the univariate analysis that persisted in the fully adjusted model (SHR: 3.90; 95% CI: 1.59-9.60; p = 0.003), with a PAF for AD of 6.11% (95% CI: 1.30%-16.17%). No significant association between 'subcases' of anxiety at baseline and AD risk was found. LIMITATIONS: Data on apolipoprotein E were not available. The hospital-based diagnosis was not completed in all cases of dementia. CONCLUSION: Late-life, clinically significant anxiety (but not subclinical anxiety) seems to increase the risk of AD, independently of the effect of several confounders, including depression. Taking into account the high prevalence of anxiety among the elderly, future studies are warranted to determine potential risk reduction of AD.