Acceleration Spectrum Analysis: A Novel Quantitative Method for Frequenc y‐Domain Analysis of the Signal‐Averaged Electrocardiogram

Background: Frequency‐domain techniques presently used for micropotential analysis in the signal averaged ECG (SAECC) have several inherent shortcomings. For example, they depend on sensitive determination of the I‐point, which becomes inaccurate in the presence of noise, or derivation of multiple, complicated statistical parameters to quantify spectral characteristics in a three‐dimensional “spectral temporal map.” While these techniques are not as well accepted clinically as the conventional time‐domain Simson method, the latter is not without limitations either. Although time‐domain SAECG analysis has a very high negative predictive value, it has low positive predictive accuracy. Furthermore, it cannot be used to analyze SAECC data in patients with conduction delay problems, such as bundle branch block. Hence, the goal of this new frequency‐domain technique is to address and solve some of these shortcomings. Methods: The Fourier transform of the second derivative signal, or “acceleration spectrum,” extracts the frequency‐domain “signature” of damaged myocardium throughout the entire QRS complex, rather than from only the late potential region. The technique i s not dependent on precise endpoint or other fiducial point determination. A “spectral change index” (SCI) for quantifying variation from 50‐300 Hz in the acceleration spectrum i s calculated. The characterization of the cut‐off values for the SCI was based on results from a study including 50 postmyocardial infarction (post‐MI) patients (25 of whom were inducible to sustained ventricular tachycardia), and 10 normal controls. Results: An SCI <20, typical of a normal, “flat” acceleration spectrum in the 50‐ to 300‐Hz band width, may indicate undamaged myocardium, while an SCI 220 corresponding to a higher degree of spectral “fragmentation” in the same bandwidth, may indicate increased myocardial tissue damage. Using this cutoff, the sensitivity, specificity, and positive and negative predictive values for this initial study were 72%, 84%, 82%, and 75%, respectively. Conclusions: Acceleration spectrum analysis (ASA) using the SCI shows promise in predicting inducibility in post‐MI patients, including those with conduction delay problems. Since it is well documented that time‐domain SAECG has a high negative predictive value and a low positive predictive value, the high positive predictive value of the newly developed ASA increases the overall value of the SAECG test.