International Journal of Obesity (2006) 30, 374–379
& 2006 Nature Publishing Group All rights reserved 0307-0565/06 $30.00
www.nature.com/ijo
ORIGINAL ARTICLE
Body mass index of healthy men compared with
healthy women in the United States
KM Flegal1,2
1
National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD, USA and 2Center for
Weight and Health, University of California, Berkeley, CA, USA
Objective: To compare the distributions of body mass index (BMI) in relatively healthy nonsmoking men and women in the
United States.
Design: Cross-sectional national survey data from the Third National Health and Nutrition Examination Survey (NHANES III).
Subjects: In total, 11 404 nonsmoking men (n ¼ 4894) and women (n ¼ 6510), ages 20 years and above, drawn from a
representative population sample.
Measurements: Increasingly stringent definitions of ‘health’ were applied, based on self-reported health, medical history,
measurements of blood pressure, blood lipids, serum glucose, glycosylated hemoglobin, and behavioral factors including
smoking and physical activity. Main outcome measures were mean and median BMI by health level, 5th and 95th percentiles of
BMI, and the prevalence of overweight and obesity.
Results: For both men and women, the distribution of BMI became less skewed at better health levels. The range of BMI values
that included 90% of healthy men and women was approximately 19.5–30 kg/m2 for men and 18–30 kg/m2 for women, with
median values of approximately 24.5 kg/m2 for men and 21.5 kg/m2 for women. The prevalence of overweight declined sharply
with increasing health level for women but varied little for men; the prevalence of obesity declined at higher health levels for
both men and women.
Conclusions: Only about 5% of healthy younger men or women would be classified as obese by BMI levels. However, the
distribution of BMI differs between healthy men and healthy women. Relative to the distribution of BMI values for healthy men,
the distribution of BMI values for healthy women is shifted to the left and is more skewed.
International Journal of Obesity (2006) 30, 374–379. doi:10.1038/sj.ijo.0803117; published online 27 September 2005
Keywords: body mass index; NHANES; overweight; health
Introduction
A variety of body weight standards have been promulgated
and used over time by groups such as insurance companies,
NIH consensus conferences and WHO expert committees.1–5
These standards are based on criteria that are not completely
specified but which include elements of risk, predictive
power and digit preference.6–8
In the past, some weight standards, such as the Metropolitan Life Tables and the older Dietary Guidelines for
Americans used different values for men and women.1,2,9,10
Correspondence: Dr K Flegal, National Center for Health Statistics, Centers for
Disease Control and Prevention, 3311 Toledo Rd, Room 4311, Hyattsville, MD
20782, USA.
E-mail: [email protected]
The findings and conclusions in this report are those of the author and do not
necessarily represent the views of the funding agency.
Received 22 March 2005; revised 24 June 2005; accepted 5 July 2005;
published online 27 September 2005
At a given height and weight, because of gender differences
in body composition, a woman is likely to have more adipose
tissue than a man.11 However, current weight standards are
the same for men and women. The recommendations of the
NHLBI Clinical Guidelines for the Treatment of Overweight
and Obesity5 are similar for men and for women.
The purpose of this paper is to compare the distributions of
body mass index (BMI) in healthy nonsmoking men and
women in the United States, using a large representative
sample, and compare these distributions to current BMI recommendations. Identifying the BMI ranges occurring among
healthy nonsmoking men and women may help to inform
public health recommendations for achievable weight levels.
Methods
NHANES III
In the third National Health and Nutrition Examination
Survey (NHANES III), a nationally representative sample of
BMI of healthy men and women
KM Flegal
375
Table 1
Definitions of health levels
Health level Criteria
0
1
2
3
4
5
All nonsmoking respondents with measured height and weight
Met criteria for level 0 and in addition had self-reported excellent, very good or good health; no history of arthritis, diabetes mellitus, gallbladder
disease, congestive heart failure, stroke, emphysema, lupus, cancer other than skin cancer, heart attack or osteoporosis, and had no missing data for
blood pressure, total cholesterol or HDL-cholesterol
Met criteria for level 1 and in addition did not have high blood cholesterol or hypertension
Met criteria for level 2 and in addition did not have low HDL-cholesterol
Met criteria for level 3 and in addition, had serum cholesterol o200 mg/dl, systolic blood pressure o135 mmHg, diastolic blood pressure o85 mmHg,
glycosylated hemoglobin (HbA1c) o6%; had none of the following in the highest quartile for their sex and 10-year age group: serum cholesterol,
systolic blood pressure, diastolic blood pressure; did not have HDL-cholesterol in the lowest quartile for their sex and 10-year age group.
(Half-sample only). Met criteria for level 4 and in addition did not have undiagnosed diabetes or impaired fasting glucose or high LDL-cholesterol levels;
had serum triglycerides o150 mg/dl.
the US civilian noninstitutionalized population was selected
using a complex, stratified, multistage probability cluster
sampling design. A home interview was followed by a
physical examination in a mobile examination center. A
description of the plan and operation of the survey has been
published.12
Weight and height were measured using standardized
techniques and equipment. BMI was calculated as weight
(kg) divided by the square of height (m). For adults,
overweight was defined as a BMI value in the range
25–29.9 kg/m2 and obesity was defined as a BMI value
X30.0 kg/m2.4,5
Respondents who reported smoking as many as 100
cigarettes in their lifetime were asked if they smoked
cigarettes ‘now.’ Those who responded affirmatively were
considered to be current smokers. Current smokers were
excluded from the primary analyses because smoking is
often associated with a lower body weight and thus the
inclusion of smokers might confound the analyses. The
sample of nonsmoking respondents examined in the mobile
examination center in NHANES III included 4900 men and
6533 nonpregnant women, of whom six men and 23 women
had missing values for BMI, so that the initial analytic
sample size consisted of 4894 men and 6510 women An
additional 276 men and 366 women had missing values for
measured blood pressure, total serum cholesterol or HDLcholesterol and were excluded from subsequent analyses.
Definitions
Hypertension was defined as using hypotensive medication
or having a systolic blood pressure at or above 140 mmHg or
having a diastolic blood pressure at or above 90 mmHg. High
blood cholesterol was defined as using medications to lower
cholesterol or having serum cholesterol level of 6.22 mmol/l
(240 mg/dl) or greater. Low HDL-cholesterol was defined as
an HDL-cholesterol level below1.04 mmol/l (40 mg/dl) for
men and below 1.30 mmol/l (50 mg/dl) for women.
One-half of the eligible participants had been randomly
assigned to participate in a morning examination after an
overnight fast. Data on fasting glucose levels came from this
half-sample. Undiagnosed diabetes was defined as a morning
fasting plasma glucose level of 7 mmol/l (126 mg/dl) or
greater in a person without previously diagnosed diabetes.
Impaired fasting glucose was defined as a morning fasting
plasma glucose level in the range of 6.1 mmol/l (110 mg/dl)
up to 7.0 mmol/l (126 mg/dl) in a person without previously
diagnosed diabetes. Elevated LDL-cholesterol was defined as
a serum LDL-cholesterol level of 4.14 mmol/l (160 mg/dl) or
greater.
Different levels of ‘health’ were defined using sequential
exclusion criteria, so that each level included only participants who met the criteria for the previous level as well.
There is no standard definition of health; criteria included
participants’ self-report of health status, history of medical
conditions and measured lipids and blood pressure levels
(Table 1). Health 5 (the highest level) was based on a halfsample of individuals who had been randomized to a
morning examination and had fasted at least 9 h.
Data analysis and statistical methods
For adults, age was defined as age in years at the time of the
household interview, which generally preceded the examination by 2–3 weeks. The age groups used were 20–39,
40–59, 60–79 and 80 þ years.
Descriptive analyses were carried out using SAS statistical
software (Version 9.1, SAS Institute, Cary NC, USA). For the
survey, sampling weights had been calculated that took into
account the unequal probabilities of selection resulting from
the sample design and from planned oversampling of certain
subgroups. Weighted percentiles were calculated using the
sampling weights. For the Health 5 level, the sample weights
for the morning sample were used. To examine the
differences in the distribution of BMI among groups,
mean-difference plots were used.13 For graphical presentation, the distributions of BMI were smoothed using a 4253H
nonparametric smoothing algorithm, based on sequential
calculations of running medians for groups of adjacent
points of different lengths.14 The objective of this paper is to
estimate quantiles and describe the distributions graphically;
formal statistical tests were not carried out.
International Journal of Obesity
BMI of healthy men and women
KM Flegal
376
Table 2
Unweighted sample size and weighted percent of nonsmoking population by sex, agegroup and health status
Sex
Men
Women
Health level
0
1
2
3
4
5a
0
1
2
3
4
5a
20–39 years
1846
1375
1145
887
471
207
2478
1551
1410
893
503
256
(100%)
(78%)
(66%)
(48%)
(27%)
(27%)
(100%)
(67%)
(62%)
(43%)
(26%)
(25%)
40–59 years
1189
668
375
254
96
41
1745
692
457
304
129
59
60–79 years
(100%)
(60%)
(36%)
(26%)
(10%)
(8%)
(100%)
(46%)
(31%)
(23%)
(11%)
(11%)
1369
348
122
93
21
7
1705
286
66
42
7
2
80+ years
(100%)
(27%)
(11%)
(7%)
(2%)
(1%)
(100%)
(21%)
(6%)
(5%)
(1%)
(1%)
485
103
33
27
6
3
576
117
17
9
0
0
(100%)
(23%)
(7%)
(6%)
(1%)
(2%)
(100%)
(21%)
(4%)
(2%)
(0%)
(0%)
a
Health 5 was limited to a half-sample of respondents who had been randomized to a morning examination and had fasted at least 9 h.
The unweighted sample sizes and the weighted population
percent are shown in Table 2 for each health level. The
number of respondents decreased progressively for each
added health category. For purposes of this paper, estimates
of the 5th and 95th percentiles are not provided for
categories where the unweighted sample size is less than
100 respondents, and estimates of mean and median BMI are
not provided for categories where the unweighted sample
size is less than 20 respondents. These restrictions affected
primarily those ages 60 years and above in the higher health
categories.
Table 3 Mean and median BMI and 5th and 95th percentiles for men, by
age and health level
Health
level
20–39
years
40–59
years
60–79
years
80+
years
Mean BMI
0
1
2
3
4
5
26.1
25.8
25.4
25.0
24.4
25.1
27.9
27.1
26.4
26.0
25.4
24.2
27.6
26.9
26.7
25.5
24.4
–
25.4
25.0
24.9
24.8
–
–
Median BMI
0
1
2
3
4
5
25.4
25.2
24.9
24.4
24.0
24.7
27.1
26.4
25.9
25.2
24.9
24.5
27.3
26.5
26.6
25.6
24.4
–
25.4
25.2
24.6
24.1
–
–
5th percentile
0
1
2
3
4
5
20.2
20.2
19.9
19.8
19.7
19.4
22.0
21.9
21.4
20.9
–
–
21.1
21.0
20.4
–
–
–
19.5
19.7
–
–
–
–
95th percentile
0
1
2
3
4
5
34.7
33.5
32.7
31.5
29.9
32.7
36.9
34.6
32.7
32.7
–
–
35.2
33.8
33.5
–
–
–
32.1
30.3
–
–
–
–
Results
Percentiles of BMI
The distribution of BMI values in these samples is shown in
Table 3 (men) and Table 4 (women), which include the
mean, the median, the 5th percentile and the 95th percentile
values for health levels 0–5. In general, as more and more
stringent definitions of health were applied, the distribution
of BMI shifted to the left, as shown by decreases in the mean,
the median and the 5th and 95th percentiles.
For men, sample sizes were adequate to estimate mean and
median BMI for all health levels only for ages 20–39 and 40–
59 years. In those age groups, both mean and median tended
to decline as health level increased. The magnitude of this
shift was less for the younger men than for the 40–59 years
old men, for whom the median declined almost three BMI
units between health level 0 and health level 5. For both
these age groups, the mean declined slightly more than the
median, reflecting the decreased skewness in the distributions at higher health levels. The decline in the 5th
percentile of BMI was slight, less than one BMI unit for the
younger men. For both age groups, however, the decline in
the 95th percentile was greater than the decline in the
median, also indicating a decrease in skewness. Although at
older ages, sample sizes were too small to estimate these
International Journal of Obesity
percentiles for all health levels, the trends were broadly
similar to those seen for men ages 40–59 years.
At health level 4 for men ages 20–39 years, the median BMI
was B24.5 kg/m2, the 5th percentile was B19.5 kg/m2 and
the 95th percentile was B30 kg/m2. At all health levels the
values for men ages 40–59 years were slightly higher than
those for men ages 20–39 years.
The BMI values for women also declined with increasing
health level, with the decline being greater at the 95th
percentile and lower at the 5th percentile than at the
BMI of healthy men and women
KM Flegal
377
20–39
years
40–59
years
60–79
years
80+
years
Mean BMI
0
1
2
3
4
5
25.5
24.6
24.2
23.0
22.4
21.9
27.9
25.9
24.8
24.1
23.0
23.0
27.6
25.7
25.0
24.7
–
–
25.5
25.0
–
–
–
–
Median BMI
0
1
2
3
4
5
23.5
23.1
22.9
22.1
21.7
21.4
26.4
24.7
23.6
23.3
22.2
21.7
26.6
25.4
25.2
25.2
–
–
25.3
24.5
–
–
–
–
5th percentile
0
1
2
3
4
5
18.5
18.4
18.4
18.0
18.0
17.6
19.9
19.4
18.9
18.9
18.6
–
20.3
19.9
–
–
–
–
19.1
19.5
–
–
–
–
95th percentile
0
1
2
3
4
5
38.6
36.6
35.8
30.7
30.6
29.6
40.1
36.1
34.0
32.8
29.5
–
38.6
32.7
–
–
–
–
33.5
33.2
–
–
–
–
10
15
20
25
30
35
Body mass index
40
45
50
Figure 1 Smoothed frequency distribution of BMI values for nonsmoking
men ages 20–39 years who were defined as healthy (solid line) or not healthy
(dotted) at health level 4.
Percent of population
Health
level
Percent of population
Table 4 Mean and median BMI and 5th and 95th percentiles for women, by
age and health level
10
15
20
25
30
35
40
45
50
Body mass index
median. The estimates were greater for women ages 40–59
years than for women ages 20–39 years for the mean, median
and 5th percentile, but not always for the 95th percentile.
For women in both these age groups the declines tended to
be greater than for men in the corresponding age groups. The
declines were particularly large at the 95th percentile, where
the value of BMI declined by nine units. Even at the 5th
percentile, the decline was almost one BMI unit.
Two examples of the shifts in the distribution of BMI are
shown graphically in Figures 1 and 2 for men and women
ages 20–39 years for health level 4. Because these are
nonparametrically smoothed distributions, they may not
correspond exactly to the percentile estimates shown in the
tables. The smoothed distribution of BMI for men ages 20–39
years defined as healthy (health 4) is shown in Figure 1 along
with the smoothed distribution of BMI for men of the same
age who were defined as not healthy at level 4. As may be
seen, the distribution of BMI in healthy men is shifted to the
left relative to the distribution of BMI in less healthy men
and is also less skewed. Similar results were seen for other
health levels and were confirmed by mean-difference plots
(not shown) that indicated both a leftward shift and a
decrease in skewness in almost all cases.
Similar smoothed distributions of BMI for women are
shown in Figure 2. For women, the same trends are evident.
However for women, both distributions are more skewed
than the corresponding distributions for men.
Figure 2 Smoothed frequency distribution of BMI values for nonsmoking
women ages 20–39 years who were defined as healthy (dashed line) or not
healthy (dotted line) at health level 4.
The distribution of BMI differed somewhat between men
and women. With a few exceptions at the 95th percentile,
values were almost always lower for women than for men at
a given age and health level. For women at higher health
levels the mean and median BMI values were B21.5 kg/m2,
the 5th percentile was B18 kg/m2 and the 95th percentile
was B30 kg/m2 The smoothed distributions of BMI for men
and for women at health level 4 are shown in Figure 3.
Compared with the distribution of BMI values for healthy
men, the distribution of BMI values for healthy women is
shifted to the left and is more skewed.
Prevalence of overweight and obesity
The prevalence of BMI values 25–29 kg/m2 (overweight) and
X30 kg/m2 (obese), as well as overall prevalence of BMI
X25 kg/m2 are shown in Figure 4 for men. The prevalence of
BMI X30 kg/m2 (obesity) dropped markedly with increasing
health levels. For example, for men ages 20–39 years the
prevalence of BMI X30 kg/m2 dropped from 15.8% (Health
0) to 6.8% (Health 5). The prevalence of BMI 25–29.9 kg/m2
(overweight) also dropped but remained fairly high, with the
International Journal of Obesity
BMI of healthy men and women
KM Flegal
378
70
Percent of population
60
50
40
30
20
10
0
10
15
20
25
30
35
40
45
50
Body mass index
0 1 2 3 4 5
0 1 2 3 4 5
0 1 2 3
0 1
20-39 y
40-59 y
60-79 y
80 + y
Figure 5 Prevalence of BMI 25–29.9 kg/m2 (light bar) and BMI X30 kg/m2
Figure 3 Smoothed frequency distribution of BMI values for nonsmoking
(dark bar) by age and health level for nonsmoking women.
men (solid line) and women (dashed line) ages 20–39 years who were defined
as healthy at health level 4.
prevalence of BMI 25–29.9 kg/m2 was 38.1% for men versus
11.8% for women. As a result, the prevalence of overweight
and obesity combined (BMI X25 kg/m2) was also lower for
women than for men.
80
70
60
50
40
30
20
10
0
0 1 2 3 4 5
20-39 y
0 1 2 3 4 5
40-59 y
0 1 2 3
60-79 y
0 1 2
80 + y
Figure 4 Prevalence of BMI 25–29.9 kg/m2 (light bar) and BMI X30 kg/m2
(dark bar) by age and health level for nonsmoking men.
lowest value being 32.5% (men ages 20–39 years, Health 4).
For men ages 20–39 years, the prevalence of BMI 25–29.9
kg/m2 was almost identical in the healthiest group (38.1%)
and in the overall group (38.3%). The prevalence of BMI 25–
29.9 kg/m2 ranged from 32.5–46.8% over all age ranges and
health statuses. The overall prevalence of overweight and
obesity combined (BMI X25 kg/m2) was relatively high in all
age groups regardless of health status. Even for the youngest
men at Health 5, the prevalence was 44.9%.
Comparable results for women are shown in Figure 5. For
women ages 20–39 and 40–59 years, at health level 0, the
prevalence of BMI X30 kg/m2 was greater than the prevalence of BMI 25–29.9 kg/m2. However, the drop in the
prevalence of BMI 25–29.9 kg/m2 and of BMI X30 kg/m2
with increasing health level was greater than for men. For
example for women ages 20–39 years, the prevalence of BMI
X30 kg/m2 dropped from 20.7% (Health 0) to 4.2% (Health
5) and the prevalence of BMI 25–29.9 kg/m2 dropped from
19.3% to 11.8%. At all health levels, the prevalence of BMI
X30 kg/m2 was similar for men and women, but the
prevalence of BMI 25–29.9 kg/m2 was considerably lower
for women than for men. For example, for men and women
ages 20–39 years at health level 5, the prevalence of BMI
X30 kg/m2 was 6.8% for men and 4.2% for women but the
International Journal of Obesity
Additional analyses
These analyses were repeated using additional criteria. In one
set of analyses, health that was self-described as ‘good’
(rather than as very good or excellent) was considered
unhealthy rather than healthy. Another set of analyses was
limited to those who were physically active, defined as
engaging in moderate activity at least five times a week or
vigorous activity as least three times a week. Analyses were
also carried out for non-Hispanic whites only. A further
analysis combined all three of the above criteria. These
additional criteria had little effect on the estimates, generally
only 0.1 or 0.2 BMI units, and the variations were not
systematic (results not shown). Analyses were also carried
out for smokers only. Results for male smokers only were
lower than for nonsmokers, with smokers having a distribution of BMI shifted to the left by 1–2 BMI units relative to
nonsmokers (results not shown). At health level 1 and above,
there was little difference in the distribution of BMI between
female smokers and female nonsmokers.
Discussion
These results present the distribution of BMI in a large
nationally representative sample of nonsmoking healthy
men and women, with ‘healthy’ defined at five levels using
progressively more stringent criteria. The criteria were based
on self-reported health, reports of selected chronic conditions and medications, and measured values of blood
pressure, blood lipids including cholesterol, HDL-cholesterol, LDL-cholesterol and triglycerides, plasma glucose level
and glycosylated hemoglobin level. The criteria were
sufficiently stringent that at the highest level only 27% of
BMI of healthy men and women
KM Flegal
379
nonsmoking men and 25% of nonsmoking women ages
20–39 years were considered ‘healthy.’
At progressively higher levels of ‘health’ the distribution of
BMI shifted to the left and became less skewed and more
normally distributed for both men and women. The shifts
were minor at the median and larger at the upper end.
The distribution of BMI in this apparently healthy
population had a median value of about 24.5 kg/m2 for
men, but a lower value of about 21.5 kg/m2 for women,
depending on age. The range of BMI values from the 5th to
the 95th percentile, covering 90% of the healthy population,
was approximately 19.5–30 kg/m2 for men and approximately 18–30 kg/m2 for women. The distribution of BMI
among healthy women was more skewed than the corresponding distribution among healthy men. It would be of
interest to see whether this pattern has changed over time or
is similar to that observed in other countries.
People with BMI values over 25 kg/m2 who appear to be
healthy at younger ages may be at increased risk of
developing chronic conditions or cardiovascular risk factors
at older ages, relative to similar people with BMI values
under 25 kg/m2. The median BMI of the highly selected
‘healthy’ population tended to be higher in older age groups.
The ‘healthy’ population described here may have less risk
than would an unselected less healthy population and thus
show more uniform changes with age across all levels of BMI.
At younger ages relatively few people develop conditions
such as hypertension or hypercholesterolemia; at older ages
many people do regardless of weight.15
A limitation of this study is in the definition of ‘healthy.’
There is no standard set of criteria by which it can be
determined that a given individual is healthy. The criteria
applied here cover self-reported overall health status, selfreported medical history of a set of conditions, a set of
standardized physiological measurements, primarily those
related to cardiovascular risk factors, and several behavioral
factors (smoking and physical activity). Self-reported overall
health status may to some extent control for other
unspecified health conditions not included in the medical
history questionnaire. Additional unmeasured factors might
distinguish those within the ‘healthy’ group who are more or
less healthy. On the other hand, the criteria used here
already exclude almost 3/4 of younger men and women,
suggesting that the use of additional criteria would result in a
definition of health so restrictive that only a small proportion even of younger men and women would be considered
healthy. These findings suggest that the distribution of BMI
values for healthy people is different for women than for
men. For healthy women, the distribution of BMI values is
shifted to the right relative to the distribution for healthy
men but is also more skewed, so that the median and 5th
percentile values for women are lower than those for men,
but the 95th percentiles differ little. This suggests that the
optimal BMI levels associated with good health may differ
between men and women. However, even though the shape
of the distribution differed between men and women, only
about 5% of healthy men or women were classified as obese
by BMI, supporting the utility of the definition of obesity as a
BMI of 30 kg/m2 or above.
References
1 Metropolitan Life Insurance Company. New weight standards for
men and women. Stat Bull 1959; 40 (November–December): 1–4.
2 Metropolitan Life Insurance Company. Metropolitan height and
weight tables. Stat Bull 1983; 64 (January–June): 2–9.
3 National Institutes of Health Consensus Development Panel on
the Health Implications of Obesity. Health implications of
obesity: National Institutes of Health Consensus Development
Conference Statement. Ann Intern Med 1985; 103: 1073–1077.
4 WHO Expert Committee on Physical Status: the Use and
Interpretation of Anthropometry. Physical status: the use and
interpretation of anthropometry. Report of a WHO expert
committee. World Health Organization: Geneva, 1995. (World
Health Organization Technical Report Series; 854).
5 NHLBI Expert Panel on the Identification Evaluation and
Treatment of Overweight and Obesity in Adults. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults – The Evidence Report. National
Institutes of Health [published erratum appears in Obes Res 1998
Nov;6(6):464]. Obes Res 1998; 6 (Suppl 2): 51S–209S Also available
at
http://www.nhlbi.nih.gov/guidelines/obesity/ob_gdlns.htm
(accessed March 9, 2005).
6 Garrow JS. Treat Obesity Seriously: A Clinical Manual. Churchill
Livingstone: Edinburgh, 1981.
7 Bray GA. In defense of a body mass index of 25 as the cut-off
point for defining overweight. Obes Res 1998; 6: 461–462.
8 Bray GA. Definition, measurement, and classification of the
syndromes of obesity. Int J Obes 1978; 2: 99–112.
9 U.S. Department of Agriculture and U.S. Department of Health
and Human Services. Nutrition and Your Health: Dietary Guidelines for Americans. Home and Garden Bulletin No. 232. US
Government Printing Office: Washington, DC, 1980.
10 U.S. Department of Agriculture and U.S. Department of Health
and Human Services. Nutrition and Your Health: Dietary Guidelines for Americans. Home and Garden Bulletin No. 232. U.S.
Government Printing Office: Washington, DC, 1985.
11 Norgan NG. The beneficial effects of body fat and adipose tissue
in humans. Int J Obes Relat Metab Disord 1997; 21: 738–746.
12 National Center for Health Statistics. Plan and Operation of the
Third National Health and Nutrition Examination Survey, 1988–
1994. National Center for Health Statistics: Hyattsville, MD,
1994. (Vital and Health Statistics Series 1; No. 32).
13 Flegal KM, Troiano RP. Changes in the distribution of body mass
index of adults and children in the US. Int J Obes Relat Metab
Disord 2000; 24: 807–818.
14 Velleman PF, Hoaglin DC. Applications, Basics and Computing of
Exploratory Data Analysis. Duxbury Press: Boston MA, 1981.
15 Brown CD, Higgins M, Donato KA, Rohde FC, Garrison R,
Obarzanek E et al. Body mass index and the prevalence of
hypertension and dyslipidemia. Obes Res 2000; 8: 605–619.
International Journal of Obesity