Adaptive frequency tracking of the baseline ECG identifies the site of atrial fibrillation termination by catheter ablation

https://doi.org/10.1016/j.bspc.2013.02.005Get rights and content

Abstract

Multiple organization indices have been used to predict the outcome of stepwise catheter ablation in long-standing persistent atrial fibrillation (AF), however with limited success. Our study aims at developing innovative organization indices from baseline ECG (i.e. during the procedure, before ablation) in order to identify the site of AF termination by catheter ablation. Seventeen consecutive male patients (age 60 ± 5 years, AF duration 7 ± 5 years) underwent a stepwise catheter ablation. Chest lead V6 was placed in the back (V6b). QRST cancelation was performed from chest leads V1 to V6b. Using an innovative adaptive harmonic frequency tracking, two measures of AF organization were computed to quantify the harmonics components of ECG activity: (1) the adaptive phase difference variance (APD) between the AF harmonic components as a measure of AF regularity, and (2) and adaptive organization index (AOI) evaluating the cyclicity of the AF oscillations. Both adaptive indices were compared to indices computed using a time-invariant approach: (1) ECG AF cycle length (AFCL), (2) the spectrum based organization index (OI), and (3) the time-invariant phase difference TIPD. Long-standing persistent AF was terminated into sinus rhythm or atrial tachycardia in 13/17 patients during stepwise ablation, 11 during left atrium ablation (left terminated patients – LT), 2 during the right atrium ablation (right terminated patients – RT), and 4 were non terminated (NT) and required electrical cardioversion. Our findings showed that LT patients were best separated from RT/NT before ablation by the duration of sustained AF and by AOI on chest lead V1 and APD from the dorsal lead V6b as compared to ECG AFCL, OI and TIPD, respectively. Our results suggest that adaptive measures of AF organization computed before ablation perform better than time-invariant based indices for identifying patients whose AF will terminate during ablation within the left atrium. These findings are indicative of a higher baseline organization in these patients that could be used to select candidates for the termination of AF by stepwise catheter ablation.

Highlights

► Modified ECG system was used (chest lead V6 placed in the patient's back). ► Contribution on the ECG of right/left atrial appendage activities was estimated. ► An adaptive harmonic frequency tracking algorithm was developed. ► Two innovative indices of organization were extracted. ► Promising indices to identify non-invasively the site of AF termination by catheter ablation.

Introduction

Atrial fibrillation (AF) affects millions of individuals in the West and is associated with an increased risk of morbidity and mortality [1]. Haissaguerre et al. showed in 1998 [2] that most triggers of paroxysmal AF are located within the pulmonary veins that can be isolated by catheter ablation. It has been observed that AF induces itself atrial changes explaining its natural evolution toward sustainability [3]. In AF sustained for more than one year (i.e. long-standing persistent AF, LS-pAF), significant electrical and structural remodeling of the atria were observed, providing diffuse substrate for AF perpetuation [3], [4], [5], [6], [7]. Importantly, the contribution of the remodeling process to AF stability remains unknown to date [6]. Sites with continuous high frequency activities, described as complex fractionated electrograms (CFAEs, [8]), and regions showing structural discontinuities in the left atrium (LA) are currently believed to be involved in AF maintenance [4], [8]. Nademanee et al. showed in 2004 that ablation of CFAEs throughout the atria could restore sinus rhythm (SR) in LS-pAF [8]. Other groups observed the incremental benefit of linear ablations during ablation of LS-pAF over pulmonary veins isolation (PVI) [9], and of stepwise catheter ablation (step-CA) on the organization of AF until the conversion to atrial tachycardia (AT) or SR. Recently, the ablation of CFAEs with given characteristics was of benefit in terms of organization over PVI [10], including right atrium (RA) CFAEs ablation when the RA AF cycle length (AFCL) became shorter than that of the LA [11]. Therefore, step-CA has become the treatment of choice for the termination of AF in patients with LS-pAF [12]. The success rate of step-CA, however, appears limited as the organization required to achieve long term SR and the amount of additional RA ablation remain unknown to date. There is therefore a strong interest in predicting the procedural outcome from the pre-ablation ECG, and the additional RA ablation required to terminate LS-pAF.

Most commonly, AFCL measured from bipolar intracardiac recordings has been used to assess AF organization during step-CA [9], [13]. It has been observed that prolongation of AFCL during step-CA is predictive of procedural AF termination [9], [13]. However, in patients with LS-pAF in whom AF continued despite PVI, AFCL did not change during step-CA, while other indices based on large intracardiac dipoles did [14]. The characterization of the fibrillatory content of the ECG signals from its AF harmonic components using time-invariant bandpass filters has been previously developed [15], [16]. If the oscillations are irregular, which is often the case in most LS-pAF, the expected harmonics may largely deviate from the actual ones. Hence, if the filter is too narrow, the oscillations may escape from the time-invariant filter passband, while the estimates may become unreliable if the filter is too wide. Therefore, the reliability of such methods may be largely affected by the temporal variability of AF waveforms. In our study, we chose to circumvent this limitation by using adaptive bandpass filters. We hypothesize that new organization indices based on an adaptive harmonic frequency tracking scheme improve the assessment of organization in LS-pAF referred for step-CA.

Our study is aimed at (1) developing innovative signal processing indices based on the surface ECG to evaluate the level of AF organization, and (2) identifying patients in whom AF can be terminated by ablation within the LA using advanced signal processing indices based on pre-ablation ECG recordings.

The paper is organized as follows: patients and data acquisition are presented in Section 2, the methods in Section 3, the results are presented in Section 4, followed by a discussion and conclusions in Sections 5 Discussion, 7 Conclusion, respectively.

Section snippets

Electrophysiological study

All patients had effective oral anticoagulation for >1 month. All antiarrhythmic drugs, with the exception of amiodarone and beta-blockers, were discontinued for ≥five half-lives before the procedure. The procedure was performed under general anesthesia. The following catheters were introduced via the right femoral vein: (1) a 3.5 mm cooled-tip catheter for mapping and ablation (Navistar, Biosense Webster); (2) a steerable decapolar catheter (2–8–2 1 mm, Biosense Webster) within the coronary

Ventricular activity subtraction

Cancelation of ventricular activity was used to ensure the reliability of surface ECG analysis during AF. QRST complexes were canceled from 12-lead ECG recordings using the single-beat method [18], where QRS and T-waves are treated separately. The T-wave cancelation is based on a dominant T-wave approach and QRS cancelation is basically an interpolation of atrial activity with weighted sinusoids. Importantly, this approach does not depend upon the length of the ECG recording, and generates no

Results

In the study population (Table 1), 77% (13) of the patients were terminated into AT (11) and SR (2) during step-CA, among which 65% (11) during LA ablation and 12% (2) during RA ablation. The remaining 4 patients (23%) stayed into AF at the end of the step-CA despite extensive ablation of similar duration compared to the other two groups as shown in Table 1. Importantly, NT patients displayed similar clinical characteristics compared to LT and RT patients including the LA volume, except for the

Discussion

Matsuo et al. [13] have recently reported that the RA activity during AF is a major contributor to the f-waves recorded from chest lead V1. Petrutiu et al. [17] extended their findings by showing that chest lead V1 reflects mostly the RA dynamics, while posterior leads were best correlated with LA activity. In our study, a strong correlation was found between the RAA activity and the AFCL measured on lead V1, and between the LAA activity and the AFCL of lead V6b. Interestingly, Berenfeld et al.

Limitations

Our study is mainly limited by the small study population. Future work should compare on a larger population the performance of our new indices with that of sustained AF duration to separate LT, NT and RT patients during the initial steps of step-CA. This limitation also prevented us from determining cut-off values from receiver–operator curves and perform a multivariable analysis. Another limitation is that the correlation between surface and intracardiac atrial activity was performed using

Conclusion

Our study suggest that using adaptive techniques, non-invasive measurements based on the harmonic dynamics of ECG f-waves can distinguish LS-pAF patients with LA termination from RT and NT patients.

The quantification of the coupling between the fundamental and its first harmonic of AF activity measured from the ECG is an important feature for characterizing AF organization and for discerning patients who successfully need single atria ablation from patients who need bi-atria ablation. Our

Acknowledgments

This study was supported by grant 205321_129876 from the Swiss National Science Foundation. The authors wish to thank Biosense Webster for their technical assistance.

References (38)

  • M.C. Wijffels et al.

    Atrial fibrillation begets atrial fibrillation. A study in awake chronically instrumented goats

    Circulation

    (1995)
  • S.Y. Ho et al.

    Left atrial anatomy revisited

    Circulation: Arrhythmia and Electrophysiology

    (2012)
  • S.M. Narayan et al.

    Clinical mapping approach to diagnose electrical rotors and focal impulse sources for human atrial fibrillation

    Journal of Cardiovascular Electrophysiology

    (2012)
  • U. Schotten et al.

    Pathophysiological mechanisms of atrial fibrillation: a translational appraisal

    Psychological Review

    (2011)
  • S.M. Narayan et al.

    Repolarization alternans reveals vulnerability to human atrial fibrillation

    Circulation

    (2011)
  • M. Haïssaguerre et al.

    Changes in atrial fibrillation cycle length and inducibility during catheter ablation and their relation to outcome

    Circulation

    (2004)
  • R.J. Hunter et al.

    Characterization of fractionated atrial electrograms critical for maintenance of AF: a randomized controlled trial of ablation strategies (The CFAE AF Trial)

    Circulation: Arrhythmia and Electrophysiology

    (2011)
  • V. Fuster et al.

    ACCF/AHA/HRS focused updates incorporated into the ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines

    Circulation

    (2011)
  • T.H. Everett et al.

    Assessment of global atrial fibrillation organization to optimize timing of atrial defibrillation

    Circulation

    (2001)
  • Cited by (14)

    • Atrial fibrillation frequency tracking in ambulatory ECG signals: The significance of signal quality assessment

      2018, Computers in Biology and Medicine
      Citation Excerpt :

      As a result, 14 patients were excluded since they did not perform any physical activity, whereas the remaining 24 performed physical activity from 13 to 46 min, see Table 1. Tracking of the DAF comprises f-wave extraction [29], frequency tracking [30], and signal quality assessment [27], see Fig. 1. Each step is briefly described in the following.

    • A High Baseline Electrographic Organization Level Is Predictive of Successful Termination of Persistent Atrial Fibrillation by Catheter Ablation

      2016, JACC: Clinical Electrophysiology
      Citation Excerpt :

      From the 2 outputs of the HFT (the fundamental and its first harmonic) two novel complementary organization measures of AF organization were derived: 1) the ECG adaptive organization index (AOI) quantifying the cyclicity of the DF and the first harmonic oscillations, bounded between zero (low organization) and one (high organization); and 2) the ECG adaptive phase index (API) quantifying the phase coupling (PC) between the DF and its first harmonic (by definition, small API values are indicative of well-organized AF oscillation). The performance of the AOI and API were compared with that of the AF cycle length (AFCL) and Everetts’s organization index (OI; computed only on the first 2 harmonics) (16,17). More details concerning the computational steps of the 4 indexes are provided in the Online Appendix.

    View all citing articles on Scopus
    View full text