Session S74.6

Vectorcardiographic T-Wave Alternans Ranking: Striking Disagreement between Two Measures of Repolarization Heterogeneity

S Man*, AC Maan, MJ Schalij, EE van der Wall, CA Swenne

Leiden University Medical Center
Leiden, Netherlands

Background: Multiple studies have demonstrated the predictive value of T-wave alternans (TWA) for lethal arrhythmias and sudden cardiac death. Methods for T-wave alternans quantification rest on two important pillars: the features selected to represent the T wave, and the method that is chosen to compute the alternans in the selected T-wave feature. Regarding T-wave feature extraction, only limited attention has been paid to the electrophysiological correlates of the parameters chosen to represent T-wave morphology. For that reason we compare here the alternans in two vectorcardiographic parameters for which computer simulation studies have demonstrated that they are indexes of repolarization heterogeneity: the magnitude of the vectorial T-wave integral (iT, in mV.ms), and the magnitude of the maximal T-loop vector (Tmax, in µV).
Methods: We analyzed the first 30 seconds of all 100 ECG recordings comprised in the Computers in Cardiology 2008 Challenge data set with our research ECG/VCG processing program LEADS (compute Cardiol 2005;32:515-518) running in the unsupervised batch mode. Briefly, LEADS detects QRS complexes, removes baseline wander, deselects noisy and premature beats, computes an average beat, and synthesizes a vectorcardiographic representation of this beat by using the inverse Dower matrix. For the purpose of this challenge, we added to LEADS the feature that all ECGs were processed twice, one time with only odd, and one time with only even beat numbers selected. LEADS computes a multitude of parameters in the average beat, among which are iT and Tmax. For each ECG, the absolute differences between the "odd" and "even" values of iT and of Tmax were taken as iT-TWA and as Tmax-TWA, respectively. Finally, we computed Pearson's correlation coefficient between the iT-TWA and Tmax-TWA values.
Results: Tmax-TWA was 14.73 ± 17.94 [0.07 - 80.25] µV, iT-TWA was 2.67 ± 3.39 [0.06 - 19.08] mV.ms, the correlation coefficient between Tmax-TWA and iT-TWA was 0.634 (p<0.01).
Discussion: The significant correlation between iT-TWA and Tmax-TWA demonstrates that there is common information in these TWA parameters. However, only 40% of the variations in iT-TWA can be explained by variations in Tmax-TWA. We intuitively propose iT-TWA as the superior TWA ranking parameter because it represents the whole T wave rather than a more noise-sensitive measurement of one instantaneous amplitude. Obviously, such a choice is to be corroborated by further follow-up studies, demonstrating the superior value of iT-TWA over Tmax-TWA in predicting lethal arrhythmias.

(Abstract Control Number: 367)