Data analysis
The tissue impedance data recorded during RF delivery were processed offline using the R statistical software (version 4.0.3) to remove noise (see supplementary materials section).
Ablation lesions were excluded if any of the following occurred: short duration of less than 5 sec (as no LSI value is generated before 5 sec of RF delivery); sudden and progressive impedance rise, suggestive of catheter tip or tissue overheating with charring formation9.
After processing of the raw data, the maximum absolute impedance drop (Max-Imp-∆), the maximum percentage impedance drop (Max-Imp-%), the mean CF, and the contact force variability (CFV) were calculated for each study ablation lesion. The maximum absolute impedance drop (Max-Imp-∆) was calculated as the difference between maximum impedance value per lesion, which corresponded to the impedance at the beginning of the ablation, and minimum impedance reached during ablation. The maximum percentage impedance drop (Max-Imp-%) was calculated by converting the Max-Imp-∆ into a percentage of the starting impedance to minimise any influence from the initial impedance value. The contact force variability (CFV) was calculated as difference between mean of the peaks and mean of the troughs of CF during ablation, as previously described 10 (Figure 1).
The ablation lesions were grouped by RF power used (20, 30 and 40W), by mean CF used (< 5, 5-10, 10-15 or >15 g) and by CFV (≤ 5g or > 5 g) and the Max-Imp-% values in each group were compared.
The variation of impedance drop over ablation time was investigated by dividing each ablation delivery into consecutive, cumulative 0.5 sec time intervals and calculating the mean and the standard deviation (SD) of all lesions Max-Imp-% at each interval compared to the initial impedance at the start of ablation. The relative plateaus of Max-Imp-% and the corresponding times at plateau for each group of study ablations were identified by finding the point after which the SD of the Max-Imp-% at each time interval became negligible (less than 0.1) and the difference between Max-Imp-% at that point and Max-Imp-% at the end of ablation was very low (less than 0.25).
The relation between impedance drop and FTI was assessed as described in the supplementary section.
The relationship between impedance drop and LSI was assessed by dividing the ablation lesions into consecutive, cumulative 0.1 LSI intervals and calculating the Max-Imp-% of each interval compared to the initial impedance at the start of ablation. The mean and SD of Max-Imp-% were calculated for each interval. The relative plateaus of impedance drop and the corresponding LSI values at plateau were identified by finding the point after which the SD of the Max-Imp-% at each LSI interval became negligible (less than 0.1) and the difference between Max-Imp-% at that point and Max-Imp-% at the end of ablation was very low (less than 0.25). To confirm the accuracy of the technique used to identify the plateau values of the FTI/LSI over time curves, both per-patient analyses and combined analyses of all-patients lesions were performed and their results were compared.