Introduction

Recent data indicate 31.5% of U.S. children are overweight or obese (1). The adverse health outcomes due to the increased prevalence of childhood obesity have not been fully realized; however, it has been reported that increased adiposity during childhood substantially increases the risk for obesity, type 2 diabetes, and cardiovascular disease as an adult (2). One consequence of childhood obesity that has not been extensively studied is the effect of increased adiposity on cardiovascular autonomic (cANS)¹ function. Previous studies using heart rate variability (HRV) methods to measure cANS function have indicated that obese children have decreased parasympathetic nervous system (PSNS) function as compared with normal-weight children (3)(4)(5)(6)(7). However, these studies have not been in agreement as to whether childhood obesity is characterized by increased or decreased sympathovagal balance.

Leptin is an adipocyte-derived protein hormone that carries out its primary task of regulating food intake and energy homeostasis [via increased sympathetic nervous system (SNS) outflow] by binding to specific leptin receptors in the hypothalamus (8)(9). Numerous isoforms of the leptin receptor exist in humans, but the soluble form of leptin receptor (sOB-R) has been most studied and has been shown to be in low concentration in obese patients (10). Leptin resistance is a physiological state observed in human obesity that is characterized by high leptin and low sOB-R concentrations. Recent findings have reported significant positive associations with leptin and insulin resistance and blood pressure in overweight children (11)(12). Indeed, leptin has been postulated to increase sympathetic nervous system activity by increasing circulating norepinephrine concentrations as evidenced in animal models (9)(13). A recent study in children indicated that leptin resistance is decreased when the level of adiposity is decreased and this reduction is associated with improved metabolic (decreased insulin) and cardiovascular function (decreased blood pressure) (11). To date, no studies have compared the relationships between cANS function and leptin resistance in normal-weight and overweight children.

One of the primary metabolic alterations with increasing and prolonged obesity is a progressive increase in insulin resistance. This metabolic maladaptation has been shown to be associated with numerous metabolic and cardiovascular abnormalities (14). Hyperinsulinemia has been proposed as one of the potential mechanisms by which differences in cANS function is shown between normal and overweight children. It has been suggested that elevated concentrations of insulin might continually activate the sympathetic nervous system to increase metabolism in an attempt to counteract the positive energy balance from overeating (15).

Obesity has been linked to a state of subclinical inflammation and increased oxidative stress as previous studies in obese adults and children have found increased levels of interleukin-6 (IL-6) (16), tumor necrosis factor- (TNF-) (17), C-reactive protein (CRP) (18), and 8-isoprostane (19). Recent data have indicated that lower HRV is associated with higher CRP in older individuals (i.e., age >46 years) (20), but this relationship was not fully explained since it was not known whether some other mechanism associated with increased age confounded the relationship. Regardless, it is well known that IL-6 and TNF- stimulate the release of glucocorticoids and adrenalin (21) from the hypothalamic-pituitary-adrenal axis in states of inflammation, which may in turn increase SNS activity potentiating the cANS dysfunction associated with obesity. Thus, it is physiologically plausible to propose potential relationships between cANS function and inflammation and oxidative stress in childhood obesity. However, this relationship has not been previously investigated.

We, therefore, hypothesized that overweight and obese children, as compared with normal-weight controls, would display cANS dysfunction that is characterized by suppressed PSNS and increased SNS modulation of cardiac function and that these differences would be associated with leptin resistance, insulin resistance, oxidative stress, and inflammation.

Research Methods and Procedures

Participant Population

A total of 36 children (18 male, 18 female; age, 11.5 0.8 years; Tanner stage, 1.9 0.8) were recruited from the greater Minneapolis metro area. All studies took place at the University of Minnesota General Clinical Research Center. For this study, overweight (OW) and obese (OB) groups were defined by an age-adjusted BMI greater than the 85th percentile but less than the 95th percentile (OW) and greater than the 95th percentile (OB), respectively (22). Despite the terminology used by the Centers for Disease Control and Prevention for increased adiposity in children (at-risk-for-overweight = 85th percentile < BMI < 95th percentile; overweight = BMI >95th percentile), we choose to use the terminology of overweight and obese to emphasize the importance of the adverse physiological profiles of these categories of children. All participants were healthy OW, OB, or normal-weight (NW; BMI <85th percentile) children between the ages of 10 and 13 years. Children who were affected by disease states that effect cardiovascular function and/or cardiac electrophysiology, which included orthostatic intolerance and unexplained syncopal episodes, were not allowed to participate in this study. It was determined whether a child met any of the above-mentioned exclusion criteria by a medical history questionnaire after the informed consent process and by performing a resting electrocardiogram before testing.

Experimental Protocol

Written informed consent from the parents/guardians and assent from the participants were obtained after all components of a standard informed consent including purpose, risks, and benefits were fully explained to each child and his or her parent(s). All methods used in the study were reviewed and approved by the University of Minnesota Institutional Review Board. Medical history was ascertained via questionnaire. All participants were studied after a 10-hour overnight fast. Fasting blood samples were then obtained. Height and weight was measured using a wall-mounted stadiometer and digital weight scale (Model 5002; Scale-Tronix, Inc., Wheaton, IL). BMI (kg/m²) was calculated as the body weight (kg) divided by height squared (m²). A DXA scan was then performed for the determination of total body fat percent fat mass, and trunk fat percent (software version 6.7; Prodigy 3M, Madison, WI). Assessment of pubertal development (Tanner score) was determined through a brief physical examination by a trained pediatrician. Participants were then prepped for electrode placement for measurement of heart rate via a 3-lead electrocardiograph (ECG). The ECG (Lead II) and blood pressure (BP) were continuously recorded using an automated blood pressure cuff and automated tonometer (Colin Pilot 7000; Colin Medical Instruments Corp., San Antonio, TX). No constraints were imposed on breathing frequency due to findings in pilot studies in our laboratory that indicated that children find it difficult to pace their breathing to a set cadence. Participants lay flat on their backs on a cushioned bed for 15 minutes to ensure that a resting state was attained. After the initial rest period, the participant continued to lie relaxed for an additional 15 minutes to record resting ECG and BP measures.

HRV Analyses

The ECG and BP waveforms were digitally recorded continuously using a desktop computer and WinDaq Pro data collection software (DATAQ Instruments Inc., Akron, OH). Each signal was sampled at 500 Hz throughout all testing. The WinDaq Pro software allowed for instantaneous analog to digital conversion of the ECG and BP signal with recordings stored for latter off-line analysis. Files were imported into a software program (Matlab; The MathWorks, Inc., Natick, MA) for computation of standard time- [standard deviation of normal inter-beat (R-R) intervals (SDRR); square root of the mean squared differences of successive R-R intervals (RMSSD)] and frequency-domain HRV variables based on current recommendations (23). The last 5 minutes of the 15 minutes resting period was utilized for calculation of all resting HRV variables. Each 5-minute segment was manually reviewed for ectopic beats or arrhythmias and segments containing such alterations of normal electrophysiological function were excluded from analysis.

Power spectral density of the R-R time series was calculated using non-parametric methods (fast-Fourier transform) after being passed through a Hamming window. Three frequency bands [very-low frequency (VLF) (0 to 0.03 Hz); low frequency (LF) (0.04 to 0.15 Hz), which, when expressed in normalized units, is indicative of primarily SNS modulation with some influence from PSNS modulation; and high frequency (HF) (0.15 to 0.4 Hz), indicative of purely PSNS modulation of cardiac function] are obtained from short-term HRV recordings, but only LF and HF were considered in our analyses because the VLF component is considered a dubious measure due to its lack of physiological meaning and interpretation. Therefore, total power within the power spectrum (0.04 to 0.40 Hz) and the normalized units of the LF and HF components [normalized unit = (LF or HF)/(total power - VLF)] were reported. In addition, the LF:HF ratio was calculated as a measure of relative sympathovagal balance (23). Normalized units and LF:HF allow for an accurate assessment of the balanced modulations of each branch of the autonomic nervous system while not being influenced by the total power within the power spectrum (5). Reproducibility data from our laboratory have shown a mean difference of 0.01 0.02 seconds for SDRR (coefficient of variation = 20.6%), 1.64 4.0 seconds for RMSSD (coefficient of variation = 4.1%), 0.05 0.1 for the LF:HF ratio (coefficient of variation = 5.7%), 1.28 2.4 for LF normalized units (LFnu; coefficient of variation = 3.1%), and 1.27 2.4 for HF normalized units (HFnu; coefficient of variation = 4.6%) when testing 5 healthy young adults one week apart (unpublished data).

Non-linear analyses of each 5-minute segment of R-R intervals were also performed with software (Matlab, The MathWorks, Inc.) for calculation of the sample entropy score with the fixed input parameters m = 2 and r = 0.2 standard deviations of the heart period data. We chose to use sample entropy as our non-linear measure of complexity because it was developed for cardiovascular and other biological signals and has been shown to provide more accurate data than other measures (i.e., approximate entropy) because its accuracy is not dependent on relatively long and stationary signals (24).

Blood Collection/Analysis

After an overnight fast and confirmation that the participant had not suffered from an injury or illness in the previous week, a blood sample was drawn by vena puncture for analysis of lipids, blood glucose, insulin, CRP, leptin, sOB-R, 8-isoprostane, adiponectin, IL-6, and TNF-. Blood glucose, cholesterol, and triglycerides were determined by calorimetric reflectance spectrophotometry. Insulin was determined by chemiluminescent immunoassay. As previously reported by Matthews et al. (25), the homeostasis model assessment (HOMA) was calculated as: (fasting glucose fasting insulin)/22.5. CRP was measured with an ultra-sensitive assay using rate nephlometry. Leptin, sOB-R, 8-isoprostane, adiponectin, IL-6, and TNF- were measured in the University of Minnesota Cytokine Reference Laboratory with enzyme-linked immunosorbent assay. Inter- and intra-assay coefficients of variation for 8-isoprostane and the adipokines were as follows (presented as inter- and intra-assay coefficients of variation): leptin (7.2 to 11.2; 4.6 to 8.0), sOB-R (5.2 to 8.6; 2.2 to 6.1), 8-isoprostane (6.0 to 18.0; 3.0 to 20.0), adiponectin (9.2 to 14.6; 2.8 to 5.6), IL-6 (5.9 to 7.9; 4.3 to 4.6), and TNF- (6.0 to 7.2; 3.7 to 4.8).

Statistical Analyses

Data were analyzed using a one-way ANOVA with Bonferroni post hoc testing when F was significant (p < 0.05) for cross-sectional comparisons of NW, OW, and OB children. Pearson's r was calculated to examine relationships between the main outcome variables of cANS function (HRV variables) and measures of leptin, sOB-R, oxidative stress (8-isoprostane), inflammation (CRP, IL-6, TNF-), and insulin resistance (fasting insulin and HOMA). For all measurements, means (standard deviation) were calculated and reported. CRP was log-transformed (i.e., Y = log X) and LFnu, HFnu, and LF:HF were transformed with the natural logarithm (i.e., Y = LnX) for data analyses due to their skewed distribution. Partial correlation analysis was performed to evaluate the potential interactions between the outcome variables while having the ability to adjust for covariates. All statistical procedures were performed using SPSS 10.0 (SPSS, Inc., Chicago, IL). An level of 0.05 was used to denote statistical significance.

Results

A summary of demographic and physiological variables is presented in Table 1. Overall, significant differences were found among groups for weight, BMI, total body fat %, insulin, HOMA, log-CRP, systolic blood pressure, and high-density lipoprotein. By design, weight, and BMI were significantly different among all three groups. However, total body fat % was not significantly different between OW and OB. Similarly, the trunk fat % was significantly different between the NW and OB groups (21.5 11.7 vs. 49.3 5.1%, p = 0.001), but not different between the OB and OW groups (43.3 5.4%). The OB group had significantly higher fasting insulin, HOMA, and systolic blood pressure than OW and NW, but there were no differences between OW and NW in these variables. Furthermore, log-CRP was significantly different among all three groups indicating a significant increase in CRP as level of adiposity increased. No gender differences were found for any of the HRV variables or for any of the blood variables (data not shown).

Table 1. - Demographics and physiologic profiles of normal-weight, overweight, and obese.

Full table

cANS Function, 8-Isoprostane, and Adipokines

The differences in HRV variables among the three groups are presented in Table 2. Significant differences among groups were found for LFnu, HFnu, and LF:HF (as well as each natural logarithm transform). Post hoc testing revealed no differences between NW and OW for all HRV variables. However, the OB group had significantly higher LFnu and LF:HF and significantly reduced HFnu as compared with NW.

Table 2. - Differences in HRV parameters of normal-weight, overweight, and obese children.

Full table

Table 3 shows the differences among groups for 8-isoprostane and adipokines. The results indicate plasma leptin levels were significantly different among groups with the highest levels being found in the OB group. 8-isoprostane was significantly increased in OB as compared with NW. No differences were noted among groups for adiponectin and TNF-.

Table 3. - Differences in 8-isoprostane and adipokines in normal-weight, overweight, and obese children.

Full table

Relationships Between cANS, Insulin Resistance, Oxidative Stress, Leptin, and Inflammation

As presented in Table 4, significant relationships were found among HRV measures, leptin, and HOMA. SDRR and HFnu were negatively related to leptin and after adjustment for fat mass, SDRR, and leptin remained negatively related. LF:HF was positively related, whereas HFnu was negatively related, to HOMA. However, partial correlation analyses indicated that these significant relationships were not significant after using fat mass as a covariate. As compared with the analysis with HOMA, similar findings were observed when examining the relationship between HRV measures and fasting insulin. BMI had a strong positive correlation with leptin (Table 4) and moderately strong negative correlation with sOB-R (r = -0.60, p = 0.01). When fat mass was used as the measure of adiposity, the relationship with leptin was strengthened (r = 0.82, p = 0.001), but the relationship with sOB-R was weakened (r = -0.49, p = 0.01). No significant relationships were noted between any of the HRV variables and sOB-R.

Table 4. - Correlation matrix among leptin, HOMA, and HRV variables.

Full table

Before adjustment, 8-isoprostane was significantly related to LFnu (r = 0.36, p = 0.04), SDRR (r = -0.44, p = 0.01), and RMSSD (r = -0.40, p = 0.02). There was a trend for a positive correlation between 8-isoprostane and LF:HF (r = 0.33, p = 0.06). After adjustment for fat mass, no significant relationships between HRV variables and 8-isoprostane were observed.

No significant relationships were noted between any of the HRV variables and IL-6 and TNF-. However, CRP was positively related to LF:HF (r = 0.40, p = 0.02) and LFnu (r = 0.38, p = 0.03) and negatively related to HFnu (r = -0.41, p = 0.02). After adjustment for fat mass, the relationships between CRP and HRV variables were not significant.

Discussion

The aim of the present study was to further clarify differences in cANS function between NW, OW, and OB children and to examine relationships between cANS function and leptin, insulin resistance, oxidative stress, and inflammation. Our main findings were: 1) cANS function was depressed in OB children, but no differences existed between OW and NW; and 2) significant relationships were found between measures of cANS function and leptin, insulin resistance (fasting insulin and HOMA), oxidative stress (8-isoprostane), and inflammation (CRP), but these relationships tended to be non-significant after adjustment for fat mass.

It was interesting that no differences were observed in HRV measures between the NW and OW children. The most pronounced differences were observed when comparing NW and children of the extreme levels of obesity (BMI > 95th percentile). Consequently, the results of this study indicate that the blunted cANS function is not fully realized in children until the obesity is substantially increased. Indeed, the majority of studies previously published in this area have compared NW with OB children. Similar to our findings, these studies have noted significantly lower RMSSD (3)(5)(6), SDRR (5)(6), HFnu (3)(4), and increased LF:HF (3)(5)(6) in OB as compared with NW children. However, Yakcini et al. reported data indicating no differences in sympathetic activity but reduced PSNS activity between OB and NW (7). Rabbia et al. showed that children who were recently OB (<4 years) had significantly increased sympathetic activation, but children who had been OB for >4 years were no different than healthy controls, which suggests that the duration of obesity might be a factor explaining the differences seen in studies assessing sympathovagal balance (5). We did not collect clinical data to examine the duration of obesity; thus, we cannot speculate as to what effect the duration of obesity had on our sample.

Despite clear trends that follow the dysfunction seen in the OB group, our data indicate that children who are OW or at risk for being OB (85th percentile < BMI <95th percentile) do not yet display significant changes in cANS function. We are unaware of previous studies that have delineated the differences in cANS function between OW and OB children. Nevertheless, these data support the need for interventions, such as a structured aerobic exercise program, which has previously been shown to reverse the abnormalities in cANS function associated with childhood obesity (26).

The authors are unaware of any previous studies assessing non-linear HRV characteristics with sample entropy in an OB, OW, and NW pediatric sample. The use of non-linear techniques is not new to HRV studies, but rather has been suggested to provide additional prognostic information above conventional measures (i.e., time and frequency domain) because it is known that, to some degree, mechanisms in cardiovascular regulation interact in a non-linear way (27). Vuksanovic and Gal showed the sensitivity of sample entropy for detecting cardiovascular disease by reporting that the change in sample entropy from rest to standing was significantly different between healthy controls and children with congenital heart failure (28). The sample entropy calculations from our novel analysis suggest a trend for reduced heart rate complexity in OB children. These data follow the similar patterns of the physiological interpretation of our time- and frequency domain HRV data indicating decreased HRV and, thus, overall diminished cANS function in OB children.

The significant differences we found for leptin and trends for differences for sOB-R among NW, OW, and OB children support previous findings and indicate that increased fat mass is associated with increased leptin and decreased sOB-R (leptin resistance). In the current study, we present novel data indicating that SDRR, a measure of overall HRV, is significantly related to leptin concentrations in children of varying levels of adiposity. Previous literature has indicated that leptin may act on the hypothalamus to induce thermogenesis via sympathetic nervous system mechanisms (9)(29). Our data indicate that as leptin concentrations increase, SDRR decreases, which is important because large epidemiological studies have shown that decreased SDRR is a predictor of adverse cardiac events and overall cardiovascular mortality (30) and associated with coronary heart disease (31).

The LF:HF ratio has been proposed to be an accurate measure of the overall sympathovagal balance of the autonomic nervous system in which higher values indicate a more sympathetically driven cardiovascular system (23)(32). We found that LF:HF was positively related to insulin resistance (fasting insulin and HOMA), inflammation (CRP), and tended to be positively related to oxidative stress (8-isoprostane). Interestingly, when adjusted for fat mass, no relationships existed between LF:HF and these measures, suggesting that fat mass or other factors associated with adiposity might be one of the physiological features that drives these relationships. The observation that fat mass may be one of the primary mediators of the relationships between cANS function and insulin resistance, inflammation, and oxidative stress underscores the importance of therapies (i.e., dietary and physical activity interventions and/or pharmacological therapies) targeted at decreasing adiposity for a given OB child. It should be noted that our data are limited in that we did not have an accurate measure of visceral fat mass, which has been implicated as the primary adipocyte (and, thus, adipokine) source relating to the adverse metabolic and cardiovascular effects of obesity (33). Further studies examining visceral adiposity and cANS function in childhood obesity is warranted.

Our study was limited in that we examined relationships among variables via correlation, and it is mindful to remember that correlation does not prove causation. It is physiologically plausible to suggest that the complex interplay of mechanisms regulating autonomic function, oxidative stress, insulin resistance, and inflammation are intricately linked to the increased cardiovascular risk observed with obesity. However, we cannot exclude the potential confounding effect of a factor that we did not measure on the results of the study. Therefore, more controlled experiments with larger sample sizes are required to elucidate the mechanisms involved in the complex relationship among cANS function, adipokines, insulin, and oxidative stress. Although the physiological significance of HRV has recently been debated (34) and has been criticized previously (35), we think that the vast amount of literature supporting its utility in accurately characterizing changes in autonomic modulation due to pharmacological interventions (36), predicting risk of sudden cardiac death after myocardial infarction (37), and identifying diabetic patients at risk of developing potentially fatal neuropathies (38) signifies the acceptability of the method for accurately assessing autonomic modulation of cardiac function. Furthermore, given the more invasive nature of alternative methods (i.e., muscle sympathetic nerve activity, plasma catecholamine levels) for estimating alterations in autonomic activity, the use of the non-invasive technique of HRV methodology in our sample of children appears justified.

In summary, the present study presents novel data indicating that OB, but not OW children, are characterized by cANS dysfunction as compared with NW and that the relationships between cANS function and leptin, insulin resistance, oxidative stress, and inflammation are primarily mediated by fat mass. We, therefore, suggest that appropriate therapies aimed at decreasing adiposity early in life are needed to potentiate the reduction in the cardiovascular disease risk factor profile of OW and OB children.

Notes

¹ Nonstandard abbreviations: cANS, cardiovascular autonomic; HRV, heart rate variability; PSNS, parasympathetic nervous system; SNS, sympathetic nervous system; sOB-R, soluble form of leptin receptor; IL-6, interleukin-6; TNF-, tumor necrosis factor-; CRP, C-reactive protein; OW, overweight; OB, obese; NW, normal-weight; ECG, electrocardiograph; BP, blood pressure; R-R, inter-beat; SDRR, standard deviation of normal R-R intervals, RMSSD, square root of the mean squared differences of successive R-R intervals; VLF, very low frequency; LF, low frequency; HF, high frequency; LFnu, LF normalized unit; HFnu, HF normalized unit; HOMA, homeostasis model assessment.

References

1. Hedley, A. A., Ogden, C. L., Johnson, C. L., Carroll, M. D., Curtin, L. R., Flegal, KM. (2004) Overweight and obesity among US children, adolescents, and adults, 1999–2002. JAMA 291: 847–850. | Article |
2. Dietz, WH. (2004) Overweight in childhood and adolescence. N Engl J Med. 350: 855–857. | Article | PubMed | ISI | ChemPort |
3. Martini, G., Riva, P., Rabbia, F., et al (2001) Heart rate variability in childhood obesity. Clin Auton Res. 11: 87–91. | Article | PubMed | ChemPort |
4. Nagai, N., Matsumoto, T., Kita, H., Moritani, T. (2003) Autonomic nervous system activity and the state and development of obesity in Japanese school children. Obes Res. 11: 25–32. | Article | PubMed |
5. Rabbia, F., Silke, B., Conterno, A., et al (2003) Assessment of cardiac autonomic modulation during adolescent obesity. Obes Res. 11: 541–548. | Article | PubMed |
6. Riva, G., Martini, G., Milan, A., Paglieri, C., Chiandussi, L., Veglio, F. (2001) Obesity and autonomic function in adolescence. Clin Exp Hypertens. 23: 57–67. | Article | PubMed |
7. Yakinci, C., Mungen, B., Karabiber, H., Tayfun, M., Evereklioglu, C. (2000) Autonomic nervous system functions in obese children. Brain Dev. 22: 151–153. | Article | PubMed | ChemPort |
8. Eikelis, N., Schlaich, M., Aggarwal, A., Kaye, D., Esler, M. (2003) Interactions between leptin and the human sympathetic nervous system. Hypertension 41: 1072–1079. | Article | PubMed | ISI | ChemPort |
9. Haynes, W. G., Morgan, D. A., Walsh, S. A., Mark, A. L., Sivitz, WI. (1997) Receptor-mediated regional sympathetic nerve activation by leptin. J Clin Invest. 100: 270–278. | Article | PubMed | ISI | ChemPort |
10. Popruk, S., Tungtrongchitr, R., Pongpaew, P., et al (2005) Relationship between soluble leptin receptor, leptin, lipid profiles, and athropometric parameters in overweight and obese Thai subjects. J Med Assoc Thai. 88: 220–227. | PubMed |
11. Reinehr, T., Kratzsch, J., Kiess, W., Andler, W. (2005) Circulating soluble leptin receptor, leptin, and insulin resistance before and after weight loss in obese children. Int J Obes Relat Metab Disord. 29: 1230–1235. | Article | ChemPort |
12. Steinberger, J., Steffen, L., Jacobs, D. R., Moran, A., Hong, C-P, Sinaiko, AR. (2003) Relation of leptin to insulin resistance syndrome in children. Obes Res. 11: 1124–1130. | Article | PubMed | ChemPort |
13. Tang-Christensen, M., Havel, P. J., Jacobs, R. R., Larsen, P. J., Cameron, JL. (1999) Central administration of leptin inhibits food intake and activates the sympathetic nervous system in rhesus macaques. J Clin Endocrinol Metab. 84: 711–717. | Article | PubMed | ChemPort |
14. Toni, R., Malaguti, A., Castorina, S., Roti, E., Lechan, RM. (2004) New paradigms in neuroendocrinology: relationships between obesity, systemic inflammation and the neuroendocrine system. J Endocrinol Invest. 27: 182–186. | PubMed | ISI | ChemPort |
15. Emdin, M., Gastaldelli, A., Muscelli, E., et al (2003) Hyperinsulinemia and autonomic nervous system dysfunction in obesity: effects of weight loss. Circulation 103: 513–519.
16. Gallistl, S., Sudi, K. M., Aigner, R., Borkenstein, M. (2001) Changes in serum interleukin-6 concentrations in obese children and adolescents during a weight reduction program. Int J Obes Relat Metab Disord. 25: 1640–1643. | Article | PubMed | ChemPort |
17. Warnberg, J., Moreno, L. A., Mesana, M. I., et al (2004) Inflammatory mediators in overweight and obese Spanish adolescents: the AVENA Study. Int J Obes Relat Metab Disord. 28(suppl): 59–63. | Article |
18. Lambert, M., Delvin, E. E., Paradis, G., O'Loughlin, J., Hanley, J. A., Levy, E. (2004) C-reactive protein and features of the metabolic syndrome in a population-based sample of children and adolescents. Clin Chem. 50: 1762–1768. | Article | PubMed | ISI | ChemPort |
19. Urakawa, H., Katsuki, A., Sumida, Y., et al (2003) Oxidative stress is associated with adiposity and insulin resistance in men. J Clin Endocrinol Metab. 88: 4673–4676. | Article | PubMed | ISI | ChemPort |
20. Araujo, F., Antelmi, I., Pereira, A. C., et al (2006) Lower heart rate variability is associated with higher serum sensitivity C-reactive protein concentration in healthy individuals aged 46 years or more. Int J Cardiol. 107: 333–337. | Article | PubMed |
21. Chrousos, GP. (1995) The hypothalamic-pituitary-adrenal axis and immune-mediated inflammation. N Engl J Med. 332: 1351–1362. | Article | PubMed | ISI | ChemPort |
22. Rosner, B., Prineas, R., Loggie, J., Daniels, SR. (1998) Percentiles for body mass index in US children 5 to 17 years of age. J Pediatr. 132: 211–222. | Article | PubMed | ISI | ChemPort |
23. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology (1996) Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation 93: 1043–1065. | PubMed | ISI |
24. Richman, J. S., Moorman, JR. (2000) Physiological time-series analysis using approximate entropy and sample entropy. Am J Physiol Heart Circ Physiol. 278: H2039–H2049. | PubMed | ChemPort |
25. Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., Turner, RC. (1985) Homeostasis model assessment: insulin resistance and beta-cell function from fasting glucose and insulin concentrations in man. Diabetologia 28: 412–419. | Article | PubMed | ISI | ChemPort |
26. Gutin, B., Howe, C. A., Johnson, M. H., Humphries, M. C., Snieder, H., Barbeau, P. (2005) Heart rate variability in adolescents: relations to physical activity, fitness, and adiposity. Med Sci Sports Exerc. 37: 1856–1863. | Article | PubMed |
27. Makikallio, T. H., Ristimae, T., Airaksinen, K. E., Peng, C. K., Goldberger, A. L., Huikuri, HV. (1998) Heart rate dynamics in patients with stable angina pectoris and utility of fractal and complexity measures. Am J Cardiol. 81: 27–31. | Article | PubMed | ChemPort |
28. Vuksanovic, V., Gal, V. (2005) Correlation properties and regularity of heart period time series: influence of posture and heart disease. Ann NY Acad Sci. 1048: 422–426. | Article | PubMed |
29. Matsumoto, T., Miyatsuji, A., Miyawaki, T., Yanagimoto, Y., Moritani, T. (2003) Potential association between endogenous leptin and sympatho-vagal activities in young japanese women. Am J Hum Biol. 15: 8–15. | Article | PubMed |
30. Tsuji, H., Venditti, F. J., Manders, E. S., et al (1994) Reduced heart rate variability and mortality risk in an elderly cohort: the Framingham Heart Study. Circulation 91: 878–883.
31. Liao, D., Cai, J., Rosamond, W. D., et al (1997) Cardiac autonomic function and incident coronary heart disease: a population-based case-cohort study. The ARIC Study. Atherosclerosis Risk in Communities Study. Am J Epidemiol. 145: 696–706. | PubMed | ChemPort |
32. Malliani, A., Pagani, M., Furlan, R., et al (1997) Individual recognition by heart rate variability of two different autonomic profiles related to posture. Circulation 96: 4143–4145. | PubMed | ChemPort |
33. Mori, Y., Hoshino, K., Yokota, K., Itoh, Y., Tajima, N. (2006) Differences in the pathology of the metabolic syndrome with or without visceral fat accumulation: a study in pre-diabetic Japanese middle-aged men. Endocrine 29: 149–153. | Article | PubMed | ChemPort |
34. Parati, G., di Rienzo, M., Castiglioni, P., Mancia, G., Taylor, J. A., Studinger, P., (2006) Point:Counterpoint: cardiovascular variability is/is not an index of autonomic control of circulation. J Appl Physiol. 101: 676–682. | Article | PubMed |
35. Eckberg, DL. (1997) Sympathovagal balance: a critical appraisal. Circulation 96: 3224–3232. | PubMed | ISI | ChemPort |
36. Montono, N., Cogliati, C., Porta, A., et al (1998) Central vagotonic effects of atropine modulate spectral oscillations of sympathetic nerve activity. Circulation 98: 1394–1399. | PubMed |
37. Kleiger, R. E., Miller, J. P., Bigger, J. T., Moss, AJ. (1987) Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. Am J Cardiol. 59: 256–262. | Article | PubMed | ISI | ChemPort |
38. Bellavere, F., Belzani, D. e., I, Masi, G, et al (1992) Power spectral analysis of heart rate variations improves assessment of diabetic cardiac autonomic neuropathy. Diabetes 41: 633–640. | Article | PubMed | ISI | ChemPort |

Acknowledgments

The authors thank Eric Williamson, Joseph Warpeha, and Brett Bruininks in data collection for this project. This study was supported in part by Minnesota Obesity Center Grant 1 P30 DK 50456–08 (to D.R.K.), and GCRC M01-RR00400, General Clinical Research Center Program, National Center for Research Resources/NIH.