Comparative study of methods for ventricular activity cancellation in atrial electrograms of atrial fibrillation

Physiol Meas. 2007 Aug;28(8):925-36. doi: 10.1088/0967-3334/28/8/014. Epub 2007 Jul 19.

Abstract

Atrial fibrillation is a very common cardiovascular disease in clinical practice. One relevant issue to understand its pathophysiological mechanisms is the analysis and interpretation of atrial electrograms (AEG). To study these signals properly, ventricular activity has to be removed from the AEG. In this work, a new application of independent component analysis (ICA) to the AEG is presented, where ventricular activity is removed from atrial epicardial recordings making use of only one reference lead. Therefore the technique is suitable when multi-lead recordings are unavailable as in atrial implantable cardioverter defibrilators. In addition to the proposed new methodology this work also presents the first comparative study, making use of unipolar epicardial AEGs, among the ICA-based technique, template matching and subtraction (TMS), and adaptive ventricular cancellation (AVC) on a database of 20 patients. A performance comparative analysis was carried out by evaluating epicardial atrial waveform similarity (S) and ventricular depolarization reduction (VDR) as a function of atrial rhythm regularity on a beat-by-beat basis. Results indicate that, when the epicardial atrial rhythm is quite organized, ICA is able to preserve the atrial waveform very precisely and better than the other methods (median S = 99.64% +/- 0.31% in contrast to 95.18% +/- 2.71% for TMS and 94.76% +/- 4.12% for AVC). Moreover, ventricular reduction is the best for ICA (median VDR = 6.32 +/- 4.41 dB in contrast to 4.98 +/- 4.48 dB for TMS and 4.12 +/- 2.72 dB for AVC). On the other hand, when the atrial activity is disorganized, TMS notably improves performance (S = 97.72% +/- 1.87%), but ICA still is the best in waveform preservation (S = 98.22% +/- 1.53%) whereas AVC remains similar (S = 93.74% +/- 4.38%). In conclusion, ICA can be considered as notably the best approach to reduce ventricular activity from unipolar atrial electrograms in organized atrial arrhythmias. On the other hand, both TMS and ICA give quite similar results when the atrial arrhythmia is disorganized.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Atrial Fibrillation / physiopathology*
  • Data Interpretation, Statistical
  • Electrocardiography / statistics & numerical data*
  • Heart / physiopathology*
  • Heart Atria / physiopathology
  • Heart Ventricles / physiopathology
  • Humans
  • In Vitro Techniques
  • Pericardium / physiopathology
  • Principal Component Analysis
  • Reproducibility of Results