Averaging real-time impedance enhances the prediction of steam pop risk and lesion characteristics

A novel impedance filtering function that averages impedance values was developed to mitigate cardiac and respiratory oscillations. We aimed to evaluate the clinical significance of averaging real-time impedance in predicting steam pops (SPs) and lesion characteristics.

Radiofrequency (RF) ablation was performed in 20 swine using a flexible-tip temperature-controlled power regulation catheter. Both unfiltered and filtered (averaged) impedance values were recorded using the EnSite™ X system. For each RF application, absolute (ΔImp-drop) and relative (%Imp-drop) impedance drops were quantified. Associations between impedance parameters and SP occurrence, atrial lesion transmurality, and ventricular lesion dimensions were evaluated. Among 959 lesions, SPs occurred in 36 applications (3.8%), all within the ventricles. Notably, 6 SPs occurred within 90 s despite RF power ≤ 40 W, with 4 during left ventricular ablation under low systolic blood pressure (<40 mmHg). Lesions with SPs exhibited significantly greater unfiltered and averaged ΔImp-drop and %Imp-drop (all P < 0.001). Averaged %Imp-drop showed the highest predictive value for SPs (AUC = 0.93), with a 20.9% cut-off yielding 88.9% sensitivity and 85.5% specificity. The time to reach the initial 10%, 15%, and 20% reduction in averaged %Imp-drop was not associated with SP occurrence. Both unfiltered and averaged impedance drops correlated with atrial transmural lesion formation. Averaged impedance drops significantly improved estimation of lesion depth, surface area, and volume compared with unfiltered values (P < 0.01).

The averaged relative impedance drop demonstrated the strongest association with SP occurrence, and averaging impedance provided a more accurate assessment of lesion characteristics than unfiltered measurements.