New Parameters in the Spectral Analysis for the Detection of Ventricular Late Potentials

Chun-Chen Lin^ad, Chih-Ming Chen^a, Ten-Fang Yang^bc, and Ing-Fang Yang^c

^aDepartment of Electrical Engineering, National Taiwan University of Science and Technology, Taipei
^bGraduate Institute of Medical Information, Taipei Medical University, Taipei
^cDepartment of Internal Medicine (Cardiology and Nephrology), Jen-Chi General Hospital, Taipei
^dDepartment of Electrical Engineering, Chin-Min College, Miaoli, Taiwan

Correspondence: Ten-Fang Yang, Graduate Institute of Medical Information, Taipei Medical University, 250 Wu Hsing Street, Taipei 110, Taiwan. E-mail: tfy@tmu.edu.tw, phone +886-2-23776730 ext 111, fax +886-2-23411903

Abstract. The standard methods for time domain signal-averaged electrocardiogram (SAECG) analysis have been established by the 1991 ESC/AHA/ACC Task Force. For frequency domain spectral analysis, there are no consensus approcahes aggreed upon by all research groups, however. In order to focus on the VLPs, this study is devoted in improving the frequency domain spectral SAECG analysis. For this purpose, the 40 ms segment at the terminal QRS complex was used in this study. The results have shown that the mean value of the root-mean-square area (RMSA) of VT patients in 40 to 250 Hz high frequency band was significantly smaller than that of a normal person (36 18 vs. 61 35, p<0.001). Additionally, the mean value of the ratio of root-mean-square area (RMSAR) of VT patients was smaller than normal (12.0 3.6 vs. 17.1 7.0, p<0.001) as well. Hence, these new parameters observed from the SAECG spectral analysis might be useful for future clinical practice.

1.  Introduction

Standard method for the analysis of VLPs in time domain has been well established but not in frequency domain [Breithardt et al., 1991]. It was demonstrated that the larger high frequency components originated from VLPs were presented in VT patients [Haberl et al., 1988; Pierce et al., 1989], but these results failed to be confirmed by others [Kulakowski et al., 1992]. The purpose of this study is to improve the spectral analysis for the detection of VLPs at the terminal QRS complex. A method was introduced to define the VLPs segment which was considered to have lower spectral amplitude in the high frequency band.

2.  Material and Methods

There were 43 normal Taiwanese (N) (12 men and 31 woman, aged 47 13, ranged from 30 to 79 years old) recruited for this study. There were 17 patients (10 men and 7 women, aged 66 16, aged from 42 to 98 years old) with sustained VT documented by 24-hour Holter ECG. All patients were suffering from chronic ischemia after clinically documented myocardial infarction.

The high resolution ECGs were recorded using a commercial Simens-Elema Megacart^®. A bipolar, orthogonal X, Y and Z lead system was used and a sample of 10 minutes raw ECG data with 12-bit resolution at 2 kHz was stored for analysis. The typical segment used in the spectral analysis is depicted as segment {A} in Fig. 1. The presence of VLPs can be demonstrated by a larger high frequency component. However, VLPs are to some extent mixed with the terminal normal QRS complex. Therefore, it is difficult to determine the onset of VLPs. For this reason, the starting point of VLPs was not chosen as the reference, the major segment selected for spectral analysis was the 40 ms segment before the QRS ends (segment {B} in Fig.1.). This method can evaluate the total amount of high frequency amplitude of terminal 40 ms QRS complex. The mean RMS40 of VT patients has been reported to be significantly smaller than normal due to the presence of VLPs in VT patients. Hence, this should lead to a smaller high frequency component in segment {B}.

The 80 ms length of Blackman Harris window with 92 dB side lobe was used to obtain signals from the terminal 40 ms QRS complex in order to prevent the edge discontinuity. The starting point of this window was set at 60 ms before the QRS ends. After windowing, this 80 ms signal was zero-padded to 256 ms. The spectral calculation based on fast Fourier Transform (FFT) was performed by the definition of discrete Fourier transformation (DFT) [Oppenheim and Schafer, 1989].

The composite spectrum of XYZ leads was performed by the method of spatial vector. The root-mean-square area (RMSA) is defined as RMS value of the spectral amplitude in 40 to 250 Hz band. The RMS area ratio (RMSAR) which is defined as 40 to 250 Hz RMS area / 0 to 250 Hz RMS area was used to evaluate the high frequency components.

Data were presented as mean standard deviation (SD). Student t test was used for comparing means of two independent variables and F test was used for the variance comparisons between parameters. Statistical significance was defined as p < 0.05.

Figure 5. Illustrations to dicpict the detection of VLPs in the late QRS complex. (a) QRS complex without VLPs, (b) QRS complex with VLPs

3.  Results

There were significant differences of all time domain parameters between VT patients and normals. The RMS area and RMS area ratio of VT patients were also significantly lower than normals.

Table 1. Summary of results in time domain and frequency domain SAECG analysis.

p^†<0.01, p^††<0.001 compared with normal subjects.

4.  Discussion

The major differences of spectral analysis method based on FFT algorithm reported form many previous studies were (1) the selection of VLPs segment, including the definition of starting point and segment length selected for analysis, and (2) the frequency band width used in evaluating the high frequency components of VLPs. However, the absence of a precise and reproducible approach in determining the starting point may be one of the key factors leading to the discrepancies of spectral analysis results. On the other hand, the data derived from the various definitions of spectral parameters (e.g. spectral area, energy and high frequency ratio) lead to further inconsistencies among these studies.

In this study, the reference point used in analyzing VLPs was the offset of QRS complex. One of the advantages in the selection of this reference point is its high reproducibility which had been documented in a number of time domain analyses. It is worth to note that, for VT patients, there were significantly lower high frequency components as illustrated by RMSA and RMSAR.

References

Breithardt G, Cain ME, el-Sherif N, Flowers NC, Hombach V, Janse M, Simson MB, Steinbeck G. Standards for analysis of ventricular late potentials using high-resolution or signal-averaged electrocardiography: a statement by a task force committee of the European Society of Cardiology, the American Heart Association, and the American College of Cardiology. Journal of the American College of Cardiology, 17:999-1006, 1991.

Haberl R, Jilge G, Pulter R, Steinbeck G. Comparison of frequency and time domain analysis of the signal-averaged electrocardiogram in patients with ventricular tachycardia and coronary artery disease: methodologic validation and clinical relevance. Journal of the American College of Cardiology, 12:150-158,1988.

Kulakowski P, Marlik M, Poloniecki J, Bashir Y, Odemuyiwa W, Farrell T, Staunton A, Camm J. Frequency versus time domain analysis of signal-averaged electrocardiograms. II. Identification of Patients With Ventricular Tachycardia After Myocardial Infaction. Journal of the American College of Cardiology, 20:135-143,1992.

Oppenheim AV, Schafer RW. Discrete-Time signal processing. Prentice-Hall, Inc., New Jersey, 1989.

Pierce DL, Easley AR Jr, Windle JR, Engel TR. Fast Fourier transformation of the entire low amplitude late QRS potential to predict ventricular tachycardia. Journal of the American College of Cardiology, 11:1731-1740, 1989.