Role of Structural Barriers in the Mechanism of

Alternans-Induced Reentry
Joseph M. Pastore, David S. Rosenbaum

Abstract—Previously, using an animal model of T-wave alternans in structurally normal myocardium, we demonstrated
that repolarization can alternate with opposite phase between neighboring myocytes (ie, discordant alternans), causing
spatial dispersions of repolarization that form the substrate for functional block and reentrant ventricular fibrillation
(VF). However, the mechanisms responsible for cellular discordant alternans and its electrocardiographic manifestation
(ie, T-wave alternans) in patients with structural heart disease are unknown. We hypothesize that electrotonic
uncoupling between neighboring regions of cells by a structural barrier (SB) is a mechanism for discordant alternans.
Using voltage-sensitive dyes, ventricular action potentials were recorded from 26 Langendorff-perfused guinea pig
hearts in the absence (ie, control) and presence of an insulating SB produced by an epicardial laser lesion. Quantitative
analysis of magnitude and phase of cellular alternans revealed that in controls, action potential duration alternated in
phase at all ventricular sites above a critical heart rate (269⫾17 bpm), ie, concordant alternans. Also, above a faster
critical heart rate threshold (335⫾24 bpm), action potential duration alternated with opposite phase between sites, ie,
discordant alternans. In contrast, only discordant but not concordant alternans was observed in 80% of hearts with the
SB, and discordant alternans always occurred at a significantly slower heart rate (by 68⫾28 bpm) compared with
controls. Therefore, the SB had a major effect on the alternans– heart rate relation, which served to facilitate the
development of discordant alternans. Whether a SB was present or not, discordant alternans produced considerable
increases (by ⬇170%) in the maximum spatial gradient of repolarization, which in turn formed the substrate for
unidirectional block and reentry. However, by providing a structural anchor for stable reentry, discordant alternans in
the presence of a SB led most often to sustained monomorphic ventricular tachycardia rather than to VF, whereas in the
absence of a SB discordant alternans caused VF. SBs facilitate development of discordant alternans between cells with
different ionic properties by electrotonically uncoupling neighboring regions of myocardium. This may explain why
arrhythmia-prone patients with structural heart disease exhibit T-wave alternans at lower heart rates. These data also
suggest a singular mechanism by which T-wave alternans forms a substrate for initiation of both VF and sustained
monomorphic ventricular tachycardia. (Circ Res. 2000;87:1157-1163.)
Key Words: repolarization 䡲 reentry 䡲 alternans

T -wave alternans is a beat-to-beat fluctuation in the

amplitude of the electrocardiographic T wave that re-
peats once every other beat and has been closely associated
with ventricular arrhythmias and sudden cardiac death. Visu-
ally apparent T-wave alternans is known to hasten cardiac
electric instability in a surprisingly wide variety of experi-
mental1– 4 and clinical5–7 conditions. More recently, using
sensitive signal-processing techniques, we found that the
detection of microvolt-level, visually inapparent T-wave
alternans is also a strong predictor of vulnerability to ventric-
ular arrhythmias in patients.8 Interestingly, T-wave alternans
has been closely associated with polymorphic ventricular
tachycardia (VT),9 ventricular fibrillation (VF),10 torsade de
pointes,7,11 and sustained monomorphic VT (SMVT)8 both in
the absence of structural heart disease and in patients with
advanced cardiomyopathy. Therefore, a greater understand-
ing of the mechanisms underlying T-wave alternans may
provide important new insights into the pathophysiology of
sudden cardiac death in a variety of clinical and experimental
circumstances.
Recently, we used an experimental model to demonstrate
that T-wave alternans is caused by primary alternations in the
repolarization phase of the action potential.10 Importantly,
above a critical heart rate threshold, repolarization of action
potentials from adjacent regions of the ventricle alternated
with opposite phase, ie, discordant alternans, causing steep
spatial gradients of repolarization that formed an electrophys-
iological substrate for functional block, reentry, and VF.10 It
is noteworthy that discordant alternans occurred under con-
ditions of normal intercellular coupling, indicating that re-

Received August 30, 2000; revision received October 6, 2000; accepted October 6, 2000.
From the Heart and Vascular Research Center and Department of Biomedical Engineering, MetroHealth Campus, Case Western Reserve University,
Cleveland, Ohio.
Correspondence to David S. Rosenbaum, MD, Director, Heart and Vascular Research Center, MetroHealth Campus, Case Western Reserve University,
2500 MetroHealth Dr, Hamman 322, Cleveland, OH 44109-1998. E-mail drosenbaum@metrohealth.org
© 2000 American Heart Association, Inc.
Circulation Research is available at http://www.circresaha.org

1157
1158 Circulation Research December 8/22, 2000
gional differences in cellular ionic properties can overcome
electrotonic effects that normally act to synchronize repolar-
ization.12,13 However, the mechanisms responsible for trig-
gering discordant alternans are unknown.
A major unresolved question is why patients with struc-
tural heart disease and T-wave alternans are at particularly
high risk for ventricular arrhythmias. Clearly, patients with
contractile dysfunction constitute the vast majority of cases of
sudden cardiac death. Among patients with structural heart
disease, those with T-wave alternans are more susceptible to
inducible SMVT8 and spontaneous arrhythmic events.8,14
However, the role of structural heart disease in the mecha-
nisms of alternans-induced reentry is unknown. Most forms
of myocardial disease are associated with fibrotic barriers,15
alterations of fiber bundle architecture,16 and maldistribution
of cardiac gap junctions,17 all of which are expected to
produce electric uncoupling between fibers. Because cell-to-
cell coupling plays an important role in synchronizing repo-
larization between cells, we hypothesize that electrotonic
uncoupling by structural barriers (SBs) facilitates the devel-
opment of discordant alternans. This may explain a mecha-
nism by which SBs promote the development of critical
dispersions of repolarization that cause functional block and
reentry.
To test this hypothesis, we used high-resolution optical
mapping in an experimental model of T-wave alternans. This
model offered several advantages, as follows. (1) Action
potentials could be recorded simultaneously while T-wave
alternans was elicited on the ECG. (2) SB could be created
with precise control using an argon ion laser in a preparation
with previously characterized heterogeneities of repolariza-
tion. (3) Propagation, functional block, and dispersions of
repolarization could be mapped directly with high spatial
resolution. (4) In this model, both monomorphic and poly-
morphic arrhythmias could be induced. We found that elec-
trotonic uncoupling produced by a SB greatly facilitates the
development of discordant alternans, creating critical hetero-
geneities of repolarization that can form the substrate for a
variety of reentrant arrhythmias. These data provide addi-
tional evidence linking T-wave alternans to a novel mecha-
nism of arrhythmogenesis.
Materials and Methods
Experiments were performed in the guinea pig model of pacing-
induced T-wave alternans10 in which a thin (⬇800 ␮m) viable rim of
left and right ventricular epicardium having normal electrophysio-
logical properties was created by endocardial cryoablation in
Langendorff-perfused hearts.10,18 After staining with voltage-
sensitive dye, hearts were placed in an imaging chamber that
maintained uniform temperature, eliminated motion artifacts without
altering electrophysiological properties, and allowed ECG record-
ings from the volume conductor bath.10,18,19 To uncouple neighbor-
ing regions of cells having known differences in cellular restitution
kinetics,20 an insulating SB was produced by a computer-driven
argon laser to produce a 2⫻10-mm SB (Figure 1) spanning from the
epicardial surface transmurally to the endocardial cryoablation.
We have developed an optical action potential mapping system
that is capable of resolving membrane potential changes as small as
0.5 mV simultaneously from 128 sites across the anterior epicardial
surface of the guinea pig ventricle.18,20,21 An optical magnification of
⫻1.2 was used that corresponded to a mapping area of 15⫻15 mm,
1.25-mm interpixel spatial resolution, and 1-ms temporal resolution.
Figure 1. Langendorff-perfused guinea pig heart was stained
with voltage-sensitive dye and action potentials were mapped
from a 1.5⫻1.5-cm mapping area (square grid). Left ventricle
was stimulated (square pulse symbol) over a broad range of
CLs in absence (ie, control) and presence of a SB. ECGs were
recorded from electrodes immersed in the coronary effluent. RA
indicates right atrium; LA, left atrium; RV, right ventricle; LV, left
ventricle; and LAD, left anterior descending coronary artery.
Repolarization time (RT), activation times, and action potential
duration (APD) were determined from previously described algo-
rithms.18,20 Alternans of cellular RT was determined by measuring
differences in local RT on consecutive beats. Because the error
associated with measurement of RT was 3⫾4 ms, only RT alternans
⬎10 ms was considered significant. A previously validated spectral
analysis technique10 was used to measure ECG T-wave alternans.
The method of Bayly et al22 was modified for optically recorded
action potential maps to accurately quantify the magnitude and
direction of conduction velocity and repolarization gradients at each
recording site.
In all hearts (n⫽26), action potentials were recorded (Figure 1)
over a broad range of steady-state pacing cycle lengths (CLs) until
one-to-one capture failed or an arrhythmia was initiated. In each
experiment, the left ventricular free wall was stimulated near the
atrioventricular groove (Figure 1) to ensure that the sequence of
epicardial activation was nearly identical before and after the
introduction of the SB. This stimulation protocol was repeated in
each experiment to confirm reproducibility of the findings. To ensure
that laser energy did not alter properties of cells outside of the
precisely demarcated SB, validation studies were performed in 15
hearts. Each heart served as its own control, as the identical
stimulation protocol (described above) was performed before (ie,
control) and after a SB was introduced. APD, repolarization gradi-
ents, conduction velocity, and ECG morphology were compared at
the baseline stimulation CL of 400 ms in the absence and presence
of a SB. The effects of SB on cellular alternans was determined in a
subset of 5 of these experiments in which quantitative analysis of the
magnitude and phase of cellular alternans was compared at each
steady-state CL both before and after the introduction of a SB.
An expanded Materials and Methods section can be found in an
online data supplement available at http://www.circresaha.org.
Results
Effect of SB on Baseline Electrophysiology
Although the electrophysiological properties of guinea pig
epicardium were described in detail previously,19,20 it was
important to confirm whether introduction of a SB changed
the baseline characteristics of the model. Conduction velocity
and APD were compared at a baseline heart rate of 150 bpm
during control and in the presence of the SB. Conduction-
velocity transverse to fibers measured during control (24⫾5
cm/second) and in the presence of the SB (25⫾7 cm/second)
did not differ significantly (P⫽0.79, Wilcoxon matched pairs
 Pastore and Rosenbaum
Figure 2. Effect of SB on repolarization alternans. A, Relation-
ship between local repolarization alternans and heart rate in
absence (ie, control, open circles) and presence (filled circles) of
SB. Optically recorded action potentials shown were recorded
from the same site (filled pixel, Figure 1) as heart rate varied.
Notice in both cases that repolarization alternans occurs above
a threshold heart rate that is lower with SB (240 bpm) compared
with control (300 bpm). Repolarization alternans is also demon-
strated by superposition of even and odd (ie, sequential) action
potentials in panel C. B, Effect of the SB on the discordant
alternans heart rate threshold from 5 experiments. Notice that
presence of SB significantly decreases the heart rate at which
discordant alternans is elicited.
test). Furthermore, the SB did not significantly change
baseline APD (Wilcoxon matched pairs test) on either the
basal (208⫾14 before compared with 222⫾11 ms after the
SB, P⫽0.07) or the apical (197⫾13 compared with 213⫾8
ms, P⫽0.08) side of the SB, nor was apex-to-base APD
gradient changed by the SB (4.57⫾1.37 compared with
4.79⫾0.93 ms/mm).
There was no evidence of ischemia either during control or
after the introduction of a SB as confirmed by the absence of
triangulated action potentials and by histological examination
using viability staining.
Effect of SB on Cellular Alternans
The effect of regional uncoupling produced by a SB on the
magnitude and heart rate threshold of action potential alter-
nans is illustrated in Figure 2. Panel A shows a representative
example of the change in cellular alternans– heart rate relation
at one ventricular site caused by a SB. Such changes were
indicative of cellular alternans at all recording sites. At
relatively slow heart rates (⬍200 bpm), no cellular alternans
is present either in control or in the presence of SB. However,
above a critical threshold heart rate, RT alternans occurs and
increases markedly as heart rate further increases. Notice that
the heart rate threshold for RT alternans is considerably lower
in the presence of a SB (240 bpm) compared with control
(300 bpm). Consequently, over a broad range of heart rates,
the magnitude of RT alternans is amplified in the presence of
a SB. The difference in cellular alternans produced by the SB
is further emphasized by the superposition of even and odd
action potentials shown in panel C. Note that the heart rate
dependence of repolarization alternans in individual cells was
profoundly influenced by the insulating SB located some
distant from the cell. The SB also decreased the heart rate
threshold for T-wave alternans measured from the ECG (not
Alternans-Induced Reentry 1159
shown), suggesting that SBs can similarly explain changes in
heart rate alternans relation manifest on the surface ECG.
Effect of SB on Discordant Alternans
Between Cells
In this investigation, the alternans heart rate threshold was
defined as the slowest heart rate at which significant alternans
was recorded at one or more ventricular sites, and the heart
rate threshold for discordant alternans occurred at the heart
rate at which 2 sites had significant alternans with opposite
phase between them. As shown in Figure 2B, the SB
significantly (P⬍0.05, Wilcoxon matched pairs test) de-
creased the heart rate threshold for discordant alternans (by
68⫾28 bpm) compared with controls. Also, in controls, a
progression from no alternans to concordant alternans to
discordant alternans always occurred as heart rate increased.
In contrast, in the presence of a SB, a progression directly
from no alternans to discordant alternans was observed in
80% of experiments. Consequently, the heart rate that pro-
duced discordant alternans in hearts with SBs (276⫾6 bpm)
was essentially the same as the heart rate that produced
concordant alternans in control hearts (269⫾17 bpm,
P⫽0.35). Therefore, a SB greatly increased the propensity for
discordant alternans. Preliminary experiments performed
over long time periods (⬎3 hours) without introduction of a
SB (ie, time controls) demonstrated stable electrophysiolog-
ical properties of this preparation and, importantly, no change
in the heart rate dependence or distribution of cellular
alternans. Therefore, changes in magnitude or phase of
cellular alternans that followed the application of the SB
could not be attributed to degradation of the preparation over
time.
We hypothesize that electrically insulating 2 regions of
cells with heterogeneous intrinsic repolarization properties
such as those known to exist across the epicardium20 will
influence the development of discordant alternans. Figure 3
demonstrates that electrically uncoupling 2 regions of func-
tionally distinct myocardium by introducing a SB creates an
anatomic substrate across which discordant alternans can
form. Shown are RT alternans phase plots delineating regions
of cells where local repolarization alternates in a short-long
phase (Figure 3, black region) or a long-short phase (Figure 3,
Figure 3. Effect of a SB on spatial orientation of discordant
alternans shown as alternans phase plots. Alternans phase plots
from 5 experiments show regions alternating in short/long fash-
ion (black region) or long/short fashion (gray region). Areas with-
out significant repolarization alternans are represented in white.
Below each plot are action potentials from representative sites
(labeled either A or B) within each region, with APDs shown in
ms. Notice that in all experiments, the SB forms a portion of the
line of separation between the short/long and long/short
regions.
1160 Circulation Research December 8/22, 2000
Figure 4. Changes in ventricular repolarization caused by the
SB. Ten-millisecond isochrone maps represent patterns of
depolarization (DEPOL), APD, and repolarization (REPOL) for 2
consecutive beats in absence (ie, control) and presence of SB
at an identical heart rate. Activation propagates as expected
from site of stimulation (square pulse symbol) in all cases. SB
produces large changes in magnitude and orientation of APD
gradients (right middle panel) that were not present during con-
trol (left middle panel). This creates large spatial gradients of
repolarization in presence of SB. These effects are evident on
the ECG (top) as visible T-wave alternans is apparent with SB.
gray region). Areas without significant RT alternans are
represented in white. All plots correspond to the slowest heart
rate at which discordant alternans was recorded. Note that in
all experiments, the SB forms at least a portion (and in
experiments 1, 3, and 5, a large portion) of the line separating
regions of cells alternating with opposite phase. Below each
phase plot are representative action potentials from both the
Figure 5. Alternans-induced ventricular arrhythmias. Top, ECG
lead showing initiation of VF with a 10-ms decrement in CL in a
heart without a SB. Bottom, ECG tracing from same heart in
which SMVT was initiated after a SB was introduced. Notice in
both cases that visible T-wave alternans precedes the arrhyth-
mia. T-wave alternans were associated with discordant action
potential alternans between cells (not shown). Summary data
shown in bar graphs demonstrate that despite the similar role of
discordant alternans in the mechanism of unidirectional block
and reentrant VF and SMVT, the SB was typically required for
SMVT.
Figure 6. Effect of discordant alternans on the maximum gradi-
ent of repolarization. Shown is the maximum repolarization gra-
dient measured within the mapping array during baseline pacing
and during discordant alternans in control hearts preceding the
initiation of VF and in hearts with a SB preceding SMVT. Notice
that the maximum gradient is significantly greater preceding a
reentrant arrhythmia compared with baseline. However, the gra-
dients preceding VF are similar to those preceding SMVT, sug-
gesting that the conditions necessary for functional block are
similar in the presence and absence of the SB.
long-short (Figure 3, site A) and short-long (Figure 3, site B)
regions demonstrating discordant alternans. Importantly, in
the absence of a SB, stimulation at the identical heart rate
produced concordant RT alternans in the same hearts. There-
fore, the SB lowered the heart rate threshold for discordant
alternans.
Because of its effect on the phase of RT alternans, the SB
substantially affected spatial gradients of repolarization.
Shown in Figure 4 are isochrone maps representing patterns
of activation, APD, and repolarization for 2 consecutive beats
in the absence and presence of a SB in the same heart and at
an identical heart rate. The pattern of activation is similar
from beat to beat both before and after the introduction of the
SB. During control, there is a gradient of APD oriented in an
apex (shortest APD)–to– base (longest APD) direction on
both beats (Figure 4, left middle panel) as reported previous-
ly.20 In contrast, the introduction of a SB produces large
changes in the orientation and magnitude of APD gradients
from beat to beat (Figure 4, right middle panel). The effect on
APD accounts for the large differences in the orientation of
repolarization gradients on consecutive beats (Figure 4, right
bottom panel) and the larger spatial dispersion of repolariza-
tion as evidenced by the crowding of RT isochrones (Figure
4, compare repolarization maps on right with those on left).
Role of SB on Alternans-Induced Arrhythmias
To determine the effect of the SB on alternans-induced
reentry, we attempted to induce discordant alternans and
arrhythmias in all 26 hearts. In 15 of these hearts, the
arrhythmia-induction protocol was performed in the absence
and presence of a SB, whereas in 11 hearts the protocol was
performed only in the absence of a SB. In all 26 experiments,
reentrant VF or SMVT was only initiated after discordant
alternans formed; ie, reentrant excitation never occurred in
the absence of discordant alternans. In the absence of a SB,
 Pastore and Rosenbaum
Figure 7. Alternans-induced reentry in presence of SB. Top,
ECG lead showing initiation of reentry by introduction of a pre-
mature stimulus (S2) in a heart with a SB (white box). Middle,
10-ms isochrone maps of depolarization for the beat preceding
the S2 and during the initiation of the arrhythmia. Notice that
discordant alternans (action potential tracings, bottom panel)
between mapping sites A and B creates a large gradient of
repolarization that produces functional block (hatched area,
middle isochrone plot) necessary for reentry around the SB.
VF was initiated in 93% of hearts, whereas SMVT was
initiated in only 7%. In contrast, the presence of a SB resulted
in reentrant SMVT in 70% of hearts and VF in 30%.
Therefore, the SB was a significant and specific factor for the
development of SMVT (␹2⫽11.1, P⬍0.001). Figure 5 illus-
trates representative examples of the induction of VF and
SMVT in the same heart before and after the introduction of
a SB. Therefore, discordant alternans occurring in the pres-
ence and absence of a SB formed the electrophysiological
substrate for SMVT or VF, respectively.
To ascertain the role of discordant alternans in forming an
arrhythmogenic substrate, the maximum repolarization gra-
dient within the mapping array was measured at baseline
(CL⫽400) and during discordant alternans at the CL imme-
diately preceding the initiation of either reentrant VF or
SMVT. We found that discordant alternans created an elec-
trophysiological substrate for unidirectional block by ampli-
fying gradients of repolarization independent of underlying
structure. Figure 6 shows the maximum gradient of repolar-
ization measured within the mapping array at a baseline heart
rate of 150 bpm as compared with during discordant alternans
immediately preceding the initiation of VF (in hearts without
a SB) and SMVT (in hearts with a SB). Notice that the
maximum repolarization gradient is significantly larger pre-
ceding SMVT or VF compared with baseline, whereas the
maximum repolarization gradients preceding SMVT and VF
are similar. These data suggest that discordant alternans
transforms minor gradients of repolarization into critical
gradients sufficient for development of functional block that
leads to both SMVT and VF. This is illustrated in the
representative example shown in Figure 7 demonstrating the
initiation of reentrant SMVT after the application of a single
premature stimulus (S2) during discordant alternans. Reentry
is immediately preceded by discordant alternans as shown by
Alternans-Induced Reentry 1161
the action potential tracings from 2 selected sites, sites A and
B, (Figure 7, bottom panel). Notice the large repolarization
gradient between these 2 sites, which are only 5 mm apart.
Although the wavefront propagates successfully along both
sides of the SB on the S1 beat, the gradient caused by
discordant alternans is of sufficient magnitude to cause
unidirectional block of the premature (S2) beat. Notice that
the wavefront fails to propagate against the repolarization
gradient created by discordant alternans on the basal side of
the SB but does propagate along the apical side. About 130
ms later, the wavefront reenters the area of functional block in
a clockwise direction and propagates around the SB forming
the first beat of reentry. Similarly, discordant alternans
produced critical gradients of repolarization, which formed
the substrate unidirectional block, which led to reentrant VF
when no SB was available to stabilize the reentrant circuit.
Therefore, in the presence of a SB, discordant alternans
underlie the mechanism of SMVT because the SB served as
an anchor around which a stable reentrant circuit could form.
Discussion
For more than three-quarters of a century, T-wave alternans
has been closely associated with susceptibility to ventricular
arrhythmias in remarkably broad patient populations both
with8 and without7 structural heart disease. Furthermore,
T-wave alternans is now emerging as a clinical tool to screen
for patients at risk for sudden cardiac death based on the
principle that the heart rate threshold for T-wave alternans is
significantly reduced in high-risk patients.8,23 Recently, we
demonstrated that T-wave alternans is linked to the mecha-
nism of reentry by discordant alternans that develops between
cells possessing different electrophysiological properties.10
Importantly, discordant alternans greatly amplifies dispersion
of repolarization, producing conditions necessary for func-
tional block and reentry. We hypothesized that electrotonic
uncoupling by a SB increases the propensity for discordant
alternans, which may explain a mechanism of T-wave alter-
nans and arrhythmias in patients with structural heart disease.
In this report, we show conclusively that discordant alternans
is a novel mechanism by which the electrophysiological
substrate for block and reentry can form and that SBs greatly
increase the propensity for discordant alternans to develop.
Although results obtained from the guinea pig pacing-
induced model of T-wave alternans must be extrapolated
cautiously to patients, the model offers several distinct
advantages. As is the case in patients, electrical alternans in
this model persists over time (ie, alternans are not transient),
does not require ischemic damage, and principally involves
the T-wave rather than the QRS complex.8 Although sponta-
neous initiation of VT or VF in patients is often not preceded
by visible T-wave alternans as seen in these experimental
studies, it is possible that microvolt-level T-wave alternans is
present but unrecognized during clinical episodes of sudden
cardiac death.8 Finally, cellular alternans elicited by steady-
state pacing produces an electrophysiological substrate for
reentrant VT and VF,10 which typically lasts indefinitely
unless actively terminated by pacing or defibrillation,
respectively.18
1162 Circulation Research December 8/22, 2000
The effect of the SB on the development of T-wave
alternans and arrhythmogenesis was ascertained using optical
action potential mapping, which served to monitor the devel-
opment of cellular alternans both in the absence (ie, control)
and presence of a SB. This barrier was produced with a laser
to create a well-defined anatomic obstacle that electrically
insulated 2 regions of myocardium without altering the
electrophysiological properties of cells surrounding the bar-
rier.18 Therefore, our observation of profound changes in
cellular repolarization and alternans caused by SBs empha-
sizes the importance of investigating cell physiology in the
intact heart.
Previously, we have shown that the orientation of discor-
dant alternans closely follows heterogeneities of restitution
properties between cells10 that are known to exist across
guinea pig epicardium,20 suggesting that the reason neighbor-
ing regions of cells alternate with opposite phase is that the
cells possess different ionic properties. One would predict
that in the presence of reduced intercellular coupling, elect-
rotonic forces that normally act to maintain synchronization
of repolarization between cells are reduced, thereby enhanc-
ing dispersion of repolarization.24 Conversely, several studies
have shown that increased coupling attenuates repolarization
gradients.12,13 The present investigation demonstrates an in-
dependent and important effect of cell-to-cell uncoupling on
repolarization. In the presence of minor heterogeneities of
repolarization properties between cells, electrotonic uncou-
pling promotes the development of discordant alternans,
which in turn causes marked gradients of repolarization.
Importantly, arrhythmias only occurred in the presence of
steep repolarization gradients caused by discordant alternans.
Discordant alternans has also been reported during isch-
emia,2,25 in which cell-to-cell coupling is impaired,26,27 fur-
ther supporting a role of uncoupling in the mechanism of
discordant alternans.
There are several lines of evidence indicating that a SB
facilitates the development of discordant alternans. First, in
structurally normal myocardium, concordant alternans always
preceded discordant alternans, whereas in the presence of a
SB, a direct transition from no alternans to discordant
alternans occurred. Second, myocardial cells alternating with
opposite phase were typically located in proximity to the SB
(Figure 3). Third, the heart rate threshold required to elicit
discordant alternans was reduced significantly by the SB
(Figure 2B), which was also reflected in the ECG by a
decrease in the T-wave alternans heart rate threshold. This
corresponds well with recent clinical observations that sug-
gest that patients in whom SMVT can be initiated during
electrophysiological study have a significantly lower heart
rate threshold for T-wave alternans compared with patients
with negative electrophysiological tests.8,23 The findings in
this study provide a possible mechanism explaining these
clinical observations.
Our data suggest that discordant alternans can potentially
form in the following 2 situations: (1) where large heteroge-
neities in cellular repolarization properties exist in structur-
ally normal myocardium and (2) where minor heterogeneities
of repolarization exist in the presence of uncoupling caused
by SBs. For example, significant repolarization heterogene-
ities may exist in structurally normal myocardium in the
congenital long-QT syndrome in which discordant alternans
has been documented28 and may explain the association of
long-QT syndrome with torsade de pointes29,30 and VF.31 In
contrast, hearts of patients with structural heart disease have
reduced cell-to-cell coupling32 in part because of interdigita-
tion of fibrous tissue between myocyte bundles.15 Even
though these patients may have only minor heterogeneities of
repolarization between cells, electrical uncoupling may pro-
mote the formation of discordant alternans. Moreover, the
precise relation between the location of a SB and repolariza-
tion gradients may also explain why a scar can predispose to
reentry in some cases but not others. Our data do not suggest
that SBs of any type or having any orientation constitute a
requisite condition for monomorphic and polymorphic reen-
try. For instance, if a barrier is made along rather than across
the gradients of repolarization, reentry is not easily induced in
this model.18 Biological variability in location of repolariza-
tion gradients relative to the SB and pacing site may explain
why discordant patterns varied between experiments (Figure
3) and why, in turn, the site of block also varied (Figure 7).
The initiation of SMVT, rather than VF, in the presence of
a SB was explained by the barrier serving as an anchor to
stabilize a reentrant circuit (Figures 5 and 7). The concept is
supported by studies of both atrial and ventricular reentry.
During atrial flutter, the reentrant circuit can stabilize on
structural discontinuities in atria such as the crista terminalis
and eustachian ridge33 or on the tricuspid annulus.34 Simi-
larly, in ventricle, a meandering core of a spiral wave anchors
to structures such as small arteries, which stabilize the
wave.35 Following this reasoning, it is not surprising that
discordant cellular alternans provided a common mechanism
for VF and SMVT in our experiments. Such findings suggest
a singular mechanism by which T-wave alternans forms a
substrate for initiation of VF and polymorphic and monomor-
phic VT.
The structural alterations produced by heart disease are
obviously much more complex than those of our experimen-
tal model. Although a requirement for unidirectional block in
these studies, discordant alternans is clearly only one com-
ponent of a more complex substrate for reentry that is
undoubtedly present in patients and other models of reentry.
For example, in these experiments a single SB served to
promote both discordant alternans and an anchor for mono-
morphic reentry, whereas in patients it is more likely that
multiple SBs produce these effects independently. In addi-
tion, our model does not account for potential disease-
induced changes in the expression of ionic currents.36,37
However, the physical principles set forth by our experiments
should apply to more complex disease states.
Acknowledgments
This study was supported by NIH Grant HL-54807 and the
American Heart Association. We are extremely grateful to Imad
Libbus for his assistance with data analysis.
References
1. Orchard C, McCall E, Kirby M, Boyett M. Mechanical alternans during
acidosis in ferret heart muscle. Circ Res. 1991;68:69 –76.
 Pastore and Rosenbaum
2. Kurz RW, Mohabir R, Ren X-L, Franz MR. Ischaemia induced alternans
of action potential duration in the intact heart: dependence on coronary
flow, preload, and cycle length. Eur Heart J. 1993;14:1410 –1420.
3. Dilly SG, Lab MJ. Electrophysiological alternans and restitution during
acute regional ischemia in myocardium of anesthetized pig. J Physiol
(Lond ). 1988;402:315–333.
4. Kurz RW, Ren XL, Franz MR. Dispersion and delay of electrical resti-
tution in the globally ischaemic heart. Eur Heart J. 1994;15:547–554.
5. Salerno JA, Previtali M, Panciroli C, Klerbsy C, Chimienti M, Regazzi
Bonora M, Marangoni E, Falcone C, Guasti L, Campana C, Rondanelli R.
Ventricular arrhythmias during acute myocardial ischemia in man: the
role and significance of R-ST-T alternans and the prevention of ischemic
sudden death by medical treatment. Eur Heart J. 1986;7:63–75.
6. Shimoni Z, Flateau E, Schiller D, Barzilay E, Kohn D. Electrical alternans
of giant U waves with multiple electrolyte abnormalities. Am J Cardiol.
1984;54:920 –921.
7. Schwartz PJ, Malliani A. Electrical alternation of the T-wave: clinical and
experimental evidence of its relationship with the sympathetic nervous
system and with the long QT syndrome. Am Heart J. 1975;89:45–50.
8. Rosenbaum DS, Jackson LE, Smith JM, Garan H, Ruskin JN, Cohen RJ.
Electrical alternans and vulnerability to ventricular arrhythmias. N Engl
J Med. 1994;330:235–241.
9. Raeder E, Rosenbaum DS, Bhasin R, Cohen R. Alternating morphology
of the QRST complex preceding sudden death. N Engl J Med. 1992;326:
272–273.
10. Pastore JM, Girouard SD, Laurita KR, Akar FG, Rosenbaum DS.
Mechanism linking T-wave alternans to the genesis of cardiac fibrillation.
Circulation. 1999;99:1385–1394.
11. Shimizu W, Antzelevitch C. Cellular and ionic basis for T-wave alternans
under long-QT conditions. Circulation. 1999;99:1499 –1507.
12. Laurita KR, Girouard SD, Rudy Y, Rosenbaum DS. Role of passive
electrical properties during action potential restitution in the intact heart.
Am J Physiol. 1997;273:H1205–H1214.
13. Tan RC, Joyner RW. Electrotonic influences on action potentials from
isolated ventricular cells. Circ Res. 1990;67:1071–1081.
14. Hohnloser SH, Klingenheben T, Zabel M, Li Y-G, Albrecht P, Cohen RJ.
T wave alternans during exercise and atrial pacing in humans. J Car-
diovasc Electrophysiol. 1997;8:987–993.
15. Pogwizd SM, Hoyt RH, Saffitz JE, Corr PB, Cox JL, Cain ME. Reentrant
and focal mechanisms underlying ventricular tachycardia in the human
heart. Circulation. 1992;86:1872–1887.
16. Ursell P, Gardner P, Albala A, Fenoglio J, Wit A. Structural and elec-
trophysiological changes in the epicardial border zone of canine myo-
cardial infarcts during infarct healing. Circ Res. 1985;56:436 – 451.
17. Peters NS, Coromilas J, Severs NJ, Wit AL. Disturbed connexin43 gap
junction distribution correlates with the location of reentrant circuits in
the epicardial border zone of healing canine infarcts that cause ventricular
tachycardia. Circulation. 1997;95:988 –996.
18. Girouard SD, Pastore JM, Laurita KR, Gregory KW, Rosenbaum DS.
Optical mapping in a new guinea pig model of ventricular tachycardia
reveals mechanisms for multiple wavelengths in a single reentrant circuit.
Circulation. 1996;93:603– 613.
19. Laurita KR, Girouard SD, Akar FG, Rosenbaum DS. Modulated dis-
persion explains changes in arrhythmia vulnerability during premature
stimulation of the heart. Circulation. 1998;98:2774 –2780.
20. Laurita KR, Girouard SD, Rosenbaum DS. Modulation of ventricular
repolarization by a premature stimulus: role of epicardial dispersion of
Alternans-Induced Reentry 1163
repolarization kinetics demonstrated by optical mapping of the intact
guinea pig heart. Circ Res. 1996;79:493–503.
21. Girouard SD, Laurita KR, Rosenbaum DS. Unique properties of cardiac
action potentials recorded with voltage-sensitive dyes. J Cardiovasc Elec-
trophysiol. 1996;7:1024 –1038.
22. Bayly PV, KenKnight BH, Rogers JM, Hillsley RE, Ideker RE, Smith
WM. Estimation of conduction velocity vector fields from epicardial
mapping data. IEEE Trans Biomed Eng. 1998;45:563–571.
23. Kavesh NG, Shorofsky SR, Sarang SE, Gold MR. Effect of heart rate on
T wave alternans. J Cardiovasc Electrophysiol. 1998;9:703–708.
24. Lesh MD, Pring M, Spear JF. Cellular uncoupling can unmask dispersion
of action potential duration in ventricular myocardium: a computer
modeling study. Circ Res. 1989;65:1426 –1440.
25. Konta T, Ikeda K, Yamaki M, Nakamura K, Honma K, Kubota I, Yasui
S. Significance of discordant ST alternans in ventricular fibrillation.
Circulation. 1990;82:2185–2189.
26. Dekker LRC, Rademaker H, Vermeulen JT, Opthof T, Coronel R, Spaan
JAE, Janse MJ. Cellular uncoupling during ischemia in hypertrophied and
failing rabbit ventricular myocardium: effects of preconditioning. Circu-
lation. 1998;97:1724 –1730.
27. Kleber AG, Riegger CB, Janse MJ. Electrical uncoupling and increases of
extracellular resistance after induction of ischemia in isolated, arterially
perfused rabbit papillary muscle. Circ Res. 1987;61:271–279.
28. Shimizu W, Yamada K, Arakaki Y, Kamiya T, Shimomura K.
Monophasic action potential recordings during T-wave alternans in con-
genital long QT syndrome. Am Heart J. 1996;132:699 –701.
29. Curran ME, Splawski I, Timothy KW, Vincent GM, Green ED, Keating
MT. A molecular basis for cardiac arrhythmia: HERG mutations cause
long QT syndrome. Cell. 1995;80:795– 803.
30. Roden DM, Lazzara R, Rosen M, Schwartz PJ, Towbin J, Vincent GM.
Multiple mechanisms in the long-QT syndrome: current knowledge, gaps,
and future directions. Circulation. 1996;94:1996 –2012.
31. Leenhardt A, Lucet V, Denjoy I, Grau F, Ngoc DD, Coumel P. Cat-
echolaminergic polymorphic ventricular tachycardia in children: a 7-year
follow-up of 21 patients. Circulation. 1995;91:1512–1519.
32. Peters NS, Green CR, Poole-Wilson PA, Severs NJ. Reduced content of
connexin43 gap junctions in ventricular myocardium from hypertrophied
and ischemic human hearts. Circulation. 1993;88:864 – 875.
33. Olgin JE, Kalman JM, Fitzpatrick AP, Lesh MD. Role of right atrial
endocardial structures as barriers to conduction during human type I atrial
flutter: activation and entrainment mapping guided by intracardiac echo-
cardiography. Circulation. 1995;92:1839 –1848.
34. Kalman JM, Olgin JE, Saxon LA, Fisher WG, Lee RJ, Lesh MD. Acti-
vation and entrainment mapping defines the tricuspid annulus as the
anterior barrier in typical atrial flutter. Circulation. 1996;94:398 – 406.
35. Pertsov AM, Davidenko JM, Salomonsz R, Baxter WT, Jalife J. Spiral
waves of excitation underlie reentrant activity in isolated cardiac muscle.
Circ Res. 1993;72:631– 650.
36. Aggarwal R, Boyden PA. Diminished Ca2⫹ and Ba2⫹ currents in
myocytes surviving in the epicardial border zone of the 5-day infarcted
canine heart. Circ Res. 1995;77:1180 –1191.
37. Pak PH, Nuss HB, Tunin RS, Kääb S, Tomaselli GF, Marbán E, Kass DA.
Repolarization abnormalities, arrhythmia and sudden death in canine
tachycardia-induced cardiomyopathy. J Am Coll Cardiol. 1997;30:
576 –584.