International Journal of Obesity (2000) 24, 1453±1458

ß 2000 Macmillan Publishers Ltd All rights reserved 0307±0565/00 $15.00

www.nature.com/ijo

Waist circumference and waist-to-height

ratio are better predictors of cardiovascular
disease risk factors in children than body mass
index

SC Savva1*, M Tornaritis1, ME Savva1, Y Kourides1, A Panagi1, N Silikiotou1, C Georgiou1 and A Kafatos2

1
Research and Education Program of Child Health, Cyprus; and 2Department of Social and Preventive Medicine, University of Crete
School of Medicine, Crete, Greece

BACKGROUND: Visceral adipose tissue is associated with increased risk for cardiovascular disease risk factors and
morbidity from cardiovascular diseases. Waist measurement and waist-to-height ratio (WHtR) have been used as
proxy measures of visceral adipose tissue, mainly in adults.
OBJECTIVE: To validate body mass index (BMI), waist circumference and WHtR as predictors for the presence of
cardiovascular disease risk factors in children of Greek-Cypriot origin.
SUBJECTS AND METHODS: A total of 1037 boys and 950 girls with mean age 11.4  0.4 y were evaluated. Dependent
variables for the study were total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholestrol (HDL-C), low-
density lipoprotein cholesterol (LDL-C), and systolic (SBP) and diastolic (DBP) blood pressure.
RESULTS: When children were divided into two groups according to the 75th percentile for BMI, waist circumference
and WHtR, all dependent variables had higher mean values in the highest percentile groups in WHtR groups and
almost all variables in BMI and waist circumference groups. Adjusted odds ratios for predicting pathological values of
cardiovascular disease risk factors were slightly higher for the highest WHtR group for predicting lipid and lipoprotein
pathological values and for the highest BMI groups in predicting high blood pressure measurement. Using stepwise
multiple regression analysis to explain the variance of the dependent variables, waist circumference was the most
signi®cant predictor for all variables both for boys and girls, whereas BMI had the lowest predictive value for the
detection of cardiovascular disease risk factors.
CONCLUSION: Waist circumference and WHtR are better predictors of cardiovascular disease risk factors in children
than BMI. Further studies are necessary to determine the cutoff points for these indices for an accurate prediction of
risk factors.
International Journal of Obesity (2000) 24, 1453±1458

Keywords: cardiovascular; BMI; waist; waist-to-height ratio; children

Introduction concentrations, blood pressure and left ventricular

Increased weight is associated with higher levels of
triglycerides and low-density lipoprotein cholesterol
(LDL-C) and lower levels of high-density lipoprotein
cholesterol (HDL-C), and at least 75% of hyperten-
sion can be attributed to obesity in adults.1 Lipid
abnormalities are related primarily to the amount of
visceral fat as measured by computed tomography.2
However, obese men with low levels of visceral fat
were found to have similar lipid levels to non-obese
subjects.2 In children, greater deposition of central fat,
as measured by dual X-ray absorptiometry, was asso-
ciated with less favorable plasma lipid and lipoprotein
*Correspondence: SC Savva, Research and Education Program
of Child Health, 138 Limassol Ave, Flat 205, Nicosia 2015, Cyprus.
E-mail: samar1@cytanet.com.cy
Received 31 August 1999; revised 22 May 2000; accepted
22 May 2000
mass,3 but there was no in¯uence of visceral fat, as
measured by magnetic resonance imaging, on serum
lipids.4
Nevertheless, dual X-ray absorptiometry and imag-
ing techniques cannot be used for mass screening or
in everyday clinical practice. Instead of these meth-
ods, a simple, inexpensive and noninvasive method is
required for the selection of children for screening for
cardiovascular disease risk factors. Body mass index
(BMI) is a useful index for the evaluation of obesity.
Although in adults a BMI value greater than 28 kg=m2
is associated with a three to four-fold increased risk of
ischemic heart disease, stroke and diabetes mellitus,5
in children there is no de®nition of obesity that relates
BMI to health measures.1 Waist circumference corre-
lates with visceral adipose tissue,6 ± 8 plasma lipids,
lipoproteins and insulin levels7 better than waist-to-
hip ratio (WHR) in adults. Waist circumference was
also found in children to be correlated with high levels
of plasma lipids and lipoprotein levels.9,10 Recently,
 Predicting cardiovascular risk in children
SC Savva et al
1454
waist-to-height ratio (WHtR) has been validated as a
useful predictor of coronary risk factors in adults.11,12
The purpose of this study was to compare BMI,
waist circumference and WHtR for the prediction of
cardiovascular disease risk factors in children.
Subjects and methods
The study included 1987 children (boys 1037 and girls
950), 10 ± 14 y of age, who had been evaluated for the
presence of cardiovascular disease risk factors by the
Research and Education Program for Child Health of
Cyprus. The evaluation took place during the school
year 1998 ± 1999. Children were from urban (29%) as
well as rural (71%) areas. Subjects were included after
informed consent was obtained from the parent or
legal guardian. The Research and Education Program
for Child Health is under the auspices of the Ministry
of Health and the Ministry of Education and Culture
of Cyprus and is funded by the Popular Bank of
Cyprus.
Anthropometry
Trained research assistants took all measurements.
Weight was measured with a portable Seca 762
scale (Vogel & Halke, Hamburg, Germany) to the
nearest kilogram. Height was measured with a porta-
ble Seca stadiometer 208 (Vogel & Halke, Hamburg,
Germany) to the nearest millimeter. Both measure-
ments were taken in the fasting state, with the child
dressed in light clothing, without shoes. The body
mass index (BMI) was calculated as weight (kg)=
height2 (m2). Every morning, the scale and stadio-
meter were calibrated with a standard weight and
height, respectively. Waist circumference was mea-
sured with a ¯exible tape, at the level of the umbili-
cus, and was recorded to the nearest millimeter.
Waist-to-height ratio was calculated as waist=height.
The inter-observer variability was >0.96 for all
measurements.
Pubertal stage
The pubertal stage (breast development and pubic
hair in females, genitalia development and pubic
hair in males) was recorded according to Tanner
classi®cation.
Blood pressure measurement
Blood pressure was measured in the right arm with the
child at the sitting position, at rest, with the mercury
sphygmomanometer Baumanometer (W.A. Baum,
Copiague, New York). The ®rst Korotkoff phase
was used to determine systolic blood pressure (SBP)
and the ®fth Korotkoff phase to determine diastolic
blood pressure (DBP). The mean of three measure-
International Journal of Obesity
ments was used for the analysis. The same researcher
took all blood pressure measurements.
Lipids and lipoproteins
Blood samples were obtained in Vacutainers after 12 h
fasting, before breakfast. All samples were analyzed
at the Clinical Chemistry Laboratory of the General
Hospital of Nicosia. The samples were analyzed the
same day, and since the study lasted for about 8
months, the same stock of reagents was not used for
all analyses. All measurements were reproducible and
validated by internal (Beringer, local) and external
(Murex, international EQAS, WHO) quality control.
Total cholesterol (TC) and triglycerides (TG) were
assayed enzymatically (Hitachi 717 analyzer) and
HDL-C with a non-immunologic enzymatic reaction.
LDL-C was calculated from the Friedewald formula
(LDL-C ˆ TC 7 (HDL-C ‡ TG=5)).
Statistical analysis
The statistical analysis was performed with SPSS for
Windows statistical package (SPSS Inc., Chicago, II).
Descriptive statistics for physical characteristics and
cardiovascular disease risk factors were calculated,
both for boys and girls. Differences between genders
were calculated with Student's t-test. The dependent
variables for this study were the established cardio-
vascular risk factors (SBP and DBP, and lipids and
lipoproteins). Independent variables were the three
tested parameters of the study (BMI, waist circum-
ference and WHtR) as well as age, sex and stage
of pubertal maturation. Sex was transformed to a
continuous variable (0 for female, 1 for male).
The relationships between the various variables
(independent, systolic and diastolic blood pressure
and lipid and lipoprotein levels) were calculated
using partial correlation coef®cients, adjusted for
age, sex and pubertal maturation. We used the 75th
percentile of our sample independent variables as a
cut-off point, to divide the sample into subgroups with
high and normal values for the three variables. These
cut-off points are shown in Table 1. Mean (  s.d.)
values for the dependent variables within these groups
were calculated and compared with Student's t-test.
Table 1 The 75th percentile values for the three study
independent variables
50th 75th
Variable percentile percentile Range
Boys
Body mass index (kg=m2) 18.53 21.37 12.35 ± 34.52
Waist circumference (cm) 68.20 76.8 53.1 ± 124.7
Waist-to height (ratio) 0.460 0.519 0.341 ± 0.733
Girls
Body mass index (kg=m2) 18.95 21.67 12.35 ± 36.54
Waist circumference (cm) 69.50 76.3 50.4 ± 113.3
Waist-to height (ratio) 0.460 0.509 0.358 ± 0.739

Stepwise multiple linear regression analysis was
used to test the in¯uence of the independent variables
on the variance of cardiovascular risk factors. Odds
ratios were calculated according to the independent
variables 75th percentile cut-off point, in order to
detect the presence of cardiovascular risk factors,
after adjustment for age, sex and pubertal maturation.
Signi®cance was selected at P values <0.05 for all
tests.
Predicting cardiovascular risk in children
SC Savva et al
1455
Mean values of various cardiovascular disease risk
factors in the two groups of children, as they were
divided by the 75th percentile cut-off point for BMI,
waist circumference and WHtR, are presented in
Tables 5 ± 7 respectively. There were statistically
signi®cant higher mean values of SBP, DBP, LDL-C
and triglycerides in the higher percentile groups of all
three independent variables. In girls, the BMI groups

Table 4 Relationship (partial correlation coef®cientsa for age,

sex and pubertal maturation) between the lipid and lipoprotein
levels
Results
Total cholesterol HDL-C LDL-C
Study subjects' characteristics are presented in Table HDL-C 0.227
2. As it would be expected due to the more advanced LDL-C 0.889 70.167
pubertal stage for this age (42% of girls were Tanner Triglycerides 0.294 70.354 0.212
stage 4 or 5 vs only 9% of boys), girls had higher a
All correlation coef®cients are signi®cant at P < 0.001.
values of body weight, height and SBP. There is a
marginal difference in WHtR, and slight differences
for lipid and lipoprotein concentrations between boys Table 5 Comparison of cardiovascular risk factors according to
and girls. BMI 75th percentile cut off point. Values are expressed as mean
(  s.d.)
The partial correlation coef®cient (adjusted for age,
sex and pubertal maturation) between systolic and Groups according
diastolic blood pressure was r ˆ 0.699, P < 0.001. to BMI 75th percentile
Similarly the relationships between the independent Variable > 75th  75th
variables and blood lipid and lipoprotein levels were
Boys (n ˆ 1037)
calculated and are shown in Table 3 and 4, respec- Systolic blood pressure (mmHg) 119.0 (11.1) 110.9 (11.1)a
tively. There is a very strong correlation between the Diastolic blood pressure (mmHg) 72.3 (8.3) 66.5 (7.8)a
independent variables (BMI vs waist circumference Total cholesterol (mg=dl) 176.6 (32.1) 168.8 (27.3)a
HDL-C (mg=dl) 56.2 (13.1) 63.9 (13.0)a
r ˆ 0.914, BMI vs WHtR r ˆ 0.923, and waist circum- LDL-C (mg=dl) 103.6 (29.7) 92.5 (25.1)a
ference vs WHtR r ˆ 0.947, all correlations signi®cant Triglycerides (mg=dl) 84.6 (49.5) 62.0 (27.2)a
at P < 0.001). Girls (n ˆ 950)
Systolic blood pressure (mm Hg) 118.2 (11.6) 112.9 (9.5)a
Diastolic blood pressure (mm Hg) 71.1 (8.8) 67.8 (7.9)a
Total cholesterol (mg=dl) 169.8 (30.8) 167.5 (25.6)
Table 2 Anthropometric and cardiovascular risk factor HDL-C (mg=dl) 54.1 (11.5) 60.5 (12.9)a
characteristics of subjects. Values are expressed as mean LDL-C (mg=dl) 96.7 (26.5) 91.8 (22.7)b
(  s.d.) Triglycerides (mg=dl) 95.4 (46.3) 75.3 (34.8)a
a
Boys Girls P < 0.001; bP < 0.05.
Variable (n ˆ 1037) (n ˆ 950)

Age (y) 11.4 (0.4) 11.4 (0.4)

Weight (kg) 42.0 (10.0) 43.4 (10.2)a Table 6 Comparison of cardiovascular risk factors according to
Height (cm) 146.8 (7.5) 148.8 (7.3)b waist circumference 75th percentile cut off point. Values are
Body mass index (kg=m2) 19.3 (3.5) 19.4 (3.6) expressed as mean (  s.d.)
Waist circumference (cm) 70.7 (10.1) 70.7 (9.7)
Waist-to-height (ratio) 0.48 (0.06) 0.47 (0.06)c Groups according to
Systolic blood pressure (mmHg) 112.9 (10.2) 114.2 (10.3)b waist 75th percentile
Diastolic blood pressure (mmHg) 68.0 (8.3) 68.7 (8.3)
Total cholesterol (mg=dl) 170.8 (28.8) 168.1 (27.0)c Variable > 75th  75th
Triglycerides (mg=dl) 67.7 (35.6) 80.3 (38.8)b
HDL-C (mg=dl) 62.0 (13.5) 58.9 (12.9)b Boys (n ˆ 1037)
LDL-C (mg=dl) 95.3 (26.7) 93.0 (23.8)c Systolic blood pressure (mmHg) 118.4 (11.1) 111.1 (9.2)a
Diastolic blood pressure (mmHg) 71.9 (8.5) 66.7 (7.9)a
a
P < 0.005, bP < 0.001, cP < 0.05. Total cholesterol (mg=dl) 176.1 (31.5) 169.0 (27.6)b
HDL-C (mg=dl) 56.2 (12.3) 63.9 (13.3)a
LDL-C (mg=dl) 103.0 (28.9) 92.7 (25.5)a
Triglycerides (mg=dl) 84.1 (48.9) 62.3 (27.9)a
Table 3 Relationship (partial correlation coef®cientsa for age,
sex and pubertal maturation) between the independent variables Girls (n ˆ 950)
of the study Systolic blood pressure (mmHg) 118.4 (11.4) 112.8 (9.5)a
Diastolic blood pressure (mmHg) 71.3 (8.0) 67.8 (8.0)a
BMI Waist Total cholesterol (mg=dl) 168.6 (30.4) 167.9 (25.9)
HDL-C (mg=dl) 53.6 (11.4) 60.7 (12.9)a
Waist 0.914 LDL-C (mg=dl) 95.8 (26.8) 92.1 (22.7)
WHtR 0.923 0.947 Triglycerides (mg=dl) 95.6 (46.1) 75.3 (34.9)a
a a
All correlation coef®cients are signi®cant at P < 0.001. P < 0.001, bP < 0.05.

International Journal of Obesity

 Predicting cardiovascular risk in children
SC Savva et al
1456
Table 7 Comparison of cardiovascular risk factors according to Results of the multiple regression analysis for blood
WHtR 75th percentile cut-off point. Values are expressed as
mean (  s.d.)
pressure are presented in Table 8 and for lipid and
lipoprotein levels in Table 9. All three independent
Groups according to variables were signi®cant determinants for SBP and
WHtR 75th percentile
DBP. Age was also a signi®cant determinant both for
Variable > 75th  75th SBP and DBP, whereas sex and pubertal maturation
Boys (n ˆ 1037)
did not prove to be signi®cant predictors. The percen-
Systolic blood pressure (mmHg) 118.3 (11.1) 111.1 (9.2)a tage of the variance of SBP explained was 14.2% by
Diastolic blood pressure (mmHg) 71.9 (8.3) 66.7 (7.9)a waist circumference, 3.5% by WHtR and 1.8% by
Total cholesterol (mg=dl) 178.2 (32.6) 168.3 (26.9)a BMI. Similarly, for DBP the variance explained was
HDL-C (mg=dl) 55.9 (12.8) 63.9 (13.1)a
LDL-C (mg=dl) 104.9 (30.4) 92.0 (24.6)a 9.8%, 1.1% and 0.6% respectively.
Triglycerides (mg=dl) 86.3 (49.6) 61.4 (26.6)a Waist circumference was also a signi®cant predic-
Girls (n ˆ 950) tor for all lipid and lipoprotein levels except for
Systolic blood pressure (mmHg) 117.8 (11.2) 113.0 (9.7)a triglycerides, whereas WHtR was a signi®cant pre-
Diastolic blood pressure (mmHg) 71.1 (8.6) 67.8 (8.0)a
Total cholesterol (mg=dl) 171.7 (30.9) 166.9 (25.5)b dictor for all but HDL-C. On the other hand, BMI
HDL-C (mg=dl) 53.5 (11.2) 60.7 (12.9)a was a signi®cant predictor only for triglycerides. The
LDL-C (mg=dl) 98.8 (27.1) 91.1 (22.9)a percentage of variance of the lipid and lipoprotein
Triglycerides (mg=dl) 97.0 (47.3) 74.8 (34.0)a
levels explained by these parameters, however, was
a
P < 0.001; bP < 0.05. low. Thus waist circumference explained 8.8% of
HDL-C variance, and 2.8% of TC variance. WHtR ex-
plained 3.9% of LDL-C and 1.6% of TC variance and
had no difference in TC, and the waist circumference BMI explained 8.4% of the variance of triglycerides.
groups had no differences in TC and LDL-C. Only the Using the odds ratio (adjusted for age, sex and
WHtR groups had statistical signi®cant differences in pubertal maturation) for the prediction of the presence
all dependent variables both for boys and girls. of cardiovascular disease risk factors, subjects in the

Table 8 Stepwise multiple regression analysis models for explaining the variance of blood
pressure. Independent variables were body mass index (BMI), waist circumference, waist-to-
height ratio (WHtR), age, sex and pubertal stage

Variable Coef®cient s.e. r2 Change P

Systolic blood Constant 114.0 6.76 Ð < 0.001

pressure Waist 0.69 0.70 0.142 < 0.001
WHtR 7119.66 10.94 0.035 < 0.001
BMI 1.09 0.16 0.018 < 0.001
Age 71.20 0.55 0.002 0.028
Diastolic blood Constant 97.45 5.70 Ð < 0.001
pressure Waist 0.40 0.06 0.098 < 0.001
Age 73.67 0.46 0.020 < 0.001
WHtR 754.02 9.23 0.011 < 0.001
BMI 0.50 0.14 0.006 < 0.001

Table 9 Stepwise multiple regression analysis models for explaining the variance of blood lipids
and lipoproteins. Independent variables were body mass index (BMI), waist circumference, waist-
to-height ratio (WHtR), age, sex and pubertal stage

Variable Coef®cient s.e. r2change P

Total cholesterol Constant 181.9 19.9 Ð < 0.001

WHtR 211.2 34.2 0.016 < 0.001
Waist 70.95 0.21 0.028 < 0.001
Pubertal stage 71.43 0.60 0.004 0.011
Age 73.75 1.70 0.003 0.024
HDL-C Constant 87.2 2.10 Ð < 0.001
Waist 70.39 0.03 0.88 < 0.001
Sex 2.94 0.58 0.12 < 0.001
LDL-C Constant 91.80 18.03 Ð < 0.001
WHtR 199.8 30.90 0.039 < 0.001
Waist 70.72 0.19 0.020 < 0.001
Age 73.48 1.50 0.003 0.021
Pubertal stage 71.02 0.51 0.002 0.045
Triglycerides Constant 2.6 7.7 Ð 0.733
BMI 1.5 0.57 0.084 0.009
Sex 712.7 1.67 0.025 < 0.001
WHtR 102.1 34.3 0.004 0.003

International Journal of Obesity

 Predicting cardiovascular risk in children
SC Savva et al
1457
Table 10 Odds ratio (adjusted for age, sex and pubertal maturation) for the risk of the presence
of cardiovascular risk factors according to the 75th percentile cut off points for the three
independent variables

Odds ratio (95% con®dence interval)

BMI75 a WAIST75 WHtR75

Boys
SBP 125b 8.52 (5.16 ± 14.06)c 7.09 (4.36 ± 11.52)c 7.55 (4.63 ± 11.29)c
TC 170 1.62 (1.20 ± 2.17)d 1.58 (1.17 ± 2.12)d 1.69 (1.27 ± 2.27)c
LDL 110 2.31 (1.69 ± 3.16)c 2.11 (1.54 ± 2.89)c 2.35 (1.73 ± 3.20)c
Triglycerides 110 4.65 (2.96 ± 7.29)c 4.10 (2.63 ± 6.41)c 4.99 (3.19 ± 7.79)c
Girls
SBP 125 2.75 (1.73 ± 4.37)c 2.56 (1.61 ± 4.08)c 2.24 (1.42 ± 3.54)c
TC 170 1.10 (0.79 ± 1.52) 1.09 (0.79 ± 1.52) 1.18 (0.86 ± 1.62)
LDL 110 1.87 (1.28 ± 2.73)d 1.70 (1.16 ± 2.50)d 1.92 (1.36 ± 2.82)c
Triglycerides 110 2.36 (1.57 ± 3.54)c 2.36 (1.56 ± 3.56)c 2.61 (1.76 ± 3.87)c
a
BMI75: body mass index >75th percentile cut-off, waist75: waist circumference > 75th percentile
cut-off, WHtR 75: waist-to-height ratio >75th percentile cut-off.
b
SBP125: systolic blood pressure >125 mmHg, TC170: total cholesterol >170 mg=dl, LDL110:
LDL-C >110 mg=dl; triglycerides 110: triglycerides >110 mg=dl.
c
P < 0.001; dP < 0.005.

highest group for BMI had higher risk for high blood
pressure measurement and subjects in the highest
group for WHtR had the highest risk for high
plasma lipid and lipoprotein levels (Table 10). Never-
theless the risk for one or another variable was only
slightly higher than for the others.
Discussion
The purpose of this study was to compare BMI with
waist circumference and WHtR as predictors for the
presence of cardiovascular disease risk factors. BMI
was chosen because it is the most widely used index
for the evaluation of obesity both in adults1,5 and
children.13 Waist circumference correlates well with
visceral adipose tissue in adults.6 ± 8 Waist circumfer-
ence and WHtR have been validated in adults as
useful predictors for cardiovascular disease risk fac-
tors and coronary heart disease.7,11,12,14 Finally, it has
been proposed that one particular advantage of using
the waist-to-height ratio might be that unisex action
levels could be speci®ed15 and we hypothesized that
this ratio may be useful in children since it takes into
account the child's height. Body weight, height and
waist circumference are all simple measurements that
most physicians can precisely measure, while the
same does not apply to other measurements such as
the skinfold measurements.
The results of the study indicate that obese children
have higher risk for the presence of cardiovascular
disease risk factors. Children who exceed the 75th
percentile for body mass index, waist circumference
and waist-to-height ratio have signi®cantly higher
mean values for all cardiovascular risk factors,
except for HDL-C, for which mean values were
signi®cantly lower (Tables 5 ± 7). These results are
in agreement with other studies that have shown that
obesity, by means of increased BMI, is associated
with increased risk for the presence of cardiovascular
disease risk factors in children.16,17 Despite this asso-
ciation, there is no de®nition of obesity in children
that relates BMI to health outcomes.1 A possible
explanation for this weakness of BMI is that children
and adolescents with similar BMI have large differ-
ences in total body fat and percentage body fat.18 The
accumulation of visceral fat has been proven a better
predictor for adult morbidity than obesity itself,2 and
there is a correlation of adult changes in adiposity by
means of BMI, total body fat and percentage body fat
with serum lipids and lipoprotein levels.19 The use,
therefore, of indices that correlate to visceral fat
would be more justi®ed.
In particular, in our study children in the highest
WHtR percentile group had higher mean values for all
six cardiovascular disease risk factors both for boys
and girls. Waist circumference and BMI were almost
as good as WHtR, except for total cholesterol and
LDL-C for which there were no differences in the two
percentile groups in girls.
Children in the highest percentile groups for all
three indices were at signi®cantly greater risk (un-
adjusted odds ratio) of having pathological values of
cardiovascular disease risk factors. BMI was slightly
better than the other two indices for predicting high
systolic blood pressure measurement and WHtR was
slightly better in predicting increased plasma lipid and
lipoprotein levels both in boys and girls, except for
total cholesterol where there was no greater risk in
either group.
The results of stepwise multiple regression analysis
with BMI, waist circumference and WHtR as the
independent variables for the explanation of the vari-
ance of cardiovascular disease risk factors gave waist
circumference as signi®cant predictor for all risk
factors, except for triglyceride levels in girls. WHtR
was the second most signi®cant predictor since it
explained SBP, DBP, TC, LDL-C and triglycerides.
Finally, BMI could predict only SBP and DBP and
International Journal of Obesity
 Predicting cardiovascular risk in children
SC Savva et al
1458
triglycerides. The percentage of the variance of the
dependent variables explained by these indices was,
however, low. As mentioned before, because of the
simplicity and reproducibility of body weight, height
and waist circumference measurements, we included
them in the analysis, and we did not include less
reproducible measurements, such as the skinfold
thickness (which might substantially improve the
regression models). The explanation of the variance
of blood pressure was higher with waist circumference
and lowest for BMI. Waist circumference was also the
better predictor of the variance of TC and HDL-C,
WHtR was the better predictor for LDL-C and BMI
was better only for triglyceride levels.
The results of this study are in agreement with the
observation that simple anthropometric indices of
visceral adipose tissue are helpful for predicting the
presence of cardiovascular risk factors in adults and
children.7,9 ± 12 Waist circumference and waist-to-
height ratio were proven to be better predictors than
BMI. Although waist circumference has been vali-
dated as a useful predictor for cardiovascular disease
risk factors in children,9,10 we also validated waist-to-
height ratio in this study, since it takes into account a
child's height, whereas a single cut-off point cannot
be speci®ed for waist measurement in children as is
the case in adults. Expanding these observations
further in other age groups could possibly de®ne a
single cut-off value for WHtR, which will highlight
children in high risk groups for cardiovascular disease
risk factors.
In conclusion, the results of the present study
indicate that waist circumference and waist-to-height
ratio are better predictors for the presence of cardio-
vascular disease risk factors than BMI in children.
BMI seems to be useful for the prediction of high
blood pressure measurement in children but does not
help in the prediction of lipid and lipoprotein plasma
levels. Further studies are necessary to extend these
®ndings in other age groups and to determine precise
cut-off points for these indices for an accurate predic-
tion of the presence of cardiovascular disease risk
factors.
Acknowledgements
We would like to thank the Ministries of Health and
Education and Culture of Cyprus, the Popular Bank of
Cyprus, and the Clinical Chemistry Laboratory of the
General Hospital of Nicosia.
References
1 Krauss RM, Winston M, Fletcher BJ, Grundy SM. Obesity.
Impact on cardiovascular disease. Circulation 1998; 98:
1472 ± 1476.
2 Lamarche B, Lemieux S, Dagenais GR, Despres JP. Visceral
obesity and the risk of ischaemic heart disease: insights from
the Quebec cardiovascular study. Growth Hormone IGF Res
1998; 8: 1 ± 8.
International Journal of Obesity
3 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:
541 ± 545.
4 Leung SSF, Chan YL, Lam CWK, Peng XH, Woo KS,
Metreweli C. Body fatness and serum lipids of 11-year-old
Chinese children. Acta Paediatr 1998; 87: 363 ± 367.
5 Rosenbaum M, Leibel RL, Hirsch J. Obesity. New Engl J Med
1997; 337: 396 ± 407.
6 Conway JM, Chanesta FF, Wang P. Intraabdominal adipose
tissue and anthropometric surrogates in African American
women with upper- and lower-body obesity. Am J Clin Nutr
1997; 66: 1345 ± 1354.
7 Pouliot SC, Despres JP, Lemieux S, Moorjani S, Bouchard C,
Tremblay A, Nadeau A, Lupien PJ. Waist circumference and
abdominal sagittal diameter: best simple anthropometric
indexes of abdominal visceral adipose tissue accumulation
and related cardiovascular risk in men and women. Am J
Cardiol 1994; 73: 460 ± 468.
8 Taylor RW, Keil D, Gold EJ, Williams SM, Goulding A. Body
mass index, waist girth, and regional adiposity in women:
evaluation using receiver operating characteristic curves. Am J
Clin Nutr 1998; 67: 44 ± 49.
9 Flodmark CE, Sveger T, Nilsson-Ehle P. Waist measurement
correlates to a potentially atherogenic lipoprotein pro®le
in obese 12 ± 14-year-old children. Acta Paediatr 1994; 83:
941 ± 945.
10 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: 308 ± 317.
11 Hsieh SD, Yoshinaga H. Abdominal fat distribution and
coronary heart disease risk factors in men-waist=height ratio
as a simple and useful predictor. Int J Obes Relat Metab
Disord 1995; 19: 585 ± 589.
12 Hsieh SD, Yoshinaga H. Waist=height ratio as a simple and
useful predictor of coronary heart disease risk factors in
women. Intern Med 1995; 34: 1147 ± 1152.
13 Barlow SE, Dietz WH. Obesity evaluation and treatment:
Expert Committee Recommendations. Pediatrics 1998; 102:
e29.
14 Rexrode KM, Carey VJ, Hennekens CH, Walters EE, Colditz
GA, Stampfer MJ, Wilter WC, Manson JE. Abdominal adi-
posity and coronary heart disease in women. JAMA 1998;
280: 1843 ± 1848.
15 Ashwell M, Lejeune S. Ratio of waist circumference to height
may be better indicator of need for weight management. Br
Med J 1996; 312: 377.
16 Dwyer JT, Stone EJ, Yang M, Feldman H, Weber LS, Must A,
Perry CL, Nader PR, Parcel GS. Predictors of overweight and
overfatness in a multiethnic pediatric population. Am J Clin
Nutr 1998; 67: 602 ± 610.
17 Freedman DS, Serdula MK, Percy CA, Ballew C, White L.
Obesity, levels of lipids and glucose, and smoking among
Navajo adolescents. J Nutr 1997; 127: 2120S ± 2127S.
18 Pietrobelli A, Faith MS, Allison DB, Gallagher D, Chiumello
G, Heyms®eld SB. Body mass index as a measure of adiposity
among children and adolescents: a validation study. J Pediatr
1998; 132: 204 ± 210.
19 Siervogel RM, Wisemandle W, Maynard LM, Guo SS, Roche
AF, Chumlea WC, Towne B. Serial changes in body composi-
tion throughout adulthood and their relationships to changes in
lipid and lipoprotein levels. The Fels longitudinal study.
Arterioscler Thromb Vasc Biol 1998; 18: 1759 ± 1764.