From the Divisions of Endocrinology and Metabolism (M.J.S., T.S.) and
Nuclear Medicine (D.M.R., K.C.A., F.D., E.F., D.M.W., M.S.), Department of
Internal Medicine, University of Michigan, Ann Arbor, Mich, and Diabetes
Research and Treatment Center (M.A.P.), Southern Illinois University School of
Medicine, Springfield, Ill.
Methods and Results—PET using the sympathetic
neurotransmitter analogue 11C-labeled
hydroxyephedrine ([11C]-HED) was used to characterize
left ventricular sympathetic innervation in diabetic
patients by assessing regional disturbances in myocardial
tracer retention and washout. The subject groups comprised 10 diabetic
subjects without DAN, 10 diabetic subjects with mild DAN, 9 diabetic
subjects with severe DAN, and 10 healthy subjects. Abnormalities of
cardiac [11C]-HED retention were detected in 40% of
DAN-free diabetic subjects. In subjects with mild
neuropathy, tracer defects were observed only in the distal
inferior wall of the left ventricle, whereas with more
severe neuropathy, defects extended to involve the distal
and proximal anterolateral and inferior walls. Absolute
[11C]-HED retention was found to be increased by 33%
(P<0.01) in the proximal segments of the severe DAN
subjects compared with the same regions in the DAN-free subjects (30%;
P<0.01 greater than the proximal segments of the mild
DAN subjects). Despite the increased tracer retention, no appreciable
washout of tracer was observed in the proximal segments,
consistent with normal regional tone but increased sympathetic
innervation. Distally, [11C]-HED retention was decreased
in severe DAN by 33% (P<0.01) compared with the
DAN-free diabetic subjects (21%; P<0.05 lower than the
distal segments of the mild DAN subjects).
Conclusions—Diabetes may result in left ventricular
sympathetic dysinnervation with proximal hyperinnervation complicating
distal denervation. This combination could result in potentially
life-threatening myocardial electrical instability and explain the
enhanced cardioprotection from ß-blockade in these subjects.
Radiolabeled analogues of norepinephrine are actively taken
up by the sympathetic nerve terminals of the heart and thereby permit
direct regional assessment of cardiac sympathetic integrity. Cardiac
scanning with 123I-labeled
metaiodobenzylguanidine
([123I]-MIBG) has identified sympathetic
denervation in most diabetic subjects with normal
cardiovascular reflex
testing,15 16 17 with abnormalities of MIBG
retention correlating with left ventricular
dysfunction.16 The radiotracer
11C-labeled hydroxyephedrine
([11C]-HED) has recently been developed as a
norepinephrine analogue for
PET.18 19 20 This has been shown to undergo highly
specific uptake and retention in the sympathetic nerve
terminals,18 19 20 which thus facilitate the
quantitative regional characterization of sympathetic neuronal
dysfunction and loss.19 Studies using
[11C]-HED have shown that its retention is
preferentially reduced in the distal left ventricular
myocardium by diabetes.19 We
speculated that this distal denervation in DAN may elsewhere induce
regional sympathetic hyperactivity, which could potentially become a
focus of electrical and chemical instability.
In this article, we report the results of a study in which PET using
[11C]-HED was used to characterize sympathetic
neuronal density and function in the proximal myocardial segments of
diabetic subjects with and without DAN.
Diabetic subjects were categorized by cardiovascular
reflex testing into diabetic subjects without DAN (DAN free), diabetic
subjects with mild DAN, and subjects with severe DAN. These diabetic
subjects were compared with healthy age-matched nondiabetic subjects (3
men, 7 women). Clinical details are given in Table 1
Autonomic Function Testing
Heart rate variability was measured for 6 minutes with patients in the
supine position breathing at a fixed rate of 5 breaths per minute by
standardized protocol.24 First, the E/I
ratio25
(R-Rmax/R-Rmin) over 25
breaths was determined. Spectral analysis was then used to
quantify defects in sympathetic and vagal innervation of the
heart.26 The power spectrum of heart rate
variability was determined in low (0.01 to 0.05 Hz), middle (0.05 to
0.15 Hz), and high (0.15 to 0.5 Hz) frequency ranges. Abnormality was
defined as results that were below the 2.3 percentile for
age.27 A 5-minute postural study was also
performed in which the change in the blood pressure on standing was
assessed after 20 minutes of the patient lying supine. The lowest
standing systolic pressure was used to calculate postural
change in blood pressure. The Valsalva maneuver24
was then performed twice, and the mean Valsalva ratio was
calculated.
PET Studies and Image Construction
Polar Map Generation
The heterogeneity of regional left
ventricular [11C]-HED retention in
each diabetic patient was compared with the normal reference
distribution by calculating a z score,
zi=(qi-µi)/
Retention Index Calculation
For plots of time-activity data, the measured PET counts were converted
to units of megabecquerel per milliliter by multiplying by the PET
scanner calibration factor, which has units of megabecquerel per
milliliter per PET count. This calibration factor is determined weekly
by imaging a 20-cm cylindrical flood phantom filled with an aqueous
solution of [18F] at a known concentration.
Quantitative Evaluation of Regional Myocardial Perfusion
Statistical Analysis
Regional Ventricular [11C]-HED
Retention
Measurements of the absolute tissue retention index for
[11C]-HED in the proximal and distal myocardial
segments in the DAN-free diabetic subjects were not significantly
different from those of the healthy nondiabetic subjects,
consistent with previous reports.19 28
Absolute [11C]-HED retention was found to be
increased by 33% (P<0.01) in the proximal segments of the
severe DAN subjects compared with the same regions in the DAN-free
subjects (30%; P<0.01 greater than the proximal segments
of the mild DAN subjects) (Figure 2
Figure 3A
To explore neuronal synaptic vesicular integrity in the proximal and
distal myocardial segments, comparative regional time-activity curves
were generated in a severe DAN subject after administration of 20 mCi
of [11C]-HED followed 60 minutes later by 20
mCi of [11C]-labeled epinephrine
(Figure 4
Sympathetic hyperactivity and increased regional
norepinephrine levels competitively decrease the retention
of sympathetic neurotransmitter analogues, including HED, by
accelerating their washout.18 However, under
resting conditions, no significant washout of
[11C]-HED is normally observed in healthy
nondiabetic myocardium.28 Regional
[11C]-HED time-activity washout curves were
therefore used to explore whether sympathetic tone was elevated (and
hence washout was increased) in the proximal myocardial segments of the
severe DAN subjects. In these subjects, no washout of
[11C]-HED was observed in the proximal
segments, because tissue [11C]-HED activity was
constant between 6.25 and 50 minutes (Figure 5
The marked heterogeneity of regional
[11C]-HED retention in severe DAN suggests that
diabetes can provoke an extreme imbalance of left
ventricular sympathetic innervation.
[11C]-HED is taken up into the neuron by
energy-dependent uptake-1,18 29 is
nonmetabolized, and thus marks the location of functioning sympathetic
nerve terminals. The retention of [11C]-HED in
the heart is dependent on continuous recycling into and out of the
neuron,18,29 and its neuronal retention requires
intact vesicular storage. Nonneuronal [11C]-HED
retention is low, and its extraction from blood to the neuronal
axoplasm very high,29 suggesting that its
increased retention in the proximal myocardium reflects an
increased number of neuronal terminals with intact uptake-1 and
vesicular storage. Alternatively, because the net tissue uptake of
[11C]-HED is flow
limited,29 increased myocardial
[11C]-HED retention could result from an
increase in myocardial blood flow. Resting myocardial blood flow was
higher in DAN subjects than in nondiabetic and DAN-free diabetic
subjects, which may reflect the resting tachycardia that
accompanies DAN.4 5 However, comparison of blood
flow in the proximal myocardium of mild and severe DAN
subjects did not reveal a difference, suggesting that differences in
tracer delivery could not explain the increase in tracer retention in
these segments in the severe DAN subjects. The integrity of neuronal
vesicular storage in the proximal myocardial segments of severe DAN
subjects was confirmed by the demonstration that washout of
[11C]-labeled epinephrine
paralleled [11C]-HED, because nonvesicular
epinephrine would be rapidly
metabolized.29 In contrast, enhanced washout of
[11C]-HED in the distal myocardial segments of
the severe DAN subjects parallels that observed in the
myocardium of desipramine-pretreated normal
subjects28 (albeit of greater magnitude in DAN,
potentially reflecting incomplete pharmacological block of neuronal
uptake-1) and suggests that failure of distal
[11C]-HED retention in DAN specifically
reflects impaired neuronal tracer uptake resulting from uptake-1
dysfunction or neuronal loss.
Proximal myocardial hyperinnervation could result from a persistent
increase in axonal sprouting and regeneration that normally precede
organ reinnervation.30 31 Dense innervation of
the proximal versus distal myocardium may contribute to its
resistance to denervation and serve as a foundation for
ventricular reinnervation, which normally proceeds from the
base of the heart to the apex.32 In nondiabetic
animals, organ reinnervation after partial denervation requires axonal
sprouting and hyperinnervation initially within the islands of
remaining innervation,30 31 which then becomes
downregulated as distal reinnervation is completed. In severe DAN,
failure of successful distal reinnervation may lead to a sustained
increase in proximal neuronal density. Increased proximal myocardial
[11C]-HED retention was not observed in the
mild DAN subjects in whom the extent of the distal
[11C]-HED retention defect averaged only 13%
of that observed in the severe DAN subjects. It is thus tempting to
speculate that there may exist a threshold for distal left
ventricular denervation to initiate unsuccessful proximal
regeneration and repair.
We proposed that sympathetic hyperactivity may be present in the
proximal myocardium of the severe DAN subjects. In
experimental diabetes, increased sympathetic tone has been
etiologically invoked in nerve conduction
slowing,33 and increased sympathetic activity in
the rat diabetic myocardium is reported to precede chronic
sympathetic denervation.34 35 Increased regional
norepinephrine levels resulting from increased sympathetic
tone would be expected to decrease [11C]-HED
retention and accelerate washout.18 However, the
findings that [11C]-HED retention was increased
and washout was unchanged in the proximal myocardial segments in severe
DAN suggest that sympathetic tone was not abnormally elevated in these
regions.
In DAN, cardiac sympathetic dysinnervation in conjunction with
parasympathetic denervation4 5 may contribute to
the excess of cardiac deaths in diabetic
patients.2 3 4 5 6 11 12 13 14 In nondiabetic subjects,
decreased protective parasympathetic tone and heterogeneous
sympathetic cardiac innervation are thought to contribute to sudden
death.36 37 38 39 Increased regional sympathetic
innervation may predispose to ventricular fibrillation by
decreasing the arrhythmogenic threshold.39
Regional cardiac sympathetic hyperactivity increases the risk of
cardiac arrhythmias during myocardial
ischemia,10 37 and selective cardiac
sympathetic denervation or ß-blockers11 12 13 14 37
reduce the risk of both cardiac arrhythmias and sudden death in
these patients, implicating regional sympathetic imbalance as a factor
promoting arrhythmogenesis. Diabetic patients have increased mortality
after myocardial infarction,11 12 13 14 40 41 42 which
may reflect the extent of CAD or increased susceptibility to other
triggering factors,40 41 42 including autonomic
imbalance.2 3 4 5 6 7 The decrease in heart rate
variability observed in the DAN subjects in this study, despite
preservation of proximal myocardial sympathetic innervation, is
consistent with regional or global myocardial parasympathetic
denervation,4 5 which may contribute to
electrical imbalance. In diabetes, DAN has been directly implicated in
depressed diastolic filling and decreased resting left
ventricular ejection fraction.16 43
In diabetic subjects complicated by myocardial infarction, DAN is
predictive of both increased mortality and left ventricular
failure.44
Compared with nondiabetic subjects, diabetic patients experience
greater cardioprotection with ß-blockade11 12 13 14
with highly significant reductions in postinfarct
mortality13 despite poorer risk factor profiles,
implicating a role for cardiac adrenergic hyperactivity. In DAN
patients, ß-blockade with atenolol typically slows the elevated
resting heart rate,45 an effect that requires
residual sympathetic tone and is thus consistent with our
demonstration of persistent sympathetic innervation even in advanced
DAN. Potentially, proximal myocardial sympathetic hyperinnervation
could promote regional myocardial perfusion imbalances by decreasing
myocardial vascularity,46 by stimulating vascular
hyperreactivity,47 and by contributing to
paradoxical coronary vasoconstriction,48
thereby increasing cardiac mortality. In addition, increased
norepinephrine levels in these proximal segments may
contribute to malignant arrhythmogenesis by precipitating myocardial
necrosis49 secondary to increased intracellular
calcium and free radical injury.50 Therefore,
these islands of hyperinnervation may be the focus of electrical,
chemical, and vascular instability, particularly if denervation
hypersensitivity is also present.
In summary, our studies have demonstrated that diabetes may result in
left ventricular sympathetic dysinnervation with proximal
hyperinnervation complicating distal denervation. This exaggerated
sympathetic imbalance may contribute to the accelerated rate of cardiac
death in diabetes. Therefore, instead of avoiding the use of
ß-blocking drugs in DAN, we may need to reconsider their potential
benefits in this group of high-risk patients.
Received January 29, 1998;
revision received April 7, 1998;
accepted April 22, 1998.
© 1998 American Heart Association, Inc.
Clinical Investigation and Reports
Cardiac Sympathetic Dysinnervation in Diabetes
Implications for Enhanced Cardiovascular Risk
Abstract
Top
Abstract
Introduction
Methods
Results
Discussion
References
Background—Regional cardiac
sympathetic hyperactivity predisposes to malignant arrhythmias
in nondiabetic cardiac disease. Conversely, however, cardiac
sympathetic denervation predicts increased morbidity and mortality in
severe diabetic autonomic neuropathy (DAN). To unite these
divergent observations, we propose that in diabetes regional cardiac
denervation may elsewhere induce regional sympathetic hyperactivity,
which may in turn act as a focus for chemical and electrical
instability. Therefore, the aim of this study was to explore regional
changes in sympathetic neuronal density and tone in diabetic patients
with and without DAN.
Key Words: diabetes mellitus • nervous system, autonomic • imaging • ventricles
Introduction
Top
Abstract
Introduction
Methods
Results
Discussion
References
Diabetic autonomic
neuropathy (DAN) is a common complication of diabetes,
present in up to 40% of insulin-dependent diabetic patients at
diagnosis.1 It has been invoked as a cause of
sudden cardiac death in diabetic subjects both with and without
myocardial ischemia.2 3 4 5 6 7 Absent heart
rate variability secondary to DAN is predictive of both left
ventricular failure and increased
mortality.6 Left ventricular function
has been reported to be depressed in 59% of diabetic subjects with DAN
compared with only 8% of those without DAN.7
Longitudinal studies of DAN subjects have typically shown 5-year
mortality rates ranging between 
16% and
53%,2 3 4 with the highest mortality rates
corresponding to advanced cardiovascular sympathetic
denervation.4 In contrast, regional cardiac
sympathetic hyperactivity is associated with the development of
malignant ventricular arrhythmias and cardiac death
in nondiabetic humans and animals, particularly when accompanied by
reduced protective parasympathetic tone and myocardial
ischemia.8 9 10 The association of
increased cardiovascular mortality with decreased
cardiac sympathetic innervation in severe DAN appears somewhat
paradoxical, as does the enhanced protective effects of ß-adrenergic
receptor blockade in these subjects.11 12 13 14
Methods
Top
Abstract
Introduction
Methods
Results
Discussion
References
Subjects
The diabetic subjects were recruited from consecutive
presentations to the Michigan Diabetes Research and
Training Center Clinical Core. All the diabetic subjects included in
this study had undergone assessment of myocardial perfusion with PET
and [13N]NH3 at rest and
during maximal adenosine stimulation (140 µg ·
kg-1 ·
min-1)21 22 23 within a
6-month period preceding this study to confirm the absence of
coronary artery disease (CAD). The inclusion criteria for all
diabetic subjects included a diagnosis of type 1 diabetes, age
between 20 and 65 years, the absence or presence of DAN (as defined
below), and the absence of any risk factors for other causes for
neuropathy (determined by a medical history, family
history, history of medications, occupational history, history of
exposure to toxins, physical and neurological examinations, and
laboratory studies). Exclusion criteria included preexisting
cardiovascular disease, including CAD, congestive heart
failure, known arrhythmias, documented ventricular
structural abnormalities and valvular disease,
peripheral vascular disease, and uncontrolled hypertension;
a history of primary dyslipidemia requiring therapy; a
creatinine clearance <70 mL/min; or a history of previous
kidney, pancreas, or cardiac transplantation. 
. DAN-free diabetic subjects (6 men, 4
women) were selected on the basis of having 
1 abnormality on
standardized testing and no symptoms of DAN. Mild DAN subjects (5 men,
5 women) had 2 to 3 abnormal cardiac autonomic function tests (4 had
symptoms from delayed gastric emptying), and severe DAN subjects (4
men, 5 women) had 
4 abnormal cardiac autonomic function tests and
symptoms attributable to autonomic neuropathy. All severe
DAN subjects reported early satiety, nausea, periodic vomiting,
persistent constipation with periodic episodes of diarrhea, and
dizziness on standing. Four suffered from periodic episodes of urinary
retention requiring catheterization. Written informed
consent was obtained from all patients, and the study protocol was
approved by the institutional review board of the University of
Michigan.
View this table:
[in a new window]
Table 1. Clinical Characteristics of Experimental Group
Cardiovascular autonomic neuropathy
was assessed by use of a battery of autonomic function tests to
evaluate heart rate variability. All subjects were fasted and had
abstained from both prescription and nonprescription medications on the
day of assessment. Subjects were instructed to delay their morning
insulin injection until after testing and were excluded if they had
experienced a hypoglycemic episode within 8 hours of testing. Blood
glucose values were within the range of 8 to 14 mmol/L during the
assessment period.
Patients taking medications or substances (including caffeine)
known to interfere with neuronal uptake of norepinephrine
analogues abstained from such substances for 2 weeks before study.
Cardiac PET imaging was performed in a whole-body PET scanner (model
CTI 931 and Siemens/ECAT 931) with 20 mCi of
[11C]-HED and 20 mCi of
[13N]NH3.19
This scanner has 8 circular detector rings that allow the
simultaneous acquisition of 15 contiguous transaxial images
(oriented perpendicular to the sagittal and coronal planes of the body)
with a slice thickness of 6.75 mm. After introduction of a
22-gauge intravenous cannula into the antecubital vein and
placement of the tomograph with the aid of a scout image, a 15-minute
transmission study with a retractable germanium-68 ring source was
performed to correct emission data for tissue attenuation. Dynamic scan
acquisition was initiated simultaneously with the injection
of [11C]-HED. The image acquisition protocol
comprised 15 images with varying frame duration (6x30, 2x60, 2x150,
2x300, 2x600, and 1x1200 seconds). The emission data were corrected
for attenuation and reconstructed by filtered backprojection with a
Hanning filter with a cutoff frequency of 1.12 cycles per 1 cm to
produce images with a resolution of 9 to 10 mm at full width at
half-maximum in-plane. A Sun workstation (Sun Microsystems) was used to
realign the images perpendicular to the long axis of the left
ventricle, yielding 8 short-axis views (slice thickness, 0.8 cm) of
myocardial tracer distribution extending from the apex to the base of
the left ventricle. After waiting 1 hour for the
[11C] to decay after the end of data
acquisition, we evaluated resting myocardial perfusion using
[13N]NH3.
Circumferential count-profile analysis was performed on
each of the 8 short-axis images. Each short-axis slice was divided into
36 angular regions of interest ("sectors"), and the myocardial
concentration of [11C]-HED in each sector, as
determined from the mean PET counts in the sector, was calculated.
Regional variation in myocardial retention of
[11C]-HED was determined from the mean PET
counts in each of the 288 sectors (8 images times 36 sectors per image)
divided by the value found in the sector containing the maximum mean
PET counts. These normalized [11C]-HED
retention data were then displayed as polar coordinate maps of relative
tracer activity
([11C]/[13N]) from the
short-axis blood flow and the 40- to 60-minute postinjection
[11C]-HED images. The left
ventricular myocardium was depicted on the map
with the apex at the center. The mean relative tracer activity value
was determined19 within each left
ventricular region (divided into apical, anterior, septal,
inferior, and lateral myocardial segments). Apical values
were obtained by averaging together all sectors in the 2 most apical
short-axis slices. Distal values for the other segments were obtained
by averaging together the appropriate sectors in the 3 planes adjacent
to the 2 apical planes. Similarly, the corresponding proximal values
were obtained by use of the final 3 short-axis slices toward the base
of the heart. Reference values were obtained from a separate group of
normal subjects. They were processed and averaged together to determine
the homogeneity of retention in the healthy left ventricle. No regional
differences in relative [11C]-HED retention
were found in the normal subjects, which was in agreement with previous
reports19 and confirms homogeneous
uptake of [11C]-HED within the normal left
ventricle. 
i,
where qi is the relative
[11C]-HED retention value in the ith sector
value of the diabetic polar map, and µi and
i are the mean and SD of the relative
[11C]-HED retention in the ith sector of the
reference polar map. In the diabetic subjects, sectors that had a
z score >2.5 (ie, the relative
[11C]-HED retention in the patient's sector
was less than the corresponding sector mean relative
[11C]-HED retention in the reference subjects
by >2.5 SD) was defined as abnormal. Thus, the calculated z
scores represent a validated19 measure of
the individual subject's myocardial tracer retention
heterogeneity, with an increase in
heterogeneity being consistent with distal left
ventricular denervation.19 The
"extent" of the heterogeneity was expressed as the
percentage of sectors in the polar map that were abnormal, ie,
zi>2.5.
To quantify changes in regional myocardial uptake of
[11C]-HED, absolute
[11C]-HED retention in the proximal and distal
left myocardium was measured with a "retention index"
approach as previously reported19 that corrects
[11C]-HED retention for myocardial tracer
delivery. Absolute tissue retention of
[11C]-HED between 40 and 60 minutes after
injection was measured for distal and proximal anterior, septal,
inferior, and lateral myocardial segments by averaging the
appropriate sectors from the last 2 frames of the dynamic data set.
Mean tracer counts per pixel within the myocardial regions were
determined. To correct this measurement for the amount of tracer
delivered to the myocardium, retention was divided by the
total counts in the blood over the time period from injection to 60
minutes later. Total blood counts were determined from the area under
the time-activity curve for a small region of interest (5x5 pixels)
placed at the center of the left ventricular blood pool on
a basal plane. This yielded a [11C]-HED
retention index as follows: retention index (mL blood ·
min-1 · mL
tissue-1) equals tissue counts between 40 and 60
minutes divided by blood counts from time 0 to 60 minutes.
PET using [13N]NH3
allows accurate quantification of regional myocardial blood
flow,21 22 with the flow values agreeing closely
with those obtained by invasive techniques.23 The
time course of tracer distribution in myocardium and blood
was defined by dynamic image acquisition using a previously described
method.21 23 After the transmission scan, 20 mCi
of [13N]NH3 was
administered into the antecubital vein over 30 seconds. Dynamic scan
acquisition for this study was initiated with varying frame duration
(12x10 seconds/6x30 seconds/2x300 seconds). Twelve myocardial
regions per plane were defined in the 8 planes in the last time frame
of the dynamic study sequence. After correction for subject motion, the
dynamic image set was sampled, and 96 (8 planesx12 regions)
time-activity curves were stored for further analysis.
Arterial input function was determined from circular
regions at the 2 most basal planes corresponding to the large blood
pool at the center of the largest left ventricle diameter of the
resliced images. A previously validated tracer kinetic
model23 for
[13N]NH3 was used to
calculate regional myocardial blood flow in milliliters per gram per
minute. This 3-compartment model represents vascular and
extravascular [13N]NH3,
as well as metabolically trapped 13N,
which comprises glutamine.23 In this model, the
delivery and extraction of
[13N]NH3 (which is
>90%28), are used as an estimate of myocardial
blood flow.23
Statistical analysis was performed by use of Super ANOVA
(Abacus Concepts Inc). Differences among experimental groups were
detected by ANOVA, and the significance of differences between these
groups was assessed by the Student-Newman-Keuls multiple range test. If
the variances for the variables were found to differ significantly,
a logarithmic transformation was performed that corrected the unequal
variances. All analyses were then performed on the transformed
data. Data in the text are given as mean±SD. Significance was defined
at P=0.05.
Results
Top
Abstract
Introduction
Methods
Results
Discussion
References
Regional Myocardial Perfusion
The presence of regional discrete perfusion defects
consistent with occult CAD were evaluated by visual and
quantitative analyses of myocardial blood flow at rest. All
subjects had homogeneous
[13N]NH3 retention on
visual inspection of standardized color-coded blood flow images taken
from the vertical and horizontal long axes and the distal and proximal
short axes of the left ventricle. Resting regional myocardial blood
flow was similar in nondiabetic and DAN-free diabetic subjects but was
increased by 26% to 46% in the subjects with mild and severe DAN,
respectively (Table 2). Resting global
flow was higher in all myocardial segments in the DAN subjects compared
with the other subject groups, with no regional differences in flow
emerging.
View this table:
[in a new window]
Table 2. Regional Myocardial Perfusion in Different
Subject Groups
Abnormalities of regional [11C]-HED
retention in the 3 groups of diabetic subjects are shown in Figure 1. Of the 10 DAN-free subjects, 4 were
found to have regional abnormalities in
[11C]-HED retention, affecting from 4% to 8%
of the left ventricular map (5±2%). Of the 10 mild DAN
subjects, 9 were also found to have similar tracer defects affecting
from 3% to 17% of the left ventricular map (6±5%,
P=NS versus DAN-free subjects). All severe DAN subjects were
found to have extensive [11C]-HED retention
defects, ranging from 21% to 79% of the left ventricular
area (48±19%, P<0.01 versus mild DAN and DAN-free
diabetic subjects). In the mild DAN subjects, defects were observed
only in the distal inferior wall of the left ventricle,
whereas in the severe DAN subjects, defects extended to involve the
distal and proximal anterolateral and inferior walls. The
ability of the currently available reflex tests of autonomic function
to correctly classify subjects free of DAN (specificity) was 0.86, and
their ability to correctly classify those with DAN (sensitivity) was
0.67.
View larger version (10K):
[in a new window]
Figure 1. Extent of regional [11C]-HED
retention abnormalities detected in left ventricle in
DAN-free diabetic subjects and diabetic subjects with mild and severe
DAN on reflex testing. Data shown as percentage of sectors
in left ventricular (LV) polar map with relative
[11C]-HED retention >2.5 SD of reference values.
P<0.05, severe DAN versus DAN-free and mild DAN
subjects. ). The
increase in [11C]-HED retention in the proximal
segments of the severe DAN subjects appeared not to reflect increased
tracer delivery because myocardial blood flow was not significantly
different in these segments compared with the mild DAN subjects (Table 2
). Distally, [11C]-HED retention was decreased
in severe DAN by 33% compared with the DAN-free diabetic subjects
(P<0.01) (and 21% [P<0.05] compared with
mild DAN subjects) (Figure 2). The 14% reduction in absolute
[11C]-HED retention in the distal myocardial
segments of the mild DAN subjects was not statistically different from
the DAN-free diabetic subjects (P=0.1).
View larger version (44K):
[in a new window]
Figure 2. Comparison of absolute [11C]-HED
retention in proximal versus distal myocardial segments in healthy
nondiabetic subjects, DAN-free diabetic subjects, and diabetic subjects
with mild and severe DAN. *P<0.05 vs other groups;
**P<0.01 vs other groups; 
P<0.01 vs
proximal segments. LV indicates left ventricle. demonstrates the PET images
taken from a 35-year-old female subject with severe DAN who, after 33
years of diabetes, had both symptoms of DAN (gastroparesis, bladder
dysfunction, and postural hypotension) and abnormalities in all her
autonomic function tests. Blood flow images appear normal.
[11C]-HED retention, however, was markedly
heterogeneous, with only the proximal anterior cardiac
segments demonstrating visible [11C]-HED
retention. The tracer retention index was abnormally elevated in the
proximal segments (0.219±0.02 min-1) and
decreased distally (0.050±0.002 min-1) (Figure 3B).
View larger version (80K):
[in a new window]
Figure 3. HED PET images from a 35-year-old female with
severe DAN. A, Left ventricle short- and long-axis blood flow and
[11C]-HED images. Blood flow images appear normal (top),
but abnormalities of [11C]-HED retention are extensive,
with only proximal cardiac segments being visible with abnormally high
[11C]-HED retention (bottom). B, Surface plot of
z-score polar map from same subject. Calculated
z scores for each sector of the [11C]-HED
retention index polar map are plotted on the z axis of
the surface plot. z Score data are encoded in the
surface plot a second way by superimposition of a color table. Yellow
and white regions are >2.5 SD above the healthy control mean
[11C]-HED retention index. Blue regions are >2.5 SD
below the control mean. Septal and inferior walls of the
polar map are indicated along the x and y
axes, respectively. DSA indicates distal short axis; PSA, proximal
short axis; HLA, horizontal long axis; and VLA, vertical long
axis. ). Unlike HED,
epinephrine is a substrate for monoamine
oxidase,28 so its neuronal retention is largely
dependent on highly efficient vesicular
storage.28 Regional time-activity curves for
[11C]-labeled epinephrine
paralleled those generated with [11C]-HED,
suggesting that vesicular storage mechanisms are intact in the proximal
myocardial segments but defective distally.
View larger version (17K):
[in a new window]
Figure 4. Time-activity curves comparing regional washout of
[11C]-HED and [11C]-epinephrine
(EPI) in a 36-year-old male with severe DAN. Washout of
[11C]-epinephrine parallels
[11C]-HED in both proximal (base) and distal (apical)
myocardial segments. ). In contrast, in the distal myocardial
segments of the severe DAN subjects, rapid
[11C]-HED retention washout was observed over
this time period. We have previously reported that administration of
desipramine to healthy subjects decreases
[11C]-HED retention throughout the left
ventricle by 50 minutes after injection to 34% of control
values.28 Therefore, the lack of washout of
[11C]-HED in the proximal myocardial segments
of severe DAN patients paralleled that observed in healthy
myocardium, but its accelerated washout in the distal
myocardium is observed in healthy subjects only after the
administration of desipramine.
View larger version (12K):
[in a new window]
Figure 5. Comparison of [11C]-HED washout in
proximal versus distal myocardial segments of subjects with severe DAN.
Data shown as [11C]-HED tissue activity in proximal and
distal myocardial segments at 50 minutes expressed as percentage of
initial (6.25 minutes) activity. *P<0.01 vs basal and
50-minute proximal segments.
Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
References
We have used PET and [11C]-HED to directly
explore regional changes in sympathetic neuronal density and tone in
diabetic subjects with differing severity of DAN. Abnormalities of
cardiac sympathetic innervation were detected in 40% of DAN-free
diabetic subjects, with deficits beginning distally but sparing the
proximal myocardial segments even in severe DAN subjects. In severe
DAN, absolute [11C]-HED retention was found to
be paradoxically increased in the proximal myocardial segments, while
distally [11C]-HED retention was decreased. No
appreciable washout of the tracer was observed in the proximal
segments, consistent with increased sympathetic innervation but
normal regional tone. In the distal myocardium, increased
[11C]-HED washout in severe DAN reproduced that
observed in the myocardium of desipramine-treated healthy
subjects,28 suggesting distal neuronal uptake-1
dysfunction or neuronal loss. These data suggest that diabetes may
result in left ventricular sympathetic dysinnervation with
proximal hyperinnervation accompanying distal denervation, which in
combination could result in potentially life-threatening myocardial
electrical instability.
Acknowledgments
This work was supported in part by grant RO1-HL-47543 from the
National Heart, Lung, and Blood Institute, National Institutes of
Health, Bethesda, Md. Patients were recruited and characterized in the
Clinical Implementation Core of the Michigan Diabetes Research and
Training Center. We express our gratitude to the University of Michigan
Medical Cyclotron and Radiochemistry Unit for the production of
the [11C]-HED and to the PET suite
technologists for performing the scintigraphic studies. We would like
to thank Dr D.A. Greene for his editorial comments.
Footnotes
Reprint requests to Dr Martin J. Stevens, Division of Endocrinology and Metabolism, Department of Internal Medicine, University of Michigan, 5570 MSRB II, Box 0678, 1150 W Medical Center Dr, Ann Arbor, MI 48109-0678.
References
Top
Abstract
Introduction
Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  A. I. Vinik and D. Ziegler Diabetic Cardiovascular Autonomic Neuropathy Circulation, January 23, 2007; 115(3): 387 - 397. [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Raffel, R. A. Koeppe, R. Little, C.-N. Wang, S. Liu, L. Junck, M. Heumann, and S. Gilman PET Measurement of Cardiac and Nigrostriatal Denervation in Parkinsonian Syndromes J. Nucl. Med., November 1, 2006; 47(11): 1769 - 1777. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Bengel, P. Ueberfuhr, D. Schafer, S. G. Nekolla, B. Reichart, and M. Schwaiger Effect of Diabetes Mellitus on Sympathetic Neuronal Regeneration Studied in the Model of Transplant Reinnervation J. Nucl. Med., September 1, 2006; 47(9): 1413 - 1419. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Raffel, W. Chen, P. S. Sherman, D. L. Gildersleeve, and Y.-W. Jung Dependence of Cardiac 11C-meta-Hydroxyephedrine Retention on Norepinephrine Transporter Density J. Nucl. Med., September 1, 2006; 47(9): 1490 - 1496. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M.-R. Movahed, M. Hashemzadeh, and M. M. Jamal Increased Prevalence of Third-Degree Atrioventricular Block in Patients With Type II Diabetes Mellitus Chest, October 1, 2005; 128(4): 2611 - 2614. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G A Suarez, V M Clark, J E Norell, T E Kottke, M J Callahan, P C O'Brien, P A Low, and P J Dyck Sudden cardiac death in diabetes mellitus: risk factors in the Rochester diabetic neuropathy study J. Neurol. Neurosurg. Psychiatry, February 1, 2005; 76(2): 240 - 245. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Pop-Busui, I. Kirkwood, H. Schmid, V. Marinescu, J. Schroeder, D. Larkin, E. Yamada, D. M. Raffel, and M. J. Stevens Sympathetic dysfunction in type 1 diabetes: Association with impaired myocardial blood flow reserve and diastolic dysfunction J. Am. Coll. Cardiol., December 21, 2004; 44(12): 2368 - 2374. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. S. Lim, R. J. MacFadyen, and G. Y. H. Lip Diabetes Mellitus, the Renin-Angiotensin-Aldosterone System, and the Heart Arch Intern Med, September 13, 2004; 164(16): 1737 - 1748. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Y. Fang, J. B. Prins, and T. H. Marwick Diabetic Cardiomyopathy: Evidence, Mechanisms, and Therapeutic Implications Endocr. Rev., August 1, 2004; 25(4): 543 - 567. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Scott and P. L. Kench Cardiac Autonomic Neuropathy in the Diabetic Patient: Does 123I-MIBG Imaging Have a Role to Play in Early Diagnosis? J. Nucl. Med. Technol., June 1, 2004; 32(2): 66 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H Laine, J Sundell, P Nuutila, O T Raitakari, M Luotolahti, T Ronnemaa, T Elomaa, P Koskinen, and J Knuuti Insulin induced increase in coronary flow reserve is abolished by dexamethasone in young men with uncomplicated type 1 diabetes Heart, March 1, 2004; 90(3): 270 - 276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F. Beek, J. A. P. van der Sloot, M. Huikeshoven, H. J. Verberne, B. L. F. van Eck-Smit, J. van der Meulen, J. G. P. Tijssen, M. J. C. van Gemert, and R. Tukkie Cardiac denervation after clinical transmyocardial laser revascularization: Short-term and long-term iodine 123-labeled meta-iodobenzylguanide scintigraphic evidence J. Thorac. Cardiovasc. Surg., February 1, 2004; 127(2): 517 - 524. [Abstract] [Full Text] [PDF] |