New quantitative methods of ventricular repolarization analysis in patients with left ventricular hypertrophy

Left ventricular hypertrophy (LVH) is accompanied by specific changes in ventricular electrophysiology, which are potentially arrhythmogenic. Nevertheless, the electrocardiographic diagnostic signs for LVH have a relatively low predictive power for arrhythmic events and sudden death. We thought that other parameters derived from the surface ECG, not apparent at visual inspection, might be detected by specific analysis of electrocardiographic digital recordings. The purpose of our work was to analyze the surface distribution of repolarization potentials and search for subtle alterations not revealed by the usual electrocardiographic processing, which are likely to reflect ventricular repolarization heterogeneity.

Body surface potential maps were recorded from 62 chest leads in 16 patients with LVH due to aortic stenosis and in 35 normal subjects. By applying a principal component analysis of the ST-T waves, we computed the similarity index. The value of the similarity index is inversely proportional to the variability of T wave morphology and a low value is considered a marker of repolarization heterogeneity.

The similarity index was significantly lower in LVH patients than in normals both in 62 leads (0.73 +/- 0.067 vs 0.77 +/- 0.044, p = 0.03) and in 12 unipolar leads (V1- V8, V3R, VR, VL, VF) extracted from the map (0.77 +/- 0.075 vs 0.81 +/- 0.045, p = 0.03). Moreover, we computed the "late repolarization deviation index", which quantifies the instantaneous variations of surface potential distribution from peak to end of the T wave. This index was significantly higher in LVH patients than in controls (in 62 leads 0.07 +/- 0.05 vs 0.028 +/- 0.016, p = 0.005; in 12 leads 0.064 +/- 0.052 vs 0.024 +/- 0.020, p = 0.008).

The values of similarity index and of late repolarization deviation index found in LVH patients suggest a higher than normal degree of repolarization heterogeneity, not detected by the usual electrocardiographic analysis. Since both indices maintained statistical significance when calculated on the 12 leads derived from our map lead system, they could be reliably computed from digital recordings of the 12 conventional leads.