International Journal of Pediatric Obesity.

2009; 4: 150
159

ORIGINAL ARTICLE

Cardiovascular disease risk factor (CVDRF) associated waist

circumference patterns in obese-prone children

KAREN L. LEIBOWITZ1, RENEÉ H. MOORE2, ALBERT J. STUNKARD2, VIRGINIA A.

STALLINGS 1,2 , ROBERT I. BERKOWITZ 1,2 , NICOLAS STETTLER 1,2 , JESSE L.
CHITTAMS2 & MYLES S. FAITH1,2
1
The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, 2University of Pennsylvania School of Medicine,
Philadelphia, Pennsylvania

Abstract
Objectives. The present study tested whether children born at high risk (HR) compared with low risk (LR) for obesity are
more likely to have a waist circumference (WC) associated with cardiovascular disease risk factors (CVDRF-WC) and tested
whether CVDRF-WC status tracks over time. Methods. This prospective cohort study involved 71 children, three to eight
years, who were divided into two groups, LR (n37) and HR (n34), based upon maternal prepregnancy body mass index
(BMI). HR subjects were subdivided into HR normal-weight (HRNW) and HR overweight (HROW) groups, based on
BMI ]85%. Children were classified as having or not having a CVDRF-WC at each year, using age- and gender-specific
WC cut-offs. Anthropometry was assessed annually. Results. Although HR children had a significantly greater mean WC
than LR children at years 5
8 (p B0.03), these differences became non-significant after adjusting for BMI. HROW were
more likely to have a CVDRF-WC status (p 50.0001) at age 4 years (10%, 5%, vs. 58%), 5 years (3%, 10%, vs. 60%), 6
years (0%, 0%, vs. 70%), 7 years (0%, 0%, vs. 50%) to 8 years (0%, 0%, vs. 55%) than LR and HRNW. Although 60
100%
of the children tracked CVDRF-WC status, higher proportions of HROW children (0
40%) transitioned into having a
CVDRF-WC, compared with LR (0
6%) and HRNW (0
9%). Conclusions. HROW were more likely to have or develop a
CVDRF-WC. Although the effects of obesity risk on WC may be secondary to BMI, clinically assessing WC in obese-prone
children may help identify youth at risk for obesity-related complications.

Key words: Obesity, children, waist circumference, obesity predisposition, cardiovascular disease, BMI

Introduction BMI is the most appropriate screening measurement

There is increasing interest in waist circumference
(WC) as a screening tool for central adiposity in
children (1
3) and potentially for identifying risk for
obesity-related complications, such as cardiovascular
disease (CVD) (4
7). WC was significantly asso-
ciated with fat distribution in children (2,3) and, in
adults, was more strongly associated with obesity-
related health risk than body mass index (BMI, kg/
m2) (8
10).
In clinical settings, BMI percentiles or z-scores are
often used as a screening tool for identifying over-
weight children, because it predicts adult weight
status and provides an estimate of total body fat
(11,12). However, there are questions as to whether
in children for determining overweight status and
risk for obesity-related co-morbidities during child-
hood and adulthood. First, BMI measures both fat
and fat free mass and thus may not be sensitive to the
shifts in proportions of fat and fat free mass during
growth (13
16). Second, WC may be the better
indicator of CVD risk as compared with BMI. Some
studies have found WC alone to be the better
predictor of adverse CVD risk profile (5
7,17),
while others suggest that WC and BMI combined
provide the best prediction model (5,18). However,
it is questionable as to whether WC is uniquely
informative in predicting disease risk beyond BMI in
children. For example, an analysis of the Bogalusa

Correspondence: Karen L. Leibowitz, The Children’s Hospital of Philadelphia, Division of Gastroenterology and Nutrition, 34th Street and Civic Center
Boulevard, 7th Floor Main Building, Philadelphia, PA 19104-4399, USA. Fax: 1 732 325 9335. E-mail: peddm@aol.com

(Received 14 April 2008; accepted 7 October 2008)

ISSN Print 1747-7166 ISSN Online 1747-7174 # 2009 Informa UK Ltd
DOI: 10.1080/17477160802596130

Heart Study found that BMI-for-age and waist-to-
height ratio did not differ in their ability to predict
children with adverse cardiovascular risk profiles
(19).
CVD risk factor associated WC (CVDRF-WC)
refers to a WC that is associated with increased CVD
risks, such as dyslipidemia, hypertension and meta-
bolic syndrome. Various studies proposed pediatric
CVDRF-WC cut-offs in order to identify the chil-
dren who need evaluation and management for
obesity and its co-morbidities (17,20
25). While
there have been different approaches in determining
cut-offs, one noteworthy study developed age- and
gender-specific cut-offs for CVDRF-WC using data
from the National Health and Nutrition Examina-
tion Survey (NHANES) cohort (21), a nationally
representative sample.
The main objective of the present study was to test
whether children born with a familial predisposition
to obesity compared with those without are more
likely to have a CVDRF-WC. Although a familial
predisposition to obesity is an established risk factor
for an elevated weight and BMI status in children
(26
29), few studies have tested this association to
the child’s WC (1,30). In the present investigation,
infants born at low risk (LR) or high risk (HR) for
obesity based on maternal pre-pregnancy BMI were
enrolled and followed until eight years of age. The
main aim of the study was to compare the WC values
of HR and LR children at ages three to eight years,
thereby testing the hypothesis that the HR group will
have higher mean WC levels than the LR group. As
part of this main aim, we compared the proportion
of HR and LR children having a CVDRF-WC,
testing the hypothesis that a higher percentage of
HR than LR children will have a CVDRF-WC. An
exploratory aim was to test whether CVDRF-WC
status tracks, or remains at the same status, over
time.
Methods
Participants
The Infant Growth Study (31,32) is an ongoing
longitudinal investigation that was designed to assess
anthropometric, behavioral and metabolic determi-
nants of growth and development in children with
and without maternal risk for obesity. Eighty-two
healthy full-term Caucasian infants were initially
enrolled in the Infant Growth Study (33), after
which there was no further recruitment of additional
families. Thus, the families analyzed in the present
study were part of the initial enrollment cohort.
Although not all children attended each of the
annual assessments from years 3
8 (see Table II for
CVDRF-waist circumference and obesity 151
N’s at each year), no subjects formally dropped out
of the study during this period. Of the original 41 LR
families, 85%, 73%, 88%, 88%, 90% and 90% were
assessed at years 3
8, respectively, in this report. Of
the original 41 HR families, 76%, 80%, 78%, 80%,
83%, and 71% were assessed at years 3
8, respec-
tively. Children were designated as being HR or LR
for obesity based on their mothers’ pre-pregnancy
BMI status. Specifically, mothers of the HR children
had a BMI 66% and those of the LR children had
a BMI B33% based on nationally normative cut-
offs from the NHANES sample (34); mothers with
BMI between 33
66% were not enrolled during the
initial recruitment. Out of the 34 HR mothers
assessed in this report, 32 mothers had a prepreg-
nancy BMI greater than 25 (94.1%), whereas all 37
LR mothers had BMI below 25. Thus, although
mothers’ overweight/obesity status per se was not
used to define child risk group, the vast majority of
HR mothers were overweight/obese and none of the
LR mothers were overweight/obese.
Independent of obesity risk status (i.e., HR or LR),
which was based on mothers’ prepregnancy BMI,
children’s overweight status was determined based on
the children’s BMI percentiles from CDC growth
charts. Children’s height and weight were measured
at each year, thereby allowing for computation of
BMI, BMI z-score, and BMI percentiles from CDC
growth charts. Children with a BMI percentile ]
85th were considered overweight or obese (‘‘over-
weight’’ for the purpose of this report). We com-
bined all children above the 85th percentile in order
to capture both overweight and obese children; the
BMI cut-off of 85% for HROW children was chosen
based on the current recommendations from the
Expert Committee Recommendations Regarding the
Prevention, Assessment, and Treatment of Child
and Adolescent Overweight and Obesity: Summary
Report (35).
HR subjects were sub-classified into high risk
normal weight (HRNW) and high risk overweight
(HROW) at each year based upon child overweight
status (i.e., BMI below or above the 85% percentile),
because only a subset of HR children showed the
overweight phenotype at each respective year (31).
Because the LR group only had 1 or 2 subjects with a
BMI ]85% at certain ages, subdividing the LR
group would have yielded an insufficient number of
cases to analyze appropriately. Thus, we retained a
full LR group.
Maternal and infant enrollment criteria included:
1) infants born between 36 and 42 weeks gestational
age; 2) infants not small- or large- for gestational age;
3) mothers without history of gestational diabetes;
and 4) maternal age greater than 18 years. Anthro-
pometry was assessed annually on the subjects that
152 K. L. Leibowitz et al.
returned for a given study year. If triplicate measures
of weight, height and WC were not obtained or
reported, these subjects were not included in the
analyses, in order to ensure the accuracy of the data.
Additionally, subjects were required to have complete
data for at least two years, in order to be included in
the tracking analyses of CVDRF-WC status over
time. The number of children analyzed in the present
investigation by risk group and age are provided.
Procedures
Anthropometric assessments were performed at the
Growth and Nutrition Laboratory and the General
Clinical Research Center at the Children’s Hospital
of Philadelphia, starting in May 1996 and ending in
March 2003. During these visits, height, weight and
WC were obtained in triplicate by trained assistants,
using standardized techniques (36). Height and
weight were measured by a stadiometer (Holtain,
Crymych, United Kingdom) and digital scale
(model 6002; Scaletronix, Carol Strea, IL), respec-
tively. BMI (kg/m2) was calculated from the mean
height and mean weight. A trained anthropometrist
measured the WC with an inelastic tape that was
placed around the subjects at the level of the natural
waist, the narrowest portion of the torso, at the end
of expiration. The intra-measurer reliability of waist
circumference measures was 0.99 at each of the
assessment years.
Written informed consent was obtained from the
parents. The protocol was approved by the institu-
tional review boards of the University of Pennsylva-
nia and the Children’s Hospital of Philadelphia. All
applicable institutional and governmental regula-
tions concerning the ethical use of human volunteers
were followed during this research.
CVDRF-WC status
CVDRF-WC status (i.e., have vs. not have CVDRF-
WC) of the subjects was determined at years 3
8
using the WC threshold values established by Li et
al. (21). Using the NHANES data that were
obtained over four cycles between 1988 and 2004,
Li et al. (21) established age- and gender-specific
WC thresholds at the 90th percentile to define
CVDRF-WC values. Because this study was based
on a large sample of children from the United Sates
and was one of the few studies that had age- and
gender-specific WC cut-off values, Li et al. was
applied to this cohort. Furthermore, because appro-
priate WC cut-offs to identify a child at increased
risk for CVD have not been established from
national data, the WC cut-off of 90th percentile
was used from Li et al. based on prior research
showing that children with a WC ]90% were more
likely to have two or more of the CVD risk factors
compared with children with a WC B90% (25). In
this previous study, the CVD risk factors that were
used to establish the CVDRF-WC criteria included
total cholesterol, low-density lipoprotein cholesterol,
high-density lipoprotein cholesterol, ApoA1/ApoB
and blood pressure.
WC was coded as either zero (no CVDRF-WC) or
one (presence of CVDRF-WC) at each year, de-
pending on whether the WC value fell above or
below the WC threshold. As this criterion used age-
and gender-specific thresholds for CVDRF-WC, age
and gender were not adjusted for via inclusion as
independent covariates in these statistical analyses.
Data analysis
At each year (ages three to eight), mean differences
in height, weight, BMI, and BMI z-score between
the two groups were examined using t-tests or the
non-parametric Wilcoxon rank-sum test. To test the
WC differences between the HR and LR subjects, a
two (Risk Groups)six (Years) repeated measures
mixed model analysis was employed (37,38). This
procedure allows for the comparison of risk groups
while examining WC as a continuous measure and
adjusting for the correlation among repeated mea-
surements. We then re-ran this analysis controlling
for child BMI z-score (at each year) and the
interaction between child BMI z-score and year.
This analysis allowed us to examine the extent to
which any effect of risk group on child waist
circumference may be secondary to child BMI z-
score.
A three (Risk Groups) six (Years) repeated
measures analysis also was conducted to test for
differences in WC when classifying children accord-
ing to the three risk groups (LR, HRNW, HROW).
Due to the collinearity between the two variables,
maternal BMI and child risk status were not
included in the same model simultaneously as
predictors. Least squares means and standard errors
were generated at each time point for both mixed
models.
To test whether the proportion of children with a
CVDRF-WC differed by risk group, generalized
estimating equations (GEE) (38) and Fishers Exact
Test (39) were used. As GEE uses all available data
and takes into account the correlation among long-
itudinal repeated measurements, it is the preferred
method to analyze the proportion of children with a
CVDRF-WC. When GEE did not provide stable
estimates (e.g., small sample sizes), Fisher’s Exact
Tests were utilized. In addition to testing the main
comparison of HR and LR children at each year,
 CVDRF-waist circumference and obesity 153

additional analyses were conducted comparing the
three risk groups (LR, HRNW, HROW) in terms of
proportions with a CVDRF-WC.
To test whether the presence of CVDRF-WC
tracks over time, the percent of children who
remained in the same CVDRF-WC status versus
transitioned to the other status at later years was
computed. Specifically, long-term tracking was de-
fined as maintaining the same CVDRF-WC status at
each respective year prior to year 8 (i.e., ages three,
four, five, six, and seven) to the CVDRF-WC status
at year 8. The number and percentage of children
who transitioned from not having a CVDRF-WC to
having a CVDRF-WC (‘‘No Tracking
UP’’) or vice
versa (‘‘No Tracking
DOWN’’) were also com-
puted. Analyses were conducted by the three risk
groups (LR, HRNW and HROW).
All analyses were performed using the statistical
software package SAS, version 9.1. Significant alpha
level was considered to be B0.05 for all analyses.
Results
Sample characteristics
Table I provides the general characteristics and
anthropometric measurements for the two risk
groups at each study year. LR and HR mothers
had significantly different mean9standard deviation
(SD) BMIs of 19.591.1 and 30.594.4, respectively.
Weight and BMI did not significantly differ between
groups at year three. However, at year four and all
subsequent years, the HR group had significantly
higher mean BMI and mean weight measurements

Table I. Anthropometric measures of children by risk group and age (year).

Variable Low Risk: Mean (SD) High Risk: Mean (SD) p-value

Year 3
N 35 31
Height (cm) 94.4 (2.9) 95.20 (3.1) 0.325
Weight (kg) 14.0 (1.2) 15.80 (1.3) 0.313
BMI (kg/m2) 15.7 (0.9) 15.90 (1.5) 0.473
BMI Z-score 0.23 (0.8) 0.12 (1.3) 0.675
Year 4
N 30 33
Height (cm) 101.3 (3.4) 102.50 (3.9) 0.206
Weight (kg) 15.8 (1.3) 17.40 (2.8) 0.010
BMI (kg/m2) 15.6 (0.9) 16.50 (2.0) 0.036
BMI Z-score 0.01 (0.7) 0.54 (1.4) 0.082
Year 5
N 36 32
Height (cm) 108.5 (3.8) 109.80 (4.5) 0.221
Weight (kg) 18.2 (1.7) 20.40 (4.5) 0.009
BMI (kg/m2) 15.5 (0.9) 16.80 (3.0) 0.016
BMI Z-score 0.05 (0.7) 0.58 (1.4) 0.057
Year 6
N 36 33
Height (cm) 115.7 (4.3) 117.40 (5.2) 0.155
Weight (kg) 20.1 (1.9) 23.40 (6.4) 0.005
BMI (kg/m2) 15.0 (1.0) 16.80 (3.5) 0.006
BMI Z-score 0.33 (0.9) 0.37 (1.4) 0.016
Year 7
N 37 34
Height (cm) 121.8 (4.8) 122.90 (5.4) 0.354
Weight (kg) 22.9 (2.5) 27.40 (7.9) 0.002
BMI (kg/m2) 15.5 (1.1) 18.00 (4.2) 0.002
BMI Z-score 0.08 (0.8) 0.68 (1.2) 0.004
Year 8
N 37 29
Height (cm) 128.0 (5.0) 130.60 (5.3) 0.050
Weight (kg) 26.3 (3.2) 32.40(10.2) 0.002
BMI (kg/m2) 16.0 (1.3) 18.90 (5.0) 0.002
BMI Z-score 0.01 (0.8) 0.71 (1.2) 0.005

Note: BMIBody mass index; SD Standard deviation.

154 K. L. Leibowitz et al.
75
than the LR group (p B0.05). Height was higher in

Waist Circumference (cm)

the HR group only at year eight. Mean BMI z-score 70 LR HRNW HROW

differed at years four, six, seven, and eight. The 65

number of male and female subjects in the three risk 60

groups at each year is presented in Table II. 55
50

Risk Group differences in WC by age
The two (Risk Groups) six (Years) mixed model
analysis showed a significant main effect of risk
group (P B0.01) and a significant age-by-risk group
interaction (p 0.024). This indicates that the
increase in WC over time was greater for the HR
than the LR children. The mean WC of the HR
children was significantly greater than that of the LR
children at ages five (56.6 cm vs. 53.9 cm, p B0.03),
six (59 cm vs. 55.2 cm, pB0.02), seven (61.8 cm vs.
56.6 cm, pB0.005), and eight years (64.7 cm vs.
58.2 cm, pB0.002). Results did not change when
including gender as a covariate in the model. This
divergent WC trend between the two risk groups
began after a significantly steeper slope from year
three to four for HR WC compared with LR WC
(p 0.03). However, when re-running the full 2 6
mixed model adjusting for BMI z-score and its
interaction with year, the risk group main effect
(p 0.12) and age-by-risk group interaction (p  Figure 2. Because of the large number of cells having
0.25) for WC were no longer statistically significant.
Additionally, BMI z-score was found to have a
stronger correlation with WC than with risk group
status.
The three (Risk Groups)six (Years) mixed
model analysis indicated significant main effects
and interaction terms (all p B0.0001). The mean
WC of the LR and HRNW children did not
significantly differ at any year (p 0.16); however,
the mean WC of the HROW children was signifi-
cantly greater than that of the LR and HRNW
children at ages 3 years (p 0.04), 4 years (p  At age three, all three groups had a relatively small
0.006), 5 years (p 0.0002), 6 years (p B0.0001),
45
Yr 3 Yr 4 Yr 5 Yr 6 Yr 7 Yr 8
Age (Years)
Figure 1. Means with standard errors of waist circumference (cm)
by risk group and age.
Notes: LRLow Risk, HRNWHigh Risk Normal Weight,
HROWHigh Risk Overweight. The means with their standard
error bars were estimated from the mixed-model analysis. The
effects of risk, age, and the risk*age interaction were all significant
in the mixed-model analyses, each having a p-value B0.0001.
While there was no significant difference of the mean WC between
LR and HRNW groups, there was a significant difference of the
mean WC between HRNW and HROW groups (pB0.04).
7 years (p B0.0001) and 8 years (p B0.0001)
(Figure 1).
Risk Group differences in CVDRF-WC status by age
The proportion of children with a CVDRF-WC in
the three risk groups at each age is presented in
no children with a CVDRF-WC, the Fisher Exact
Test was used for the two and the three risk group
comparisons. When comparing the two risk groups,
the proportion of CVDRF-WC children was signifi-
cantly higher in the HR group than the LR group at
ages five to eight (p B0.01 for each year). The three
risk group comparison revealed significant differ-
ences in the proportion of children with a CVDRF-
WC at ages four to eight (p-value 50.0001 for each
year), with the proportions consistently highest
among HROW children (50 to 70% across years).
proportion of children with CVDRF-WC (0 to 13%),

Table II. Total number of subjects by risk group and gender at each year.

Number of Subjects

3 years 4 years 5 years 6 years 7 years 8 years

Risk group M F M F M F M F M F M F

LR 18 17 13 17 17 19 17 19 18 19 18 19
HR 14 17 16 17 15 17 16 17 16 18 15 14
HRNW 10 13 9 11 11 9 14 9 10 10 11 7
HROW 2 3 6 6 3 7 5 7 5 7 4 7

Notes: LRLow Risk; HRHigh Risk; HRNWHigh Risk Normal Weight; HROWHigh Risk Overweight. MMales, FFemales.
HR was determined based on the maternal prepregnancy BMI, while HRNW and HROW groups were established based on the subject’s
BMI at a given year. If a HR subject was missing weight or height for a particular year, that participant would not be able to be used in the
HRNW or HROW analyses. Therefore, the HRNW and HROW will not always add up to the total number of HR group in this Table.
 CVDRF-waist circumference and obesity 155
% with CVDRF-WC 70
LR HRNW HROW Tracking of CVDRF-WC status.
60
50
The proportion of children showing long-term
40 tracking of their CVDRF-WC status is presented
30 in Table III. The majority of the LR and HRNW
20 children tracked well, ranging from 84 to 100%
10 (Table III), with later years (i.e., closer to age 8)
0 tracking at higher percentages compared with earlier
Yrs 3 Yrs 4 Yrs 5 Yrs 6 Yrs 7 Yrs 8
Ages (Years) years (i.e., further away from age 8). For the HROW
group, the majority of children tracked across years
Figure 2. Proportion (%) of children with a cardiovascular disease (60 to 100%), although the proportion of tracking
risk factor associated waist circumference (CVDRF-WC) by age
and risk group. was generally smaller compared with LR and
Notes: CVDRF-WC cut-offs were age- and gender-specific that HROW. Once again, the proportion of HROW
corresponded to 90th percentile for WC using NHANES data children who showed tracking increased in later
(21). LRLow Risk, HRNWHigh Risk Normal Weight, years.
HROWHigh Risk Overweight. For years 3, 6, 7 and 8, the
missing bar indicates that there were no subjects with CVDRF-
Figure 3 displays the changes in raw WC levels,
WC in that specific risk group. from initial to follow-up assessment, for each HROW
child in each tracking analysis. The figures help to
and there were no differences among the three groups put these findings in context, as a change in only a
(p 0.38). However, from age three to four the few subjects can yield a relatively large percentage of
proportion of children with a CVDRF-WC in the children who do not track given the sample size.
HROW group jumped from 0% to 58% and subse-
quently ranged from 50 to 70% for ages five to eight Discussion
years. The LR and HRNW groups, by contrast, had
low proportions of CVDRF-WC at each of the years, Our main finding was that HROW children were
ranging from 0
13%. more likely to have an elevated WC and a WC that

Table III. Number (proportion) of children tracking and not tracking of cardiovascular disease risk factor associated waist circumference
(CVDRF-WC) status comparing age 3, 4, 5, 6 and 7 to age 8 years.

CVDRF-WC Tracking status

Years Risk group N Tracking* No tracking- UP** No tracking- DOWN***

3&8
LR 35 34 (97%) 0 ( 0%) 1 ( 3%)
HRNW 19 16 (84%) 2 (11%) 1 ( 5%)
HROW 5 3 (60%) 2 (40%) 0 ( 0%)
4&8
LR 30 27 (90%) 0 ( 0%) 3 (10%)
HRNW 18 16 (89%) 1 ( 6%) 1 ( 6%)
HROW 10 7 (70%) 1 (10%) 2 (20%)
5&8
LR 36 35 (97%) 0 ( 0%) 1 (3%)
HRNW 17 16 (94%) 0 ( 0%) 1 (6%)
HROW 9 7 (78%) 1 (11%) 1 (11%)
6&8
LR 36 36 (100%) 0 ( 0%) 0 ( 0%)
HRNW 20 20 (100%) 0 ( 0%) 0 ( 0%)
HROW 8 8 (100%) 0 ( 0%) 0 ( 0%)
7&8
LR 37 37 (100%) 0 ( 0%) 0 ( 0%)
HRNW 17 17 (100%) 0 ( 0%) 0 ( 0%)
HROW 10 9 (90%) 1 (10%) 0 ( 0%)

Notes: *Number and percent (%) of children who maintain the same CVDRF-WC status each respective age prior to age 8 years compared
with the status at age 8. ** Number and percent (%) of children who do not maintain the same CVDRF-WC status, transitioning from not
having a CVDRF-WC to having a CVDRF-WC. *** Number and percent (%) of children who do not maintain the same CVDRF-WC
status, transitioning from having a CVDRF-WC to not having a CVDRF-WC. LR Low Risk; HRNWHigh Risk Normal Weight;
HROWHigh Risk Overweight; WCWaist Circumference.
156 K. L. Leibowitz et al.
85
A B
105
80

Waist Circumference (cm)

Waist Circumference (cm)

75 95
70
85
65

60 75

55 65
50
55
45

40 45
Year 3 Year 8 Year 4 Year 8

C 105 D 105
100 100
Waist Circumference (cm)

Waist Circumference (cm)

95 95
90 90
85 85
80 80
75 75
70 70
65 65
60 60
55 55
Year 5 Year 8 Year 6 Year 8

E 105
Waist Circumference (cm)

100
95
90
85
80
75
70
65
60
55
Year 7 Year 8

Figure 3. Tracking of cardiovascular disease risk factor associated waist circumference CVDRF-WC (cm) for high risk overweight (HROW)
children at each year to year 8.
Notes: Panel A reflects tracking from year 3 to 8; Panel B reflects tracking from year 4 to 8; Panel C reflects tracking from year 5 to 8; Panel
D reflects tracking from year 6 to 8; Panel E reflects tracking from year 7 to 8. Each line represents one HROW subject. The thin straight
line indicates that that particular subject remained in the same CVDRF-WC status (either above or below the cut-off) for both that given
year and year 8. The dotted line represents a subject that transitioned up from below the CVDRF-WC cut-off at a given year to above the
cut-off at year 8. The bolded line indicates that a subject transitioned down from above the cut-off at a given year to below it at year 8.

was associated with an increased risk for CVD group mixed model analysis revealing that the
compared with LR and HRNW children. This HROW group had the greatest WC values, as well
significant divergence of WC patterns among risk as categorical analyses showing that HROW group
groups started as early as four years of age. While was most likely to have a CVDRF-WC. Thus, the
previous studies have shown that maternal prepreg- effects of risk group on WC per se may be secondary
nancy BMI or early pregnancy body weight was to child BMI z-score. Future research needs to
associated with an increased risk of overweight in the establish the extent to which there are independent
offspring (28,40
42), limited information on this genetic and environmental influences on child waist
relationship with the WC of children has been circumference independent of BMI. The findings
published (1). The significant finding in the WC of also underscore and reflect the high multicollinearity
the HROW group suggests that a familial predis- that exists among BMI and WC in children.
position to obesity in conjunction with a higher BMI The divergence of the mean WC between the risk
may lead to an increased risk for an elevated WC. groups followed a significantly steeper WC slope of
A noteworthy finding was that the effect of risk the HR group from three to four years of age. This
group on WC became non-significant when control- significant transition is similar to the concept of
ling for BMI z-score. This was consistent with the 3- ‘‘adiposity rebound’’ for BMI, defined as BMI

declining and reaching a minimum level around 5
6
years of age prior to gradually increasing to adult
BMI status. Furthermore, early adiposity rebound,
where the BMI nadir occurs before 5 years of age,
may be a marker or precursor to the onset of adult
obesity (43
46) although this issue remains contro-
versial (47). The transition in WC from year three to
four, observed in the HROW youth, may reflect a
similar phenomenon, such that this subgroup may
be more likely to develop adult obesity than the LR
or HRNW children. Future research examining
longer term tracking of WC should address this
issue.
To characterize the CVD risk status of the present
sample, CVDRF-WC cut-offs established from
NHANES were applied. The proportions of the
HROW children with a CVDRF-WC (50 to 70%)
were significantly higher than those in the LR (0 to
10%) and HRNW (0 to 10%) groups at ages four to
eight years. Moreover, the exploratory long-term
tracking analyses indicated that HROW children
were more likely to transition into having a
CVDRF-WC when compared with the other risk
groups, particularly during the transition from ages
three to four. The HROW group increased more in
the proportions of children with CVDRF-WC, from
0% at age three to 58% at age four. This further
supports the suggestion that this may be a critical
period for differentiation of WC in this sample of
high-risk children.
Most children tracked within a CVDRF-WC
status, which is similar to the tracking of BMI in
children in the literature (48,49). These results also
are consistent with two other studies that reported
moderate to strong tracking of WC in children over a
two and five year period (30,50), although these
studies utilized different approaches to analyze
tracking. When a child was above the CVDRF-WC
cut-off, even as early as four years of age, the
CVDRF-WC status tends to persist over a four-
year period. While it has been demonstrated that
childhood BMI is associated with BMI in young
adults (48,51,52), minimal data have addressed the
issue of body composition tracking during childhood
and adolescence (53). This topic warrants more
research. The tracking of CVDRF-WC suggests that
early intervention efforts may be important in
preventing the onset of elevated WC, especially
given the likelihood of CVDRF-WC status persisting
over time.
The present findings should be interpreted in light
of the substantial study limitations, which include a
relatively small sample, use of a single ethnic group,
lack of CVD risk factors for this cohort, and lack of
accepted standard of care criteria for CVDRF-WC
in children. In addition, although data on intra-rater
CVDRF-waist circumference and obesity 157
reliability were provided, data on inter-rater relia-
bility were not collected. Due to the small sample,
when subjects transitioned from one CVDRF-WC
status to another, a small number of subjects
transitioning could lead to a relatively large change
as a percentage. To further examine the relationship
between WC and CVDRF-WC, biomarkers are
currently being collected in this cohort for future
analyses. Additionally, the inconsistencies in the
literature, including the use of different types of
measuring tapes and various locations on the body to
measure the waist circumference, are important
limitations to this field. Further research is necessary
to establish a consistent method of measuring waist
circumference. Strengths of this study include use of
a longitudinal, prospective cohort study whose body
habitus has been carefully measured since three
months of age.
In conclusion, children who are at risk for
becoming overweight or obese, based upon familial
predisposition for obesity, were more likely to have
an increased WC, which may be associated with
CVD risk factors. This divergence of WC began
between three to four years of age, following a
significantly steeper slope of increasing WC from
ages three to four years between HR than LR
children. The effect of family risk on WC may be
secondary to BMI, and future research needs to
establish whether there are unique effects of WC on
disease risk independent of BMI. Still, WC may be
an important health screening measurement, in
conjunction with BMI, to identify children who are
at risk for obesity and CVD risk factors and who may
be in need of further evaluation and management.
Acknowledgements
This work was funded by a National Institute of
Health grant DK068899, General Clinical Research
Center grant RR00240, General Clinical Research
Center/Clinical Translational Research Center grant
UL1-RR-024134, and the Nutrition and Growth
Laboratory of Children’s Hospital of Philadelphia.
Declaration of interest: The authors report no
conflicts of interest. The authors alone are respon-
sible for the content and writing of the paper.
References
1. Garnett SP, Cowell CT, Baur LA, Shrewsbury VA, Chan A,
Crawford D, et al. Increasing central adiposity: the Nepean
longitudinal study of young people aged 7
8 to 12
13 y. Int J
Obes (Lond) 2005;29(11):1353
60.
/ /
2. Daniels SR, Khoury PR, Morrison JA. Utility of different
measures of body fat distribution in children and adolescents.
Am J Epidemiol. 2000;152(12):1179
84.
/ /
158 K. L. Leibowitz et al.
3. Taylor RW, Jones IE, Williams SM, Goulding A. Evaluation
of waist circumference, waist-to-hip ratio, and the conicity index
as screening tools for high trunk fat mass, as measured by
dual-energy X-ray absorptiometry, in children aged 3
19 y.
Am J Clin Nutr. 2000;72(2):490
5. / /
4. Cook S, Weitzman M, Auinger P, Nguyen M, Dietz WH.
Prevalence of a metabolic syndrome phenotype in adoles-
cents: findings from the third National Health and Nutrition
Examination Survey, 1988
1994. Arch Pediatr Adolesc Med.
2003;157(8):821
7.
/ /
5. Lee S, Bacha F, Gungor N, Arslanian SA. Waist circumfer-
ence is an independent predictor of insulin resistance in black
and white youths. J Pediatr. 2006;148(2):188
94.
6. Freedman DS, Serdula MK, Srinivasan SR, Berenson GS.
Relation of circumferences and skinfold thicknesses to lipid
and insulin concentrations in children and adolescents: the
Bogalusa Heart Study. Am J Clin Nutr. 1999;69(2):308
17.
7. Daniels SR, Morrison JA, Sprecher DL, Khoury P, Kimball
TR. Association of body fat distribution and cardiovascular
risk factors in children and adolescents. Circulation. 1999;
99(4):541
5. /
8. Janssen I, Katzmarzyk PT, Ross R. Body mass index, waist
circumference, and health risk: evidence in support of current
National Institutes of Health guidelines. Arch Intern Med.
2002;162(18):2074
9.
/ /
9. Janssen I, Katzmarzyk PT, Ross R. Waist circumference and
not body mass index explains obesity-related health risk. Am J
Clin Nutr. 2004;79(3):379
84.
/ /
10. Zhu S, Wang Z, Heshka S, Heo M, Faith MS, Heymsfield SB.
Waist circumference and obesity-associated risk factors
among whites in the third National Health and Nutrition
Examination Survey: clinical action thresholds. Am J Clin
Nutr. 2002;76(4):743
9.
/ /
11. Mei Z, Grummer-Strawn LM, Pietrobelli A, Goulding A,
Goran MI, Dietz WH. Validity of body mass index compared
with other body-composition screening indexes for the
assessment of body fatness in children and adolescents. Am
J Clin Nutr. 2002;75(6):978
85. / /
12. Pietrobelli A, Faith MS, Allison DB, Gallagher D, Chiumello
G, Heymsfield SB. Body mass index as a measure of adiposity
among children and adolescents: a validation study. J Pediatr.
1998;132(2):204
10.
/ /
13. Freedman DS, Wang J, Maynard LM, Thornton JC, Mei Z,
Pierson RN, et al. Relation of BMI to fat and fat-free mass
among children and adolescents. Int J Obes (Lond) 2005;
29(1):1
8. /
14. Maynard LM, Wisemandle W, Roche AF, Chumlea WC,
Guo SS, Siervogel RM. Childhood body composition in
relation to body mass index. Pediatrics 2001;107(2):344
50.
15. Must A, Anderson SE. Body mass index in children and
adolescents: considerations for population-based applica-
tions. Int J Obes (Lond) 2006;30(4):590
4.
16. Demerath EW, Schubert CM, Maynard LM, Sun SS,
Chumlea WC, Pickoff A, et al. Do changes in body mass
index percentile reflect changes in body composition in
children? Data from the Fels Longitudinal Study. Pediatrics.
2006;117(3):487
95.
/ /
17. Savva SC, Tornaritis M, Savva ME, Kourides Y, Panagi A,
Silikiotou N, et al. Waist circumference and waist-to-height
ratio are better predictors of cardiovascular disease risk factors
in children than body mass index. Int J Obes Relat Metab
Disord. 2000;24(11):1453
8. / /
18. 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. / NC: SAS Institute Inc.; 2006.
19. Freedman DS, Kahn HS, Mei Z, Grummer-Strawn LM,
Dietz WH, Srinivasan SR, et al. Relation of body mass index
and waist-to-height ratio to cardiovascular disease risk factors
in children and adolescents: the Bogalusa Heart Study. Am J
Clin Nutr. 2007;86:33
40. / /
20. Katzmarzyk PT, Srinivasan SR, Chen W, Malina RM,
Bouchard C, Berenson GS. Body mass index, waist circum-
ference, and clustering of cardiovascular disease risk factors in
a biracial sample of children and adolescents. Pediatrics.
2004;114(2):e198
205.
/ /
21. Li C, Ford ES, Mokdad AH, Cook S. Recent Trends in Waist
Circumference and Waist-Height Ratio Among US Children
/ /
and Adolescents. Pediatrics. 2006;118:1390
8. / /
22. McCarthy HD, Jarrett KV, Crawley HF. The development of
waist circumference percentiles in British children aged 5.0-
16.9 y. Eur J Clin Nutr. 2001;55(10):902
7. / /
/ /
23. Higgins PB, Gower BA, Hunter GR, Goran MI. Defining
health-related obesity in prepubertal children. Obes Res.
2001;9(4):233
40.
/ /
/
24. Moreno LA, Pineda I, Rodriguez G, Fleta J, Sarria A, Bueno
M. Waist circumference for the screening of the metabolic
syndrome in children. Acta Paediatr. 2002;91(12):1307
12. / /
25. Maffeis C, Pietrobelli A, Grezzani A, Provera S, Tato L. Waist
circumference and cardiovascular risk factors in prepubertal
children. Obes Res. 2001;9(3):179
87. / /
26. Heude B, Kettaneh A, Rakotovao R, Bresson JL, Borys JM,
Ducimetiere P, et al. Anthropometric relationships between
parents and children throughout childhood: the Fleurbaix-
Laventie Ville Sante Study. Int J Obes (Lond) 2005;29(10):
/ /
1222
9.
27. Agras WS, Hammer LD, McNicholas F, Kraemer HC. Risk
factors for childhood overweight: a prospective study from
birth to 9.5 years. J Pediatr. 2004;145(1):20
5.
/ /
28. Whitaker RC. Predicting preschooler obesity at birth: the role
of maternal obesity in early pregnancy. Pediatrics. 2004;
/
114(1):e29
36.
/
29. Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH.
Predicting Obesity in Young Adulthood from Childhood and
Parental Obesity. N Engl J Med. 1997;337:869
73. / /
30. Psarra G, Nassis GP, Sidossis LS. Short-term predictors of
abdominal obesity in children. Eur J Public Health. 2006; /
16(5):520
5.
/
31. Berkowitz RI, Stallings VA, Maislin G, Stunkard AJ. Growth
/
of children at high risk of obesity during the first 6 y of life:
implications for prevention. Am J Clin Nutr. 2005;81(1):140
/ /
6.
32. Faith MS, Berkowitz RI, Stallings VA, Kerns J, Storey M,
/ /
Stunkard AJ. Parental feeding attitudes and styles and child
body mass index: prospective analysis of a gene-environment
interaction. Pediatrics. 2004;114(4):e429
36. / /
/ /
33. Stunkard AJ, Berkowitz RI, Stallings VA, Cater JR. Weights of
parents and infants: is there a relationship? Int J Obes Relat
Metab Disord. 1999;23(2):159
62. / /
34. Frisancho AR. Anthropometric Standards for the Assessment
of Growth and Nutritional Status. Ann Arbor, MI: University
of Michigan Press, 1990.
35. Barlow SE. Expert committee recommendations regarding
the prevention, assessment, and treatment of child and
adolescent overweight and obesity: summary report. Pedia-
trics 2007;120 Suppl4:S164
92.
/ /
36. Lohman T, Roche A, Martorell R. Anthropometric standar-
dization reference manual. Champaign, IL: Human Kinetics,
1988.
37. Littell R, Milliken GA, Stroup WW, Wolfinger RD, Scho-
/
benberger O. SAS for Mixed Models, Second Edition: Cary,

38. SAS Institute Inc SAS/STAT 9.1 User’s Guide: Cary, NC:
SAS Institute Inc.; 2004.
39. Stokes M, Davis CS, Koch GG. Categorical Data Analysis
Using the SAS System: Cary, NC: SAS Institute Inc.; 1995.
CVDRF-waist circumference and obesity 159
48. Deshmukh-Taskar P, Nicklas TA, Morales M, Yang SJ, Zakeri
I, Berenson GS. Tracking of overweight status from child-
hood to young adulthood: the Bogalusa Heart Study. Eur J
Clin Nutr. 2006;60(1):48
57.
/ /

40. Salsberry PJ, Reagan PB. Dynamics of early childhood 49. Wang Y, Ge K, Popkin BM. Tracking of body mass index
overweight. Pediatrics. 2005;116(6):1329
38.
/ /

from childhood to adolescence: a 6-y follow-up study in

41. Li C, Kaur H, Choi WS, Huang TT, Lee RE, Ahluwalia JS. China. Am J Clin Nutr. 2000;72(4):1018
24.
/ /

Additive interactions of maternal prepregnancy BMI and 50. Wardle J, Brodersen NH, Cole TJ, Jarvis MJ, Boniface DR.
breast-feeding on childhood overweight. Obes Res. 2005; /

Development of adiposity in adolescence: five year long-

13(2):362
71.
/

itudinal study of an ethnically and socioeconomically diverse

42. Frisancho AR. Prenatal compared with parental origins of
sample of young people in Britain. BMJ. 2006;332(7550):
/ /

adolescent fatness. Am J Clin Nutr. 2000;72(5):1186
90.

1130
5.
/ /

43. Whitaker RC, Pepe MS, Wright JA, Seidel KD, Dietz WH.
51. Freedman DS, Khan LK, Serdula MK, Dietz WH, Srinivasan
Early adiposity rebound and the risk of adult obesity.
SR, Berenson GS. The relation of childhood BMI to adult
Pediatrics. 1998;101(3):E5.
adiposity: the Bogalusa Heart Study. Pediatrics. 2005;115(1):
/ /

44. Taylor RW, Goulding A, Lewis-Barned NJ, Williams SM.

/ /

Rate of fat gain is faster in girls undergoing early adiposity 22
7.

52. Janssen I, Katzmarzyk PT, Srinivasan SR, Chen W, Malina
rebound. Obes Res. 2004;12(8):1228
30.
/ /

45. Williams SM. Weight and height growth rate and the timing
of adiposity rebound. Obes Res. 2005;13(6):1123
30.
/ /
46. Rolland-Cachera MF, Deheeger M, Bellisle F, Sempe M,
RM, Bouchard C, et al. Utility of childhood BMI in the
prediction of adulthood disease: comparison of national and
international references. Obes Res. 2005;13(6):1106
15.
/ /

Guilloud-Bataille M, Patois E. Adiposity rebound in children:
a simple indicator for predicting obesity. Am J Clin Nutr.
1984;39(1):129
35.
/ /
53. Mueller WH, Dai S, Labarthe DR. Tracking body fat
distribution during growth: using measurements at two
occasions vs. one. Int J Obes Relat Metab Disord. 2001;25: / /

47. Cole TJ. Children grow and horses race: is the adiposity 1850
5.
rebound a critical period for later obesity? BMC Pediatr.
2004;4:6.
/ /