Credits:Rakesh Latchamsetty, MD; Abraham G Kocheril, MD
Running Title: Dominant Frequency in Atrial Fibrillation
Address for Correspondence: Rakesh Latchamsetty, University of Illinois at Chicago, Section of Cardiology , 4766 Park Road, Ann Arbor, MI 48103.
Significant advancements have been
made in the technology and approach to catheter ablation of atrial fibrillation
(AF). Pulmonary vein isolation has emerged as the dominant strategy in this
procedure and has fueled innovations in catheter design as well as various
mapping and navigation systems. Mapping and targeting of complex fractionated
atrial electrograms has also emerged as an additional or alternate strategy employed
by some ablationists. Recently, attention is being drawn to a new approach
targeting atrial sites with high dominant frequencies (DF) derived from their
electrograms. This article is a review of the basic concepts of DF, the
relevant literature behind DF analysis in AF, and the potential clinical
applicability of DF analysis for catheter ablation.
Key Words : Dominant Frequency, Atrial
Abbreviations : AF - atrial fibrillation;
CFAE – complex fractionated atrial electrogram; DF - dominant frequency
Despite the vast extent of
resources devoted to its understanding and treatment, atrial fibrillation (AF) remains
the most significant clinical arrhythmia in terms of its morbidity and
continues to challenge us to fully uncover its elusive mechanisms. This is not
to say, however, that there has not been significant progress made. On the
contrary, remarkable strides have been made both in the basic science
laboratory to uncover the mechanisms by which the arrhythmia is generated and
sustained, and in clinical studies of treatment. Catheter ablation has emerged
as an important therapy in drug refractory AF and has brought with it
continuously improving technology in areas including catheter design, mapping
systems, and overall ablation strategy.
Catheter-based pulmonary vein
isolation as a means to electrically isolate triggers of AF has significantly
improved success at maintaining sinus rhythm (particularly in paroxysmal AF)
but has not eliminated AF altogether. Permanent maintenance of pulmonary vein
electrical isolation can often prove to be challenging, and important sources outside
the pulmonary veins and predominantly in the left atrium remain. Targeting
areas of complex fractionated atrial electrograms (CFAE), although reported with
varying degrees of efficacy, is believed to modify important sites for the
maintenance of AF. As an adjunct to pulmonary vein isolation, CFAE ablation
has gained some favor as a means to tailor ablation to a given patient .
Recently, significant attention has
been directed toward utilizing dominant frequency (DF) analysis to highlight sites
of rapid and consistent periodicity which are believed to represent likely targets
for maintenance of AF . This article provides an overview of the basic
concepts behind DF analysis, reviews the pertinent data regarding its
application to AF, and discusses the current and future status of its clinical
Multiple theories regarding the
initiation and maintenance of AF have been introduced and evolved our
understanding of this arrhythmia since at least the beginning of the twentieth
century when Winterberg, et al. in 1907 described AF as due to multiple rapidly
firing atrial foci [3, 4]. In 1914, Mines
promoted the theory of circus movement as the key reentrant mechanism to
perpetuate AF . In 1964, Moe et al. described AF as a self
sustaining process of multiple randomly propagating wavelets made possible by
an atrial substrate of heterogeneous refractoriness [6, 7]. Challenging
this, at least in some cases of AF, has been the theory of mother rotors, described
by Jalife, et al. as discrete self sustaining reentrant foci that may provide
the engine for the maintenance of AF. These rotors tend to anchor to various
anatomical substrates, thereby maintaining a relatively fixed location and giving
off multiple randomly circulating wavelets. These wavelets then propagate
throughout the atria with varying amounts of degradation (fibrillatory
conduction ) as they encounter the heterogeneous tissue of
the left and right atria and their connections . The presence of these rotors have been studied
through optical mapping of the left atrium and their periodicities have been
shown to match with electrogram frequencies at those locations . Identifying the sites of these rotors by their rapid and periodic deflections on electrograms is the aim of frequency
In its application with respect to
AF, the fundamental goal of DF analysis for any given location is to find the
activation rate of the dominant atrial signal at that site. The standard
approach during electrophysiology studies utilizes a time domain analysis where
the amplitude of the signal as seen on the electrogram is plotted against
time. However, during AF, the varying amplitude and morphologies of the atrial
signals often preclude accurate measurement of atrial cycle lengths. DF
analysis, however, dissects the electrogram into components of varying
frequencies and creates a power spectrum based on their amplitudes.
Ultimately, the “dominant” or highest amplitude frequency is identified and
used to determine the activation rate of the primary atrial signal from that
location . Below is a more detailed description on the
specific algorithms used in DF analysis.
Initially, an electrogram from a
given atrial site is obtained over a period of about five seconds . This “time domain” signal then undergoes a
Fast Fourier Transform to display a power spectrum of its frequencies and
ultimately to identify a DF. The fundamental principle behind the Fourier
Transform is that any time series (such as an atrial electrogram) can be
portrayed as a sum of a discrete set of sinusoidal waves of specific
frequencies, amplitudes, and phase shifts. After the signal is broken into
these compository waves, a power spectrum of their frequencies is created and a
DF can be identified (figure 1). The
term “Fast” Fourier transformation implies that the time sample analyzed is of
a power of 2 which lends itself to a more efficient and quick analysis.
Figure 1: Frequency spectrum derived from an atrial electrogram.
A. Bipolar left atrial electrogram from a patient in atrial fibrillation. This signal, after undergoing several processing steps
(including bandpass filtering and rectification) is broken down by Fourier transformation to its component sinsusoidal waves.
B. Examples of sine waves of varying frequency, amplitude and phases are pictured.
C. Ultimately the power spectrum of the different frequencies are plotted and the dominant
frequency is identified.
In order to provide a cleaner or
more discrete frequency power spectrum from biphasic electrogram recordings,
signals need to undergo several processing steps including bandpass filtering,
rectification, and signal tapering. Band-pass filtering serves to attenuate
signal “noise” outside a specified desired frequency range and highlights deflections
that represent local atrial depolarization. Rectification converts the
biphasic signal to a monophasic one more easily represented by a sinusoidal
wave; and signal tapering reduces to baseline the signals at the two ends of a
specified time “window” to prevent incompletely recorded deflections from
affecting the data [10, 12]. All these
steps serve to “clean” the frequency spectrum and help elucidate the dominant
atrial waveform and its periodicity.
Any atrial site can be examined
utilizing the methods detailed above and a DF can be assigned to each of these
sites. Ultimately, the implication in targeting sites with high DFs for
ablation is that they potentially represent the location of rotors that may be
responsible for the maintenance of AF.
Several studies have suggested that
in both animal and human hearts in AF, left atrium and pulmonary vein sites tend
to have higher DFs than the right atrium, thereby representing a left to right DF
gradient [11, 13, 14]. Lazar et al. demonstrated the presence of
this gradient correlated with a higher probability of successful ablation
through pulmonary vein isolation . However, whether high DF sites in the pulmonary
veins correlate with their identification as triggering foci remains unclear as
the tissue characteristics have not yet been fully described. In an animal
model of Langendorff-perfused sheep hearts in AF, Jalife et al. elegantly demonstrated
this left to right atrium gradient and showed its path of decrementation as it
crossed Bachman’s bundle from the left atrium into the right atrium . All these findings support the theory that
periodic and relatively stationary high frequency sources to maintain AF may be
present in the left atrium and can therefore represent potential targets for therapy
Based on the concept of high DF
sites representing potential sources maintaining AF, its specific clinical
application is still being determined. Sanders et al. demonstrated in a study
where the operators during AF ablation were blinded to the DF analysis,
ablation at a high DF site was more likely to prolong the AF cycle length . Also, in paroxysmal AF the majority of AF
termination occurred while ablating a high DF site. Furthermore, a difference
in DF distribution was seen where patients with paroxysmal AF had high DF sites
that were more likely to lie in the pulmonary vein whereas in persistent AF,
left atrial DFs were often higher. Some groups are also beginning to perform
real time analysis of DF to guide ablation. When targeting high DF sites, Atienza
et al. have seen a higher probability of remaining free from (arrhythmia or) AF
when successfully ablating DFmax sites or abolishing the left to right atrium
DF gradient .
While a definitive role for DF
analysis to guide ablation is yet to be established and varying accounts of its
efficacy have been presented, this concept is certainly one that warrants
further clinical evaluation. Targeting of high DF sites may ultimately have an
expanded role in AF ablation, particularly as an adjunct to pulmonary vein
isolation. The exact strategy for DF based ablation is still being developed
and this process will certainly face a number of challenges. Among these is
the dynamic nature of DF spectra during PVI and left atrial ablation which may
require remapping of DF sites during the procedure. Software and catheter
design may also need to be improved to provide more efficient and accurate real
time DF maps. Also, different strategies may prevail between chronic and
paroxysmal AF patients, especially in relation to its use with pulmonary vein
isolation. Other potential obstacles to accurate DF analysis such as far-field
signals, signal alteration around scars, and prior ablation lesions setting up local
reentry will also need to be addressed.
However, with innovations such as catheter
systems that can deliver a large number of electrodes to the atria and
noncontact mapping, evaluating the substrate for high DF sites will likely become
more efficient. Furthermore, ongoing clinical investigation is advancing our
understanding of the clinical significance of high DF sites and strategies on
how and when to map and ablate at these sites are being developed. If these
advancements in design and refinement of strategies to target DF sites prove to
increase success rates, curative ablation of AF may be accomplished with less
atrial damage than current approaches.
Oral, H., Chugh, A., Good, E., et al. A tailored approach to catheter ablation of paroxysmal atrial fibrillation. Circulation. 2006 Apr 18;113(15):1824-31.
Atienza, F., Jalife, J. Reentry and atrial fibrillation. Heart Rhythm. 2007 Mar;4(3 Suppl):S13-6.
Jalife, J., Berenfeld, O., Mansour, M. Mother rotors and fibrillatory conduction: a mechanism of atrial fibrillation. Cardiovascular research. 2002 May;54(2):204-16.
Winterberg, H. Uber die wirkung des N. vagus und accelerans auf das Flimmen des Herzens. Arch Physiol. 1907;117:34.
Mines, G. R. On dynamic equilibrium in the heart. The Journal of physiology. 1913 Jul 18;46(4-5):349-83.
Moe, G. K., Abildskov, J. A. Atrial fibrillation as a self-sustaining arrhythmia independent of focal discharge. American heart journal. 1959 Jul;58(1):59-70.
Moe, G. K., Rheinboldt, W. C., Abildskov, J. A. A Computer Model of Atrial Fibrillation. American heart journal. 1964 Feb;67:200-20.
Ryu, K., Khrestian, C. M., Matsumoto, N., et al. Characterization of the critical cycle length of a left atrial driver which causes right atrial fibrillatory conduction. Conf Proc IEEE Eng Med Biol Soc. 2004;6:3960-3.
Skanes, A. C., Mandapati, R., Berenfeld, O., et al. Spatiotemporal periodicity during atrial fibrillation in the isolated sheep heart. Circulation. 1998 Sep 22;98(12):1236-48.
Ng, J., Goldberger, J. J. Understanding and interpreting dominant frequency analysis of AF electrograms. Journal of cardiovascular electrophysiology. 2007 Jun;18(6):680-5.
Stiles, M. K., Brooks, A. G., John, B., et al. The effect of electrogram duration on quantification of complex fractionated atrial electrograms and dominant frequency. Journal of cardiovascular electrophysiology. 2008 Mar;19(3):252-8.
Botteron, G. W., Smith, J. M. A technique for measurement of the extent of spatial organization of atrial activation during atrial fibrillation in the intact human heart. IEEE transactions on bio-medical engineering. 1995 Jun;42(6):579-86.
Lazar, S., Dixit, S., Marchlinski, F. E., et al. Presence of left-to-right atrial frequency gradient in paroxysmal but not persistent atrial fibrillation in humans. Circulation. 2004 Nov 16;110(20):3181-6.
Mansour, M., Mandapati, R., Berenfeld, O., et al. Left-to-right gradient of atrial frequencies during acute atrial fibrillation in the isolated sheep heart. Circulation. 2001 May 29;103(21):2631-6.
Sanders, P., Berenfeld, O., Hocini, M., et al. Spectral analysis identifies sites of high-frequency activity maintaining atrial fibrillation in humans. Circulation. 2005 Aug 9;112(6):789-97.
Atienza, F., Almendral, J., Jalife, J., et al. Real-time dominant frequency mapping and ablation of dominant frequency sites in atrial fibrillation with left-to-right frequency gradients predicts long-term maintenance of sinus rhythm. Heart Rhythm. 2009 Jan;6(1):33-40.