Right Hemispheric Predominance of Brain Infarcts in Atrial Fibrillation: A Lesion Mapping Analysis.

Dear Sir:

Atrial fibrillation (AF) increases the risk of an ischemic stroke three- to five-fold [1], and cardioembolic strokes are associated with worse outcomes than strokes of other origin [2]. Covert brain infarcts occur in more than a third of patients with AF and have been associated with cognitive impairment [3]. Regional distribution of brain infarcts may explain why most patients with AF and brain infarcts on neuroimaging have no clinical history of stroke or transient ischemic attack (TIA). Previous studies on a possible side-predominance of brain infarction in patients with AF have yielded conflicting evidence and were mostly done in patients with clinically manifest stroke [4]. In the present study, we systematically analyzed the anatomical distribution and localization of covert and overt brain infarcts according to different vascular territories in a large cohort study of patients with established AF.

This is a cross-sectional analysis of the Swiss-AF cohort, a prospective, multicenter cohort study of unselected patients with AF across 14 centers in Switzerland [5]. In total, the study has enrolled 2,415 participants with AF. At baseline visits, brain magnetic resonance imaging (MRI) was acquired in all patients (without contraindications for MRI) on either a 1.5 or a 3.0 Tesla scanner, depending on the participating site. The most frequent reason for missing MRI was the presence of an implanted cardiac device (n=461). The 1,716 patients with eligible baseline brain MRI data (mean age 73±8 years, 27% women, 20% with history of stroke or TIA) were included in the present study (Supplementary Table 1). Detailed methodological descriptions are provided in the Supplementary methods [3,5-11]. Briefly, we manually segmented brain infarcts on MRI. Lesions were classified into large non-cortical or cortical infarcts (LNCCIs) and small non-cortical infarcts (SNCIs) [7]. LNCCIs were considered infarcts of potentially embolic or atherothrombotic origin and included large non-cortical infarcts with a diameter of >20 mm, and cortical infarcts of any size. SNCIs were considered infarcts of potentially microangiopathic origin and defined as lesions with a diameter of ≤20 mm not involving the cortex. Lesion masks were co-registered separately to an age-specific brain template and overlaid to compute a voxel-based probability map.

We observed LNCCIs in 386 (22%) patients. Visually, in the voxel-based probability map, a clear right-hemispheric preponderance was observed (Figure 1A). Results of the statistical comparison of lesion counts, average lesion volumes, and percentages affected per vascular territory are reported in Supplementary Table 2. LNCCIs were most often located in the territory of the middle cerebral artery (MCA). While, in the MCA territory, LNCCI counts were evenly distributed between the hemispheres (Wilcoxon matched-pair rank test; P=0.796), the average volume of infarcts was significantly larger in the right compared to the left hemisphere (left [total]: 512.4 mL vs. right [total]: 1,240.1 mL; Wilcoxon matched-pair rank test; P=0.002). Similarly, the percentage of the MCA territory affected by LNCCIs (calculated by the proportion of lesional voxels per vascular territory per patient) was significantly larger in the right than in the left hemisphere (Wilcoxon matched-pair rank test; P=0.008). No side-preponderance of LNCCIs was detected for the vascular territories of the anterior and posterior cerebral arteries. No side-difference was observed for SNCIs (Figure 1B).

Figure 1.

The distribution of infarcts. Figure depicting the voxel-based sum of all large cortical and non-cortical infarcts and small non-cortical infarcts overlaid on an age-specific standard brain template. The color scale indicates the number of patients with an infarct in a given voxel.

Our findings might indicate, that large emboli preferentially hit the right brain hemisphere. We hypothesize that both vascular anatomy (i.e., the right artery originating more proximally from the heart) and the physical properties of cardiac emboli (i.e. their size) contribute to the right-sided preponderance for embolic cerebral infarcts. The right common carotid artery originates more proximally from the heart and with less angulation from the aortic arch, which in turn increases the likelihood of these large emboli to enter the right carotid artery. This may especially be the case for larger cardiac emboli which are dragged to the outer curvature of the aorta. This hypothesis is supported by a study using anatomically accurate models of the human aorta and their branch arteries [12].

As no information on competing stroke etiologies was collected in SWISS-AF, we cannot know how many LNCCIs might have been of potential atheroembolic origin, which represents a major limitation of our study. However, it is unlikely that the observed asymmetry in LNCCIs is explained by atherosclerosis of the aortic arch or supra-aortic arteries.

A better understanding of the localization in relation to the size of brain infarcts might be helpful in clinical routine. Focal symptoms and signs originating from right hemispheric lesions may easily be missed by patients and physicians. The proportion of patients with a history of clinically manifest stroke or TIA was consistent with previous population-based studies in AF [1], but 37,1% of patients had an ischemic infarct on MRI (SNCI or LNCCI combined). We have previously shown that covert brain infarcts have a significant impact on cognitive decline in AF [3], and are associated with an increased risk of future strokes [13]. Whether routine brain MRI screening in patients with AF improves risk stratification remains to be elucidated by future prospective studies.