 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* School of Physical and Health Education and Department of Community Health and Epidemiology, Queen's University, Kingston, Ontario, Canada
Department of Epidemiology, Tulane Center for Cardiovascular Health, Tulane University School of Public Health and Tropical Medicine, New Orleans, Louisiana
Tarleton State University, Stephenville, Texas
|| Human Genomics Laboratory, Pennington Biomedical Research Center, Baton Rouge, Louisiana
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods. Subjects included 2597 black and white, 5- to 18-year-old, male and female youths. Outcome measures included 7 CAD risk factors. In the first analysis step, BMI and WC were used as continuous variables to predict CAD risk factors. In the second analysis step, participants were placed into normal-weight, overweight, and obese BMI categories and, within each BMI category, CAD risk factors were compared for groups with low and high WC values.
Results. When BMI and WC were included in the same regression model to predict CAD risk factors, the added variance above that predicted by BMI or WC alone was minimal, which indicated that BMI and WC did not have independent effects on the risk factors. For example, for systolic blood pressure, BMI alone explained 7.3% of the variance, WC alone explained 7.7% of the variance, and the combination of BMI and WC explained 8.1% of the variance. When BMI and WC values were categorized with a threshold approach, WC provided information on CAD risk beyond that provided by BMI alone, particularly when the categories were used to predict elevated CAD risk factor levels. For instance, in the overweight BMI category, the high-WC group was 
2 times more likely to have high triglyceride levels, high insulin levels, and the metabolic syndrome, compared with the low-WC group.
Conclusion. These findings provide some evidence that a combination of BMI and WC should be used in clinical settings to evaluate the presence of elevated health risk among children and adolescents.
Key Words: obesity • body mass index • waist circumference • cholesterol • metabolic syndrome
Abbreviations: CAD, coronary artery disease • IOTF, International Obesity Task Force • CDC, Centers for Disease Control and Prevention • WC, waist circumference • HDL, high-density lipoprotein • LDL, low-density lipoprotein • OR, odds ratio • CI, confidence interval
The BMI is a predictor of numerous coronary artery disease (CAD) risk factors among children and adolescents,1,2 and its clinical utility in pediatric populations has been endorsed by numerous committees and organizations.3–6 Moreover, the International Obesity Task Force (IOTF)3 and US Centers for Disease Control and Prevention (CDC)4 have developed age- and gender-specific BMI cutoff points that can be used to classify children and adolescents as normal-weight, overweight, or obese. The IOTF cutoff points are tied to adult overweight (25 kg/m2) and obesity (30 kg/m2) thresholds,3 whereas the CDC cutoff points are based on a distributional approach in which the 85th and 95th percentiles of the population denote "at risk of overweight" and "overweight" thresholds, respectively.
Waist circumference (WC) also predicts CAD risk factors among young people.7–9 Whereas BMI is thought to be an indicator of overall adiposity, WC has been advocated as an indicator of abdominal fat content. At present, WC is not a routinely used measure in the pediatric setting, in part because no organizations have developed or endorsed WC cutoff points for children and adolescents. However, reference data for WC are available for Canada,10 Cuba,11 Italy,12 Spain,13 the United Kingdom,14 and the United States,15 and age- and gender-specific WC cutoff points for classifying 5- to 18-year-old youths as having either a low WC or high WC, based on relationships with CAD risk factors, were reported recently.15 These WC cutoff points provide an alternative to BMI for classifying obesity-related health risks among youths.
The weighted evidence for adults indicates that WC predicts health risk beyond that predicted by BMI alone.16–20 For a given BMI value or category, adults with higher WC values have a greater health risk than do adults with lower WC values. The combined influence of BMI and WC on obesity-related health outcomes among adults has been recognized in the US National Institutes of Health2 and Health Canada21 obesity classification systems. Both classification systems indicate that obesity-related health risk increases in a graded manner with the move from normal-weight (18.5–24.9 kg/m2) to overweight (25–29.9 kg/m2) to obese (
30 kg/m2) BMI categories, and, that within each of these BMI categories, adults with high WC values (>102 cm for men and >88 cm for women) are at greater health risk than adults with low WC values.
Far less is known about the combined influence of BMI and WC on health outcomes among children and adolescents. Although some studies indicated that WC is a better marker of CAD risk factors than BMI among school-aged youths,9,22 it is unclear whether BMI coupled with WC predicts CAD risk factors better than either measure alone. Therefore, there is a need to establish whether BMI and WC have independent effects on obesity-related health risk among young people and to assess the clinical utility of using a combination of BMI and WC in this age group. Answering these unknown questions could have important implications for determining the manner in which BMI and WC are used to classify overweight and obesity status among young people. Thus, the objectives of this investigation were to determine whether BMI and WC have independent effects on CAD risk factors among children and adolescents and to assess the clinical utility of incorporating WC in addition to BMI to identify children and adolescents with elevated CAD risk factors.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
22000) biracial (35% black and 65% white) community in Louisiana. Participation rates for the surveys are >80% for children and 60% for young adults. The present analysis was limited to a cross-sectional sample of 2597 individuals, 5 to 18 years of age, who were examined between 1992 and 1994. Informed consent was obtained from all participants, and study protocols were approved by the human subjects review committees of the Louisiana State University School of Medicine and the Tulane University School of Public Health and Tropical Medicine.
General Examination
Height and weight were measured in duplicate, to the nearest 0.1 cm and 0.1 kg, respectively, and the average of the 2 measurements was used to calculate BMI. WC was measured in triplicate, midway between the lowest rib and the superior border of the iliac crest. The average of the 3 measurements was used. Systolic and diastolic (fifth phase) blood pressure levels were measured in 6 replicates by 2 nurses on the right arm of the participants, who were in a relaxed sitting position. The means of the 6 measurements were used.
Laboratory Analyses
Participants fasted for 12 hours before a blood sample was obtained for determination of blood lipid, glucose, and insulin levels. Cholesterol and triglyceride levels in whole serum and the fraction containing high-density lipoprotein (HDL) cholesterol were determined with enzymatic procedures,23,24 with an Abbott VP instrument (Abbott Laboratories, North Chicago, IL). Serum low-density lipoprotein (LDL) cholesterol and HDL cholesterol levels were analyzed with a combination of heparin-calcium precipitation and agar-agarose gel electrophoresis procedures.25 The laboratory was monitored by the CDC surveillance program. Plasma glucose levels were measured with a glucose oxidase method, with a glucose analyzer (Beckman Instruments, Fullerton, CA). Plasma insulin concentrations were measured with a radioimmunoassay (Phaadebas insulin kit; Pharmacia Diagnostics, Piscataway, NJ).
Definition of Terms and Groups
Age-Adjusted BMI and WC Values
Because BMI and WC increase as a function of normal growth and maturation, age-adjusted values were created for analysis. BMI and WC were each regressed up to a full cubic polynomial in age (age, age2, and age3) within the gender-by-race groups, with forward stepwise regression. Variables were allowed to enter or leave the model at P < .05. The standardized residuals were retained, and these values represented the age-adjusted values.
BMI Categories
Subjects were divided into 3 BMI categories according to the IOTF BMI cutoff points for children.3 These age- and gender-specific cutoff points were derived from a large international sample with regression techniques, by passing a line through the adult cutoff points at 18 years. Participants with BMI values corresponding to an adult BMI of <25 kg/m2 were classified as normal weight, participants with BMI values corresponding to an adult BMI of 25 to 29.9 kg/m2 were classified as overweight, and participants with BMI values corresponding to an adult BMI of 30 kg/m2 were classified as obese.
WC Categories
There are currently no WC cutoff points that have been developed for use within BMI categories for children and adolescents. Although age- and gender-specific WC cutoff points for youths were developed in the Bogalusa Heart Study cohort26 and a number of nationally representative samples,10–15 without exception these cutoff points were not designed to be used within BMI categories but rather as an alternative to BMI. Virtually all obese youths have WC values above the threshold developed in the Bogalusa Heart Study, and virtually all normal-weight youths have WC values below these thresholds. It has been proposed that the age- and gender-specific 75th percentile be used for children and adolescents to denote a high WC.27,28 However, virtually all obese youths in this study had WC values above the 75th percentile for the American population.15 Therefore, the available WC cutoff points would provide little or no additional discriminatory ability in the normal-weight and obese BMI categories and are not appropriate for the present study. Finally, the use of a single WC cutoff point to define groups with low and high WC values, as performed for adults, is confounded by the fact that WC increases during normal growth and maturation.
Because the original intention for the use of WC within BMI categories in adult populations was to identify individuals with higher levels of abdominal fat than would be expected on the basis of BMI alone,2 BMI was used to develop the thresholds for denoting low and high WC values among children and adolescents in the present study. That is, individuals with WC values that were lower than expected, on the basis of their BMI and age, were categorized into the low-WC groups, whereas those with WC values that were higher than expected were categorized into the high-WC groups.
BMI and age were used in a stepwise regression model to predict WC within the gender-race groups. The regression equations were as follows: white male subjects: WC (cm) = [BMI (kg/m2) x (2.41 ± 0.03)] + [age (years) x (0.99 ± 0.05)] + [9.8 ± 0.7]; black male subjects: WC (cm) = [BMI (kg/m2) x (2.09 ± 0.03)] + [age (years) x (0.96 ± 0.05)] ± [14.4 ± 0.7]; white female subjects: WC (cm) = [BMI (kg/m2) x (2.14 ± 0.03)] + [age (years) x (0.45 ± 0.05)] + [17.3 ± 0.7]; black female subjects: WC (cm) = [BMI (kg/m2) x (1.96 ± 0.03)] + [age (years) x (0.59 ± 0.05)] + [19.4 ± 0.6].
The standardized residuals from the regression analyses were retained. Individuals with a negative residual had a WC that was lower than expected on the basis of their BMI and age and were categorized into the low-WC group. Individuals with a positive residual had a WC that was higher than expected and were categorized into the high-WC group.
Although we used the residual approach described above for our analyses, in Fig 1 we provide an illustration of how the relationship between BMI and WC could be used to determine high and low WC values in a clinical setting, in which case the use of a regression algorithm would be cumbersome. Within each of the gender-by-race groups, BMI was regressed against WC for 5 small age ranges (2- to 3-year ranges). The regression lines represent the middle of a given age range (eg, for the age range of 5–7 years, the regression line represents 6.5 years of age). For a given BMI value, individuals falling above the regression line would have a high WC value, whereas individuals falling below the regression line would have a low WC value. The BMI values in Fig 1 range from the 3rd to 97th percentiles of the US population4 for the given age ranges.
|
Statistical Analyses
All analyses were conducted with SAS software and procedures (SAS version 8, SAS Institute, Cary, NC). Descriptive statistics were computed for all variables of interest and are expressed as mean ± SD. The correlations between BMI and WC were determined with partial correlations, controlling for age. In the first phase of the analysis, the variance in CAD risk factors explained by BMI and WC was determined with stepwise regression. Variables were allowed to enter or leave the model at P < .05. Initially, the R2 was determined for a base model based on gender, race, and a full cubic polynomial in age (age, age2, and age3). Then, a full cubic polynomial for BMI (BMI, BMI2, and BMI3) and/or WC (WC, WC2, and WC3) was added to the base model, to determine the additional variance above the base model that was explained by the anthropometric variables. Logistic-regression analyses were also used to examine the independent and combined effects of BMI and WC on elevated CAD risk factors. The odds ratios (ORs) were computed for each 1-SD change in the age-adjusted BMI and WC values.
In the second phase of the analysis, subjects were placed into normal-weight, overweight, and obese BMI categories and into low- and high-WC categories. Within each of the BMI categories, unpaired t tests were used to compare the mean values of the CAD risk factors in the low- and high-WC groups. Logistic-regression analyses were used to compare the likelihood of having elevated CAD risk factors in the low- and high-WC groups within each BMI category. The low-WC group was used as the reference group (OR: 1).
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
|
|
95th percentile; replaces IOTF obesity category).4 With few exceptions, the results based on the CDC classification system (data not shown) were the same as those based on the IOTF classification system.
Consistency of Results Based on Gender and Race
For the analyses presented in Tables 2 through 4 and Fig 2, similar patterns of results were found within each of the 4 gender-by-race groups (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The second purpose of this study was to assess the clinical utility of incorporating WC in addition to BMI to predict CAD risk among children and adolescents. In the clinical setting, BMI (normal weight, overweight, or obese) and WC (low or high) are categorized with a threshold approach. When BMI and WC values are categorized, the strength of the association between them is reduced, compared with that when the association is based on continuous BMI and WC values. Therefore, BMI and WC are more likely to have independent effects on CAD risk factors when a clinical approach is used, because of the reduced covariance.
Because BMI and WC change during normal growth and maturation, at times rapidly, age-specific cutoff points are needed to classify adiposity status among children and adolescents. In this study, the IOTF age- and gender-specific BMI cutoff points3 were used to classify BMI status for the 5- to 18-year-old Bogalusa Heart Study participants. For WC, however, we were unable to use a preexisting classification system and were thus required to develop our own cutoff points. Because the original intention behind using WC within BMI categories in adult populations was to identify individuals with higher levels of abdominal fat than would be expected on the basis of BMI alone,2 BMI was used to develop the thresholds for denoting low and high WC values for children and adolescents in the present study. That is, individuals with WC values that were lower than expected on the basis of their BMI and age were categorized into the low-WC groups, whereas those with WC values that were higher than expected were categorized into the high-WC groups. With this approach, 50% of the individuals within each BMI category had a high WC. Whereas the mean BMI values were similar in the low- and high-WC groups within each BMI category, the mean WC values differed by 4.0 to 9.7 cm.
When the aforementioned clinical classification system was used in this study and the CAD risk factors were examined on a continuous scale (eg, LDL cholesterol level), the mean CAD risk factor values were only slightly different in the low- and high-WC groups. For example, in the normal-weight BMI category, the differences in mean CAD risk factor values in the low- and high-WC groups, although statistically significant for 4 of the 7 risk factors, were small (range: 0–8% difference) and of little clinical significance. To further explore the potential combined effect of BMI and WC, the odds of having elevated CAD risk factors (eg, high LDL cholesterol levels) were compared in the low- and high-WC groups within a given BMI category. The results of this analysis indicated that WC provided useful information for predicting individuals at elevated risk. That is, for 8 of the 21 ORs tested across the 3 BMI categories, the likelihood of having elevated CAD risk factors was significantly greater for children and adolescents with high WC values, compared with those with low WC values. For instance, in the overweight BMI category, the high-WC group was 2 times more likely to have high triglyceride levels, high insulin levels, and the metabolic syndrome, compared with the low-WC group. This observation indicates that BMI and WC have independent effects in predicting elevated CAD risk among youths when these anthropometric variables are categorized with a clinical approach.
The finding that children and adolescents with high WC values were more likely to have elevated CAD risk factors, compared with those with low WC values, within a given BMI category is comparable to previous observations among adults. Within normal-weight, overweight, and class I obese BMI categories of a representative sample of US men and women, individuals with high WC values (>102 cm in men and >88 cm in women) were more likely to have a number of metabolic disorders, compared with those with normal WC values.16 Similar results were found in a representative sample of Canadian women, although the added effect of WC was not as apparent for Canadian men.17
More recently, BMI and WC contributed independently to the prediction of nonabdominal (eg, subcutaneous fat in the arms and legs), abdominal subcutaneous, and visceral fat for 341 men and women, for whom total and regional fat was measured with MRI.30 Although we are unaware of a comparable study involving children and adolescents, this may be a mechanistic explanation for why youths with high WC values in the current study were more likely to have elevated CAD risk factors, compared with those with low WC values, within a given BMI category. In the aforementioned study,30 WC was only a modest predictor of nonabdominal and abdominal subcutaneous fat after controlling for BMI. Conversely, WC was a strong predictor of visceral fat, after controlling for BMI. Although the relative contribution of specific abdominal fat depots to obesity-related health risk is unclear,31,32 it is well documented that visceral fat is a predictor of numerous CAD risk factors among children and adolescents.33–35 Therefore, a higher level of visceral fat might have explained in part the increased odds of elevated CAD risk factors in the high-WC groups in the present study.
This study has notable strengths, specifically the large battery of CAD risk factor measurements and the use of a large biracial sample of 5- to 18-year-old male and female youths. The study population was, however, a nonrepresentative sample of youths. Additional studies in other populations are needed to confirm the generalizability of the classification approach used here.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
The Bogalusa Heart Study is a joint effort of many investigators and staff members, whose contributions are acknowledged gratefully.
|
FOOTNOTES |
|---|
Reprint requests to (I.J.) School of Physical and Health Education, Queen's University, Kingston, ON, Canada K7L 3N6. E-mail: janssen{at}post.queensu.ca
No conflict of interest declared.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. Zhu, W. Yan, D. Ge, F. A. Treiber, G. A. Harshfield, G. Kapuku, H. Snieder, and Y. Dong Relationships of Cardiovascular Phenotypes With Healthy Weight, at Risk of Overweight, and Overweight in US Youths Pediatrics, January 1, 2008; 121(1): 115 - 122. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Utriainen, J. Jaaskelainen, J. Romppanen, and R. Voutilainen Childhood Metabolic Syndrome and Its Components in Premature Adrenarche J. Clin. Endocrinol. Metab., November 1, 2007; 92(11): 4282 - 4285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Kelishadi, M. M. Gouya, G. Ardalan, M. Hosseini, M. Motaghian, A. Delavari, R. Majdzadeh, A. Heidarzadeh, M. S. Mahmoud-Arabi, M. M. Riazi, et al. First Reference Curves of Waist and Hip Circumferences in An Asian Population of Youths: CASPIAN Study J Trop Pediatr, June 1, 2007; 53(3): 158 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J Hussey, C Bell, K Bennett, J O'Dwyer, and J Gormley Relationship between the intensity of physical activity, inactivity, cardiorespiratory fitness and body composition in 7-10-year-old Dublin children Br. J. Sports Med., May 1, 2007; 41(5): 311 - 316. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Mukamal, I. Kawachi, M. Miller, and E. B. Rimm Body Mass Index and Risk of Suicide Among Men Arch Intern Med, March 12, 2007; 167(5): 468 - 475. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. I Brixner, Q. Said, P. K Corey-Lisle, A V. Tuomari, G. J L'Italien, W. Stockdale, and G. M Oderda Naturalistic Impact of Second-Generation Antipsychotics on Weight Gain Ann. Pharmacother., April 1, 2006; 40(4): 626 - 632. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2005 American Academy of Pediatrics. All rights reserved. |
||
|
||
