Assessment of the Effects of Age, Gender, and Exercise Training on the Cardiac Sympathetic Nervous System Using Positron Emission Tomography Imaging

Abstract

Background:

Using positron emission tomography (PET) imaging, we sought to determine whether normal age or exercise training cause changes in the cardiac sympathetic nervous system function in male or female healthy volunteers.

Methods:

Healthy sedentary participants underwent PET studies before and after 6 months of supervised exercise training. Presynaptic uptake by the norepinephrine transporter-1 function was measured using PET imaging of [¹¹C]-meta-hydroxyephedrine, a norepinephrine analog, and expressed as a permeability–surface area product (PS_nt in mL/min/mL). Postsynaptic function was measured as β-adrenergic receptor density (β′_max in pmol/mL tissue) by imaging the β-receptor antagonist [¹¹C]-CGP12177. Myocardial blood flow (MBF in mL/min/mL tissue) was measured by imaging [¹⁵O]-water.

Results:

At baseline, there was no age difference in β′_max or MBF but PS_nt declined with age (1.12±0.11 young vs 0.87±0.06 old, p = .036). Before training, women had significantly greater MBF (0.87±0.03 vs 0.69±0.03, p < .0001) and PS_nt (1.14±0.08 vs 0.75±0.07, p < .001) than men. Training increased VO₂ max by 13% (p < .0001), but there were no training effects on β′_max, PS_nt, or MBF. Greater MBF in females and a trend to increased PS_nt post-training persisted.

Conclusion:

With age, presynaptic uptake as measured by PS_nt declines, but there were no differences in β′_max. Endurance training significantly increased VO₂ max but did not cause any changes in the measures of cardiac sympathetic nervous system function. These findings suggest that significant changes do not occur or that current PET imaging methods may be inadequate to measure small serial differences in a highly reproducible manner.

One of the hallmarks of normal cardiovascular age among men is a reduced response to β-adrenergic stimulation; this is manifested as decreased heart rate, ejection fraction, and cardiac output responses to the β-adrenergic stimulant isoproterenol and to exercise stress (1–7). By comparison, when old and young men exercise in the presence of β-adrenergic blockade, the age-related differences in myocardial contractility, heart rate, and cardiac output are blunted; this further suggests that the age-associated declines during exercise are a manifestation of reduced β-adrenergic responsiveness (8,9). Endurance exercise training reverses, in part, the age-associated changes in cardiovascular function and improves responsiveness to β agonists in young men, but whether similar favorable changes occur in others is largely unknown (10–13). No in vivo human studies of the effect of exercise training on the cardiac β-adrenergic system exist.

The reduced β-adrenergic responsiveness of age is a major cause of the decline in cardiovascular responses to stress with age and may increase susceptibility to age-associated cardiovascular disease (8,9,14). Sympathetic function is abnormal in congestive heart failure patients, who demonstrate increased sympathetic nerve activity, norepinephrine (NE) plasma levels, cardiac NE spillover, and depleted cardiac NE stores (14). The basis or mechanism of the reduced β agonist responsiveness with age has not been well elucidated, but changes in the β-effector system could be secondary to altered presynaptic or postsynaptic cardiac sympathetic function (14).

Studies of animal cardiac β-adrenergic receptors and human peripheral lymphocyte β-adrenergic receptors do not reflect changes in human cardiac β-adrenergic activity (15–20). Many animal studies of myocardium were conducted in rats, animals with myocardium primarily comprise α-1 receptors (21–32). In contrast, human myocardium predominantly comprises β-1 receptors. Comparisons of studies of human peripheral lymphocytes to human myocardium are also limited in that human peripheral lymphocytes do not contain β-1 receptors nor are they innervated (33–35). Thus, prior studies in both animals and human lymphocytes have limited applicability to the human heart. In addition, conclusions from prior studies of human cardiac tissue have been limited by the subpopulation examined (14,20,36,37). Furthermore, results from these studies of diseased and in vitro transplant myocardium are conflicting, highlighting the need for study of healthy individuals’ in vivo myocardium.

Finally, gender differences in resting cardiovascular function exist with age. One study showed women to have a decrease in resting cardiac output and increased systemic vascular resistance with age; this change was not observed in men (38). Gender differences in autonomic nervous system regulation of the heart and circulation at rest may exist and at least partly explain these differences in cardiovascular function (39). Whether gender differences in autonomic function occur with age, exercise, or exercise training has not been well studied.

Hypothesis

We hypothesized that cardiac β-receptor downregulation and presynaptic upregulation occur with age in vivo. We further hypothesized that there are no gender differences in β-adrenergic reception density (β′_max) and NE transport-1 function. Finally, we hypothesized that endurance training improves cardiac β-adrenergic responsiveness by alterations in the cardiac sympathetic nervous system (SNS) measured by positron emission tomography (PET).

Materials and Methods

Two volunteer populations, healthy young (18–33 years) and healthy older (65–80 years) individuals were studied. These healthy men and women were screened to exclude (i) significant or exercise limiting cardiovascular disease including myocardial infarction, angina, claudication, significant valvular heart disease, left ventricular hypertrophy, stroke, or transient ischemic attacks; (ii) weight >150% ideal body weight; (iii) significant or exercise limiting pulmonary or orthopedic disease (including severe osteopenia); (iv) active or limiting arthritis; (v) diabetes mellitus (fasting plasma glucose > 140mg/dL); (vi) significant hypertension (>140/90, untreated); (vii) psychiatric disease or dementia that would limit participation or cooperation; (viii) excessive alcohol intake (greater than two drinks/day); (ix) tobacco use; (x) abnormal liver function or other screening tests; (xi) regular exercise in the last 12 months (greater than two times per week, 20 minutes per session); (xii) pregnancy; (xii) abnormal resting electrocardiogram; and (xiv) chronic use of any prescription medication except for treated hypothyroidism and estrogen replacement therapy, which all older females received.

All had unremarkable blood counts, urinalysis, and chemistries, including cholesterol and thyroid-stimulating hormone, as well as normal two-dimensional and Doppler echocardiograms for age. The old group also had normal maximal postexercise SPECT ^99mTc-sestamibi images. Twenty-four young volunteers, 10 males (mean age ± SD, 24.9±3.6 years) and 14 females (mean age ± SD, 26.1±4.0 years) participated. Thirty-three older volunteers, 14 males (mean age ± SD, 70.3±4.45 years) and 19 females (mean age ± SD, 70.7±4.5 years) were studied. All of our participants were Caucasian. All older females were receiving hormone replacement therapy. This study was approved by the University of Washington Human Subjects, Radioactive Drug Research and Radiation Safety Committees. All participants gave written informed consent.

Study Protocol

Radiotracer synthesis

¹⁵O-water (40), [¹¹C]-meta-hydroxyephedrine (¹¹C-mHED) (41), and the β-adrenergic receptor (BAR) antagonist (S)-(–)-¹¹C-CGP12177 ((S)-4-(3′-t-butylamino-2′-hydroxypropoxy)-benzimidazol-2-¹¹C-one), ¹¹C-CGP12177 (¹¹C-CGP) (42–44) were synthesized according to published methods (45,46). Metabolites of ¹¹C-mHED were measured as reported (47). (S)-(–)-CGP12177 is not metabolized in humans (48). The amount of nonradioactive material in the ¹¹C-CGP and the ¹¹C-mHED injectates was measured using high-performance liquid chromatography with mass spectrometry (ES+) detection (Waters 2190 and Micromass ZMD) (49).

Imaging protocol

Heart rate and blood pressure were monitored continuously and recorded each minute from 5 minutes before to 15 minutes after each radiotracer injection. The participants were positioned in the tomograph (Advance; GE Healthcare) using 1-minute transmission scans to localize the heart (35 planes, 15cm). Twenty-minute transmission scans were acquired using a rotating ⁶⁸Ge rod source. The imaging protocol consisted of sequential injections of high-specific-activity ¹¹C-CGP followed approximately 20 minutes later by injection of low-specific-activity ¹¹C-CGP as described previously (46) followed by ¹⁵O-water and ¹¹C-mHED. Activity (mCi) of each injection was measured using a calibrated ion chamber.

For the first CGP injection, dynamic PET images (60s × 1, 5s × 2, 10s × 6, 15s × 6, 30s × 4, 60s × 15) were acquired for 21 minutes, starting 1 minute before the first CGP injection (CGP1). Low-specific-activity CGP (CGP2) was injected 25 minutes after the CGP1 injection. The CGP1 dynamic image sequence was repeated and continued for 45 minutes by adding 5-minute frames. Myocardial blood flow (MBF) was measured with an intravenous injection of ¹⁵O-water. Dynamic PET images were acquired for 5 minutes. After 15 minutes for ¹⁵O decay, ¹¹C-mHED was infused with the dynamic PET image acquisition sequence as for CGP2. The ¹¹C-CGP and ¹¹C-mHED were injected over 1 minute. Participants remained in the tomograph for the entire imaging protocol. Heparinized plasma samples at ~5, 10, 20, 30, and 40 minutes after ¹¹C-mHED injection were analyzed for ¹¹C-mHED and metabolites (47).

The dynamic PET image sets were reconstructed, reoriented into short-axis cardiac projections, and analyzed. Images were decay corrected to the time of each radiotracer injection, except for the CGP2 images, which were corrected to the start of the CGP1 injection. Myocardial and left atrial (LA) regions of interest (ROIs) were placed on static short-axis images from each set. These ROIs were applied to the dynamic images and time–activity curves generated for 3 middle LA planes and 96 myocardial ROIs per heart (12 slices with 8 sectors per slice). Three LA planes were averaged to provide a single LA time–activity curve for each image set; this was used as the input function for the respective model analysis described below. The three most apical left ventricle planes were excluded from analysis to avoid any partial-volume effect from this region. After quantitative analysis of the remaining individual ROIs, the slice data were averaged to give three cross-section slices labeled apical, mid, and basal. Then, two adjacent sector’s ROIs were averaged to give four regions per slice: anterior, lateral, inferior, and septal. This resulted in 12 left ventricle ROIs and a global ROI (average of the 12) per participant for each of the 4 radiopharmaceutical injections.

The ¹⁵O-water time–activity curves (cpm/pixel) from the LA and from each of the individual left ventricle ROIs were modeled to obtain MBF (50). The ¹¹C-CGP time–activity curve data were converted to pmol/mCi at the time of first injection based on the specific activity of the injectate.

Supervised training protocol

All participants participated in walking, jogging, and/or bicycling sessions supervised by an exercise physiologist 3 days per week for 6 months. Sessions began with 10–15 minutes of stretching followed by a 5-minute warm-up of slow walking, jogging, or bicycling; this was followed by a 30- to 45-minute period of training at the specified training intensity. The ratio of walk/jog to bicycling was 2:1. A 10-minute cooling period that included more stretching ended each session.

The target heart rate for each participant was a specified percentage of their heart rate reserve based on the baseline maximal test. Participants began training at 50%–60% of their heart rate reserve, calculated according to the following formula: initial training heart rate reserve = resting heart rate + 0.5 (maximal heart rate − resting heart rate). Training intensity was increased over a 10-week period as tolerated to 80%–85% of heart rate reserve. This schedule was created according to the guidelines of the American College of Sports Medicine. Heart rates were monitored with rate alarms individually set for each participant.

Maximal oxygen consumption was measured using a maximal treadmill exercise test. VO₂ max was obtained by standard procedures as described previously (51). The test ended when the participant could no longer continue walking or running. Mean expiratory respiratory exchange ratio on the pretest and post-test indicated good effort. Post-training imaging was performed on all participants 36–48 hours after the last bout of exercise to reduce the effects of acute exercise.

Data Analysis

¹¹C-mHED has been shown to have a high affinity for the reuptake one mechanism that recycles ~70% of the NE released from the sympathetic nerve terminals (52). ¹¹C-mHED metabolites, expressed as the fraction of the whole-blood activity, contribute to the blood PET signal (53). These fractions were curve fit to a rising time-dependent exponential function of the form, f(t) = y ⁰ + (a*[1 − exp(–b*t)]) using Sigma Plot (SYSTAT Software). Using these derived values, a metabolite-corrected LA cavity time–activity curve was generated by multiplying the LA cavity activity by the function (1 − f[t]).

¹¹C-mHED kinetics were modeled by blood tissue exchange using the MBF value from the ¹⁵O-water study to provide the flow parameter in the model (45,49). This model expresses ¹¹C-mHED kinetics in terms of the permeability–surface area product (PS_nt, mL/min/mL tissue) for ¹¹C-mHED uptake into the nerve terminal from the interstitial space through the NE transport-1 process and PS_ves for release of ¹¹C-mHED back into the interstitial space through exocytosis. The neuronal storage of ¹¹C-mHED is expressed as a virtual volume (V′_nt, mL/mL tissue) and the rate of vesicular storage of mHED is expressed as G _seq (mL/min/mL). Transport rates and volumes are expressed as milliliter of tissue without conversion for any tissue density. In the context of a three-compartment mode, our parameters PS_nt, V′_nt, and G_seq, and PS_ves, represent k₃, the combined cellular and vesicular reuptake processes. The parameter PS_nt most closely represents ¹¹C-mHED uptake into the nerve by NE transport-1 and is the process most likely to be affected by myocardial ischemia or injury. BAR density (B′_max) was estimated for each myocardial ROI using the Delforge graphical method (46,54).

Statistics

Hemodynamic measurements, MBF, and SNS measures, B′_max and PS_nt, were compared between young and old participants as well as male and female cohorts using a two-tailed, unpaired Student’s t test. Paired sample t tests were used to compare pretraining and post-training measurements. Analysis of variance was used for repeated measurements. Statistical analyses used SPSS.

Results

Baseline Characteristics

Average heart rate (± SD) pretraining in all participants was 68±8.8 and post-training significantly dropped to 64±8.7. Average systolic blood pressure pretraining in all participants was 120±15 and post-training dropped to 117±17.

Average peak VO₂ (± SD) pretraining in all participants was 29±8.3 and post-training increased significantly to 34±9.7.

Average weight (± SD) pretraining in all groups was 169±15kg and post-training was 165±15kg. Average baseline body mass index (± SD) in all participants was 25±2.7 and post-training was 25±4.9 (see Table 1).

Table 1.

Baseline and Post-training HR, SBP, Peak VO₂, and BMI

	All Participants	Young	Old	Female	Male
Pre-HR	68±8.8	69±9.4	67±8.6	70±6.3	66±10
Post-HR	64±8.7	63±8.3	64±9.2	66±6.6	61±9.7
p Value	<.0001	.002	.005	.01	<.0001
Pre-SBP	120±15	113±12	125±11	117±14	122±11
Post-SBP	117±17	110±14	123±14	113±15	120±14
p Value	.45	.74	.49	.72	.49
Prepeak VO2	29±8.3	36±7	24±4.8	25±5.2	33±8.5
Postpeak VO2	34±9.7	42±8	28±5.6	29±6.8	38±10
p Value	<.0001	<.0001	.0001	<.0001	<.0001
Pre-BMI	25±2.7	25±2.7	25±2.1	25±2.3	25±3.2
Post-BMI	25±4.9	23±4.5	27±4.5	23±3.8	27±5.1
p Value	.86	.0006	.004	.002	.02

Note: BMI = body mass index; HR = heart rate; SBP = systolic blood pressure.

Effects of Age

At baseline, there was a nonsignificant trend for a higher mean β-adrenergic receptor density (β′_max density) in the young (15.1±5.5 vs old 13.1±4.3, p = .20; see Tables 2 and 5). NE reuptake (PS_nt) declined significantly with age (1.12±0.10 vs 0.87±0.32, p = .036; see Tables 3 and 5). There was no significant difference in mean MBF with age (0.76±0.03 young vs 0.81±0.14 older, p = .24; see Tables 4 and 5).

Table 2.

Global β′_max Density* for Participants Before and After Exercise Training

Training Status	N	Mean β′max*	SD
Pretraining	53	13.9	4.9
Young adults	22	15.1	5.5
Older adults	31	13.1	4.3
Men	21	14.3	4.5
Women	32	13.6	5.2
Post-training	36	13.2	5.3
Young adults	14	13.1	6.4
Older adults	22	13.3	4.7
Men	15	11.7	5.4
Women	21	14.3	5.0

Note: β′_max = β-adrenergic receptor density.

*Postsynaptic function was measured as β′_max (in pmol/mL tissue) by imaging the β receptor antagonist [¹¹C]-CGP12177.

Table 5.

Summary by Group Including Mean With Standard Deviations

	Young	Old	p	Male	Female	p	Overall
Pretraining PSnt	1.12±1.0	0.87±0.32	.036	0.75±0.57	1.14±0.83	<.001	0.98±0.74
Post-training PSnt	1.09±0.64	0.82±1.8	.127	0.76±0.63	1.04±1.9	.085	0.91±1.5
Pretraining MBF	0.76±0.03	0.81±0.14	.24	0.69±0.12	0.86±0.14	<.0001	0.79±0.14
Post-training MBF	0.75±0.19	0.78±0.19	.63	0.67±0.14	0.85±0.16	<.001	0.77±0.17
Pretraining βʹ max	15.1±5.5	13.1±4.3	.20	14.3±4.5	13.6±5.2	.87	13.9±4.9
Post-training βʹ max	13.1±6.4	13.3±4.7	.90	11.7±5.4	14.3±5.0	.20	13.2±5.3

Note: MBF = myocardial blood flow. At baseline, the only significant differences noted were a lower PS_nt with age (p = .036) and a gender difference in MBF and PS_nt, where men had significantly lower values than women (p < .0001 and p < .001, respectively). The trend for a lower PS_nt with age persisted after training (p = .127) as did gender differences in MBF and PS_nt (p < .001 and p = .085, respectively).

Table 3.

PS_nt* for Participants Before and After Exercise Training

Training Status	N	Mean PSnt*	SD
Pretraining	56	0.98	0.7
Young adults	24	1.12	1.0
Older adults	32	0.87	0.32
Men	24	0.75	0.57
Women	31	1.14	0.83
Post-training	40	0.91	1.5
Young adults	14	1.09	0.64
Older adults	26	0.82	1.8
Men	18	0.76	0.63
Women	22	1.04	1.9

Note: PET = positron emission tomography.

*Presynaptic uptake by the norepinephrine transporter-1 function was measured using PET imaging of [¹¹C]-meta-hydroxyephedrine, a norepinephrine analog, and expressed as a permeability–surface area product (PS_nt in mL/min/mL).

Table 4.

MBF* for Participants Before and After Exercise Training

Training Status	N	Mean MBF*	SD
Pretraining	57	0.79	0.14
Young adults	24	0.76	0.14
Older adults	33	0.81	0.14
Men	24	0.69	0.12
Women	33	0.86	0.14
Post-training	41	0.77	0.17
Young adults	15	0.75	0.19
Older adults	26	0.78	0.19
Men	19	0.67	0.14
Women	22	0.85	0.16

Note: MBF = myocardial blood flow.

*MBF (in mL/min/mL tissue) was measured by imaging [¹⁵O]-water.

Effects of Gender

At baseline, there was no significant difference in β′_max density by gender (men 14.3±4.5 vs women 13.6±5.2 where p = .87; see Tables 2 and 5). PS_nt was significantly lower among men at baseline (men 0.75±0.57 vs women 1.14±0.83 where p < .001; see Tables 3 and 5). At baseline, MBF was significantly lower among men (men 0.69±0.12 vs women 0.86±0.14 where p < .0001; see Tables 4 and 5).

Training Effects

Overall, exercise training increased VO₂ uptake by 13% (p < .0001), reduced resting heart rate by 5% (p < .01), and caused a small decline in body weight (170±25 to 168±25 lbs, p = .03). In the overall group, there was no effect of exercise training on mean β′_max density, PS_nt, or MBF (see Tables 1 and 5). After training, there remained no significant difference between men and women’s mean β′_max density.

The trend for a lower PS_nt with age persisted after training (1.09±0.64 vs 0.82±1.8 where p = .127) as did the trend for a higher PS_nt in women than men (1.04±1.9 vs 0.76±0.63 where p = .085). The difference in MBF by gender persisted after training (men 0.67±0.14 vs women 0.85±0.16 where p < .001).

Discussion

This study investigated the effects of age, gender, and exercise training on the cardiac SNS.

The main findings of this study are (i) evidence of a decrease in presynaptic function with age whereas postsynaptic function and MBF were unchanged with age and (ii) in women evidence of increased presynaptic function and increased MBF compared with men; both of these gender differences persisted after training. Postsynaptic function was not affected by age or gender. Exercise training did not affect cardiac sympathetic function or MBF.

Age

There was no statistically significant difference in β′_max density with age. There are only very limited in vivo data about β′_max density in normal participants. One prior study found young mean β′_max density to be 11.15±2.2 (55), somewhat lower than our values; there are no prior studies in older healthy adults. Prior studies have shown a greater response to β-adrenergic stimulation in healthy young versus old men (1–6,56–58); our results suggest that differences in beta receptor density do not account for these differences in responsiveness to beta adrenergic stimulation.

No prior studies have examined NE reuptake (PS_nt) with age; however, one study showed increased cardiac NE spillover with reduced extraction of ³H NE across the heart with age (59). Our results would be consistent if increased cardiac NE spillover were due to a decrease in reuptake.

Mean MBF did not change with age in our group. In a study done by Czernin and coworkers, a small but significant increase in MBF with age was noted (60). In contrast, Hachamovitch and coworkers reported a reduction in MBF as well as decreased contractile performance (61). Czernin and coworkers assumed homogenous myocardial wall thickness of 1cm; if thickness had been increased by 1mm in their oldest participants, MBF in his study would have been overestimated by 6%.

Gender

There was no significant difference in β′_max density by gender at baseline or following training. Both PS_nt and MBF were significantly lower among men at baseline. The resting heart rate was 4–5 bpm higher in the women than men before and after training, whereas the systolic pressure was 5–7 mmHg lower in the women than men, so these two determinants of MBF by themselves are unlikely to explain the higher MBF in women. A study in 2007 showed no MBF difference by gender among young participants (62), whereas another study found that older women had a significantly higher MBF than men, similar to our findings (63). Older women account for the difference in MBF by gender (data not shown).

Training

Supervised exercise training led to a significant training effect as shown by the 13% increase in peak VO₂ (p < .0001) and reduction in resting heart rate. Taken together, these findings indicated that participants were adequately trained during the study period. Despite the evidence of a significant training effect, we found no difference in β′_max density or in mean PS_nt as a result of exercise training. There was also no training induced difference in MBF. A prior study by our research group showed no significant change in isoproterenol response following exercise training in men, which would be consistent with no change in receptor number. Other studies demonstrated that exercise reduces level of G protein–coupled receptor kinase-2 that is known to be involved in beta receptor dysregulation (57). The trend for a reduction in PS_nt with age persisted after training, as did the PS_nt difference by gender.

Overall, training did not affect MBF. This result is consistent with one prior study that examined the effect of training on MBF in older adults (63). A difference in MBF by gender persisted after training.

β′_max Density

An in vitro study of donor hearts showed a profound decrease in cardiac β-adrenergic responsiveness with age occurring by multiple mechanisms including downregulation and decreased agonist binding of β₁ receptors, uncoupling of β₂ receptors, and abnormal G protein–mediated signal transduction (64). Phosphorylation of the β-adrenergic receptor by members of the G protein–coupled receptor kinase family could occur. Alternatively, inhibitory Gi proteins or post-cAMP defects could affect β-adrenergic responsiveness with age or exercise training. This is in contrast with our findings which did now show any changes in postsynaptic function with age suggesting that imaging techniques are not yet capable of capturing these changes. Conversely, although the donor hearts were considered normal, the events preceding death likely led to alterations in the SNS.

PS_nt

A significant decline in NE reuptake (PS_nt) was observed with age. These results are consistent with another study demonstrating impaired NE reuptake with age (65). Our study showed that the decline persisted after training. PS_nt was significantly lower among men; this result persisted after training.

Limitations

Our sample size was small, and our findings could be subject to beta statistical error. As demonstrated in the tables, there were some missing data that had led to varying sample sizes in different groups. In addition, our study examining age differences was cross-sectional. Individual changes may occur across the lifespan, but a longitudinal study was not feasible. Reproducibility of PET measurements of cardiac sympathetic function was not studied in this investigation. Finally, we did not include a group without exercise training due to additional expense. Prior studies have examined the impact of exercise training relative to remaining sedentary on hemodynamic responses including VO₂ max, and have found no change in these parameters with continued sedentary lifestyle (66,67).

Finally, all postmenopausal women were receiving hormone replacement therapy at the time of the study. We do not believe this affected the results of our study; however, we do not have a control group that was on no replacement therapy.

Conclusion

The only significant differences noted were a decline in PS_nt with age and a gender difference in MBF and PS_nt, where men had significantly lower values than women. The trend for a reduction in PS_nt with age persisted after training as did gender differences in MBF and PS_nt. With these exceptions, little difference in cardiac SNS function was found with age or as a result of endurance training, suggesting that such changes do not occur or that current PET imaging methods may be inadequate to measure small serial differences in a highly reproducible manner.

Funding

This research was supported by grants NIH HL50239 and AG 15462 by VA Research and Development.

Conflict of Interest

None of the authors have any known conflicts of interest.