Vol. 18 No. 2
Printed in Great Britain
International Journal of Epidemiology
© International EpidemioJoglcal Association 1989
Body Fat Distribution and Self-Reported
Prevalence of Hypertension, Heart
Attack, and Other Heart Disease in
Older Women
AARON R FOLSOM, RONALD J PRJNEAS, SUSAN A KAYE AND JOHN T SOLER
The relationship of blood pressure to body weight, as
well as to other measures of adiposity, is well established.
Numerous epidemiological studies have demonstrated
increasing adiposity is associated with higher levels
of blood pressure.'~ 3 Weight loss reduces blood
pressure/
The association of weight to other cardiovascular
diseases is less clear-cut. Some prospective studies have
shown a positive linear relationship of weight to coronary
heart disease. 5 Other investigations have demonstrated
a wide range of associations including a negative
association, no association, a U-shaped relationship, and
a threshold effect.6
Over the past few years, reports have indicated that
blood pressure levels or hypertension prevalence is
related to the distribution of body fat beyond overall
fatness.7"12 Increasing abdominal adiposity increases
the risk of high blood pressure. Similarly, a few studies
have suggested that greater abdominal fat is associated
with increased risk of other cardiovascular diseases. Q~16
Lapidus, et al reported that the waist-to-hip
circumference ratio (WHR) was significantly and
positively associated with the 12-year incidence of
myocardial infarction, angina pectoris, stroke, and death
in women.13 The association of WHR with incidence of
myocardial infarction remained significant in multivariate
analysis, but did not remain significant for the other
endpoints. Among men from the same community,
WHR was positively associated with incidence of stroke
and ischaemic heart disease, after accounting for overall
fatness, but was not related to these endpoints in
multivariate analysis. u In the Paris Prospective Study,
trunk adiposity, measured by multiple skinfold thicknesses, was positively and significantly associated with
coronary heart disease incidence in middle-aged men. u
The association remained significant in multivariate
analysis. Gillum found an index of abdominal adiposity
was significantly and positively associated with higher
blood pressure, hypertension, and hypertensive heart
disease independent of multiple confounders. a The
association of Gillum's index with coronary heart disease
prevalence was not significant after adjustment for other
confounders.
The purpose of the present study was to confirm the
association of fat distribution with prevalence of
hypertension, heart attack, and other heart disease based
on self-reports in a large survey of post-menopausal
women.
Division of Epidemiology, School of Public Health, University of
Minnesota, Stadium Gate 27, 611 Beacon St SE, Minneapolis, MN
55455, USA.
361
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
Folsom A R (Division of Epidemiology, School of Public Health, University of Minnesota, Stadium Gate 27, 611
Beacon St SE, Minneapolis, MN 55455, USA), Prineas R J, Kaye S A and Soler J T. International Journal of
Epidemiology 1989,18: 361-367.
The associations of self-reported body mass and fat distribution with setf-reported prevalence rates of
hypertension, heart attack, and other heart disease were examined in a sample of 40 000 women, aged 55-69
years. Fat distribution was measured by the waist-to-hip circumference ratio (WHR), which had a mean ± SD of
0.838 ± 0.085. Prevalence of hypertension was significantly and positively associated with both body mass index
(BMI) and waist-to-hip ratio. The prevalence rate ratio for hypertension in the highest versus the lowest fertile of
body mass index and walst-to-hip ratio (considered jointly) was 2.7. Prevalence rates of heart attack and other heart
disease were significantly and positively associated with waist-to-hip ratio but not with body mass index. The
prevalence rate ratios were 2.2 for heart attack and 1.4 for other heart disease in the highest versus the lowest tertile
of body mass index and waist-to-hip ratio. Findings were substantiated using multiple logistic regression. These
results support the hypothesis that a significant relationship exists between body fat distribution and the
occurrence of cardiovascular disease in older women.
362
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Non-Caucasians, 0.8% of the population, were
excluded from analyses. Twenty-eight per cent of those
who reported other heart disease also reported having
had a heart attack. To provide mutually exclusive
categories, these subjects were included in the heart
attack category but were excluded from analysis of the
other heart disease variable. The heart attack category
was thus intended to be specific for coronary artery
disease, whereas the other heart disease category is a
mixture of less definite coronary disease, arrhythmias,
congestive heart failure, and other heart conditions. Data
were analysed univariately first by comparing characteristics of those with the self-reported disease category (ie
hypertension, heart attack, other heart disease) versus
those without. BMI and WHR were divided into tertiles
to examine the joint relationship of these factors to the
disease categories. Prevalence rate ratios were calculated
as a measure of association, with the lowest BMI
multiplied by WHR fertile as the reference category.
Finally, multiple logistic regression analysis was used
to determine the independent association of WHR to
each disease category. Regression models were built
by first forcing in age, BMI, and WHR, retaining
them regardless of significance, and allowing all other
first order covariates (categorical variables shown in
Table 1) to compete for entry using forward stepping.
Then, quadratic terms and two-way interactions involving
the variables age, BMI, and WHR were added, if
significant. Terms that increased the log likelihood chisquare by at least 25.0 were taken as being significant.
SAS19 was used for analyses.
RESULTS
Thirty-eight per cent of the Iowa women aged 55—69
reported having a history of hypertension, 3.2% reported
a previous heart attack, and 6.9%reportedother heart
disease (but no heart attack). Table 1 shows the selfreported prevalence of these conditions by several
participant characteristics. History of heart attack was
three times as common and other heart disease was twice
as common among hypertensives. All three conditions
were much higher in women reporting diabetes than
those reporting no diabetes, in those with lower levels
of formal education, and those who were formerly
married. Cigarette smokers had a higher prevalence of
heart attack and other heart disease, but not
hypertension. Physical activity was inversely associated
with hypertension and other heart disease, but seemingly
unrelated to heart attack. Women who lived on a farm
tended to have lower prevalence rates of all three
conditions.
The mean (± SD) BMI and WHR in the Iowa women
were 34.2 ± 6.4 kg/m u and 0.838 ± 0.085, respectively. The product moment correlation between
BMI and WHR was 0.40. Participants who reported
hypertension, heart attack, or other heart disease were
slightly older, had a higher mean BMI and a higher mean
WHR than those without these diseases (Table 2).
Figures 1-3 indicate the joint relationships of BMI
and WHR to the disease categories. The tertiles for WHR
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
METHODS
As part of a longitudinal study on cancer incidence in
the state of Iowa, we recruited a sample of 41 837 Iowa
women aged 55—69 years.17 Subjects were selected
from a list of driver's license holders in the state of Iowa.
Approximately 100 000 women, aged 55—69 years, were
mailed a questionnaire for completion and return.
Forty-three per cent of subjects returned the questionnaire. We compared survey respondents with nonrespondents using information from the driver's license
tape and US census data by county of residence.
Respondents were slightly younger (0.3 years) and had
a slightly lower body mass (0.4 kg/m2), based on
driver's license information. In addition, respondents
tended to be from counties which, by 1980 census data,
were slightly more rural and had residents with slightly
fewer years of education and income.
The survey asked for self-reports of both medication
use and physician-diagnosed medical conditions
including: diabetes mellitus, heart attack, 'angina or
other heart disease' (here called other heart disease), and
high blood pressure. Those reporting ever having been
on antihypertensive medication were also included as
having high blood pressure. Subjects were also asked
to report their current height (without shoes and to the
nearest inch) and weight (without clothes and to the
nearest pound). A body mass index (BMI) was calculated
as weight (kg)/height (m)1-3. In addition, a paper tape
measure, verified for accuracy, was enclosed for the
measurement of several body circumferences, including
waist and hip. The reliability and accuracy of these
measures have been previously reported.18 Briefly,
measurements were to be made in duplicate with the
help of another person. Abdominal circumference was
measured to the nearest quarter inch at a point one inch
above the umbilicus. Hips were measured to the nearest
quarter inch at the maximal protrusion of the buttocks.
The duplicate measurements were averaged for analysis,
and the ratio of waist-to-hip circumference (WHR) was
calculated. Marital status, education level, smoking
status, urban/rural residence, and physical activity were
obtained by self report. The two questions on physical
activity enquired about regular participation in moderate
and vigorous activity. Subjects were categorized as
having low, medium, or high activity based on the
combined frequency of these types of activities.
363
BODY FAT DISTRIBUTION AND CARDIOVASCULAR DISEASE
were <0.795, 0.795 to 0.869, and >0.869; the tertiles for
BMI were <30.8, 30.8 to 35.8, and >35.8 kg/m.1-3
Hypertension prevalence was significantly associated
with BMI, showing about a twofold increase in
prevalence across tertiles. Hypertension was also
associated with WHR (slightly less than a twofold
prevalence increase across tertiles). The rate ratio of
TABLE 1 Self-reported prevalence of hypertension, prior heart attack,
and other heart disease according to participant characteristics, Iowa
women, aged 55—69 years.
Prevalence Rales (per 100)
Participant
charatceristic
Hypertension
Other heart
disease
5.6
1.7
10.9
4.6
Hypertensive
Yes
No
Diabetes mellituj
Yes
No
67.3
36.2
10.2
2.7
16.0
6.4
Cigarette smoking
Ever
Never
36.7
39.0
4.6
2.4
7.5
6.6
Physical activity level
Low
Medium
High
40.2
37.1
35.1
3.3
2.9
3.3
7.3
6.6
6.3
Education level
<High school graduate
High school graduate
>High school graduate
43.0
39.3
34.7
4.6
3.7
2.7
6.6
6.3
Marital status
Never married
Currently married
Formerly married
36.0
37.4
41.1
1.9
2.9
4.6
6.7
6.7
7.9
36.6
37.6
2.6
3.5
5.8
6.8
41.6
3.7
7.4
38.7
3.4
6.7
38.0
3.3
7.2
37.9
3.2
7.3
Residence
Farm
Rural, not farm
City/town, population
<1000
City/town, population
1000-2499
City/town, population
2 5 0 0 - 1 0 000
City/town, population
>10 000
FIGURE 1 Self-reported prevalence* (per 100) and prevalence rate
ratios^ of hypertension by tertiles of BMI and WHR, Iowa women,
aged 55—69 yean.
* Block heights and numbers at top of blocks indicate prevalence rates.
+ Numbers in three-by-three table indicate prevalence rate ratios.
8.5
TABLE 2 Mean (±SEM) age, body mass index (BMI), and waist-to-hip ratio (WHR) of Iowa women, aged 55—69 years, with and without a
self-reported history of hypertension, heart attack, and other heart disease
Hypertension
Characteristic
Age (years)
BMI (kg/m1-5)
WHR
No
62.2±0.03
32.9±0.04
0.824±0.0005
25 120
Heart attack
Yes
63.1 ±0.03*
36.3±O.O6*
0.861±0.0007»
15 514
•p^.0001 for overall difference in means
No
62.5±0.02
34.1 ±0.03
0.837±O.O0O4
39 014
Yes
64.2±0.11»
35.5±0.20»
0.868±0.0025»
1287
No
Other heart disease
Yes
62.4±0.02
34.1±0.03
0.836±0.0004
36 443
63.5±O.O8*
34.8±0.13*
0.847±0.0016*
2713
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
Heart
attack
hypertension in the group with the highest versus lowest
tertiles of WHR and BMI was 2.7. For heart attack, there
was a significant increase in prevalence across tertiles
of WHR, whereas BMI seemed to be inconsistently
related to heart attack history. Being in the highest versus
lowest tertile of WHR afforded about a twofold increase
in heart attack prevalence ratio. Other heart disease was
associated equally strongly with BMI and WHR. The
odds of other heart disease in the group with the highest
versus lowest tertiles of BMI and WHR was 1.4.
364
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
FIGURE 2 Self-reported prevalence* (per 100) and prevalence rate
ratios* of heart attack by tertilei ofBMI and WHR. Iowa women, aged
55-69 yean.
* Block heights and numbers at top of blocks indicate prevalence rates,
t Numbers in three-by-three table indicate prevalence rate ratios.
In the final model for heart attack (Table 3), the BMI
term became non-significant (p X).O5) with the
introduction of the hypertension term. However, the
WHR term remained significant (p <0.0001). A 0.085
unit increase in WHR increased the predicted odds of
heart attack, 1.2-fold (versus 1.5-fold in the model with
only age, BMI, and WHR). In addition, the predicted
odds of heart attack was 3.9 for hypertension, 2.5 for
diabetes, 2.5 for ever smoking, 1.6 for vigorous physical
activity, and 0.71 for greater than a high school education.
In the model for other heart disease, the BMI term
became non-significant (p >0.05) after introduction of
the hypertension term. There was significant interaction
(p <0.0001) between WHR and age, meaning that the
WHR effect must be interpreted by age. At age 55, a
0.085 unit increase in WHR was associated with a
0.78-fold decrease in the predicted odds of other heart
disease. At age 69, the same increase in WHR increased
the predicted odds of other heart disease 1.3-fold.
Moreover, the predicted odds of other heart disease was
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
To examine the independence of WHR and BM1
effects, multiple logistic regression models were
constructed for the three disease categories. In three
factor models with age, WHR, and BMI as main effects,
the logistic regression coefficients ± SEM for WHR
were: 3.04 ± 0.15 for hypertension, 4.75 ± 0.35 for heart
attack, and 3.22 ± 0.30 for other heart disease (all
p <0.0001). For BMI, the regression coefficients were
0.069 ± 0.002, 0.035 ± 0.006, and 0.025 ± 0.004,
respectively (all p <O0001). The final models, including
other covariates, are shown in Table 3. Hypertension
prevalence (Table 3) was significantly related (p <O0001)
to WHR, WHR2, and BMI. At the mean WHR (0.838),
a 0.085 unit increase in WHR (one SD) increased the
predicted odds of hypertension 1.3—fold. A 6.4 kg/mIJ
increase in BMI (one SD) increased the predicted odds
of hypertension 1.5-fold. In addition, the predicted odds
of hypertension was increased 2.4-fold for diabetes and
was 0.89 for greater than a high school education versus
other categories of education.
FIGURE 3 Self-reported prevalence* (per 100) and prevalence rate
ratios* of other heart disease by teniles of BM1 and WHR. Iowa
women, aged 55-69 years.
* Block heights and numbers at top of blocks indicate prevalence rates,
t Numbers in three-by-three table indicate prevalence rate ratios.
365
BODY FAT DISTRIBUTION AND CARDIOVASCULAR DISEASE
TABLE 3 Multiple logistic regression modelsforprevalence of hypertension, heart attack, and other heart disease on participant characteristics,
Iowa women, aged 55—69 years
Hypertension
Independent variable
Beta
0.01
0.96
0.53
0.20
0.05
0.02
2.8 for hypertension, 2.1 for diabetes, 1.4 for ever
smoking, 0.73 for being currently married, and 0.75 for
greater than a high school education.
DISCUSSION
This study of a random sample of white women, aged
55—69, demonstrated that distribution of body fat is
associated with self-reported prevalence of hypertension,
heart attack, and other heart disease. Specifically, risks
of these diseases were increased in those with greater
levels of self-measured WHR, which reflects greater
adiposity in the abdomen relative to the hips. The
associations of WHR to disease were moderate in
strength, generally had a dose-response pattern, and
were independent of other measured risk factors.
Moreover, WHR was associated quadratically with
Jiypertension prevalence and displayed an interaction
with age in relation to other heart disease prevalence.
Self-reported BMI also was significantly related to
prevalence of heart attack and other heart disease in the
simplest logistic regression models. However, this
association became non-significant when the strong
association of BMI with hypertension prevalence was
accounted for.
Beyond the associations with body size measures, we
found all three disease categories were positively
associated with self-reports of diabetes mellitus and with
lower levels of education. Prevalence of heart attack and
other heart disease were also positively associated with
hypertension and smoking. We found other heart disease
associated negatively with currently being married,
consistent with other reports.20 Surprisingly, however,
prevalence of heart attack was positively associated with
Beta
SE
Other heart disease
Beta
SE
1.66
-1.27
2.12
0.30
0.23
0.39
1.30
-1.31
-25.88
0.42
0.21
0.16
4.60
0.07
0.68
0.93
1.35
0.91
0.59
0.10
0.08
0.08
0.36
0.75
1.03
0.35
0.46
0.09
0.06
0.06
0.44
0.08
-0.32
0.06
-0.29
0.06
-0.34
0.08
high levels of physical activity. The validity of this
finding, based on self-reports in a cross-sectional study,
must be questioned. Exercise might have a detrimental
effect in older adults; however, most evidence is to the
contrary.21 Verification of this finding in prospective
studies of older women is warranted.
The possible drawbacks of this study must be
considered to place these results in perspective. This
study was a cross-sectional postal survey of prevalence
rates, meaning that fetal cases were excluded and that
appropriate temporality cannot be established between
WHR and disease. However, others have reported
abdominal adiposity is associated with incidence, as well
as prevalence, of cardiovascular disease. 7 " 16 The
response rate of our survey was low, but response was
unrelated to body mass and age. It would have been
useful to have blood cholesterol measurements in this
sample, but this was unfortunately not possible. The
validity of the body size measurements and self-reported
disease endpoints might be questioned. We believe the
girth measurements are accurate and reliable,18 and
others have confirmed that the bias in self-reported
height and weight is small.22-23 Self-reports of cardiovascular disease have moderate validity,2425 and we are
currently attempting to document validity in this sample
of women. To increase specificity, we defined the heart
attack variable to be mutually exclusive from the other
heart disease variable. The other heart disease category
is somewhat difficult to interpret, because it is a mixture
of several heart conditions. Finally, if physicians happen
to look more intensively for cardiovascular conditions
in people with abdominal adiposity, our results, based
on self-reports of physician diagnoses, might suffer from
a diagnostic detection bias.
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
Age (years)
0.05
Age^lOO
WHR
9.10
WHRX age
WHR*
-3.57
BMI*100 (kg/m IJ )
6.45
Diabetes (1 =yes, 0=no)
0.87
Hypertension (1 =yes, 0=no)
Smoker (1—ever, 0=never)
Physical activity
(l=high, 0-=low/med)
Marital status
(l=current, 0 = never, formerly)
Education (1 =GT high school,
0=LE high school)
-0.12
Heart attack
SE
366
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
ACKNOWLEDGEMENTS
The authors thank Ms Karen Johnson, Ms Laura
Kemmis, Drs Robert Wallace and Peter Isacson for their
contributions to this project. The authors especially
thank the women who kindly volunteered for study.
Supported by National Cancer Institute Grant 1R01-CA
39742-01.
REFERENCES
Stamler R, Stamler J, Riedlinger W F, Algera G, Roberts R H.
Weight and blood pressure: Findings in hypertension screening
of one million American!. JAMA 1978; 240: 1607-10.
2
Kannel W B, Brand N, Skinner i 1 Jr, Dawber T R, McNamara
P M. The relation of adiposity to blood pressure and
development of hypertension. Ann Intern Med EX>7; 67:48 -59.
3
Dyer A R, Stamler J, Shekelle R B, a al. Relative weight and blood
pressure in four Chicago epidemiologic studies. J Chron Dis
1982; 35: 897-908.
"Tuck M L, Sowers J, DomfeW L, Kledzik G, Maxwell M. The
effect of weight reduction on blood pressure, plasma renin
activity, and plasma aldosterone levels in obese patients. N Engl
J Med 1981; 304: 930-3.
3
Hubert H B, Feinleib M, McNamara P M, Castelli W P. Obesity
as an independent risk factor for cardiovascular disease: A
26-year follow-up of participants in the Framingham Heart
Study. Circulation 1983; 67: 968-77.
6
Barrett-Connor E L. Obesity, atherosclerosis, and coronary artery
disease. Ann Intern Med B85; 103 (Pt 2): 1010-9.
1
7
Hartz A J, Rupley D C, Rimm A A. The association of girth
measurements with disease in 32 836 women. Am J Epidemiol
1984; 110: 71-80.
•Blair D, Habicht J-P, Sims E A H, Sylvester D, Abraham S.
Evidence for an increased risk for hypertension with centrally
located body fat and the effect of race and sex on this risk.
Am J Epidemiol 1984; 119: 526-40.
'Stallones L, Mueller W H, Christensen B L. Blood pressure,
fatness, and fat patterning among USA adolescents from two
ethnic groups. Hypertension 1982; 4: 483-6.
"Katt S H, Hedigo- M L, Zemel B S, Parks J S. Blood pressure, body
&t, and dehydroepiandrosterone sulfate variation m adolescence.
Hypertension 1986; 8: 277-84.
n
Weinsier R L, Norris D J, Birch R, et al. The relative contribution
of body fat and fat pattern to blood pressure level. Hypertension
1985; 7: 578-85.
a
Gillum R F. The association of body fit distribution with hypertension, hypertensive heart disease, coronary heart disease,
diabetes and cardiovascular risk factors in men and women aged
18-79 years. J Otron Dis 1987; 40: 421-8.
0
Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E, Sjostrom
L. Distribution of adipose tissue and risk of cardiovascular
disease and death: a 12 year follow up of participants in the
population study of women in Gothenburg, Sweden. Br Med
J 1984; 289: 1257-61.
14
Larsson B, Svardsudd K, Welin L, Wilhelmsen L, Bjomtorp P,
Tibblin G. Abdominal adipose tissue distribution, obesity, and
risk of cardiovascular disease and death: 13 yearfollowup of
participants in the study of men born in 1913. Br Med J 1984;
288: 1401-4.
° Ducimetiere P, Richard J, Cambien F. The pattern of subcutaneous
fat distribution in middle-aged men and the risk of coronary
heart disease: The Paris Prospective Study. IntJ Obesity 1986;
W: 229-40.
16
Donahue R P, Abbott R D, Bloom E, Reed D M, Yano K. Central
obesity and coronary heart disease in men. Lancet 1987; 1:
821-4.
17
Kaye S A, Folsom A R, Kushi L H, Prineas R J, Seal U S, Wallace
R B. Relationship of body fat distribution to self-reported
prevalence of diabetes mellitus and fasting concentrations of
glucose and insulin in post-menopausal women. Submitted to
J Epidemiol Community Health, 1989.
B
Kushi L H, Kaye S A, Rjlsom A R, Soler J T, Prineas R J. Accuracy
and reliability of self-measurement of body girths. Am J
Epidemiol 1988; 128: 740-a
S
SAS User's Guide: Statistics. Cary, North Carolina: SAS Institute,
Inc, 1985.
20
Berkman L F. Measures of social networks and social support:
Evidence and measurement In: OttfeU A M, Eater E D (eds).
Measuring Psychosodal \briables in Epidemiologic Studies of
Cardiovascular Disease. USDHHS (NTH) Pub No 8 5 - 2 2 7 0 ,
Washington, DC: US Government Printing Office, 1985.
21
Paffenbarger R S, Hyde R T. Exercise in the prevention of coronary
heart disease. Prev Med 1984; 13: 3 - 2 2 .
22
Pirie P, Jacobs D, Jeffery R, Hannan P. Distortion in self-reported
height and weight data. J Am Diet Assoc 1981; 7 8 : 6 0 1 - 6 .
a
Palta M , Prineas R J, Berman R, Hannan P. Comparison of selfreported and measured height and weight. Am J Epidemiol 1982;
115: 223-30.
24
Shaper A G, Cook D G, Walker M, MacFarlane P W. Recall of
diagnosis by men with ischemic heart disrasc. Br Heart 11984;
51: 606-11.
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
Despite the limitations of this study, our findings have
biological plausibility. Increased abdominal adiposity is
associated with an unfavourable cardiovascular risk
profile — greater hypertension7"12 and greater blood
concentrations of glucose, insulin, and atherogenic
lipids. ct726 The hyperinsulinaemia associated with
increased WHR might contribute to high blood
pressures.27 Moreover, abdominal adipocytes have
higher lipolytic activity than femoral adipocytes28 and
greater abdominal adiposity would increase portal
delivery of fatty acids.14
The inclusion of terms for hypertension and diabetes
in the final multivariate model for heart attack clearly
understates the effect of BMI and probably understates
the WHR effect as well, in that the relationship of these
variables to coronary disease may be mediated by blood
pressure or glucose. The fact that the predicted heart
attack rate ratio was 1.5 for a one SD change in WHR
in the model with only age, BMI, and WHR, versus 1.2
for the final model, illustrates that this may be the case.
Future studies of obesity as a risk factor for
hypertension and heart disease should consider fat distribution to help elucidate possible causal mechanisms. Fat
distribution probably is genetically influenced29 and
therefore modifiable only to a degree. However, those
interested in disease prevention might use WHR as a
means of identifying individuals at high risk of
cardiovascular disease.
BODY EAT DISTRIBUTION AND CARDIOVASCULAR DISEASE
23
367
a
National Centers for Health Statistics. Net Differences in Interview
Ostman J, Amer P, Engfeldt P, Kager L. Regional differences in the
Data on Chronic Conditions and Information Derived
from
control of liporysis in human adipose tissue. Metabolism 1979;
Medical Records. Vital and Health Statistics. Series 2, No 57.
28: 1198-1203.
Washington, US Government Printing Office, 1973.
»Bouchard C Inheritence of fat distribution and adipose tissue
26
Soter J T, Folsom A R, Kushi L H, Prineas R J, Seal U S.
metabolism. In: Metabolic Complications of Human Obesities,
Association of body fat distribution with plasma lipids,
j •Nfcgue, et al, eds. Amsterdam, Hsvier, 1985.
lipoproteins, apolipoproteins AI and B in Postmenopausal
Wamen. / Clin Epidemiol 1988; « : 1075-81.
27
DeFronzo R A . The effect of insulin on renal sodium metabolism:
A review with clinical implications. Diabetologia 1981: 21:
165-71.
(Revised version received May 1988)
Downloaded from http://ije.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016