Advances in Arrhythmia and Electrophysiology
Cardiac Memory
Diagnostic Tool in the Making
Alexei Shvilkin, MD; Henry D. Huang, MD; Mark E. Josephson, MD
C lassic definition of the term cardiac memory (CM) refers
to the persistent T-wave changes on the ECG after a
period of wide QRS rhythms that become evident once nor-
mal ventricular activation pattern is restored. It is related to
the term ventricular electric remodeling sometimes used in
basic science literature. Although CM itself is considered as
an adaptive reaction to the change in the ventricular activa-
tion sequence, its manifestations (usually T-wave inversions,
TWIs) are often confused with pathological conditions mani-
festing with TWI, such as myocardial ischemia or infarc-
tion. Although originally CM was described after restoration
of the normal ventricular activation (narrow QRS), recently
it was described and studied in wide QRS rhythms signifi-
cantly expanding the clinical relevance of this phenomenon.
Although complex molecular mechanisms of CM have been
previously reviewed extensively, practical clinical applica-
tion of this phenomenon received much less attention. In this
article, we will focus on the clinically relevant aspects of CM
including its diagnostic use in narrow and wide QRS rhythms.
We will also examine how the fundamental CM property of
being initiated by the changes in the geometry of the ventricu-
Downloaded from http://ahajournals.org by on August 7, 2019
lar contraction might open new perspectives in wide QRST
morphology analysis by establishing parallels between elec-
tric and mechanical functions of the heart.
History, Properties, and Adaptive
Nature of Cardiac Memory
In 1915, White1 was the first to observe the phenomenon of
transient TWI after single ventricular premature beats. Later
in the 1940s, TWIs after conversion to sinus rhythm were
described after paroxysmal tachycardias.2 Since then abnor-
mal T waves of various duration have been documented after
intermittent ventricular pre-excitation,3–5 ventricular pacing,6,7
intermittent left bundle branch block (LBBB),8,9 ventricular
tachycardia,10 and even QRS widening associated with sodium
channel blocker toxicity.11
In 1982, Rosenbaum et al10 introduced the term heart mem-
ory and presented the first unified hypothesis of how abnor-
mal ventricular activation could lead to the development of
T-wave abnormalities regardless of the wide QRS cause by
a process he referred to as electrotonic modulation. In this
seminal article, 3 principles of CM were formulated: (1) the
Received January 14, 2015; accepted February 5, 2015.
From the Department of Medicine/Cardiology Division, Beth Israel Deaconess Medical Center, Boston, MA.
Correspondence to Alexei Shvilkin, MD, Baker4/Cardiology, Beth Israel Deaconess Medical Center, 185 Pilgrim Rd, Boston, MA 02215. E-mail
ashvilki@bidmc.harvard.edu
(Circ Arrhythm Electrophysiol. 2015;8:475-482. DOI: 10.1161/CIRCEP.115.002778.)
© 2015 American Heart Association, Inc.
Circ Arrhythm Electrophysiol is available at http://circep.ahajournals.org DOI: 10.1161/CIRCEP.115.002778
475
direction of the T waves in sinus rhythm follows (remembers)
the direction of the QRS complex during preceding episode
of abnormal activation; (2) the amplitude of memory T waves
increases the longer abnormal conduction continues, and (3)
repeat episodes of abnormal activation after complete normal-
ization of T waves result in more rapid and prominent accu-
mulation of T-wave changes (hence the term memory). The
authors hypothesized that CM represents adaptation of myo-
cardial repolarization to the new activation sequence.
The adaptive nature of CM was subsequently confirmed
in both animal and human studies. Costard-Jäckle et al12 and
later Sosunov et al13 in an isolated rabbit heart demonstrated
the presence of an inverse relationship between local activa-
tion and repolarization times resulting in homogeneity of the
activation-recovery intervals throughout the ventricle during
normal activation. Initiation of ventricular pacing disrupted
this relationship, but within 60 to 120 minutes of pacing, this
relationship was re-established suggesting adaptation of the
repolarization to the new activation pattern. After the reversal
to normal conduction, it took another 60 minutes for repolar-
ization to readjust to the now new old activation sequence.
In humans, 1 week of ventricular pacing resulted in sig-
nificant QT interval shortening during continuous pacing.
However, when normal conduction was restored and QRS
became narrow, QT interval became prolonged compared with
the baseline, indicating that the heart has adapted to the new
activation sequence.14,15 No evidence of hypertrophy, changes
in myocardial blood flow, and hemodynamics were observed
in the setting of CM after 21 days of ventricular pacing in dogs
further suggesting the adaptive nature of this process.16
Electrophysiology, Molecular
Mechanisms, and Triggers of CM
Repolarization Gradients
The ST segment and T waves on the 12-lead ECG are con-
current with the repolarization phase of the cardiac cycle.
The concordance of QRS and T wave in a normal human
ECG for many years was explained by the presence of a
transmural repolarization gradient resulting from a shorter
epicardial action potential duration (APD) so that epicar-
dial cells being depolarized last repolarize first. Not surpris-
ingly such 1-dimensional explanation applied to a complex
 476  Circ Arrhythm Electrophysiol  April 2015

3-dimensional structure as the heart seemed oversimplified Molecular Mechanisms

and was called into question.17 Multiple regional repolariza-
tion gradients (apical-basal, right-to-left ventricle) have been
discovered and their relative contributions along with the
transmural gradient to the normal T-wave morphology remain
the subject of an ongoing debate.17,18
Similar to the theories of normal T-wave morphology ini-
tially inverted T waves of CM were explained by the change
in the transmural repolarization gradient because of prefer-
ential epicardial APD prolongation.16,19 Later, it was discov-
ered that not only the transmural repolarization gradient in
CM is site-specific,20 but also other gradients exist including
apical-basal and right-to-left ventricle.21,22 There is a gen-
eral agreement across the studies that CM is associated with
prolongation of repolarization in the early activated areas.
However, repolarization changes at the late activated sites
were variable ranging from shortening,23 minimal, or no
change20,22 to significant APD prolongation.24,25 Discrepancy
in these findings is a subject of an ongoing debate and is
likely related to the differences in experimental models,
pacing site, species, and methods of measurement. In addi-
tion, novel factors influencing local APD, such as Purkinje
fiber––myocyte junctions have been described and merit
further investigation.26 From the clinical perspective, the
human data on QT interval shortening during continuous
ventricular pacing14,15 are most consistent with reciprocal
changes in APD between early (prolongation) and late-acti-
vated (shortening) areas.
Downloaded from http://ahajournals.org by on August 7, 2019
Molecular mechanisms of CM have been extensively reviewed
recently.27,28
In brief, CM comprises a spectrum of 2 distinct but over-
lapping phenomena which differ in the mechanisms by which
repolarization changes are achieved: short-term and long-term
CM. Short-term CM observed within minutes of ventricular
pacing is thought to occur from modulation and modification
of existing proteins and channel trafficking. It is relatively
short-lived and dissipates within minutes. Long-term CM is
seen after longer periods of abnormal activation and is longer
lasting (days to weeks). It includes changes in gene transcrip-
tion and protein synthesis.
Molecular changes observed in the setting of CM include
alterations of multiple ion channels, receptors, and cell
coupling, including the transient outward current, Ito, IKr,
ICa,27,28 Na/Ca exchanger,25 AT1 receptors, stretch-activated
receptors,29 and gap junction redistribution.23,30
Triggers of CM
Perhaps the most important discovery in the physiology of CM
was establishing its relationship to the mechanical function of
the heart. The notion that it is the change in ventricular con-
traction pattern and local ventricular wall stress (as oppossseed
to electrotonic modulation originally proposed by M. Rosen-
baum) that triggers CM was first made on the basis of CM inhi-
bition by the renin-angiotensin system blockade.31 It was later
confirmed in elegant experiments with excitation–contraction

Figure 1. A, Development and partial resolution of cardiac memory (CM) in a chronic canine model of epicardial left ventricular pacing.
Leads I, aVF, and frontal plane vectorcardiogram shown at baseline, during ventricular pacing, 1 hour after interruption of ventricular pac-
ing for 7, 21, and 3 days after cessation of ventricular pacing. Leads I and aVF are shown (top). At baseline (left column), the T waves dur-
ing sinus rhythm are positive in both leads. Left ventricular pacing (column 2) results in wide QRS with secondary T-wave abnormalities
and discordant T waves. Pacing interruption after 7 and 21 days results in the development of progressive (accumulation) T-wave inver-
sions during sinus rhythm that persist during day 3 of recovery. The polarity of the T waves corresponds to the polarity of the paced QRS
complex. Bottom, The corresponding frontal plane vectorcardiograms are displayed demonstrating gradual alignment of the T-vector
loop during sinus rhythm with the QRS loop during ventricular pacing. Note the increase in the T-vector amplitude and rotation toward the
direction of the paced QRS vector (arrows). Cross-hair represents calibration at 1 mV. B, Quantitative assessment of CM. The baseline
and memory T-vector loops superimposed. CM can be measured by the change in the T-vector direction, amplitude, and 3-dimensional
distance (T peak displacement [TPD] measured in mV) between the T-vector peaks (T control represents the baseline T-vector loop and T
memory represents the T-vector loop after CM induction). TPD is the preferred method of CM measurement. Adapted from Yu et al19 with
permission of the publisher. Copyright © 1999, Wolters Kluwer Health. Adapted from Plotnikov et al32 with permission of the publisher.
Copyright © 2001, European Society of Cardiology. Authorization for this adaptation has been obtained both from the owner of the copy-
right in the original work and from the owner of copyright in the translation or adaptation.

uncoupling demonstrating that pacing-induced CM only devel-
ops if the change in the electric activation sequence is accom-
panied by the active mechanical contraction, whereas CM-like
changes can develop as the result of local mechanical myocar-
dial stretching alone without electric activation change.13
We think this fundamental property of CM provides an
opportunity to consider CM as a potential indicator of myo-
cardial contractility disturbance in clinical practice, as will be
shown later.
Role of Vectorcardiography in Evaluation of CM
As originally noted by Rosenbaum et al,10 not only the T
waves follow the direction of the wide QRS complex in each
ECG lead, but also the whole T-wave 3-dimensional vec-
tor aligns with the wide QRS vector when assessed using
vectorcardiography.14
This 3-dimensional nature of repolarization changes in CM
made vectorcardiography a preferred tool of its assessment
so that instead of tracking disparate T-wave polarities and
amplitudes in individual leads the change in the whole T-wave
vector can be observed and measured. The vectorcardiogra-
phy-based measurement of CM, T-vector peak displacement
was developed.32 It allowed to express CM magnitude as a sin-
gle numeric value and track its changes under different experi-
mental conditions (Figure 1). T-vector peak displacement has
been used in several experimental and clinical studies further
advancing CM principles on a new quantitative level.14,20,32
Clinical Applications of Cardiac
Downloaded from http://ahajournals.org by on August 7, 2019
Memory in Narrow QRS Rhythms
Why Was CM Initially Described Only During
Normal Ventricular Activation
Traditionally, CM has been associated with TWI during sinus
rhythm with narrow QRS for several reasons:
First, normal T waves during sinus rhythm are positive in
most leads, whereas TWI is usually associated with patho-
logical conditions (eg, ischemia, strain, cerebral hemorrhage).
Therefore, T-wave changes are more likely to be noticed when
TWI is present. Second, in the most common situations lead-
ing to CM development (LBBB, right ventricular apical pac-
ing), wide QRS deflections initiating CM are negative in the
majority of the leads and therefore produce TWI.
Third, large secondary repolarization changes during wide
QRS rhythms mask CM until the QRS becomes narrow
(Figure 2).
However, as both depolarization and repolarization of the
heart occur in a 3-dimensional space, CM always produces a
mixture of positive and negative T-wave changes based on the
wide QRS vector polarity. Although positive T-wave changes
are often overlooked, they are as important as TWI in diagnos-
tic applications of CM as will be seen below.
Detecting the Presence and Localizing Intermittent
Wide QRS Rhythms
CM signature consisting of the specific combinations of
negative and positive T waves can give a clue to the pres-
ence and localization of intermittent wide QRS rhythms
Shvilkin et al   Cardiac Memory   477
Figure 2. ECG during narrow (AAI pacing, A) and wide QRS
(DDD pacing, B) before (baseline) and after cardiac memory (CM)
induction (day 7). A, At baseline, the T waves during AAI pacing
are positive in most leads. On day 7 during AAI pacing, inverted
T waves in multiple leads assume the polarity of the paced QRS
complex consistent with CM. B, During wide QRS rhythm, CM
manifests as a decrease in the T-wave amplitude with no change
in polarity on day 7 compared with the baseline. Unless quantita-
tive vectorcardiography methods are used to measure the T-wave
loop, this change usually goes unnoticed on a 12-lead ECG. Sub-
tle changes in the QRS axis and amplitude between the baseline
and day 7 are most likely positional. Adapted from Shvilkin et
al14 with permission of the publisher. Copyright © 2009, Elsevier.
Authorization for this adaptation has been obtained both from the
owner of the copyright in the original work and from the owner of
copyright in the translation or adaptation.
including pre-excitation, ventricular tachycardia, and conduc-
tion abnormalities.
Figure 3 demonstrates CM signature after ablation of a left
posteroseptal bypass tract. Figure 3A shows baseline ECG
with manifest pre-excitation. ECG in Figure 3B obtained
30 minutes after the bypass tract ablation demonstrates tall
peaked T waves across the precordium as well as inferior TWI
corresponding to the vector of the delta wave. Tall positive
memory T waves in the right precordial leads can be easily
confused with hyperkalemia or ischemia. ECG in Figure 3C
obtained 1 week after ablation shows a significant decrease
in the amplitude of both positive and negative T-wave deflec-
tions. ECG completely normalized in 6 months (not shown).
A year later the patient was again noted to have T-wave abnor-
malities on his routine ECG. Follow-up Holter monitoring
showed intermittent pre-excitation without any arrhythmias.
In this case, CM was the only sign of the partial recurrence of
antegrade conduction over the bypass tract.
Another patient with structurally normal heart and frequent
episodes of palpitations presented with T-wave abnormalities
during sinus rhythm on his ECG (Figure 4, left). Based on the
presence of the inferior TWI and tall peaked T waves in the
right precordial leads in the absence of ischemia, one can sus-
pect the presence of a wide QRS rhythm originating in the infe-
rior septal aspect of the left ventricle. Indeed, the patient had
verapamil-sensitive ventricular tachycardia (Figure 4, right).
 478  Circ Arrhythm Electrophysiol  April 2015

Figure 3. Cardiac memory after ablation of a left

posteroseptal accessory bypass tract. A, Before
ablation. Left posteroseptal bypass tract is pres-
ent. B, Thirty minutes after ablation. Note tall posi-
tive peaked T waves in the right precordial leads
as well as inferior T-wave inversions. C, One week
after ablation. Both positive and negative T waves
decreased in amplitude but have not normalized
completely.

CM signature of intermittent LBBB will be discussed in the
following section along with right ventricular pacing.
CM-related T-wave changes after intermittent right bundle
branch block have not been described. The likely reason for that is
that the delayed activation of the right ventricle produces mostly
positive deflections in the right precordial leads resulting in an
increase in the positive T-wave amplitude that goes unnoticed.
Distinguishing Between TWI From CM and
Ischemia
Since the time CM was first described, its similarity to isch-
emic TWI made it a significant confounder in the diagnosis of
myocardial ischemia often resulting in excessive cardiac test-
ing. In particular, precordial TWI because of intermittent right
Downloaded from http://ahajournals.org by on August 7, 2019
attributable to the proximal left anterior descending coronary
artery lesion involves the left ventricle, the ischemic precordial
TWI has a rightward axis in the frontal plane and is character-
ized by TWI in leads I and aVL (Figure 5, left). The rare cases
of right coronary artery ischemia producing precordial TWI
with the leftward axis (Figure 5, middle) have deeper TWI in
inferior leads than that in the precordial leads. However, right
ventricular apical pacing and LBBB produce QRS vectors pos-
itive in leads I and aVL resulting in positive T waves in these
leads on resumption of normal conduction consistent with CM
principles. It also produces precordial TWI deeper than the
inferior TWI (Figure 5, right).
The combination of

ventricular pacing or LBBB (by far the most common presen-

tations of CM encountered in clinical practice) can be easily
confused with the Wellens’ syndrome characteristic of tran-
sient proximal left anterior descending coronary artery occlu-
sion, a condition requiring prompt coronary intervention.33
In general, the direction of the ischemic T-waves points
away from the area of ischemia. Because the ischemic region
1. positive T in lead aVL and positive/isoelectric T in lead
I; and
2. precordial TWI>inferior TWI produces a unique pacing-
induced CM signature that was 92% sensitive and 100%
specific in differentiating pacing-induced TWI from
ischemia in a retrospective study.34

Figure 4. Cardiac memory (CM) in a setting of recurrent monomorphic ventricular tachycardia. Left, Inferolateral T-wave inversions (TWI
lead III>TWI lead II) and unusually tall T waves in leads V2–V3 during sinus rhythm suggest CM due to intermittent wide QRS originating
in the inferoseptal area of the left ventricle. Right, Spontaneous monomorphic ventricular tachycardia observed later in the same patient.
Bottom, Rhythm strip (lead II) during tachycardia demonstrating capture beat.
 Shvilkin et al   Cardiac Memory   479

Figure 5. T-wave inversions (TWI) because of isch-

emia and cardiac memory (CM). Left, Left anterior
descending coronary artery (LAD) ischemia. Note:
TWI in leads I and aVL. Middle, Dominant right
coronary artery (RCA) ischemia. Although the TWI
pattern is similar to CM because of the right ven-
tricular pacing (positive T waves in leads I and aVL),
the ratio of inferior/precordial TWI is different (maxi-
mal inferior TWI>maximal precordial TWI). Right,
TWI because of CM. T waves in leads I and aVL are
positive; precordial TWI>inferior TWI.

Although not studied in larger prospective trials, these easily missed) in contrast to the narrow QRS rhythm, where T vec-
applicable criteria have been used successfully in clinical prac- tor becomes larger and rotates toward the paced QRS. Using
tice to distinguish between CM and ischemia.35 Because the axis quantitative vectorcardiograpic analysis, CM can be detected in
during right ventricular pacing and LBBB is similar with posi- paced rhythm within minutes after the onset of pacing.
tive QRS complex in leads I and aVL, the criteria can also be
extended to the cases of TWI because of intermittent LBBB.36
One of the study limitations is its relatively small size.
Although not observed in the original study, some distal left
anterior descending coronary artery ischemia might not pro-
duce enough T-wave right axis deviation to cause negative T
waves in leads I and aVL.
The other limitation in using this diagnostic tool is a con-
cern that T-wave changes because of CM can counterbalance
or obscure ischemic TWI when CM and ischemia occur simul-
taneously in the same patient. No systematic data are available
in patients with T-wave abnormalities because of coronary
artery disease but our preliminary results in this population
indicate that CM does not reverse ischemic TWI minimizing
Downloaded from http://ahajournals.org by on August 7, 2019

this risk. Notably, when TWI in leads I and aVL are present at
baseline as the result of the structural heart disease, T waves
usually stay inverted, despite the development of CM (T vec-
tor never fully aligns with the paced QRS vector).
The proposed diagnostic criteria in the future can be further
improved by the use of vectorcardiography techniques.

Cardiac Memory in Wide QRS Rhythms

Despite the fact that processes underlying CM do not develop
suddenly on QRS normalization but rather progress gradually
during abnormal ventricular activation, for a long time it was
thought that CM cannot be detected until QRS becomes nar-
row. Secondary T-wave changes in wide QRS rhythms occur
immediately and are dominant, whereas CM-induced T-wave
changes are initially subtle and while gradually increasing over
time are obscured by the secondary changes. The assessment
of CM by surface ECG in the setting of wide QRS is limited
because of the dominance of the secondary T-wave changes.
Repolarization changes because of CM in this situation have
been largely overlooked, being masked by the large discordant Figure 6. Vectorcardiogram of the patient in Figure 2 during AAI
and DDD pacing in frontal (A) and transverse (B) projections
T waves. Nevertheless, vectorcardiogram can readily detect
before (baseline) and after the induction of cardiac memory
CM in this situation. Figure 6 presents the vectorcardiogram of (CM; day 7). On day 7 during AAI pacing, the T vector assumes
the patient shown in Figure 2 demonstrating the 3-dimensional the direction of the paced QRS complex while increasing in mag-
T-vector displacement after 7 days of right ventricular pacing nitude. At the same time in DDD mode, the T-vector magnitude
decreases with no change in direction. Black arrows indicate the
during the narrow (AAI pacing) and wide (DDD pacing) QRS. direction and magnitude of the projection of T-peak displace-
The actual T-vector change (T peak displacement) in the nar- ment as the result of CM and are similar in AAI and DDD modes.
row and wide QRS is nearly identical in both magnitude and Sagittal plane (not shown) showed similar changes. Adapted from
Shvilkin et al14 with permission of the publisher. Copyright © 2010,
direction.14 As seen on both Figures 2 and 6, CM in wide QRS
Elsevier. Authorization for this adaptation has been obtained both
rhythms presents as a decrease in T-vector magnitude with- from the owner of the copyright in the original work and from the
out directional change (another reason why they are usually owner of copyright in the translation or adaptation.
 480  Circ Arrhythm Electrophysiol  April 2015
Clinical Application of CM in Wide QRS: Age
Determination of LBBB
The finding that CM decreases discordant T-wave amplitude
in wide QRS rhythms has important clinical implications. For
Downloaded from http://ahajournals.org by on August 7, 2019
example, it can be used to determine whether LBBB is old or
new which becomes valuable in the setting of chest pain.
For many years, American College of Cardiology/American
Heart Association guidelines for the management of patients
with ST-segment–elevation myocardial infarction considered
new or presumed new LBBB, a class I indication for reperfu-
sion therapy when associated with symptoms suggestive of
ischemia creating a decision-making dilemma. Recently, this
recommendation has been withdrawn37 largely because of the
inability to determine the LBBB age which resulted in overly
aggressive management of the chronic LBBB patients.
In clinical practice determining whether LBBB is actually
new is virtually impossible as prior ECGs for comparison
even if available are usually dated well beyond the relevant
period of time (from the onset of symptoms which ranges
from minutes to a few hours). LBBB in the vast majority of
patients presenting with chest pain is old.
The criterion for LBBB age determination was developed38 on
the premise that new LBBB would have larger T vector (and taller
T waves on a 12-lead ECG) compared with the old one in accor-
dance with CM development in wide QRS rhythm (Figure 7). As
the duration of LBBB increases the discordant secondary T waves
become smaller (Figure 7A). A retrospective analysis >1700
LBBB ECGs showed that indeed the new onset LBBB (defined
as <24-hour duration) was rare (3%) and had significantly larger
T waves (as well as smaller QRS vector magnitude) compared
with the chronic LBBB (Figure 7B and 7C). Although the best
results in LBBB age determination were achieved using quantita-
tive vectorcardiography technique, a simplified 12-lead ECG cri-
terion using the ratio of the maximal precordial S wave/maximal
Figure 7. Use of the maximal precordial S/T ratio in
left bundle branch block (LBBB) age determination.
S/T represents the ratio of the maximal precordial S
wave/maximal precordial T-wave amplitude.
A, As the duration of LBBB increases, the maximal
precordial S/T ratio increases from 1.64 at 6 hours
(=new) to 3.22 at 15 days (=old). B, A typical exam-
ple of resolution of a new LBBB (<8-hour duration).
S/T ratio is 1.61. Resolution of LBBB results in no
apparent T-wave abnormalities during narrow QRS.
C, A typical example of resolution of an old LBBB
(>3-day duration). S/T ratio is 2.93. Typical changes
of cardiac memory with precordial TWI are evident
during narrow QRS. D, A new onset LBBB (<75 min-
utes) in a setting of an acute left anterior descending
coronary artery thrombosis. S/T ratio is 1.68 consis-
tent with the new LBBB. Note ST segment changes
in the baseline tracing consistent with ischemia and
loss of R waves after resolution of LBBB. Adapted
from Shvilkin et al38 with permission of the publisher.
Copyright © 2010, Elsevier. Authorization for this
adaptation has been obtained both from the owner
of the copyright in the original work and from the
owner of copyright in the translation or adaptation.
precordial T-wave amplitude (as approximations of the QRS and
T vectors, respectively) was also developed. A conservative cutoff
of S/T<2.5 allowed to detect 100% of the new LBBBs.
The T-vector magnitude in LBBB changes significantly dur-
ing the first 24 hours. In cases of painful LBBB syndrome when
ECG is recorded within seconds/minutes of symptoms onset
(often during an exercise stress test), S/T ratio can be as low as
1.4. The majority of the T-wave changes occur within the first
24 hours of LBBB persistence when it assumes chronic QRST
configuration.10 In the true chronic LBBB, the S/T ratio is close
to ≥3.0 (Figure 7A and 7C). It is important to recognize that
LBBB is often a dynamic phenomenon and can be intermittent
or rate-dependent. Repeated episodes of intermittent LBBB can
Figure 8. Correlation between cardiac memory (CM) magnitude
(measured during DDD pacing) and the left ventricular dys-
synchrony (difference in time to the maximal longitudinal strain
between septal and lateral left ventricular walls). With the increase
in dyssynchrony CM magnitude decreases. Adapted from
Shvilkin41 with permission of the publisher. Copyright © 2013,
Springer. Authorization for this adaptation has been obtained both
from the owner of the copyright in the original work and from the
owner of copyright in the translation or adaptation.

cause accumulation of CM-related T-wave changes, sometimes
making the distinction between new and old LBBB difficult.
The new LBBB S/T criterion held true in a small subgroup
of patients with acute coronary syndrome38 (Figure 7D) but
needs to be confirmed in larger clinical trials.
Using this criterion in low-risk patients can help justify
conservative management and avoid unnecessary coronary
interventions.
Cardiac Memory in the Setting of Biventricular
Pacing
Dynamic T-wave changes have been reported in the setting
of biventricular pacing.39,40 This situation is of a particular
interest as it demonstrates that the CM can develop on top of
already existing CM-induced repolarization changes and the
heart is capable of relearning. Whether the new T waves fol-
low the biventricular or left ventricular paced QRS direction
as well as clinical significance of these changes is the subject
of further research.
Through a Looking Glass: Cardiac Memory as a
Reflection of Ventricular Function
As discussed above, CM is initiated by the changes in local
myocardial contractility. Understandably, a question arises
whether CM magnitude can be used as a reflection of pacing-
induced contractile pattern disruption, and whether wide varia-
tion of pacing-induced CM magnitude between individuals
(approximately 8-fold)14 can be explained by the differences in
the left ventricular contraction pattern during ventricular pac-
Downloaded from http://ahajournals.org by on August 7, 2019
ing. If true, this can open an opportunity to assess the risk of
pacing-induced ventricular dysfunction based on ECG analy-
sis. Indeed, the preliminary data suggest that pacing-induced
CM magnitude is tightly correlated with the synchrony of left
ventricular contraction during right ventricular pacing. Patients
with more pronounced septal to lateral wall dyssynchrony
developed less CM (Figure 8). The dyssynchrony was associ-
ated with accelerated septal contraction (septal flash) resulting
in the smaller stroke volume.41
To further extend this concept given the positive correlation
between CM magnitude and QRS/T ratio at the onset of ventric-
ular pacing,14 wide complex QRST morphology (and in particu-
lar QRS/T vector magnitude ratio) potentially can be viewed as
the reflection of the ventricular contractile disturbance and used
for risk assessment of dyssynchronous heart failure.
Of further interest, recently vectorcardiography-derived
T-wave area in LBBB was found to be predictive of response
to cardiac resynchronization.42 Further studies are needed to
confirm relationships between wide QRST morphology and
contractile function of the heart.
Conclusions
Cardiac memory is a universal adaptive property of the heart
reflecting the adjustment of repolarization to the new activation
sequence and manifesting by T-wave changes. CM can be detected
during both narrow and wide QRS rhythms, and the knowledge
of its features improves ECG diagnosis. Further research in CM
and wide QRST morphology can provide insight into mechanical
function of the heart and potentially be used in risk assessment of
dyssynchronous heart failure and its correction.
Shvilkin et al   Cardiac Memory   481
Disclosures
None.
References
1. White P. Alteration of the pulse: a common clinical condition. Am J Med
Sci. 1915;150:82–96.
2. Campbell M. Inversion of T waves after long paroxysms of tachycardia.
Br Heart J. 1942;4:49–56.
3. Nicolai P, Medvedowsky JL, Delaage M, Barnay C, Blache E, Pisapia A.
Wolff-Parkinson-White syndrome: T wave abnormalities during normal
pathway conduction. J Electrocardiol. 1981;14:295–300.
4. Poole JE, Bardy GH. Further evidence supporting the concept of T-wave
memory: observation in patients having undergone high-energy direct cur-
rent catheter ablation of the Wolff-Parkinson-White syndrome. Eur Heart
J. 1992;13:801–807.
5. Nirei T, Kasanuki H, Ohnishi S, Tamaki A, Toyoshima Y, Hosoda S.
Cardiac memory in patients with intermittent Wolff-Parkinson-White syn-
drome. J Electrocardiol. 1997;30:323–329.
6. Chatterjee K, Harris A, Davies G, Leatham A. Electrocardiographic
changes subsequent to artificial ventricular depolarization. Br Heart J.
1969;31:770–779.
7. Gould L, Venkataraman K, Goswami MK, Gomprecht RF. Pacemaker-
induced electrocardiographic changes simulating myocardial infarction.
Chest. 1973;63:829–832.
8. Engel TR, Shah R, DePodesta LA, Frankl WS, Krause RL. T-wave ab-
normalities of intermittent left bundle-branch block. Ann Intern Med.
1978;89:204–206.
9. Gould L, Reddy CV, Singh B, Zen B. T-wave changes with intermittent
left bundle branch block. Angiology. 1980;31:66–68.
10. Rosenbaum MB, Blanco HH, Elizari MV, Lázzari JO, Davidenko JM.
Electrotonic modulation of the T wave and cardiac memory. Am J Cardiol.
1982;50:213–222.
11. Wylie JV Jr, Zimetbaum P, Josephson ME, ShvilkinA. Cardiac memory induced
by QRS widening due to propafenone toxicity. Pacing Clin Electrophysiol.
2007;30:1161–1164. doi: 10.1111/j.1540-8159.2007.00830.x.
12. Costard-Jäckle A, Goetsch B, Antz M, Franz MR. Slow and long-lasting
modulation of myocardial repolarization produced by ectopic activation
in isolated rabbit hearts. Evidence for cardiac “memory”. Circulation.
1989;80:1412–1420.
13. Sosunov EA, Anyukhovsky EP, Rosen MR. Altered ventricular stretch
contributes to initiation of cardiac memory. Heart Rhythm. 2008;5:106–
113. doi: 10.1016/j.hrthm.2007.09.008.
14. Shvilkin A, Bojovic B, Vajdic B, Gussak I, Zimetbaum P, Josephson ME.
Vectorcardiographic determinants of cardiac memory during normal
ventricular activation and continuous ventricular pacing. Heart Rhythm.
2009;6:943–948. doi: 10.1016/j.hrthm.2009.03.025.
15. Wecke L, Rubulis A, Lundahl G, Rosen MR, Bergfeldt L. Right ventricu-
lar pacing-induced electrophysiological remodeling in the human heart
and its relationship to cardiac memory. Heart Rhythm. 2007;4:1477–1486.
doi: 10.1016/j.hrthm.2007.08.001.
16. Shvilkin A, Danilo P Jr, Wang J, Burkhoff D, Anyukhovsky EP, Sosunov
EA, Hara M, Rosen MR. Evolution and resolution of long-term cardiac
memory. Circulation. 1998;97:1810–1817.
17. Janse MJ, Coronel R, Opthof T, Sosunov EA, Anyukhovsky EP, Rosen
MR. Repolarization gradients in the intact heart: transmural or api-
co-basal? Prog Biophys Mol Biol. 2012;109:6–15. doi: 10.1016/j.
pbiomolbio.2012.03.001.
18. Patel C, Burke JF, Patel H, Gupta P, Kowey PR, Antzelevitch C, Yan GX.
Is there a significant transmural gradient in repolarization time in the intact
heart? Cellular basis of the T wave: a century of controversy. Circ Arrhythm
Electrophysiol. 2009;2:80–88. doi: 10.1161/CIRCEP.108.791830.
19. Yu H, McKinnon D, Dixon JE, Gao J, Wymore R, Cohen IS, Danilo P
Jr, Shvilkin A, Anyukhovsky EP, Sosunov EA, Hara M, Rosen MR.
Transient outward current, Ito1, is altered in cardiac memory. Circulation.
1999;99:1898–1905.
20. Coronel R, Opthof T, Plotnikov AN, Wilms-Schopman FJ, Shlapakova IN,
Danilo P Jr, Sosunov EA, Anyukhovsky EP, Janse MJ, Rosen MR. Long-
term cardiac memory in canine heart is associated with the evolution of
a transmural repolarization gradient. Cardiovasc Res. 2007;74:416–425.
doi: 10.1016/j.cardiores.2007.02.024.
21. Ghosh S, Rhee EK, Avari JN, Woodard PK, Rudy Y. Cardiac memory in pa-
tients with Wolff-Parkinson-White syndrome: noninvasive imaging of ac-
tivation and repolarization before and after catheter ablation. Circulation.
2008;118:907–915. doi: 10.1161/CIRCULATIONAHA.108.781658.
 482  Circ Arrhythm Electrophysiol  April 2015
22. Marrus SB, Andrews CM, Cooper DH, Faddis MN, Rudy Y.
Repolarization changes underlying long-term cardiac memory due to
right ventricular pacing: noninvasive mapping with electrocardiographic
imaging. Circ Arrhythm Electrophysiol. 2012;5:773–781. doi: 10.1161/
CIRCEP.112.970491.
23. Spragg DD, Akar FG, Helm RH, Tunin RS, Tomaselli GF, Kass DA.
Abnormal conduction and repolarization in late-activated myocardium
of dyssynchronously contracting hearts. Cardiovasc Res. 2005;67:77–86.
doi: 10.1016/j.cardiores.2005.03.008.
24. Jeyaraj D, Wilson LD, Zhong J, Flask C, Saffitz JE, Deschênes I, Yu X,
Rosenbaum DS. Mechanoelectrical feedback as novel mechanism of
cardiac electrical remodeling. Circulation. 2007;115:3145–3155. doi:
10.1161/CIRCULATIONAHA.107.688317.
25. Jeyaraj D, Wan X, Ficker E, Stelzer JE, Deschenes I, Liu H, Wilson
LD, Decker KF, Said TH, Jain MK, Rudy Y, Rosenbaum DS. Ionic
bases for electrical remodeling of the canine cardiac ventricle. Am J
Physiol Heart Circ Physiol. 2013;305:H410–H419. doi: 10.1152/
ajpheart.00213.2013.
26. Walton RD, Martinez ME, Bishop MJ, Hocini M, Haïssaguerre M, Plank
G, Bernus O, Vigmond EJ. Influence of the Purkinje-muscle junction on
transmural repolarization heterogeneity. Cardiovasc Res. 2014;103:629–
640. doi: 10.1093/cvr/cvu165.
27. Ozgen N, Rosen MR. Cardiac memory: a work in progress. Heart Rhythm.
2009;6:564–570. doi: 10.1016/j.hrthm.2009.01.008.
28. Jeyaraj D, Ashwath M, Rosenbaum DS. Pathophysiology and clinical im-
plications of cardiac memory. Pacing Clin Electrophysiol. 2010;33:346–
352. doi: 10.1111/j.1540-8159.2009.02630.x.
29. Kooshkabadi M, Whalen P, Yoo D, Langberg J. Stretch-activated receptors
mediate cardiac memory. Pacing Clin Electrophysiol. 2009;32:330–335.
doi: 10.1111/j.1540-8159.2008.02240.x.
30. Patel PM, Plotnikov A, Kanagaratnam P, Shvilkin A, Sheehan CT,
Xiong W, Danilo P Jr, Rosen MR, Peters NS. Altering ventricular acti-
vation remodels gap junction distribution in canine heart. J Cardiovasc
Electrophysiol. 2001;12:570–577.
31. Ricard P, Danilo P Jr, Cohen IS, Burkhoff D, Rosen MR. A role for the re-
nin-angiotensin system in the evolution of cardiac memory. J Cardiovasc
Electrophysiol. 1999;10:545–551.
Downloaded from http://ahajournals.org by on August 7, 2019
32. Plotnikov AN, Shvilkin A, Xiong W, de Groot JR, Rosenshtraukh L,
Feinmark S, Gainullin R, Danilo P, Rosen MR. Interactions between anti-
arrhythmic drugs and cardiac memory. Cardiovasc Res. 2001;50:335–344.
33. de Zwaan C, Bär FW, Janssen JH, Cheriex EC, Dassen WR, Brugada
P, Penn OC, Wellens HJ. Angiographic and clinical characteristics
of patients with unstable angina showing an ECG pattern indicating
critical narrowing of the proximal LAD coronary artery. Am Heart J.
1989;117:657–665.
34. Shvilkin A, Ho KK, Rosen MR, Josephson ME. T-vector direc-
tion differentiates postpacing from ischemic T-wave inversion in
precordial leads. Circulation. 2005;111:969–974. doi: 10.1161/01.
CIR.0000156463.51021.07.
35. Chen-Scarabelli C, Scarabelli TM. T-wave inversion: cardiac memory or
myocardial ischemia? Am J Emerg Med. 2009;27:898.e1–898.e4. doi:
10.1016/j.ajem.2008.10.037.
36. Byrne R, Filippone L. Benign persistent T-wave inversion mimicking isch-
emia after left bundle-branch block–cardiac memory. Am J Emerg Med.
2010;28:747.e5–747.e6. doi: 10.1016/j.ajem.2009.10.003.
37. O’Gara PT, Kushner FG, Ascheim DD, Casey DE Jr, Chung MK, de
Lemos JA, Ettinger SM, Fang JC, Fesmire FM, Franklin BA, Granger CB,
Krumholz HM, Linderbaum JA, Morrow DA, Newby LK, Ornato JP, Ou
N, Radford MJ, Tamis-Holland JE, Tommaso CL, Tracy CM, Woo YJ,
Zhao DX, Anderson JL, Jacobs AK, Halperin JL, Albert NM, Brindis RG,
Creager MA, DeMets D, Guyton RA, Hochman JS, Kovacs RJ, Kushner FG,
Ohman EM, Stevenson WG, Yancy CW; American College of Cardiology
Foundation/American Heart Association Task Force on Practice Guidelines.
2013 ACCF/AHA guideline for the management of ST-elevation myocar-
dial infarction: a report of the American College of Cardiology Foundation/
American Heart Association Task Force on Practice Guidelines. Circulation.
2013;127:e362–e425. doi: 10.1161/CIR.0b013e3182742cf6.
38. Shvilkin A, Bojovic B, Vajdic B, Gussak I, Ho KK, Zimetbaum P,
Josephson ME. Vectorcardiographic and electrocardiographic crite-
ria to distinguish new and old left bundle branch block. Heart Rhythm.
2010;7:1085–1092. doi: 10.1016/j.hrthm.2010.05.024.
39. Wecke L, van Deursen CJ, Bergfeldt L, Prinzen FW. Repolarization
changes in patients with heart failure receiving cardiac resynchronization
therapy-signs of cardiac memory. J Electrocardiol. 2011;44:590–598. doi:
10.1016/j.jelectrocard.2011.06.006.
40. Padeletti L, Fantappiè C, Perrotta L, Ricciardi G, Pieragnoli P, Chiostri M,
Valsecchi S, Porciani MC, Michelucci A, Fantini F. Cardiac memory in hu-
mans: vectocardiographic quantification in cardiac resynchronization ther-
apy. Clin Res Cardiol. 2011;100:51–56. doi: 10.1007/s00392-010-0209-8.
41. Shvilkin A. Cardiac memory: from electrical curiosity to clinical diagnos-
tic and research tool. In:Gussak I, Antzelevitch C, eds. Electrical Diseases
of the Heart. London: Springer; 2013:415–433.
42. Engels EB, Végh EM, VAN Deursen CJ, Vernooy K, Singh JP, Prinzen
FW. T-wave area predicts response to cardiac resynchronization therapy
in patients with left bundle branch block. J Cardiovasc Electrophysiol.
2015;26:176–183. doi: 10.1111/jce.12549.
Key Words: bundle-branch block ◼ electrocardiography ◼ vectorcardiography
◼ ventricular remodeling