Spatial Relationship Between Left Peri-Atrial Fat with Underlying Arrhythmogenic Substrate in Patients with Atrial Fibrillation - Innocent Bystander or Culprit?

Abstract Introduction Peri-atrial fat may have pro-inflammatory and pro-fibrotic effects on atrial myocardium and increase the risk of developing atrial fibrillation. Consistent with prior evidence, we have previously shown that peri-atrial fat volume, but not fat attenuation (an imaging biomarker of inflammation), was associated with prevalent atrial fibrillation. Whether this association is secondary to "outside-to-inside" tissue cross talk with local adverse electrical remodelling of adjacent atrial cardiomyocytes is not known. Purpose In this novel proof of concept study, we aimed to identify a spatial relationship between left peri-atrial fat and potentially arrhythmogenic substrate (low voltage areas and slow conduction pathways) in the left atrial myocardium. Methods Contrast-enhanced cardiac computed tomography (CT) imaging and electroanatomic mapping, during coronary sinus pacing, were performed pre-ablation in atrial fibrillation patients (n=27). The left atrium and surrounding peri-atrial fat were segmented and quantified using seg3D2 and CEMRGapp. We performed registration of the electroanatomic map onto the CT-derived left atrial mesh using OpenEP and EP Workbench. Mean fat volume and attenuation around each electroanatomic point (within a 3mm radial region of interest) were correlated with local voltage and conduction velocity. Results A total of 24,839 electroanatomic points were registered onto the 3D CT-derived left atrial mesh, with a median of 935 points (range 673 to 1150) per patient. The average bipolar voltage and conduction velocity in the left atrium was 1.92±1.91 mV and 0.66±0.47 ms-1 respectively. In comparison to areas with little/no surrounding periatrial fat (Quartile 1, ≤0.93 mm3 fat per electroanatomic point), areas with the highest burden of fat (Quartile 4, ≥17.6 mm3) had significantly lower bipolar voltage (1.69±1.7 vs 2.09±2.0 mV, p<0.001) and conduction velocity (0.627±0.43 vs 0.700±0.47 ms-1, p<0.001). However, the overall correlation between periatrial fat volume with bipolar voltage (r=-0.06, p<0.001) and conduction velocity (r=-0.07, p<0.001) was weak. High, not low, fat attenuation (Quartile 4; >-46.5HU vs Quartile 1; ≤-75.5HU) was associated with increased bipolar voltage (2.23±2.1 vs 1.63±1.6 mV) and conduction velocity (0.703±0.42 vs 0.623±0.39 ms-1 respectively). Conclusion There is a weak correlation between periatrial fat and localized adverse electrical remodelling. The relationship between fat attenuation and electrophysiological properties was inverted. These findings confirm an association between peri-atrial fat burden but call question as to whether peri-atrial fat inflammation is implicated in atrial fibrillation pathogenesis.