ABSTRACT
BACKGROUND: Atrial fibrillation increases risk of stroke, and thus risk of cognitive impairment and dementia. Emerging evidence suggests an association also in the absence of stroke. We aimed to examine the association between atrial fibrillation and incident dementia, with and without exclusion of individuals with stroke, and if sex and genetic factors modify the possible association. METHODS: In 2000-2001, a population-based sample of 70-year-olds (N = 561) underwent comprehensive somatic and neuropsychiatric examinations, as part of the Gothenburg H70 Birth Cohort Studies. Participants were followed up at age 75 and 79. Atrial fibrillation at baseline was identified through ECG, proxy-reports and the National Patient Register (NPR). Stroke at baseline and follow-up was identified through self-reports, proxy-reports and the NPR. Dementia at baseline and follow-up was diagnosed according to the DSM-III-R criteria based on neuropsychiatric examinations, proxy-reports and the NPR. RESULTS: Individuals with atrial fibrillation had an almost threefold increased risk of dementia during 12-year follow-up (HR 2.8; 95% CI 1.3-5.7; P = 0.004), and this risk remained after excluding individuals with stroke at baseline and follow-up. After stratification for sex, the association was only found amongst men (HR 4.6; 95% CI 1.9-11.2; P < 0.001, interaction sex*atrial fibrillation; P = 0.047) and noncarriers of the APOE ε4 allele (HR 4.2; 95% CI 1.8-9.7; P < 0.001, interaction APOE*atrial fibrillation; P = 0.128). Population attributable risk for dementia resulting from atrial fibrillation was 13%. CONCLUSION: The relevance for atrial fibrillation as an indicator of subclinical brain vascular risk needs to be further explored. In addition, patients with atrial fibrillation should be screened for cognitive symptoms.
