| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Medicine (C.S., G.J.E., J.H.P.) and Pediatrics (R.R.S.), Indiana University, Purdue University at Indianapolis, and the Richard L. Roudebush Veterans Administration Medical Center (J.H.P.), Indianapolis, Indiana 46202
Address all correspondence and requests for reprints to: Dr. R. Ravi Shankar, Indiana University Medical Center, Division of Endocrinology and Metabolism, 541 N. Clinical Drive, Indianapolis, Indiana 46202-5111. E-mail: rshankar{at}iupui.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
95th] in children has become a major public health problem. The age when overweight begins and how it progresses are mostly unknown. Such information would be important for the optimal timing of prevention. We conducted a survival analysis on time to overweight and compared survival curves by race and sex. Data from a cohort of 924 children recruited from schools in Indianapolis, IN, were analyzed. Blacks were at greater risk for becoming overweight than whites. Similar findings were obtained when at risk of overweight (BMI percentile 85th and <95th) and overweight were considered as a single category. Twenty-five percent of blacks were overweight or at risk of overweight at or before age 7 yr, whereas it was age 11 yr in white females and age 10 yr in white males when 25% became overweight or were at risk of becoming overweight. The overall overweight-free survival curve for black females was significantly different from that for white females (P < 0.001), and black males were significantly different from white males (P = 0.04). There was no sex difference. The time to overweight during childhood and adolescence varies by race, indicating the need for race-specific timing of interventions. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
95th] has become increasingly common (1). Overweight, even at this early age, poses a real risk for diabetes, hypertension, and dyslipidemia (2, 3, 4). A high prevalence of the metabolic syndrome was recently reported for overweight children and adolescents (5). In addition, overweight children are more likely to be obese adults (6, 7, 8, 9, 10); the odds ratio for adult obesity is more than 20 times greater for overweight children than for nonoverweight children (11). If the current trend toward childhood and adolescence overweight continues, it is logical to anticipate in adults an even greater increase in overweight-related health problems than already exists. Because sustainable weight loss is not achieved by most obese adults (12), preventing and treating obesity much earlier may be useful in reversing the current trends. Although the Bogalusa Heart Study (13) showed that in both children and adults, black females had a greater rate of increase in adiposity than white females, to our knowledge, no prospective cohort studies have been performed to identify the age(s) at which overweight begins and how it progresses during childhood and adolescence. Such information could assist with the timing of interventions aimed at prevention (14, 15). Interventions designed to reduce the prevalence of overweight may differ based on race and sex; thus, we also looked for an influence of these two factors on the age when overweight first appears.
We estimated the probability of becoming overweight at or before age 6 yr through age 18 yr in a cohort of children studied prospectively with frequent measurements of weight and height. We used survival analyses to compare time for becoming overweight and a statistical method that took into account individuals who were overweight before entering the study. A similar approach was used to analyze time to overweight or at risk of overweight (BMI percentile 85th and <95th) as a single category.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The subjects were participants in a longitudinal study of blood pressure in which anthropometric data were collected. They were recruited from schools in Indianapolis from 1985–2001, as described previously (16). Most were from two large public school systems, one more representative of a lower socioeconomic status and the other of a more affluent population. All eligible students between ages 5 and 18 yr were invited to participate. The exclusion criteria were history of renal or cardiac disease, hypertension, or diabetes mellitus. Informed consent was obtained from each subject and his or her parent or guardian. The institutional review board of Indiana University-Purdue University at Indianapolis approved the study. Measurements were made at the schools every 6 months; at less than 5% of the visits subjects had measurements made in an out-patient facility of the General Clinical Research Center. On the average, each subject had 8.7 ± 7.0 (mean ± SD) measurements (range, 1–27). There was no difference between males and females in the number of measurements, but whites had, on the average, 2.6 more sets of measurements than blacks. However, all four groups had a similar number of subjects across the age range of 6–18 yr (P = 0.23). There were 187 black and 267 white males, and 210 black and 260 white females.
A digital electronic scale (Measurements Specialties, Inc., Huntsville, AL) was used to measure weight, and an anthropometer (GPM Anthropological Instruments, Zurich, Switzerland) was used to measure height. BMI [weight (kilograms)/(height (meters)2] was used to compute the age- and sex-adjusted BMI percentile. Definitions for normal weight (BMI percentile <85), at risk of overweight, and overweight are based on the 2000 Centers for Disease Control and Prevention growth charts for the United States (17). The Committee on Clinical Guidelines for Overweight in Adolescent Preventive Services recommended the use of similar definitions in adolescents (18).
Statistical methods
SAS codes from the Centers for Disease Control (www.CDC.gov) were used to compute the age- and sex-adjusted BMI percentile. Descriptive statistics (mean ± SD) were computed to describe the baseline characteristics of the study subjects and were modeled as a function of age, race, and sex to compare these variables between blacks and whites as well as between males and females. An iterative procedure based on the product-limit method of Kaplan and Meier that takes into account double censoring (left and right) was used to compute the overweight-free survival curves (19, 20). This method includes subjects with one observation available, because for estimating the probability of becoming overweight at or before a specific time point, the method requires data only at that specific time point. Subjects were left-censored if they were already overweight at the time of their entry into the study and were right-censored if they were lost to follow-up, dropped out of the study, or were not overweight by age 18 yr. The distributions of age for the left-censored subjects were similar for the four race- and sex-based groups (P = 0.23). Survival curves for overweight generated for the four groups (black males, white males, black females, and white females) were then compared using a generalized two-sample Wilcoxon test for doubly censored data (21, 22). Similar procedures were used to generate and compare survival curves using overweight or at risk of overweight as one category (BMI percentile 85th). Programs to estimate and compare the survival curves were written in SAS and verified for accuracy using published examples (19, 22).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Table 1
depicts the baseline characteristics of subjects. The males and females were of comparable age (P = 0.94), although the whites were slightly younger than the blacks (P < 0.001) and had a lower age-adjusted height (P < 0.001). The blacks were heavier, as assessed by age-adjusted weight (P < 0.001), BMI (P < 0.001), and BMI percentile (P < 0.001). The females and males had comparable weights (P = 0.26) and BMI (P = 0.88). The birth weights of males were significantly greater than those of females (P < 0.001), and the birth weights of whites were significantly greater than those of blacks (P < 0.001). The parents of black children had less formal education than the parents of white children (P < 0.001). Although the reported median family income in the white families was more than twice that for the blacks, the median income for blacks was similar to the median income for blacks at the national level in 1986, the median subject enrollment time of our study (23). The median family income of the white participants was slightly higher than the median income of whites at the national level. Thus, the samples used appeared to be reasonably representative of the general population.
|
The overweight-free survival curves for each of the four groups (black males, black females, white males, and white females) are presented in Fig. 1. Over the age range of 6–18 yr, the blacks were at greatest risk for becoming overweight. The blacks were also more likely to become overweight earlier than the whites. Within each race group, males and females had similar risks of being overweight, but after the age of 11 yr, white males were at greater risk than white females. Twenty-five percent of black females vs. only 10% of white females became overweight by age 10 yr. Similarly, 25% of black males became overweight by age 13 yr, but it was age 16 yr when 25% of white males became overweight. The rate for becoming overweight in the preadolescent period (<12 yr of age) was 28% in black females, 22% in black males, 11% in white females, and 17% in white males.
|
Time to overweight or at risk of overweight
Time to the development of overweight or at risk of overweight was also analyzed in all four groups. Figure 2 shows the overweight or at risk of overweight-free survival curves for the four race- and sex-based groups. In these analyses, the survival curves for black females and males were very similar. However, there was a consistent difference between white females and males. Once again, the blacks were likely to become overweight or at risk of overweight earlier than the whites. White females had the lowest risk of being at risk of overweight or overweight. Twenty-five percent of black females and males were already at risk of overweight or overweight by age 7 yr, whereas 25% of white females and males were at risk of overweight or overweight by age 11 and 10 yr, respectively. The median age for becoming overweight or at risk of overweight (i.e. the age when 50% of the group became overweight or at risk of overweight) was 15 yr for black females and 14 yr for black males. Less than 50% of whites became overweight or at risk of overweight before the age of 18 yr. The rate of becoming at risk of overweight or overweight before adolescence (<12 yr of age) was 41% for black females, 44% for black males, 29% for white females, and 35% for white males.
|
Socioeconomic status and risk of becoming overweight
The risk of becoming overweight during the subject’s follow-up period was modeled by logistic regression. Blacks had significantly higher odds of becoming overweight even after adjustments were made for income, birth weight, father’s education, and mother’s education. However, there was no difference between males and females. Higher income and lower birth weight were associated with a lower risk of becoming overweight. Father’s education had a significant inverse association with the risk of being overweight, but mother’s education did not have any significant effect on the risk of becoming overweight.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
It has been reported that, in general, there are no racial differences in the risk of being overweight between prepubescent black females and age-matched white females (26, 27), but by early adulthood, black females are significantly heavier than white females (28, 29). In addition, the Bogalusa Heart Study (13) showed that in both children and adults, black females had a higher rate of change in adiposity than white females, but no race difference was observed for males. These previous data suggested that the racial disparity observed in adult females occurs during adolescence. However, in our cohort of subjects, we observed a significant racial difference for both males and females in the time to overweight or at risk of overweight at a much earlier age. The difference became further magnified during adolescence in females, but not in males.
Our study was primarily descriptive in nature, and we can only speculate on the mechanisms for the racial differences in our observations. A multitude of variables can contribute to racial differences in the rate of weight gain and the prevalence of overweight in children and adolescents, such as differences in resting energy expenditure (30), caloric intake and patterns of physical activity (31, 32), socioeconomic status (33), and maternal attitudes toward breastfeeding (34). A logistic regression analysis on the risk of becoming overweight showed that blacks had significantly higher likelihood of becoming overweight even after adjustments were made for family income, birth weight, and parents’ level of education.
A limitation to the present study was that the subjects’ voluntary participation might have introduced an element of ascertainment bias. One might anticipate, however, that heavier children would be less inclined to volunteer for a study in which repeated measurements of body size would take place and, if anything, the selection of subjects might result in an underestimation of the prevalence of obesity. An additional limitation was that we did not have information on pubertal staging; an earlier onset of puberty, which is more typical in blacks (35), could contribute to an earlier emergence of weight gain.
Because the condition of childhood overweight is not a benign problem, appropriate preventive measures are now recognized as being exceedingly important. To date, data from the Child and Adolescent Trial for Cardiovascular Health suggest that although children in the intervention group reported more physical activity and better food choices, there were no objective differences in anthropometric measures between the intervention and control schools (36). Similarly, data from the pilot trials for the Girls Health Enrichment Multi-Site Studies suggest that although short-term (12 wk) interventions after school result in increased physical activity and improved food choices, changes in BMI did not reach statistical significance (37, 38, 39, 40). The lack of impact on weight in the Child and Adolescent Trial for Cardiovascular Health may be a consequence of the fact that recruits were older (9 yr). For the Girls Health Enrichment Multi-Site Studies, black girls between the ages of 8–10 yr have been enrolled in a study to prevent overweight. However, by that age, in our study more than 25% of black girls had already become overweight or at risk of overweight, and prevention of overweight in the remainder may not be feasible, because by this age patterns that lead to becoming overweight may already be set.
Our study highlights the fact that blacks become overweight at much earlier ages than whites and suggests that interventions directed at prevention should begin at an earlier age in blacks than is generally appreciated. Programs to prevent overweight in black children might best begin in infancy (possibly, they should be included in the anticipatory guidance provided to families at well-baby visits in infancy) and continue through the school-age years. The promising nature of early intervention is borne out by the results of a pilot study using home visits to counsel overweight and obese mothers of toddlers that reported a trend toward lower weight for height z-scores as early as wk 16 in the group that received counseling on obesity prevention compared with an increase in the control group (41). Currently, there may not be any programs directed at preventing overweight in preschool-aged black children. In contrast, because whites become overweight at a slower rate and later than blacks, programs catering to a predominantly white population could afford to wait until later to start focusing on preventive measures.
|
Footnotes |
|---|
First Published Online February 22, 2005
Abbreviation: BMI, Body mass index.
Received November 10, 2004.
Accepted February 10, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Find additional patient-related information at:
This article has been cited by other articles:
 |
|
 |
|
  Disparity Starts Early for Obesity Risk DOC News, August 1, 2005; 2(8): 14 - 14. [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
