Download Different Vagal Modulation of the Sinoatrial Node and AV Node in

58
Heart Rate Variability
15
Different vagal modulation of the shoatrial node and
AV node in patients w i t h congestive heart failure
PETER KOWALLIK'.', MD, ROBERT F. GIWUR'.', Jr., PhD,
SUSANNE FLEISCHER', and MALl'E MEESMANN',', MD.
'Department of Medicine, Wiirzburg University, Germany.
'Department of Physiology, Cornell University, Ithaca,
NY, U.S.A., and Helmholtz Institute for the Study of
Complex Systems in Biology and Medicine, Wiirzburg,
Germany.
s-rY
1. We have previously shown that in healthy young
men autonomic control of the sinoatrial (SA) and AV
node may be independent during sleep. It is
conceivable, that this independence is lost in
patients with high sympathetic activity. This would
be in analogy to exercise in normal subjects, where
an increase in sinus rate is associated with a
shortening of the PR interval.
2. The aim of this study was to investigate whether
this independence of SA and AV nodal autonomic
modulation is maintained in patients with congestive
heart failure.
3. For analysis of heart rate variability (HRV) the
ECG was online digitized from 10 pm to 6 am in six
patients with congestive heart failure (EF<40%). The
onset of P-waves and QRS-complexes was recognized by
a computer algorithm with an accuracy of t1 ms. Power
spectra of PR intervals and PP intervals were
calculated for consecutive 256 second segments. The
power in the high frequency component (HF, 0.15-0.4
Hz) of PP intervals was used as an index of vagal
drive to the SA node. The vagal input to the AV node
was determined by the spectral power of the
corresponding PR intervals.
4. All patients showed the typical spectral peak in
the HF band, both in PP and PR. The power spectral
density of HF varied over time with different
patterns for PP and PR. The ratio of the HF power
derived from PP and PR was calculated for each
segment. This ratio was not constant, but showed a
distinct time course.
5. Congestive heart failure did not abolish the
independence of vagal modulation of SA and AV node,
as assessed by the HF power derived from PP and PR
intervals. Thus, the difference in vagal traffic to
the SA and AV node was maintained even in the setting
of high background sympathetic activity. Further
investigation is needed to analyze potential factors
responsible for this difference in patterns and the
clinical relevance of this finding.
Introduction
Several experimental studies have shown that the
autonomic nervous system (ANS) can have differential
effects on the electrophysiological properties of the
sinus (SA) node, the atrioventricular (AV) node and
ventricular muscle [l-71. This result is believed to
reflect both differences in the distribution of the
sympathetic and parasympathetic nerves to various
regions of the heart, as well as differences in the
sensitivity of each region to the effects of
autonomic neuromodulators [31 .
We have previously shown [El, that in healthy
young men autonomic control of the SA and AV node may
be independent during sleep. This was based on
changes in the PR interval during sleep, when
spontaneous, transient periods of heart rate
acceleration occurred [9,10]. Acceleration of the
heart rate during sleep could result from activation
of sympathetic tone, withdrawal of parasympathetic
tone, or a combination of the two. Secondary to the
direct positive drmotropic effects of sympathetic
stimulation or parasympathetic withdrawal on the AV
node [ll-131, PR interval should shorten. In some
instances, however, heart rate acceleration was
accompanied by prolongauon of the PR interval, even
after consideration of the passive recovery effects
on AV nodal conduction. Thus, ANS input to the SA
node was greater than that to the AV node in this
cases L14.151 .
It is known, that in patients with congestive
heart failure the sympathetic nervous system is 'mre
activated than in healthy subjects [161, and
parasympathetic tone was reduced. It is conceivable,
that this independence is lost in patients with high
sympathetic activity. This would be in analogy to
exercise in normal subjects, where an increase in
sinus rate is associated with a shortening of the PR
interval. We therefore investigated whether the
independent autonomic modulation of the SA and AV
nodes was preserved in patients with congestive heart
failure.
Methods
ECG recording:
The study population was comprised of 6 patients
(age 27 to 67 years) with congestive heart failure.
All patients had an ejection fraction below 40% (mean
34%), and the classification according to NYHA was
11-111. All patients were informed about the study
protocol and gave their verbal consent.
ECG-recording was started around 10 pm and lasted
eight hours, during which the patients kept strict
bed rest. Three orthogonal ECG leads X, Y, and Z were
recorded continuously usin? the PREDICTOR@I system
(Dr. Kaiser Medizintechnik GmbH, Bad Hersfeld,
Germany). The filter settings were 0.05 Hz for highpass and 300 Hz for low-pass. The signals were
digitized on-line with 16 bit accuracy at a sampling
rate of 500 Hz/ channel and stored on hard disk.
Automatic ECG analysis:
Analysis was performed off-line. The onset of QRScomplexes and P-waves were identified automatically
using a computer algorithm, as described previously
[El.
QRS-complex: For an approximate detection of the
QRS onset a high pass filtered version of one ECG
channel was scanned. The exact position of each QRScomplex was determined by calculation of autocorrelations. These calculations were performed in
the vicinity of the approximate QRS onset and a
visually defined QRS-pattern to determine the time of
the best match.
P-wave: A time window of 200 m s , starting 360 ms
prior to each QRS-complex was scanned for the P-wave.
For that purpose a 100 ms ECG segment including the
P-wave was selected as the pattern for P-wave
recognition. Seventeen typical ECG segments were
averaged and that pattern was used for autocorrelation thereafter. There was a clear peak of the
autocorrelation in each window allowing for precise
measurement of P-wave onset. Thus, the resolution of
the onset of the P-wave and the QRS-complex was flms.
Consequently, the resolution of the intervals was
t2ms. Because of movement artifacts or noise in the
ECG recording the onset of P-waves, that could not be
determined was between 0.1% and 1.75% per subject.
PR and PP interval calculation:
PR and PP intervals were visually checked, if the
automatic recognition had failed or there pronounced
beat to beat changes. In addition, randcm samples
were taken. In case of artifacts the time of onset
was manually corrected. In total, all QRS-complexes
and 99.12% of the P-waves were available for
analysis.
Heart rate variability analysis:
Because of fluctuations in the autonomic input,
the total recording time of 8 hours was divided into
Heart Rate Variability
consecutive segments of 256 seconds for each patient.
For frequency domain analysis of HRV an instantaneous
heart rate signal was constructed from PP intervals
according to Berger et al. [171. For PR analysis, the
instantaneous PR signal was constructed from linear
interpolation of the PR intervals timed by the
corresponding PP intervals. Linear trends in the data
of each segment were subtracted. Gaps in the interval
series were filled according to a linear spline
interpolation. Resampling with 4 Hz resulted in 1024
data points for each segment. A fast fourier
transform was computed by a program written in
Borland C for OS/2" according to Numerical Recipes
s
'
, was
in C 1181. The HF power, expressed as lo-'
integrated between 0.15 Hz - 0.4 Hz.
Data analysis:
Statistical analysis was performed using c m e r cia1 software (Systat", SYSTAT Inc., hranston, IL,
USA, and Mathematica" for OS/2. Wolfram Research,
Inc). To test the hypothesis that PR depends on PP,
for each 256-second-segment a linear regression
analysis (least-squaresfit) was performed according
to
(1).
P,R, = a, + a,*(P,P,,,)
Here, P, and R,, denote the onset of the P wave and the
QRS complex of the nch beat. The ensuing beat is
indexed by n+l, and the previous beat by n-1.
Changes in PR interval correspond to changes in AV
nodal conduction time, which reflects both the raterelated conduction properties of the AV node and its
neural modulation. To differentiate this modulation
of AV nodal conduction by the ANS from the passive
conduction properties, the equation
PnR,,= b, + b,*(P,P,.,) + b,/(R,,.,P,)
(2)
was used following Leffler et al. (191. In this
model, a hyperbolic term (see last term of equation
2 ) accounts for the rate-related properties of the AV
node. The coefficient a, (equation 1) represents the
overall dependence of PR on PP, whereas b, (equation
2) reflects the neural modulation after correction
for the rate-related effects. A two-tailed t-test was
performed to test the coefficients to be different
from zero. A p-value of less than 0.05 was considered
significant.
1
0.5 -0.4
0'3
0.2
0.1
00
-0.1
iI
0 '
I -
ooo
-
Results
Dependence of PR on PP under consideration Of A"
nodal recovery effects
For all patients, the linear dependence of PR On
PP was calculated according to equation 1 for each
256-second segment. Shortening of PR with decreasing
PP resulted in a positive slope of the linear
regression line, i.e. a,>O. This was the case in most
of the analyzed segments of all patients, as depicted
for patient #2 in Fig. 1. Nevertheless, in some
segments PR increased with shortening PP (a,eO,
circles in the left part of Fig. 1).
To account for the passive properties of the AV
node, the dependence of PR on PP and RP was
calculated according to equation 2. The slopes a, of
the simple linear regression (equation 1) were
compared with the slopes b, resulting from equation 2
for all segments. For patient #2, the relation
between the slopes is depicted in Fig. 1. If a, was
positive, then b, was usually positive, too, i.e. as
Pig. 2: Time course of €IF spectral p e r of PP
intervals. The power in the high frequency (HF) band
of spectral analysis derived from PP intervals was
calculated for all consecutive 256 second segments of
one patient ( # 2 ) . The variation in HF power is
clearly visible within the 8 hourrecoang period.
PP decreases, PR as well as recovery-adjusted PR
decreases. This means that the autonomic effects on
AV nodal conduction- time exceeded the recovery
properties of the AV node. If a, was negative, the PP
decrease was associated with an increase of PR
(circles in the left quadrants of Fig. 1). In this
case, b, was always positive (circles in the upper
left quadrant of Fig. 1). i.e. recovery-adjusted PR
I
0
O
59
O
0
'8"
I
tc
--
-0.1 0.0 0.1' 0.2
0.3
0.4
0.5
a1
Fig. 1: Comparison of slopes from the linear
regression of PR or recwery adjusted PR on PP. For
all segments the simple linear dependence of PR on PP
(PR=a,+a,*PP) and the dependence of recovery adjusted
PR interval on PP (PR=b,+b,*PP+b,/RP) was calculated
(patient #2). The slopes a, and b, resulting from both
equations were compared, if the p-values were less
than 0.05 for a, and b,. A free space along the
horizontal zero-axis resulted, as a, was not
significantly different from zero in this range.
.
oL_L__L_L-a
0
20
10
;
I
'
60
80
100
120
Scgrwnl In1
Fig. 3 : Time course of €IF spectral power of PR
intervals. The power in the high frequency (HF) band
of spectral analysis derived from PR intervals was
calculated for all consecutive 256 second segments
(same patient as in Fig. 2). Note the fluctuations in
HF wwer over time.
Heart Rate Variability
always decreased. The sometimes observed lengthening
of the PR interval was dominated by the passive
properties of the AV node (a, negative and b,
positive) . Pronounced differences in autonomic modulation of SA and AV node were not found, which would
be expected in the case of recovery adjusted PR interval prolongation during PP increase (missing
circles in the left lower quadrant).
Analysis of heart rate variability
Spectral analysis was used to determine the
activity of the parasympathetic tone on the SA and AV
node to analyze subtle fluctuations in the autonomic
tone during the 8 hour recording period. All patients
showed the typical spectral peak in the HF band in PP
and PR. As shown in Fig. 2 , the power spectral
density of HF derived from PP intervals varied over
time.This variation occurred in all patients. Within
the recording period 3 to 6 local peaks in HF could
be detected in addition to some random scatter.
Comparing the patients, these local peaks occurred at
different times.
Spectral analysis of PR intervals also showed
L
L
V
0
20
40
00
F3
100
170
Segnnent In1
Fig. 4: Time course of the ratio of PR and PP high
frequency power. The ratio was calculated of the high
frequency power (HF) derived from PP interval and PR
interval analysis (same patient as in Figs. 2 and 3 ) .
peaks in the HF band. Similar to PP analysis, for
each patient the HF power varied over time within the
recording period (Fig. 3). The total power as well as
the power in the HF band of PR intervals was,
however, much smaller than that of the PP intervals.
Again, within this variation of HF power local peaks
could be observed.
To look for a dependence between the changes in HF
power from PP and PR interval analysis, the ratio was
calculated for each segment. This ratio showed a
distinct time course (Fig. 4 ) for all patients.
Discussion
There are two major findings of this study. In
contrast to healthy subjects during sleep [ E l , sinus
node automaticity and AV node conduction time are
always positively correlated with one another in
patients with congestive heart failure. Thus,
increases in heart rate are associated with decreases
in recovery adjusted PR interval. Despite of this
similar response, the strength of the parasympathetic
tone on SA and AV nodes appears to be independent of
one another, as assessed from the differing fluctuations in the HF power of PP and PR intervals over
time.
Dependence of PR on PP under consideration of AV
nodal recovery effects
For healthy young men we have shown, that during
the transient shortening of PP intervals, the PR
intervals did not show a uniform behavior [ E l . The PR
intervals either shortened, remained unchanged or
were prolonged. These changes in PR interval
correspond to changes in AV nodal conduction time,
which reflects both the passive properties of the AV
node and the modulation of the AV nodal conduction by
the ANS. With respect to the passive properties of
the AV node, shorter PP interval inputs to the AV
node may encroach on AV nodal refractoriness and
thereby lengthen the PR interval, i.e. there is an
inverse relation between the RP and PR intervals
[19,20].
This effect can be counteracted by a
concomitant activation of sympathetic input or
withdrawal of parasympathetic input to the AV node.
The net effect on PR interval depends on the relative
strength of these forces.
Given the above, shortening of the PR interval
during shortening of the PP interval (a,>O, circles
in the right quadrants of Fig. 1) indicates parallel
increase in sympathetic input and/or decrease in
vagal input to the SA and AV nodes. Similarly, an
unchanged PR interval during shortening of PP
interval also indicates coactivation of the SA and AV
nodes by the ANS, with the expected rate-dependent
slowing of AN nodal conduction being offset by
sympathetic activation and vagal withdrawal.
The lengthening of the PR interval during
shortening of the PP interval (a,cO, circles in the
left quadrants of Fig. 1) could result either from
rate-dependent slowing of AV nodal conduction or froni
markedly dissimilar sympathetic and parasympathetic
inputs to the SA and AV nodes. To differentiate
between these two possibilities, we corrected for the
potential effects of the Rp interval on the PR
interval using equation 2 [191.
The corrected PR interval was positively correlated with the PP interval in all cases (bl>O,
circles in the upper left quadrant of Fig. 1 1 , which
indicates parallel ANS inputs to the SA and AV nodes.
This was contrary to our findings in healthy young
men, where in some instances the corrected PR
interval was negatively correlated with the PP
interval (b,cO, which would result in circles in the
lower left quadrant of Fig. 1 ) . This result suggests
that the presumed activation of sympathetic tone
and/or withdrawal of parasympathetic tone in the SA
node associated with the increase in heart rate was
accompanied by similar or only slightly different
combination of sympathetic and parasympathetic tone
in the AV node. The decrease in independent autonomic
modulation of SA and AV node in patients with
congestive heart failure might be caused by the
relatively higher sympathetic tone and reduced
parasympathetic tone. Similarly, Leffler et al. 1191
found only a positive relationship between the
recovery-adjusted PR interval and the PP interval in
healthy subjects, which were asked to breath in
synchrony to a metronome. This task might have caused
a higher background sympathetic activation through
mental stress, too (211.
Analysis of HRV
Usually, the autonomic tone is estimated by
frequency domain analysis of HRV. The power in the HF
band of spectral analysis is considered to reflect
the parasympathetic tone f 2 2 . 2 3 1 . It is well known,
that the amplitude of the spectral components may
vary with time. This was shown for 5 minute segments
within a period of one hour 1241 and within 24 hours
[251
as well as for repeated 24 hour holter
recordings 7 days apart 1261. In concordance with
these studies, HF power of PP intervals varied over
time in the present study. Such changes would not be
unexpected, given the known complexity and lability
of ANS activity during sleep [27,281. Interestingly,
these fluctuations can also be found in the HF power
of PR intervals. If strength of parasympathetic tone
to the SA node closely paralleled that to the AV
node, corresponding fluctuations of PP and PR derived
HF power would be expected. However, this was not the
case, since the local maxima and minima of HF power
occurred at different times for PP and PR. Given the
Heart Rate Variability
power in the RF band as a quantitative measure of
parasympathetic tone, the strength of vagal drive on
SA and AV node is different, This allows for a more
detailed picture of the autonomic drive to the heart.
Acknowledgments
This study was supported by the Wilhelm SanderStiftung, Miinchen, Germany. We wish to thank Rainer
Scharf, PhD, for help in statistical analysis.
Address for correspondence:
Dr. Peter Kowallik
Medizinische Universitatsklinik Wiirzburg
Josef-Schneider-StraBe2
D-97080 Wiirzburg, Germany
Phone: +(49)-931-201-2774
Fax:
+ ( 4 9 ) -931-201-2291
e-mail: medk280~3rzbox.uni-wuerzburg.de
61
actions on SA and AV nodal function. British Journal
of Pharmacology 1990; 30: 490-492.
13. Atterh6g JH, Locgna E. Negative nonlinear model
relating heart rate and PR. Journal of Electrocardiology 1977; lO(4): 331-336.
14. Lehmann MH, Denker S, Mahmud R, Akhtar M.
Patterns of human atrioventricular nodal accomodation to a sudden acceleration of atrial rate.
American Journal of Cardiology 1984; 53: 71-76.
15. Wiener I, Kunkes S, Rubin D, Kupersmith J, Packer
M, Pitchon R. Schweitzer P. Effects of sudden change
in cycle length on human atrial, atrioventricular
nodal and ventricular refractory periods. Circulation
1981; 64: 245-248.
16. Meredith IT, Eisenhofer G, Lambert GW, Dewar EM,
Jennings GL, Esler MD. Cardiac sympathetic nerve
activity in congestive heart failure. Evidence for
increased neuronal norepinephring release and
preserved neuronal uptake. Circulation 1993; 88:
References
1. Nayebpour M. Talajic M. Nattel S . Effects of
P-adrenergic receptor stimulation and blockade on
rate-dependent atrioventricular nodal properties.
Circulation Research 1992; 70: 902-911.
2. Warner MR, DeTarnowsky JM, Whitson CC, Loeb JM.
Beat-by-beat modulation of AV conduction 11.
Autonomic neural mechanisms. American Journal of
Physiology 1986; 251: H1134-H1142.
3. Inoue H, Zipes DP. Changes in atrial and
ventricular refractoriness and in atrioventricular
nodal conduction produced by combinations of vagal
and sympathetic stimulation that result in a constant
spontaneous sinus cycle length. Circulation Research
136-145.
17. Berger RD, Akselrod S, Gordon D, Cohen RJ.
1987; 60: 942-951.
4. Schuessler RB, Boineau JP, Wylds AC, Hill DA,
Wu D, Denes P, Rosen I(,
Arzbaecher R. A quantitative description of normal AV
nodal conduction curve in man. Journal of Applied
Physiology 1976; 40: 74-78.
21. Jiang W, Hayano J, Coleman ER, et al. Relation of
the cardiocascular responses to mental stress and
cardiac vagal activity in coronary artery disease.
American Journal of Cardiology 1993; 72: 551-554.
22. Akselrod S, Gordon D, Ubel FA, Shannon DC, Barger
AC, Cohen RJ. Power spectrum analysis of heart rate
fluctuation: a quantitative probe of beat-to-beat
cardiovascular control. Science 1981; 213: 220-222.
23. Hayano J , Sakakibara Y, Yamada A, et al. Accuracy
of assessment of cardiac vagal tone by heart rate
variability in normal subjects. American Journal of
Cardiology 1991; 67: 199-204.
24. Brown DR, Randall DC, Knapp CF, Lee KC, Yingling
JD. Stability of the heart rate power spectrum over
time in the conscious dog. Faseb Journal 1989; 3:
Miller CB, Roeske WR. Effect of canine cardiac nerves
on heart rate, rhythm, and pacemaker location.
American Journal of Physiology 1986; 250: H630-H644.
5. Urthaler F, Neely BH, Hageman GR, Smith LR.
Differential sympathetic-parasympatheticinteractions
in sinus node and AV junction. American Journal of
Physiology 1986; 250: H43-H51.
6. Salata JJ, Gill RM, Gilmour RF. Jr., Zipes DP.
Effects of sympathetic tone on vagally induced phasic
changes in heart rate and atrioventricular node
conduction in the anesthetized dog. Circulation
Research 1986; 38: 584-5Y4.
7. Thomas J, Randall WC. Autonomic influences on
atrioventricular conduction in conscious dogs.
American Journal of.Physiology 1983; 244: H102-H108.
8. Kowallik P, Meesmann M. Independent autonomic
modulation of the human sinus and AV node: evidence
from beat-to-beat measurements of PR- and PP-intervals during sleep. J Cardiovasc Electrophysiol 1995;
in press:
9. Alihanka J. Sleep movements and associated
autonomic nervous activities in young male adults.
Acta Physiologica Scandinavicae 1982; 511 (Suppl):
1-85.
10. Viitasalo M, Halonen L, Partinen M, Iivanainen M,
Kala R, Eisalo A. Sleep and cardiac rhythm in healthy
men. Annals of Medicine 1991; 23: 135-139.
11. Luceri RM, Brownstein SL, Vardeman L, Goldstein
S. PR interval behavior during exercise: implications
for physiological pacemakers. Pacing and Clinical
Electrophysiology 1990; 13: 1719-1723.
12. Danter WR, Carruthers SG. The heart rate-PR
interval relationship: A model for evaluating drug
An
efficient algorithm for spectral analysis of heart
rate variability. IEEE Transactions on Biomedical
Engineering 1986; 33: 900-904.
18. Press WH, Flannery BP, Teukolsky SA, Vetterling
WT. Fourier and spectral applications. In: Numerical
recipes in C. The art of scientific computing.
Cambridge: Cambridge University Press, 1992: 537-606.
19. Leffler CT, Saul JP, Cohen RJ. Rate-related and
autonomic effects on atrioventricular conduction
assessed through beat-to beat PR interval and cycle
length variability. J Cardiovasc Electrophysiol 1994;
5: 2-15.
20. Teague S, Collins S,
1644-1650.
25. Saul JP.
Albrecht P, Berger RD, Cohen RJ.
Analysis of long term heart rate variability:
methods, l/f scaling and implications. Comp Cardiol
1988; 14: 419-422.
26. Klingenheben T, Zabel M, Just H, Hohnloser SH.
Reproduzierbarkeit von HerzfrequenzvariabilitatsMessungen in wiederholten 24-Stunden-Langzeit-EKGAufzeichnungen. Zeitschrift far Kardiologie 1993; 82:
302-308.
Mancia G. Autonomic modulation of the cardiovascular system during sleep. The New England Journal
of Medicine 1993; 328: 347-349.
28. Baccelli G, Guazzi M, Mancia G, Zanchetti A.
Neural and nonneural mechanisms influencing circulation during sleep. Nature 1969; 223: 184-185.
27.