Vol. 277, Issue 6, H2129-H2135, December 1999
Dynamic counterbalance between direct and indirect vagal
controls of atrioventricular conduction in cats
Shi-Liang
Chen,
Toru
Kawada,
Masashi
Inagaki,
Toshiaki
Shishido,
Hiroshi
Miyano,
Takayuki
Sato,
Masaru
Sugimachi,
Hiroshi
Takaki, and
Kenji
Sunagawa
Department of Cardiovascular Dynamics, National Cardiovascular
Center Research Institute, Suita, Osaka 565-8565, Japan
 |
ABSTRACT |
The vagal system regulates the
atrioventricular conduction time
(TAV) via two
opposing mechanisms: a direct effect on the atrioventricular node and
an indirect effect through changes in heart period
(TAA). To
evaluate how dynamic vagal activation affects TAV, we
stimulated the vagal nerve with frequency-modulated Gaussian white
noise and estimated the transfer function from vagal stimulation to the
TAV response
under conditions of no pacing and constant pacing in anesthetized cats.
The effect of changes in
TAA on
TAV was estimated
by a random-pacing protocol. The transfer function from vagal
stimulation to
TAV has low-pass
filter characteristics. Constant pacing increased the maximum step
response in TAV
(2.4 ± 1.2 vs. 6.3 ± 2.2 ms/Hz,
P < 0.01). The time constant did not differ between the vagal effect on
TAV and that on
TAA (2.9 ± 1.2 vs. 2.3 ± 0.5 s). Because changes in
TAA reciprocally
affected TAV
without significant delay, the direct and indirect effects were
dynamically counterbalanced and exerted stable
TAV transient response during vagal stimulation under normal sinus rhythm.
systems analysis; Gaussian white noise; interaction; atrioventricular node; transfer function
| |
INTRODUCTION |
THE VAGAL SYSTEM EXERTS negative effects
not only on the sinus node but also on the atrioventricular (AV) node.
Vagal stimulation prolongs the AV conduction time (A-V interval) even
when its negative chronotropic effect on the sinus node is prevented
through constant atrial pacing. On the other hand, prolongation of the
heart period (A-A interval) shortens the A-V interval through
electrophysiological mechanisms. Thus, during normal sinus rhythm,
vagal stimulation affects the A-V interval via two opposing mechanisms
(Fig. 1): prolonging the A-V interval by directly
affecting the AV node and indirectly shortening the A-V interval by
prolonging the A-A interval.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 1.
Schema of vagal control of A-V interval
(TAV). Vagal
stimulation affects
TAV both directly
and indirectly. Indirect effect consists of chronotropic effect of
vagal stimulation on A-A interval
(TAA) and
electrophysiological effect of
TAA changes on
TAV. Thick arrows
indicate that vagal stimulation prolongs
TAA and
TAV; thin arrow
indicates that prolongation of
TAA shortens
TAV.
|
|
Although numerous investigations on the vagal control of the A-V
interval have been conducted, including investigation of the
interactions between the direct and indirect mechanisms (15, 16, 22),
the way in which dynamic vagal activation affects the A-V interval
remains to be elucidated. Because vagal nerve activity changes
dynamically in response to environmental stresses, quantification of
the A-V interval response to dynamic vagal stimulation is essential for
understanding dynamic A-V interval regulation by the vagal system.
Leffler et al. (9) investigated dynamic characteristics of autonomic
and heart rate effects on P-R interval in humans using spectral
analysis and transfer function analysis. However, the transfer function
from vagal nerve activity to P-R interval was not measured, because
recording of vagal nerve activity was unavailable in the human study.
An animal experiment is mandatory to complement the human study and to
characterize the direct and indirect effects of dynamic vagal
stimulation on the A-V interval. To this end, we applied transfer
function analysis using frequency-modulated band-limited Gaussian white
noise stimulation (1, 6, 7, 14, 18) in anesthetized cats. The results
indicated that the direct and indirect effects were dynamically
counterbalanced and exerted stable transient A-V interval response
during vagal stimulation.
| |
MATERIALS AND METHODS |
Animal preparation.
Animals were cared for in accordance with the guidelines set by the
Physiological Society of Japan. Seven adult cats of either sex weighing
2.0-3.5 kg were anesthetized with pentobarbital sodium (30-35
mg/kg ip). Supplemental doses of pentobarbital sodium were injected (2 mg/kg iv) as necessary to maintain an appropriate level of anesthesia.
The animals were intubated and mechanically ventilated with room air
mixed with oxygen. Body temperature was maintained at ~37°C with
a heating pad and a heat lamp.
The bilateral vagal nerves were sectioned through a midline cervical
incision. We attached a pair of bipolar platinum electrodes to the
cardiac end of each vagal nerve for electrical stimulation. To prevent
drying and to provide insulation, the stimulation electrodes and nerves
were soaked in a mixture of white petrolatum (Vaseline) and liquid
paraffin. The chest was opened transversely at the second intercostal
space. The bilateral stellate ganglia were sectioned to eliminate
sympathetic nerve activity to the heart. We sutured a pair of bipolar
stainless electrodes to the right atrium through the right fifth
intercostal space to permit pacing. The femoral artery and the femoral
vein were cannulated for monitoring of arterial pressure, administering
anesthetics, and maintaining fluid balance. A 5-Fr bipolar
electrode catheter was introduced in a retrograde manner via the right
common carotid artery to the noncoronary cusp of the aortic valve to
record activity from the His bundle (12). The His bundle
electrogram was band-pass filtered in the frequency range of
50-1,000 Hz.
Stimulation protocols.
The study consisted of the following protocols. In
protocol 1, to estimate the transfer
function from vagal stimulation to the A-V and A-A intervals, we
dynamically stimulated the vagal nerves for 10 min. The stimulation
frequency was switched every second according to Gaussian white noise,
with a central frequency of 5 Hz and a standard deviation of 2 Hz. The
pulse duration was set at 2 ms. The baseline heart rate was 155 ± 16 beats/min, and the amplitude of vagal stimulation (4-8 V) was
adjusted in each animal to decrease heart rate to ~100 beats/min
under 10-Hz tonic vagal stimulation. The efficacy of vagal stimulation
on the heart rate did not change >10% throughout the experiment.
Dynamic vagal stimulation was repeated under conditions of constant
atrial pacing to abolish the influence of the A-A interval changes on
the A-V interval. The pacing rate (158 ± 17 beats/min) was set to a
level slightly higher than the baseline heart rate. In
protocol
2, to identify the transfer function
from the A-A interval to the A-V interval, we randomly paced the heart
for 10 min according to Gaussian white noise with a pacing interval
distribution of 300 ± 50 (SD) ms in the absence of vagal
stimulation. The pacing interval was bounded in the range of
200-400 ms to allow the heart to capture every pacing stimulus.
We used different sequences of Gaussian white noise for different
animals. The order of the experimental runs was randomized to reduce
the likelihood of bias or of systematic error in our identification
approach. The stimulation (or pacing) signal and the His bundle
electrogram were digitized at 2,000 Hz through a 12-bit
analog-to-digital converter and stored on the hard disk of a dedicated
laboratory computer system (NEC PC-9801FA).
Data analysis.
AV conduction time was assessed from the His bundle electrogram. The
A-H interval was defined as the time from the earliest deflection of
the atrial wave to the peak of the His potential. We used a maximum
negative point of first derivative of the signal (dV/dt)
in the earliest deflection as the A wave position. The H-V
interval was defined as the time from the His potential to the earliest
deflection of the ventricular wave. We used a time point of maximum
negative dV/dt
of the earliest deflection as the V wave position. We manually selected
the templates for A, H, and V waves and subsequently detected matched
signals through an adaptive template-matching algorithm. Discrete time
series data were then obtained from the measured A-A, A-V, A-H, and H-V intervals so as to follow the principle of causality, e.g., the A-A
interval values were positioned at the time points of each corresponding second A wave and the A-V interval values were positioned at the time points of each corresponding V wave. Finally, we linearly interpolated the respective discrete time series data to a frequency of
8 Hz. In two of seven animals, we could not record the His potential.
However, because the other five animals did not show significant H-V
interval changes under any of the protocols in the present study, we
represented AV conduction time by the A-V interval rather than the A-H
interval and pooled the results from all seven animals.
The transfer functions representing the dynamic system characteristics
were calculated as follows. We segmented the 8-Hz resampled input-output data pairs into six 50%-overlapping bins of 1,024 points
each (20). For each bin, a linear trend was removed and a Hanning
window was applied. We then performed fast Fourier transformation (2)
to obtain the frequency spectra of the input
X(f)
and output Y(f).
Next, we ensemble averaged, over the six bins, the power of the input,
SXX(f),
the power of the output,
SYY(f),
and the cross power between the two,
SYX(f).
Finally, we estimated the transfer function,
H(f),
using the following equation (14)
We also calculated a magnitude-squared coherence function. The
coherence function, Coh(f), is
a frequency-domain measure of linear dependence between the input and
output signals. It was calculated according to the following equation
(14)
In protocol 1, to estimate the
transfer function from vagal stimulation to the A-A interval
(HV-AA), we
treated the vagal stimulation frequency as the input and the A-A
interval as the output of the system. When estimating the transfer
function from vagal stimulation to the A-V interval
(HV-AV), we
treated the A-V interval as the output. Because we switched the vagal
stimulation frequency every second, the input power spectrum was fairly
constant up to 0.5 Hz, decreased to 1/10 by ~0.8 Hz, and dropped to
noise levels at 1 Hz. We displayed
HV-AA and
HV-AV only in the
frequencies <0.8 Hz, because the estimation of the transfer functions
was unreliable in the higher frequency range because of a lack of input
power. We estimated
HV-AV during
normal sinus rhythm
(HV-AV,N) and
constant atrial pacing
(HV-AV,P).
HV-AA represents
the vagally mediated chronotropic effect on the sinus node.
HV-AV,P
represents the direct effect of dynamic vagal stimulation on the A-V
interval. HV-AV,N
represents the total effect (i.e., the combined direct and indirect
effects) of dynamic vagal stimulation on the A-V interval.
In protocol
2, to estimate the transfer function
from the A-A interval to the A-V interval
(HAA-AV), we
treated the A-A interval as the input and the A-V interval as the
output of the system. We changed every A-A interval randomly, with a
mean value of 300 ms. This yielded a relatively flat input power
spectrum beyond 1 Hz. We displayed
HAA-AV in the
frequencies up to 1 Hz.
HAA-AV represents
the dynamic characteristics of the electrophysiological effect of
changes in the A-A interval on the A-V interval.
Statistical analysis.
To facilitate interpretation of the determined dynamic characteristics,
we calculated a 30-s step response associated with each transfer
function. We first applied inverse Fourier transformation to the
transfer function and obtained an impulse response. The step response
was obtained by a time integral of the impulse response.
To quantitatively examine the differences among
HV-AA,
HV-AV,N and
HV-AV,P, we
calculated a maximum response and a time constant in their
corresponding step responses by means of nonlinear least-squares fitting. We also calculated a 90%-rise time at which 90% of the maximum response was attained. We tested the differences in parameters by Friedman's test for repeated-measures nonparametric comparison. If
there was a significant difference, we applied the Student-Newman-Keuls test based on ranks to identify the difference between any two of the
three groups (3).
| |
RESULTS |
Figure 2A
shows representative time series data obtained from
protocol
1 with no pacing. The top panel shows
the vagal stimulation frequency according to the Gaussian white noise
command signal; the middle panel shows the associated A-A interval
response. Increasing vagal stimulation frequency prolonged the A-A
interval, whereas decreasing it shortened the A-A interval. The bottom
panel of Fig. 2A shows the A-V, A-H,
and H-V interval responses. Although changes in the A-V interval were
proportional to changes in the A-A interval, the magnitude of those
changes was much smaller. Because vagal stimulation did not perceivably
alter the H-V interval, the dynamic response of the A-H interval to
vagal stimulation was identical to that of the A-V interval.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 2.
Representative time series data showing vagal stimulation protocols
with no pacing (A) and constant
pacing (B) in 1 cat.
Top panels, vagal stimulation
frequency according to Gaussian white noise command signal.
Middle panels, A-A interval in
response to vagal stimulation. Bottom
panels, A-V, A-H, and H-V intervals in
response to vagal stimulation. A-H and H-V interval responses are shown
by thin lines. Both A-A and A-V intervals were prolonged by vagal
stimulation. Constant pacing enhanced A-V and A-H responses to vagal
stimulation. H-V interval was unchanged by vagal stimulation.
|
|
Illustrated in Fig. 2B are time series
data obtained from the same animal under constant atrial pacing. The
vagal stimulation frequency, the A-A interval at a constant pacing
interval, and the A-V, A-H, and H-V interval responses are shown in
Fig. 2B (top,
middle, and
bottom, respectively). Constant pacing
enhanced the A-V interval response to dynamic vagal stimulation. The
H-V interval remained unchanged. As a result, the dynamic response of
the A-H interval to vagal stimulation was identical to that of the A-V interval.
Figure 3A
shows the averaged transfer functions associated with vagal stimulation
pooled from all the animals.
HV-AA,
HV-AV,N, and
HV-AV,P are shown
in Fig. 3A
(left,
center, and
right, respectively). The top and
middle panels are the gain and phase plots of the transfer functions,
respectively, and the bottom panels are the coherence functions in Fig.
3A. The gain and phase plots reveal low-pass filter characteristics in all three transfer functions. The
phase approaches zero radians in the lowest frequency, reflecting the
fact that vagal stimulation prolongs both
TAA and
TAV. The coherence values are ~0.7 in the frequencies <0.4 Hz. The gain was
smaller in HV-AV
than in HV-AA.
The gain of HV-AV was smaller with
no pacing (Fig. 3A,
top center) than with constant
pacing (Fig. 3A, top
right).
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 3.
Transfer functions (A) and step
responses (B) associated with vagal
stimulation averaged from all experimental animals.
HV-AA, transfer
function from vagal stimulation to A-A interval;
HV-AV,N, transfer
function from vagal stimulation to A-V interval estimated under
conditions of normal sinus rhythm;
HV-AV,P, transfer
function from vagal stimulation to A-V interval estimated under
conditions of constant pacing; all transfer functions showed low-pass
characteristics with an in-phase relationship between input and output
signals. Coh, coherence function; Step Resp, step response associated
with each transfer function. Vertical axis for A-V interval response is
scaled only 1/8 that for A-A interval response. Solid and dashed lines
indicate mean and mean + SD values, respectively.
|
|
Figure 3B illustrates the calculated
step responses associated with
HV-AA,
HV-AV,N, and
HV-AV,P. The
maximum response was much smaller in the A-V interval (Fig.
3B,
right) than in the A-A interval
(Fig. 3B,
left). The maximum response in the
A-V interval was smaller under conditions of no pacing (Fig.
3B,
center) than constant pacing (Fig.
3B,
right).
Table 1 summarizes the parameters of step
responses shown in Fig. 3B.
Simultaneous comparison indicated that the maximum step responses were
significantly smaller in the A-V interval than in the A-A interval. The
maximum step response in the A-V interval was significantly smaller
when there was no atrial pacing. Although the time constant associated
with HV-AV,N
tended to be greater than that associated with
HV-AV,P or
HV-AA, there were no statistically
significant differences among the responses. The 90%-rise time
associated with
HV-AV,N was
significantly longer than that associated with
HV-AA but was not
significantly different from that associated with
HV-AV,P.
Figure 4 represents the time series data
obtained from protocol
2. The top panel in Fig. 4 shows
changes in the A-A interval according to the Gaussian white noise
command signal, and the bottom panel shows changes in the A-V, A-H, and
H-V intervals. Although the A-A and A-V intervals showed a reciprocal
relationship, the magnitude of the A-V interval change was about
one-tenth of that of the A-A interval change. The H-V interval remained
unchanged by the random pacing.
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 4.
Representative time series data showing random pacing protocol without
vagal stimulation in 1 cat. Top,
changes in A-A interval according to Gaussian white noise command
signal. Bottom, A-V, A-H, and H-V
intervals in response to changes in A-A interval. A-H and H-V interval
responses are shown by thin lines. A-V and A-H intervals changed
reciprocal to A-A interval; H-V interval was unaltered.
|
|
Figure 5A
shows the average
HAA-AV obtained
from all the animals. The gain plot
(top), the phase plot
(middle), and the coherence function
(bottom) are shown in Fig.
5A. The gain was 0.15 ± 0.02 at
0.008 Hz and decreased gradually to 0.09 ± 0.02 at 1 Hz. The phase
plot indicated an out-of-phase relationship between the A-A and A-V
intervals in the frequency range of 0.008 to 1.0 Hz in the absence of
vagal stimulation. In this frequency range, the coherence values were
>0.8.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 5.
Transfer function (A) and step
response (B) obtained from random
pacing protocol averaged from all experimental animals. A-A and A-V
intervals show out-of-phase relationship. Transfer function had
relatively frequency-independent characteristics. Solid and dashed
lines indicate mean and mean + SD values, respectively.
|
|
Figure 5B shows the step response of
the A-V interval to the A-A interval derived from
HAA-AV. The step
response reached >60% of the maximum negative response within a
second and then gradually approached the steady-state value of
0.13 ± 0.02 ms.
| |
DISCUSSION |
We have shown that the dynamic characteristics from vagal stimulation
to the A-V interval response approximated a low-pass filter. The time
constant of the A-V interval step response to vagal stimulation was not
significantly different from that of the A-A interval step response.
Constant atrial pacing increased the maximum response of the A-V
interval step response to vagal stimulation while not affecting the
time constant.
Transfer characteristics from vagal stimulation to A-A interval.
The transfer function representing the vagally mediated chronotropic
effect on the sinus node,
HV-AA, can be
described as a low-pass filter (Fig.
3A,
left). Taking the reciprocal
relationship between the A-A interval and heart rate into account,
these characteristics are comparable to those estimated for the
transfer function from vagal stimulation to heart rate in dogs (1) and
in rabbits (6, 7). The 90%-rise time of the A-A interval step response (Table 1) was, however, longer than that estimated in rabbits, suggesting species differences in the vagally mediated chronotropic effect on the sinus node. According to our previous investigation on
the determinants of the transfer function from vagal stimulation to
heart rate (18), one of the possible candidates for such species
differences is a level of acetylcholinesterase activity. Difference in
the effective vagal stimulation intensity among studies might also
affect a corner frequency of the transfer function (1) and thus the
90%-rise time of the corresponding step response.
Inconsistent with our previous results (6, 7, 18), the phase plot of
the present study did not show significant lag time between vagal
stimulation and the A-A interval response. Difference in input data
processing might have affected the estimation of the lag time. In the
present study, we measured the stimulation interval between two
consecutive pulses and positioned the stimulation frequency (an inverse
of the stimulation interval) at the time point of the second pulse, to
provide consistent data processing between stimulation signal and the
His bundle electrogram. Thus the stimulation signal used for
calculation of the transfer function was, on average, 200 ms (5 Hz)
behind the command signal. If the command signal were used for the
calculation of the transfer function as in our previous studies, the
lag time would be estimated as ~200 ms.
Transfer characteristics from vagal stimulation to A-V interval.
Similar to the transfer function from vagal stimulation to the A-A
interval response, the transfer function characterizing the direct
effect of dynamic vagal stimulation on the A-V interval, HV-AV,P, also
approximated characteristics of a low-pass filter (Fig. 3,
right). The step responses
associated with
HV-AA and HV-AV,P revealed
similar time constant and 90%-rise time (Fig. 3B,
left and
right; Table 1). Leffler et al. (9)
demonstrated similar distributions of power spectra of R-R and P-R
intervals using random respiration, suggesting similar dynamic
characteristics of autonomic control of R-R and P-R intervals. The
present study confirmed and extended their results in relation to the
vagal control of the A-A and A-V intervals. The similarities between the A-A and A-V interval responses suggest that the same kinds of
signal transduction sequences are involved in the negative effects of
vagal stimulation on the sinus node and on the A-V node.
Constant atrial pacing enhanced the dynamic A-V interval response to
vagal stimulation (Fig. 3A,
center and
right; Table 1). In other words, the
dynamic A-V interval response was markedly attenuated when the normal
sinus rhythm was maintained. Such dynamic characteristics are
consistent with the steady-state A-V interval response in a human study
in which baroreflex-induced A-V interval responses were smaller under
normal sinus rhythm than under constant pacing (13). The
A-V interval response is known to affect ventricular filling (17). A
longer A-V interval prematurely terminates left ventricular filling
through early mitral valve closure associated with atrial relaxation,
whereas a shorter A-V interval reduces the contribution of atrial
contraction to ventricular filling. The optimal A-V interval is within
the normal A-V interval in dogs (17). However, further experiments are
needed to clarify whether the attenuation of A-V interval changes
during normal sinus rhythm is more beneficial for overall cardiac
function, compared with what is attained by the combination of normal
sinus rhythm and the A-V interval changes observed during constant pacing.
Transfer characteristics from A-A interval to A-V interval.
The transfer function characterizing the effect of changes in the A-A
interval on the A-V interval,
HAA-AV, indicates
a reciprocal relationship between the A-A and A-V intervals in the
absence of autonomic nervous activity (Fig.
5A). The frequency-dependent decrease in gain was much less than that observed in
HV-AV,P. Thus the
electrophysiological effect of changes in the A-A interval on the A-V
interval is much quicker than the direct effect of vagal stimulation on
the A-V interval. The dynamic characteristics of the indirect effect of
vagal stimulation on the A-V interval are determined by a serial
connection of the chronotropic effect and the electrophysiological
effect (Fig. 1). Because the electrophysiological effect approximates
an inverter with an all-pass filter in the frequency range of the
present study, the low-pass characteristics of the indirect effect are
mainly determined by the chronotropic effect. Therefore, similar time
courses in the chronotropic effect (HV-AA) and the
direct effect
(HV-AV,P)
result in similar time courses in the direct and indirect effects of
dynamic vagal stimulation on the A-V interval. The indirect effect,
however, counterbalances the direct effect via the inverter-like
characteristics of the electrophysiological effect.
To explore the physiological meaning of the similarity in the time
course between the direct and indirect effects, we simulated the A-V
interval response to vagal stimulation while varying the time constant
for the direct and indirect effects. Figure
6A shows the simulation of the A-V interval step response to vagal stimulation with fixed time constant (3 s) for the direct effect
(
direct) while the time
constant for the indirect effect
(indirect) changed in the
range from one-eighth to eight times
direct. If
indirect is much shorter than
direct, the A-V interval
response becomes oscillatory. Figure
6B shows the A-V interval response
with fixed indirect (3 s) while
direct changed in the range
from one-eighth to eight times
indirect. If
direct is much shorter than
indirect, the A-V interval step
response becomes underdamped. Therefore, if the time constants for the
direct and indirect effects do not match, the resultant A-V interval
response to vagal stimulation becomes unstable. Under physiological
conditions, because the direct and indirect effects have similar time
courses, vagal stimulation manifests stable monophasic transient A-V
interval response (Fig. 6, A and
B, thick lines). Impairment of any one
of the direct effects, the chronotropic effect and the
electrophysiological effect (Fig. 1), would cause imbalance in the time
course between the direct and indirect effects of dynamic vagal
stimulation on the A-V interval, thereby leading to unstable A-V
interval regulation by the vagal system. Although outcomes of such an
imbalance between the direct and indirect vagal effects remain
speculative, Fig. 6 indicates that 10-Hz vagal stimulation would
prolong the A-V interval by >50 ms if the imbalance exists. Because
50 ms is >50% of the A-V interval under normal conditions, such a
prolongation could reduce the effectiveness of AV blood transfer by
20% (17). In addition, the A-V interval could show a variety of step
responses if the time courses of direct and indirect vagal effects do
not match. If the variation occurs among multiple pathways in the AV
node, including fast and slow pathways (8), it might provide substrates
for AV nodal reentrant tachycardia.
View larger version (46K):
[in this window]
[in a new window]
|
Fig. 6.
Simulated A-V interval response to unit change in vagal stimulation
frequency under a variety of combinations of time constants for direct
and indirect vagal effects. A: time
constant for direct effect
(direct) is fixed at 3 s.
Time constant for indirect effect
(indirect) varies from 1/8 to
8 times value of direct.  AV,
changes in A-V interval. B:
indirect is fixed at 3 s.
direct varies from 1/8 to 8 times value of indirect. Thick
line in each panel indicates direct = indirect = 3 s. -Axis is
reversed in B to provide a good view
of thick line.
|
|
Limitations.
Several variables that can potentially affect the A-V interval have not
been considered in the present study. First, we obtained all data from
animals under anesthesia. If data had been obtained from conscious
animals, the results might have been different. However, because we cut
efferent pathways of both vagal and sympathetic systems, the effects of
anesthesia on the central nervous system should have had little effect
on our conclusions.
Second, there was a difference in the operating range of the A-A
interval during the estimation of
HAA-AV,
HV-AA, and
HV-AV. Because
the effect of the A-A interval changes on the A-V interval is known to
be greater in shorter A-A interval ranges (19), the difference in the
operating range of the A-A interval would have affected the absolute
gain values of the estimated transfer functions.
Third, the A-V interval response to changes in the A-A interval depends
not only on the operating range of the A-A interval but also on the
rate and direction of heart rate change (12). However, the rate and
direction of the A-A interval change were considerably randomized in
protocol
2 through the pacing command of
Gaussian white noise. Therefore,
HAA-AV might represent average transfer characteristics from the A-A interval to the A-V interval.
Fourth, AV conduction is regulated both by the vagal system and by the
sympathetic system. For instance, the sympathetic effects may
predominate over the vagal effects during exercise. Although previous
studies (11, 21) demonstrated minimal interactions between the vagal
and sympathetic systems in regulating AV conduction, further studies
are clearly needed to characterize the dynamic interactions between the
vagal and sympathetic systems.
Finally, the stimulating pattern of Gaussian white noise is different
from the physiological discharge of vagal nerves (5). The asynchronous
nature of Gaussian white noise stimulation relative to each heartbeat
would mask a phase-dependent sensitivity of vagal effects on the A-A
and A-V intervals (4, 10). According to a study in dogs (10), in which
the baseline heart rate was 154 beats/min, steady-state heart rate
during vagal stimulation ranged from 79 to 95 beats/min depending on
the phase at which the vagal stimulation was applied. Thus the
phase-dependent sensitivity (95 79 = 16 beats/min) in
the steady-state heart rate response to vagal stimulation is ~30% of
the phase-independent response (154 95 = 59 beats/min). The
identified transfer functions in the present study would represent
average dynamic characteristics of the phase-independent vagal effects
on the A-A and A-V intervals. A somewhat different limitation persists,
however: because we stimulated the whole bundle of vagal nerve at once,
the response observed cannot account for the possible regional
differences in function of individual nerve fibers that make up the
nerve bundle.
In conclusion, the transfer function representing the direct effect of
dynamic vagal stimulation on the A-V interval approximated a low-pass
filter similar to that representing vagally mediated chronotropic
effect on the sinus node except for absolute gain values. Under normal
sinus rhythm, the direct effect of vagal stimulation on the A-V
interval is counterbalanced dynamically by the indirect effect through
changes in the A-A interval. The direct and indirect effects of vagal
stimulation had similar time constants, leading to a manifestation of
stable transient A-V interval response.
| |
ACKNOWLEDGEMENTS |
This study was supported by Research Grants for Cardiovascular
Diseases (6A-4, 7C-2, 7A-1, 9C-1) from the Ministry of Health and
Welfare of Japan, by a Grant from the Science and Technology Agency of
Japan, Encourage System of COE (Center of Excellence), by a Grant from
the Ministry of Health and Welfare of Japan, Research on Advanced
Medical Technology, and by a Grant from Ground-Based Research
Announcement for the Space Utilization promoted by NASDA (National
Space Development Agency of Japan) and Japan Space Forum.
| |
FOOTNOTES |
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement"
in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Address for reprint requests and other correspondence: T. Kawada, Dept.
of Cardiovascular Dynamics, The National Cardiovascular Center Research
Inst., 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan.
Received 7 May 1999; accepted in final form 23 July 1999.
| |
REFERENCES |
1.
Berger, R. D.,
J. P. Saul,
and
R. J. Cohen.
Transfer function analysis of autonomic regulation. I. Canine atrial rate response.
Am. J. Physiol.
256 (Heart Circ. Physiol. 25):
H142-H152,
1989[Abstract/Free Full Text].
2.
Brigham, E. O.
FFT transform applications.
In: The Fast Fourier Transform and Its Applications. Englewood Cliffs, NJ: Prentice-Hall, 1988, p. 167-203.
3.
Glantz, S. A.
Primer of Biostatistics (4th ed.). New York: McGraw Hill, 1997.
4.
Jalife, J.,
and
D. C. Michaels.
Phase-dependent interactions of cardiac pacemakers as mechanisms of control and synchronization in the heart.
In: Cardiac Electrophysiology and Arrhythmias, edited by D. P. Zipes,
and J. Jalife. New York: Grune and Stratton, 1985, p. 109-119.
5.
Katona, P. G.,
J. W. Poitras,
G. O. Barnett,
and
B. S. Terry.
Cardiac vagal efferent activity and heart period in the carotid sinus reflex.
Am. J. Physiol.
218:
1030-1037,
1970
6.
Kawada, T.,
Y. Ikeda,
M. Sugimachi,
T. Shishido,
O. Kawaguchi,
T. Yamazaki,
J. Alexander, Jr.,
and
K. Sunagawa.
Bidirectional augmentation of heart rate regulation by autonomic nervous system in rabbits.
Am. J. Physiol.
271 (Heart Circ. Physiol. 40):
H288-H295,
1996[Abstract/Free Full Text].
7.
Kawada, T.,
M. Sugimachi,
T. Shishido,
H. Miyano,
Y. Ikeda,
R. Yoshimura,
T. Sato,
H. Takaki,
J. Alexander, Jr.,
and
K. Sunagawa.
Dynamic vagosympathetic interaction augments heart rate response irrespective of stimulation patterns.
Am. J. Physiol.
272 (Heart Circ. Physiol. 41):
H2180-H2187,
1997[Abstract/Free Full Text].
8.
Leffler, C. T.,
and
J. P. Saul.
NAPSE Young investigator awardee
1993. Computer model of the atrioventricular node predicts reentrant arrhythmias.
Pacing Clin. Electrophysiol.
17:
113-130,
1994[Medline]
9.
Leffler, C. T.,
J. P. Saul,
and
R. J. Cohen.
Rate-related and autonomic effects on atrioventricular conduction assessed through beat-to-beat PR interval and cycle length variability.
J. Cardiovasc. Electrophysiol.
5:
2-15,
1994[Medline]
10.
Levy, M. N.,
P. J. Martin,
T. Iano,
and
H. Zieske.
Paradoxical effect of vagus nerve stimulation on heart rate in dogs.
Circ. Res.
25:
303-314,
1969[Abstract/Free Full Text].
11.
Levy, M. N.,
and
H. Zieske.
Autonomic control of cardiac pacemaker activity and atrioventricular transmission.
J. Appl. Physiol.
27:
465-470,
1969[Free Full Text].
12.
Loeb, J. M.,
J. M. Detarnowsky,
M. R. Warner,
and
C. C. Whitson.
Dynamic interactions between heart rate and atrioventricular conduction.
Am. J. Physiol.
249 (Heart Circ. Physiol. 18):
H505-H511,
1985.
13.
Mancia, G.,
O. Bonazzi,
L. Pozzoni,
A. Ferrari,
M. Gardumi,
L. Gregorini,
and
R. Perondi.
Baroreceptor control of atrioventricular conduction in man.
Circ. Res.
44:
752-758,
1979[Abstract/Free Full Text].
14.
Marmarelis, P. Z.,
and
V. Z. Marmarelis.
The white noise method in system identification.
In: Analysis of Physiological System. New York: Plenum, 1978, p. 131-221.
15.
Martin, P.
Paradoxical dynamic interaction of heart period and vagal activity on atrioventricular conduction in the dog.
Circ. Res.
40:
81-89,
1977[Abstract/Free Full Text].
16.
Martin, P.
The influence of the parasympathetic nervous system on atrioventricular conduction.
Circ. Res.
41:
593-599,
1977[Free Full Text].
17.
Meisner, J. S.,
D. M. McQueen,
Y. Ishida,
H. O. Vetter,
U. Bortolotti,
J. A. Strom,
R. W. M. Frater,
C. S. Peskin,
and
E. L. Yellin.
Effects of timing of atrial systole on LV filling and mitral valve closure: computer and dog studies.
Am. J. Physiol.
249 (Heart Circ. Physiol. 18):
H604-H619,
1985.
18.
Nakahara, T.,
T. Kawada,
M. Sugimachi,
H. Miyano,
T. Sato,
T. Shishido,
R. Yoshimura,
H. Miyashita,
and
K. Sunagawa.
Cholinesterase affects dynamic transduction properties from vagal stimulation to heart rate.
Am. J. Physiol.
275 (Regulatory Integrative Comp. Physiol. 44):
R541-R547,
1998[Abstract/Free Full Text].
19.
Nayebpour, M.,
M. Talajic,
C. Villemaire,
and
S. Nattel.
Vagal modulation of the rate-dependent properties of the atrioventricular node.
Circ. Res.
67:
1152-1166,
1990[Abstract/Free Full Text].
20.
Oppenheim, A. V.,
and
R. W. Schafer.
Power spectrum estimation.
In: Digital Signal Processing. Englewood Cliffs, NJ: Prentice-Hall, 1975, p. 532-574.
21.
Wallick, D. W.,
P. J. Martin,
Y. Masuda,
and
M. N. Levy.
Effects of autonomic activity and changes in heart rate on atrioventricular conduction.
Am. J. Physiol.
243 (Heart Circ. Physiol. 12):
H523-H527,
1982.
22.
Wallick, D. W.,
R. G. Xu,
and
P. J. Martin.
Dynamic interaction of vagal activity and heart rate on atrioventricular conduction.
Am. J. Physiol.
262 (Heart Circ. Physiol. 31):
H792-H798,
1992[Abstract/Free Full Text].
Am J Physiol Heart Circ Physiol 277(6):H2129-H2135
0002-9513/99 $5.00
Copyright © 1999 the American Physiological Society
TOP
ABSTRACT
INTRODUCTION
