(Circulation. 1998;98:607-608.)
© 1998 American Heart Association, Inc.
Atrial Flutter Cycle Length Oscillations and Role of the Autonomic Nervous System
Flavia Ravelli, PhD
Medical Biophysics Division CMBM,
Trento, Italy
To the Editor:
In a recent issue of Circulation, Stambler
et al1 provide an intriguing but possibly
misleading set of results regarding the variability of atrial flutter
cycle length (AFCL) and the role of the autonomic nervous system in
AFCL modulation. Previous studies2 3 4 5 identified
in the AFCL variability the existence of 2 rhythmic
oscillations, 1 controlled by the ventricular
rate and 1 by respiratory activity, suggesting a modulatory effect of
atrial stretch. Using spectral analysis, Stambler et al claimed
the existence of an additional low-frequency oscillation
(band 1) mediated by sympathetic tone. Their conclusions are not
supported by their study design and results.
In fact, the newly proposed low-frequency oscillation
band was concentrated in the extremely low-frequency region of the
spectrum and on average was centered around 0.006±0.002 Hz (Table 2 of
the article by Stambler et al). This should correspond to an AFCL
oscillation with a period of 166 seconds, clearly longer
than the 2 minutes of analyzed recording time. This
means that we are dealing with a baseline trend rather than a rhythmic
oscillation. In other words, band 1 is simply a DC
component of the spectrum reflecting the nonstationary nature of the
analyzed time series and thus should be avoided in the
interpretation of the power spectra.6
In their study, Stambler et al also examined the role of the
autonomic nervous system in the modulation of AFCL by studying the
effect of pharmacological interventions on the power-band distribution
of AFCL spectra. Stimulation and inhibition of sympathetic tone
obtained by intravenous infusion of isoproterenol and
long-term therapy with ß-adrenergic blockers affected only the newly
proposed low-frequency band, leaving unchanged the respiratory and
ventricular rate–related oscillations. These
spectral changes were accompanied in patients receiving isoproterenol
by a decrease of the mean AFCL. From these results, the authors
conclude that band 1 is mediated by sympathetic tone.
By the above methodological considerations, isoproterenol and
ß-blockers, by changing only the DC component of AFCL spectra, seem
to be effective only on the stationarity of the time series by slowly
changing the level of AFCL. Thus, it seems probable that the increase
of power in band 1 caused by isoproterenol may simply reflect the trend
of shortening of the mean AFCL rather than indicate a modulatory role
of sympathetic tone on short-term AFCL variability. This is clearly
evident in Figure 2 of the article by Stambler et al, which shows the
effects of isoproterenol on band 1, in which the spectral estimator is
not able to attribute a frequency different from zero to this spectral
component (see the numerical data inserted in the figure). On the other
hand, modulatory sympathetic activity on the
cardiovascular variables is reported to be active
at frequencies around 0.1 Hz,7 at least 1 order
of magnitude greater than the average frequency of the band 1 given by
Stambler et al. In addition, the authors point out that vagal
stimulation by intravenous edrophonium does not modify
either overall AFCL variability or power-band distribution.
Thus, from these results it seems hard to conclude a possible role of
the autonomic nervous system in beat-to-beat modulation of AFCL. On the
other hand, the additional data given by the authors (that in
transplant patients, well-defined AFCL oscillations exist,
although the global variability was less than in control patients) do
not seem to offer any additional element to support the hypothesis of
an autonomic modulation of AFCL. In fact, the existence in transplant
recipients of AFCL fluctuations may strengthen the hypothesis of a
mechanical modulation of AFCL. The authors suggest that the diminished
variability may reflect the lack of innervation in transplant
recipients compared with preserved autonomic function in control
patients, indicating a role of autonomic tone in AFCL variability.
However, it is known that orthotopic heart transplant results in an
altered anatomy and function of the atria. In the hypothesis
that AFCL variability is caused by atrial
stretch,2 3 4 the reduced AFCL variability in
transplant recipients may also be explained by impaired mechanical
atrial function.
In conclusion, we believe that spectral analysis can be a
powerful tool for studying cardiovascular variability
series.6 7 8 However, it should be used with full
knowledge of the methodology, especially if commercially available
systems for heart rate variability analysis are used. In the
application of this technique to a novel variability series such as
AFCL, generated by the modulation of a reentrant mechanism clearly
different from the physiological mechanism of sinus
rhythm, attention should be paid to avoid tout court
transposition of the heart rate variability analysis as the
strained search of a sympathetic-mediated oscillation as
well as the arbitrary introduction of a sympathovagal balance index
(page 2517: Methods; page 2519: Results1 ).
However, we believe that if the identification of rhythms and their
physiological interpretation is correctly
performed, spectral analysis may provide some insight into the
intriguing mechanisms of AFCL oscillations.
References
1.
Stambler BS, Ellenbogen KA. Elucidating the
mechanisms of atrial flutter cycle length variability using power
spectral analysis techniques. Circulation. 1996;94:2515–2525.[Abstract/Free Full Text]
2.
Lammers WJEP, Ravelli F, Disertori M, Antolini R,
Furlanello F, Allessie MA. Variations in human atrial flutter cycle
length induced by ventricular beats: evidence of a
reentrant circuit with a partially excitable gap. J
Cardiovasc Electrophysiol. 1991;2:375–387.
3.
Ravelli F, Disertori M, Cozzi F, Antolini R,
Allessie MA. Ventricular beats induce variations in cycle
length of rapid (type II) atrial flutter in humans: evidence of leading
circle reentry. Circulation. 1994;89:2107–2116.[Abstract/Free Full Text]
4.
Waxman MB, Yao L, Cameron DA, Kirsh JA. Effects
of posture, Valsalva maneuver and respiration on atrial flutter rate:
an effect mediated through cardiac volume. J Am Coll
Cardiol. 1991;17:1545–1552.[Abstract]
5.
Ravelli F, Disertori M, Del Greco M, Antolini R.
Paradoxical modulation of atrial flutter cycle length by respiratory
activity. J Auton Nerv Syst. 1993;43(suppl):105.
Abstract.
6.
Task Force of the European Society of
Cardiology and the North American Society of Pacing and
Electrophysiology. Heart rate variability: standards of measurements,
physiological interpretation, and clinical use.
Circulation. 1996;93:1043–1065.[Free Full Text]
7.
Malliani A, Pagani M, Lombardi F, Cerutti S.
Cardiovascular neural regulation explored in the
frequency domain. Circulation. 1991;84:482–492.[Abstract/Free Full Text]
8.
Nollo G, Del Greco M, Ravelli F, Disertori M. Evidence
of low- and high-frequency oscillations in human AV
interval variability: evaluation with spectral analysis.
Am J Physiol. 1994;267:H1410–H1418.[Abstract/Free Full Text]
Response
Bruce S. Stambler, MD
West Roxbury VA Medical Center Harvard Medical School,
West Roxbury, Mass
Kenneth A. Ellenbogen, MD
Medical College of Virginia McGuire VA Medical Center,
Richmond, Va
In response to the comments offered by Ravelli, our
study1 applied power spectral analysis
techniques to characterize the oscillations in atrial
flutter cycle length variability in both the time and frequency
domains. Previous attempts2 3 4 to investigate the
mechanisms governing atrial flutter cycle length variability have only
used time-domain methods and importantly have concluded that the
beat-to-beat variability in atrial flutter cycle length could be
explained entirely by an effect of the ventricular rate
mediated via atrial stretch. Thus, previous studies had failed to
consider the role of other potential control mechanisms, such as
respiration and the autonomic nervous system.
In our study, DC filtering was used to remove the DC component of the
data before power spectral analysis was performed. A detailed
analysis of 3 frequency bands in the spectra was performed
(band 1, <0.18 Hz; band 2, 0.18 to 0.60 Hz; and band 3, 0.6 to 2.2 Hz)
at baseline and in response to a variety of interventions
(isoproterenol, long-term ß-adrenergic blockade, edrophonium, heart
transplantation, and changes in respiratory rate).
The findings of our study provided additional confirmatory evidence of
the importance of the ventricular contraction in mediating
atrial flutter cycle length variability but also suggested that other
mechanisms may be involved. We showed that band 3, which accounted for
>60% of total spectral power, was significantly correlated with the
ventricular rate. However, the lower-frequency bands (ie,
<0.6 Hz) were not related to or altered by changes in the
ventricular rate. Furthermore, we also demonstrated that
band 2, which accounted for on average 
25% of total spectral power,
appeared to be related to respiration because it could be modified in a
similar direction as changes in respiratory rate (ie, increases in
respiratory rate shifted the band 2 frequency peak to higher
frequencies). Finally, we also found that on average 10% of total
spectral power was in frequencies <0.18 Hz. In about half the patients
studied, this was associated with a distinct spectral peak in these
very low frequencies, although in other patients no distinct peak was
observed at these frequencies. On the basis of several observations, we
concluded that although activity at frequencies <0.18 Hz was a
relatively minor component of atrial flutter cycle length variability,
it appeared to be related to changes in autonomic or sympathetic tone.
These observations included the following findings: (1) isoproterenol
altered the spectral characteristics of the band 1 component (ie,
increased the power in band 1); (2) patients receiving long-term
therapy with ß-adrenergic blockers had a similar atrial flutter cycle
length as patients not receiving ß-blockers but virtually absent band
1 spectral activity; and (3) heart transplant recipients had markedly
diminished atrial flutter cycle length variability.
Thus, in conclusion, our data are in agreement with Ravelli's comments
that atrial flutter cycle length variability is primarily modulated by
both the ventricular rate and respiration. However, we do
not believe that a potential role for autonomic nervous system effects
has been fully excluded. We hope and expect that future studies will
continue to examine this issue using the methodologies that we have
newly applied to analysis of atrial flutter.
References
1.
Stambler BS, Ellenbogen KA. Elucidating the
mechanisms of atrial flutter cycle length variability using power
spectral analysis techniques. Circulation. 1996;94:2515–2525.
2.
Lammers WJEP, Ravelli F, Disertori M, Antolini R,
Furlanello F, Allessie MA. Variations in human atrial flutter cycle
length induced by ventricular beats: evidence of a
reentrant circuit with a partially excitable gap. J
Cardiovasc Electrophysiol. 1991;2:375–387.
3.
Waxman MB, Kirsh JA, Yao L, Cameron DA, Asta JA.
Slowing of the atrial flutter rate during 1:1
atrioventricular conduction in humans and dogs: an
effect mediated through atrial pressure and volume. J
Cardiovasc Electrophysiol. 1992;3:544–557.
4.
Ravelli F, Disertori M, Cozzi F, Antolini R, Allessie
MA. Ventricular beats induce variations in cycle length of
rapid (type II) atrial flutter in humans. Circulation. 1994;89:2107–2116.
