Download Slow and Long-Lasting Modulation of

1412
Slow and Long-Lasting Modulation of
Myocardial Repolarization Produced by
Ectopic Activation in Isolated Rabbit Hearts
Evidence for Cardiac "Memory"
Angelika Costard-Jackle, MD, Bettina Goetsch, Matthias Antz,
and Michael R. Franz, MD
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Prolonged ventricular pacing induces T wave polarity changes that persist long after cessation of
pacing. To examine how ventricular repolarization is modulated by prolonged changes in activation
sequence, we studied the effect of ectopic pacing on the distribution of action potential durations
(APDs) in nine isolated Langendorif-perfused rabbit hearts. A contact electrode probe was used to
map right and left ventricular-epicardial monophasic action potentials during three consecutive
changes in stimulus site, that is, 1) during 45 minutes of right atrial pacing, 2) during 120 minutes
of right ventricular pacing, and 3) again, during 60 minutes of right atrial pacing. During each of
these phases, the effect of activation sequence on repolarization was examined by linear-regression
analysis ofAPD on activation time (AT). Results averaged for all nine hearts showed that during the
initial atrial-pacing phase, APD was inversely related to AT (slope [S]= -1.63; r=0.76), indicating
that sites with earlier activation repolarized later. With the onset of ventricular pacing, the inverse
correlation between AT and APD disappeared (S= -0.19; r=0.31). Continuing ventricular pacing,
however, produced slow changes in APD that restored the inverse relation (S -0.71; r=0.68 at 120
minutes; p<0.0002 versus 5 minutes). Switching from ventricular to atrial pacing again perturbed
the inverse correlation (S=0.28; r=0.35) but at 60 minutes of atrial pacing, a significant inverse
correlation was reestablished (S= -1.25; r=0.53;p<0.01 versus 5 minutes). An inverse correlation
between AT and APD tends to synchronize repolarization time (RT, the sum of AT and APD).
Accordingly, the variance coeffiicient for RT at all recording sites, a measure of repolarization
dispersion, was significantly greater shortly after a change in pacing site (4.7±0.3 at 5 minutes of
ventricular pacing versus 2.4±0.3 at 45 minutes of atrial pacing, p<0.0001) and decreased
progressively as pacing at the new site was maintained (3.4±0.4 at 120 minutes of ventricular
pacing,p<0.02 versus at 5 minutes). These data indicate that the sequence of ventricular activation
modulates the sequence of ventricular repolarization by a yet unidentified process with very slow
onset and offset characteristics. This phenomenon of cardiac "memory" may explain T wave
changes following prolonged changes in ventricular activation, as seen in intermittent left bundle
branch block, after ventricular tachycardia, and in preexcitation syndromes. Short-term changes in
ventricular activation (as used during electrophysiologic studies) may not allow for sufficient
adaptation of APD to AT, producing greater and possibly arrhythmogenic dispersion of repolarization. (Circulation 1989;80:1412-1420)
From the Cardiology Division, Stanford University School of
Medicine, Stanford, California.
Supported in part by the American Heart Association, California Chapter. A.C.J. supported by a fellowship grant from the
Deutsche Forschungs gemeinschaft, Bonn Bad Godesberg, German Federal Republic. M.R.F. is a recipient of a FIRST award of
the National Heart, Lung, and Blood Institute.
Presented in part at the Young Investigator Award Competition of the North-American Society of Pacing and Electrophysiology (NASPE).
Address for reprints: Michael R. Franz, MD, PhD, Cardiac
Arrhythmia Unit, Cardiology Division, CVRC 293, Stanford University School of Medicine, 300 Pasteur Drive, Stanford, CA 94305.
Received July 25, 1988; revision accepted July 13, 1989.
wave changes can be classified as either
primary or secondary. Primary T wave
changes have been defined as the result of
uniform or nonuniform changes in action potential
T
duration (APD) in the absence of changes in the
sequence of activation. Secondary T wave changes
have been defined to reflect changes in the sequence
of repolarization that result solely from changes in
the sequence of activation, without any abnormalities in the duration and shape of action potentials.1-3
However, Chatterjee et a14 and Rosenbaum et a15
Costard-Jackle et al Myocardial Memory
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
demonstrated that prolonged alteration in the
sequence of activation caused by ventricular pacing
not only caused secondary T wave changes during
the period of pacing but that conspicuous T wave
changes persisted for a long time after pacing was
terminated and a normal supraventricular activation
pattern had resumed. Other, apparently related
clinical phenomena are transient T wave alterations
after intermittent left bundle branch block,6,7 ventricular tachycardia,8'9 extrasystoles,10 or ventricular preexcitation.11 In each, T wave alterations
persist after normalization of ventricular activation
and, for the most part, in the absence of known
myocardial disease that would allow these T wave
changes to be classified as primary.
Rosenbaum et a15 noted that prolonged ventricular pacing modulates ventricular repolarization such
that the T wave again becomes concordant to the R
wave, as during normal supraventricular activation.
They interpreted this intriguing finding by attributing to the myocardium the properties of accumulation or "memory" by which the heart "learns" to
adjust its repolarization to an altered activation
sequence and by which it retains the adapted state
long after the activation sequence is normalized.
However, basic electrophysiology has not yet provided direct validation that myocardial tissue is
capable of such long-lasting "memory" of the past
sequence of activation or repolarization.
The present study was designed to examine
whether pacing-induced changes in repolarization
are indeed due to an intrinsic myocardial property
and can occur without extracardiac influences. In
isolated Langendorff-perfused rabbit heart preparations (which lack neurohumoral influences), we
mapped the distribution and interrelation of activation time (AT), APD, and repolarization time (RT)
during consecutive, long-lasting changes in ventricular activation sequence. Our objectives were 1) to
determine whether prolonged changes in activation
sequence can produce persistent changes in ventricular repolarization, 2) to determine the time course
with which these changes develop and dissipate,
and 3) to determine how these changes influence the
synchrony of repolarization in the whole ventricle.
Methods
Isolated Rabbit Heart Preparation
Twenty-one New Zealand White rabbits of either
sex, weighing 5.0-6.6 kg, were anticoagulated with
75 units/kg heparin i.v. and stunned by a blow to the
occiput. After a median sternotomy, the heart was
removed quickly and chilled in ice-cold modified
Krebs-Henseleit solution of the following composition (mM): NaCl, 118; KCl, 4.7; CaCI2, 2.5; MgSO4,
1.2; KH2PO4, 1.2; NaHCO3, 25; glucose, 5.5; and
Na pyruvate, 2.0. Human albumin (0.0006 mM) was
added to the solution. These small amounts of
albumin have been shown to significantly increase
the stability of the preparation so that the frequency
1413
of ventricular fibrillation is drastically delayed, and
the effective refractory period remains stable over
4-5 hours of artificial perfusion.12 The cut aortic
stump was cannulated and the heart transferred to a
warmed Langendorff apparatus. Nonrecirculating
perfusion with warmed (34.9±+ 0.20 C), oxygenated
(95% 02-5% C02), modified Krebs-Henseleit solution of the composition previously given (pH
7.40+0.02) was initiated at a constant flow rate of 20
ml/minute using a flow roller-pump system (Micro
Pump model 31152). Average heart weight was
8.0±0.6 g. The time from killing of the rabbit to
initiation of the perfusion was less than 60 seconds.
A small needle was inserted through the left ventricular free wall for drainage of thebesian flow. The
perfusion pressure was monitored through a branch
of the aortic cannula with a sphygmomanometer
and was within a range of 60-100 mm Hg throughout the experiment. Bipolar copper wire electrodes
were inserted into the right atrial appendage and
into the right ventricular free wall near the apex and
connected to a custom-built stimulator. The stimulus output could be switched between the two
pacing sites without interruption of the pacing frequency. Bipolar platinum wire electrodes were
inserted at the ventricular base, and the recorded
electrogram was monitored continuously on a Tektronix 5111A oscilloscope.
Experimental Protocol
Immediately after the heart was mounted to the
Langendorff apparatus, pacing was started through
bipolar platinum pacing wires inserted into the right
atrial appendage. Another pacing dipole was inserted
in the right ventricular apex to be used later in the
experiment. Stimuli of 1-msec duration were delivered with a constant current isolated stimulus unit
(World Precision Instruments, New Haven, Connecticut) at twice diastolic threshold strength that
initially averaged 0.6+0.2 mA and remained constant within 0.2 mA throughout the experiment. The
pacing rate was set at 180 beats/min and kept
constant throughout the experiment. After an equilibration period of 30 minutes, monophasic action
potentials (MAPs) were recorded from 12-20 different epicardial sites with a bipolar contact electrode similar to that already described.13 This electrode has been shown to reproduce with fidelity the
activation time and the time course for repolarization of adjacent cells.14 Recordings were made
sequentially from each site, using a detailed map of
the epicardial surface, to identify the locations from
which each recording was made. On average, two
thirds of the recordings were made from the left
ventricle and one third was made from the right
ventricle, with recordings sites spread out evenly
over each ventricle's surface as judged visually.
When the center electrode was pressed slightly
against the surface of the heart, monophasic action
potentials developed and stabilized within 2-6 beats,
allowing us to complete MAP mapping at up to 20
1414
Circulation Vol 80, No 5, November 1989
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
sites within 5 minutes. MAP recordings were amplified with a high-impedance, direct-current coupled,
differential preamplifier (model 1001, EP Technologies, Mountain View, California) that had a bandwidth of 0 to 5000 Hz. The preamplified MAP and
the ventricular electrogram were recorded on a
two-channel Gould 2200 S chart recorder at a paper
speed of 200 mm/sec.
After mapping studies had been performed at 30
and 45 minutes of atrial pacing, pacing was switched
to ventricular pacing and MAPs were recorded from
the same epicardial sites as during atrial pacing at 5,
30, 60, 90, and 120 minutes after the onset of
ventricular pacing. By simply switching the oscillator from one (the atrial) stimulus output channel to
another (the ventricular) output channel, the pacing
rate of 180 beats/min was maintained at exactly the
same value during the transition from atrial to
ventricular pacing. After 120 minutes of ventricular
pacing, the stimulus output was switched back to
the atrial-pacing wires (again under maintenance of
the same pacing rate), and MAP recordings were
repeated 5, 30, and 60 minutes after resumption of
atrial pacing. During each subsequent mapping study
within one experiment, the drawn epicardial surface
map was used to reproduce the recording sites of
the initial mapping study.
Control Studies
In four experiments, hearts were paced continuously from the right atrium alone, for 4 hours, and
epicardial MAP maps were obtained at the same
intervals as previously described. The purpose of
these control studies was twofold: first, to test the
reproducibility of the individual recording sites by
repeated measurements from the same sites using
the ventricular map (see "Results"), and second, to
examine the change of AT and APD in an isolated
Langendorff-perfused heart, in the absence of any
alteration of pacing site.
Metabolic and electrophysiologic stability of the
preparation was a major prerequisite for this study.
Flow, pressure, pH, temperature of the perfusate,
and temperature at five to eight epicardial sites were
measured at least every 30 minutes. Experiments in
which pH was below 7.32, above 7.48, or varied by
more than 0.07 and experiments in which temperature varied by more than 0.50 C between any epicardial sites or throughout the experiment were
excluded from analysis (n=2). Experiments also
were excluded from analysis if premature beats
(n=1), atrioventricular block during atrial pacing
(n=1), sinus-rhythm interference due to ventriculoatrial block during ventricular pacing (n=1), or
sinus frequency higher than the paced frequency
occurred (n=3). Before data analysis, the entire set
of MAP recordings obtained in a single experiment
was assessed for stability and quality. Stable MAP
recordings throughout the entire protocol from at
least 12 different epicardial sites were required to
include an experiment in statistical analysis; this
condition was not met in two experiments that were
subsequently discarded. These stringent criteria
allowed only 11 out of 21 preparations to be submitted to data analysis.
Data Analysis
Monophasic action potentials were analyzed for
AT, APD, and RT. AT during atrial pacing was
measured from the onset of QRS in the electrocardiogram, to the time when local activation occurs at
the ventricular recording site (upstroke of the MAP),
and during ventricular pacing from the stimulus
spike to the upstroke of the MAP. To compare AT
during atrial and ventricular pacing in any given
experiment, AT at every single epicardial site was
also analyzed as the delay from the site with earliest
activation. APD was measured from the intrinsic
deflection of the upstroke to the level of 90%
repolarization.9 All data from each site are the
average of five consecutive MAP recordings. Local
RT was calculated as the sum of AT and APD. The
accuracy of the measured intervals was within 2
msec as determined by repeated measurements
made by two independent observers. The two-tailed
Student's t test for paired data was used to compare
mapping data at different times of the pacing protocol in the same experiment. The degree of heterogeneity of ventricular repolarization (dispersion of
RT) was estimated from the coefficient of variance
of RT between all recordings obtained in a given
heart (minimum number of recordings, 10).
Results
Reproducibility and Stability of AT and APD, and
Their Interrelation During Baseline Conditions
Repeated mapping of the right and left ventricle
at intervals of 10 minutes, using the drawn diagram
of the heart, demonstrated good reproducibility of
AT and APD measurements at each of the 12-20
individual recording sites. AT differed between the
two consecutive measurements by only 1.0-1.5
msec (1.2+0.2 msec), and APD differed by only
2.1-3.0 msec (2.4+0.3 msec). To assess the electrophysiologic stability of the colloid-perfused isolated heart preparation, constant-rate right atrial
pacing was continued for a total of 4 hours and
measurements were repeated at 30 and 60 minutes,
and at 2, 3, and 4 hours. Throughout this period,
MAPs remained stable in amplitude and configuration (Figure 1), and AT measurements at the preselected sites remained constant within 2±2 msec
(range, 0-5 msec). APD remained unchanged (average, -0.1 msec) during the first hour of continuous
pacing but shortened by an average of 4±3 msec
during the second hour. At 3 and 4 hours, mean
APD was 11+5 msec and 20+5 msec, respectively,
shorter as compared with beginning pacing measurements. The standard deviations of the change in
APD from the start of pacing to 4 hours later ranged
from 2.0 to 3.8 msec (average, 2.7±0.8 msec),
Costard-Jdckle et al Myocardial Memory
1415
Atrial pacing
240 min
180 min
120 min
60 min
min
5
A 5
240 min
FIGURE 1.
Stability of monophasic action
potentials (MAPs) during 4 hours of atrial
pacing at rate of 180 beats/min. Depicted
A
1
Ms
169
1
17
are
MA4Ps
recorded
hourly from four dif-
ferent epicardial sites (A-D) in
experiment.
X
B
178
175
D
171ii
170
AA-D
15
13
171
\
t
165
173
170
ms
14
1
165
mns
14
160
X
160
14
<
< ms
12
12
one
given
recordings indi-
cate action potential duration (APD) at level
of90% ofrepolarization. MAPs are stable in
configuration and amplitude but shorten
slightly
over
This
time.
effect
is
relatively
homogenous (range ofAPD shortening after
4 hours of atnial pacing, 14-17 msec), maintaining a constant difference between longest (MAP A) and shortest (AMP D) MAP of
12-15 msec.
MS
157
Numbers within
14
ms
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
indicating that APD shortening occurring during
control experiments was relatively uniform over the
heart (Figure 1).
When APD was plotted as a function of AT, there
was a consistent inverse correlation between AT
and APD, such that shorter AT was associated with
longer APD and vice versa. This relation remained
significant throughout the entire right atrial-pacing
protocol. Measured at 30 and 60 minutes and then
every hour after the start of pacing, the slopes of the
linear regression between AT and APD remained
within a range of -1.05 to -1.41, and the range of
the corresponding linear-correlation coefficients was
0.71-0.67 (Table 1). Thus, although slight electrophysiologic deterioration of the preparation occurred
during the experiment, these changes were uniform
over the ventricular surface.
in the control experiments, APD was inversely correlated to AT with a linear-regression slope of greater
than -1. In the representative example given in
Figure 2 (panel A), the linear-regression slope for the
relation between AT and APD at 45 minutes of atrial
pacing was -1.73 (r, 0.87,pp<0.0001), indicating that
usually a site with later activation repolarized before a
site with earlier activation. Five minutes after the
switch to ventricular pacing, however, the inverse
linear correlation was lost (S, 0.16, r=0.25, NS) (panel
B). With continuing ventricular pacing, a significant
linear and inverse correlation between AT and APD
again developed (panels C and D), and the magnitude
of the negative slope increased progressively (ventricular pacing for 120 minutes: S, -0.51, r=0.73,
p<0.OOl).
Two hours after continuous constant-rate right
ventricular pacing was begun, the stimulus output
was switched back to the right atrial pacing wires to
reestablish the original supraventricular activation
sequence. Mapping 5 minutes after resumption of
atrial pacing again demonstrated loss of the inverse
correlation between AT and APD (S, +0.91, r=0.52,
p<O.OS). After 60 minutes of atrial pacing, however, a significant inverse correlation was again
restored (S, -0.63, r=0.56, p<0.02).
Table 2 summarizes the changes in correlation
coefficient, and Table 3 summarizes the slope
Changes in the AT/APD Relation at Various
Intervals During Atrial and Ventricular Pacing
The effect of an alteration in ventricular activation
sequence on ventricular repolarization was determined by repeated mapping after a change in pacing
site. The first two mapping studies were performed
after 30 and 45 minutes of continuous right atrial
pacing at a rate of 180/minute. This phase presumably
continued the supraventricular activation sequence
experienced by the ventricles before explantation. As
TABLE 1. Linear-Regression Analysis for Relation Between AT and APD in Control Experiments (n=4)
30 min
r
1
2
3
4
Mean
SEM
r
0.73
0.66
0.71
0.73
0.71
0.02
S
-0.84
-0.90
-1.19
-1.25
-1.05
0.10
45 min
r
0.81
0.82
0.64
0.73
0.75
0.04
r, correlation coefficient; S, slope.
S
-0.90
-1.64
-1.41
-1.69
-1.41
0.18
1 hr
r
0.68
0.83
0.58
0.62
0.68
0.06
S
-0.91
-1.4
-1.02
-1.15
-1.12
0.11
3 hr
2 hr
r
0.65
0.67
0.64
0.72
0.67
0.02
S
-1.05
-1.16
-1.38
-1.92
-1.38
0.19
r
0.73
0.65
0.63
0.72
0.68
0.03
S
-1.42
-1.02
-1.51
-1.28
-1.31
0.11
4 hr
r
0.66
0.72
0.64
0.67
0.67
0.02
S
-0.83
-1.19
-1.16
-1.38
-1.14
0.11
Circulation Vol 80, No 5, November 1989
1416
A 2001
EB 200-
1. Atrial pacing 45 min
Ventricular pacing 5 min
190
190
E
E
C) 180-
in 180-
*
170-
170
160
160 t
0
1
5
15
10
y -180 +0.16 x R=-0.25 NS
20
AT (ms)
10
AT (ma)
C
200.
D 200 l
Ventricular pacing 60 min
190
30
40
Ventricular pacing 120 min
190
FIGURE 2. Plots of linear-regression analysis of relation between activation time (AT)
and action potential duration (APD) at various times of right atrial-right ventricularright atrial pacing protocol.
m
a 180-
180'
0.
4Z
170-
170 -
*
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
a=800
y = 180
037 x
m
R =0.51
p
160
E
y=187 - 0.51 x R = 0.73 p < .001
<.05
16U
30
20
AT (ms)
10
*
F
200'
2. Atrial pacing 5 min
.
10
40
190-
200-
20
AT (ms)
30
40
2. Atrial pacing 60 min
190'
E
E
a 180'
*
a.
180'
a
#'U
170-
170-
y=168 0.91 x R=0.52 p<.05
itn -L
160
+
160 .i
0
.t
10
5
15
c
y=176-0.63x R=0.56
AT (ms)
p<
10
5
.02
15
AT (ms)
changes of the linear-regression analysis for the
relation between AT and APD at each of the nine
mapping studies during the entire pacing protocol.
These data confirm the findings in the example
shown in Figure 2, that is, immediately after alteration of the ventricular activation sequence, the
inverse correlation between AT and APD diminished and only gradually recovered with prolonged
pacing from the ectopic site, and the reverse phenomenon occurred when the supraventricular activation sequence was reestablished.
Cause and Effect in the Changing Correlation
Between AT and APD
Abruptly changing the pacing site from a right
atrial to a right ventricular site causes a rapid
change in the ventricular activation sequence, but
once changed, the ventricular activation sequence
should remain stable for as long as the pacing site
is kept constant. Accordingly, changes in the
linear regression that occurred during the period
of pacing at a constant site should reflect changes
in APD rather than changes in AT. This was tested
by analyzing the degree of change in AT and APD
at each of the epicardial mapping sites at the
beginning and end of each right atrial and right
ventricular pacing period. The data presented in
Table 4 confirm that switching the pacing site from
right atrial to right ventricular, and vice versa,
produced a sudden marked change in the average
AT, with relatively little change in average APD,
whereas during continuing pacing from either the
right atrium or right ventricle, only APD underwent
progressive, significant change, and AT remained
essentially constant. These data indicate that the
progressive increase in the negative slope and the
improvement of the correlation coefficient of the
linear regression analysis between AT and APD
during prolonged alterations in activation sequence
cannot be explained by changes in AT but are caused
by a slow adaptation of APD to the altered activation
sequence.
Figure 3 further illustrates the slow time course
with which APD at two different sites (sites A and
B) changed when the activation sequence was
abruptly altered. Altering the activation sequence
by switching to right ventricular pacing lengthened
the AT at site A (left ventricular, midseptal) by 17
msec and shortened it at site B (right ventricular,
basal) by 4 msec. Changes in AT occurred immedi-
Costard-Jackle
et
al Myocardial Memory
1417
TABLE 2. Correlation Coefficients From Linear-Regression Analysis for Relation Between AT and APD During Changes in Ventricular
Activation Sequence (n=9)
Atrial pacing (2)
Ventricular pacing
Atrial pacing (1)
60 min
30 min
90 min
120 min
60 min
30 min
5 min
45 min
5 min
30 min
Rabbit
0.43
0.24
0.65
0.15
0.82
0.67
0.57
0.64
0.54
1
0.57
0.48
0.82
0.52
0.75
0.62
0.71
0.49
0.59
...
2
0.71
0.48
0.06
0.30
0.21
0.82
0.65
0.20
0.73
0.07
0.73
3
0.62
0.85
0.55
0.66
0.73
...
...
0.31
0.36
...
4
0.71
0.44
0.57
0.43
0.75
0.52
0.42
0.85
0.50
0.87
5
0.40
0.62
0.77
0.42
0.83
0.82
0.74
0.83
0.05
0.62
6
0.61
0.43
0.70
0.14
0.51
0.12
0.39
0.39
0.86
0.84
7
0.56
0.46
0.73
0.52
0.87
0.55
0.51
0.83
0.25
0.15
8
0.31
0.50
0.83
0.69
0.18
0.67
0.32
0.42
0.42
0.63
9
0.53§
0.41
0.68t
0.354
0.72
0.58t
0.76
0.31*
0.41
Mean
0.75
0.04
0.04
0.06
0.06
0.06
0.04
0.07
0.06
0.04
0.04
SEM
*p<0.05 versus 45 min atrial pacing; tp<0.05 versus 5 min ventricular pacing; tp<0.05 versus 120 min ventricular pacing;
versus 5 min second atrial pacing.
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
ately after the change from atrial to ventricular
pacing, and once changed, AT remained constant as
long as the pacing site was kept constant. In contrast, changes in APD developed very slowly over
the ensuing 120 minutes of constant right ventricular pacing. There was a gradual, small increase in
APD at site B, which had undergone a small decrease
in AT, and there was a slow but more pronounced
decrease in APD at site A, which had undergone a
greater increase in AT. When, after 120 minutes of
ventricular pacing, pacing was switched back to the
atrium, AT at sites A and B immediately returned to
values very similar to those occurring during the
original supraventricular activation sequence. In
contrast, APD at both sites changed very slowly
and in a direction opposite to that measured during
the previous, reverse change in activation sequence.
Effect of APD Adaptation on Dispersion of
Ventricular Repolarization
An inverse correlation between AT and APD
should make RT more uniform within the heart.
§p<0.05
Figure 4 summarizes the changes in the coefficient
of RT variance, used as an estimate of repolarization dispersion at the beginning and end of each
right atrial and right ventricular pacing period,
averaged from all nine experiments. Immediately
after the switch to ventricular pacing, the variability of RT among the mapping sites markedly
increased, indicating that the loss of the inverse
correlation between AT and APD leaves the repolarization sequence in a less organized, more
dispersed state. With continuing pacing from the
same right ventricular site, variability of RT
decreased significantly, increased again when right
atrial pacing was resumed, and again diminished
with continued right atrial pacing. As seen by
comparison (Figure 4 and Table 3), a decrease in
dispersion parallels an improvement in the inverse
correlation between AT and APD.
Discussion
This study presents several intriguing findings.
The first is the inverse correlation between AT and
TABLE 3. Slopes From Linear-Regression Analysis for Relation Between AT and APD During Changes in Ventricular Activation Sequence
(n=9)
Atrial pacing (2)
Ventricular pacing
Atrial pacing (1)
60 min
5 min
30 min
120 min
5 min
30 min
60 min
90 min
45 min
30 min
-1.40
-0.88
-0.97
-0.80
-1.12
-0.86
-1.00
-0.96
-0.69
-0.71
1.79
-3.00
-0.90
...
-0.60
-0.76
-0.42
-0.37
-3.00
-3.01
-0.10
-0.87
-0.37
-0.29
-0.55
-0.69
-1.38
-0.07
0.23
-1.38
-1.10
-0.80
..
-1.10
-0.89
-0.27
-0.82
...
-0.35
...
-1.82
-1.07
-0.97
-0.96
-0.70
-0.38
-0.32
-0.38
-1.12
-1.15
-0.95
1.17
2.01
-0.82
-0.57
-0.71
-0.03
-0.55
-1.10
-1.46
-1.41
-1.13
-0.34
-0.43
-0.46
0.40
-0.32
0.14
-2.29
-1.62
0.56
0.91
0.46
-0.51
-0.39
-0.37
0.09
0.16
-1.73
-2.13
-1.67
-0.76
-0.51
-0.42
-0.37
-0.67
0.00
0.00
-1.69
-2.08
-0.82
0.28t
-0.21
-0.71t
-1.25§
-0.19*
-1.69
-1.63*
Mean
-0.71t
-0.55t
0.17
0.36
0.37
0.10
0.07
0.07
0.11
0.25
0.26
0.25
SEM
ventricular
120
versus
ventricular
min
versus
5
atrial
min
versus
45
§p<0.05
pacing;
min
pacing;
tp<0.05
pacing; tp<0.05
*p<0.05
versus 5 min second atrial pacing.
Rabbit
1
2
3
4
5
6
7
8
9
Circulation Vol 80, No 5, November 1989
1418
Anterior
atrial
pacing
Site B
ATatr pac
ATvtr. pac.
Site A
ATatr pac. = 5 ms
ATvtr. pac. = 22 ms
= 16
12
FIoURE 3. Upper panel: Epicardial surface diagram
=
with recording sites A (left ventricular, midseptal) and B
(right ventricular, basal). Switch in pacing site from atrial
to ventricular pacing caused lengthening of activation
time (AT) at site A and shortening at site B. Lower panel:
Slow time course of adaptation of action potential duration (APD) to sudden change in ATafter switch in pacing
site. (For further explanation, see text.)
ventricular
pacing
-a
C:
c
CL
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
.5
c
a
E
5'
45'
Atrial pacing
30' 60' 90'
Ventricular pacing
120'
60 minutes
30
Atrial pacing
5'
APD in ventricles that are depolarized along their
normal supraventricular activation pathway. The
second more intriguing finding is that the inverse
relation between AT and APD appears to be the
result of a very slow mechanism that conditions
ventricular repolarization according to the sequence
of ventricular activation. If ectopic ventricular pacing is sustained long enough, APD adjusts to the
new activation sequence in such a way as to correlate again with AT in an inverse fashion. Thus, the
opposing direction of activation and repolarization
sequence, which is the basis of the normal concordant T wave, does not seem to be an a priori
relation but is acquired by the heart during constant
activation sequence.
An inverse relation between AT and indexes of
repolarization has been reported previously from
isolated and in vivo canine heart studies15-18 and
for both the endocardial and epicardial surface of
the human heart.19 The significance of this inverse
relation is that it helps explain the concordance of
8-
6-
TABLE 4. Average Change in Activation Time (AT) and Action
Potential Duration (APD) During Right Atrial-Right VentricularRight Atrial Pacing Protocol
AT
A
18-+-8
pc0.0001
APD
5+5
B
22
NS
12±7
C
T
c2)
CD
0
4.
D
rc_
19+9
pc0.0001
5+5
NS
1+1
NS
8+6
p<0.0001
p<0.005
p<0.02
A, change from end of first atrial pacing to 5 min after the onset
of right ventricular pacing; B, change from 5 min to 120 min of
continuous ventricular pacing; C, change from 120 min of
ventricular pacing to 5 min of second atrial pacing; D, change
from 5 to 60 min of continuous second atrial pacing. Change is
measured in absolute terms. Data are expressed in msec+SD.
Each average is based on data from a total of 124 epicardial sites
(nine experiments, each 10 to 18 sites).
2
0O
45
atrial pacing
5
ventricular
1
20
pacing
5
atrial
60
minutes
pacing
FIGURE 4. Bar graph of coefficient of variance of repolarization time (RT) at various times ofpacingprotocoL
Data represent mean from nine experiments.
Costard-Jackle et al Myocardial Memory
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
the T wave. In the human heart, the slope of this
inverse correlation averaged -1.32, which corresponds well with the relation between AT and
APD in the present isolated rabbit heart study;
during the initial atrial pacing phase, the average
linear-regression slope was -1.63. Similar to
humans, the rabbit body surface electrocardiogram has a concordant T wave20 consistent
with an inverse relation between depolarizing and
repolarizing electrical forces.
Mechanisms of Activation-Dependent Modulation
of Repolanzation
A mechanism proposed to explain the effect of
activation sequence on APD or repolarization is
electrotonic interaction. Studies in isolated ventricular muscle and Purkinje fiber preparations have
shown that repolarizing (anodal) currents applied
during repolarization shorten the action potential21-23
and that depolarizing (cathodal) currents applied
during repolarization prolong the action potential.22
Because electrotonic interaction is a passive electrical phenomenon with only minimal (capacitative)
delay, this mechanism cannot alone be responsible
for the slowly developing and persistent effects on
repolarization. One therefore has to postulate
another factor that modulates (strengthens or weakens) the degree of electrotonic interaction between
myocardial cells. Cardiac cells are electrically coupled by intercellular gap junctions that provide low
resistance pathways.24,25 It is conceivable that
repeated current flow in the same direction through
these junctions may decrease their resistance even
further, thereby amplifying the electrotonic effect.
If conditioning of cellular electrical communication
through repeated, uniform use is indeed responsible
for the slow modulation of repolarization (and its
retention), myocardium can be thought to possess
some form of rudimentary "memory." Further basic
investigation is needed to identify the cellular or
molecular basis of this previously unrecognized
form of intercellular communication in the heart.
Study Limitations
This study has several limitations. First, because
recordings were made only from the epicardium, we
could not assess the influence of transmural activation on the epicardial APD. Second, differences in
AT between different epicardial sites of the rabbit
heart were relatively small, which is not surprising
considering that the entire rabbit ventricle is depolarized in less than 20 msec.20 Although we attempted
to "perfect" our analysis of MAP recordings by
using high-speed recordings and stringent criteria13
for determining the onset of local activation and
repolarization, the short time frame within which
the rabbit ventricle depolarizes probably increased
the "noise level" of our data. Third, the overall
observation period in our preparations was limited
to a maximum of 4 hours. Slight but progressive
shortening of APD in the isolated rabbit heart
1419
preparation prevented us from extending the recording period. The clinical study of Rosenbaum et a15
showed that pacing-induced primary T wave changes
require 24-48 hours to fully develop, and that an
even longer time of normal activation sequence is
required for them to disappear. It is possible that
the development of the inverse relation between
depolarization and repolarization would have progressed even farther (steepening its slope and improving its correlation coefficient), if pacing could have
been continued for longer times in our Langendorff
heart preparation.
Clinical Implications
The dependence of APD on activation sequence,
and the long time required to establish such close
correlation, has several important implications. An
inverse correlation between AT and APD tends to
compensate for the successive delay along the ventricular activation pathway and, consequently, tends
to synchronize ventricular repolarization. Commensurate with this postulate, we found less dispersion
of repolarization after prolonged pacing from the
same site, when the inverse correlation between AT
and APD was highly significant, than we found
shortly after a change in ventricular activation
sequence, when the inverse correlation disappeared. Accordingly, a heart that is activated along
the same activation pathway for a long time (as is
the normal heart with supraventricular impulse origin and conduction over the His-Purkinje system)
can be expected to have a more synchronized global
repolarization of the ventricular myocardium than a
heart shortly after a change in activation sequence,
as in ventricular extrasystoles or short runs of
ventricular tachycardia. By the same rationale, the
short bursts of ventricular pacing commonly used
during electrophysiologic testing would perturb the
natural synchrony of ventricular repolarization. The
resulting increase in dispersion of ventricular repolarization could add to the arrhythmia propensity of
the ventricles.26
The activation sequence-dependent modulation
of ventricular repolarization may explain electrocardiographic repolarization abnormalities after a period
of aberrant conduction or ectopic activity. Patients
with intermittent left bundle branch block often
demonstrate T wave inversion on resumption of a
normal ventricular excitation,7,8 and this persistent
T wave change may reflect adaptation of ventricular
APD to the previous aberrant conduction pattern.
Similarly, patients with ventricular preexcitation,11,27
tachycardia,8,9 and repeated uniform extrasystoles10,27,29 may demonstrate T wave changes long
after normalization of their rhythm. Our study suggests that, contrary to previous contentions, these
fascinating phenomena are based on a physiologic
property of the myocardium and do not necessarily
imply pathophysiologic conditions such as ischemia, autonomic reflex alterations, or pacinginduced myocardial injury.
1420
Circulation Vol 80, No 5, November 1989
Finally, as already implied by Rosenbaum and
coworkers,5,27 our study suggests that accumulation and retention of repolarization changes dictated by the activation sequence plays a dominant
role in the creation of the ventricular gradient.
Secondary T wave changes, formerly believed to
be simply a reflection of an abnormal activation
sequence without changes in APD, thus may not
be limited to the period of abinormal activation but
may also involve changes in APD, provided that
abnormal activation lasts long enough to condition
these changes. Accordingly, a dogmatic categorization of T wave changes into "primary" and
"secondary" can no longer be upheld.
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Acknowledgments
The authors thank Byron W. Brown, PhD, for
valuable advice during the statistical analysis,
and Cecil Profitt and Robert Kernoff for their
technical assistance.
References
1. Wilson FN, MacLeod AG, Barker PS: The T deflection of
the electrocardiogram. Trans Assoc Am Physicians 1931;
46:29-38
2. Surawicz B: The pathogenesis and clinical significance of
primary T wave abnormalities, in Schlant RC, Hurst JW
(eds): Advances in Electrocardiography. New York/
London, Grune & Stratton, 1972, pp 377-421
3. Wilson FN, MacLeod AG, Barker PS, Johnston F-D: The
determination and the significance of the areas of the ventricular deflections of the electrocardiogram. Am Heart J
1934;10:46-61
4. Chatterjee K, Harris A, Davies G, Leatham A: Electrocardiographic changes subsequent to artificial ventricular depolarization. Br Heart J 1969;31:770-779
5. Rosenbaum MB, Blanco HH, Elizari MV, Lazzari JO,
Davidenko JM: Electrotonic modulation of the T wave and
cardiac memory. Am J Cardiol 1982;50:213-222
6. Denes P, Pick A, Miller RH, Pietras RJ, Rosen KM: A
characteristic precordial repolarization abnormality with intermittent left bundle-branch block. Ann Intern Med 1978;
89:55-57
7. Surawicz B: Transient T wave abnormalities in intermittent
bundle branch block. Am J Cardiol 1982;50:363-364
8. Lepeschkin E: Modern Electrocardiography. Baltimore, Williams & Wilkins, 1951, pp 239, 420.
9. Kernohan RJ: Post-paroxysmal tachycardia syndrome. Br
Heart J 1969;31:803 -806
10. Levine HD, Lown B, Streeper RB: The clinical significance of postextrasystolic T wave changes. Circulation
1952;6:538-548
11. Nicolai P, Medvedovsky JL, Delaage M, Barnay C, Blach E,
Pisapia A: Wolff-Parkinson White syndrome: T-wave abnormalities during normal pathway conduction. JElectrocardiol
1981;14:295-300
11. Hoffman BF: Electrotonic modulation of the T
wave. Am J
Cardiol 1982;50:361-362
12. Kates RE, Yee YG, Hill I: Effect of albumin on the electrophysiological stability of isolated perfused rabbit hearts. J
Cardiovasc Pharmacol 1989;13:168-172
13. Franz MR: Long-term recording of monophasic action
potentials from human endocardium. Am J Cardiol 1983;
51:1629-1634
14. Franz MR, Burkhoff D, Spurgeon H, Weisfeldt ML, Lakatta
EG: In vitro validation of a new cardiac catheter technique
for recording monophasic action potentials. Eur Heart J
1984;7:34-41
15. Abildskov JA: The sequence of normal recovery of excitability in the dog heart. Circulation 1975;52:442-446
16. Toyoshima H, Burgess MJ: Electrotonic interaction during
canine ventricular repolarization. Circ Res 1978;43:348- 356
17. Autenrieth G, Surawicz B, Kuo CS: Sequence of repolarization on the ventricular surface in the dog. Am Heart J 1975;
89:463-469
18. Toyoshima H, Lux RL, Wyatt RF, Burgess MJ, Abildskov
JA: Sequences of early and late phases of repolarization on
dog ventricular epicardium. JElectrocardiol 1981;14:143 -152
19. Franz MR, Bargheer K, Raffienbeul W, Haverich A, Lichtlen: Monophasic action potential mapping in human subjects
with normal electrocardiograms: Direct evidence for the
genesis of the T wave. Circulation 1987;75:379-386
20. Slapak L, Hermanek P: Beobachtungen uber das Elektrokardiogramm des Kaninchens. Z Kreislaufforschung 1957;
46:136-146
21. Weidmann S: Effect of current flow on the membrane
potential of cardiac muscle.JPhysiol (Lond) 1951;115:227-236
22. Cranefield PF, Hoffman BF: Propagated repolarization in
heart muscle. J Gen Physiol 1958;41:633-649
23. Vassalle M: Analysis of cardiac pacemaker potentials using a
"voltage clamp" technique. Am J Physiol 1966;210:1335-1341
24. De Mello WC: Intercellular communication in cardiac muscle. Circ Res 1982;50:2-35
25. Berger WK: Correlation between the ultrastructure and
function of intercellular contacts, in De Mello WC (ed):
Electrical Phenomena in the Heart. New York/London,
Academic Press, 1972, pp 63-88
26. Kuo CS, Munakata K, Reddy CP, Surawicz B: Characteristics and possible mechanism of ventricular arrhythmia
dependent on the dispersion of action potential durations.
Circulation 1983;61:1356-1367
27. Rosenbaum MB, Sicouri SJ, Davidenko JM, Elizari MV:
Heart rate and electrotonic modulation of the T wave: A
singular relationship, in Zipes DP, Jalife J (eds): Cardiac
Electrophysiology. New York/London, Grune & Stratton,
1985, pp 485-488
28. Engel JR, Shah R, DePodesta LA, Franke WS, Krause RL:
T wave abnormalities of intermittent left bundle-branch
block. Ann Intern Med 1978;89:204-206
29. Rosenbaum MB, Blanco HH, Elizari MV, Lazzari JO,
Vetulli HM: Electrotonic modulation of ventricular repolarization and cardiac memory, in Rosenbaum MB, Elizari MV
(eds): Frontiers of Cardiac Electrophysiology. Boston, Martinus Nijhoff Publishers, 1983, pp 67-99
KEY WORDS * action potential duration * T wave * cardiac
memory * ventricular function
Slow and long-lasting modulation of myocardial repolarization produced by ectopic
activation in isolated rabbit hearts. Evidence for cardiac "memory".
A Costard-Jäckle, B Goetsch, M Antz and M R Franz
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Circulation. 1989;80:1412-1420
doi: 10.1161/01.CIR.80.5.1412
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1989 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/80/5/1412
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/