Session PC6.2

Rest ECG Amplitude Changes Quantified by the ECG Variability Contour Method

G Dori*, M Gershinski, BS Lewis, H Bitterman

Carmel Medical Center
Haifa, Israel

A method for evaluating ECG amplitude changes during stress (compared with rest) was recently introduced. Briefly, after pre-processing procedures, the mean rest ECG complex was subtracted from all complexes in the rest and stress data sets, resulting in residue matrices. The amplitude of the residues depended on the difference of the samples composing the ECG complexes from the mean complex. The contour of the rest residue matrix (termed: rest ECG variability contour, rEVC) was used as a reference for evaluating stress ECG amplitude changes, instead of using standard voltage criteria. The residues of the stress residue matrix were cumulatively summed and normalized (NCS) to result in a graphic description of the trends of these residues. If NCS were contained within rEVC, amplitude changes would be considered “physiological” or insignificant. In contrast, if NCS crossed rEVC, amplitude changes would be considered significant and consistent. Hitherto, real ECG data were not analyzed with the rEVC method.
Goal: Before analyzing real stress and rest data, this study aimed to characterize ECG amplitude changes physiologically occurring at rest. Methods: 1052 rest ECG leads were recorded from 92 examinees undergoing myocardial perfusion imaging scan. 5 minute rest ECG was recorded by each standard ECG lead, from which three 1-minute-long traces were extracted to form the 3 rest data sets. Each rest data set served as a reference for the other two. For each set a mean complex was calculated and subtracted from the complexes composing the three data sets. Nine residue matrices were thus created. Three variables were calculated for the rest residue matrices: the area under curve (AUC) of the rEVC, which reflected “physiologic variability” of the reference data set. The AUC of NCS of the residues and the fraction of NCS lying outside the rEVC, which reflected consistent amplitude changes of the signal.
Results: On the average, <1 sample (0.5%) of the NCS was located outside the rEVC. The mean maximal variability of the AUC of the rEVC was 45.7%. In the majority of leads (73%) the AUC of NCS increased with the temporal distance between the reference and the compared data set.
Conclusions: 1) consistent ECG amplitude changes during rest were negligible using this method, 2) consistent ECG amplitude changes were detected and accurately localized to the corresponding ECG component, 3) subtle, yet consistent ECG changes which do not fulfill criteria of significance may also be detected, 4) given the length of the data sets (5 minute long), the difference between sets increased with time, i.e. closer data sets exhibited less difference.

(Abstract Control Number: 55)