Download Optimal ventricular rate slowing during atrial fibrillation - AJP

yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts

Coronary artery disease wikipedia, lookup

Management of acute coronary syndrome wikipedia, lookup

Heart failure wikipedia, lookup

Cardiac surgery wikipedia, lookup

Cardiac contractility modulation wikipedia, lookup

Hypertrophic cardiomyopathy wikipedia, lookup

Myocardial infarction wikipedia, lookup

Jatene procedure wikipedia, lookup

Electrocardiography wikipedia, lookup

Quantium Medical Cardiac Output wikipedia, lookup

Arrhythmogenic right ventricular dysplasia wikipedia, lookup

Ventricular fibrillation wikipedia, lookup

Atrial fibrillation wikipedia, lookup

Heart arrhythmia wikipedia, lookup

Am J Physiol Heart Circ Physiol 282: H1102–H1110, 2002.
First published November 23, 2001; 10.1152/ajpheart.00738.2001.
Optimal ventricular rate slowing during atrial fibrillation
by feedback AV nodal-selective vagal stimulation
Department of Cardiovascular Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio 44195
Received 17 August 2001; accepted in final form 19 November 2001
heart rhythm; arrhythmia; hemodynamics; autonomic control; atrioventricular node
ATRIAL FIBRILLATION (AF) has been long recognized as one
of the most frequent chronic arrhythmias. Although
the restoration and maintenance of normal sinus
rhythm is the ultimate goal, it is frequently not achievable. Therefore, ventricular rate (VR) control during
AF remains the only realistic long-term solution in a
majority of patients (27). Although a VR from 60 to 90
beats/min has been recommended for patients with AF
(45, 47), no specific experimental or clinical data are
available to support this recommendation (22).
The VR-hemodynamics relationship could be determined in several experimental models. In choosing the
one implemented in this study, we rejected those that
are either associated with marked changes of the nor-
Address for reprint requests and other correspondence: T. N.
Mazgalev, Research Institute FF1-02, Cleveland Clinic Foundation,
9500 Euclid Ave., Cleveland, OH 44195 (E-mail: [email protected]).
mal physiology, are inherently imprecise, or are difficult to implement.
First, the atrioventricular node (AVN) can be destroyed (ablated), followed by implantation of ventricular pacemaker. This approach results in retrograde
ventricular activation starting with depolarization of
the apex. On the positive side, the pacemaker permits
maintenance of any desired constant retrograde VR.
Because our scope was limited to physiologically normal anterograde conduction, this method was not investigated in the present study.
Second, the AVN conduction function could be depressed by various drugs. This approach preserves the
normal anterograde propagation. However, the use of
drugs to determine an optimal VR has limitations
because the decrease in the VR during treating of AF
occurs gradually and unpredictably (22). In addition,
most drugs have direct effect on cardiac contractility
Third, partial ablation (modification) of the slow
AVN pathway has been used to slow the VR. However,
a progressive attenuation of the rate by this method is
virtually impossible and the procedure frequently results in complete AVN block (10, 25).
A fourth, novel, approach has emerged recently. It is
well established that stimulation of parasympathetic
nerves, which selectively innervate the AVN, can slow
AV conduction (2, 6, 12, 13, 21, 28–30). In our recent in
vitro study, we found that vagally induced slowing of
the VR during AF is feasible by using postganglionic
endocardial nerve stimulation (23), whereas similar
effects were also obtained by utilizing epicardial (44) or
endocardial (36–39) nerve stimulation in vivo. This
approach is nondestructive and associated with maintenance of physiological anterograde propagation of
the cardiac beat. The vagal effects can be turned on and
off almost instantaneously and can be repeated multiple times.
We hypothesized that selective atrioventricular node
(AVN) vagal stimulation (AVN-VS), by varying the
nerve stimulation intensity, could achieve precise
graded VR slowing and thus permit evaluation of the
optimal VR during AF. Accordingly, the aims of the
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. Section 1734
solely to indicate this fact.
0363-6135/02 $5.00 Copyright © 2002 the American Physiological Society
Downloaded from by on May 3, 2017
Zhang, Youhua, Kent A. Mowrey, Shaowei Zhuang,
Don W. Wallick, Zoran B. Popović, and Todor N.
Mazgalev. Optimal ventricular rate slowing during atrial
fibrillation by AV nodal-selective vagal stimulation. Am J
Physiol Heart Circ Physiol 282: H1102–H1110, 2002. First
published November 23, 2001; 10.1152/ajpheart.00738.
2001.—Although the beneficial effects of ventricular rate
(VR) slowing during atrial fibrillation (AF) are axiomatic, the
precise relationship between VR and hemodynamics has not
been determined. We hypothesized that selective atrioventricular node (AVN) vagal stimulation (AVN-VS) by varying
the nerve stimulation intensity could achieve precise graded
slowing and permit evaluation of an optimal VR during AF.
The aims of the present study were the following: 1) to
develop a method for computerized vagally controlled VR
slowing during AF, 2) to determine the hemodynamic
changes at each level of VR slowing, and 3) to establish the
optimal anterograde VR during AF. AVN-VS was delivered to
the epicardial fat pad that projects parasympathetic nerve
fibers to the AVN in 14 dogs. Four target average VR levels,
corresponding to 75%, 100%, 125%, and 150% of the sinus
cycle length (SCL), were achieved by computer feedback
algorithm. VR slowing resulted in improved hemodynamics
and polynomial fit analysis found an optimum for the cardiac
output at VR slowing of 87% SCL. We conclude that this
novel method can be used to maintain slow anterograde
conduction with best hemodynamics during AF.
present study were the following: 1) to develop a
method for computerized vagally controlled VR slowing
during AF, 2) to determine the hemodynamic changes
at each level of VR slowing, and 3) to establish the
optimal anterograde VR during AF.
AJP-Heart Circ Physiol • VOL
feedback ⫽ pgain 䡠 error ⫹ igain 䡠
error 䡠 dt ⫹ dgain 䡠
where pgain is the proportional correction, igain is the integral
correction, and dgain is the derivative correction.
The first component determined the feedback amplification based on the absolute value of the difference between
TRR and each current ARR. The second component adjusted
the feedback amplification based on the cumulative value of
the error over all preceding beats. The third component was
included to account for the specific dynamics of R-R changes.
The “igain” was the major source of control. In most cases, we
used igain ⫽ 10,000, with the “pgain” and the “dgain” being set
to zero.
This PID1 feedback voltage (range ⫾ 10 V) was used to
determine the instantaneous amplitude of the impulses applied for VS. Specifically, bursts containing 20 pulses (6 ms
apart, 1 ms in duration) were synchronized with the right
ventricular apex electrogram (Fig. 1). The polarity of each
consecutive pulse in a burst was alternated to diminish any
polarization effects. The impulses were delivered to the animal via optically isolated constant-current devices (Isolator10, Axon Instrument).
Study protocol. The experiments consisted of three steps.
First, after the surgical preparation and at least 30 min of
stabilization, all electrical and hemodynamic data (surface
ECG, right atrial and right ventricular epicardial electrograms, blood pressures, and AoF) were collected during sinus
rhythm as the baseline values. During the data collection
period, the sinus cycle length (SCL) was determined by
averaging 100 consecutive cycles by a computer. Second, AF
was induced by rapid right atrial pacing (20 Hz, 1 ms) and
thereafter maintained by subthreshold stimulation of the
right pulmonary vein-atrial junction fat pad (selective VS to
the sinus node and surrounding atrium). The latter stimula-
Fig. 1. A right ventricular apex (RVA) complex (top trace) triggers
the delivery of 20 impulses (bottom trace) of vagal stimulation (VS)
stimuli (6 ms apart, 1 ms in duration, and the polarity is alternated
to diminish any polarization effect). The amplitude of atrioventricular node (AVN)-VS impulses is determined with the use of computer
feedback control program (see METHODS).
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
The study was approved by the Institutional Animal Research Committee and is in compliance with the National
Institutes of Health Guide for the Care and Use of Laboratory
Surgical preparation. Fourteen adult mongrel dogs (body
wt 21–30 kg) were premedicated with thiopental sodium (20
mg/kg iv) and intubated and ventilated with room air supplemented with oxygen as needed to maintain normal arterial blood gases by a respirator (NARKOMED 2, North American Drager; Telford, PA). Anesthesia was then maintained
with 1–2% isoflurane throughout the experiment. The left
external jugular vein was cannulated to infuse normal saline
at 100–200 ml/h to replace spontaneous fluid loses. Standard
surface electrocardiogram (ECG) leads (I, II, and III) were
continuously monitored. Intermittent arterial blood gas measurements were taken and ventilator adjustments were
made to correct any metabolic abnormalities. Body temperature was monitored with a rectal probe (model TM-2400,
Electromedics), and an electrical heating pad under the animal and operating-room lamps were used to maintain a body
temperature of 36°-37° C.
The right femoral artery was cannulated, and a micromanometer-tipped catheter pressure transducer (Millar; Houston, TX) was inserted and advanced to the thoracic aorta to
monitor the systemic aortic blood pressure (AoP). For left
ventricular (LV) pressure (LVP) measurements, another Millar catheter was inserted through the left carotid artery and
advanced so that the tip was in the left ventricle. Before the
insertion, both Millar catheters were soaked in warm saline
for 30 min and precalibrated. After the chest was opened
through a median sternotomy, a pericardium cradle was
created to support the heart. Custom-made Ag-AgCl quadripolar plate electrodes were sutured to the high right atrium
and right ventricular apex for bipolar pacing and recording.
Similar bipolar plate electrodes were also sutured to two
epicardial fat pads that contain parasympathetic neural
pathways selectively innervating the sinus node (SN) and the
AVN, the right pulmonary vein-atrial junction fat pad, and
the inferior vena cava-left atrium fat pad, respectively (44).
The ascending aorta was isolated, and a flow probe was
placed around it and connected to a flowmeter (16A/20A, HT
207, Transonic System; Ithaca, NY) to display and measure
aortic flow (AoF). All signals (surface ECG, right atrial and
right ventricular electrocardiogram, aortic blood pressure,
LVP, and AoF signals) were amplified, filtered, and displayed
on our recording system (GE Marquette Medical Systems). In
addition these signals along with calibrations were simultaneously recorded on magnetic tape (model 400A; Vetter Digital) for later computer analysis with AxoScope (Axon Instruments) and a custom-written software program.
Delivery of selective AV nodal vagal nerve stimulation. To
achieve different levels of VR slowing, a computer-controlled
feedback program was developed. The program was personal
computer based using an analog-to-digital (A/D) board from
Microstar Laboratories (Bellevue, WA). Data acquisition and
stimulation output were controlled from Microsoft Excel via a
Microsoft Visual Basic interface with the A/D board. All
desired measurements were written to the spreadsheet in
real time. Development was centered on an existing Mi-
crostar command (PID1) that implemented a classic, proportional-integral derivative, closed loop process control. Specifically, a R-R value was defined as a target ventricular
interval (TRR), such as the R-R interval during spontaneous
sinus rhythm. The actual R-R interval (ARR) was measured
in real time and compared with the TRR. The difference
between the two was the control error. The PID1 command
would then compute the feedback necessary to move the
measured R-R toward the TRR. For example, if TRR ⬎ ARR
in the current beat, the amplitude of AVN-VS was increased
during the next beat, and vice versa. The following algorithm
was used
Y ⫽ Y avg ⫹ 共b1 ⫻ ␰1兲 ⫹ 共b2 ⫻ ␰2兲 ⫹ · · · ⫹ 共bn ⫻ ␰n兲
where Yavg is the average value obtained with all levels of VR
slowing, n is the order of fitted polynomial, bn is a fitting
parameter, and ␰n is determined as
␰ n ⫽ 关共%SCL ⫺ 100兲/25兴n ⫹ mn ⫺ 2 ⫻ 关共%SCL ⫺ 100兲/25兴n ⫺ 2
⫹ · · · ⫹ mn ⫺ 2k ⫻ 关共%SCL ⫺ 100兲/25兴n ⫺ 2k
where m is a scaling factor (42) and k is an integral value
ranging from 1 to n/2. The maximum value of Y determined
the optimal %SCL.
A P value ⬍0.05 was required for statistical significance.
VR during sinus rhythm, AF alone, and graded VR
slowing by AVN-VS. AF normally resulted in irregular
and rapid ventricular responses that rendered an average VR substantially faster than the spontaneous
sinus rate (the average R-R during AF was 61 ⫾ 10% of
the SCL, n ⫽ 14). In 2 of the 14 dogs, however, the
average R-R intervals during AF were 78% and 81% of
the SCL, respectively. In these two cases, the target of
75% SCL could not be achieved because AVN-VS always prolongs the R-R intervals. In two other dogs, a
150% SCL slowing was not achievable at the maximal
available intensity of AVN-VS. These four dogs were
excluded from the analysis, and the reported results
are based on 10 dogs, in which the average R-R during
AF was 58 ⫾ 7% of the SCL and all four target VR
levels (faster and slower than the sinus rate) were
The combined averaged VR achieved and maintained in all dogs are shown in Table 1. As expected,
compared with sinus rate the VR was significantly
faster during AF (125 ⫾ 22 vs. 214 ⫾ 33 beats/min).
Computer feedback-assisted AVN-VS resulted in a
graded predetermined VR slowing. The feedback control was highly effective, as evident from the closeness
of the target and achieved levels (numbers in parentheses in Table 1). Even in the case of the 150% SCL
target, the VR was slowed to 86 beats/min, which
equals 145% SCL.
Figure 2 shows a representative experiment, in
which the four different levels of VR are illustrated. In
this figure, episodes of 500 R-R intervals at sinus
rhythm, AF, and four different levels of AVN-VS were
generated as an uninterrupted sequence for illustration purpose. In the real experiment, as previously
stated, each episode contained up to 1,000 beats, was
followed by a 5-min pause and the order was randomized. It can be seen in Fig. 2 that although the desired
average levels of VR slowing were closely achieved, a
substantial variability was observed on a beat-to-beat
Hemodynamic responses during sinus rhythm, AF
alone, and AF with different levels of VR slowing.
Figure 3 shows representative hemodynamic traces
during AF (Fig. 3A), and VR slowing at 75%, 100%,
Table 1. VR during spontaneous sinus rhythm, induced AF, and during AF with 4 target levels of feedbackcontrolled VR slowing by selective AVN-VS in 10 dogs
Slowing of VR by AVN-VS during AF
75% SCL
100% SCL
125% SCL
150% SCL
R-R interval, ms
VR, beats/min
494 ⫾ 87
125 ⫾ 22
287 ⫾ 47
214 ⫾ 33
371 ⫾ 66 (75%)
166 ⫾ 29
488 ⫾ 86 (99%)
126 ⫾ 20
607 ⫾ 108 (123%)
102 ⫾ 18
715 ⫾ 117 (145%)
86 ⫾ 15
Values are means ⫾ SD. AF, atrial fibrillation; AVN, atrioventricular node; VS, vagal stimulation; VR, ventricular rate; SCL, spontaneous
sinus cycle length. The achieved average R-R intervals during the AVN-VS slowing are shown in milliseconds as well as a percentage of the
SCL (in parentheses) for 4 target levels. P ⬍ 0.001, statistically significant differences were observed between sinus rhythm and the
corresponding values shown in columns 3, 4, 6, and 7, as well as between any pair of columns 3 to 7.
AJP-Heart Circ Physiol • VOL
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
tion utilized continuous sequence of very brief impulses with
duration of 50 ␮s, amplitude of 5–10 mA, and frequency of 20
Hz. In several cases, AF was maintained by continuing the
rapid right atrial pacing. After the hemodynamics stabilized
for at least 15 min, data were recorded during AF for at least
500 cardiac cycles. Third, while maintaining the AF, we
initiated the feedback program to deliver the AVN-VS and to
slow the VR. Four target levels were preprogrammed that
were 75%, 100%, 125%, and 150% of the corresponding spontaneous SCL. The order of the four different levels was
randomized. A rest period of 5 min was allowed after each of
the AVN-VS periods, although the AF was maintained uninterrupted. Hemodynamic data were collected during up to
1,000 ventricular beats in each period, and beats numbered
200–700 were included in the analysis (this was done to
exclude transients at the start of the AVN-VS).
Data acquisition and analysis. Tape-recorded data were
played back off-line and digitized at 1 kHz per channel by
AxoScope (Axon Instrument). For each of the three steps of
the protocol, the R-R intervals, systolic blood pressure (SBP)
and diastolic blood pressure (DBP) from the AoP signal, LV
systolic pressure (LVSP) and LV end-diastolic pressures
(LVEDP) from the LVP signal, and stroke volume (SV) were
calculated beat-by-beat during 500 beats and averaged by a
custom-written software. The SV for individual beats was
determined by integrating the AoF signal over the duration
of the corresponding R-R interval. The ⫾d(LVP)/dt was derived from LVP, and cardiac output (CO) was calculated by
multiplying the heart rate by the SV. The CO determined in
this fashion represented the total LV outflow reduced by the
coronary flow.
Statistical analysis. Data are expressed as means ⫾ SD.
Hemodynamic differences during sinus rhythm, AF alone,
and AF plus different levels of AVN-VS were evaluated by
single factor repeated measurements analysis of variance, followed by post hoc Tukey’s honestly significant difference test.
To calculate the optimal average R-R intervals (expressed
as %SCL), a contrast analysis was used to detect significant
orthogonal polynomial components for each hemodynamic
parameter Y (expressed as a function of %SCL) that were
fitted to the regression equation (42)
125%, and 150% SCL (Fig. 3, B–E). Notably, rapid
irregular electrical activation of the ventricles during
AF resulted in many aborted beats (asterisks in Fig.
3A) along with deterioration of other parameters. The
term “aborted beat” (sometimes referred to as “pulse
deficit”) means that there was electrical activation but
there was no subsequent SV. In such beats, LV contraction resulted in an insufficient LV pressure, less
than the diastolic arterial blood pressure level. As a
result, the aortic valve did not open and no blood was
ejected from the left ventricle. In this study, we considered the beat to be aborted if the electrical activation was followed by ⬍1 ml of SV. With AVN-VS
producing progressive VR slowing, these aborted beats
were dramatically reduced (asterisks in Fig. 3, B–D)
together with the improvement of other indexes. There
were almost no aborted beats when VR was slowed to
150% SCL level (Fig. 3E).
The composite hemodynamic responses in all 10 dogs
during sinus rhythm, AF alone, and AF accompanied
with different levels of VR slowing are summarized in
Table 2. AF resulted in a fast VR with deterioration of
all hemodynamic parameters compared with sinus
rhythm (P ⬍ 0.001). Selective AVN-VS, by prolonging
the R-R intervals, improved CO, SBP, DBP, LVSP,
LVEDP, ⫾d(LVP)/dt, as well as SV and decreased the
number of aborted beats (Table 2, compare the asterisk-labeled values with the values for AF).
It should be noted that in contrast to the sinus rate,
AF was associated with smaller LVSP than SBP values
(63 ⫾ 12 vs. 77 ⫾ 17 mmHg). This apparent “discrepancy” resulted from the presence of large (43 ⫾ 9%)
number of aborted beats, for which LVSP is always less
than SBP. Because an aborted beat produced no AoF,
its (apparent) SBP was defined as equal to the DBP of
previous cycle. On the other hand, because the aortic
valve did not open, it is obvious that the LVSP for the
AJP-Heart Circ Physiol • VOL
aborted beat was far below the DBP of the previous
beat. Slowing of the VR by the VS reduced the number
of aborted beats from 43% to 2% (Table 2) and resulted
in a progressive reduction of the difference between the
averaged LVSP and SBP.
A more detailed analysis of the dependency of the
hemodynamics on the VR is shown in Fig. 4, where the
experimental data points are superimposed on curves
obtained with a polynomial fit procedure for each of the
measured parameters (see METHODS). The regression
lines for CO, SBP, LVSP, ⫾d(LVP)/dt, and DBP had a
discernible maximum. The maximum occurred between 83% and 140% of SCL but was most clear for CO
with a maximum at 87% SCL (Fig. 4A). The remaining
parameters exhibited rather monotonic relationships.
Although VR slowing in the entire range from 75% to
150% SCL tended to improve each parameter, there
was no statistically significant additional beneficial
effect at R-R⬎100% SCL for SBP, LVSP, ⫾d(LVP)/dt,
DBP, and LVEDP (Fig. 4, D–I, and Table 2). Statistically significant increase in this range was evident only
for the SV and especially for the number of aborted
beats [Fig. 4, B and C, and Table 2 (compare the
†-labeled values in columns 4, 6, and 7 with the values
in column 5)]. Importantly, despite the increased SV,
Fig. 3. Representative hemodynamic traces during AF (A), AF with
VR slowing to 75% (B), 100% (C), 125% (D), and 150% (E) SCL.
Electrocardiogram (ECG), standard surface ECG lead II; RA, right
atrial electrogram; RV, right ventricle electrogram; AoP, aortic blood
pressure; LVP, left ventricular pressure; AoF, aortic blood flow. The
asterisk marks the aborted beats (see RESULTS). The delivery of
AVN-VS could be seen by the artifacts in surface ECG at the end of
QRS (B–E).
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
Fig. 2. R-R intervals during spontaneous sinus rhythm, atrial fibrillation (AF), and AF with different levels of ventricular rate (VR)
slowing achieved by selective AVN-VS. Each episode contains 500
R-R intervals. Beats 1–500, 501–1,000, 1,001–1,500, 1,501–2,000,
2,001–2,500, and 2,501–3,000 were collected during sinus rhythm,
AF, and AF with VR slowing to 75%, 100%, 125%, and 150% of the
spontaneous sinus cycle length (SCL), respectively. The SCL was 463
ms. The corresponding average R-R intervals in each subsequent
period were 270 ms (AF), 348 ms (achieved 75% SCL), 462 ms
(achieved 99.9% SCL), 577 ms (achieved 124.6% SCL), and 691 ms
(achieved 149.2% SCL), respectively.
Table 2. Hemodynamic changes (compared with the spontaneous sinus rhythm) during AF alone, and during
AF with 4 different levels of feedback-controlled VR slowing by selective AVN-VS in 10 dogs
Sinus Rhythm
TL ⫽ 75% SCL
TL ⫽ 100% SCL
TL ⫽ 125% SCL
TL ⫽ 150% SCL
SBP, mmHg
DBP, mmHg
LVSP, mmHg
⫹d(LVP)/dt, mmHg/s
⫺d(LVP)/dt, mmHg/s
SV, ml/beat
% Aborted beats
CO, l/min
100 ⫾ 16*
77 ⫾ 14*
101 ⫾ 12*
5.4 ⫾ 2.8*
1,691 ⫾ 364*
⫺1,789 ⫾ 373*
22.8 ⫾ 8.0*
2.7 ⫾ 0.7*
77 ⫾ 17
61 ⫾ 14
63 ⫾ 12
7.6 ⫾ 3.0
1,327 ⫾ 379
⫺952 ⫾ 356
8.6 ⫾ 2.9
43 ⫾ 9
1.8 ⫾ 0.4
96 ⫾ 21*
77 ⫾ 19*
80 ⫾ 11*
6.0 ⫾ 2.4*
1,624 ⫾ 342*
⫺1,315 ⫾ 429*
14.3 ⫾ 5.6*†
33 ⫾ 10*†
2.2 ⫾ 0.7*
98 ⫾ 20*
74 ⫾ 17*
89 ⫾ 13*‡
6.2 ⫾ 2.2
1,683 ⫾ 414*
⫺1,448 ⫾ 461*‡
19.0 ⫾ 6.7*‡
20 ⫾ 8*‡
2.3 ⫾ 0.6*‡
96 ⫾ 17*
69 ⫾ 14
91 ⫾ 10*
5.5 ⫾ 2.0*
1,635 ⫾ 340*
⫺1,435 ⫾ 310*
21.4 ⫾ 7.4†*
10 ⫾ 8†*
2.1 ⫾ 0.5†*
97 ⫾ 21*
66 ⫾ 17†
95 ⫾ 12*
5.7 ⫾ 2.2*
1,674 ⫾ 376*
⫺1,539 ⫾ 484
23.6 ⫾ 7.1*†
2 ⫾ 6*†
1.9 ⫾ 0.5†
the slowing of the VR beyond 100% SCL resulted in a
significant decline of the CO [Fig. 4A, and Table 2
(compare the †-labeled values in columns 6 and 7 with
the values in column 5)]. This negative outcome was a
combined result of the increased SV and the substantially decreased VR.
Thus slowing of VR to 100% SCL increased most
hemodynamic parameters and produced the best CO.
However, these improvements were still insufficient to
reach the values measured during the spontaneous
sinus rhythm [(Table 2 (compare ‡-labeled values in
column 5 with column 2)]. This, at least in part, may be
because the slowing of the VR by AVN-VS did not
eliminate the irregularity of the heart rhythm. In fact,
the example illustrated in Fig. 2 demonstrates that in
absolute terms there was larger beat-to-beat variability at the slower VR levels. The summarized data from
all 10 dogs are shown in Fig. 5. The relative irregularity is represented by the normalized standard deviation of the 500 measured R-R intervals (as a percentage
of SD/R-R, closed symbols). The CO, normalized versus
the values at sinus rate, is also shown (as percent, open
symbols). Compared with AF, it is clear that the CO
increased from 66% in AF to 85% of the maximum
when AVN-VS slowed the rate to 100% SCL. At that
level of slowing, there was an SD/R-R ⫽ 25%. It is
interesting to note that further slowing of the VR to
150% SCL actually reduced CO to 70% despite the fact
that the variability declined to 17%. Although the
present experimental protocol does not permit direct
evaluation of the hemodynamic contributions of irregularity (see DISCUSSION), the data suggest a predominant role of the optimal slowing of the VR in maintaining high CO during AF.
Major findings. This study confirmed that selective
AVN parasympathetic stimulation through the epicardial fat pad approach could be used to control (i.e., to
AJP-Heart Circ Physiol • VOL
slow down to a predetermined level) the average VR
during AF. AF worsened all hemodynamic parameters
compared with those observed during sinus rhythm.
We identified a range of vagally induced VR slowing
that was associated with improvement of hemodynamics. Although a substantial improvement (vs. AF) of all
major hemodynamic parameters was observed at any
level of AVN-VS, the best results were achieved when
VR was maintained equal to, or slightly faster than,
the spontaneous sinus rate. In particular, the CO,
which is a product of the SV and VR, was suboptimal at
75% SCL due to insufficient SV, as well as at 125% and
150% SCL due to excessive slowing of VR. Therefore,
we concluded that 100% SCL should be a preferable
target during AF.
Fast VR during AF and its effect on the hemodynamics. Multiple mechanisms are responsible for the detrimental hemodynamic consequences during AF.
These include the increase in heart rate, the loss of
atrial contribution to ventricular filling, the reduced
interval for passive diastolic filling, and the irregular
ventricular rhythm (3, 7, 9, 15, 24, 26, 33, 35, 40).
Clinically, VR slowing is considered to be the first
therapeutic step in a majority of patients with both
acute and chronic AF (16). The importance of this
parameter is further accentuated by the observation
that in patients with successful conversion of AF, CO
increase was observed only when the conversion was
associated with slowing of the VR (1, 20, 32).
The detrimental hemodynamic response during AF
is a complex function of the rapid irregular sequence of
shorter and longer R-R intervals. Mechanistically, this
produces interaction of two phenomena. First, after a
short R-R interval, the inotropic response is decreased
due to incomplete mechanical restitution reflecting the
dynamics of calcium release from the sarcoplasmic
reticulum (46). In addition, according to the FrankStarling principle (4), the short R-R intervals are associated with reduced end-diastolic volume that fur-
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
Values are means ⫾ SD. TL, target level of the average R-R interval as a percentage of the SCL. SBP and DBP: systolic and diastolic blood
pressures determined from the aortic pressure (AoP) signal; LVSP and LVEDP, left ventricular systolic and left ventricular end-diastolic
pressures determined from the LVP signal; ⫾d(LVP)/dt, ⫾ maximal rate of change of LV pressure; SV, stroke volume; CO, cardiac output.
P ⬍ 0.05 was considered to be significant. * Statistically significant differences vs. AF; † statistically significant differences between the
values at 75%, 125%, and 150% SCL vs. the values at 100% SCL; ‡ statistically significant differences between the values at 100% SCL vs.
the values at sinus rhythm.
ther attenuates the subsequent SV. Second, when a
longer R-R interval occurs after a shorter one, the
mechanical response after the former exhibits a relative potentiation (34, 46). The random occurrence of
shorter and longer R-R intervals during AF results in a
complex interaction between the above two phenomena. The overall decreased cardiac index (15) indicates
that the persistent, adverse effects of shorter R-R intervals on ventricular performance dominate the hemodynamic outcome.
Although the present study was not designed to
directly evaluate the effects of irregularity, the data in
Fig. 5 suggest that vagally induced slowing of the VR
during AF was associated with an increased standard
deviation of the R-R intervals. However, the present
observations preclude conclusion about the separate
AJP-Heart Circ Physiol • VOL
Fig. 5. Normalized cardiac output in percent vs. the values at the
spontaneous sinus rate (E) and heart rate variability represented as
a normalized standard deviation of 500 measured R-R intervals (in
%SD/R-R, F) during AF and 4 levels of VR slowing to 75%, 100%,
125%, and 150% of the SCL. AVN-VS increased the normalized CO
from 66% during AF to 85% at slowing of 100% SCL, although this
was accompanied by an increase in irregularity from 17% to 25%
(P ⬍ 0.001 for both). See RESULTS for details.
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
Fig. 4. Polynomial contrasts analysis of the measured hemodynamic parameters during AF and 4 levels of VR
slowing expressed as %SCL. The VR corresponding to the AF is considered an additional level (the experimental
value for it in 10 animals was 58 ⫾ 7% SCL). Solid circles represent mean experimental data (see Table 2). The
lines represent the fit obtained with a combination of significant linear, quadratic, cubic, and quartic components.
Optimal %SCL, where detected, is indicated with dashed lines. The optimum in the CO curve at 87% SCL (A).
Excessive VR slowing (R-R ⬎100% SCL), despite the progressive increase in the SV (B) and the decline in the
number of aborted beats (C), resulted in a decrease of the CO (A). See RESULTS for details.
AJP-Heart Circ Physiol • VOL
well as in dogs in vivo (44). The present study extended
these findings and established that precise VR slowing
can be achieved by feedback-controlled AVN-VS.
The superiority of this approach is that it avoids the
unwanted effects carried by drugs. Moreover, in our
experiments, AVN-VS exerted its effects exclusively on
the AVN. That is, when applied during sinus rhythm it
did not change the sinus rate (44), and when applied
during atrioventricular sequential pacing it did not
alter any hemodynamic indexes (not shown in the
reported data). Furthermore, AVN-VS allowed VR reduction to several predetermined levels to be achieved
promptly during AF. Notably, by preserving the anterograde His-Purkinje conduction, we were able to
evaluate the hemodynamic response in a model much
closer to the real natural AF compared with models
utilizing retrograde ventricular pacing (22).
Implications of the reported findings and study limitations. The anatomic basis of the epicardial approach
for selective parasympathetic stimulation has been
well established in animal models (2, 6, 12, 13, 21,
28–30). Although only endocardial stimulation has
been so far employed to produce dromotropic effects in
humans (11, 17–19, 43), it is possible that the technique used in the present study could be applied in
some patients, especially postoperative patients with
AF. The optimal VR level observed in this study could
be considered as a clinical target VR level during AF.
However, due to the nature of animal studies, certain limitations should be considered. First, these are
acute experiments. Whether the acutely determined
rates would also apply to a chronic stage, and whether
enhanced sympathetic tone during exercise would compete with the AVN-VS remain to be determined in
chronic experiments. Second, whether the optimal rate
derived from healthy animals can be recommended in a
diseased state is still not clear. This is important because clinical AF is frequently accompanied by other
heart diseases (i.e., coronary heart disease and heart
failure). Third, the prolonged effects of AVN-VS during
AF need further investigation in chronic experiments.
Finally, the slowed ventricular rhythm observed in this
study was irregular. It is interesting that the slight increase of variability with the slowing of the VR (Figs. 2
and 5) has also been documented in patients undergoing
slow pathway AVN modification during AF (41). Specifically, the ablation prolonged the average R-R interval
from 481 to 640 ms and increased the average SD/R-R
from 16.8% to 20%. Although the mechanism of this
phenomenon is not known, it might represent an inherent property of conduction of the depressed AVN during
AF. A comparison of the hemodynamic outcome during
controlled slowing of the anterograde irregular ventricular activation (as achieved in the present experiments)
with complete AVN ablation and retrograde but regular
pacing at several comparable slow rates has never been
done and deserves a separate investigation. Such studies
are needed to evaluate whether or not the price paid in
AVN destruction and pacemaker dependency is worth a
potential hemodynamic benefit.
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
negative hemodynamic contribution of this factor. In
particular, the best CO was achieved at 100% SCL
(Table 2 and Fig. 5), despite the fact that at this level
of AVN-VS the irregularity was higher than at lesser
(75% SCL) or higher (150% SCL) degrees of VR slowing
(Fig. 5).
Optimal anterograde VR slowing during AF. Although the beneficial effect of the slower VR during AF
is axiomatic, the relationship between the VR and the
parameters of cardiac function in AF has not been
clearly quantified. In this study, we were able to
achieve different levels of VR slowing during AF by
utilizing vagal modulation of AVN conductivity without directly affecting ventricular performance. Our results clearly showed that the detrimental, rapid, irregular ventricular response could be successfully
attenuated by selective AVN-VS. Slowing the VR to the
sinus rate or slightly faster (average R-R interval in
the range of 75%-100% of the SCL) appeared to yield
best overall results (Table 2). The polynomial fit indicated a presence of an optimum for the CO at 87% of
the SCL (Fig. 4). Slowing to 125% and 150% of the SCL
further improved some of the other measured parameters, although only for SV and the number of aborted
beats the additional changes reached statistical significance (Table 2, Fig. 4). However, the overall hemodynamic effectiveness as reflected in the CO decreased
due to excessive reduction of the number of cardiac
contractions per unit time. This is in agreement with
previous studies suggesting that lengthening of the
diastolic interval beyond ⬃700 ms does not appreciably
increase the LV end-diastolic volume, whereas shortening of the diastolic interval to ⬍500 ms impairs LV
filling and SV (14). Interestingly, the empirically recommended target rate of 60–90 beats/min for humans during AF (45, 47) corresponds closely to the data obtained in
this study (i.e., for an average sinus rate of 70 beats/min,
or SCL ⫽ 857 ms, the recommended target of 75–100%
SCL during AF would be 70–93 beats/min).
Selective AVN-VS as a useful novel method for slowing VR during AF. As previously indicated, the use of
drugs to slow the VR during AF has limitations because VR changes are unpredictable in response to
treatment (22), and agents such as ␤-blockers, verapamil, diltiazem, and digitalis, have a direct effect on
cardiac function (31). Moreover, drugs lack efficacy in
some subjects and are intolerable by others (8).
The approach used in the present study was based on
the well-established dromotropic effects of the vagus on
AVN. The cell bodies of the postganglionic parasympathetic nerve fibers, which selectively innervate limited
regions of the heart, have been identified for both
animals and humans (2, 5, 6, 12, 13, 21, 28–30). For
example, the fibers that innervate the AVN course
through “. . .a smaller fat pad overlying epicardium at
the junction of inferior vena cava-inferior left atrium”
(29). It is known that stimulation of these parasympathetic projections can slow AV conduction without affecting other cardiac functions (2, 6, 12, 13, 21, 28–30).
Using this method, we have previously demonstrated
the possibility to slow VR during AF in vitro (23) as
The authors thank Donald G. Hills and William J. Kowalewski for
expert help during the surgical preparation and assistance in the
This study was supported in part by National Heart, Lung, and
Blood Institute Grant RO1-HL-60833-01A1.
AJP-Heart Circ Physiol • VOL
19. Keim S, Mehdirad AA, Rist K, Mazgalev T, and Tchou P.
Subthreshold burst stimulation of the AV node unmasks latent
preexcitation in a concealed ectopic nodoventricular accessory
pathway (Abstract). Pacing Clin Electrophysiol 17: 336, 1994.
20. Killip T and Baer RA. Hemodynamic effects after reversion
from atrial fibrillation to sinus rhythm by precordial shock.
J Clin Invest 45: 658–671, 1966.
21. Lazzara R, Scherlag BJ, Robinson MJ, and Samet P. Selective in situ parasympathetic control of the canine sinoatrial
and atrioventricular nodes. Circ Res 32: 393–401, 1973.
22. Liau CS, Chen MF, Lin FY, Tsai CH, and Lee YT. Relationship between ventricular rate and cardiac output in mimic experimental atrial fibrillation. J Electrocardiol 27: 163–168,
23. Mazgalev TN, Garrigue S, Mowrey KA, Yamanouchi Y,
and Tchou PJ. Autonomic modification of the atrioventricular
node during atrial fibrillation: role in the slowing of ventricular
rate. Circulation 99: 2806–2814, 1999.
24. Mitchell JH and Shapiro W. Atrial function and the hemodynamic consequences of atrial fibrillation in man. Am J Cardiol
23: 556–567, 1969.
25. Morady F, Hasse C, Strickberger SA, Man KC, Daoud E,
Bogun F, Goyal R, Harvey M, Knight BP, Weiss R, and
Bahu M. Long-term follow-up after radiofrequency modification
of the atrioventricular node in patients with atrial fibrillation.
J Am Coll Cardiol 29: 113–121, 1997.
26. Naito M, David D, Michelson EL, Schaffenburg M, and
Dreifus LS. The hemodynamic consequences of cardiac arrhythmias: evaluation of the relative roles of abnormal atrioventricular sequencing, irregularity of ventricular rhythm and atrial
fibrillation in a canine model. Am Heart J 106: 284–291, 1983.
27. Prystowsky EN, Benson DW Jr, Fuster V, Hart RG, Kay
GN, Myerburg RJ, Naccarelli GV, and Wyse DG. Management of patients with atrial fibrillation. A statement for healthcare professionals. The Subcommittee on Electrocardiography
and Electrophysiology, American Heart Association. Circulation
93: 1262–1277, 1996.
28. Randall WC, Ardell JL, Calderwood D, Miloslavljevic M,
and Goyal SC. Parasympathetic ganglia innervating the canine
atrioventricular nodal region. J Auton Nerv Syst 16: 311–323,
29. Randall WC, Ardell JL, O’Toole MF, and Wurster RD.
Differential autonomic control of SAN and AVN regions of the
canine heart: structure and function. Prog Clin Biol Res 275:
15–31, 1988.
30. Randall WC, Miloslavljevic M, Wurster RD, Geis GS, and
Ardell JL. Selective vagal innervation of the heart. Ann Clin
Lab Sci 16: 198–208, 1986.
31. Reiffel JA. Drug choices in the treatment of atrial fibrillation.
Am J Cardiol 85: 12–19, 2000.
32. Resnekov L. Haemodynamic studies before and after electrical
conversion of atrial fibrillation and flutter to sinus rhythm. Br
Heart J 29: 700–708, 1967.
33. Rodman T, Pastor BH, and Figueroa W. Effect on cardiac
output of conversion from atrial fibrillation to normal sinus
mechanism. Am J Med 41: 249–258, 1966.
34. Samet P. Hemodynamic sequelae of cardiac arrhythmias. Circulation 47: 399–407, 1973.
35. Samet P, Bharati S, and Levy D. Significance of the atrial
contribution to ventricular filling. Am J Cardiol 15: 195–202,
36. Schauerte P, Scherlag BJ, Scherlag MA, Goli S, Jackman
WM, and Lazzara R. Ventricular rate control during atrial
fibrillation by cardiac parasympathetic nerve stimulation: a
transvenous approach. J Am Coll Cardiol 34: 2043–2050, 1999.
37. Schauerte P, Scherlag BJ, Scherlag MA, Jackman WM,
and Lazzara R. Transvenous parasympathetic nerve stimulation in the inferior vena cava and atrioventricular conduction.
J Cardiovasc Electrophysiol 11: 64–69, 2000.
38. Schauerte PN, Scherlag BJ, Scherlag MA, Goli S, Jackman W, and Lazzara R. Transvenous parasympathetic cardiac
282 • MARCH 2002 •
Downloaded from by on May 3, 2017
1. Abildskov JA, Millar K, and Burgess MJ. Atrial fibrillation.
Am J Cardiol 28: 263–267, 1971.
2. Ardell JL and Randall WC. Selective vagal innervation of
sinoatrial and atrioventricular nodes in canine heart. Am J
Physiol Heart Circ Physiol 251: H764–H773, 1986.
3. Benchimol A. Significance of the contribution of atrial systole to
cardiac function in man. Am J Cardiol 23: 568–571, 1969.
4. Braunwald E, Frye RL, Atgen AA, and Gilbert JW. Studies
on Starling’s law on the heart. III. Observations in patients with
mitral stenosis and atrial fibrillation on the relationships between left ventricular end-diastolic segment length, filling pressure, and the characteristics of ventricular contraction. J Clin
Invest 39: 1874–1884, 1960.
5. Carlson MD, Geha AS, Hsu J, Martin PJ, Levy MN, Jacobs
G, and Waldo AL. Selective stimulation of parasympathetic
nerve fibers to the human sinoatrial node. Circulation 85: 1311–
1317, 1992.
6. Chiou CW, Eble JN, and Zipes DP. Efferent vagal innervation
of the canine atria and sinus and atrioventricular nodes. The
third fat pad. Circulation 95: 2573–2584, 1997.
7. Clark DM, Plumb VJ, Epstein AE, and Kay GN. Hemodynamic effects of an irregular sequence of ventricular cycle
lengths during atrial fibrillation. J Am Coll Cardiol 30: 1039–
1045, 1997.
8. Crijns HJ, Van Gelder IC, and Lie KI. Benefits and risks of
antiarrhythmic drug therapy after DC electrical cardioversion of
atrial fibrillation or flutter. Eur Heart J 15, Suppl: 17–21, 1994.
9. Daoud EG, Weiss R, Bahu M, Knight BP, Bogun F, Goyal R,
Harvey M, Strickberger SA, Man KC, and Morady F. Effect
of an irregular ventricular rhythm on cardiac output. Am J
Cardiol 78: 1433–1436, 1996.
10. Feld GK. Radiofrequency catheter ablation versus modification
of the AV node for control of rapid ventricular response in atrial
fibrillation. J Cardiovasc Electrophysiol 6: 217–228, 1995.
11. Fromer M and Shenasa M. Ultrarapid subthreshold stimulation for termination of atrioventricular node reentrant tachycardia. J Am Coll Cardiol 20: 879–883, 1992.
12. Furukawa Y, Wallick DW, Carlson MD, and Martin PJ.
Cardiac electrical responses to vagal stimulation of fibers to
discrete cardiac regions. Am J Physiol Heart Circ Physiol 258:
H1112–H1118, 1990.
13. Gatti PJ, Johnson TA, Phan P, Jordan IK, Coleman W, and
Massari VJ. The physiological and anatomical demonstration of
functionally selective parasympathetic ganglia located in discrete fat pads on the feline myocardium. J Auton Nerv Syst 51:
255–259, 1955.
14. Gosselink AT, Blanksma PK, Crijns HJ, Van Gelder IC, De
Kam PJ, Hillege HL, Niemeijer MG, Lie KI, and Meijler
FL. Left ventricular beat-to-beat performance in atrial fibrillation: contribution of Frank-Starling mechanism after short
rather than long RR intervals. J Am Coll Cardiol 26: 1516–1521,
15. Herbert WH. Cardiac output and the varying R-R interval of
atrial fibrillation. J Electrocardiol 6: 131–135, 1973.
16. Jung F and DiMarco JP. Treatment strategies for atrial
fibrillation. Am J Med 104: 272–286, 1998.
17. Keim S, Mazgalev T, Kinn R, and Tchou P. Non captured
burst stimulation near the AV node: evidence for vagal stimulation as the mechanism for modulation of AV conduction in
humans (Abstract). Pacing Clin Electrophysiol 15: 519, 1992.
18. Keim S, Mazgalev T, and Tchou P. Physiologic effects of
subthreshold burst stimulation on the human AV node (Abstract). Pacing Clin Electrophysiol 16: 158, 1993.
nerve stimulation: an approach for stable sinus rate control.
J Cardiovasc Electrophysiol 10: 1517–1524, 1999.
Scherlag MA, Scherlag BJ, Yamanashi W, Schauerte P,
Goli S, Jackman WM, Reynolds D, and Lazzara R. Endovascular neural stimulation via a novel basket electrode catheter: comparison of electrode configurations. J Interv Card Electrophysiol 4: 219–224, 2000.
Shapiro W and Klein G. Alterations in cardiac function immediately following electrical conversion of atrial fibrillation to
normal sinus rhythm. Circulation 38: 1074–1084, 1968.
Simpson C, Yee R, Lee JK, Murgatroyd FD, Basta M, and
Krahn AD. The effect of catheter AV node modification on heart
rate irregularity in atrial fibrillation (Abstract). Circulation 98:
I-181, 1998.
Snedecor GW and Cochran WG. Statistical Methods. Ames,
IA: Iowa State University Press, 1989, p. 409.
Tchou P, Keim S, Kinn R, and Mazgalev T. Non captured
burst stimulation near the AV node: selective influence on slow
pathway conduction in AV nodal reentrant tachycardia in humans (Abstract). J Am Coll Cardiol 19: 144A, 1992.
Wallick DW, Zhang Y, Tabata T, Zhuang S, Mowrey KA,
Watanabe J, Greenberg NL, Grimm RA, and Mazgalev TN.
Selective AV nodal vagal stimulation improves hemodynamics
during acute atrial fibrillation in dogs. Am J Physiol Heart Circ
Physiol 281: H1490–H1497, 2001.
Wood DL, Hammill SC, and Kopecky SL. Supraventricular
arrhythmias. In: Mayo Clinic Practice of Cardiology, edited by
Giuliani ER, Gersh BJ, McGoon MD, Hayes DL, and Schaff HV.
St. Louis, MO: Mosby, 1996, p. 748–779.
Yue DT, Burkhoff D, Franz MR, Hunter WC, and Sagawa
K. Postextrasystolic potentiation of the isolated canine left ventricle. Relationship to mechanical restitution. Circ Res 56: 340–
350, 1985.
Zipes DP. Specific arrhythmias: diagnosis and treatment. In:
Heart Disease–A Textbook of Cardiovascular Medicine, edited by
Braunwald E. Philadelphia, PA: Saunders, 1992, p. 667–715.
Downloaded from by on May 3, 2017
AJP-Heart Circ Physiol • VOL
282 • MARCH 2002 •