Abstract
Objective
To determine whether waist circumference (WC) and family history of disease increase the predictive utility of body mass index (BMI) for adult metabolic syndrome (MetS).
Study design
A subsample of 161 men and women from the Fels Longitudinal Study with childhood and adulthood measures were analyzed. Using logistic regression, childhood BMI categories (50th, 75th, and 85th percentiles), WC categories (75th and 90th percentiles), and family history of type 2 diabetes mellitus or cardiovascular disease were modeled separately and in combinations to predict adult MetS. Predicted probabilities and c-statistics were compared across models.
Results
The addition of family history to BMI improved the predicted probability of adult MetS from 29% to 52% (Δc-statistic = 0.13). The combination of WC and BMI was more predictive than BMI alone but did not outperform the combination of family history and BMI. In 3 of the 4 models with a combination of family history, WC, and BMI, the predicted probability of adult MetS did not exceed that from the combination of family history and BMI.
Conclusions
Family history of type 2 diabetes or cardiovascular disease is a useful addition to BMI in childhood to predict the future risk of adult MetS.
Although body mass index (BMI) and waist circumference (WC) are highly correlated, their respective role in the screening of children at risk for metabolic disorders continues to be debated. Age- and sex-adjusted BMI generally has been used to ascertain obesity in children in light of the relative ease of obtaining weight and height measurements and the ready availability of standardized growth charts. In contrast, the measurement of WC in children is not well standardized. But because WC is a proxy of visceral fat, the type of adipose tissue known to predispose individuals to high risk of metabolic disease,1 there is much interest in researching whether WC can replace or improve BMI as a screening tool.
Findings regarding BMI versus WC have been inconsistent across studies. Some reports have indicated WC as a predictor for insulin resistance, high blood pressure, and dyslipidemia, independent of BMI.2–5 In addition, Lee et al4 reported that in African-American and Caucasian children, the addition of WC to BMI percentile increased the explained variance in systolic blood pressure by 15% and in triglycerides and high-density lipoprotein cholesterol (HDL-C) levels by 3% and 7%, respectively. More recently, Rubin et al6 reported that in a sample of normal and overweight adolescents, WC was superior to BMI in terms of correlations with levels of adipocytokines, including adiponectin, resistin, and interleukin-6.
In contrast to the foregoing findings, however, Garnett et al7 reported that a high BMI at age 8 years resulted in a 7-fold increased risk for cardiovascular outcomes at age 15 years, compared with a 4-fold increase risk associated with a high WC at this age. A study of adolescents in Hong Kong found that BMI and WC demonstrated similar sensitivity and specificity levels in relation to metabolic syndrome (MetS).8 In the Fels Longitudinal Study, in which participants were followed from birth through adulthood, the divergence of BMI and WC values occurred at around the same age (age 6 to 8 years in boys and 13 years in girls) in adults with and without MetS.9 Finally, Janssen et al10 found that in children age 5 to 18 years in the Bogalusa Heart Study, including BMI and WC in the same regression model to predict coronary arterial disease risk factors resulted in only minimal additional variance above that predicted by BMI or WC alone.
The inconsistent findings in the literature may be the result of collinearity in regression models. Because BMI and WC are so highly correlated, estimates of their standard errors may be unstable when both are entered simultaneously into the same regression model. As such, the relative significance of BMI or WC may not be determinable using this approach. For example, in the report of Janssen et al,10 although the combination of BMI and WC did not seem to improve the prediction of BMI or WC alone, high WC appeared to increase the risk of coronary arterial disease within BMI subgroups in stratified analysis. Another methodological problem in the literature is that most studies reporting on this issue have not examined childhood data in relation to adult outcomes, so it is uncertain whether correlations between BMI or WC and metabolic risk factors in childhood and adolescence translate to the same patterns of prediction when adult outcomes are considered. Finally, when the goal is to improve the screening of at-risk children beyond that can be achieved by BMI, whether other measures, such as family history, can serve this purpose equally well or better than WC is unclear. Positive family history has been shown to independently predict type 2 diabetes mellitus (T2D) in a population cohort followed for 30 years.11
In the present study, we used data from the Fels Longitudinal Study to examine the predictive probabilities of BMI, WC, family history of T2D and cardiovascular disease, and combinations thereof in relation to adult MetS. Our use of predictive probabilities as a point of comparison eliminated the collinearity problem between BMI and WC, because standard errors were not at issue.
Methods
Data from a subsample of 161 male and female subjects of the Fels Longitudinal Study were examined. Details of the Fels Longitudinal Study have been reported previously.12 All subjects underwent at least one examination during childhood (age 6 to 20 years) and adulthood (25 to 55 years). Further inclusion criteria included the ability to be classified according to MetS status during adulthood and complete childhood measurements, including weight (kg) and height (cm) to calculate BMI (kg/m2), WC (cm), and a family history for either T2D or cardiovascular disease. BMI and WC were measured using standard techniques.12 BMI percentile was computed using the Centers for Disease Control and Prevention’s 2000 growth charts.13 WC was classified as either above or below the 75th percentile and, separately, above or below the 90th percentile, using both the reference values of Fernandez et al14 and Cook et al (see this Supplement). Family history of T2D or cardiovascular disease was determined from questionnaire data reported in the Fels Longitudinal Study by parents and siblings of the subjects. This direct information was supplemented by the subjects’ self-report of disease status for their parents and siblings. A positive family history was recorded as “yes,” a negative history as “no.” T2D was reported directly, but a combination of conditions was used to define cardiovascular disease, including stroke, congestive heart failure, angina, heart attack, and angioplasty or bypass surgery.
Adult MetS was chosen as the outcome measure for the present study because of the limited cases of T2D and cardiovascular disease in the study group. Adult MetS, comprising high WC, hyperglycemia, hypertriglyceridemia, low HDL-C, and hypertension, was defined based on Adult Treatment Panel III guidelines,15 consistent with most US population-based reports. All blood chemistry measurements were obtained with the subjects in the fasting state.
Data Analysis
BMI percentile was grouped into < 50th percentile, 50th to 85th percentile, and > 85th percentile. WC percentile was grouped as > 75th percential or > 90th percentile, as described previously. These variables were used together and separately and with or without family history in logistic regressions to predict adult MetS status. Predictive values for each BMI or WC group and family history status were computed, and corresponding c-statistics for each model were recorded.
Results
Characteristics of the study group are summarized in Table I. Overall, 8.7% of the sample was overweight in childhood, and 19.3% had MetS in adulthood. Using the reference values of Fernandez et al,14 14.9% of the sample had a WC above the 75th percentile and 3.7% had a WC above the 90th percentile in childhood. Using the reference values of Cook et al (see this Supplement), these prevalences were 13.7% and 1.2%, respectively. A positive family history of T2D or cardiovascular disease was seen in 37% of the subjects. Mean values of childhood anthropometric and adult anthropometric and blood chemistry data are given in Table II.
Table I.
Sample characteristics
| n | % | |
|---|---|---|
| Sex | ||
| Male | 69 | 42.9 |
| Female | 92 | 57.1 |
| Race | ||
| Caucasian | 161 | 100.0 |
| Family history | ||
| Cardiovascular disease | ||
| Yes | 42 | 26.1 |
| No | 119 | 73.9 |
| T2D | ||
| Yes | 30 | 18.6 |
| No | 131 | 81.41 |
| Combined | ||
| Yes | 60 | 37.3 |
| No | 101 | 62.7 |
| Adult MetS status | ||
| Yes | 31 | 19.3 |
| No | 130 | 80.8 |
| BMI percentile | ||
| < 50th | 82 | 50.9 |
| 50th to 84th | 64 | 39.8 |
| 85th to 94th | 14 | 8.7 |
| 95th | 1 | 0.6 |
| WC | ||
| Fernandez et al values | ||
| > 90th percentile | 6 | 3.7 |
| > 75th percentile | 24 | 14.9 |
| Cook et al values | ||
| > 90th percentile | 2 | 1.2 |
| > 75th percentile | 22 | 13.7 |
Table II.
Means of childhood and adulthood data
| n | Mean | SD | |
|---|---|---|---|
| Childhood | |||
| Age, years | 161 | 14.52 | 2.61 |
| WC, cm | 161 | 71.55 | 8.43 |
| Weight, kg | 161 | 52.98 | 13.38 |
| Height, cm | 161 | 162.24 | 13.78 |
| BMI, kg/m2 | 161 | 19.81 | 2.98 |
| BMI percentile | 161 | 48.33 | 25.28 |
| Adulthood | |||
| Age, years | 161 | 35.89 | 5.75 |
| WC, cm | 161 | 92.61 | 14.37 |
| BMI, kg/m2 | 161 | 26.24 | 5.54 |
| Systolic blood pressure, mm Hg | 160 | 111.15 | 11.98 |
| Diastolic blood pressure, mm Hg | 160 | 72.41 | 9.94 |
| Blood chemistry | |||
| HDL-C | 155 | 51.17 | 13.08 |
| Triglycerides | 155 | 145.06 | 120.60 |
| Glucose | 156 | 91.79 | 10.38 |
The predicted probabilities of BMI, family history, WC, and different combinations of these measures in relation to adult MetS are presented in Table III. The predicted probability of adult MetS was 18% to 19% in children with BMI < 85th percentile and 29% in overweight children (model 1). The inclusion of family history increased this predicted probability to 30% in normal-weight children and 52% in overweight children (model 2). Adding high WC (> 90th percentile based on the reference values of Fernandez et al14) to BMI yielded predicted probabilities of 34% in those with a childhood BMI between the 50th and 84th percentiles and 41% in those with a childhood BMI > 85th percentile (model 4). Comparable models were estimated using Cook et al’s reference value for high WC (model 8), where the predicted probability of adding WC > 90th percentile to overweight status reached 50%.
Table III.
Predicted probabilities of adult MetS
| Model | c-statistic | Predicted probability(95% CI) |
|---|---|---|
| 1 BMI percentile group | 0.53 | |
| <50th | 0.18 (0.11–0.28) | |
| 50th–84th | 0.19 (0.11–0.30) | |
| 85th–94th | 0.29 (0.11–0.56) | |
| 2 BMI percentile + Hx | 0.66 | |
| <50, Hx − | 0.11 (0.06–0.21) | |
| <50, Hx + | 0.29 (0.18–0.45) | |
| 50–84, Hx − | 0.11 (0.05–0.22) | |
| 50–84, Hx + | 0.30 (0.17–0.47) | |
| 85–94, Hx − | 0.25 (0.09–0.52) | |
| 85–94, Hx + | 0.52 (0.22–0.81) | |
| WC by Fernandez et al | ||
| 3 BMI percentile + WC>75th | 0.53 | |
| <50, WC<75 | 0.18 (0.11–0.28) | |
| <50, WC>75 | 0.22 (0.05–0.64) | |
| 50–84, WC<75 | 0.18 (0.10–0.31) | |
| 50–84, WC>75 | 0.22 (0.06–0.58) | |
| 85–94, WC<75 | 0.24 (0.04–0.71) | |
| 85–94, WC>75 | 0.29 (0.11–0.56) | |
| 4 BMI percentile + WC>90th | 0.54 | |
| <50, WC<90 | 0.18 (0.11–0.28) | |
| <50, WC>90 | 0.34 (0.05–0.82) | |
| 50–84, WC<90 | 0.19 (0.11–0.30) | |
| 50–84, WC>90 | 0.34 (0.06–0.82) | |
| 85–94, WC<90 | 0.23 (0.07–0.55) | |
| 85–94, WC>90 | 0.41 (0.10–0.81) | |
| 5 BMI percentile + Hx + WC>75th | 0.66 | |
| <50, Hx −, WC<75 | 0.11 (0.05–0.21) | |
| <50, Hx −, WC>75 | 0.17 (0.03–0.58) | |
| <50, Hx +, WC<75 | 0.30 (0.18–0.45) | |
| <50, Hx +, WC>75 | 0.42 (0.09–0.84) | |
| 50–84, Hx −, WC<75 | 0.11 (0.05–0.22) | |
| 50–84, Hx −, WC>75 | 0.17 (0.04–0.51) | |
| 50–84, Hx +, WC<75 | 0.29 (0.16–0.47) | |
| 50–84, Hx +, WC>75 | 0.41 (0.11–0.79) | |
| 84–94, Hx −, WC<75 | 0.16 (0.02–0.62) | |
| 84–94, Hx −, WC>75 | 0.25 (0.09–0.52) | |
| 84–94, Hx +, WC<75 | 0.39 (0.07–0.85) | |
| 84–94, Hx +, WC>75 | 0.53 (0.22–0.81) | |
| 6 BMI percentile + Hx + WC>90th | 0.66 | |
| <50, Hx −, WC<90 | 0.11 (0.06–0.21) | |
| <50, Hx −, WC>90 | 0.18 (0.02–0.69) | |
| <50, Hx +, WC<90 | 0.29 (0.17–0.45) | |
| <50, Hx +, WC>90 | 0.42 (0.07–0.87) | |
| 50–84, Hx −, WC<90 | 0.11 (0.05–0.22) | |
| 50–84, Hx −, WC>90 | 0.18 (0.02–0.69) | |
| 50–84, Hx +, WC<90 | 0.30 (0.17–0.47) | |
| 50–84, Hx +, WC>90 | 0.42 (0.07–0.87) | |
| 84–94, Hx −, WC<90 | 0.22 (0.06–0.53) | |
| 84–94, Hx −, WC>90 | 0.32 (0.07–0.76) | |
| 84–94, Hx +, WC<90 | 0.47 (0.16–0.81) | |
| 84–94, Hx +, WC>90 | 0.61 (0.18–0.92) | |
| WC by Cook et al | ||
| 7 BMI percentile + WC>75th | 0.54 | |
| <50, WC<75 | 0.18 (0.11,0.28) | |
| <50, WC>75 | 0.28 (0.07,0.69) | |
| 50–84, WC<75 | 0.18 (0.10,0.30) | |
| 50–84, WC>75 | 0.27 (0.08,0.63) | |
| 85–94, WC<75 | 0.19 (0.03,0.61) | |
| 85–94, WC>75 | 0.29 (0.11,0.57) | |
| 8 BMI percentile + WC>90th | 0.53 | |
| <50, WC<90 | 0.18 (0.11,0.28) | |
| <50, WC>90 | 0.40 (0.03,0.94) | |
| 50–84, WC<90 | 0.19 (0.11,0.30) | |
| 50–84, WC>90 | 0.41 (0.03,0.94) | |
| 85–94, WC<90 | 0.25 (0.08,0.55) | |
| 85–94, WC>90 | 0.50 (0.06,0.94) | |
| 9 BMI percentile + Hx + WC>75th | 0.66 | |
| <50, Hx −, WC<75 | 0.11 (0.05,0.21) | |
| <50, Hx −, WC>75 | 0.21 (0.04,0.62) | |
| <50, Hx +, WC<75 | 0.30 (0.18,0.45) | |
| <50, Hx +, WC>75 | 0.48 (0.13,0.86) | |
| 50–84, Hx −, WC<75 | 0.10 (0.05,0.21) | |
| 50–84, Hx −, WC>75 | 0.20 (0.05,0.55) | |
| 50–84, Hx +, WC<75 | 0.28 (0.15,0.46) | |
| 50–84, Hx +, WC>75 | 0.46 (0.14,0.82) | |
| 84–94, Hx −, WC<75 | 0.13 (0.01,0.54) | |
| 84–94, Hx −, WC>75 | 0.25 (0.09,0.53) | |
| 84–94, Hx +, WC<75 | 0.35 (0.06,0.81) | |
| 84–94, Hx +, WC>75 | 0.53 (0.23,0.82) | |
| 10 BMI percentile+ Hx + WC>90th | 0.66 | |
| <50, Hx −, WC<90 | 0.11 (0.06–0.21) | |
| <50, Hx −, WC>90 | 0.10 (0.004–0.77) | |
| <50, Hx +, WC<90 | 0.30 (0.17–0.45) | |
| <50, Hx +, WC>90 | 0.27 (0.02–0.90) | |
| 50–84, Hx −, WC<90 | 0.11 (0.05–0.23) | |
| 50–84, Hx −, WC>90 | 0.10 (0.004–0.78) | |
| 50–84, Hx +, WC<90 | 0.30 (0.17–0.48) | |
| 50–84, Hx +, WC>90 | 0.28 (0.02–0.90) | |
| 84–94, Hx −, WC<90 | 0.25 (0.08–0.55) | |
| 84–94, Hx −, WC>90 | 0.23 (0.02–0.84) | |
| 84–94, Hx +, WC<90 | 0.53 (0.19–0.84) | |
| 84–94, Hx +, WC>90 | 0.50 (0.06–0.94) |
Hx, family history.
Values in bold are imputed due to lack of sample for those cells.
When BMI, family history, and WC were all considered simultaneously (models 5, 6, 9, and 10), the highest predicted probability hovered around 50% in all cases of overweight status combined with a positive family history and high WC (75th or 90th percentile cutoff) in all models except model 6, in which the probability reached 61% for the combination of BMI > 85th percentile, positive family history, and WC > 90th percentile (using reference values of Fernandez et al14).
Discussion
This longitudinal study has examined whether adding family history and/or WC to BMI in childhood improves the predictive probability of MetS in adulthood. Our statistical approach avoided the collinearity problem in estimating standard errors in regression models while still allowing us to ascertain whether clinically significant benefits can be derived from using multiple tools for risk screening in children. Our results demonstrate a marked increase in risk for adult MetS when BMI exceeded the 85th percentile in childhood. The addition of family history to BMI significantly improved the predicted probability of adult MetS. The combination of high WC (> 90th percentile) and BMI was more predictive than BMI alone but did not outperform the combination of family history and BMI. In 3 of the 4 models in which BMI, family history, and WC were all included simultaneously, the predicted probability of adult MetS was not better than that derived from the combination of family history and BMI.
Consistent with the results of Janssen et al,10 we found that in children with BMI between the 50th and 84th percentiles and between the 85th and 94th percentiles, WC—particularly at the 90th percentile cutpoint—increased the predicted probability of adult MetS. Another study found evidence of the utility of WC in predicting metabolic abnormalities in normal-weight children.16 These findings suggest that WC might be useful within stratified BMI categories to discriminate children at lower versus higher metabolic risk, given a particular BMI level; however, this evidence likely is obscured by comparing only the P values from regression models with both BMI and WC entered simultaneously.
In 2007, an Expert Committee on childhood obesity guidelines17 recommended blood screening of all children and adolescents with BMI > 85th percentile. But pediatricians have been reluctant to accept this recommendation, given the time and financial costs involved in screening all children in the overweight but not obese category. Our results show that among overweight but not obese children (BMI between the 85th and 94th percentile), those with no family history of T2D or cardiovascular disease had a 25% risk of adult MetS, but that this risk more than doubled, to > 50%, in the presence of a positive family history. Thus, in light of our findings, although adding WC to BMI appears to improve prediction of adult MetS, family history seems to be an even better additional measure with BMI. In most cases, family history can be assessed more easily and accurately than WC, given the latter variable’s lack of standardized technique and reference values. Thus, adding family history to BMI might be useful to help identify children who should undergo more extensive screening, to reduce the clinical burden and facilitate screening adherence among pediatricians.
The present study is unique because it is the first to use extensive childhood and adulthood follow-up data, with family history and WC modeled in combination with BMI in childhood for the prediction of MetS in adulthood. However the study’s long-term follow-up limited our sample size. Our sample included only 1 subject with BMI > 95th percentile in childhood, which limited our ability to compare the role of family history and WC in children with frank obesity. Furthermore, whereas others have shown differences in WC by sex and ethnicity given a specific BMI level,18 our study included only Caucasian children, and our modest sample size did not allow for meaningful sex comparisons.
In conclusion, despite our sample’s limitations, our results suggest that WC can augment BMI to refine the screening of metabolically at-risk children, but a family history of T2D and cardiovascular disease appears to increase the predicted probability of adult MetS even more than WC when considered in conjunction with BMI in childhood. Given the relative ease of assessing family history compared with WC, this may have important implications for identifying children who are overweight but not yet obese requiring additional detailed screening for metabolic risks.
Glossary
- BMI
Body mass index
- HDL-C
High-density lipoprotein cholesterol
- MetS
Metabolic syndrome
- T2D
Type 2 diabetes mellitus
- WC
Waist circumference
Footnotes
The contents of this article do not necessarily represent the views and policies of the National Institutes of Health.
Author Disclosures
The following authors have no financial arrangement or affiliation with any corporate organization or manufacturer of any product discussed in this supplement: Christine M. Schubert, Stephen Cook, Shumei S. Sun, Terry T.-K. Huang.
References
- 1.Calabro P, Yeh ET. Intra-abdominal adiposity, inflammation, and cardiovascular risk: new insight into global cardiometabolic risk. Curr Hypertens Rep. 2008;10(1):32–8. doi: 10.1007/s11906-008-0008-z. [DOI] [PubMed] [Google Scholar]
- 2.Genovesi S, Antolini L, Giussani M, Pieruzzi F, Galbiati S, Valsecchi MG, et al. Usefulness of waist circumference for the identification of childhood hypertension. J Hypertens. 2008;26(8):1563–70. doi: 10.1097/HJH.0b013e328302842b. [DOI] [PubMed] [Google Scholar]
- 3.Hirschler V, Aranda C, de Calcagno ML, Maccalini G, Jadzinsky M. Can waist circumference identify children with the metabolic syndrome? Arch Pediatr Adolesc Med. 2005;159(8):740–4. doi: 10.1001/archpedi.159.8.740. [DOI] [PubMed] [Google Scholar]
- 4.Lee S, Bacha F, Arslanian SA. Waist circumference, blood pressure, and lipid components of the metabolic syndrome. J Pediatr. 2006;149(6):809–16. doi: 10.1016/j.jpeds.2006.08.075. [DOI] [PubMed] [Google Scholar]
- 5.Lee S, Bacha F, Gungor N, Arslanian SA. Waist circumference is an independent predictor of insulin resistance in black and white youths. J Pediatr. 2006;148(2):188–94. doi: 10.1016/j.jpeds.2005.10.001. [DOI] [PubMed] [Google Scholar]
- 6.Rubin DA, McMurray RG, Harrell JS, Hackney AC, Haqq AM. Do surrogate markers for adiposity relate to cytokines in adolescents? J Investig Med. 2008;56(5):786–92. doi: 10.2310/JIM.0b013e3181788cf1. [DOI] [PubMed] [Google Scholar]
- 7.Garnett SP, Baur LA, Srinivasan S, Lee JW, Cowell CT. Body mass index and waist circumference in midchildhood and adverse cardiovascular disease risk clustering in adolescence. Am J Clin Nutr. 2007;86(3):549–55. doi: 10.1093/ajcn/86.3.549. [DOI] [PubMed] [Google Scholar]
- 8.Ng VW, Kong AP, Choi KC, Ozaki R, Wong GW, So WY, et al. BMI and waist circumference in predicting cardiovascular risk factor clustering in Chinese adolescents. Obesity (Silver Spring) 2007;15(2):494–503. doi: 10.1038/oby.2007.588. [DOI] [PubMed] [Google Scholar]
- 9.Sun SS, Liang R, Huang TT, Daniels SR, Arslanian S, Liu K, et al. Childhood obesity predicts adult metabolic syndrome: the Fels Longitudinal Study. J Pediatr. 2008;152(2):191–200. doi: 10.1016/j.jpeds.2007.07.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Janssen I, Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM, Bouchard C, et al. Combined influence of body mass index and waist circumference on coronary artery disease risk factors among children and adolescents. Pediatrics. 2005;115(6):1623–30. doi: 10.1542/peds.2004-2588. [DOI] [PubMed] [Google Scholar]
- 11.Morrison JA, Friedman LA, Wang P, Glueck CJ. Metabolic syndrome in childhood predicts adult metabolic syndrome and type 2 diabetes mellitus 25 to 30 years later. J Pediatr. 2008;152(2):201–6. doi: 10.1016/j.jpeds.2007.09.010. [DOI] [PubMed] [Google Scholar]
- 12.Roche AF. Growth, Maturation, and Body Composition: The Fels Longitudinal Study 1929–1991. Cambridge, UK: Cambridge University Press; 1992. [Google Scholar]
- 13.Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS, Wei R, et al. CDC growth charts: United States. Adv Data. 2000;314:1–27. [PubMed] [Google Scholar]
- 14.Fernandez JR, Redden DT, Pietrobelli A, Allison DB. Waist circumference percentiles in nationally representative samples of African-American, European-American, and Mexican-American children and adolescents. J Pediatr. 2004;145(4):439–44. doi: 10.1016/j.jpeds.2004.06.044. [DOI] [PubMed] [Google Scholar]
- 15.Grundy SM. Metabolic syndrome scientific statement by the American Heart Association and the National Heart, Lung, and Blood Institute. Arterioscler Thromb Vasc Biol. 2005;25(11):2243–4. doi: 10.1161/01.ATV.0000189155.75833.c7. [DOI] [PubMed] [Google Scholar]
- 16.Plachta-Danielzik S, Landsberg B, Johannsen M, Lange D, Muller MJ. Association of different obesity indices with blood pressure and blood lipids in children and adolescents. Br J Nutr. 2008;100(1):208–18. doi: 10.1017/S0007114508882980. [DOI] [PubMed] [Google Scholar]
- 17.Barlow SE. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics. 2007;120(Suppl 4):S164–92. doi: 10.1542/peds.2007-2329C. [DOI] [PubMed] [Google Scholar]
- 18.Lee S, Kuk JL, Hannon TS, Arslanian SA. Race and gender differences in the relationships between anthropometrics and abdominal fat in youth. Obesity (Silver Spring) 2008;16(5):1066–71. doi: 10.1038/oby.2008.13. [DOI] [PubMed] [Google Scholar]