Recent evidence suggests that patients with atrial fibrillation are at increased risk for cognitive dysfunction and dementia, even those who don't have a clinical history of stroke. In a multicenter study in Switzerland, researchers assessed the relationship between cognitive function and clinically known and unknown (silent) vascular brain lesions in a large sample of patients with atrial fibrillation.

A total of 1,737 patients (mean age, 73 years) who had brain magnetic resonance imaging (MRI) at baseline were included in the analysis. Ninety percent were taking anticoagulants at baseline. Cognitive function was assessed using a standardized test, on which higher scores indicated better cognitive function.

At least one large noncortical or cortical infarct was found on MRI in 22% of patients, small noncortical infarcts were observed in 21%, microbleeds in 22%, and white matter lesions in 99%. When patients with a history of stroke or transient ischemic attack were excluded, 15% and 18% of patients still had evidence of silent large noncortical or cortical infarct and silent small noncortical infarct, respectively.

The cognitive function score was lower in patients with a large noncortical or cortical infarct than in those without such an infarct. In a multivariable model including all vascular brain lesion parameters, large noncortical or cortical infarct count and volume were the strongest predictors of cognitive function scores. In this model, small noncortical infarct volume, white matter lesion volume, and microbleed count weren't independently associated with cognitive function. Anticoagulation wasn't related to a higher microbleed count.

The authors advise that future research develop risk scores for patients with atrial fibrillation to determine which patients may benefit from brain imaging to guide antithrombotic treatment. Because of the cross-sectional study design, the authors note, it wasn't possible to determine causality or directionality of effect. Also, most of the participants were white and had health insurance, so the applicability of these results to other populations is unknown.