Background and purpose: Atrial fibrillation (AF) has been associated with cognitive decline independent of stroke, suggesting additional effects of AF on the brain. We aimed to assess the association between AF and brain function and structure in a general elderly population.
Methods: This is a cross-sectional analysis of 4251 nondemented participants (mean age, 76 ± 5 years) in the population-based Age, Gene/Environment Susceptibility-Reykjavik Study. Medical record data were collected for the presence, subtype, and time from first diagnosis of AF; 330 participants had AF. Brain volume measurements, adjusted for intracranial volume, and presence of cerebral infarcts were determined with magnetic resonance imaging. Memory, speed of processing, and executive function composites were calculated from a cognitive test battery. In a multivariable linear regression model, adjustments were made for demographic factors, cardiovascular risk factors, and cerebral infarcts.
Results: Participants with AF had lower total brain volume compared with those without AF (P<0.001). The association was stronger with persistent/permanent than paroxysmal AF and with increased time from the first diagnosis of the disease. Of the brain tissue volumes, AF was associated with lower volume of gray and white matter hyperintensities (P<0.001 and P = 0.008, respectively), but not of white matter hyperintensities (P = 0.49). Participants with AF scored lower on tests of memory.
Conclusions: AF is associated with smaller brain volume, and the association is stronger with increasing burden of the arrhythmia. These findings suggest that AF has a cumulative negative effect on the brain independent of cerebral infarcts.