Multidimensional analysis of atrial arrhythmias from the surface ECG.

摘要

Atrial arrhythmias are abnormal rhythms occurring in the two upper chambers of the heart. Many of these rhythms, including atrial fibrillation, atrial flutter, and focal atrial tachycardias, can be readily diagnosed from the surface electrocardiogram (ECG). However, much about how atrial arrhythmias are reflected in the ECG is not well understood, partially due to the fact that the arrhythmias themselves are not well understood.; This thesis demonstrates how information from ECG signals during arrhythmias can be gleaned through "multi-dimensional" analysis, namely from the temporal, spectral and spatial domains. Taking advantage of redundancy of information from multi-channel ECGs during atrial fibrillation, an algorithm was developed to "subtract" ventricular components to isolate atrial activity, which showed improvement over traditional techniques. Spectral analysis of isolated atrial activity of atrial fibrillation ECGs recorded as part of the multi-center AFFIRM trial showed that peak frequencies reflected anti-arrhythmic drug use, atrial fibrillation duration, and age, consistent with expected effects from electrophysiologic and structural remodeling.; Analysis of synthesized vectorcardiograms derived from the 12 lead ECG showed that the planes of vector loops from common isthmus-dependent flutter corresponded anatomically to the reentry circuit around the tricuspid valve. Counter-clockwise and clockwise isthmus-dependent flutter could be readily distinguished using the left anterior oblique projection of the vector loops. Details of the loops reflected the area of slow conduction on the atrial floor and the rapid conduction along the crista terminalis. The vector loops also provided insight to the phase relationship of flutter waves in different leads of the ECG. Additionally, this method also showed promise for differentiating left atrial flutter and focal tachycardias at various origins. Vector analysis of atrial fibrillatory activity in the surface ECG showed evidence of organization and transient linking and demonstrated that activation sequences during fibrillation tended occur in the sagittal plane, suggesting that fibrillation is not entirely random and the activation sequences may be constrained by anatomic factors.; The clinical implication of this work is that an improved understanding of atrial arrhythmias and the surface ECG can aid in directing specific treatment.