SEARCH

SEARCH BY CITATION

Keywords:

  • cardiovascular risk factors;
  • follow-up study;
  • reference;
  • classification;
  • metabolic syndrome

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Objective: To determine whether the U.S. Centers for Disease Control and Prevention (CDC; CDC Reference) or International Obesity Task Force (IOTF; IOTF Reference) BMI cut-off points for classifying adiposity status in children are more effective at predicting future health risk.

Research Methods and Procedures: The sample (N = 1709) included 4- to 15-year-old (at baseline) boys and girls from the Bogalusa Heart Study. Overweight and obesity status were determined using both the CDC Reference and IOTF Reference BMI cut-off points at baseline. The ability of childhood overweight and obesity, determined from the two BMI classification systems, to predict obesity and metabolic disorders in young adulthood (after a 13- to 24-year follow-up) was then compared.

Results: Independently of the classification system employed to determine adiposity based on childhood BMI, the odds of being obese and having all of the metabolic disorders in young adulthood were significantly (p < 0.05) higher in the overweight and obese groups by comparison with the nonoverweight groups. Childhood overweight and obesity, determined by both the CDC Reference and IOTF Reference, had a low sensitivity and a high specificity for predicting obesity and metabolic disorders in young adulthood. Overweight and obesity as determined by the CDC Reference were slightly more sensitive and slightly less specific than the corresponding values based on the IOTF Reference.

Discussion: Overweight and obesity during childhood, as determined by both the CDC and IOTF BMI cut-off points, are strong predictors of obesity and coronary heart disease risk factors in young adulthood. The differences in the predictive capacity of the CDC Reference and IOTF Reference are, however, minimal.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

There is a general consensus that BMI can be used to determine overweight and obesity status in children in the clinical setting and in population-based studies (1, 2, 3, 4). Because BMI increases with normal growth and maturation after the adiposity rebound (5), age- and gender-specific BMI cut-off points are needed for the proper classification of overweight and obesity in children.

Historically, age- and gender-specific 85th and 95th percentiles of national reference data have been employed as cut-off points to determine overweight and obesity status, respectively, in children. This approach was originally employed in 1975 by Garn et al. (6), who used the 85th and 95th age- and gender-specific percentiles of a sample of boys and girls from 10 states to define overweight and obesity within that sample. In 1991, Must et al. (7) developed nationally representative age- and gender-specific 85th and 95th BMI percentiles for 6- to 17-year-old children using data from the first U.S. National Health and Nutrition Examination Survey (NHANES).1 For a decade, these cut-off points were used as reference values to classify overweight and obesity status in both the research and clinical setting (3, 8). The U.S. Centers for Disease Control and Prevention (CDC) have recently provided updated age- (by month) and gender-specific 85th and 95th BMI percentiles for ages 2 to 19 that were derived using data from five large representative American surveys (2).

In 1998, the World Health Organization proposed a universal classification system for defining overweight and obesity in adults (9). In this classification system, BMI cut-off points of 25 and 30 kg/m2 denote overweight and obesity, respectively. In 1999, an International Obesity Task Force (IOTF) Working Group recommended the development of BMI cut-off points for defining overweight and obesity in 2- to 18-year-old children that are linked to the adult health-related cut-off points of 25 and 30 kg/m2 (10). These cut-off points were subsequently developed and published by Cole et al. (1). They obtained BMI data in large representative surveys from six countries. In these cross-national samples, the adult BMI cut-off points at age 18 were regressed back through the growth curve, providing age- (6-month intervals) and gender-specific overweight and obesity cut-off points for children.

At present, both the CDC Reference and IOTF Reference BMI cut-off points are routinely employed to assess childhood obesity. Thus, researchers and clinicians are faced with the decision of determining which cut-off points to use. The logical choice would be to employ the cut-off points that are the most accurate at predicting health risk. Overweight and obesity in childhood are related to elevated levels of cardiovascular disease risk factors (11, 12, 13). However, given that children rarely develop overt cardiovascular disease, there is growing interest in identifying children and adolescents who have an increased risk of future health problems as adults. To our knowledge, no studies have compared the ability of the CDC and IOTF BMI classification systems to predict future health outcomes. Therefore, the aim of this paper was to determine whether the CDC Reference or IOTF Reference BMI classification system better predicts future health risk. To this end, overweight and obesity status in 4- to 15-year-old boys and girls from the Bogalusa Heart Study were determined using both the CDC and IOTF References. The ability of childhood overweight and obesity, determined from the two different BMI classification systems, to predict obesity and metabolic disorders in young adulthood (after a 13- to 24-year follow-up) was then compared.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Study Population and Design

The Bogalusa Heart Study consists of a series of cross-sectional surveys (1976 to 1996) of cardiovascular disease risk factors in school-aged children and young adults from Bogalusa, a semirural (total population ∼ 22, 000) biracial (∼35% black, 65% white) community in Louisiana. Participation rates in the surveys ranged from over 80% for school-aged children to over 60% for the young adult cohort. Nested within the cross-sectional surveys is a cohort of individuals who have been measured on multiple occasions and are the focus of the present investigation.

Children (N = 3865) 4 to 15 years old were examined as part of the 1976 to 1977 cross-sectional survey. These individuals were subsequently invited to participate in cross-sectional surveys of young adults in 1988 to 1991 and 1995 to 1996, at which time they were 19 to 38 years of age. The cohort for the present study was selected from those who participated in the 1976 to 1977 survey and either the 1988 to 1991 or 1995 to 1996 survey and who met the following criteria: completion of an overnight fast for the adult exam and absence of a reported pregnancy. Accordingly, 1709 individuals (44% males, 66% white) with a follow-up interval of 13 to 24 years were selected for the study. Informed consent was obtained from all participants. Study protocols were approved by review committees of the Louisiana State University School of Medicine and the Tulane University School of Public Health and Tropical Medicine.

Measures

Baseline Survey (1976 to 1977)

Height and weight were measured in duplicate to the nearest 0.1 cm and 0.1 kg, respectively, and the average of the two measurements was used to calculate BMI (kilograms per meter squared). Exact chronological ages were calculated as the date of the examination minus birth date.

Follow-up Surveys (1988 to 1991 and 1995 to 1996)

Height and weight were measured in duplicate to the nearest 0.1 cm and 0.1 kg, respectively, and the average of the two measurements was used to calculate BMI. Waist circumference was measured in triplicate midway between the lowest rib and the superior border of the iliac crest using a flexible tape. The average of the three waist circumference measurements was used. Systolic and diastolic (fifth phase) blood pressure (BP) levels were measured in six replicates by two nurses on the right arm of the participants while they were in a relaxed, sitting position. The means of the six BP measurements were used.

Cholesterol and triglyceride levels of whole serum and the fraction containing high-density lipoprotein (HDL) were determined using enzymatic procedures (14, 15) on the Abbott VP instrument (Abbott Laboratories, North Chicago, IL). Serum low-density lipoprotein (LDL)-cholesterol and HDL-cholesterol levels were analyzed by a combination of heparin-calcium precipitation and agar-agarose gel electrophoresis procedures (16). The laboratory was monitored by the CDC (Atlanta, GA) surveillance program. Measurement error, estimated from the coefficient of variation of 189 pairs of blind duplicate determination, was 3.9% for triglycerides, 3.6% for LDL-cholesterol, and 4.9% for HDL-cholesterol (17).

Definition of Groups and Terms

Subjects were divided into three BMI groups as children (baseline) and as young adults (follow-up). As young adults, subjects were classified as non-overweight (≤24.9 kg/m2), overweight (25.0 to 29.9 kg/m2), or obese (≥30.0 kg/m2) according to the World Health Organization (9) and NIH (18) obesity guidelines. As children, subjects were classified according to both the CDC Reference (2) and IOTF Reference (1) BMI cut-off points.

The CDC Reference BMI cut-off points are based on five nationally representative American surveys conducted from 1963 to 1994, including the National Health Examination Survey II (1963 to 1965, 6- to 11-year-old subjects), National Health Examination Survey III (1966 to 1970, 12- to 17-year-old subjects), NHANES I (1971 to 1974, 2- to 19-year-old subjects), NHANES II (1976 to 1980, 2- to 19-year-old subjects), and NHANES III (1988 to 1994, 2- to 5-year-old subjects) (2). The CDC Reference BMI growth charts provide age (in 1-month intervals, compared with 6-month intervals for IOTF) and gender-specific percentiles for ages 2 to 19 years. Using the CDC approach, the 85th and 95th BMI percentiles were used to classify subjects as non-overweight (≤84th percentile), overweight (85th to 94th percentiles), and obese (≥95th percentile). Although the CDC guidelines use the terms “at risk for overweight” and “overweight” rather than overweight and obese, respectively, to remain consistent with the terminology used for adults and that used by the IOTF Reference in children, we have used the terms overweight and obesity for the CDC Reference in this study.

The IOTF Reference BMI cut-off points were derived from a large international data set comprised of nationally representative samples of children collected over a number of years from the U.S. (1963 to 1980, 2- to 18-year-old subjects), Brazil (1989, 2- to 18-year-old subjects), Great Britain (1978 to 1993, 2- to 18-year-old subjects), Hong Kong (1993, 2- to 18-year old subjects), the Netherlands (1980, 2- to 18-year-old subjects), and Singapore (1993, 6- to 18-year-old subjects). Regression techniques were used to create smoothed percentile BMI curves that passed through the adult cut-off points of 25 (overweight) and 30 (obese) kg/m2 at 18 years. Age- (by 6-month intervals) and gender-specific BMI cut-off points were produced for ages 2 to 18 years. Using the IOTF Reference approach, children with BMI values corresponding to an adult BMI ≤ 24.9 kg/m2 were classified as non-overweight, children with BMI values corresponding to an adult BMI of 25 to 29.9 kg/m2 were classified as overweight, and children with BMI values corresponding to an adult BMI ≥ 30 kg/m2 were classified as obese.

During young adulthood, abdominal obesity was defined according to the NIH obesity guidelines as a waist circumference >102 cm in men or >88 cm in women (18). Hypertension was defined according to the guidelines of the Joint National Committee on Detection, Evaluation, and Treatment of High BP as a systolic BP ≥ 140 mm Hg or diastolic BP ≥ 90 mm Hg (19). High LDL-cholesterol (≥160 mg/dL), low HDL-cholesterol (<40 mg/dL), high triglycerides (≥200 mg/dL), and the metabolic syndrome were defined according to the National Cholesterol Education Program guidelines (20). The metabolic syndrome, which represents a clustering of coronary heart disease risk factors, was defined as including three or more of the following: triglycerides ≥ 150 mg/dL, HDL-cholesterol < 40 mg/dL in men or <50 mg/dL in women, systolic BP ≥ 130 mm Hg or diastolic BP ≥ 85 mm Hg, fasting glucose ≥ 110 mg/dL, or waist circumference >102 cm in men or >88 cm in women (20).

Statistical Analysis

Because the distribution of triglyceride values was positively skewed, they were log transformed before further analysis. Differences among non-overweight, overweight, and obese groups (based on childhood/baseline BMI) in young adulthood (follow-up) were tested using analysis of covariance and χ2 statistics. Logistic regression was used to predict the likelihood of obesity and metabolic disorders in young adulthood based on childhood BMI groups. Dummy variables were created to compute odds ratios (ORs) for these factors. The non-overweight group was used as the reference group (OR = 1.00), and all three adiposity groups were included in the same logistic regression models. Age, race, gender, and follow-up length were used as covariates for the analysis of covariance and logistic regression analyses. To determine the utility of childhood obesity to predict obesity and metabolic disorders in young adulthood, sensitivities, specificities, and areas under the receiver operating characteristic curve (AUCs) were computed. AUCs were compared for the CDC Reference and IOTF Reference using the methods described by Hanley and McNeil (21).

All of the aforementioned analyses were performed twice in the entire study cohort: once based on the CDC Reference and a second time based on the IOTF Reference. The receiver operating characteristic analyses were conduced with SPSS (version 12, SPSS Inc., Chicago, IL), and all other analyses were conducted with SAS (version 8, SAS Institute, Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

The percentages of the sample at baseline (4 to 15 years old) within the non-overweight, overweight, and obese BMI groups determined using the CDC and IOTF References are shown in Table 1. A slightly smaller percentage of the participants was classified as overweight based on the CDC Reference compared with the IOTF Reference, whereas a slightly greater percentage of the participants was classified as obese based on the CDC Reference compared with the IOTF Reference.

Table 1. . Prevalence (percentage) of study participants who were non-overweight, overweight, and obese in childhood (baseline) according to the CDC Reference and IOTF Reference BMI cut-off points
 Adiposity status in childhood
 CDC ReferenceIOTF Reference
 Non-overweightOverweightObeseNon-overweightOverweightObese
  1. Data presented as prevalences (percentage).

All subjects83.79.96.484.711.14.3
Black females81.811.07.282.711.65.7
Black males87.95.66.489.65.64.8
White females86.29.04.885.711.13.2
White males79.812.37.982.313.34.8

Table 2 shows the characteristics of the sample in young adulthood (follow-up, 19 to 38 years old) according to BMI status in childhood determined 13 to 24 years earlier. All of these statistical comparisons were controlled for age, gender, race, and length of follow-up. Independently of the BMI classification system employed, those who were overweight in childhood had higher weight, BMI, and waist circumference values in young adulthood compared with those who were non-overweight in childhood (p < 0.001). In turn, participants who were obese in childhood had higher weight, BMI, and waist circumference values in young adulthood compared with participants who were overweight in childhood (p < 0.001). Without exception, the BP and blood lipid values were higher (HDL lower) in young adulthood in the groups that were overweight and obese as children compared with the group that had normal BMI as children (p < 0.05). The BP values also tended to be higher in young adulthood in the group that was obese as children compared with the group that was overweight as children (p < 0.05). The findings for the BP and blood lipid variables were consistent whether adiposity status was based on the CDC Reference or IOTF Reference. Furthermore, for any given BMI level (non-overweight, overweight, obese), the mean values of the anthropometric and metabolic variables were comparable whether the CDC or IOTF classification system was employed. For example, systolic BP values (mean ± SD) in the obese groups were 119 ± 13 mm Hg based on the CDC Reference and 120 ± 13 mm Hg based on the IOTF Reference.

Table 2. . Participant characteristics in young adulthood according to childhood (baseline) adiposity status determined using the CDC Reference and IOTF Reference BMI cut-off points
  Adiposity status in childhood
  CDC ReferenceIOTF Reference
 All SubjectsNon-overweightOverweightObeseNon-overweightOverweightObese
  • Data presented as means ± SD. All statistical comparisons have been adjusted for age, gender, race, and follow-up length.

  • *

    Different from non-overweight group (analysis of covariance, p < 0.05).

  • Different from overweight group (analysis of covariance, p < 0.05).

Height (cm)169.2 ± 9.4169.0 ± 9.3169.5 ± 9.5171.1 ± 9.2*169.1 ± 9.3169.3 ± 9.6170.6 ± 9.0
Weight (kg)75.9 ± 19.471.9 ± 16.491.2 ± 17.0*105.0 ± 23.0*,72.1 ± 16.592.0 ± 18.3*108.7 ± 23.3*,
BMI (kg/m)26.4 ± 6.225.1 ± 5.131.7 ± 5.4*35.9 ± 7.7*,25.1 ± 5.132.1 ± 5.7*37.4 ± 7.8*,
Waist circumference (cm)84.7 ± 14.881.8 ± 12.796.1 ± 13.4*105.5 ± 16.8*,81.9 ± 12.897.3 ± 14.5*107.2 ± 16.7*,
Systolic BP (mm Hg)113.0 ± 11.2112.3 ± 10.7115.4 ± 12.1*118.8 ± 13.1*,112.4 ± 10.7115.1 ± 12.2*120.2 ± 13.4*,
Diastolic BP (mm Hg)73.4 ± 9.172.7 ± 8.776.1 ± 10.1*77.9 ± 10.6*72.8 ± 8.775.8 ± 10.1*79.3 ± 11.3*,
LDL-cholesterol (mg/dL)119.3 ± 33.6117.3 ± 32.7130.7 ± 35.3*127.4 ± 38.0*117.5 ± 32.8129.5 ± 35.9*128.6 ± 38.0*
HDL-cholesterol (mg/dL)50.9 ± 14.351.8 ± 14.546.9 ± 12.4*45.4 ± 11.3*51.7 ± 14.546.8 ± 11.9*45.8 ± 12.2*
Triglycerides (mg/dL)116.7 ± 114.8109.4 ± 86.8153.8 ± 188.8*156.3 ± 221.7*109.7 ± 86.6116.5 ± 235.0*134.5 ± 107.3*

The prevalences of obesity and metabolic disorders in young adulthood according to BMI status determined in childhood are provided in Table 3. Independently of the BMI classification criteria employed, the prevalences of obesity, abdominal obesity, hypertension, high LDL-cholesterol, low HDL-cholesterol, high triglycerides, and the metabolic syndrome in young adulthood increased in a graded fashion (P for trend < 0.05) when moving from non-overweight, to overweight, to obesity as determined during childhood. For each of the three BMI categories, the prevalences were similar whether the CDC Reference or IOTF Reference BMI cut-off points were used. With the former classification system, 26.4% of the obese group developed the metabolic syndrome, whereas with the latter classification system, 27.4% of the obese group developed the metabolic syndrome.

Table 3. . Prevalences and ORs for obesity and metabolic disorders in young adulthood according to childhood (baseline) adiposity status determined using the CDC Reference and IOTF Reference BMI cut-off points
 Adiposity status in childhood
 Prevalence (%)OR (95% Confidence invervals)
 Non-overweightOverweightObeseNon-overweightOverweightObese
  • Non-overweight subjects were used as the referent category (ORs = 1.00), and all three adiposity groups were included in the same regression models. All ORs were adjusted for age, race, gender, and follow-up length.

  • *

    Trend for increasing prevalences when moving from the non-overweight, to overweight, to obese BMI category (χ2, p < 0.01).

  • Different from non-overweight group (logistic regression, p < 0.05).

  • Trend for increasing ORs when moving from the non-overweight, to overweight, to obese BMI category (logistic regression, p < 0.05).

CDC Reference      
 Obesity14.759.873.6*1.009.65 (6.81 to 13.78)16.87 (10.77 to 27.11)†,
 Abdominal obesity15.052.169.1*1.006.92 (4.88 to 9.87)13.95 (8.96 to 22.14)†,
 Hypertension1.83.08.2*1.002.04 (0.70 to 5.12)4.39 (1.83 to 9.72)†,
 High LDL-cholesterol8.719.514.6*1.002.47 (1.58 to 3.75)1.63 (0.89 to 2.81)
 Low HDL-cholesterol18.233.732.7*1.002.23 (1.55 to 3.21)2.05 (1.31 to 3.16)
 High triglycerides8.316.018.2*1.001.89 (1.16 to 2.97)2.24 (1.28 to 3.78)†,
 Metabolic syndrome6.720.726.4*1.002.53 (2.26 to 5.42)4.62 (2.81 to 7.45)†,
IOTF Reference      
 Obesity15.259.880.8*1.009.59 (6.85 to 13.53)24.08 (13.43 to 46.06)†,
 Abdominal obesity14.957.772.6*1.008.81 (6.27 to 12.45)15.73 (9.17 to 28.00)†,
 Hypertension1.73.79.6*1.002.65 (1.01 to 6.16)4.65 (1.72 to 11.23)†,
 High LDL-cholesterol8.918.015.1*1.002.18 (1.41 to 3.30)1.73 (0.84 to 3.27)
 Low HDL-cholesterol18.432.834.2*1.002.13 (1.50 to 3.01)2.37 (1.38 to 3.99)†,
 High triglycerides8.416.916.4*1.001.99 (1.27 to 3.06)2.13 (1.04 to 4.05)†,
 Metabolic syndrome6.821.727.4*1.003.69 (2.38 to 5.45)4.87 (2.70 to 8.52)†,

Table 3 also lists the associations of childhood BMI level with obesity and metabolic disorders in young adulthood. All of the ORs were adjusted for age, gender, race, and length of follow-up. Independently of the classification criteria employed to determine adiposity based on childhood BMI, the likelihood of obesity, abdominal obesity, and the various metabolic disorders was significantly (p < 0.05) higher in the overweight and obese groups by comparison with the non-overweight (referent) groups with the exception of high LDL-cholesterol in the obese groups. Further analysis revealed that the ORs for obesity, abdominal obesity, hypertension, low HDL-cholesterol (IOTF Reference only), high triglycerides, and the metabolic syndrome increased in a graded fashion (P for trend < 0.05) when moving from the non-overweight, to overweight, to obese groups. In addition to BMI status, race and gender contributed significantly to most of the logistic regression models (data not shown). Stratified analyses indicated similar patterns of ORs for overweight and obesity in the racial and gender groups, as shown for the metabolic syndrome in Figure 1.

image

Figure 1. The likelihood of having the metabolic syndrome in young adulthood based on adiposity status determined during childhood. Adiposity status in childhood was determined using the CDC Reference (left) and IOTF Reference (right) BMI cut-off points. Within each race and gender group, non-overweight subjects were used as the referent group (ORs = 1.00). (•) ORs for the overweight groups. (○) ORs for the obese groups. Error bars denote 95% confidence intervals.

Download figure to PowerPoint

The sensitivities and specificities of the CDC and IOTF overweight and obesity cut-off points are presented in Table 4. For the overweight analysis, the study participants were classified as non-overweight or overweight (overweight including obese for this analysis) based on childhood BMI, whereas for the obesity analysis, the study participants were classified as obese or non-obese (normal weight + overweight) based on childhood BMI. Childhood overweight and obesity, determined by both the CDC Reference and IOTF Reference BMI classification systems, had a low sensitivity and high specificity for predicting obesity and metabolic disorders in young adulthood. The CDC overweight and obesity cut-off points were slightly more sensitive (0.0% to 2.2% for overweight, 2.9% to 6.1% for obesity) and slightly less specific (0.6% to 1.1% for overweight, 1.1% to 2.1% for obesity) than the IOTF cut-off points. Stratified analyses indicated similar patterns of findings in each of the racial and gender groups (data not shown).

Table 4. . Capacity of childhood overweight and obesity, determined using the CDC Reference and IOTF Reference BMI cut-off points at baseline, to predict obesity and metabolic disorders in young adulthood
 Adiposity status in childhood
 Overweight*Obesity 
 SensitivitySpecificitySensitivitySpecificityAUC
  • Sensitivity = true positive/(true positive + false negative) × 100. Specificity = true negative/(true negative + false positive) × 100.

  • *

    Overweight + obese vs. non-overweight.

  • Obese vs. non-overweight + overweight.

  • Non-overweight, overweight, and obese included in ROC analysis as three groups.

CDC Reference     
 Obesity46.492.620.797.80.70
 Abdominal obesity43.491.420.197.40.68
 Hypertension35.984.123.194.00.61
 High LDL-cholesterol28.285.09.293.90.56
 Low HDL-cholesterol26.386.310.294.50.56
 High triglycerides28.385.012.094.20.57
 Metabolic syndrome40.086.118.194.80.63
IOTF Reference     
 Obesity43.993.215.198.90.69
 Abdominal obesity42.992.514.098.50.68
 Hypertension35.985.217.996.10.61
 High LDL-cholesterol25.985.96.396.00.56
 Low HDL-cholesterol24.687.17.196.50.56
 High triglycerides26.585.97.296.00.56
 Metabolic syndrome38.187.012.596.60.63

The AUCs for predicting obesity and metabolic disorders in young adulthood based on adiposity status (non-overweight, overweight, or obese) in childhood for both the CDC and IOTF References are presented in Table 4. Without exception, the AUCs were not different for the CDC and IOTF classification systems (p > 0.05).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

A number of previous reports from the Bogalusa Heart Study have demonstrated that the level of childhood adiposity is a predictor of obesity and coronary heart disease risk factors in young adulthood (17, 22, 23, 24, 25, 26, 27). This study builds on these earlier findings by examining and comparing the predictive ability of the CDC Reference and IOTF Reference BMI classification systems for children, both of which are routinely employed in the clinical and research settings. Our results indicate that overweight and obesity in black and white boys and girls, as determined by both classification systems, are strong predictors of obesity and the development of coronary heart disease risk factors in young adulthood. Differences in the predictive capacity of the CDC Reference and IOTF Reference BMI cut-off points are minimal.

To our knowledge, no prior studies have compared the utility of the CDC Reference and IOTF Reference child obesity cut-off points to predict cardiovascular risk factors. However, a large body of evidence has demonstrated that childhood overweight and obesity, determined from the previous U.S. national reference BMI cut-off points (e.g., the cut-off points that were used before the development of the CDC Reference in 2002), are related to health outcomes in childhood and the development of morbidity from childhood to adulthood in men and women of different racial origins (11, 12, 28). Further, recent studies have shown that adiposity status, based on the IOTF Reference, predicts health risk (13, 29). The prospective findings from a large biracial cohort in the present study are consistent with these reports (11, 12, 13, 28, 29). Together, the results indicate that childhood overweight and obesity, whether defined according to the CDC or IOTF classification systems, are strong risk factors for the development of metabolic disorders. Thus, the clinical evaluation of BMI status in children and adolescents is useful for identifying high-risk patients who may benefit from interventions aimed at preventing coronary heart disease and diabetes.

Our results suggested that childhood overweight and obesity, determined by both the CDC Reference and IOTF Reference BMI cut-off points, had a high specificity (84% to 99%) for predicting obesity and metabolic disorders in young adulthood. That is, only a small percentage of the participants who were in the low-risk groups in young adulthood were overweight or obese as children. However, childhood overweight and obesity, determined by both methods, had a relatively low sensitivity (6% to 46%) for predicting obesity and metabolic disorders in young adulthood. That is, a number of the Bogalusa participants developed obesity and metabolic disorders in young adulthood even though they had a healthy BMI as a child. This observation is consistent with population statistics that have shown a substantial increase in the prevalence of adult obesity in the U.S. between the 1970s (when the baseline data for this study were collected) and the 1990s (when the follow-up data for this study were collected) (30). The increase in adult obesity may have artificially lowered the sensitivity of the CDC and IOTF references. The finding of a low sensitivity is also in line with population-based statistics in the U.S., which indicate that the prevalences of obesity (31, 32) and metabolic disorders such as hypertension (33, 34) are higher in adults than in children. Thus, although childhood overweight and obesity are strong risk factors for obesity and metabolic disorders in young adulthood, adiposity status during childhood does not account for 100% of the adult risk, and clearly a number of non-overweight children will develop obesity and heath problems in the adult years.

It is noteworthy that the two methods employed in this study for defining adiposity status in childhood provide slightly different overweight and obesity prevalences. For example, 25.0% of the 15- to 17-year-old girls from NHANES III were overweight (pre-obese + obese) based on the IOTF Reference, whereas 23.0% were overweight based on the CDC Reference (35). The corresponding prevalences of obesity in that nationally representative sample were 6.6% and 8.7%, respectively (35). In the present study, we also reported that a slightly larger percentage of the participants were classified as overweight in childhood based on the IOTF Reference compared with the CDC Reference (11.1% vs. 9.9%), whereas a slightly smaller percentage of the participants were classified as obese in childhood based on the IOTF Reference compared with the CDC Reference (4.3% vs. 6.4%). Moreover, we found that a larger percentage of those who were obese at 4 to 15 years of age based on the IOTF Reference were obese 14 to 24 years later as young adults in comparison with those who were obese at 4 to 15 years of age based on the CDC Reference (80.6% vs. 73.6%). The implication of these observations is that adiposity status and associated health risk will be classified differently in a small percentage of children depending on whether the CDC Reference or IOTF Reference BMI cut-off points are employed. This may have a significant impact in the clinical setting for those patients who are classified differently within the two systems (e.g., obese based on CDC Reference but overweight based on IOTF Reference) but would have little or no impact on statistical findings in the research setting.

There are several strengths to this study, such as the large number of clinical risk factor measurements, the longitudinal nature of the study, and the use of a large biracial sample of boys and girls ranging from 4 to 15 years of age. The study population was, however, a non-representative sample of youth, and additional studies in other populations are needed to support the results reported here. Furthermore, we were unable to control for physical activity level and socioeconomic status, two variables that are related to obesity and related comorbid conditions, in our analyses. Finally, our analyses are based entirely on cardiovascular risk factors, and childhood obesity is also related to a number of other important outcomes such as bullying (36) and social stigmatization (37).

The findings of this study underscore the importance of weight control early in life and support the view that BMI should be employed in the clinical setting to evaluate the presence of elevated health risk in children and adolescents. This is particularly true given the high prevalences of overweight and obesity among children and adolescents in industrialized countries (9, 35, 38). The CDC Reference and IOTF Reference BMI cut-off points are commonly used and are effective guidelines for assessing adiposity status in children and adolescents. Our results suggest that neither of these approaches has a clear advantage over the other. Both the CDC and IOTF classification systems have advantages and limitations and should be employed with that in mind. Future studies are needed to further refine the methods for defining overweight and obesity in children and adolescents.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

The Bogalusa Heart Study is a joint effort of many investigators and staff whose contributions are gratefully acknowledged. This research was supported by Heart and Stroke Foundation of Ontario Grant T4946; by National Heart, Lung, and Blood Institute Grant HL-38844; by National Institutes of Child and Health Development Grant HD-043820; and by National Institute on Aging Grant AG-16592. The funding organizations had no role in the design, conduct, interpretation, analysis, review, or approval of the manuscript. I.J. was funded by a postdoctoral fellowship from the Canadian Institutes of Health Research, and C.B. is partially supported by the George A. Bray Chair in Nutrition. Conflicts of interest for C.B. include: Baylor Children's Nutrition Research Center or U.S. Department of Agriculture (external advisory board member), Boston Obesity and Nutrition Research Center (external advisory board member), Bristol Myers Squibb (unrestricted grant), Mars Inc. (Nutritional Research Council Member), Sanofi-synthelabo (honorarium), The Cooper Institute for Aerobic Research (scientific advisory board member), and Weight Watchers International (scientific advisory board member).

Footnotes
  • 1

    Nonstandard abbreviations: NHANES, National Health and Nutrition Examination Survey; CDC, Centers for Disease Control and Prevention; IOTF, International Obesity Task Force; BP, blood pressure; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio; AUC, area under the receiver operating characteristic curve.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References
  • 1
    Cole, T. J., Bellizzi, M. C., Flegal, K. M., Dietz, WH. (2000) Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 320: 12401243.
  • 2
    Kuczmarski, R. J., Ogden, C. L., Guo, S. S., et al (2000) CDC Growth Charts for the United States: methods and development. Vital Health Stat 11 2002: 1190.
  • 3
    Himes, J. H., Dietz, WH. (1994) Guidelines for overweight in adolescent preventive services: recommendations from an expert committee: The Expert Committee on Clinical Guidelines for Overweight in Adolescent Preventive Services. Am J Clin Nutr. 59: 307316.
  • 4
    Barlow, S. E., Dietz, WH. (1998) Obesity evaluation and treatment: Expert Committee recommendations: The Maternal and Child Health Bureau, Health Resources and Services Administration and the Department of Health and Human Services. Pediatrics 102: E29
  • 5
    Rolland-Cachera, M. F., Deheeger, M., Guilloud-Bataille, M., Avons, P., Patois, E., Sempe, M. (1987) Tracking the development of adiposity from one month of age to adulthood. Ann Hum Biol. 14: 219229.
  • 6
    Garn, S. M., Clark, DC. (1975) Nutrition, growth, development, and maturation: findings from the ten-state nutrition survey of 1968-1970. Pediatrics 56: 306319.
  • 7
    Must, A., Dallal, G. E., Dietz, WH. (1991) Reference data for obesity: 85th and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thickness. Am J Clin Nutr. 53: 839846.
  • 8
    World Health Organization Expert Committee (1995) Physical status: the use and interpretation of anthropometry: report of a WHO Expert Committee. WHO Tech Rep Ser. 854: 1452.
  • 9
    World Health Organization (1998) Obesity: Preventing and Managing the Global Epidemic World Health Organization Geneva, Switzerland.
  • 10
    Bellizzi, M. C., Dietz, WH. (1999) Workshop on childhood obesity: summary of the discussion. Am J Clin Nutr. 70: 173S175S.
  • 11
    Freedman, D. S., Dietz, W. H., Srinivasan, S. R., Berenson, GS. (1999) The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa Heart Study. Pediatrics 103: 11751182.
  • 12
    Morrison, J. A., Barton, B. A., Biro, F. M., Daniels, S. R., Sprecher, DL. (1999) Overweight, fat patterning, and cardiovascular disease risk factors in black and white boys. J Pediatr. 135: 451457.
  • 13
    Katzmarzyk, P. T., Tremblay, A., Perusse, L., Despres, J. P., Bouchard, C. (2003) The utility of the international child and adolescent overweight guidelines for predicting coronary heart disease risk factors. J Clin Epidemiol. 56: 456462.
  • 14
    Allain, C. C., Poon, L. S., Chan, C. S., Richmond, W., Fu, PC. (1974) Enzymatic determination of total serum cholesterol. Clin Chem. 20: 470475.
  • 15
    Bucolo, G., David, H. (1973) Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem. 19: 476482.
  • 16
    Srinivasan, S. R., Berenson, GS. (1983) Serum lipoproteins in children and methods for study. In: Lewis, LA eds. CRC Handbook of Electrophoresis: Lipoprotein Methodology and Human Studies 3: 185204. CRC Boca Raton, FL.
  • 17
    Srinivasan, S. R., Bao, W., Wattigney, W. A., Berenson, GS. (1996) Adolescent overweight is associated with adult overweight and related multiple cardiovascular risk factors: the Bogalusa Heart Study. Metabolism 45: 235240.
  • 18
    National Institutes of Health National Heart Lung and Blood Institute (1998) Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Obes Res. 6: S51S210.
  • 19
    Chobanian, A. V., Bakris, G. L., Black, H. R., et al (2003) The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 Report. JAMA 289: 25602571.
  • 20
    Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults (2001) Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 285: 24862497.
  • 21
    Hanley, J. A., McNeil, BJ. (1983) A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 148: 839843.
  • 22
    Srinivasan, S. R., Myers, L., Berenson, GS. (2002) Predictability of childhood adiposity and insulin for developing insulin resistance syndrome (syndrome X) in young adulthood: the Bogalusa Heart Study. Diabetes 51: 204209.
  • 23
    Freedman, D. S., Khan, L. K., Serdula, M. K., Dietz, W. H., Srinivasan, S. R., Berenson, GS. (2004) Inter-relationships among childhood BMI, childhood height, and adult obesity: the Bogalusa Heart Study. Int J Obes Relat Metab Disord. 28: 1016.
  • 24
    Freedman, D. S., Srinivasan, S. R., Valdez, R. A., Williamson, D. F., Berenson, GS. (1997) Secular increases in relative weight and adiposity among children over two decades: the Bogalusa Heart Study. Pediatrics 99: 420426.
  • 25
    Srinivasan, S. R., Myers, L., Berenson, GS. (2001) Rate of change in adiposity and its relationship to concomitant changes in cardiovascular risk variables among biracial (black-white) children and young adults: The Bogalusa Heart Study. Metabolism 50: 299305.
  • 26
    Coughlin, S. S., Metayer, C., McCarthy, E. P., et al (1995) Respiratory illness, beta-agonists, and risk of idiopathic dilated cardiomyopathy: The Washington DC, Dilated Cardiomyopathy Study. Am J Epidemiol. 142: 395403.
  • 27
    Valdez, R., Greenlund, K. J., Wattigney, W. A., Bao, W., Berenson, GS. (1996) Use of weight-for-height indices in children to predict adult overweight: the Bogalusa Heart Study. Int J Obes Relat Metab Disord. 20: 715721.
  • 28
    Dietz, WH. (1998) Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics 101: 518525.
  • 29
    Janssen, I., Katzmarzyk, P. T., Boyce, W. F., Pickett, W. (2004) The independent influence of physical inactivity and obesity on health complaints in 6th to 10th grade Canadian youth. J Phys Activity Health 1: 331343.
  • 30
    Flegal, K. M., Troiano, RP. (2000) Changes in the distribution of body mass index of adults and children in the US population. Int J Obes Relat Metab Disord. 24: 807818.
  • 31
    Flegal, K. M., Carroll, M. D., Ogden, C. L., Johnson, CL. (2002) Prevalence and trends in obesity among US adults, 1999-2000. JAMA 288: 17231727.
  • 32
    Ogden, C. L., Flegal, K. M., Carroll, M. D., Johnson, CL. (2002) Prevalence and trends in overweight among US children and adolescents, 1999-2000. JAMA 288: 17281732.
  • 33
    Hajjar, I., Kotchen, TA. (2003) Trends in prevalence, awareness, treatment, and control of hypertension in the United States, 1988-2000. JAMA 290: 199206.
  • 34
    National High Blood Pressure Working Group on Hypertension Control in Children (1996) Update on the 1987 task force on high blood pressure in children and adolescents: a working group report from the National High Blood Pressure Education Program. Pediatrics 98: 650658.
  • 35
    Flegal, K. M., Ogden, C. L., Wei, R., Kuczmarski, R. L., Johnson, CL. (2001) Prevalence of overweight in US children: comparison of US growth charts from the Centers for Disease Control and Prevention with other reference values for body mass index. Am J Clin Nutr. 73: 10861093.
  • 36
    Janssen, I., Craig, W. M., Boyce, W. F., Pickett, W. (2004) Associations between overweight and obesity with bullying behaviors in school-aged children. Pediatrics 113: 11871194.
  • 37
    Goldfield, A., Chrisler, JC. (1995) Body stereotyping and stigmatization of obese persons by first graders. Percept Mot Skills 81: 909910.
  • 38
    Lissau, I., Overpeck, M. D., Ruan, W. J., Due, P., Holstein, B. E., Hediger, ML. (2004) Body mass index and overweight in adolescents in 13 European countries, Israel, and the United States. Arch Pediatr Adolesc Med. 158: 2733.