T-wave frequency content evaluation in healthy subjects and patients affected by myocardial infarction

Abnormalities in the cardiac repolarization and, thus, in the electrocardiographic (ECG) T wave, are known to be associated to several cardiovascular diseases. For this reason, cardiac repolarization has recently became matter of major interest and several T-wave indexes have been proposed in the literature in the attempt to noninvasively discriminate pathological from healthy conditions. In the present chapter we designed two new indexes based on the T-wave frequency content at low and high frequencies, respectively. Abilities of such indexes in automatically detecting abnormal cases of repolarization were tested and compared using ECG recordings of 39 control healthy (CH) subjects and 39 acute myocardial infarction (AMI) patients. T-wave frequency evaluation was performed on the ECG vector magnitude signal (VMS) by computing the normalized power spectrum and the normalized energy according to the Fourier transformation. In both populations most of the T-wave frequency content was comprised within 10 Hz, frequency at which the normalized energy relative to the CH subjects and AMI patients reached 97.5±2.7% and 96.3±2.6%, respectively (P<0.01). Nevertheless, the normalized energy distribution relative to the two populations was significantly different. Indeed, the T-wave energy curve of the AMI patients was significantly (P<0.01) higher in correspondence of any frequency between 2 Hz and 5 Hz, due to a reduced repolarization variability observed over this frequency range. Consequently, the area under the normalized energy curve in the 2-5 Hz band (AUNE2-5) was found to be significantly greater for the AMI patients (AUNE2-5=248±15Hz) than the CH subjects (AUNE2-5=230±15 Hz; P<10-5), and allowed a discrimination between the two populations with a sensitivity and specificity of 79.5% and 74.4%, respectively. Instead, for frequencies comprised between 10 Hz and 35 Hz, the normalized T-wave energy curve of the AMI patients was significantly (P<0.01) lower than that of the CH subjects, due to an increased high-frequency variability observed in the AMI spectra. Thus, the area under the normalized energy curve in the 10-35 Hz band (AUNE10-35) was found to be significantly lower for the AMI patients (AUNE10-35=2467±32 Hz) than the CH subjects (AUNE10-35=2491±28 Hz; P<10-8), and allowed a discrimination between the two populations with a sensitivity and specificity of 94.9% and 74.4%, respectively. In conclusion, the differences in the T-wave frequency content of CH subjects and AMI patients can be described using our new AUNE2-5 and AUNE10-35 indexes. The latter, however, showed a greater ability to discriminate abnormal repolarization associated to the acute myocardial infarction.