Fast fourier transformation of the entire low amplitude late QRS potential to predict ventricular tachycardia

Abstract

Signal-averaged electrocardiograms (X, Y and Z leads) were acquired from 24 patients with coronary artery disease and recurrent ventricular tachycardia, 24 control patients with coronary artery disease and 23 normal subjects to assess the discriminant value of fast Fourier transformation of the entire late potential period of the QRS complex. Analysis of the vector magnitude in the temporal domain (25 to 250 Hz bandpass filters) measured high frequency QRS duration, the duration of terminal signals less than 40 microV and the root mean square voltage of the last 40 ms. Late potentials were defined as terminal signals greater than 25 Hz that were less than 40 microV. Analysis in the frequency domain used a 120 ms window that encompassed (had onset with) all of the late potential, but the mean value was first subtracted to eliminate a direct current component. High frequency spectral areas (60 to 120 Hz) and the percent high frequency (100 x [60 to 120 Hz/0 to 120 Hz]) were calculated. Results in both temporal and frequency domains were similar in control patients with coronary artery disease and normal subjects. Patients with ventricular tachycardia had a longer high frequency QRS complex (p less than 0.0001) and longer high frequency terminal signals less than 40 microV (p less than 0.0004), but not significantly lower voltage in the last 40 ms. The most useful temporal domain measurement was high frequency QRS duration (if greater than or equal to 120 ms, odds ratio = 8.2). Patients with ventricular tachycardia had increased high frequency spectral areas (p less than 0.0002) in the late potential, and the percent high frequency was especially increased (p = 0.0000; if percent high frequency greater than 3.1%, odds ratio = 18.4). The odds ratio and the area under the receiver operating characteristic curve were both greater for percent high frequency than for high frequency QRS duration (p less than 0.03). All patients with ventricular tachycardia had a high frequency QRS complex greater than or equal to 107 ms or percent high frequency greater than or equal to 3.1% (sensitivity 100%). For a high frequency QRS complex greater than or equal to 107 ms and percent high frequency greater than or equal to 3.1%, specificity was 96%. Therefore, high frequencies in late potentials, not their duration or reduced voltage, most usefully identify patients with coronary artery disease who are prone to ventricular tachycardia.

Low-frequency component of body surface potential maps identifies patients at risk for ventricular tachycardia

“…Further analysis shows that, relative to the NORM group, low-frequency power increases for the MI/VT+ group during QRS and then decreases for both myocardial infarction groups (relative to the NORM group) in the ST-T interval. A number of studies have focused on spectral analysis of the terminal QRS and ST segment [11,[24][25][26] , with sometimes conflicting results. Since the data segments investigated in these studies were short and often windowed, the low-frequency range of the power spectrum was often ignored.…”

“…The low-frequency region of the PS [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] obtained from analysis of the QRS-complex data segment most accurately identified patients with a history of risk for ventricular tachycardia, with a predictive accuracy of 74 6%. Non-windowed data also revealed significant lowfrequency differences, in 18 of 30 group-mean power spectra.…”

“…For all patients in the group we then calculated the average of the spectra from each of these leads and designated it the group mean power spectrum for each of the frequency bands: PS 1-7 ( Fig. 2(b)), PS [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] , PS [25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40] , PS 40-60 , PS 60-80 , PS 1-100 , PS , PS . The subscripts indicate from which isoharmonic map the power spectrum was extracted; and each group-mean power spectrum spans the entire frequency range (0·05-125 Hz).…”

Improved frequency-domain analysis of ventricular late potentials

“…The pre-defined segment of SAECG enables the spectral area in the high-frequency band to be used along with the spectral area ratio to identify high-risk VT patients [4][5][6][7]. This approach has been claimed to be superior to analysis in the time domain.…”

Comparison of the new acceleration spectrum analysis with other time- and frequency-domain analyses of the signal-averaged electrocardiogram

“…While this is still the most frequently used technique, and has proven highly reproducible [12][13][14] , it nonetheless suffers from significant shortcomings: high-pass filters are required, discrimination between noise and late potentials may be difficult, and its results in patients with bundle branch block are uncertain [15][16][17] . To overcome these limitations, several frequency-domain techniques have been developed in recent years [18][19][20][21][22][23][24] . The first frequencydomain method was the area ratio of the terminal QRS described by Cain et al [18] ; however, the reproducibility of this method is very weak [25] ; more recently, Cain et al [19] proposed another method, spectral and temporal analysis of the entire cardiac cycle.…”