Peak frequency can be effectively used to characterize atrial remodelling in atrial fibrillation

Abstract

Abstract Introduction Characterizing atrial fibrillation (AF) substrate can guide ablation strategies, particularly as current efficacy leaves room for substantial improvement. Low voltage zones (LVZs) correlate with atrial fibrosis and scar and can independently predict AF recurrence. Ablation of these zones improves freedom from arrhythmia, but the chaotic nature of atrial excitation frequently results in overestimation of LVZs limiting accurate substrate identification. Purpose While omnipolar voltage (OV) mapping has emerged to address this issue by closely correlating with sinus rhythm (SR) voltage maps, our study builds upon this by investigating a novel parameter, peak frequency (PF), to further refine AF substrate characterization. Methods Patients undergoing catheter ablation for persistent AF were enrolled. OV maps and PF maps were created during AF while bipolar voltage (BV) maps were created in SR with atrial pacing at pacing intervals (PIs) of 600ms and 250ms. Areas were classified as fixed scar (consistent LVZs across all maps), functional remodelling (LVZs in AF OV and SR BV 250ms maps) and non-LVZs (normal voltage across all maps). PF was assessed in these areas and defined as the highest frequency detected in the electrogram. Results Among 20 patients, the average voltage in AF OV maps differed significantly from SR BV 600ms (0.49±0.76mV in AF OV vs. 1.12±0.97mV SR BV 600ms, p<0.001) but not SR BV 250ms maps (0.49±0.76mV in AF OV vs. 0.52±0.84mV SR BV 250ms, p=0.10). Comparing co-registered points, there was a significant difference between AF OV and SR BV 600ms maps (0.21±0.09mV, p<0.001) but not between AF OV and SR BV 250ms maps (0.06±0.04mV, p=0.10). These findings were also consistent with regards to the proportion of LA area occupied by LVZs (42.4±8.2% AF OV, 24.3±7.4% SR BV 600ms and 39.4±6.7% SR BV 250ms; p=0.001). AF OV maps thereby correlated better with SR BV 250ms maps hence representing both fixed scar and functional remodelling. PF values ≥244Hz were predictive of non-LVZs (OR 3.91, 95% CI 3.77-4.07,) with an AUC of 0.71. A PF of ≤214Hz was predictive of fixed scar (OR 17.67, 95% CI 16.82-18.56, p<0.001) with an AUC of 0.90. A PF between 215-236Hz was predictive of functional remodelling (OR 2.83, 95% CI 2.71-2.95, p<0.001) with an AUC of 0.60. The majority of LVZs identified exclusively in AF OV maps exhibited PFs akin to those in non-LVZs highlighting the ability of PF to pinpoint potential overestimations in LVZs. Conclusions PF mapping can effectively discern between areas of fixed scar, functional remodelling, and potential overestimation of LVZs. PF may enhance substrate characterization in AF utilising data solely from electrograms.Trial Design and Zone AssignmentPeak Frequency and Voltage Zones