Journal of the American College of Cardiology Vol. 51, No.

23, 2008
© 2008 by the American College of Cardiology Foundation ISSN 0735-1097/08/$34.00
Published by Elsevier Inc. doi:10.1016/j.jacc.2008.02.062

PRECLINICAL RESEARCH

Abnormal Sympathetic Innervation

of Viable Myocardium and the Substrate
of Ventricular Tachycardia After Myocardial Infarction
Tetsuo Sasano, MD,* M. Roselle Abraham, MD,* Kuan-Cheng Chang, MD,*
Hiroshi Ashikaga, MD, PHD,* Kevin J. Mills, BS,* Daniel P. Holt, BS,† John Hilton, PHD,†
Stephan G. Nekolla, PHD,‡ Jun Dong, MD,* Albert C. Lardo, PHD,* Henry Halperin, MD,*
Robert F. Dannals, PHD,† Eduardo Marbán, MD,* Frank M. Bengel, MD†
Baltimore, Maryland; and Munich, Germany

Objectives The aim of this study was to characterize the relationship between impaired sympathetic innervation and ar-
rhythmia with noninvasive biologic imaging in an animal model of post-infarct ventricular tachycardia (VT).

Background Innervation might be abnormal in the normally perfused borderzone of myocardial infarction, contributing to
myocardial catecholamine overexposure and arrhythmogenic risk.

Methods Myocardial infarction was induced by mid-left anterior descending coronary artery balloon occlusion in 11 pigs.
Positron emission tomography (PET) of tissue perfusion and catecholamine uptake and storage was performed
with [13N]-ammonia and [11C]-epinephrine 4 to 12 weeks later. Magnetic resonance imaging and invasive elec-
trophysiology (electroanatomic mapping, basket catheter, VT inducibility) were performed within 1 week of PET.

Results When compared with a normal database of 9 healthy animals, reduced perfusion was observed in 37 ⫾ 7% of
the left ventricle (LV). Epinephrine retention was reduced in 44 ⫾ 7% of LV, resulting in a perfusion/innervation
mismatch of 7 ⫾ 4% LV. Sustained monomorphic VT was inducible in 7 of 11 animals. These animals showed a
larger perfusion/innervation mismatch (10 ⫾ 4% vs. 4 ⫾ 2% LV for animals without VT; p ⫽ 0.02). Regionally,
the degree of perfusion/innervation mismatch did not correlate with wall thickness or thickening but showed a
significant correlation with reduced myocardial voltage (r ⫽ 0.93; p ⫽ 0.001) and with the site of earliest VT
activation (chi-square 13.1; p ⬍ 0.001).

Conclusions Noninvasive mapping of cardiac sympathetic nerve terminals reveals regionally impaired catecholamine uptake
and storage in the normally perfused borderzone after experimental myocardial infarction. These areas might be
useful to characterize the individual risk for ventricular arrhythmia. (J Am Coll Cardiol 2008;51:2266–75)
© 2008 by the American College of Cardiology Foundation

Despite the high success rate of coronary revasculariza-
tion, life-threatening ventricular arrhythmia remains an
important cause of mortality after myocardial infarction
(MI) (1,2). Improved characterization of the arrhythmo-
genic substrate is thus important for optimized identifi-
cation of individuals at risk and improved therapeutic
decision making.
Impairment of sympathetic innervation is thought to be
critically involved in ventricular arrhythmogenesis. Overex-
From the Divisions of *Cardiology and †Nuclear Medicine, Johns Hopkins University
School of Medicine, Baltimore, Maryland; and the ‡Nuklearmedizinische Klinik,
Technical University of Munich, Munich, Germany. This study was supported by the
Donald W. Reynolds Foundation.
Manuscript received November 15, 2007; revised manuscript received February 13,
2008, accepted February 19, 2008.
posure to catecholamines and impaired pre-/post-synaptic
balance seem to promote electromechanical instability
(3–5). Nuclear imaging with radiolabeled catecholamine ana-
logs allows for noninvasive assessment of the integrity of
sympathetic nerve terminals in the myocardium. With this
approach, it has been demonstrated that sympathetic nerve
terminals are more susceptible to ischemic damage than
cardiomyocytes (6,7). After MI, regional impairments of
neuronal catecholamine uptake frequently exceed the area of
scar (8 –12). This pattern moderately correlated with non-
invasive electrophysiological abnormalities (11,12). In other
studies, an association of impaired innervation with ventric-
ular refractoriness was described (13), and alterations of
catecholamine uptake have been observed in patients with
arrhythmia but without coronary artery disease (3,4,14).
JACC Vol. 51, No. 23, 2008 Sasano et al. 2267
June 10, 2008:2266–75 Impaired Innervation and Post-Infarct VT

A direct link between the presence and extent of impaired QUANTITATIVE PET DATA PRO- Abbreviations
sympathetic innervation, regional myocardial electrophysi- CESSING. Attenuation-corrected and Acronyms
ological alterations, and the presence and site of inducible transaxial PET images were re-
BV ⴝ bipolar voltage
ventricular tachycardia (VT) has not yet been established constructed by filtered back-
EAM ⴝ electroanatomical
after MI. In an effort to identify the contribution of projection. The PET data were mapping
impaired sympathetic innervation to the arrhythmogenic analyzed by an observer blinded
EPS ⴝ electrophysiological
substrate, we employed a previously established, well- to electrophysiological results. studies
controlled animal model of MI (15) and performed regional With volumetric sampling of the ERP ⴝ effective refractory
characterization with detailed invasive electrophysiological [13N]-ammonia perfusion study, period
studies (EPS) and high-end noninvasive biologic and func- myocardial radioactivity was de- ICD ⴝ implantable
tional imaging. fined in 460 LV sectors and de- cardioverter-defibrillator
picted in a polarmap (18). Polar- LAD ⴝ left anterior
Methods maps were normalized to their descending coronary artery
maximum and used for assess- LV ⴝ left
Porcine model of MI. Myocardial infarction was induced
ment of regional perfusion. ventricle/ventricular
in 15 young farm pigs (25 to 35 kg), as previously described
Myocardial segments were trans- MI ⴝ myocardial infarction
(15). Under general anesthesia (induction with ketamine/
ferred to the [11C]-epinephrine MRI ⴝ magnetic resonance
xylazine/telazol, maintenance with 1.2% to 2.0% isoflurane),
images. The initial frame of the imaging
balloon occlusion of the mid left anterior descending coro-
sequence was checked to rule out PET ⴝ positron emission
nary artery (LAD), immediately distal to the second diag-
residual activity from the previ- tomography
onal branch, was performed for 150 min under fluoroscopic 13
ous [ N]-ammonia injection. RV ⴝ right ventricle
guidance. To prevent fatal arrhythmia, lidocaine was pro-
Arterial input function was de- VT ⴝ ventricular
phylactically given, and ventricular fibrillation was treated
rived from a small circular region tachycardia
by rapid cardioversion. Post-operative treatment included
of interest in the LV cavity and
narcotics and nonsteroidal anti-inflammatory drugs (ketoro-
corrected for the presence of
lac tromethamine). One animal died owing to ventricular
[11C]-labeled epinephrine metabolites. An epinephrine re-
fibrillation during balloon occlusion, and 3 animals died
from sudden death in the follow-up period, without com- tention index, measured in %/min, was calculated by nor-
pletion of further procedures. malizing myocardial activity at 40 min to the integral
Eleven pigs completed the study. Positron emission under the metabolite-corrected arterial blood time activity
tomography (PET) was performed at 4 to 12 weeks (6.6 ⫾ curve and was depicted in a parametric polarmap for the
3.2 weeks) after the procedure; magnetic resonance imaging entire LV.
(MRI) and EPS were performed within 1 week of PET. GLOBAL POLARMAP ANALYSIS. Individual polarmaps were
The experimental protocol was approved by the Institu- compared with reference normal databases created from
tional Animal Care and Use Committee. Animals were results of 9 healthy age- and weight-matched pigs under-
maintained in accordance with the guiding principles of the going the same PET protocol. Myocardial perfusion defect
American Physiological Society. was described as percent area of the polarmap depicting
PET. DATA ACQUISITION. The [11C]-epinephrine was
relative uptake values below 2.5 SD of the mean of healthy
synthesized as previously described (16). The PET imaging
control subjects. Similarly, innervation defect was described
was performed with a GE Advance scanner (GE Medical
as an epinephrine retention index below 2.5 SD of the mean
Systems, Milwaukee, Wisconsin), during general anesthe-
of control subjects (Fig. 1). A global perfusion/innervation
sia. After supine positioning, a transmission scan of 10 min
mismatch was calculated by subtracting perfusion from
was acquired for correction of photon attenuation. To
measure myocardial perfusion at rest, a static image of 15 innervation defect size. Additionally, global catecholamine
min was acquired 5 min after injection of 400 to 600 MBq storage capacity was described as mean LV epinephrine
of [13N]-ammonia. After a break of 40 min to allow for retention index of the whole polarmap.
radioactivity decay, 600 to 800 MBq of [11C]-epinephrine REGIONAL POLARMAP ANALYSIS. A region-of-interest
with a specific activity ⬎1 Ci/mmol were injected, and a model of 9 myocardial segments was applied, as previously
dynamic imaging sequence (14 frames; 6 ⫻ 30, 2 ⫻ 60, 2 ⫻ established (19). Segments comprised the apex and basal as
150, 2 ⫻ 300, 2 ⫻ 600 s) was acquired over 40 min. well as distal portions of septal, anterior, lateral, and inferior
METABOLITE ANALYSIS. To determine the contribution of wall. For accurate matching with subsequent tests, septum
[11C]-labeled metabolites to blood activity, venous blood was defined as the area between anterior and posterior
samples were drawn at 1, 5, 10, 20, and 40 min after insertion of the right ventricle (RV) on short-axis slices, and
injection of [11C]-epinephrine. Plasma metabolites were the other myocardial walls were defined accordingly (Fig. 2A).
assayed with Sep-Pak cartridges (Waters Corp, Milford, Mas- For perfusion and innervation polarmaps, average SDs from
sachusetts) as previously described (17). normal values were recorded in each segment. The degree of
2268 Sasano et al.
Impaired Innervation and Post-Infarct VT
Polarmaps of the Mean and SD of
Figure 1
Innervation and Perfusion in 9 Healthy Pigs
Ant ⫽ anterior wall; Inf ⫽ inferior wall; Lat ⫽ lateral wall; Sep ⫽ septal wall.
perfusion/innervation mismatch was expressed as the differ-
ence between SDs from normal subjects for both studies.
MRI. Magnetic resonance imaging was performed with
a 1.5-T scanner (Signa, GE Healthcare Technologies,
Milwaukee, Wisconsin) and a 4-channel cardiac phased-
array coil. An electrocardiography (ECG)-gated cine
mode acquisition was used for assessment of ventricular
function and geometry (FIESTA, flip angle 50, echo time
[TE] 1.0 ms, repetition time [TR] 4.0 ms, field-of-view
[FOV] 26 cm, slice thickness 8 mm, 30 frames). Five animals
underwent delayed enhancement imaging after intravenous
administration of gadolinium-diethyltriaminepentaacetic acid
(inversion-recovery fast-gradient-echo pulse sequence, flip an-
gle 20, TE 4.2 ms, TR 7.6 ms, FOV 26 cm, slice thickness 8
mm, contrast dose 0.2 mmol/kg, imaging 10 min after injec-
tion). With CINETOOL software (GE Healthcare Technol-
ogies), the endocardium and epicardium were manually con-
toured at end-diastole and -systole in each short-axis level to
calculate left ventricular ejection fraction (LVEF), volumes,
and mass along with regional wall thickness and thickening
and segmental scar percentage in animals undergoing delayed
enhancement imaging. For accurate matching with PET,
septum was defined as the area between anterior and posterior
insertion of the RV, and the other myocardial walls were
defined accordingly. The apex, which is not appropriately
imaged by short-axis images, was excluded.
EPS. VT INDUCIBILITY. After induction of anesthesia and
surgical venous access, a quadripolar electrophysiology cath-
eter was placed in the RV for programmed stimulation as
JACC Vol. 51, No. 23, 2008
June 10, 2008:2266–75
previously described (15); pacing was performed at twice the
pacing threshold. Extra stimuli were delivered after 8 drive
beats (cycle length 250, 300, 350 ms). The first stimulus
(S2) was initially set at 200 to 220 ms after the last pacing
stimulus of the drive train (S1). The S2 was then delivered
at progressively shorter coupling intervals, scanning in
10-ms steps until the effective refractory period (ERP) was
reached. If no arrhythmias were observed, S2 was reset to 30
ms outside ERP. A second stimulus (S3) was added 170 to
200 ms after S2, and scanning in 10-ms decrements was
repeated until S2 and S3 were both refractory or equal to
140 ms. If no arrhythmias were induced, a third stimulus
(S4) was similarly introduced. After reaching a coupling
interval of 140 ms or refractoriness with all extra stimuli,
burst pacing was performed by delivering 20 beats at a cycle
length of 280 ms. If no arrhythmias were observed, burst
cycle length was progressively scanned in 10-ms increments
until arrhythmia induction, 2:1 conduction, or a minimum
cycle length of 200 ms. If no arrhythmias were observed, the
catheter was repositioned in the RV outflow tract and the
aforementioned pacing protocols were repeated. The end
point was either completion of the protocol to refractoriness
or consistent induction of sustained monomorphic VT (at
least twice with the same morphology).
BASKET CATHETER RECORDING. If VT was induced, a 64
electrode basket catheter was inserted into the LV. Local
activation time at any site was defined as time from the
onset of the surface QRS to the time the first deflection of
the local electrogram crossed the baseline. The site of
earliest electrical activation was identified accordingly, and
the corresponding anatomical site was identified by fluoro-
scopic identification of the electrode with the earliest
activation time. This anatomical site was assigned to 1 of
the 9 myocardial segments.
ELECTROANATOMICAL MAPPING (EAM). Electroanatomical
mapping was performed with the CARTO system
(CARTO XP, Biosense-Webster Inc., Diamond Bar, Cal-
ifornia). During sinus rhythm, LV and RV endocardium
was fully mapped to achieve ⬍15 mm and ⬍20 mm filling.
For voltage mapping, bipolar voltage (BV) ⬍0.5 mV was
defined as infarct scar, and BV between 0.5 and 1.5 mV was
considered infarct border zone, according to previously
established criteria (20). Co-registration with PET and MR
data was achieved by identification of septum on BV maps
and definition of other myocardial walls accordingly (Fig.
2B). Scar and borderzone were expressed as a percent of the
entire area assigned to each myocardial segment. If VT was
induced and sustained, EAM was performed to obtain
propagation mapping for visualization of the earliest acti-
vation site (n ⫽ 5).
Statistical analysis. All data are shown as mean ⫾ SD. A
p value ⬍0.05 was considered statistically significant, unless
specified otherwise. Pearson’s correlation with Fisher r-to-z
was used to describe the relationship between continuous
variables. Receiver-operating characteristic (ROC) analysis
JACC Vol. 51, No. 23, 2008 Sasano et al. 2269
June 10, 2008:2266–75 Impaired Innervation and Post-Infarct VT

LA SA Polarmap

A
Myocardial
Segments
basal
anterior

distal
anterior
dist. bas.
dist. late- late-
bas. apical
sep- ral ral
sep-
tal
tal
Base distal
inferior

basal
inferior
RV RV
Septum
LV LV
B

Apex

3D Anterior 3D Apical

Figure 2 Segmental Analysis of LV Myocardium

With septum and cardiac long axis as anatomic landmarks, data are assigned to 9 myocardial segments in positron emission tomography (A) and CARTO (Biosense-
Webster Inc., Diamond Bar, California) electroanatomical mapping (B). LA ⫽ long axis; LV ⫽ left ventricle; RV ⫽ right ventricle; SA ⫽ short axis; other abbreviations as
in Figure 1.

was performed with MedCalc Software (Mariakerke, Bel-
gium). All other statistical analysis was carried out with
Statview 5.0 (SAS Institute, Inc., Cary, North Carolina).
The Student t test was used for comparison between
groups of animals with/without VT inducibility and (as-
suming independence of segments) for comparison of seg-
ments with/without VT inducibility. Assuming indepen-
dence of segments, 1-way factorial analysis of variance was
used for comparison of a continuous imaging variable from
EPS, MRI, or PET between 9 segments, and the post hoc
Bonferroni-corrected t test was used to assess all possible
2-way comparisons (36 in total), resulting in a corrected
p value of 0.0014 for significance. Assuming independence
of segments, the chi-square test was used to evaluate the
relationship between VT inducibility and the presence/
absence of a perfusion/innervation mismatch in segments.
To support the assumption of independence of myocar-
dial segments, a correlation analysis was performed with the
most important PET variable, the mismatch between per-
fusion and innervation abnormality. Values in the distal
anterior wall segment (the predominant site of VT induc-
ibility) did not correlate significantly with corresponding
values from any of the 8 other segments. Additionally, there
is biological support for the independence of segments from
a prior publication using a similar animal model (21), where
multiple different VT morphologies were observed between
and within animals, supporting that the segment showing
VT inducibility can be treated as an independent entity.
Results
Electrophysiological characterization. During EPS, sus-
tained monomorphic VT was inducible in 7 of 11 animals.
Eight sites of earliest VT activation were identified in the 7
animals. All VTs were reproducibly induced by triple
extrastimuli pacing, and a subgroup showed entrainment
during overdrive pacing, suggesting a re-entry mechanism.
Four were terminated by burst pacing; the remaining 4
needed cardioversion to be terminated. On 12-lead surface
ECG, the QRS configuration of the VT showed a superior
axis and right bundle branch block pattern in 7 (Fig. 3) and
a superior axis and left bundle branch block pattern in 1 VT.
According to endocardial mapping, the earliest activation
site and thus the regional VT exit site was frequently located
in the distal anterior wall segment (Figs. 3 and 4A). Time
between infarction and imaging was not different between
inducible and noninducible animals (48 ⫾ 24 days vs. 43 ⫾
23 days; p ⫽ 0.80).
The EAM localized very low voltage areas (⬍0.5 mV)
corresponding to scar predominantly in the distal septal
2270 Sasano et al. JACC Vol. 51, No. 23, 2008
Impaired Innervation and Post-Infarct VT June 10, 2008:2266–75

Figure 3 Electrophysiological Characterization of Inducible VT in a Representative Animal

(A) A 12-lead surface electrocardiogram during ventricular tachycardia (VT) shows a superior axis and right bundle branch block pattern, which is compatible with distal
anterior origin in the pig heart. (B) A 64-polar endocardial basket catheter indicates the earliest endocardial VT activation site (arrow). The corresponding electrode on
the basket catheter, as identified fluoroscopically, is marked by a circle and indicates a distal anterior location.

wall, whereas the borderzone with mildly reduced voltage of
0.5 to 1.5 mV was localized in the distal anterior wall
(matching the earliest activation site of VT) but also to a
lesser degree in the distal septal wall and apex (Fig. 4A).
LV function and geometry. The LVEF was 33 ⫾ 4% at
MRI and not significantly different between animals with
and without VT inducibility (p ⫽ 0.08) (Table 1). No
differences between groups were detected for end-diastolic
volume (p ⫽ 0.35), end-systolic volume (p ⫽ 0.20), and LV
mass (p ⫽ 0.75).
Regional end-diastolic wall thickness and wall thickening
were lowest in the distal septum, although differences were
small. Wall thickness was otherwise homogeneous, whereas
wall thickening showed a gradient between basal and distal
segments (Fig. 4B). In the 5 animals undergoing delayed
enhancement imaging, infarct location coregistered with
PET defects and reduced CARTO voltage. On a segmental
basis, presence and transmural extent of infarct correlated
significantly with PET perfusion reduction (r ⫽ 0.69; p ⬍
0.001), epinephrine reduction (r ⫽ 0.77; p ⬍ 0.001), and
reduced voltage (r ⫽ 0.78; p ⬍ 0.001) but not with
perfusion/innervation mismatch (r ⫽ ⫺0.06; p ⫽ 0.71).
Myocardial perfusion and sympathetic innervation. On a
per-animal level, PET perfusion defect size was 37 ⫾ 7% of
LV and innervation defect size was 44 ⫾ 7% of LV,
resulting in a mismatch of 7 ⫾ 4% of LV. Absolute defects
for perfusion (Table 1) (p ⫽ 0.32) and innervation (p ⫽
0.82) were not different in animals with and without VT
inducibility, but the extent of perfusion/innervation mis-
match (p ⫽ 0.019) (Fig. 5) was significantly different.
Receiver-operating characteristic analysis revealed an area
under the curve of 0.95 (95% confidence interval 0.64 to
0.98) for the perfusion/innervation mismatch to identify
inducible VT, which was higher than that of any other
global parameter (Table 1).
On a segmental level, regional analysis showed that
myocardial perfusion deviated most significantly from nor-
mal subjects in the distal septum and apex, whereas seg-
mental epinephrine retention index deviated significantly in
distal septum, apex, and distal anterior wall. This resulted in
the largest perfusion/innervation mismatch in the distal
anterior wall, the segment where VT most frequently
originated and where areas of mildly reduced voltage sug-
gestive of the infarct borderzone were identified by voltage
mapping (Fig. 4C). Figure 6 shows representative imaging
results in an animal with VT inducibility.
Segmental analysis and correlations. Table 2 summarizes
results in the 8 segments that were identified as sites of
earliest activation of VT. These segments showed a signif-
icantly higher degree of perfusion/innervation mismatch
and a higher percentage of mildly reduced voltage of 0.5 to
1.5 mV. Additionally, the earliest site of VT activation was
significantly associated with presence of a segmental perfu-
sion/innervation mismatch, if a 1.5-SD reduction of inner-
vation below perfusion was chosen as the threshold to
classify a segment as mismatch (chi-square 13.1; p ⬍ 0.001).
JACC Vol. 51, No. 23, 2008 Sasano et al. 2271
June 10, 2008:2266–75 Impaired Innervation and Post-Infarct VT

Site of VT Voltage<0.5mV (% of segment) Voltage 0.5-1.5mV (% of segm.)

0 4±7 12 ± 13

6 13 ± 20 48 ± 26 ‡
A 1 0 32 13±
0
±
2
± 33 40±
1
±
13
±
0 0 0
EPS 0
1 ±
± 14 0 7 ±
± 18‡ 3 18
34 ‡ 17
0 0
0 1±1 11 ± 17

0 0±0 0±0

ED Wall Thickness (mm) Wall Thickening (%)

7.9±1.2 51 ± 12

7.0±1.5 20 ± 11
B 7.9
5.9
7.5 7.4
± ± 50
14
18
±
42
±
* p<0.0014 vs 8 other segments
† p<0.0014 vs 7 other segments
±
MRI ±
±
1.7
0.9 1.7 ±
21
12 22 ‡ p<0.0014 vs 6 other segments
1.1 21
6.4±1.1 21± 24

8.4±1.9 30 ± 29

Perfusion (SD from Normals) EPI Retention (SD from Normals) Mismatch (SD from Normals)

-0.5±1.0 -1.3±0.9 0.8±1.2

-1.9±1.1 -4.6±1.3‡ 2.7±0.6*

C -1.1
-5.4 -4.6± ±
-0.1 -1.4
± -0.7
-4.6 -4.9± ±
-0.8 -1.9
± -0.4
-0.9 0.3±
0.7 0.5
± ±
± 0.8† 0.9 1.4 ± 0.8‡ 0.6 0.6 ±
PET ±
1.6†
±
0.9‡
±
1.2
0.7 0.9 1.2
0.8 0.7 0.9
-0.9±0.8 -1.8±0.6 0.9±0.6

-1.2±1.2 -0.9±0.7 -0.3±1.0

Figure 4 Regional Results of EPS, MRI, and PET

Regional results of EPS (A), MRI (B), and PET (C) in the 9 myocardial segments described in Figure 2. Shown are mean ⫾ SD of all animals, except for site of ventricu-
lar tachycardia (VT) inducibility, which is shown as an absolute number. For each variable except site of VT, values in segments were compared with each other with
analysis of variance, assuming independency of segments. Symbols indicate that the respective segment, highlighted in orange, differed significantly from a large num-
ber of other segments (*8, †7, and ‡6, respectively) after a Bonferroni-corrected post hoc t test, with a p value of 0.0014 to define significance after adjustment for
multiple comparisons. ED ⫽ end-diastolic; EPI ⫽ [11C]-epinephrine; EPS ⫽ electrophysiological studies; MRI ⫽ magnetic resonance imaging; PET ⫽ positron emission
tomography; VT ⫽ ventricular tachycardia.

Finally, correlation analysis was performed for mean
values in the 9 myocardial segments. Abnormality of seg-
mental innervation correlated significantly with reduced
segmental voltage, giving further support to the electrophys-
iological implications of impaired innervation (Fig. 7).
Discussion
In summary, our data show that impairment of myocardial
catecholamine uptake and storage exceeds the reduction of
tissue perfusion after experimental MI. This dysinnervation
of normally perfused viable myocardium occurs in the
infarct borderzone, is more extensive if sustained VT is
inducible, and correlates regionally with reduced voltage and
the earliest endocardial activation site of VT. These results
suggest a contribution of impaired sympathetic innervation
to the substrate of post-infarct VT.
Using several different radiolabeled catecholamine ana-
logs, previous studies have suggested that sympathetic nerve
terminals are more susceptible to ischemic damage than
cardiomyocytes. Significant impairments of innervation
have been observed after short periods of experimental
ischemia (7), and abnormally reduced catecholamine uptake
has been observed in hibernating viable myocardium of pigs
(22) and in patients with extensive coronary artery disease
but without evidence of MI (23). After MI, several studies
reported innervation impairments exceeding the extent of
scar (8 –12), and regional correlation between innervation
defect and the ischemic area at risk before revascularization
has been described (10). Few of these studies reported some
association of impaired innervation with noninvasive elec-
trophysiological parameters such as frequency of ventricular
arrhythmia (11), heart rate variability (24), and depolariza-
tion and repolarization (12). The present study is the first to
substantiate the link between abnormal innervation of viable
myocardium after MI and ventricular arrhythmia by a
detailed regional correlation between impaired myocardial
catecholamine handling, invasive electrophysiology, and the
site of inducible VT.
Sympathetic innervation is believed to play a key role in
ventricular arrhythmogenesis (25). Impairment of catechol-
amine reuptake contributes to reduced clearance of neuro-
transmitter from the synaptic cleft and thus to myocardial
 2272
Impaired Innervation and Post-Infarct VT
Sasano et al.
Individual and Grouped Results From EPS, PET, and MRI
Table 1 Individual and Grouped Results From EPS, PET, and MRI

PET
MRI
EPS VT Perfusion Innervation Mean EPI Perfusion/Innervation
Animal # Inducibility Defect (%LV) Defect (%LV) Retention (%/min) Mismatch (%LV) LVEF (%) EDV (ml) ESV (ml) LV Mass (g)
1 Yes 28 42 15.7 14 34 108 71 85
2 Yes 36 50 14.3 14 31 71 49 71
3 Yes 30 35 16.4 5 36 81 52 72
4 No 49 50 14.6 1 36 100 64 78
5 Yes 42 50 15.9 8 37 97 61 78
6 Yes 33 40 18.2 7 35 72 46 43
7 No 47 52 15.4 5 27 129 95 53
8 Yes 36 43 16.5 7 30 45 31 61
9 No 33 39 15.8 6 27 71 52 61
10 Yes 41 52 15.2 11 37 82 52 60
11 No 30 33 17.1 3 30 74 52 —
All (n ⫽ 11) — 37 ⫾ 7 44 ⫾ 7 16 ⫾ 1 7⫾4 33 ⫾ 4 84 ⫾ 23 57 ⫾ 16 66 ⫾ 13
VT yes (n ⫽ 7) — 35 ⫾ 5 45 ⫾ 6 16 ⫾ 1 10 ⫾ 4* 34 ⫾ 3 79 ⫾ 20 52 ⫾ 12 67 ⫾ 14
VT no (n ⫽ 4) — 40 ⫾ 10 44 ⫾ 9 16 ⫾ 1 4⫾2 30 ⫾ 4 93 ⫾ 27 65 ⫾ 20 64 ⫾ 13
AUC for VT, mean (95% CI) — 0.64 (0.31–0.89) 0.55 (0.24–0.84) 0.57 (0.26–0.85) 0.95 (0.64–0.98) 0.79 (0.45–0.96) 0.61 (0.28–0.87) 0.68 (0.35–0.91) 0.62 (0.28–0.88)

Summed data are shown as mean ⫾ SD. *p ⫽ 0.02 versus VT no.

AUC ⫽ area under receiver-operating characteristic curve; CI ⫽ confidence interval; EDV ⫽ end-diastolic volume; EPI ⫽ [11C]-epinephrine; EPS ⫽ electrophysiological studies; ESV ⫽ end-systolic volume; LV ⫽ left ventricle; LVEF ⫽ left ventricular ejection fraction; MRI ⫽
magnetic resonance imaging; PET ⫽ positron emission tomography; VT ⫽ ventricular tachycardia.

JACC Vol. 51, No. 23, 2008

June 10, 2008:2266–75
JACC Vol. 51, No. 23, 2008 Sasano et al. 2273
June 10, 2008:2266–75 Impaired Innervation and Post-Infarct VT

To get the most accurate information about the pathobio-

60 logical interrelationship between arrhythmia and impaired
Inducible VT (n=7)
50 No VT (n=4) sympathetic innervation, we chose the tracer [11C]-
% of Left Ventricle

40
epinephrine as a radiolabeled natural catecholamine and PET
as the currently most powerful noninvasive biologic imaging
30
technique. Advantages of this approach over the more fre-
20 *p=.02 quently used conventional scintigraphy with [ 123 I]-
10 metaiodobenzylguanidine, which might show low uptake in
0
reduced ejection fraction (6,22), are higher spatial resolution,
Perfusion Innervation Perfusion/ improved regional myocardial assessment, and the more phys-
Defect Defect Innervation
Mismatch iological tracer characteristics. Epinephrine reflects neuronal
uptake, metabolism, and storage and has thus been considered
Figure 5 PET Defect Sizes in Animals a more physiological tracer than the more popular PET
With and Without Inducible VT compound [11C]-hydroxyephedrine (HED), which primarily
Defects are measured in percentage of left ventricular myocardium, which traces uptake-1 capacity (17). This physiological behavior at
shows tracer retention below 2.5 SDs from the mean of healthy control ani- the same time represents a challenge because of enzymatic
mals. PET ⫽ positron emission tomography; VT ⫽ ventricular tachycardia.
metabolism by monoaminooxidases, which is why metabolite
correction needs to be performed. Although our high-end
catecholamine overexposure. The resulting pre-/post- approach yields conclusive results, future studies might focus
synaptic imbalance is thought to contribute to electrome- on the reproducibility of our findings with other catecholamin-
chanical instability and thus arrhythmogenesis. This is ergic tracers to simplify neuronal imaging strategies for a
supported by the observation of impaired innervation in the broader-scale clinical setting.
absence of coronary artery disease in a variety of arrhyth- It is also important to recognize the biological/functional
mogenic diseases such as idiopathic ventricular fibrillation nature of the nuclear imaging signal. Although some prior
(4), RV outflow tract tachycardia (3), and Brugada syn- histologic studies reported focally increased density of nerves in
drome (26). In these studies, however, as in other studies the peri-infarct area and postulated nerve sprouting and hy-
after MI, no regional correlation between innervation status perinnervation as a mechanism contributing to VT (27), these
and the origin of arrhythmia was performed. morphologic results are not conflicting with the functional

Figure 6 Study Results in a Representative Animal With Inducible VT

Positron emission tomography polarmaps (left) show regional distribution of [13N]-ammonia and [11C]-epinephrine retention (top), along with perfusion (white) and inner-
vation defects (blue; bottom). Segmental analysis (bottom right) shows that innervation defect exceeds perfusion defect predominantly in the distal anterior wall seg-
ment, highlighted in yellow. Three-dimensional electroanatomical maps, depicted in anterior view on top right, show reduced bipolar endocardial voltage in apex and
distal septum/anterior wall. Propagation maps show earliest activation of VT in the infarct borderzone in the distal anterior wall (red arrow). Abbreviations as in Figures
1, 2, and 4.
2274 Sasano et al.
Impaired Innervation and Post-Infarct VT
Characterization of Myocardial Segments With Earliest VT Activation
Table 2 Characterization of Myocardial Segments With Earliest VT Activation
Earliest Activation of VT
Perfusion abnormality (SD from normal subjects)
EPI retention abnormality (SD from normal subjects)
Perfusion/innervation mismatch (SD)
End-diastolic wall thickness (mm)
Wall thickening (%)
Voltage ⬍0.5 mV (% of segment)
Voltage 0.5–1.5 mV (% of segment)
*p ⬍ 0.001; †p ⫽ 0.004 versus other segments (assuming independence of segments).
EPI ⫽ [11C]-epinephrine; VT ⫽ ventricular tachycardia.
results of the present study. Higher density of nerves at 1 site
might be associated with lower density at several other sites,
and the sum of all microscopic changes might still be a
macroscopic reduction of catecholamine uptake and storage
capacity identified by PET; in addition, even if histologic
density of nerves is increased, their functionality might still be
abnormal and the overall regional potential for catecholamine
uptake and storage might be reduced.
There is clinical interest in using noninvasive cat-
echolaminergic imaging to identify individuals at high risk
for life-threatening arrhythmia and to assist in localization
of the regional substrate of ventricular arrhythmia for
targeted therapy. The problem associated with implantable
cardioverter-defibrillator (ICD) implantation is that some
patients with sudden cardiac death do not have a high-risk
profile and therefore are missed by currently applied criteria
(e.g., low LVEF). Conversely, some so-called high-risk
patients are not really at high risk, and their ICDs never fire.
The quest for better selection of candidates for ICDs
(28,29), together with the increasing interest in targeted
ablation (30,31) or molecular therapy (15), emphasize the
need for noninvasive diagnostic tests to risk-stratify indi-
viduals for ventricular arrhythmias. The present study sup-
Figure 7 Correlation of Segmental Results
of PET and Electroanatomic Mapping
Shown is a plot for average values of reduced voltage versus EPI retention
abnormality in 9 myocardial segments. Standard deviations for each segment
for both variables are indicated by horizontal/vertical lines originating from
each data point. Abbreviations as in Figure 4.
JACC Vol. 51, No. 23, 2008
June 10, 2008:2266–75
Not Earliest Activation
(n ⴝ 8) (n ⴝ 91)
⫺1.3 ⫾ 1.6 ⫺1.9 ⫾ 2.1
⫺3.5 ⫾ 1.8 ⫺2.3 ⫾ 1.8
2.2 ⫾ 1.5* 0.4 ⫾ 1.2
7.4 ⫾ 1.4 7.3 ⫾ 1.6
18 ⫾ 13 32 ⫾ 25
4⫾6 7 ⫾ 18
40 ⫾ 29† 16 ⫾ 21
ports a relationship between regionally impaired sympa-
thetic innervation and VT inducibility and thus provides a
mechanistic foundation for future clinical trials to assess the
use of noninvasive innervation imaging in the work-up of
patients for post-infarct VT risk. Previous observations,
along with the sudden death of a small fraction of our
animals, suggest that our pig model not only shows induc-
ible VT at programmed stimulation but also develops
spontaneous arrhythmia. Death of our animals, however,
occurred without completion of imaging procedures. Some
clinical data support the usefulness of VT inducibility to
predict spontaneous VT (28), but the topic of discussion is
currently controversial. Whether innervation imaging can
predict not only inducible but also spontaneous ventricular
arrhythmia in the clinical setting needs to be investigated in
prospective clinical trials. Our results not only support but
also encourage such clinical studies.
Study limitations. Some limitations of this study need to
be considered. First, our results show an association between
impaired innervation of the viable infarct borderzone and
VT inducibility but do not establish a cause/effect relation-
ship. This, however, has been shown in a prior study (13),
where myocardial regions with reduced innervation on PET
showed a longer effective refractory period. It is also
indirectly supported by retrospective clinical observations of
scintigraphically impaired innervation as an independent
predictor of cardiac death (32,33). Second, although other
tomographic imaging techniques such as computed tomog-
raphy or MRI can be integrated with electroanatomical
mapping with software-based approaches (31), such meth-
odology is not yet available for PET. We thus had to limit
our regional analysis to 9 myocardial segments that were
assigned by blinded observers with a standardized technique
to achieve co-registration. Some of the variability in our
study might thus be attributed to inaccuracies of localiza-
tion, but results remain conclusive and software advances
and the application of PET-computed tomography for
bio-morphologic imaging might refine regional assessment
in the future. Finally, VT in our model was electrophysi-
ologically characterized by a re-entry mechanism. We were
not able to map the entire re-entry circuit in our study,
probably owing to a deeper intramural course. Therefore, we
JACC Vol. 51, No. 23, 2008
June 10, 2008:2266–75
chose the earliest activation site and thus the endocardial
exit site as the best available parameter for regional VT
localization, which is in line with general clinical practice.
Conclusions
In conclusion, mapping of cardiac sympathetic nerve termi-
nals provides regional biological information about the
substrate of ventricular arrhythmia in ischemic heart disease.
The site of inducible VT correlated with the area of viable
but denervated myocardium measured by PET. Our results
provide a rationale for clinical trials to evaluate the use of
noninvasive innervation imaging in the work-up of post-
infarct ventricular arrhythmia.
Acknowledgment
The authors thank Brian Caffo, PhD, for biostatistical
advice.
Reprint requests and correspondence: Dr. Frank M. Bengel,
Director, Cardiovascular Nuclear Medicine, Division of Nuclear
Medicine/PET, 601 North Caroline Street, Johns Hopkins Out-
patient Center 3225, Baltimore, Maryland 21287. E-mail:
fbengel1@jhmi.edu.
REFERENCES
1. The Antiarrhythmics versus Implantable Defibrillators (AVID) Inves-
tigators. A comparison of antiarrhythmic-drug therapy with implant-
able defibrillators in patients resuscitated from near-fatal ventricular
arrhythmias. N Engl J Med 1997;337:1576 – 83.
2. Moss AJ, Hall WJ, Cannom DS, et al., Multicenter Automatic
Defibrillator Implantation Trial Investigators. Improved survival with
an implanted defibrillator in patients with coronary disease at high risk
for ventricular arrhythmia. N Engl J Med 1996;335:1933– 40.
3. Schafers M, Lerch H, Wichter T, et al. Cardiac sympathetic inner-
vation in patients with idiopathic right ventricular outflow tract
tachycardia. J Am Coll Cardiol 1998;32:181– 6.
4. Schafers M, Wichter T, Lerch H, et al. Cardiac 123I-MIBG uptake
in idiopathic ventricular tachycardia and fibrillation. J Nucl Med
1999;40:1–5.
5. Verrier RL, Antzelevitch C. Autonomic aspects of arrhythmogenesis:
the enduring and the new. Curr Opin Cardiol 2004;19:2–11.
6. Luisi AJ Jr., Suzuki G, deKemp R, et al. Regional 11C-
hydroxyephedrine retention in hibernating myocardium: chronic in-
homogeneity of sympathetic innervation in the absence of infarction.
J Nucl Med 2005;46:1368 –74.
7. Schwaiger M, Guibourg H, Rosenspire K, et al. Effect of regional
myocardial ischemia on sympathetic nervous system as assessed by
fluorine-18-metaraminol. J Nucl Med 1990;31:1352–7.
8. Allman KC, Wieland DM, Muzik O, et al. Carbon-11 hydroxyephed-
rine with positron emission tomography for serial assessment of cardiac
adrenergic neuronal function after acute myocardial infarction in
humans. J Am Coll Cardiol 1993;22:368 –75.
9. Dae MW, Herre JM, O’Connell JW, et al. Scintigraphic assessment of
sympathetic innervation after transmural versus nontransmural myo-
cardial infarction. J Am Coll Cardiol 1991;17:1416 –23.
10. Matsunari I, Schricke U, Bengel FM, et al. Extent of cardiac sympathetic
neuronal damage is determined by the area of ischemia in patients with
acute coronary syndromes. Circulation 2000;101:2579 – 85.
11. McGhie AI, Corbett JR, Akers MS, et al. Regional cardiac adrenergic
function using I-123 meta-iodobenzylguanidine tomographic imaging
after acute myocardial infarction. Am J Cardiol 1991;67:236 – 42.
Sasano et al. 2275
Impaired Innervation and Post-Infarct VT
12. Simoes MV, Barthel P, Matsunari I, et al. Presence of sympathetically
denervated but viable myocardium and its electrophysiologic correlates
after early revascularised, acute myocardial infarction. Eur Heart J
2004;25:551–7.
13. Calkins H, Allman K, Bolling S, et al. Correlation between scintigraphic
evidence of regional sympathetic neuronal dysfunction and ventricular
refractoriness in the human heart. Circulation 1993;88:172–9.
14. Mitrani R, Klein L, Miles W, et al. Regional cardiac sympathetic
denervation in patients with ventricular tachycardia in the absence of
coronary artery disease. J Am Coll Cardiol 1993;22:1344 –53.
15. Sasano T, McDonald AD, Kikuchi K, et al. Molecular ablation of
ventricular tachycardia after myocardial infarction. Nat Med 2006;12:
1256– 8.
16. Chakraborty PK, Gildersleeve DL, Jewett DM, et al. High yield
synthesis of high specific activity R-(-)-[11C]epinephrine for routine
PET studies in humans. Nucl Med Biol 1993;20:939 – 44.
17. Munch G, Nguyen NTB, Nekolla S, et al. Evaluation of sympathetic
nerve terminals with [11C]epinephrine and [11C]hydroxyephedrine
and positron emission tomography. Circulation 2000;101:516 –23.
18. Nekolla SG, Miethaner C, Nguyen N, et al. Reproducibility of polar
map generation and assessment of defect severity and extent assess-
ment in myocardial perfusion imaging using positron emission tomog-
raphy. Eur J Nucl Med Mol Imaging 1998;25:1313–21.
19. Bengel FM, Permanetter B, Ungerer M, et al. Relationship between
altered sympathetic innervation, oxidative metabolism and contractile
function in the cardiomyopathic human heart; a non-invasive study
using positron emission tomography. Eur Heart J 2001;22:1594 – 600.
20. Marchlinski FE, Callans DJ, Gottlieb CD, et al. Linear ablation
lesions for control of unmappable ventricular tachycardia in patients
with ischemic and nonischemic cardiomyopathy. Circulation 2000;
101:1288 –96.
21. Ashikaga H, Sasano T, Dong J, et al. Magnetic resonance-based
anatomical analysis of scar-related ventricular tachycardia: implications
for catheter ablation. Circ Res 2007;101:939 – 47.
22. Luisi AJ Jr., Fallavollita JA, Suzuki G, et al. Spatial inhomogeneity of
sympathetic nerve function in hibernating myocardium. Circulation
2002;106:779 – 81.
23. Bulow HP, Stahl F, Lauer B, et al. Alterations of myocardial
presynaptic sympathetic innervation in patients with multi-vessel
coronary artery disease but without history of myocardial infarction.
Nucl Med Commun 2003;24:233–9.
24. Bengel FM, Barthel P, Matsunari I, et al. Kinetics of I-123-
metaiodobenzylguanidine after acute myocardial infarction and reper-
fusion therapy. J Nucl Med 1999;40:904 –10.
25. Zipes DP, Rubart M. Neural modulation of cardiac arrhythmias and
sudden cardiac death. Heart Rhythm 2006;3:108–13.
26. Wichter T, Matheja P, Eckardt L, et al. Cardiac autonomic dysfunc-
tion in Brugada syndrome. Circulation 2002;105:702– 6.
27. Chen P-S, Chen LS, Cao J-M, et al. Sympathetic nerve sprouting,
electrical remodeling and the mechanisms of sudden cardiac death.
Cardiovasc Res 2001;50:409 –16.
28. Buxton AE, Lee KL, Fisher JD, et al., Multicenter Unsustained
Tachycardia Trial Investigators. A randomized study of the prevention
of sudden death in patients with coronary artery disease. N Engl J Med
1999;341:1882–90.
29. Moss AJ, Zareba W, Hall WJ, et al. Prophylactic implantation of a
defibrillator in patients with myocardial infarction and reduced ejec-
tion fraction. N Engl J Med 2002;346:877– 83.
30. Brunckhorst CB, Delacretaz E, Soejima K, et al. Identification of the
ventricular tachycardia isthmus after infarction by pace mapping.
Circulation 2004;110:652–9.
31. Dong J, Calkins H, Solomon SB, et al. Integrated electroanatomic
mapping with three-dimensional computed tomographic images for
real-time guided ablations. Circulation 2006;113:186 –94.
32. Nakata T, Miyamoto K, Doi A, et al. Cardiac death prediction and
impaired cardiac sympathetic innervation assessed by MIBG in patients
with failing and nonfailing hearts. J Nucl Cardiol 1998;5:579–90.
33. Paul M, Schafers M, Kies P, et al. Impact of sympathetic innervation
on recurrent life-threatening arrhythmias in the follow-up of patients
with idiopathic ventricular fibrillation. Eur J Nucl Med Mol Imaging
2006;33:866 –70.