Impact of filter configurations on bipolar EGMs: An optimal filter setting for identifying VT substrates

The impact of filtering on bipolar electrograms (EGMs) has not been systematically examined. We tried to clarify the optimal filter configuration for ventricular tachycardia (VT) ablation.

Fifteen patients with VT were included. Eight different filter configurations were prospectively created for the distal bipoles of the ablation catheter: 1.0-250, 10-250, 100-250, 30-50, 30-100, 30-250, 30-500, and 30-1000 Hz. Pre-ablation stable EGMs with good contact (contact force > 10 g) were analyzed. Baseline fluctuation, baseline noise, bipolar peak-to-peak voltage, and presence of local abnormal ventricular activity (LAVA) were compared between different filter configurations.

In total, 2276 EGMs with multiple bipolar configurations in 246 sites in scar and border areas were analyzed. Baseline fluctuation was only observed in the high-pass filter of (HPF) ≤ 10 Hz (p < .001). Noise level was lowest at 30-50 Hz (0.018 [0.012-0.029] mV), increased as the low-pass filter (LPF) extended, and was highest at 30-1000 Hz (0.047 [0.041-0.061] mV) (p < .001). Conversely, the HPF did not affect the noise level at ≤30 Hz. As the HPF extended to 100 Hz, bipolar voltages significantly decreased (p < .001), but were not affected when the LPF was extended to ≥100 Hz. LAVAs were most frequently detected at 30-250 Hz (207/246; 84.2%) and 30-500 Hz (208/246; 84.6%), followed by 30-1000 Hz (205/246; 83.3%), but frequently missed at LPF ≤ 100 Hz or HPF ≤ 10 Hz (p < .001). A 50-Hz notch-filter reduced the bipolar voltage by 43.9% and LAVA-detection by 34.5% (p < .0001).

Bipolar EGMs are strongly affected by filter settings in scar/border areas. In all, 30-250 or 30-500 Hz may be the best configuration, minimizing the baseline fluctuation, baseline noise, and detecting LAVAs. Not applying the 50-Hz notch filter may be beneficial to avoid missing VT substrate.